Dg As Ancillary Service

Ancillary Service Provision from Distributed Generation CONTRACT NUMBER: DG/CG/00030/00/00 URN NUMBER: 04/1738

The DTI drives our ambition of ‘prosperity for all’ by working to create the best environment for business success in the UK. We help people and companies become more productive by promoting enterprise, innovation and creativity. We champion UK business at home and abroad. We invest heavily in world-class science and technology. We protect the rights of working people and consumers. And we stand up for fair and open markets in the UK, Europe and the world.

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION CONTRACT NUMBER DG/CG/00030/00/00 URN NUMBER 04/1738

This work was commissioned and managed by the DTI's Distributed Generation Programme in support of the Technical Steering Group (TSG) of the Distributed Generation Co-ordinating Group (DGCG). The DGCG is jointly chaired by DTI and Ofgem, and further information can be found at www.distributed-generation.gov.uk

Contractor

Ilex Energy Consulting with the Manchester Centre for Electrical Energy, UMIST The work described in this report was carried out under contract as part of the DTI Technology Programme: New and Renewable Energy, which is managed by Future Energy Solutions. The views and judgements expressed in this report are those of the contractor and do not necessarily reflect those of the DTI or Future Energy Solutions.

First published 2004  Crown Copyright 2004

Project Steering Group Summary Introduction and Background: In support of DGCG TSG, Workstream 5, the Department of Trade and Industry’s New and Renewable Energy Programme commissioned Ilex Energy Consulting Ltd. to undertake an investigation and report on the potential for an Ancillary Service market to be developed within the distribution network. The work was to be carried out from a distributed generation (DG) prospective. The final report is titled ‘Ancillary Service Provision from Distributed Generation’. The investigation was undertaken during the second and third quarter of 2004 and was subject to review and commentary from the TSG WS5 P06 Project Manager and the P06 Project Steering Group whilst in progress. The Project Steering Group was made up of a diversity of people, including Distribution Network Operators (DNO’s), NGT, generators and consultants. The final report was submitted to Future Energy Solutions (FES) in September 2004. FES are the DTI’s New & Renewable Energy Programme’s managing contractor. Objectives: The aims and objectives of the study were to investigate the potential for creating ancillary service markets at the distribution level in Great Britain. Specifically the study sought to: Investigate any existing arrangements for distribution level Ancillary Service markets worldwide. Review the high level options for the design of ancillary service markets and identify any regulatory and legislative changes that might be required. Examine the prospects and opportunities for the different forms of distributed generation and assess whether the creation of different services would incentivise generation to connect to the distribution network. Investigate the commercial framework and technical procedures that might be required. Explore the infrastructure requirements. Assess the impact on different market participants. The scope of the project included a consideration of the opportunities for DG to contribute to existing Transmission System Operator (TSO) ancillary services and an investigation of the potential for DG to contribute to new DNO services that could develop in the short to medium term. Findings/Conclusion: This work area is relatively new and without a market presently at the DNO level it can be difficult to form concrete conclusions. Generally however, the report successfully developed on existing work carried out in the area of distribution level ancillary services and highlighted a number of key areas of interest. Some of these require further consideration as the subject area moves forward. The report was comprehensive and the important conclusions were:

o

o

o

o o o

o o

o

A value based assessment approach was used to determine the attractiveness of each ancillary service. This concluded that the value of the most feasible ancillary services will be relatively low and would represent incremental revenue opportunities for DG. Investing in DG on the basis of ancillary service income only, is therefore unlikely. Only TSO frequency response, TSO regulating and standing reserve and DNO security of supply contributions represent realistic opportunities for DG in the short to medium term. The extent of opportunities for DNO security services will largely relate to load growth and asset replacement profiles. TSO reactive power, DNO quality of supply and DNO voltage and power flow managements services were deemed to have little potential over the same period. CCGT and DFIG wind generators were the most promising technologies for the provision of TSO frequency services, whereas CCGT’s, diesel standby generators and perhaps micro CHP were best placed to provide reserve services. DNO security of supply services could be provided to a varying degree by most existing DG types. The majority of existing DG has not been installed with the necessary infrastructure to provide ancillary services. To extend aggregation opportunities for DG to participate in the standing reserve market (limited by cost of infrastructure), new low cost communication and monitoring arrangements should be evaluated. The most appropriate commercial arrangements for response and reserve services appear to be market-based mechanisms, potentially extending some of the TSO’s current arrangements. For DNO security of supply services, the most appropriate arrangement appears to be bilateral contracts. Registered Power Zones (RPZs) could provide an initial platform for the development of some ancillary service contracts. In the future, with increased levels of DG, the opportunities for ancillary services should increase. However, as an active network develops, there is also likely to be network constraints and delivery uncertainty of some ancillary services may increase. This is a major concern for DNOs as they are exposed to the risk of non-delivery. The impact could be financial, regulatory or legal in nature and requires further exploration. The majority of network security is secured through capital expenditure at present rather than operating expenditure. This relates to the current regulatory framework and thus further work is necessary to establish a funding mechanism for network security and support.

The study did not cover the longer term potential (beyond 2015) for ancillary services that were deemed to have little or no potential in the short to medium term. However, due to the relative low levels of DG penetration at present and the absence of a distribution level ancillary service market (accepting that some bilateral arrangements do exist and that DG can participate already at the transmission level), it is accepted by the project manager that further work in this area will probably not be feasible until DG penetration increases and the future structure of the power system is better understood. Next Steps: As already touched on, significant further work on the WS5 P06 area is probably a little premature until DG penetration increases and the active nature of the future power system becomes more established.

Bilateral Agreements are likely to continue to be used in any developing ancillary service market in the short to medium term and if further studies are to be carried out at this stage, the focus should be on developing commercial frameworks/agreements. The Distribution Commercial Forum appears well placed to develop the necessary commercial arrangements and RPZs could represent ideal test-beds for such services. Nigel Turvey Project Manager, TSG WS5 P06 11th October 2004

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Disclaimer While ILEX considers that the information and opinions given in this work are sound, all parties must rely upon their own skill and judgement when making use of it. ILEX does not make any representation or warranty, expressed or implied, as to the accuracy or completeness of the information contained in this report and assumes no responsibility for the accuracy or completeness of such information. ILEX will not assume any liability to anyone for any loss or damage arising out of the provision of this report.

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

CONTENTS EXECUTIVE SUMMARY

I

1. INTRODUCTION

1

2. BACKGROUND REVIEW

6

3. THE SCOPE FOR NEW ANCILLARY SERVICES AT THE DISTRIBUTION LEVEL

23

4. EXPLORATION OF ANCILLARY SERVICES FROM DISTRIBUTED GENERATION

32

5. PROSPECTS FOR DISTRIBUTED GENERATION

75

6. COMMERCIAL & TECHNICAL FRAMEWORK

104

7. IMPACT ASSESSMENT

128

8. CONCLUSIONS AND RECOMMENDATIONS

139

ANNEX A – TABLE 1 OF ER P2/5

144

ANNEX B – CONTRIBUTORS TO THE STUDY

146

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

TABLES TABLE 1 − SUMMARY OF RENEWABLE TECHNOLOGY CAPABILITIES

V

TABLE 2 − SUMMARY OF NON-RENEWABLE TECHNOLOGY CAPABILITIES

V

TABLE 3 − SUMMARY OF THE MAIN ANCILLARY SERVICES PROCURED BY NGC IN E&W

8

TABLE 4 − NGC PAYMENTS FOR STANDING RESERVE IN 2003/2004

14

TABLE 5 − ESTIMATED ADDITIONAL REQUIREMENTS FOR CONTINUOUS FREQUENCY RESPONSE

29

TABLE 6 − COST AND MARKET-EXTRAPOLATION BASED ESTIMATES OF THE VALUE ADDITIONAL CONTINUOUS RESPONSE

30

TABLE 7 − ADDITIONAL RESERVE REQUIRED FOR VARIOUS LEVELS OF WIND PENETRATION

34

TABLE 8 − ESTIMATES OF NET VOLUME OF ADDITIONAL BID OFFER ACCEPTANCES

34

TABLE 9 − MARKET VALUE OF ADDITIONAL RESERVE

35

TABLE 10 − DISTRIBUTION OF LOAD ACROSS VOLTAGE LEVELS

39

TABLE 11 − THE TOTAL DNO REACTIVE IMPORT [MVAR] FROM THE TRANSMISSION NETWORK IN MVAR AT PEAK FOR DIFFERENT LEVEL PENETRATION OF DG AND VARIOUS POWER FACTORS (P.F.)

40

TABLE 12 − MAXIMUM DISTANCES OF THE OHL AND UC CIRCUITS

61

TABLE 13 − ACTIVE POWER (AS A % OF THE CIRCUIT CAPACITY) THAT CAN BE TRANSPORTED

61

TABLE 14 − ADDITIONAL PLANT CAPACITY REQUIRED TO TRANSPORT REACTIVE POWER

62

TABLE 15 − RECOMMENDED VALUES FOR MINIMUM PERSISTENCE TIMES (TM) FOR SWITCHING ACTIONS

68

TABLE 16 − SUMMARY OF RENEWABLE TECHNOLOGY CAPABILITIES

87

TABLE 17 − SUMMARY OF NON-RENEWABLE TECHNOLOGY CAPABILITIES

87

TABLE 18 - F FACTORS OF IDENTICAL NON-INTERMITTENT UNITS OF 86% AVAILABILITY WITH THE PROBABILITIES OF DELIVERING THESE CONTRIBUTIONS

104

TABLE 19 − COMPARISON OF COSTS OF DELIVERING ADDITIONAL SECURITY CONTRIBUTIONS

112

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

FIGURES FIGURE 1 − CONTINUOUS AND OCCASIONAL FREQUENCY RESPONSE SERVICES

8

FIGURE 2 − GENERIC DISTRIBUTION MODEL WITH DG CONNECTED AT 11KV AND 33KV

39

FIGURE 3 − EXAMPLE OF ER P2/5 COMPLIANCE WITHOUT GENERATION CONTRIBUTIONS

43

FIGURE 4 − POSSIBLE SOLUTIONS TO A NETWORK SECURITY SHORTFALL UNDER P2/5 & P2/6

44

FIGURE 5 − DISTRIBUTIONS OF CIS AND CMLS ACCORDING TO DISTRIBUTION VOLTAGE LEVEL

47

FIGURE 6 − (A) TRADITIONAL NETWORK LAYOUT (B) POTENTIAL FOR DG TO PROVIDE NETWORK SERVICES WILL DEPEND ON POINT OF CONNECTION

48

FIGURE 7 − COMPARISON OF CML PERFORMANCE FOR TWO NETWORKS

49

FIGURE 8 − USE OF DG FOR VOLTAGE SUPPORT AT THE END OF THE FEEDER DUE TO LOAD GROWTH

54

FIGURE 9 − USE OF VOLTAGE REGULATOR TO MITIGATE VOLTAGE VIOLATION AT THE END OF THE FEEDER DUE TO LOAD GROWTH

55

FIGURE 10 − USE OF CAPACITIVE COMPENSATION TO SUPPORT VOLTAGE

55

FIGURE 11 − USE OF DG TO SUPPORT VOLTAGE AT PEAK TIMES

56

FIGURE 12 − OPTIONS FOR SOLVING VOLTAGE DIP PROBLEM AT POINT A (ASSUMING OUTAGE SECTION BETWEEN SUBSTATION 1 AND A FOLLOWED BY CLOSURE OF NORMALLY OPEN POINT)

57

FIGURE 13 − USE OF DG TO MITIGATE PARTIAL CIRCUIT OVERLOAD DUE TO LOAD GROWTH

58

FIGURE 14 − USE OF DG TO MITIGATE CIRCUIT OVERLOAD FOLLOWING CLOSURE OF NOP

59

FIGURE 15 − SCHEMATIC DIAGRAM OF THE POWER SYSTEM WITH VARIOUS FORMS OF DG TECHNOLOGIES CONNECTED TO DISTRIBUTION NETWORKS

63

FIGURE 16 − TYPICAL SCHEMATIC FOR A LARGE SCALE CHP INSTALLATION

74

FIGURE 17 − COMMUNICATION PROCESS FROM TSO TO MICRO CHP VIA THE AGGREGATOR

80

FIGURE 18 − RESPONSE SCHEDULING OF GROUPS OF MICRO CHP UNITS

83

FIGURE 19 − CONTRACTING OPTIONS FOR AGGREGATION

100

FIGURE 20 − OPTIONS TO ADDRESS A 5MW SECURITY SHORTFALL AT A 33/11 KV, 20 MVA SUBSTATION

111

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

GLOSSARY OF TERMS ADSL

Asymmetric Digital Subscriber

MV

Medium Voltage

AVR

Automatic Voltage Regulator

MVA

Mega Volt-Ampere

BETTA

British Electricity Trading & Transmission Arrangements

MW

Megawatt

BOA

Bid/Offer Acceptance

NETA

New Electricity Trading Arrangements

BM

Balancing Mechanism

NGC

National Grid Company

BMU

Balancing Mechanism Unit

OCGT

Open Cycle Gas Turbine

BSC

Balancing & Settlement Code

OFGEM

Office of Gas & Electricity Markets

CAPEX

Capital Expenditure

OHL

Overhead Line

CCGT

Combined Cycle Gas Turbine

O&M

Operation and Maintenance

CHP

Combined Heat & Power

OPEX

Operational Expenditure

CI

Customer Interruptions

OS

Overall Standards

CML

Customer Minutes Lost

PN

Physical Notification

CUSC

Connection & Use of System Code

PSTN

Public Switched Telephone Network

DFIG

Doubly Fed Induction

PTO

Public Telephone Operators

DG

Distributed Generation

PV

Photo Voltaic

DGCG

Distributed Generation Coordination Group

ROC

Renewables Obligation Certificate

DNO

Distribution Network Operator

RPZ

Registered Power Zone

DTI

Department of Trade &

SMS

Short Message Service

DUoS

Distribution Use of System

SVC

Static Voltage Compensator

GB

Great Britain

SYS

Seven Year Statement

GS

Guaranteed Standard

TNUoS

Transmission Network Use of System

GSM

Global System for Mobiles

TSG

DGCG Technical Steering

IIP

Information & Incentives

TSO

Transmission System

KV

Kilovolt

UC

Underground Cable

KW

Kilowatt

UMTS

Universal Mobile Telecommunications System

LAN

Local Area Network

UoSA

Use of System Agreement

LV

Low Voltage

VAr

Volt-Ampere reactive

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

EXECUTIVE SUMMARY This report describes a joint study undertaken as part of the DTI’s New and Renewable Energy Programme by ILEX Energy Consulting and the University of Manchester Institute of Science and Technology (UMIST) with contributions sourced from many interested stakeholders. The aims and objectives of the study were to investigate the potential for creating ancillary service markets at the distribution level in Great Britain (GB). Specifically, the study sought to: •

investigate any existing arrangements for distribution level ancillary services markets worldwide;

•

review the high level options for the design of ancillary service markets;

•

examine the prospects and opportunities for the different forms of Distributed Generation (DG);

•

investigate the commercial frameworks and technical procedures that might be required;

•

explore the infrastructure requirements; and

•

assess the impact on different market participants.

The scope of this project was to: •

consider the opportunities for DG to contribute to existing Transmission System Operator (TSO) ancillary services; and,

•

investigate the potential for DG to contribute to new Distribution Network Operator (DNO) services that could develop in the short to medium term.

The study investigated the potential for distribution level ancillary services to be provided by generators, in-line with the anticipated increase in electricity generation from distributed resources. Whilst renewable electricity generation connected to distribution networks represents a key component of Government energy policy and targets, this study has sought to evaluate the distribution ancillary service market opportunities applicable to both renewable and non-renewable forms of distributed generation. A pre-requisite for the detailed development of operational and commercial models was that any new services should be financially material to the distributed generator whilst remaining economically and operationally attractive to network operators. Consequently, value based approaches were adopted for each ancillary service with i

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION a view towards improving the attractiveness of distributed and renewable generation projects. The services for which potential arrangements have been explored are: •

TSO Frequency Response;

•

TSO Regulating and Standing Reserve;

•

TSO Reactive Power;

•

DNO Security of Supply contributions;

•

DNO Quality of Supply Services; and

•

DNO Voltage and Power Flow Management Services.

TSO Frequency Response Frequency Response services are required by the TSO to maintain the system frequency within statutory tolerances. Frequency control is achieved through the real-time matching of supply to demand. Distribution connected Combined Cycle Gas Turbine (CCGT) plant already provide this service to TSOs. A key feature of TSO frequency response provision is the requirement for generators to be part-loaded. It is unlikely that TSO frequency response services will be provided regularly by renewable generation, as the opportunity cost of operating part-loaded will be relatively high. This is because the compensation for part-loading would not only need to recover the cost of reduced energy revenues but also the costs associated with the loss of Renewables Obligation Certificate (ROC)1 revenue. It is therefore unlikely that renewable generation will be able to compete effectively in frequency response markets. Although mandatory frequency response capabilities may become a technical requirement for large distribution connected wind farms, thereby ‘resolving’ any infrastructure constraints, the extent to which the TSO will utilise such capabilities is likely to be very limited. The value of TSO Frequency Response is estimated to vary between £0.40/kW per annum for wind generation and £2.50/kW per annum for CCGT technology (excluding holding costs).

1

The Renewables Obligation Order 2002, Statutory Instrument No. 914, 2002 ii

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

TSO Regulating and Standing Reserve Reserve energy is required to provide rapid access to generation, to accommodate errors in demand forecasting, to provide contingency arrangements for generation failures and to restore frequency response capabilities. The key differences between frequency response and reserve services relate to delivery timescales. Typically, reserve services are manually initiated and involve longer lead times. A consequence of simplified service initiation procedures is a reduction in the sophistication of control requirements, thus making reserve more attractive to smaller providers. It is unlikely that synchronised reserve will be provided by renewable generation, as the compensation for part-load operation would also need to recover the loss of ROC revenue. Non-renewable distributed generation already provides standing reserve services to the TSO at a value of approximately £7/kW per annum. Increased DG participation could be facilitated by expanded aggregation services.

TSO Reactive Power TSO reactive power can be sourced from distributed generators, especially those connected at 132 kV, for transmission system voltage regulation. Reactive power sourced at lower distribution voltages will reduce the reactive power required from transmission-connected generation (at peak loads). DG connected at lower voltage levels can make a significant impact on the amount of reactive power exchanged between TSO and DNO systems. A simple generic model was developed to illustrate DNO reactive power import reductions at different levels of DG penetration. The value of DG derived reductions in DNO reactive imports was estimated to be approximately £1.20/kW per annum. The impact of DG on TSO reactive power market will be driven by many different variables. More work would be required to determine the impact of DG on DNO reactive requirements. The impact of reactive power management on the transport capabilities of distribution circuits was also investigated. DG connected close to loads could extend the transport capabilities of existing circuits. The value of this service would be limited by the low cost of power factor compensation equipment. It is unlikely that this would represent significant income for DG. High DG availabilities would be needed for DNOs to consider such services.

iii

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

DNO Security of Supply Contributions The proposed planning recommendation ER P2/6 could broaden opportunities for DNOs to consider contributions to network security from DG. However, as DNO networks should currently comply with ER P2/5, the requirement for security contributions from DG may be limited in the short-term. In the medium to long term, load growth and asset replacements could increase opportunities for DG to provide network support services. The value of security provided by non-intermittent DG can be related to the avoided or deferred costs of network reinforcement. DG can also substitute for network automation facilities. This is particularly relevant when considering security contribution of intermittent generation such as wind. A number of examples were used to illustrate the potential value of network security services. For non-intermittent generators, values in the range of £1/kW to £12/kW per annum were derived, depending upon the complexity of the network solution avoided. It is anticipated that most reinforcements would be at the lower end of this range. Because of the drive to reduce Customer Interruptions (CI) and Customer Minutes Lost (CML), DNOs have made considerable investments in 11kV and 0.4kV networks. A result of this investment is that distribution networks in GB are generally “over compliant” with planning and security standards. For the foreseeable future, the scope for DG to provide security services at these voltages could be limited.

DNO Quality of Supply Services In the future, there could be opportunities for DG to improve service quality on 11kV and 0.4 kV networks, given the contribution of such networks to Quality of Supply statistics. In order for DG to improve service quality on such networks, the generation must also be connected at 11kV or 0.4kV, thus restricting opportunities to relatively small sized generation. A key requirement for DG, to reduce the impact of outages, is islanded operating capability. Analysis suggests that the annual benefit of islanding operation was approximately £1.40/kW/annum and £19/kW/annum for residential and commercial customers respectively. Due to the complexity of islanding, it is unlikely that DG will be able to significantly reduce CIs and CMLs in the short or medium term. It is hoped that this work will provide input to another DTI project, currently evaluating the feasibility of island operation.

iv

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

DNO Voltage and Power Flow Management services Our analysis revealed that voltage control and flow management problems are essentially network planning related issues as they relate to supply restoration times following network failures (ER P2/5 or P2/6). Because of the relatively low availability of DG compared to network components and the UK’s deterministic voltage standards, opportunities may be limited for DG to provide voltage support or overload reduction. Generally, non-intermittent DG would be suitable for such applications. Inverter connected renewable generation, such as Doubly Fed Induction Generators (DFIG) or Photo Voltaics (PV), represent an exception, as reactive power is generally independent of active power output. Opportunities to provide voltage and power flow management services will improve with increased penetrations of DG due to the higher collective availability. The value of these services was estimated to be the order £1.50/kW/annum.

Ancillary Service Capabilities of Different Generation Technologies The ancillary service capabilities of renewable and non-renewable technologies are summarised in Table 1 and Table 2.

v

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Table 1 − Summary of renewable technology capabilities DG DG Technology TechnologyType Type Ancillary Ancillary Service Service

Wind Wind non-DFIG non-DFIG

Wind Wind DFIG* DFIG*

Size Size

<<50 50MW MW

>50 >50MW MW

Frequency Frequency

HF HFonly only

Reserve Reserve

Possible Possible

Biomass Biomass

Possible Possible

Land LandFill Fill Gas Gas

Solar SolarPV PV

Hydro Hydro

<100 <100kW kW

>>1MW 1MW

Possible Possible

Possible Possible

1–100MW 1–100MW 11––10MW 10MW HF HFonly only

HF HFonly only

Possible Possible

Possible Possible

Reactive Reactive Network Network Support Support

Limited Limited Future Future islanding? islanding?

Black BlackStart Start

Future Future islanding? islanding?

Future Future islanding? islanding?

* Wind Farms <50 MW may employ DFIG machines in future Table 2 − Summary of non-renewable technology capabilities DG DG Technology TechnologyType Type Ancillary Ancillary Service Service

CCGT CCGT

Large LargeCHP CHP

Micro MicroCHP CHP

Diesel Diesel&& OCGT OCGT (Standby) (Standby)

Size Size

>100 >100MW MW

1-100 1-100MW MW

11––55kW kW

<<50 50MW MW

Frequency Frequency

Limited Limited

Reserve Reserve

Possible Possible

Limited Limited Possible: Possible:High High penetrations penetrations

Reactive Reactive Network Network Support Support Black BlackStart Start

Possible: Possible:High High penetrations penetrations Possible Possible

Future Futureisland island opportunity? opportunity?

Future Futureisland island opportunity? opportunity?

Whilst all of the above services were explored in detail, only TSO Frequency Response, TSO Regulating and Standing reserve and DNO Security of Supply contributions represent realistic opportunities for distributed generators in the short or medium term. Combined Cycle Gas Turbines (CCGT) and DFIG wind farms were the most promising technologies for the provision of TSO Frequency Response services whereas CCGTs, diesel standby generators and perhaps micro CHP were best placed to provide reserve services. It was found that, to varying degrees, DNO Security of Supply services could be provided by most existing distributed generation technologies. As the majority of existing DG has been installed for electricity supply purposes, very few generators are equipped with the infrastructure necessary to provide ancillary services. Such infrastructure includes governors, automatic voltage regulators, resynchronisation facilities, appropriate protection, monitoring and communication facilities.

vi

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Commercial, Technical and Policy Implications The most appropriate commercial arrangements for response and reserve services appear to be market-based mechanisms. Ideally the TSO’s current arrangements could be extended. Expanded aggregation arrangements, utilising lower cost infrastructure, would facilitate increased participation from small generators. The most appropriate commercial arrangements for DNO Security of Supply services appear to be bilateral contracts due to the local and site-specific nature of security requirements. Opportunities for DG to provide ancillary services will undoubtedly increase as DG penetrations and availabilities increase. The analysis undertaken suggests that the value of the most feasible ancillary services will be relatively low. Consequently, such services will represent incremental revenue opportunities for DG. In general, it would not be possible to develop business cases for investing in DG solely on the basis of ancillary service income. Niche opportunities will emerge for DG to provide ancillary services, usually in circumstances where constraints restrict network development, e.g. environmental, planning and terrain related constraints. In an era with significantly increased levels of DG operating on active distribution networks, the opportunities for DG to provide ancillary services may increase. However, on active networks there is an increased likelihood that due to distribution network constraints, certain modes of operation may not be permitted by the DNO. Consequently there could be increased delivery uncertainty regarding the provision of TSO ancillary services from distributed generators connected to active networks. In circumstances where a distributed generator receives conflicting instructions regarding the provision of different ancillary services, local services should take precedence over national services. Higher penetrations of DG will increase DNO options regarding network operation and development decisions, which could (in certain situations) lead to lower overall costs. Increased penetration of DG could also enhance competition in TSO markets for frequency response and reserve. This could be particularly relevant should demand for these services increase with intermittent generation.

vii

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION As the output from distributed generation is largely purchased by suppliers and settled through supplier demand accounts within NETA, suppliers must ensure they are aware of generator operating regimes and also whether generator operation is likely to be influenced by ancillary service provision. Supplier concerns will relate to imbalance exposures in the Balancing Mechanism (BM) and the fulfilment of ROC targets. Suppliers will require notification of ancillary service provision, in order to suitably revise demand forecasts. Ancillary service instructions issued post Gate Closure will inevitably impact upon a supplier’s imbalance exposure and potentially reduce the value of the energy supplied. The introduction of Registered Power Zones (RPZs), created to encourage DNOs to develop and demonstrate new, more cost effective ways of connecting and operating generation, could provide an initial platform for the development of appropriate ancillary services contracts. In order for new technical solutions to become widely accepted, an appropriate contractual framework will need to be established, and RPZs could be used to develop such arrangements. It is important to stress that provision of ancillary services from DG should not jeopardise or degrade security of supply and may even contribute to its enhancement in future. It should be recognised that the provision of ancillary services from DG should not impact negatively upon DG contributions towards Government climate change targets. In addition, the provision of ancillary services from distributed generation could, in niche situations, avoid any negative impact of network investment in environmentally sensitive areas. Although this work explored ancillary services opportunities for DG in the short to medium term, long-term approaches also need to be investigated given the fundamental changes to the structure of power systems that may emerge in future, particularly in the context of the Technical Architecture initiative.

Recommendations The extent of opportunities on DNO networks will largely relate to load growth and asset replacement profiles. Whilst it has not been possible to quantify the relative magnitudes of these opportunities within this project, such information will be critical to evaluating service materiality under alternative future development scenarios and should be explored further.

viii

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION In order that a consistent and transparent set of arrangements can emerge to facilitate increased network security contributions from DG, it will be necessary to establish principles (and potentially standardise commercial arrangements) for procurement processes and valuation methodologies. Indeed, the Distribution Commercial Forum might consider the development of a standardised ‘model form’ contract suitable for localisation by individual DNOs. A major concern of DNOs regarding the reliance on distributed generators to provide network support services will relate to nondelivery risk exposures. Such exposures could be financial, regulatory or legal in nature. Consequently, the issues associated with service non-delivery require further exploration. Whist the current aggregation arrangements have been successful in encouraging non-BM participants into the standing reserve market, the costs of the associated infrastructure could deter wider participation. In order to extend aggregation opportunities further, new low-cost communication and monitoring arrangements should be evaluated. A potential problem for DNOs relates to the current regulatory framework in terms of capital expenditure (CAPEX) and operational expenditure (OPEX) funding distinctions. At present, network security is procured through CAPEX, which does not currently accommodate generation contributions. The current arrangements for OPEX are not ideal as DNOs could be financially penalised for funding ancillary services through this route. More work will be required to establish a suitable funding mechanism for network security and support.

ix

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

1. 1.1

INTRODUCTION

This report describes a joint study undertaken as part of the DTI’s New and Renewable Energy Programme by ILEX Energy Consulting and the University of Manchester Institute of Science and Technology (UMIST) with contributions sourced from many interested stakeholders including: •

DNOs

•

TSO

•

Renewable and non-renewable plant developers and operators

•

Micro-CHP developers

•

Academics

•

Ofgem

Thanks are offered to all of the individuals, outlined in Annex B, who provided input to this work. 1.2

The study investigated the potential for distribution level ancillary services to be provided by generators, in-line with the anticipated increase in electricity generation from distributed resources.

1.3

Whilst renewable electricity generation connected to distribution networks represents a key component of Government energy policy and targets, this study has sought to evaluate the distribution ancillary service market opportunities applicable to both renewable and non-renewable forms of DG.

Aims and Objectives 1.4

The aims and objectives of the study were to: •

1.5

Investigate the potential for creating ancillary service markets at the distribution level in Great Britain (GB).

Specifically, the study sought to: •

investigate any existing arrangements for distribution level Ancillary Services markets worldwide;

•

review the high level options for the design of ancillary service markets;

•

examine the prospects and opportunities for the different forms of distributed generation;

•

investigate the commercial framework and technical procedures that might be required; 1

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION •

explore the infrastructure requirements; and

•

assess the impact on different market participants.

With the overall intention of revealing the costs, benefits, challenges and opportunities of establishing ancillary services market(s) at a distribution level. 1.6

1.7

The scope of this project is to: •

consider the opportunities for DG to contribute to existing TSO ancillary services; and

•

investigate the potential for DG to contribute to new DNO services that are likely to develop in the short to medium term.

In the context of the recent Technical Architecture initiative, it should be noted that these opportunities and the technical/commercial frameworks would change if there were fundamental changes to the structure of power systems in the long term.

Methodology and approach 1.8

In the UK and in other international electricity markets, the term ‘Ancillary Services’ is widely used to describe many of the activities required to maintain the stability of the transmission system. In many liberalised energy markets, TSOs usually procure such services from generators, although there is increasing scope for demand-side participation. TSO procured ancillary services often include frequency response, reactive power and a variety of energy reserve services.

1.9

From the outset, this study has adopted a broad definition for ancillary services at a distribution level, which, whilst remaining primarily generation focused, could encompass a wide range of new and existing services. This approach deliberately avoided any constraints arising from a strict interpretation of the term as applied in the UK transmission context.

1.10

A pre-requisite for the detailed development of operational and commercial models was that any new services should be financially material to the distributed generator whilst remaining economically and operationally attractive to network operators. Consequently, value based approaches were adopted for each ancillary service with a view towards improving the attractiveness of distributed and renewable generation projects.

Consultation and collaboration 1.11

In undertaking this study, we have benefited from the assistance of a large number of individuals from within the electricity supply

2

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION industry. We are also grateful for the assistance received in undertaking this work, particularly from EDF Energy. 1.12

Throughout, the project consulted to the Technical Steering Group (TSG) of the Distributed Generation Co-ordination Group (DGCG) via a dedicated workstream comprising representatives from Distributions Network Operators (DNOs), distributed generators, Future Energy Solutions, the DTI and Ofgem.

1.13

The approach adopted reflects the consensus of workstream members who were regularly updated regarding study findings and direction, although the results presented in this report may not necessarily reflect the opinions of individual workstream members or the companies they represent.

Outline of the report 1.14

In Section 2 we present a background review describing the state of development of distribution ancillary services in liberalised electricity markets accommodating relatively high penetrations of distributed generation. This section extensively reviews the ancillary services managed at a transmission level by NGC in Great Britain.

1.15

Section 3 outlines the new ancillary service opportunities, which could be provided from Distributed Generation (DG) to DNOs. This Section also contrasts the common characteristics and key differences between TSO and DNO ancillary services.

1.16

Section 4 provides a detailed description of the features, characteristics, constraints and delivery issues associated with the provision of a range of existing and new ancillary services from distributed generation. Also within this section, the distinctions between local and national services are explored.

1.17

Section 5 sets out the prospects for distributed generation in terms of the capabilities of different technologies. It also contains a summary of distributed generation capabilities in terms of technology type, ratings, infrastructure requirements to provide ancillary services and renewable considerations. An important output from this study relates to the different technical capabilities of each type of distributed generation.

1.18

In Section 6, the technical and commercial frameworks for the most attractive distribution level ancillary services are expanded to describe the commercial interfaces and trading options for the different market participants. The section also discusses interactions between transmission and distribution level ancillary services. This section also provides a high level analysis of the interaction between

3

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION ancillary service provision and industry codes and agreements. The section outlines the potential procurement mechanisms and commercial arrangements for different services. 1.19

Section 7 describes the impacts on the different market participants in terms of costs, benefits, operational complexity and regulatory issues to provide an overall assessment of the feasibility and attractiveness of each service to different providers. Key to this section are the valuation examples.

1.20

Section 8 provides the conclusions and recommendations arising from the study and also outlines relevant areas for further analysis.

4

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

[This page is intentionally blank]

5

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

2. 2.1

BACKGROUND REVIEW

Part of the remit of this study was to determine the extent to which ancillary services are already being provided by distribution system connected generation, both within Great Britain (GB) and internationally. This requirement was addressed through: •

Liaison with DNOs forming part of wider international groups operating distribution assets in Europe and North America, e.g. EdF, E.ON, RWE, Western Power Distribution and Scottish Power.

•

A review of the academic and commercial literature available through a desk study concentrating on liberalised energy markets, especially those with high penetrations of renewable and distributed generation.

•

Detailed analysis of the transmission-level ancillary services market as managed and operated by the National Grid Company in England and Wales.

GB Experience 2.2

Within GB it has been confirmed that there are already a number of distributed generators providing ancillary services to the TSOs. These services predominantly relate to frequency response from large-scale conventional plant, such as 132 kV connected Combined Cycle Gas Turbine (CCGT) generators, and Standing Reserve from standby generators embedded within industrial customer premises.

2.3

Experience regarding the provision of ancillary services to DNOs, as opposed to the TSO, was found to be limited although niche applications were identified on remote and islanded distribution networks, e.g. the Channel Islands and the Isles of Scilly.

2.4

One issue complicating the role of distributed generation in relation to the provision of ancillary services to TSOs, is the lack of a consistent definition as to what constitutes transmission and distribution connection voltages. In England and Wales, the transmission system is generally regarded as operating at 275 kV and 400 kV whereas in Scotland, the transmission system is deemed to operate at 132 kV and above.

2.5

Similarly, variations also exist between Grid Codes regarding the obligations on generators to provide ancillary services. In GB, these variations depend on the size of the generator relative to the overall size of the transmission system, rather than connection voltage.

6

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Under BETTA2, the GB TSO is seeking mandatory ancillary services capabilities from: •

>100 MW generators in England and Wales;

•

>30 MW generators in Scottish Power’s transmission system area; and

•

>5 MW generators in Scottish Hydro Electric’s transmission system area.

2.6

One implication of these different generator size thresholds may be that, in future, NGC will seek ancillary service capabilities from distribution-connected generators.

2.7

A feature of TSO ancillary service provision in England and Wales has been a gradual move away from centrally administered mandatory solutions towards more flexible, commercial arrangements based upon market mechanisms. Such approaches often utilise tender processes, have less prescriptive technical requirements and encourage wider participation from the demand-side and distributed generation.

International Experience 2.8

In international liberalised energy markets, the provision of frequency response and reserve services from distributed generators was found to be commonplace. Typically generators rated in tens of Megawatts (MW) would be required to provide such services by many European TSOs.

2.9

In markets with high penetrations of renewable generators, particularly those highly reliant upon intermittent energy resources, it is increasingly common for TSOs to seek frequency response capabilities from distributed generators. In Denmark, Germany and Ireland, the respective TSOs have established Grid Code obligations requiring wind farms to have frequency response capabilities to ensure that system stability can be maintained. Similar requirements are currently being negotiated in GB.

2.10

The provision of TSO related ancillary services from smaller scale generation (<10 MW) was found to be less widespread with TSOs seeming to prefer the procurement of such services from larger-scale conventional generators.

2

British Electricity Trading and Transmission Arrangements

7

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 2.11

As in GB, the provision of ancillary service from distributed generators to DNOs was less well developed than those for TSOs, although many of the utilities approached expressed interest in the findings of this study.

2.12

In North America, a number of supply quality related services have evolved in the form of Premium Power Parks where sensitive industrial loads can co-locate with distributed generators to reduce customer exposures to supply interruptions and brown-outs. Such initiatives have largely been a consequence of the less interconnected nature of North American distribution systems, with longer circuit lengths, which can result in a lower average quality of supply compared to that experienced on many European networks.

Transmission level ancillary services in England & Wales 2.13

Since privatisation, the requirements for transmission level ancillary services have continually been refined by NGC through regulatory incentives and wider industry developments such as NETA. The main ancillary services procured by NGC can be categorised as: •

Frequency Response;

•

Reserve Energy;

•

Reactive Power;

•

Fast Start; and,

•

Black Start;

2.14

In recent years, NGC has been increasingly encouraged to reduce the costs of system operation through regulatory incentives. NGC’s System Operator incentive scheme achieves such incentivisation by establishing an annual target for the costs of system operation (including ancillary services). Where the costs of system operation are lower than the agreed target, NGC is allowed to retain a proportion of the savings. Similarly, where the costs of system operation exceed the target, NGC is exposed to a proportion of any overspend. The target for the SO incentive scheme in 2004/2005 is £415 million including losses.

2.15

A summary of the main ancillary services procured by NGC is illustrated in Table 3.

Frequency Response 2.16

Frequency Response services are required by TSOs to maintain the system frequency within statutory tolerances. Frequency control is achieved through the real-time matching of supply to demand, i.e.

8

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION generation to consumption. Consequently, frequency regulation can be achieved by either adjusting generator outputs or by demand groups regulating their energy consumption patterns within short timescales in response to changes in system frequency. 2.17

Frequency falls below 50 Hz when demand is greater than generation and rises above 50 Hz when generation is greater than demand. For the system to operate satisfactorily, the frequency must be maintained continuously within narrow limits around 50Hz.

9

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Table 3 − Summary of the main ancillary services procured by NGC in E&W

Frequency is managed by a combination of a) continuous and b) occasional response services. These two services are illustrated in Figure 1 and described further below. Figure 1 − Continuous and occasional frequency response services F re q u e n c y R e s p o n s e F re q u e n c y C o n tro l C o n tin u o u s S e r v ic e 5 0 .2 Frequency (Hz)

2.18

5 0 .0

10 s

30 s

60 s

1 0 m in s

T im e

4 9 .8 P r im a r y

S e c o n d a ry (to 3 0 m in s )

R e s e rv e 4 9 .5

4 9 .2

O c c a s io n a l S e r v ic e

Source: NGC Seven Year Statement

10

T r a n s m is s io n B u s in e s s T r a n s m is s io n S e r v ic e s

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

2.19

Continuous response is provided by generation equipped with

appropriate governing systems that control their outputs to neutralise the frequency fluctuations that may arise from relatively modest changes in demand and generation. Traditionally, large synchronised generators instructed to operate in frequency-sensitive mode have provided this service. 2.20

The objective of occasional response is to contain significant and abnormal frequency excursions caused by sudden mismatches in the generation/demand balance (e.g. loss of generation.) Large synchronised generators also provide occasional response services although demand reductions can also contribute when initiated through frequency sensitive relays.

2.21

Although all large generators are required by the Grid Code to have continuous and occasional frequency response capabilities, at any instant, only a relatively small number of participants are actually instructed by the TSO to operate in frequency sensitive mode.

2.22

Theoretically, as system frequency regulation involves real-time energy balancing, the highly interconnected nature of the GB transmission systems means that frequency response can be regarded as national, non-locational service. This implies that so long as transmission system integrity can be maintained and operational constraints on networks can be minimised, the location of the frequency response provider would be immaterial to TSOs. Indeed, the non-locational nature of frequency response delivery can be extended to include distribution connected generators and loads.

2.23

NGC procure frequency response services from both generators and demand-side participants on a mandatory or a commercial basis.

Mandatory frequency response services 2.24

Mandatory frequency response services are only provided by generators and are segregated into Primary, Secondary or High categories. Each of these services differs in terms of the speed and duration of the response required. The E&W Grid Code defines the technical parameters for each mandatory service and also stipulates that synchronous generators rated at over 100 MW must provide such services. 10% of Primary and 10% of Secondary response (machine ratings) must be provided at minimum stable generation. These technical parameters also effectively specify the control infrastructure requirements for generators.

11

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 2.25

Primary frequency response requires the most rapid generator response, albeit over the shortest timescales. The key requirements of Primary frequency response are that generators must be capable of increasing their active power output within 10 seconds of predefined system frequency excursions, and be capable of maintaining this response for a further 20 seconds.

2.26

Secondary frequency response services, whilst requiring a slower initial response must be maintained for longer periods of time. The key requirements of the Secondary frequency response are that generators must be capable of increasing their active power output within 30 seconds of predefined system frequency excursions and be able to maintain this response for a further 30 minutes.

2.27

A consequence of the short delivery timescales for both Primary and Secondary response services is the requirement for sophisticated dynamic control arrangements involving automatic generator governor action.

2.28

Generators providing Primary and Secondary response must be capable of maintaining their output and response characteristics throughout system disturbances to avoid exacerbating problems on the transmission system. These dynamic control and fault-ridethrough characteristics are also specified in the Grid Code and negotiations are underway to extend these requirements to ‘large’ non-synchronous generators connected at both transmission and distribution voltages.

2.29

A fundamental feature of both Primary and Secondary frequency response services is the requirement for generators to have ‘headroom’ in place in order to increase output. By implication, the generators providing such services will, to a lesser or greater extent, be part-loaded. This will significantly impact upon the attractiveness of such services to renewable generators and this issue is explored in greater detail in section 4. Historically, the providers of Primary and Secondary frequency response services tend to be the flexible, albeit economically marginal, large-scale generating units.

2.30

High frequency response services are required in the event of high system frequency scenarios, requiring generators to either reduce output or to cease generating altogether. High frequency services are also initiated through automatic generator governor action.

2.31

Payments for mandatory services are made according to Holding (availability) and Response Energy Payments (delivered energy). Holding Payments to generators are currently based upon cost reflective £/MW/h payments. Response Energy Payments, flowing to and from generators, are remunerated on a £/MWh basis according to 12

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION a monthly average of system buy and sell prices obtained from the Balancing Mechanism (BM) under NETA. 2.32

It should be noted that payments for variations in generator output whilst providing mandatory frequency response services do not relate to Bids or Offers in the BM unless the TSO requires a generator to de-load before undertaking frequency sensitive operation.

2.33

There are a number of options currently being discussed, (within the Balancing Services Standing Group) which could apply market mechanisms to the scheduling and remuneration of mandatory frequency response services.

Commercial frequency response services 2.34

Both generators and demand-side participants can provide commercial frequency response services beyond Grid Code requirements.

2.35

NGC can contract for commercial frequency response on a firm or optional basis. Firm arrangements commit the provider to deliver the service upon NGC’s instruction within time periods defined within the contract. Optional arrangements are more flexible in that the provider has more discretion as to whether the service is delivered.

2.36

Again, commercial frequency response services are remunerated according to Holding (availability) and Response Energy Payments (delivered energy). The key difference between firm and optional arrangements being that holding payments are restricted to those periods that the provider makes the service available.

2.37

Commercial frequency response from demand-side participants are usually sourced from large flexible industrial customer loads capable of changing demand characteristics within short timescales. Such demand-side services can be initiated automatically through the use of low frequency relays, which automatically shed load in the event of pre-determined low frequency excursions.

2.38

A more recent development has been the facilitation of smaller demand-side participation through a single point of contact managed by an aggregating agent, such as Gaz de France.

2.39

The value of the mandatory and commercial frequency response markets for 2003/2004 was approximately £45m. Commercial response services accounted for £26m of this total.

13

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Reserve energy markets 2.40

Reserve energy is also required to provide rapid access to generation, or demand reductions, to accommodate errors in demand forecasting, to provide contingency arrangements for generation failures and to restore frequency response capabilities.

2.41

The arrangements for energy reserves are similar to those for frequency response in that the provision of reserve further ensures that an energy balance can be achieved on the transmission system and thus system frequency and stability can be maintained.

2.42

The key differences between frequency response and reserve services relate to delivery timescales. Typically, reserve services are manually initiated following TSO instructions and involve longer lead times. A consequence of these simplified service initiation requirements is a reduction in the sophistication of the associated control infrastructure making this service more attractive to smaller players.

2.43

In England and Wales, NGC segregate reserve services into the following different categorises: •

regulating reserve;

•

standing reserve;

•

warming and hot standby; and

•

fast reserve.

Regulating reserve 2.44

Regulating reserve is a commercial service provided at the discretion of the generator. Synchronised, part-loaded generators, capable of changing output according to TSO requirements, provide regulating reserve services. TSO instructions are usually issued in accordance with the bid / offer ladders visible in the BM.

2.45

Revenues for regulating reserve are earned in accordance with BM rules whereby the extent of the instructed deviation from a generator’s Physical Notification (PN), when multiplied by the corresponding price, submitted as bids and offers in the BM, determine the level of payment to the service provider.

2.46

The BM enables NGC to readily schedule the different generators offering regulating reserve services on a least cost basis.

2.47

Again, the providers of regulating reserve have historically tended to be the flexible, albeit economically marginal, larger-scale generating units.

14

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Standing reserve 2.48

The key difference between regulating and standing reserve is that the latter is sourced from non-synchronised generators capable of starting, synchronising and providing the TSO’s instructed level of output within 20 minutes.

2.49

In order to provide standing reserve services, the TSO requires the reliable delivery of at least 3 MW of generation (or corresponding load reduction) through a single point of contact. This requirement does not preclude the aggregation of generation (or demand) by an agent or directly by a customer.

2.50

Other technical requirements for standing reserve providers are that the instructed level of service must be provided for at least 20 minutes, the recovery period must not be more than 20 hours and the participant should be capable of providing the service at least 3 time per week.

2.51

The requirement for standing reserve varies according to the time of year, the day of the week and time of day. NGC divides the year into five seasons, for both working and non-working days, and specifies the periods within each day that standing reserve is required. These periods are referred to as availability windows3.

2.52

NGC’s contracting for standing reserve services is an annual activity subject to a competitive tendering process.

2.53

Standing reserve agreements are currently segregated according to whether (or not) the service provider is a participant in the BM.

3

•

For BM participants, standing reserve is remunerated according to availability and utilisation. Direct availability payments are made by NGC to the service provider although utilisation payments are made via ‘Bid/Offer Acceptances’ (BOAs) in the BM.

•

For non-BM participants, standing reserve is again remunerated according to availability and utilisation although the service providers can opt to deliver on a committed or a flexible basis. A committed service provider undertakes to offer service availability for all the required availability windows in each season and NGC commits to accept and buy all services offered. A Flexible service provider is not obliged to offer services in all Availability Windows and National Grid is not obliged to accept and buy all the services offered.

www.nationalgrid.com/uk/indinfo/balancing/pdfs/Standing_Reserve_Introduction.pdf

15

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 2.54

Availability payments are remunerated according to capacity and time on a £/MW/h basis whereas utilisation payments relate to energy on a £/MWh basis.

2.55

Historically, the market participants tendering to provide standing reserve services have experienced high tender success rates (>85%). This applies to BM and non-BM providers (both flexible and committed).

2.56

For the 2004/2005 financial year, NGC has entered contracts to secure nearly 2500 MW of standing reserve from both BM & non-BM participants. Non-BM providers represent 764 MW of this total with 550 MW being provided on a flexible basis. A non-linear relationship exists between site numbers and the amount of standing reserve procured as over 40% of the contracted capability is sourced from only 10% of the sites.

2.57

Whilst the non-BM providers represent approximately 30% of NGC’s contracted MW capability, the BM providers typically provide the majority of standing reserve. In terms of utilisation, only 15% of NGC’s requirements are actually sourced from the non-BM participants by volume.

2.58

Whilst the larger BM generators (>50 MW) provide the majority of standing reserve at present, their corresponding site numbers are low when compared with the smaller participants. It is estimated that over 80% of standing reserve energy is provided by less than less than 15% of the market participants.

2.59

This disparity between site numbers and service utilisation is perhaps understandable in the context of an electricity system characterised by large-scale generation. NGC’s requirements for standing reserve will largely be driven by generation shortfalls caused by equipment failures and these will often be measured in multiples of 100 MW. Consequently, scheduling replacement generation from the larger standing reserve providers appears pragmatic from a network coordination perspective, assuming the prices of standing reserve to be comparable between BM and non-BM participants.

2.60

In terms of market value, again the standing reserve market is heavily biased towards the BM segment as illustrated in Table 4.

16

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Table 4 − NGC Payments for Standing Reserve in 2003/20044

Balancing Mechanism Participants

Non-Balancing Mechanism Participants

Availability £M

Utilisation £M

Availability £M

Utilisation £M

24

14

4

1.4

2.61

As can be seen, the value of the standing reserve market equated to approximately £43 million in 2003/2004.

2.62

Operationally, NGC issued an average of approximately 250 standing reserve instructions per month during 2003/2004 of which approximately 25% were issued to non-BM providers. It is interesting to note that the number and proportion of instructions issued to nonBM participants rose throughout this financial year.

2.63

Scatter plots of prices tendered for standing reserve services for 2004/2005 showed wide variations for both availability and utilisation. Availability prices ranged between £1.50 - £12.75/MW/h, with an average price of approximately £3.50/MW/h. Similarly, wide ranges of utilisation prices were also witnessed of between £20 - £370/MWh with the largest grouping in the £70 - £150/MWh band.

2.64

The providers of standing reserve range from large to small generators and demand customers able to offer demand reductions in response to TSO instructions. The larger generators tend to be BM participants whereas the smaller generators and demand customers are usually connected within distribution systems and hence do not have a direct BM interface.

2.65

Typically the technology types providing standing reserve tend to be conventional, flexible but marginal steam plant, Open Cycle Gas Turbines (OCGTs) and reciprocating internal combustion engines.

2.66

For industrial customers with on-site standby generation requiring regular testing, the standing reserve market can represent an attractive means of fulfilling this requirement whilst simultaneously providing the owner with an additional revenue stream. In order to participate, such customers need to make at least 3 MW of standby generation available to NGC via a single point of contact providing remote start capabilities.

4

National Grid Transco Procurement Guidelines Report, May 2004

17

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 2.67

The TSO advise that one complication of utilising ‘embedded’ standby generation for standing reserve is the possibility that the same generation could also be contracted to provide Transmission Network Use of System (TNUoS) avoidance services to suppliers and customers. In such circumstances, there is an increased likelihood that such generation could already be operating during the periods of transmission system stress and so unable to offer additional reserve services to TSO during critical periods. Such TNUoS avoidance activities effectively undermine the attractiveness of standing reserve contracts.

Warming and Hot Standby 2.68

Warming and Hot Standby services are required to ensure that there is sufficient flexible steam plant available to maintain the desired levels of system stability. Warming and Hot Standby involves maintaining a steam generator’s state of readiness to rapidly deviate from zero output and participate in the BM.

2.69

The value of the Warming and Hot Standby services are significant at approximately £21 million per annum. Whilst Warming & Hot Standby contracts position thermal generators to participate in the Standing Reserve market, it should be noted that any BM bid acceptances, requiring generator output to deviate from zero, cancel any associated warming and hot standby payments. In such situations, generators must recover their costs through the BM.

2.70

Due to the technology types employed and the typical operating regimes of most forms of distributed generation, it is unlikely that Warming and Hot Standby activities will be relevant or applicable to distributed generators.

Fast reserve 2.71

Fast reserve is a fast acting, flexible service from providers capable of increasing energy output or reducing energy consumption, following receipt of electronic instructions from NGC.

2.72

The technical parameters for fast reserve are onerous in that power delivery must start within 2 minutes of instruction at a ramp rate of at least 25MW/minute, and the reserve energy should be sustainable for a minimum of 15 minutes. In addition, fast reserve providers must be able to supply at least 50 MW of reserve power.

2.73

The number of generators capable of meeting these technical requirements is small, to the extent that the bulk of fast reserve is consistently provided less than 10 generating units, e.g. large oil fired generators and pump storage installations.

18

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 2.74

NGC procure fast reserve through a tender process and bilateral contracts. In total, during the 2003/2004 financial year, NGC spent approximately £21 million on fast reserve services.

2.75

The technical requirements of fast reserve largely preclude any participation from distributed generation.

Reactive power 2.76

NGC is required to manage the voltage on the transmission system within technical limits to ensure quality of supply. Due to the impedance characteristics of the transmission system, the ability to decouple and vary reactive power independently of active power, provides a useful means of voltage regulation for the TSO.

2.77

Consequently NGC must ensure that sufficient variable reactive power resources are available on a localised basis to meet the needs of the system, including contingencies.

2.78

NGC procures reactive power through both market-based tender processes and the default arrangements outlined in the Grid Code. The Grid Code defines obligatory reactive power requirements for all generators rated at over 100 MW and default arrangements procure reactive power accordingly. The default arrangements remunerate generators for reactive power according to utilisation on a £/MVArh basis.

2.79

Obligatory reactive power services are also procured through the market-based tender process, which enable generators to influence the attractiveness of their reactive power capabilities. Within the tender process, there is also scope for generators to offer enhanced reactive power services, which go beyond the minimum requirements of the Grid Code although generators seldom exercise this option. Indeed during the last 5 tender rounds, there have been no offers from generators to provide enhanced reactive services.

2.80

Generator remuneration for the market-based reactive power differs from the default arrangements in that both availability (£/MVAr/h) and utilisation (£/MVArh) payments are made.

2.81

In recent years, the value of the transmission level obligatory reactive power markets has been evenly split between the market-based approach and the default arrangements. In 2003/2004 NGC spent £16.9 million on the market based procurement of reactive power and £16.4 million on the default arrangements.

2.82

One issue influencing the size and value of the transmission-level reactive power market is the use of NGC’s Static Voltage

19

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Compensation (SVC) equipment. The installation of such equipment is treated as capital expenditure and associated revenues are recovered through TNUoS charges rather than featuring within the system operator incentive arrangements. Through the installation and utilisation of SVC equipment, NGC can meet local reactive power requirements and thereby avoid the requirement to procure services from generators. 2.83

Whilst theoretically, the market-based and default reactive power arrangements are open to both BM and non-BM participants, all tenders received in recent years have originated from BM participants only. Therefore, the current commercial arrangements for TSO reactive power can accommodate large-scale distributed generators. Indeed, the Grid Code already ensures that such generators are equipped with the necessary infrastructure to participate.

2.84

It is conceivable that TSO reactive power could be sourced from distributed generators, especially those connected at 132 kV, which would effectively be used for transmission system voltage regulation. Reactive power sourced at lower distribution connection voltages will effectively reduce reactive power from transmission-connected generation.

Fast start 2.85

Fast start is the special service provided by OCGT plant to start rapidly from a no-load condition and to deliver full rated power automatically within 5-7 minutes of a TSO instruction. Fast Start services are usually initiated by fully remote control infrastructure.

2.86

In recent years the requirements for fast start services have been consistent and NGC is not currently seeking any additional fast start capacity.

2.87

Fast start providers are remunerated according to capability payments (£/h) and NGC typically spends approximately £3 million per annum on such services.

2.88

The low materiality and the limited market potential for additional fast start services is unlikely to be attractive to distributed generators.

Black start 2.89

In the event that all or part of the transmission system becomes deenergised, NGC must have the capability to re-energise the transmission network to restore supplies. The process of restoring the power system is known as black start.

20

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 2.90

The providers of black start services tend to be the operators of OCGT plant located at large conventional thermal generation sites. In a black start scenario, the onsite OCGT plant is utilised in an islanded mode to re-power main generator essential services, which then reenergise the transmission system, i.e. the OCGT plant does not reenergise the transmission system directly.

2.91

There are currently 18 power stations maintaining black start capabilities although NGC is not currently seeking additional providers for this service. Availability payments form the basis of black start remuneration with NGC typically spending approximately £10 million pounds per annum on such services.

2.92

The low materiality of TSO black start services and the distant nature of much distributed generation to the transmission system (coupled with relatively low electrical ratings) mean that the TSO unlikely to seek black start services from distributed generation.

21

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

[This page is intentionally blank]

22

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

3.

THE SCOPE FOR NEW ANCILLARY SERVICES AT THE DISTRIBUTION LEVEL 3.1

In line with the anticipated increase in distributed generation, both in terms of the number of generators and energy output, it appears reasonable to assume that the scope for existing ancillary services from distributed generators could also increase.

3.2

Similarly, moving from an era with relatively few distributed generators to one with significantly higher densities could also increase the opportunities for new ancillary services to be provided to DNOs. The variety of generator sizes and technologies could present opportunities for distributors to rethink their approach to certain aspects of network management.

Passive and Active distribution network management 3.3

Historically DNOs have planned and managed their distribution networks on a passive basis. These passively managed networks are usually planned to accommodate single direction power flows, from the transmission system to demand customers, over a range of different supply voltages.

3.4

The primary assets (transformers, switchgear, overhead lines and cables) on passive networks are specified to accommodate all anticipated operating conditions, ensuring the technical parameters of supply (e.g. voltage, thermal rating and fault level) are maintained within statutory and safe tolerances, without the requirement for proactive network monitoring and reconfiguration.

3.5

Although the initial capital expenditure requirements for passive approaches can be high, the benefits include robust distribution systems in terms of fault level and uncomplicated, hence low cost, network operation.

3.6

In the event that network usage deviates from the initial planning assumptions, passive networks can prove inflexible to new operating regimes, often requiring significant redesigns and upgrades.

3.7

As much of the envisaged renewable and CHP generation will require connection to distribution systems, there is a concern that the costs of reconfiguration and reinforcement of passive networks could present a barrier to the deployment of distributed generation, especially in situations where power flows are reversed.

23

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 3.8

Consequently, many DNOs are pursing increasingly active approaches to network management to maximise the amounts of generation that can be connected to their systems without the need for costly asset replacements. Active approaches to network management typically require real-time network monitoring with proactive (automated) network reconfiguration arrangements and sophisticated voltage control.

Planning Standards and Network Security 3.9

Prior to privatisation, the Electricity Council was responsible for setting and maintaining a range of common technical and economic guidance documents, including Engineering Recommendation P2/5. ER P2/5 was intended to be used as a guide to system planning and design. The fundamental principle outlined within ER P2/5 is that there should be sufficient connections and capacity in the system such that, in outage situations, customers continue to receive a supply or have it restored within an acceptable time period.

3.10

Historically, the structure of electricity distribution networks was driven by an overall design philosophy developed to support largescale generation technologies. The level of security in distribution networks is defined in terms of the time taken to restore power supplies following a predefined set of outages. Consistent with this concept, security levels on distribution systems are graded according to the total amount of power that can be lost. In general, networks have been specified according to a principle that the greater the amount of power which can be lost, the shorter the recommended restoration time. This philosophy is formalised in the ER P2/5.

3.11

ER P2/5 contains two important tables. Table 1 of ER P2/5 states the minimum demand that must be met after certain specified circuit outages. This level is dependent on the Group Demand or Class of Supply. The amount of demand that can be supplied depends upon the available circuit capacities and critically, the contributions from local distributed generators. Table 2 in ER P2/5 specifies the contribution that can be attributed to generation connected to a particular load group.

3.12

Unfortunately, as ER P2/5 was developed in the era of centralised generation planning, it does not currently recognise many of the modern forms of distributed generation currently being connected into distribution networks in the pursuit of Government climate change targets. Consequently, it is not possible to recognise the security contributions of many new forms of distributed generation. It is anticipated that ER P2/5 will be superseded by ER P2/6 in the near future which will facilitate the inclusion of security contributions from distributed generation during network planning. 24

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 3.13

It should be noted that ER P2/5 is a design standard as opposed to an operational standard. The significance of this distinction is important. A design standard simply implies that, if the design conforms to a specified set of conditions, the designer will have complied with the planning requirement. This does not guarantee that conditions will not arise during network operation which could result in loss of supplies, i.e. planning standards cannot guarantee that operational standards can be satisfied at all times.

3.14

Although situations may arise in which the required level of network security is insufficient, this does not automatically imply that the design standard has been violated. Security shortfalls may occur in situations where insufficient generation is available during a particular network outage, e.g. generator output was less than assumed during the design phase. Similarly, security shortfalls could also be attributed to insufficient network capacity being available due to unforeseen circuit outages. It therefore follows that whilst a design standard can minimise operational standard violations, it cannot eliminate them.

3.15

The primary purpose of ER P2/5 is to provide planning guidance and a minimum set of network security requirements. ER P2/5 states that the capacity of a network to meet a group demand should be assessed as the aggregate of: •

the appropriate cyclic rating of the remaining transmission or distribution circuits which normally supply a group demand, following outage of the most critical circuit(s);

•

PLUS any transfer capacity which can be made available from alternative sources; and

•

PLUS, for demand groups containing generation, the effective contribution of the generation to network capacity as specified in updated Table 2 of ER P2/5.

3.16

ER P2/5 does not differentiate between different sources of network security, e.g. distribution circuits and generation sources. Indeed, ER P2/5 facilitates direct comparisons between these different approaches to network security. ER P2/5 enables network planners to quickly evaluate whether networks comply with the minimum planning standards.

3.17

In an era of significantly increased levels of distributed generation, with a suitably updated planning standard recognising a wide range of distributed generation technologies, there will be significant opportunities for generators to provide security contributions to network planners.

25

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Network Operational Standards 3.18

Since privatisation, a number of operational standards have been developed by successive regulatory authorities, which safeguard supply quality. These operational standards include the Guaranteed Standards and Overall Standards of Performance. Guaranteed Standards set service levels that must be met in individual cases. If the DNO fails to provide the level of service specified, it must make a penalty payment to the customer affected upon request.

3.19

At present, the main security related Guaranteed Standard (GS) focuses upon supply restoration times. To ensure that inconvenience to customers is kept to a minimum, this GS requires DNOs to restore supplies within 18 hours of becoming aware of a fault on their system. Additional performance measures have been developed as part of the Information and Incentives Project (IIP).

3.20

Overall Standards (OS) address other aspects of service quality where it is not appropriate to give individual guarantees, but where customers have a right to expect predetermined minimum levels of service from DNOs. No penalty payments are made if these levels are not achieved and it has been proposed that OS should be removed or replaced.

3.21

Whilst planning standards such as ER P2/5 must be satisfied at the design stage, the ability to satisfy the relevant GS and OS only becomes apparent after system operation has commenced, when sufficient data has become available for analysis.

3.22

There has been a shift of focus in the regulation of distribution networks, from asset based to performance-oriented regulation. The Information and Incentive Project (IIP) was established to strengthen incentives and increase the quality of outputs. As the IIP is based on financial penalties and rewards, the overall distribution revenue is a function not only of the operating and capital costs incurred by the network owner in providing the service, but also depends on the quality of service to the customer. The indices used to measure supply quality are the number of interruptions per 100 connected customers per year (CIs), the number of customer minutes lost per connected customer per year (CMLs) and details of the worst served customers.

Localised network support through generator voltage and power flow management 3.23

Historically, voltage regulation on passively managed distribution networks has been achieved through the robust specification of primary network infrastructure, on-line transformer tap changing

26

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION equipment and fixed tap distribution transformers. Also DNOs have ensured that network components are operated within thermal limits. 3.24

The robust specification of passive networks effectively minimises voltage variations across a wide range of operating conditions, e.g. from no load to full load. Also, as most passive networks have only been required to accommodate power flows in a single direction, the primary concern of DNOs has been to manage voltage drops. One consequence of managing voltage drops and thermal ratings has been that many voltage control arrangements have asymmetric tapping ranges.

3.25

The magnitude of voltage variations can also depend upon localised network characteristics. For example, voltage variations on highly interconnected, cable based urban networks are typically small whereas the corresponding variations on rural, radial Overhead Line (OHL) circuits can be much more significant.

3.26

Maintaining voltages and thermal ratings within tolerances represent absolute technical obligations for DNOs. The implications to a DNO of voltages being outside statutory limits are severe as damage to customer equipment can occur. Similarly, prolonged operation of equipment beyond thermal ratings can lead to component failures on DNO networks. Consequently, short and long-term derogations are not appropriate for such out-of-tolerance conditions. If voltages are found to be outside statutory limits, DNOs must remedy such situations immediately or interrupt supplies to customers.

3.27

The Electricity Safety, Quality and Continuity Regulations5 of 2002 state that, unless otherwise agreed in writing, the permitted variations in voltages on distribution networks are as follows: •

In the case of a low voltage supply, a variation not exceeding 10 per cent above or 6 per cent below the declared voltage at the declared frequency.

•

In the case of a high voltage supply operating at a voltage below 132,000 volts, a variation not exceeding 6 per cent above or below the declared voltage at the declared frequency.

•

In the case of a high voltage supply operating at a voltage of 132,000 volts or above, a variation not exceeding 10 per cent above or below the declared voltage at the declared frequency.

3.28

The trend towards increasingly actively managed distribution networks in conjunction with increased penetrations of distributed

5

The Electricity Safety, Quality and Continuity Regulations 2002

27

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION generation could give rise to wider ranges of power flows (including reversals) and hence wider voltage variations. Such variations could be outside the operating ranges of existing voltage regulating equipment, thus requiring DNOs to evaluate revised voltage control arrangements. 3.29

The network areas where DNOs will have the greatest exposure to fluctuating supply voltages caused by distributed generation is on rural 11 kV circuits. This can be attributed to lower fault levels, long OHL circuits and limited voltage regulation infrastructure.

3.30

In an era of significantly higher levels of distributed generation, there will be increased scope for distribution network voltages to be controlled through the regulation of generator outputs. Such approaches would be similar to the technique adopted for voltage control on transmission networks, although DNOs would probably need to regulate both active and reactive power flows in order to achieve the desired voltage changes.

3.31

Another opportunity for DNOs to utilise generator power flow management will be in network outage situations where circuit overloads can be mitigated through generator output in order that supplies to customers can be maintained.

3.32

Where generation based solutions can be adopted for voltage control or overload avoidance, the benefits to DNOs could be measured in terms of avoided network reinforcement costs. Similarly, such techniques could also benefit generators in terms of reduced costs of connection.

3.33

One possible concern regarding the use of generation based solutions for distribution network voltage control relates to the availability and reliability of generation equipment. Difficulties would arise in a scenario where a DNO requires a specific level of output although the generator is unavailable. As voltage represents an absolute technical requirement in terms of supply quality, DNOs would be incentivised to seek contingency arrangements should a particular generator fail, i.e. the attractiveness of such voltage control schemes will increase in line with the number of generators in a particular network locality.

3.34

Undoubtedly, those generators with the highest availabilities, the broadest and most flexible output ranges for both active and reactive power will represent the most attractive service providers of voltage management network overload services to DNOs in the future.

3.35

In special circumstances DG could occasionally provide short duration ancillary services to support networks in situations where

28

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION maintenance, refurbishment or replacement activities impose network constraints. Some of these services are already provided by transportable standby generation, typically mobile diesel generators. This form of services is not however considered to be of primary concern of this project and is hence not elaborated further.

Common characteristics and key differences between distribution and transmission sourced ancillary services 3.36

Having reviewed the ancillary services provided at a transmission level and the new services, which could be provided at a distribution level, it is worth contrasting the similarities and differences between these potential new markets for distributed generation.

3.37

In terms of market structure, the transmission level arrangements benefit from a single party being responsible for co-ordinating services on a national basis. This ensures that localised variations are limited regarding commercial arrangements and technical standards, thus enabling generators to contract on a consistent basis.

3.38

The development of mandatory grid codes at the transmission level also serves to clarify the responsibilities and obligations of different parties with respect to service provision. Whilst Grid Code requirements are prescriptive, they ensure the provision of infrastructure, which facilitates the provision of some ancillary services. Whilst the Distribution Code is a national document, it does not yet contain any detailed requirements regarding the provision of distribution level ancillary services.

3.39

Obviously, with 14 DNOs managed by 7 different companies, there is scope for divergence regarding the service definitions and requirements relevant to distributed generators. From a commercial perspective, this could represent a sub-optimal outcome for distributed generators operating nationally, as a variety of different contractual structures may need to be accommodated. Such divergence could impose increased costs on generators, and to the industry as a whole, whilst simultaneously undermining transparency. A downside of too much standardisation could be reduced incentives to develop innovative service offerings.

3.40

The procurement of frequency response and reserve by the TSO is simplified by the non-locational nature of these services, which broadens the choice of service providers and facilitates the development of market based commercial arrangements. By contrast, TSO reactive power services, distribution network security, distribution voltage support and quality of supply related services are

29

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION all highly locational in nature which impacts upon potential commercial arrangements as discussed in Section 6. 3.41

The sizes and types of generators providing services to the TSO differ considerably from those predicted for distributed generation. Typically, transmission connected generation is rated in hundreds of megawatts, based upon well-proven Rankin, combined cycle or hydro-electric technology and is connected to either 275 or 400 kV networks (132 kV in Scotland). The TSO arrangements are further simplified by the widespread usage of synchronous electrical generators. In terms of operational interfaces, the TSO needs only to communicate with a maximum of approximately 400 generators, the vast majority of which have a 24 hour manned presence.

3.42

By comparison, distributed generation employs many different technologies and fuels utilising a variety of synchronous and nonsynchronous electrical generators. As the size variations between the different forms of distributed generation are considerable (<10 kW & >100 MW), this also results in a wider range of connection voltages (0.4 kV – 132 kV). The technical capabilities of the different types of distributed generation also vary considerably so generic standards for service provision would be more difficult to develop.

3.43

Whilst present numbers of distributed generators are relatively low, significant increases will be required to achieve Government targets for renewable and CHP generation (potentially many thousand generators per DNO). Most distributed generators do not have a 24 hour manned presence and the associated monitoring, communication and control infrastructure is often basic. Features such as remote start-up, automatic voltage regulation, fault ride through, parallel running, governor control, islanded operation and real time metering have historically not been features of distributed generation design although many of these are key to the provision of ancillary services.

3.44

In future, the situation regarding the technical requirements applicable to large distributed generators may change as the TSO seeks to place similar obligations on these distributed generators as already apply to transmission connected generation.

3.45

The expectation that distribution networks will become increasing actively managed will create technical and operational challenges for DNOs more accustomed to managing passive networks. Such considerations do not apply to the TSO as the transmission system has evolved into an active network and is resourced and structured accordingly.

30

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

[This page is intentionally blank]

31

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

4.

EXPLORATION OF ANCILLARY SERVICES FROM DISTRIBUTED GENERATION

4.1

This section explores the features of the most significant ancillary services, which may be provided by the different forms of distributed generation.

4.2

The services for which potential arrangements have been explored are:

4.3

•

TSO Frequency Response;

•

TSO Regulating and Standing Reserve;

•

TSO Reactive Power;

•

DNO Security of Supply contributions;

•

DNO Quality of Supply Services; and

•

DNO Voltage and Power Flow Management Services.

The more specialist TSO services outlined in section 2, including Warming & Hot Standby, Fast Reserve, Fast Start and Black Start, have been excluded due to low materiality, limited market potential and/or inapplicability to distributed generation.

Features of the ancillary services required by service recipients 4.4

When assessing the potential for each of the selected ancillary services to be provided by distributed generators, it is important to consider the following generic requirements of each service: •

Running status;

•

Start-up times;

•

Availability;

•

Intermittency;

•

Scale of opportunity; and

•

Valuation methodology.

TSO Frequency Response Services 4.5

As already discussed, frequency is regulated nationally, on a realtime basis and within tight tolerances by NGC as TSO. Frequency response can be provided by large-scale generators (predominantly transmission connected) and demand-side participants.

32

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 4.6

An essential feature of frequency response services from generators is the requirement for part-load operation, which ensures that rapid changes (especially increases) in generator output are achievable. The requirement to regulate generator outputs within short timescales necessitates sophisticated control infrastructure. This control infrastructure is a mandatory Grid Code requirement for all generators rated at more than 100 MW, including distributed generators.

4.7

Whilst frequency response services from distributed generation are comparatively rare, there are a limited number of highly flexible CCGT generators capable of providing these services by regulating the output of the gas turbine. The size of these modules is typically 300 – 400 MW, which necessitates connection to 132 kV networks.

4.8

Some commercial frequency response services are also procured from distribution-connected customers. Such services tend to be facilitated through the use of Low Frequency relays, which can be arranged to automatically trip flexible industrial demand, according to predefined frequency excursions, within very short timescales.

4.9

One possibility for increased DG participation in future frequency response markets could be to utilise similar low frequency relay technology to initiate increased generator output rather than demand reduction.

4.10

Whilst the delivery of frequency response services from contracted generators is important to NGC, the national nature of the service coupled with many potential service providers, means the TSO can normally secure contingency arrangements from a variety of generators at short notice. Repeated service non-delivery could either place generators in breach of Grid Code obligations or bilateral contractual arrangements. NGC reserve the right to interrupt payments to service providers in the event of consistent poor service delivery. NGC also has access to sophisticated monitoring infrastructure to confirm services have been provided and also initiates tests to confirm providers are maintaining their service capability.

Impact of intermittent generation on frequency control 4.11

With the anticipated rise in the levels of intermittent renewable energy (e.g. wind), the GB power system will potentially be required to accommodate larger imbalances between generation and demand, thus increasing the demand for balancing services. This will

33

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION inevitably increase the volume of continuous frequency response services required6.

Impact of wind generation on the need for continuous services 4.12

The effect of rapid variations in the output of individual wind generators will be relatively minor, as the level of correlation between the fluctuating outputs of individual wind farms will be very small in the time horizons considered (several seconds to a minute).

4.13

However, as the volume of intermittent generation increases, the error in the forecast of its aggregate output will also increase. This will result in an overall increase in the forecast error of the demand/generation balance, which will place an additional duty on the remaining generating plant to provide additional capabilities to contain system frequency within allowable limits. Estimates of additional continuous service requirements with respect to wind power penetration are as presented in Table 5. These have been derived according to the same methodology as adopted in SCAR7. Table 5 − Estimated additional requirements for continuous frequency response8

Installed Wind Capacity (GW)

0

5

10

15

20

Additional continuous response requirement (MW)

0

45

160

320

500

Estimates of future value of continuous response services 4.14

Assuming that all generating plant is Grid Code compliant and hence adequately equipped to provide continuous frequency regulation, the direct costs of its provision are driven by fuel cost, the efficiency of plant when operated at part load and the O&M cost arising from

6

The requirement for occasional response services could also expand if intermittent generators are derogated from compliance with Grid Code requirements regarding faultride-through capabilities.

7

ILEX/UMIST, System Costs of Additional Renewables, a report to the DTI, October 2002

8

Estimates derived as part of this project

34

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION additional wear and tear. Thermal units operate less efficiently when part-loaded, with an average efficiency loss of about 20%. 4.15

Assuming an average energy cost of £20/MWh, estimates for the additional fuel costs associated with holding this dynamic response are presented in Table 6. An additional assumption was that the delivery of 1 MW of dynamic response would require generators to be de-loaded by approximately 1.5 MW.

4.16

Estimates of the value of additional continuous service requirements are also made on the basis of present expenditure associated with the service. The current value of the market associated with frequency control is estimated at about £50m per year. Making the simplifying assumption that the increased volume of service would be provided at the current average price, the additional value of the services with respect to various levels of penetration of wind generation is very close to the cost based figures.

35

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Table 6 − Cost and market-extrapolation based estimates of the value additional continuous response9

Installed wind capacity (GW)

0

5

10

15

20

Cost based estimate (£M/annum)

0

3

10

21

33

Market-extrapolation based estimates (£M/annum)

0

3

12

24

39

Impact of wind generation on occasional response services 4.17

The amount of occasional frequency response required is primarily driven by the size of the largest in-feed to the system (1320 MW, Sizewell B). Another determining factor is system inertia, which establishes the initial rate of change of frequency following a disturbance, e.g. following a major loss of generation. The requirement for fast acting response during high loading conditions is therefore less than that needed for light loading situations. However, no requirements are being imposed regarding minimum levels of inertia for individual plant.

4.18

Wind generation is at present based on induction machine technology. This is different to conventional synchronous generation and, at present, is not able to readily provide similar levels of support to system operation and stability10. Current Grid Code proposals from the GB transmission system operators will require new wind plant to provide system support services.

4.19

The need for additional occasional response will be driven by the degree of robustness of wind generation to withstand disturbances, their exact location on the system, demand and wind conditions and the ability of the control systems of wind generators to make their inertia available to the system. Should the new Grid Code proposals regarding fault-ride-through be adopted, it is unlikely that there will

9

Estimates derived as part of this project.

10

It is a Grid Code requirement that generation will operate in a stable manner during and after faults on the transmission network (fault-ride-through capability).

36

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION be a significant increase in the need for occasional frequency response. 4.20

The requirement for continuous frequency regulation will increase from the current average level of 600MW with wind penetration due to wind forecasting errors. The requirement for occasional frequency response could increase from the current level of 1320MW, if some wind generation receives derogations from Grid Code requirements regarding fault-ride-through capability.

Opportunities for DG 4.21

As discussed above, frequency regulation services are currently provided by large generators equipped with appropriate governing systems and customers capable of rapid demand reductions triggered by low frequency relays. The system operator instructs individual generators to operate in frequency sensitive mode. If a generator, operating at full load, is instructed to run in frequency sensitive mode, the system operator would also need to position the unit for the delivery of the response service by accepting corresponding bids in BM in order to de-load the generator.

4.22

DG could in principle contribute to frequency regulation services, although in order to provide low frequency services, DG plant would be required to operate in frequency sensitive mode and also run partloaded. DG technologies that could potentially provide continuous service include CCGT, biomass, CHP, doubly fed induction wind generators (if appropriate modifications could be made to control algorithms and infrastructure). The capabilities of each technology are further discussed in section 5.

4.23

From our analysis we conclude it is unlikely that these services would be supplied regularly by renewable generation, as the opportunity cost of operating part-loaded would be relatively high due to the requirement to recover lost ROC revenue.

4.24

It is conceivable that in low load situations, wind farm outputs could be required and compensated (via the BM), to de-load for system stability purposes. In such situations, wind farms could then provide low frequency regulating services.

4.25

Similarly, in addition to such low frequency response services, renewable generation could provide high frequency response services. However, the value of such services is relatively low and is unlikely to be sufficiently attractive to renewable generation.

4.26

In future, assuming appropriate modifications to wind turbine control systems, doubly fed induction generators could also contribute

37

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION inertia effects which would reduce the rate of frequency fall following losses of generation. Although there are benefits from providing inertia, the mechanisms for rewarding generators are not yet established. 4.27

It is also important to recognise that frequency response services are currently provided by a relatively small number of generators. Clearly, if the number of participants increases significantly, the complexity and cost of operation and monitoring of this market will increase similarly therefore devaluing the disaggregated service.

4.28

The most straightforward means of expanding frequency response markets to include greater numbers of distributed generators will be to extend the existing arrangements for commercial frequency response services. However, such opportunities for distributed generation may be constrained by two considerations: •

infrastructure requirements; and

•

size of generator.

4.29

The infrastructure required by the TSO for generators to provide low frequency response services are mandated for ‘large’ generators by the Grid Code. Historically, distributed generation has not been fitted with such infrastructure although retrofit might be possible in some cases. The rationale to retrofit such infrastructure to distributed generation would depend on the cost of the equipment and the likelihood of the service being attractive to the TSO.

4.30

From a TSO perspective, generator size is important to ensure the stability of the system can be maintained. Operationally, NGC will seek meaningful response contributions relative to the overall size of the transmission system. The individual contributions of small generators may not be particularly useful to the TSO unless such contributions could be aggregated through a single point of contact, thereby increasing the scale of response whilst simultaneously minimising operational complexity.

4.31

The combination of these infrastructure and generator size considerations imply that frequency response services will be attractive to the larger, non-renewable (yet still synchronised) generators unless an innovative, low-cost means of scheduling automated mass responses from highly flexible small plant can be developed. If such a scenario does arise, it is likely that the scope for aggregation services will increase.

4.32

Although mandatory frequency response capabilities may become a technical requirement for large distribution connected wind farms,

38

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION thereby ‘resolving’ any infrastructure constraints, the extent to which the TSO will utilise such capabilities is unclear.

TSO regulating and standing reserve markets 4.33

As outlined in section 2, reserve services are managed centrally by the TSO to balance electricity generation and demand nationally. Reserve is sourced from both synchronised and standing providers.

4.34

The monitoring and control requirements for the provision of reserve services are less onerous than those required for frequency response owing to the longer timescales for service delivery. The delivery of reserve is initiated by TSO instruction rather than through continuous dynamic control loops.

4.35

The majority of synchronised regulating reserve providers are generators with individual Balancing Mechanism Units (BMUs), which are readily able to participate in the NETA balancing mechanism. Similar to low frequency response services, a key requirement for regulating reserve providers is to operate in a partloaded mode, which can result in thermal efficiency losses of about 20%.

4.36

Standing reserve is currently provided by a wide range of different generating technologies connected at both transmission and distribution voltages. Providers of standing reserve services do not necessarily require a dedicated BMU, which facilitates participation by smaller distributed generators either through a supplier BMU or contracts directly between the TSO and the provider.

4.37

Whilst NGC requires a minimum contribution of 3 MW for standing reserve services, participation by even smaller distributed generators is also facilitated through aggregation. As long as sufficient generation can be aggregated through a single point of contact to exceed the 3 MW threshold, multiple small distributed generators can contribute reserve energy. Indeed the TSO can accommodate aggregation from either 3rd parties (e.g. suppliers, dedicated aggregators or DNOs) or customers directly.

4.38

The allocation of reserve between synchronised and standing plant is a trade-off between the cost of efficiency losses of part-loaded synchronised plant (plant with relatively low marginal costs but running at all times) and the cost of operating less efficient standing plant (plant with relatively high marginal cost but running only occasionally).

39

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 4.39

Regulating reserve services from distributed generation are more common than frequency response services with a number of CCGT generators participating. As mentioned previously, the size of such plant necessitates connection to 132 kV networks. It is envisaged that large distributed providers will continue to participate in both the regulating and standing reserve markets.

4.40

Similar to frequency response, although the reliable delivery of reserve services from generators is important to the TSO, the national nature of such services coupled with a choice of service providers, means the TSO can usually secure sufficient reserve from other providers in the event of the failure of one generator. Again, repeated non-delivery could place generators in breach of contractual obligations and NGC accordingly reserve the right to interrupt payments to unreliable service providers. NGC also has access to sophisticated monitoring infrastructure to confirm services have been provided and also initiates tests to confirm providers are maintaining their service capability.

4.41

When assessing the future role of distributed generation with respect to the provision of reserve services, it is important to consider how TSO requirements for such services will change in the future.

4.42

A critical factor influencing the size of the reserve markets in future will be the level and penetration of intermittent generation. The magnitude of output fluctuations from intermittent sources will depend, not only on the level of penetration, but also upon the time horizon considered. Statistical analysis of wind output fluctuations over various time horizons can be performed to characterise the uncertainty of wind output.

4.43

When analysing the need for additional reserve (considering all forms of reserve together), a time horizon of up to 3 - 4 hours is typically considered. This is driven by the assumption that time horizons of more than 3 - 4 hours will involve the start-up of additional units, which should be within the dynamic capabilities of conventional gas fired technologies.

4.44

In order to manage uncertainty in the present system over time horizons of 3 - 4 hours, the TSO holds approximately 2,400MW of various types of reserve11. By quantifying the wind forecasting error, as a function of wind power installed (assuming no correlation between the two), total reserve requirements can be estimated.

11

SCAR, ILEX/UMIST, October 2002

40

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 4.45

Table 7 presents the amount of reserve required to accommodate changes in output at various levels of wind penetration, using a fourhour time horizon. Table 7 − Additional reserve required for various levels of wind penetration12

4.46

Installed wind capacity (GW)

0

5

10

15

20

Additional reserve requirement (MW)

0

500

1600

3000

4400

The penetration of wind power will increase overall imbalances in the market as the forecast errors increase. Given one hour Gate Closure, the volume of additional BM offers for various penetrations of wind generation are given in Table 8. Table 8 − Estimates of net volume of additional bid offer acceptances13

Installed wind capacity (GW)

0

5

10

15

20

Estimated volume of extra BM offers (GWh)

0

125

375

750

1175

Estimated value of additional reserve 4.47

Estimates of the value associated with future additional reserve requirements can be made on the basis of present levels of expenditure on reserve services. The present annual costs of these two forms of reserve are in the order of £60 million per annum14. Table 9 presents estimates for the additional market value of reserve assuming that these are procured at the same price as the current average market price.

12

Estimates derived as part of this project

13

Estimates derived as part of this project

14

Procurement Guidelines Report for 1 May 2002 to 30 April 2003, National Grid, 30 May 2003.

41

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Table 9 − Market value of additional reserve

4.48

Installed wind capacity (GW)

0

5

10

15

20

Additional value of reserve (£M/annum)

0

11

37

67

101

Furthermore, it is expected that flexible plant will benefit from the increased value in the BM or short-term power exchanges, caused by the fluctuations in wind output creating the imbalance volumes illustrated in Table .

Opportunities for DG 4.49

DG could in principle contribute to reserve services. A number of DG technologies could provide synchronised reserve including partloaded CCGTs, biomass and wind generation, while standing reserve could be provided by diesel engines, OCGTs, and in some circumstances CHP including micro schemes (provided that suitable communication and control infrastructure were available). Indeed, if CHP schemes were to provide standing reserve, ‘heat-dump’ facilities may also be required. The capabilities of each technology are further discussed in section 5.

4.50

It is however unlikely that synchronised reserve would be supplied regularly by renewable generation, as the compensation from providing the services would need to recover the loss in ROC and energy revenues due to part-load operation. It is therefore unlikely that DG technologies relevant to achieving Government targets would be able to effectively compete in this market.

4.51

If in infrequent situations with excess wind power being available during low load conditions, wind generation may need to be deloaded, and the wind farms might then provide reserve services as for low frequency response services.

4.52

A number of distribution connected industrial and commercial customers have installed standby generation to increase reliability of supplies to high-value loads. Some of this plant is already participating in the TSO’s standing reserve market.

4.53

The most straightforward means of expanding reserve markets to include greater numbers of distributed generators will be to extend

42

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION the TSO’s existing arrangements for reserve and to encourage suppliers and 3rd parties to refine and develop aggregation services. 4.54

NGC’s Seven Year Statement (SYS) presents useful information regarding the providers of TSO services. It contains indicative estimates of future capacity margins up to 2010/2011 against a number of backgrounds. For margins calculated using customerbased demands, the ‘SYS background’ shows margins in excess of 20% for all years, whilst margins presented the ‘Existing and Under Construction’ background indicate significantly lower values in later years (12.8% in 2009 and 9.5% in 2010)15. In the latter case, capacity margins would be significantly lower than the 24% considered as secure by the CEGB.

4.55

If generation margins reduce significantly below 20% in the short to medium term, and the volume of ancillary services offered by conventional generation reduce, this may present an opportunities for DG to provide increased amounts of TSO ancillary services, potentially at higher prices. More work will however be required to quantify the impact of the potential reduction in central generation margins and the related opportunities that may emerge for DG.

TSO Reactive Power 4.56

In any ac power system the voltage and current may not be in phase and hence reactive power will flow. Reactive power may be considered as being absorbed by inductive components (e.g. transformers, distribution overhead lines, induction motors) and generated by over-excited synchronous machines and capacitors.

4.57

Transporting significant amounts of reactive power over predominantly inductive circuits will cause relatively large voltage drops. Therefore, given the electrical characteristics of transmission circuits, and in order to keep voltage fluctuation within allowable limits, reactive power demanded by loads will need to be supplied from local sources.

4.58

Management of reactive power flows is critical to maintain adequate voltage profiles on the system. The main sources of regulating reactive power are large conventional generators, in addition to compensation devices that include SVCs and switched reactors and capacitors.

15

It should be noted however that the generation backgrounds take no account of future plant closures, over and above the Nuclear Magnox sites

43

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 4.59

Management of network voltage profile is complicated by the transmission system consuming or generating significant amounts of reactive power, depending on circuit parameters. Given that voltage variations in transmission system are relatively small (say ±10%), reactive power generated by a transmission line, due to shunt capacitance effect, is almost constant. However, the amount of reactive power that the line will absorb, due to series reactance, will vary significantly with loading condition. Hence, during light loading condition long transmission lines may generate significant amount of reactive power. In order to ensure that the network voltages do not exceed the allowable limit, this (unwanted) reactive power will need to be absorbed by generators and compensators. Conversely, during heavy loading conditions, transmission lines will absorb reactive power that would need to be supplied by (local) generators and compensation facilities to ensure that voltages across the network do not fall below allowable limit.

4.60

Furthermore, under normal operating conditions, generators and other reactive compensation devices are not normally run at their maximum capability. There will be a need for a margin of reactive reserve to be held. These reactive reserves are maintained primarily to provide additional reactive power support in the event of an outage. For instance, with the loss of a transmission circuit, the network configuration changes can increase system impedance, which in turn increases reactive demand due to increased reactive losses. This increase is supplied form reactive reserves. System operation must therefore, ensure that sufficient reactive reserve is held for all credible contingencies. As indicated above, due to localised requirement of reactive support, these reserves must be appropriately distributed across the network.

4.61

In addition to the vital contribution to system security, voltage control strategies and reactive power management may have significant impact on both capital expenditure and system operation costs. Reactive power support and reserve are critical for efficient use of the existing transmission capacity, because the system may become voltage limited if the reactive loading is not managed appropriately, and this could lead to under utilisation of capital investments. Simultaneously, inadequate reactive support would inevitably cause additional operating costs, such as voltage constraints requiring redispatch of generation and hence increase balancing costs. Furthermore, inadequate reactive support is likely to lead to an increase in active power losses. On the other hand, excessive reactive support would expose the users of the transmission system to an increase in prices for using the system due to an unnecessary increase in the corresponding capital expenditure.

44

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 4.62

Costs of providing reactive power support have generally two components: (i) capital costs associated with the provision of reactive capability and (ii) operating and maintenance costs, which depend on the level of reactive output relative to capabilities. These cost components are discussed in more detail in the references given in the footnote16.

4.63

As discussed earlier, the TSO procures obligatory reactive power services from generators through the reactive power market or the default arrangements and has an incentive to manage total expenditure for the provision of reactive power service. The selection process takes into account location of generators, various network configurations and the costs of competing options. However, NGC’s own compensating plants do not participate in either market.

Impact of DG on TSO requirements 4.64

4.65

16

Distributed generation can make a significant impact on the amount of reactive power exchanged between TSO and DNO systems. This will be driven by a number of factors including: •

the level of penetration of distributed generation;

•

the voltage level at which the generation is connected;

•

the technology and operating regime of generation;

•

electrical network characteristics of distribution circuits; and

•

variations in loading conditions.

With respect to the various different types of DG operating regimes, a number of scenarios can be considered: •

DG generates active power only: by generating active power in distribution networks, distributed generation will reduce corresponding amounts of power imported from the transmission networks. This reduction in flow will reduce reactive consumption (losses) of distribution circuits and hence less reactive power will be imported from the transmission network.

•

DG generates active and reactive power: by generating reactive power locally, distributed generation can supply some of the reactive demand to local loads and contribute to the supply of

L. Mogridge, ‘Economics of provision of a reactive power/voltage control service by Generating plant’, paper no. 500-01, CIRGE Symposium on Open Access, Tours, France, 8-12 June 1997. IEE Colloquium on ‘Economic provision of reactive power for system voltage control’, London, October, 1996

45

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION reactive losses in distribution circuits. This would normally result in a more significant reduction in the amount of reactive power imported from the transmission network. •

4.66

DG generates active and absorbs reactive power: by absorbing reactive power, DG will tend to increase the demand for reactive power. The net effect will be driven by the overall balance between the increase of reactive power demand by DG and reduction caused by exporting active power.

In order to illustrate how DNO demand for reactive power varies with different levels of DG penetration (over a range of DG power factors), a simple generic model of the UK distribution was created. The schematic diagram of the system is shown in Figure 2. The distribution of load across the network and the assumed peak load power factors are provided in Table 10. Figure 2 − Generic distribution model with DG connected at 11kV and 33kV. Trans miss ion Net wor k

132kV c irc uit

132/ 33 kV Sub

33 kV c ircuit

33/11 kV Sub

DG - 33 kV

Load - 33kV

11 kV c ircuit DG - 11 kV

11/0.4 kV Sub

Load - 33 kV

Load - 0.4 kV

Table 10 − Distribution of load across voltage levels

Voltage Level

Percentage of load

46

Power factor

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 0.4 kV

65%

0.995

11 kV

30%

0.98

33 kV

5%

0.98

4.67

As a simplification, all DG was assumed to be connected at 11kV or 33kV with an equal split each and an even distribution across the country17.

4.68

Electrical parameters of the generic network (lines and transformers) are assumed to have typical values. Results are summarised in Table 11. Table 11 − The total DNO reactive import [MVAr] from the transmission network in MVAr at peak for different level penetration of DG and various power factors (p.f.)

Aggregate Peak DNO Reactive Imports (MVArs) Penetration of DG [GW]

p.f. = 0.975 (lagging)

p.f. = 1

p.f. = 0.975 (leading)

0

19,981

19,981

19,981

2.5

18,107

18,790

19,477

5

16,325

17,671

19,033

7.5

14,629

16,620

18,647

10

13,013

15,635

18,319

4.69

From the above table we observe that the output from distributed generation would tend to reduce the amount of reactive power that will be imported from the transmission network during system peak. Furthermore, given the assumptions regarding the distribution of generation across DNO voltage levels, the amount of reactive power imported will reduce with increased amounts of distributed

17

This exercise is carried out for illustrative purposes rather than for quantifying changes in DNOs reactive power demand, and hence simplistic assumptions are considered to be appropriate.

47

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION generation connected, irrespective of the power factor at which DG operates. Clearly, reduction in active power flows in high voltage distribution networks caused by distributed generation will reduce reactive power consumed by lines (reactive losses) and hence reduce the need for reactive power import. 4.70

As expected, however, the largest reduction in reactive power import would be achieved where the distributed generation also injected reactive power18. The reactive power output generated would be used to supply local reactive loads, reducing the need for the import of reactive power from the transmission network. On the other hand, reductions in reactive imports would not be very significant when the generation operates with a leading power factor.

4.71

The reduction in the import of reactive power from the transmission network would lead to a corresponding reduction in reactive power that needs to be produced by conventional generation, and hence this would reduce TSO expenditure on reactive power. This corresponds to a benefit provided to the TSO from distributed generation. In order to examine the materiality of the potential savings in transmission system operating costs, it is appropriate to consider the unity power factor scenario.

4.72

By examining the results of the analysis presented above, we conclude that the reduction in TSO reactive power requirements is between 430MVAr and 470MVAr per GW of distributed generation installed19. Taking the average to be 450MVAr/GW and assuming that the savings apply for one quarter of each year, this would equate to a value of approximately £1.2 /year for each kW of distributed generation installed20.

4.73

As the value of the savings achieved do not appear to be very significant, TSO reactive power displacement is unlikely to represent a particularly attractive opportunity for distributed generation. However, given the local nature of TSO reactive power services, there may be specific situations, which represent higher value opportunities for DG to reduce to reactive requirements21.

18

In this high level analysis we ignore potential problems associated with voltage rise effect.

19

Note for example that a reduction in reactive import of 2,310MVAr can be achieved for 5GW of capacity of distribution generation, giving a saving of 462MVAr/GW.

20

This assumes a price for reactive energy at 1.2£/MVArh.

21

For example, reduction of reactive import would tend to be more valuable in South of the country, given the magnitude of North to South flows.

48

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 4.74

It is however important to stress that this analysis is only indicative and that more detailed work would be required to establish the impact that various distribution generation growth scenarios may have on DNO reactive imports, the need for reactive procurement by the TSO and how the value of this may be assessed.

49

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

DNO Network security contributions from distributed generators 4.75

As discussed in Section 2, there are likely to be increasing opportunities for distributed generators to provide security contributions for network planning as a consequence of increased numbers of distributed generators connected to distribution networks and the introduction of revised planning standards.

4.76

If a particular area of a DNO network is not ER P2/5 compliant, there are various ways in which compliance can be achieved: •

increase the capacity of existing circuits (e.g. re-conductoring)

•

build additional circuits (reinforcement)

•

use distributed generation to substitute for the capacity of circuits on outage.

4.77

All three options are permitted under ER P2/5. For circuit-based solutions, additional security requirements can easily be determined through simple power flow studies. In the case of generation-based solutions, security contributions can be determined by reference to Table 2 in ER P2/5 although the types of generation contained with this table are limited. To date, the limited numbers of continuously operated distributed generators has restricted the provision of such security contributions and DNOs have largely been reliant upon network infrastructure based solutions.

4.78

The basic principle underpinning ER P2/5 (and ER P2/6) is to assess the time taken to restore group demand following an interruption using Table 1 and to give a capacity credit to any distributed generation connected to the demand point using Table 2. The generation contribution to security is estimated by comparing the reliability of generation with the reliability of an equivalent network using the Expected Energy Not Supplied (EENS) criterion.

4.79

Recently, a methodology was developed to assess the security contributions from a variety of modern distributed generation technologies in order that Table 2 of ER P2/5 could be updated. It is anticipated that this revised Table 2 will be used as a basis for the new planning standard (ER P2/6), which will accommodate more generation types and systems containing multiple generators with

50

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION different unit availabilities and capacities22. A draft of the new ER P2/6 planning standard is undergoing consultation. 4.80

This new planning recommendation could significantly broaden opportunities for DNOs to consider generator contributions to network security in the future. However, as DNO networks should currently comply with ER P2/5, the requirements for security contributions from distributed generators may be limited in the shortterm. In the medium to long term, load growth and asset replacements could increase opportunities for DG to provide network support services.

4.81

The simple example set out in Figure 3 illustrates how a DNO would currently comply with ER P2/5 planning guidance in a 33/11 kV demand group. As can be seen, the group demand of 100 MW can be supplied through either distribution circuit such that if one circuit were to fail, the group demand could be accommodated through the remaining circuit. In this example, it is not necessary to include any security contribution from the generator. Indeed, unless the generator type was specifically recognised in ER P2/5, no security contribution could be allocated. Figure 3 − Example of ER P2/5 compliance without generation contributions

P2/5 Example

100MW

100MW

G 50 MW

Max Group Demand = 100 MW

• Two distribution circuits rated at 100 MW each • One generator with 50 MW export capability • Maximum group demand 100 MW

22

In developing Table 2 of P2/5 and its update P2/6, the generation credit to be given was estimated by comparing the reliability of generation, using expected energy not supplied (EENS) as a criterion, with the reliability of an equivalent network. This was considered to be an acceptable approach when P2/5 was developed in the 1970s.

51

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

4.82

If the group demand in the above example were to subsequently grow to say 105 MW, the supplying network would no longer be ER P2/5 compliant, thus requiring the DNO to seek additional security contributions.

4.83

Figure 4 illustrates two different approaches available to DNOs, which would secure sufficient additional security for network compliance under ER P2/5 and the proposed ER P2/6.

4.84

Under ER P2/5, assuming no generator contribution could be utilised, the DNO could be forced to seek a network solution to provide the necessary security shortfall. In the example below, this could be secured through the installation of a 3rd distribution circuit. The rating of this third circuit would be at the discretion of the DNO but would need to accommodate any further load growth forecast during the life of the assets.

4.85

Under the proposed ER P2/6, it would be possible to recognise the security contribution of the generator. Assuming the availability of the 50 MW generator resulted in a security ‘F’ factor23 of 60%, an overall security contribution of 30 MW could be recognised for network planning purposes. Through the addition of the generator security contribution, it can be seen that the original network becomes compliant and hence the requirement for DNO investment in a network solution is avoided. In such a situation, the inclusion of the generator’s security contribution could be regarded as a valuable service from the generator to the DNO.

23

Ron Allan, Goran Strbac, Predrag Djapic and Keith Jarrett, FES Project DG/CG/0023/00/00, “Developing the P2/6 Methodology”, April 2004.

52

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Figure 4 − Possible solutions to a network security shortfall under P2/5 & P2/6

P2/5 Solution

P2/6 Solution

• Neither circuit can supply group demand alone

• P2/6 recognises generator contribution • Generator contribution limited by ‘F’ factor, e.g. 60%

• Network solution required for continued compliance, e.g. install an additional circuit

100

100

T3

• Up to 30 MW contribution available • Original network adequate 30MW Contribution 100

G

100

G 50MW Export

Group Demand 105 MW

Group Demand 105 MW

4.86

Table 1 of ER P2/5, included as Annex A, defines the minimum demand to be met after first and second circuit outage conditions as well as the timescales in which the specified portion of group demand must be restored and the time in which full group demand should be restored.

4.87

ER P2/5 prescribes maximum times to restore supplies. For demands less than 1MW, this can be the associated repair time of the system. For group demands of up to 12MW, most of the demand should be restored within 3 hours and for group demands between 12MW and 60MW, much of the demand should be restored within 15 minutes. These are therefore prescribed maxima for demand restoration times, i.e. the longest outage times before restoration must take place.

4.88

It is important to note that the capability assessment needs to be done for each of the time periods specified in Table 1 of P2/5. For instance, in the case of Class C, the two time periods of concern are the demand that must be recovered in 15 min and the demand that must be recovered in 3 hr. Both periods must be assessed separately since the required demand and the number of circuits and the amount of generation could be different in each period. Intermittent generation, such as wind, is expected to contribute significantly only in the shorter periods.

4.89

Generation that makes contributions to security needs to become available within these times, i.e. they represent the maximum values of start-up times or reconnection times. Generation needs to be

53

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION persistent for a period of time after these response times during which other means of supply are not available or not fully adequate. These response time requirements will drive infrastructure requirements and further enable generation to contribute to network security. 4.90

The response timescales outlined in Annex A are favourable to many distributed generators, as the restoration times for a substantial portion of group demand are not onerous, exceeding the normal start-up times of conventional standby DG. Hence, the communication infrastructure requirements would not be time critical.

4.91

In addition to substituting for distribution network primary infrastructure, generation could also displace automation and remote control facilities. A typical example of this is when an increase in demand in a specific group causes it to change from a Class B to a Class C group. In such a situation, the presence of generation (could even be intermittent) may allow manual switching to remain instead of replacing this with remote or automatic switching.

4.92

Consider the case of a group demand growing from 12 MW to 15 MW. In this case, following a circuit outage, 3MW of this demand should be restored within 15 minutes, while the rest (12MW) should be restored within 3 hours. Hence, a generation system able to provide 3 MW of effective contribution for 3 hours would make the system ER P2/5 compliant, provided that the remaining network after being switched in can support the group demand in full (15 MW). In this case, manual switching could remain instead of being replaced by remote or automatic switching. This would be particularly applicable for intermittent sources such as wind generation that are able to provide persistent output for relatively short periods of time. Similar logic could be applied to other Classes of group demand24.

4.93

In some situations, restoration activities cannot be completed in one step and a sequence of switching steps is required. In this case the generation can be used to support the group demand and prevent or reduce overloads that may occur during the switching sequence. This would allow earlier restoration of demand before switching is complete without encountering excessive overloads.

4.94

It should also be recognised that not all generator security opportunities will necessarily be related to planning standard compliance. There will inevitably be areas on distribution networks,

24

R Allen, G Strbac, P Djapic, K Jarret, Developing P2/6 Methodologies, April 2004

54

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION which whilst ER P2/5 or ER P2/6 compliant, could remain vulnerable to particular modes of failure, potentially causing disruption to many customers. In such cases, if DNOs perceive the risks associated with loss of supply to be sufficiently high, additional security contributions could be sought from generators, thus taking overall security levels beyond the minimum levels required by planning standards. Indeed, some DNOs are already known to be considering generation-based approaches to enhanced network security. 4.95

One means of valuing the security contributions provided by generators will be for DNOs to compare the costs of generation and network solutions for a particular security problem. In many instances, network solutions involving OHLs are likely to be of lower cost than generation solutions procured solely to resolve a security issue25. This is evaluated further in Section 7.

4.96

A consequence of the lower cost of network solutions will effectively cap the maximum amount a DNO will be prepared to pay for a generation security solution (otherwise there would be no benefit in the DNO selecting a generation solution over a network solution). Such an approach to valuing the contribution of generation to network security problems could thus be based upon the DNO’s avoided cost of network investment.

4.97

Whilst such an approach represents but one valuation methodology, there are a number of commercial and regulatory issues, which must be considered when structuring security contracts. These are more fully evaluated in Section 7.

4.98

Other considerations incentivising DNOs to pursue generator based security solutions could include increasing levels of environmental, planning and/or terrain related constraints. Such constraints could restrict or prevent DNOs from selecting network based approaches and could thus lead to a variety of niche applications for generators.

DNO Quality of Supply Related Services 4.99

As outlined earlier, electricity distribution networks have been designed in accordance with planning standards stipulating certain standards of security. Security has been traditionally graded according to the total amount of power that can be lost and networks have been specified according to the principle that the greater the amount of demand that can be lost, the shorter the recommended

25

It should be noted that there will be other revenue opportunities available to generators, e.g. energy sales

55

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION restoration time (as specified in ER P2/5). This implicitly determines the level of assets required to comply with minimum security standards. Such assets are usually in the form of OHLs, cables, switchgear and transformers but compliance can also be achieved through the use of network automation or operational support services. 4.100 As planning standards do not require redundancy of assets on low voltage (LV) networks, the duration of outages caused by LV faults is determined by component repair and replacement times. Medium Voltage (MV) networks are generally configured such that interruptions caused by single faults can be restored much more quickly, e.g. through switching. 4.101 These network design practices have effectively determined the characteristics regarding quality of service as experienced by end customers. The impact of these design practices is reflected in the network performance statistics recorded under Ofgem’s Information and Incentives Project (IIP) as shown below. Figure 5 summarises the distribution of Customer Interruption (CI) and Customer Minutes Lost (CML) according to voltage level. Figure 5 − Distributions of CIs and CMLs according to distribution voltage level

Customer Interruptions (CI)

Customer Minutes Lost (CML)

Source: Ofgem IIP statistics

4.102 Clearly, the performance of medium and low voltage networks has a dominant effect on the overall quality of service. The vast majority of CIs (85%) and CMLs (93%), have their cause in LV and MV networks. In GB, the average number of CIs is approximately 85 for every 100 customer each year. For CMLs, the average figure is approximately 85

56

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION minutes per annum. These statistics are primarily driven by the radial design of these networks.

IIP Opportunities for DG 4.103 An important issue for consideration is whether the interactions between historic network planning philosophies and Ofgem’s quality of service incentive arrangements will present opportunities for DG to provide new network support services in future. The network performance statistics above demonstrate that there could be significant opportunities for DG to improve service quality on 11kV and 0.4 kV networks, given the contribution of these networks to the overall system performance. 4.104 However, for DG to improve service quality on these networks, the generation must also be connected at 11kV or 0.4kV, i.e. “below” the point of fault. For example, supply interruptions caused by faults in 0.4 kV networks cannot be readily impacted by generation connected at or above 11kV. The requirement for DG to be connected at relatively low voltage levels in order to impact upon service quality could restrict such opportunities to relatively small sized generation. 4.105 DG could, in principle, reduce the impact of outages in networks operating at or above the voltage level that a particular generator is connected should it be possible to operate locally islanded systems. This may represent an interesting opportunity for DG, particularly if DG can offer a cost effective solution to support local load growth and/or substitute for replacement of upstream assets. This is schematically presented in Figure 6, which shows that DG connected at LV networks, close to the local load, would have, a higher value in the context of offering network support and performance related services26.

26

T Bopp, G Strbac, R Allen, Economic evaluation, of islanded operation, Power Engineer, 2004.

57

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Figure 6 − (a) Traditional network layout (b) Potential for DG to provide network services will depend on point of connection

(a)

(b)

4.106 For the purpose of this section it is interesting to derive some indicative figures that quantify the potential value of islanded operation. This in turn gives indications regarding the level of expenditure that can be justified to enable islanded operation. 4.107 Some of the established indices used to measure the quality of service, as seen by the end customers, approximately address the frequency and duration of outages27. The average number of interruptions per customer per year and the total annual duration of interruptions per individual customer are now used as the indicators of performance of electricity systems in the regulatory context.

Examining the value of DG contributions to reduced CMLs 4.108 The system shown in Figure 7 (a) has two circuits or transformer feeders (CT) supplying a demand of 50MW. Assuming the system supplies a residential area with an average peak demand of 2 kW per household, this would correspond to 25,000 customers. Each feeder is assumed to be rated at 50MW, and have an availability of 99%. Figure 7 − Comparison of CML performance for two networks

50 MW 27

50

50

50

50

R Billinton and R Allan, ‘Reliability evaluation of Power Systems’, Plenum 1984.

58

Group Demand 50 MW

Group Demand 50 MW

G

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

4.109 The system shown in Figure (b) has the same circuits or transformers together with a generator (G) having an output capacity of 50MW, an availability of 40% that is significantly lower than the availability of the circuits. 4.110 Let us now compare CML performance of these two systems. Customers in system (a) will be interrupted only when both infeeds fail simultaneously. The probability of this happing is 0.01x0.01 = 0.0001. The average duration of this event is therefore 0.0001x8760 = 0.876 hours, or 52 min per annum. Given that this affects all connected customers CML for this load group will be 52min/year. 4.111 Assuming now that the generator in system (b) can operate in an island mode, there will be 40% chance that the generator will be available in the situation when both of these transformers are out of service. Hence, CML in system (b) will be 0.6x52 = 31 min/year, achieving a savings of 21min/year, for 25,000 customers. This is clearly a massive reduction in CMLs experienced by these customers. It is important to note that a generator with relatively poor availability can make such a significant contribution to CMLs. 4.112 Assuming further that the DNO has 2,500,000 customers, the overall reduction of CMLs at the company level will be 0.21 min/year. Assuming a value £150,000 for each CML saved, the value of this reduction in CMLs will be in the order of £31,500. For a generator of 50MW, this would amount to £0.63/kW/annum. However, this would be relatively small part of the required income necessary to cover investment cost. If we assume that the plant would need to earn £45/kW/year (typical for OCGT) to cover its capital cost, for a 50MW generator this would amount to £2.25m per year. If the generator were renewable, operating at 40% load factor, ROC income (at £50/MWh) would amount to approximately £9m.

59

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 4.113 As discussed earlier, the reliability performance of the whole distribution system tends to be dominated by the performance of 11 kV and 0.4 kV networks. The relative contribution that a generator can make to reducing CMLs could be generally significant (in the above example the contribution was 40%), it is likely that only those generators connected to 11 kV (and 0.4 kV) would significantly contribute to the overall reduction in CMLs in absolute terms. This however would require the generators to be able to ‘ride’ through 11 kV network faults or to be quickly reconnected after such faults and would require the ability of operating in islanding mode.

Examining the value of islanding operation 4.114 CI & CML statistics provide indications of the scope for potential improvements in service quality; an average customer in the UK experiences less than one outage per year lasting for under 1.5 hours. The monetary value of service quality improvements can be investigated using the concept of interruption costs28. Such valuations take into account the frequency and duration of supply interruptions but also include the cost characteristics of specific customer classes (domestic, commercial and industrial etc.). Typically, such costs are normalised in £/kW of peak demand or £/kWh of annual energy consumption according to class of customer. 4.115 The present regulatory incentive arrangement is based on the number of customers affected by a fault rather than on the type of customer affected and thus does not recognise the different values that different classes of customers attach to service quality. Adopting a different but complementary approach, Ofgem has recently conducted a survey29 regarding the value that customers attach to service quality. 4.116 The analysis showed that the maximum annual benefit of an uninterrupted transition into islanding mode was approximately £1.4/kW and £19/kW30 for average residential and commercial customers31 respectively. Based on these values, the level of efficient expenditure that could be allocated towards enabling islanded

28

K K Kariuki, R N Allan, “Assessment of customer outage costs due to electric service interruptions: residential sector”, IEE Proceedings on Generation, Transmission and Distribution, Vol. 143, No. 2, March 1996.

29

Consumer Expectations of DNOs and Willingness to Pay for Improvements in Service, Ofgem, June 2004.

30

This also indicates the value of the services that could be available to DG.

31

T Bopp, G Strbac, R Allan, “Economic evaluation of islanding operation”, Power Engineer, 2004.

60

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION operation can then be estimated. Considering typical values for asset lifetime and discount rates, the present value of the potential benefits of islanding would amount to about £15/kW for residential and £200/kW for commercial customers. Assuming the average of peak diversified demand for domestic customers to be 2 kW, expenditure of around £30 per household would be justifiable to upgrade the LV system for islanding operation32. Assuming that a LV network (that can be run in an islanding mode) supplies a residential area of 500 households with a peak demand of 1MW, the level of justifiable expenditure for islanding operation would amount to approximately £15,000. On the other hand, 1MW of commercial demand would support an investment of around £200,000 to enable islanding33. It should be noted that these values are averages; the permissible expenditure can be higher in network areas with a lower quality of service and reliable DG, and vice versa.

Demand for islanding services 4.117 A number of industrial and commercial customers have installed standby generation to increase reliability of supplies to high value loads. It is unlikely that DG technologies that are relevant to achieving Government targets would be able to effectively provide such services. We therefore focus our discussion on the possible application of new, energy production driven DG installations in the provision of islanding services on LV networks of DNOs. 4.118 The most attractive target areas for extracting benefits of islanded operation would include those networks delivering relatively poor service quality to high value loads and/or large numbers of customers. Generally, however, the number of such areas will be limited, as under IIP, DNOs are investing in their local networks. It should be noted that the current regulatory framework incentivises DNOs to invest in network areas providing relatively poor performance to large numbers of customers. 4.119 Clearly, urban networks represent areas with relatively high load densities with large numbers of customers although it should be recognised that these tend to be supplied from interconnected underground networks rather than OHLs. As cable networks experience fewer faults than OHL networks, the scope for network support services is diminished in urban areas. Conversely, rural

32

These figures also indicate the value of the services that could be available to DG for providing the service (ignoring cost associated with making islanding possible).

33

It should be noted that a recent survey carried out by Ofgem indicated that customers value service quality more than expected.

61

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION networks tend to experience a larger number of interruptions. Although the cost of outages to individual customers supplied by rural network may be higher than in urban areas, the fact that rural networks tend to supply relatively low load densities with fewer consumers, the overall value of interruptions may not be sufficient to warrant islanding operation. 4.120 The recent survey conducted for Ofgem34 indicates that all customers value the speed of reconnection of power supply after major storms. Domestic consumers appear to have high expectations regarding quality of service. Although most consumers believe it is reasonable for a power cut to occur in major storm, two thirds believe that distributors should be doing more to reduce the impact of severe weather on their networks. Consumers expect power to be restored quickly; only 12% of customers expect restoration times to exceed 24 hours and nearly half expect power to be restored within 3 hours. This is an area in which DG could potentially offer services in the longer term. More work would be required to examine and quantify the potential benefits in this area. 4.121 This suggests that, in the short to medium term, the demand for islanding services from DG will be limited. Of course, there will be specific cases in which islanding operation may be a technically and commercially viable option. Also, it should be recognised that the demand for higher quality of supply is only likely to increase in the longer term. This issue is being explored elsewhere under the Technical Steering Group Workstream 5 programme.

Technical issues associated with management of islanded networks 4.122 Notwithstanding the above, cost effectively managing the operation of islanded operation on LV and MV networks is a challenging task. Clearly, frequency and voltage control in islanded networks will require local generators to be adequately equipped for these tasks with appropriate governors and Automatic Voltage Regulators. Managing the balance between demand and generation would be particularly challenging given the relatively large mismatches that would need to be accommodated. The management of demand and supply on islanding networks would probably need to include load management and may also require new technologies such as storage, power electronic based compensation and voltage regulators.

34

Accent Marketing & Research, Consumer Expectations of DNOs and Willingness to Pay for Improvements in Service, June 2004, 145f/04 (Report to Ofgem).

62

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 4.123 Another technical issue to be resolved is the close coupling between active and reactive power in LV and MV network. Given the relatively large active to reactive impedance ratios frequency and voltage control is complicated. Solutions established for the transmission network, based on decoupled reactive power-voltage and active power-phase angle would not be readily applicable and would require modifications. 4.124 Re-synchronisation equipment will be required to enable continued operation of islanded systems. Adequate protection schemes will need to be developed to ensure safe operation of the network under fault conditions. This may be a particular problem in networks with inverter-connected generators. 4.125 Also, it should be recognised that currently, the communication, metering and on-line monitoring & control infrastructure that would be necessary to support island operation is least developed in LV and MV networks. 4.126 The technical issues associated with the operation of highly distributed power systems, such as micro grids, are the subject of intensive research and development, supported by various national and international research programmes and new solutions are likely to emerge in future.

Existing experience 4.127 There are a number of examples of autonomous power systems. There is a significant amount of standby generation installed on a variety of industrial and commercial sites. As already discussed, this generation is installed to increase reliability of supply to high value loads where the public electricity supply system was not seen to be sufficiently reliable. Generally, in the case of failure in the main supply system, standby generation is configured to run in various modes when disconnected from the system, e.g. to supply only essential services or entire site demands. This demonstrates that the technical solutions adopted, and the levels of redundancy and automation specified, will vary according to the value of load that is being secured. 4.128 Also, there are a number of solutions adopted for small islands but these are not likely to be directly transferable to a large-scale implementation in existing distribution networks. A direct example of islanded power systems supplying very high value loads are found on ships although such approaches are likely to be cost prohibitive for a large scale implementation within distribution networks.

63

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 4.129 Whilst islanding may represent an interesting opportunity for DG in the long term, it is not very likely it will present a significant or lucrative market for DG in the short or medium term. Niche applications may emerge and further work will be required if this issue is to be explored in more detail.

DNO Voltage control services and Power Flow Management 4.130 Voltage variations outside statutory limits and overload conditions on distribution networks require immediate DNO remedial action. Contributions from distributed generation could also potentially remedy such situations, thus avoiding requirements to reinforce networks. Similarly, services from distributed generation relating to reactive power flow management could also maximise the capacity of distribution circuits, again avoiding requirements for network investment. This section explores the opportunities for distributed generation to provide such services.

Voltage management services 4.131 Two opportunities for DG to provide voltage support services in distribution systems are explored. One relates to a low voltage scenario at the end of a feeder as a result of load growth as illustrated in Figure 8. The other opportunity relates to a low voltage condition arising from circuits being removed from service. Figure 12 illustrates this second scenario, where the voltage at point A drops below statutory limits after the section of feeder between the 11 kV substation-1 and point A has been taken out of service and the Normally Open Point (NOP) has been closed. Both voltage support opportunities are evaluated further below.

Scenario 1: Use of DG to support voltage in load growth situations 4.132 If, as a result of load growth, load points D5 & D6 in Figure 8, begin to experience low voltages outside statuary limits, there are number of options available to the DNO to resolve the low voltage problem. An obvious solution is to reinforce the system. Another solution would be to use DG to inject power either at an intermediate point on the feeder or at the end of the feeder. Yet another solution would be to install a voltage regulator somewhere along the feeder or install a capacitor (see Figure 9 and Figure 10).

64

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Figure 8 − Use of DG for voltage support at the end of the feeder due to load growth 11 kV Primary substation

11 kV feeder with Tee offs B

A

D1

D2

D3

D4

C

D5

D6

Loads Distributed generator

4.133 The availability of DG will be the main issue of concern when mitigating voltage problems using this method. For this reason DG based on intermittent technologies could not be considered due its low local availability. Even a non-intermittent generator with an availability of 85% would be considered unacceptable. In reality however, voltage violations would only occur at times of peak demand (see Figure 11). From Figure 11, the DG would therefore only be required to support voltage at peak times from t1 to t2 and from t3 to t4. As these periods are relatively short, the probability of voltage violations would be reduced despite the “low – 85%” availability of DG. In the case of a DFIG based wind farm, VAR support will generally be independent from the active power output. Given that wind farms are rarely shut down completely, the availability of reactive power will be relatively high, and DNO could potentially utilise this support for voltage control purposes. 4.134 DNOs prefer to use solutions with availability that is close to 100% primarily because the statutory limits on voltage are deterministic. It is conceivable that opportunities for DG to provide voltage support in distribution networks could be enhanced considerably if the European Voltage Standard (EN-50160)35, describing voltage limits in statistical terms, were used to assess DNO compliance with statutory obligations on voltage limits. Because EN-50160 specifies how often and by how much voltage levels may be exceeded, it would not be

35

BS-EN-50160, “British Standard on Voltage characteristics of electricity supplied by public distribution systems”, 1995.

65

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION necessary to work on the basis of a ‘zero tolerance’ for voltage excursions as at present. Such voltage limits would provide DNOs with more discretion regarding the adoption of DG based solutions to manage network constraints/risks. Figure 9 − Use of voltage regulator to mitigate voltage violation at the end of the feeder due to load growth 11 kV Primary substation

11 kV feeder with Tee offs

B

C

A

D1

D2

D3

D4

Voltage regulator

D5

D6

Loads

Figure 10 − Use of capacitive compensation to support voltage 11 kV Primary substation

11 kV feeder with Tee offs C

B

A

D1

D2

D3

D4

Loads

D5

D6

Capacitive compensation

4.135 Use of capacitors to support voltage on distribution networks, as shown in Figure 10, is not commonplace in GB. There are two main reasons for this. The first one is that use of capacitive compensation is not as effective on distribution system as is the case on transmission networks due to the higher Resistance: Reactance ratio on distribution networks. The second is that distribution switchgear is often not well suited to switching of capacitive currents and there may be potential for resonance phenomena to occur. 4.136 It should be recognised that the cost of capacitive compensation, at around £20,000 per MVAr, is very low compared to that of a generation based solution.

66

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Load (MW)

Figure 11 − Use of DG to support voltage at peak times

t1

t2

t3

t4

Ti me (hr)

4.137 It can be concluded that with current UK voltage standards, the opportunities for DG to provide voltage support in cases where creeping load growth causes voltage problems on radial feeders are limited. The opportunities could be improved if there are many independent generators in a locality that can collectively provide a higher level of availability comparable to network solutions such as voltage regulators or capacitors. 4.138 If DG could be used to support voltage, the cost of the service must be compared with cost of providing the same service from a voltage regulator. As the typical cost of an 11kV voltage regulator is in the region of £20,000/MVAr, the annual cost of such a device is roughly £1500 assuming a 40-year life and a discount rate of 7.5%. Therefore the annual cost of voltage support service provided by DG would have to be equal to or below £1500 per annum36. 4.139 It is important to point out that in this application DG would not itself require an AVR (automatic voltage regulator) to support voltage, as the injection of active power along or at the end of a feeder would cause the voltage to rise. In other words use of DG in this mode takes advantage of the classic voltage rise effect caused by the output from the distributed generator. 4.140 As the service is non-time critical, the infrastructure requirements could be minimal consisting simply of telephone communication and half hourly metering. Supporting commercial arrangements would set out the contractual obligations including performance

36

(£1.50/kW assuming a 1 MVA generator).

67

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION requirements, proof of delivery and perhaps penalties for nondelivery. 4.141 Under all circumstances it would be expected that the DNO would remain accountable for voltage performance and retain liability for any breaches of statutory obligations.

Scenario 2: Use of DG to support voltage on outage of part of the network 4.142 As indicated earlier, under the scenario illustrated in Figure 12, Point A experiences a voltage drop when the section of feeder from 11kV primary substation -1 is removed from service and the loads on this feeder are supplied from substation -2 by closing the normally open point. The voltage problem could be resolved by injecting power from DG as shown below. Alternatively a voltage regulator could be installed and used to boost the voltage. Figure 12 − Options for solving voltage dip problem at point A (assuming outage section between substation 1 and A followed by closure of normally open point) 11 kV Primary substation - 1

11 kV Primary substation - 2

Voltage drop

A Loads

Normally open point

B Voltage regulator

Alte r solu native tion s Distributed generator

Power flow management services 4.143 Two similar scenarios could also arise relating to the overloading of a section of circuit in each of the above examples.

68

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Scenario 1: Overloaded section of feeder due to load growth 4.144 This section explores the opportunities for DG to provide network flow management services to DNOs, to relieve overloads in sections of a circuit. In the example given in Figure 13, section A of the 11kV feeder gets overloaded due to load growth. The overload can be relieved by either installing a new cable, shown as a dotted line, or by constructing a new 11 kV in-feed at B. Injecting power from the DG somewhere along the feeder at C could also relieve the overload. Figure 13 − Use of DG to mitigate partial circuit overload due to load growth Overloaded Section

11 kV Primary substation

B New 11 kV in feed Normally open point

A C

D1

D2

D3

D4

D5

D6

Loads Distributed generator

4.145 Use of DG for this type of application raises the same issues of availability as discussed under voltage control services. Therefore only non-intermittent DG could provide this service. The service would be required mostly at peak times. The value of the flow control provided by DG would be determined by assessing alternative network solutions as stated above. Infrastructure requirements would consist mainly of half-hourly metering for the generator output and a means of instructing the generator to start running. The latter could be achieved through a telephone call or the machines could be started remotely. Remote control would require appropriate communication. Existing SCADA based communication would suffice, thus avoiding any further infrastructure investment requirements.

Scenario 2: Relieving overloads following closure of Normally Open Points 4.146 Following an outage on a section of an 11kV circuit between Point A and substation 1 (see Figure 14), supply to loads downstream of Point A can be restored by closing the normally open point, thus enabling power to flow from substation 2 via the feeder connected to this substation. 69

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 4.147 Suppose the two feeders are rated at 4 MW each and each circuit carries a load of 3 MW at peak. When the Normally Open Point (NOP) is closed after taking out the first section of the feeder from substation 1, the loading on the feeder from substation 2 increases to 6MW, exceeding its rated capacity at peak times. This is also a security of supply issue that would have to be assessed in the context of ER P2/5 (or P2/6). In this case, the overload could not be resolved by load shedding 1 MW, as required for a class B group demand, since the remaining load of 5MW would still exceed the rating of the cable (4MW). The options open to the DNO to resolve this problem would include network reinforcement or contracting a security service from the DG. 4.148 As in the case of voltage control services, availability of DG is the main issue when assessing the use of DG to provide what is effectively a network security service. Since overload problems occur at peak times, DG availability is only critical at these times. Again, as in the case of voltage control services, only non-intermittent DG could reliably provide this service. 4.149 The infrastructure requirements would be largely determined by the size of group demand, as the primary consideration when determining the required speed of supply restoration. Most feeders at 11kV belong to either class A (group demands up to 1MW) or class B (group demand over 1MW to 12MW). Class A outages are non-critical as group demand must be restored according to repair times. Class B outages have more potential as it is stipulated that group demand minus 1MW must be restored in three hours. Any generation contracted to provide the security service must be either running or, if on standby, must be capable of starting in three hours. This is a fairly long timescale, which most non-intermittent sources could satisfy readily.

70

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Figure 14 − Use of DG to mitigate circuit overload following closure of NOP 11 kV Primary substation - 1

11 kV Primary substation - 2 3-4 MW

3-4 MW

A Loads

Normally open point

B

Distributed generator

4.150 Because of the long timescale, the infrastructure required to facilitate such services are not onerous. A detailed description of the required infrastructure as well as the means of valuing the service is given in the section on security services. 4.151 These voltage control and flow management problems are essentially network planning related issues as they relate to supply restoration times following 11 kV circuit failure between the substation-1 to point A. If no measures are taken to restore the circuits within planning standard timescales (ER P2/5 or P2/6), then the DNO would effectively be non-compliant for the group demand. The problem should therefore be dealt with in the context of planning standards. Depending on the group demand, the DNO could comply with P2/6 by shedding 1MW of load. The full group demand would then be restored in the repair time of the failed section of cable. 4.152 Because of the drive to reduce CMLs and CIs, DNOs have made considerable investments in 11kV and 0.4kV networks. A direct result of this investment is that distribution networks in GB are generally “over compliant” with planning and security standards. In general, it will be some time before the investment in MV and LV networks is outstripped by load growth resulting in non-compliance with security standards. For the foreseeable future, the scope for DG to provide security related ancillary services in such situations could be limited,

71

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION however, niche opportunities could emerge for DG to provide such services.

Reactive Power Management 4.153 The maximum transport distance of distribution circuits for given load characteristics is usually driven by the maximum voltage drop constraints. This is likely to be relevant predominantly to 11 kV networks, particularly rural overhead circuits. For 33kV and 132kV circuit this is usually not a significant issue. 4.154 Supplying loads with poor power factors can make the situation more severe. In order to illustrate this we consider an 11 kV OHL and an Underground Cable (UC), with the following per-kilometre characteristics: OHL r = 0.36 Ohms/km, x = 0.4 Ohms/km (max current ~ 300A) UC

r = 0.36 Ohms/km, x = 0.09 Ohms/km (max current ~ 220A)

4.155 Assuming a load of 4MW (lumped at the end of the circuit or evenly distributed along the circuit) and allowable voltage drop of 5%, the maximum transport distance is calculated for various distributions of the load along the circuit and various power factors. This is presented in Table 12.

72

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Table 12 − Maximum distances of the OHL and UC circuits

Power Factor Location of load

1

0.9

0.8

Lumped Distributed Lumped Distributed Lumped Distributed

Maximum distance for OHL (km)

8

16

5.2

10.4

4.3

8.6

Maximum distance for UC (km)

8

16

7.1

14.2

6.7

13.4

4.156 From the table we observe that transporting significant amounts of reactive power (to loads with poor power factor) through overhead distribution circuits may considerably reduce the distance over which the power can be transported. The distance over which the power can be transported can drop to 50% for 0.8 power factor, and hence may adversely affect circuit utilisation. For underground cables, this is significantly less important (this is driven by considerably lower reactance of cables in comparison to overhead lines). 4.157 In the case of a voltage-limited circuit, which transports significant reactive power, it may be beneficial to compensate for the reactive power, introduce in line voltage regulators or to reinforce the circuit. Compensation or voltage regulator based solution is likely to be more cost efficient, given the cost of these alternative solutions. 4.158 Furthermore, transport of reactive power occupies useful capacity of distribution circuits and may limit the amount of active power that can be transported. Table 13 illustrates this effect, showing the amount of active power (as % of the circuit capacity) that can be transported for various power factors. Table 13 − Active power (as a % of the circuit capacity) that can be transported

Power Factor

1

0.9

0.8

Active power transportable

100%

90%

80%

73

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 4.159 This may be also presented by the increase in demand for capacity of the plant supplying load with various power factors, as shown in Table 14. Table 14 − Additional plant capacity required to transport reactive power

Power Factor

1

0.9

0.8

Required increase in capacity

0%

10%

25%

4.160 From this simple analysis we observe that the power that can be transported (as % of the MVA capacity of the circuit) is directly proportional to the power factor of the load supplied and that the need to transport reactive power considerably increases the capacity requirements of network infrastructure. 4.161 Distributed generation connected close to load could supply some of the reactive needs of the load and hence increase the transport capabilities of the existing circuits. The value of this service would be limited by the cost of compensation equipment. However, it should be noted that the question of availability of this support would need to be relatively high for DNO to consider DG contribution in this context.

74

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

5.

PROSPECTS FOR DISTRIBUTED GENERATION

5.1

The impact of DG on system operation and development, including opportunities for DG to provide ancillary services, will depend on levels of penetration, the technologies employed and the point of connection within the distribution network.

5.2

Possible development scenarios could include various combinations of: •

large off-shore wind farms, perhaps over 100MW, connected to 132kV distribution or transmission networks;

•

medium size distributed generators (e.g. on-shore wind, biomass, CHP, with sizes ranging from 300 kW to about 50MW, connected to 11kV and 33kV distribution networks; and

•

micro generation connected to 0.4kV network. This is schematically presented in Figure 15.

Figure 15 − Schematic diagram of the power system with various forms of DG technologies connected to distribution networks

Central Generation Transmission HV Distribution

Large DG (off-shore wind)

MV Distribution

Medium size DG (on shore wind, biomass, CHP)

LV Distribution

Micro Generation (µ CHP, PV)

Demand

5.3

For example, large off shore wind farms connected at high voltage distribution networks (or directly connected to transmission networks) will have different impacts on system operation, than will micro generation or biomass. The connection of intermittent and unpredictable wind generation will increase the demand for 75

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION frequency regulation and reserve services significantly more than micro generation or biomass. Location and the connection voltage of DG will also be very important. For example, generation connected to low and medium voltage levels will have more opportunity to provide local network services than generation connected at higher voltage distribution network, e.g. avoidance of distribution asset replacement. 5.4

This section outlines the capabilities of existing and near commercial distributed generation technologies to provide different ancillary services described in section 3. The technologies evaluated include: •

Wind Power (including the latest doubly fed induction generators);

•

Biomass, Land Fill Gas and small hydro-electric schemes;

•

Large conventional CCGT power stations, > 100 MW;

•

Large CHP installations, >30 MW;

•

Micro CHP installations, typically rated between 1 – 3 kW; and

•

Standby Generation, typically rated between 200 kW & 50 MW.

5.5

In addition to the above list of thermal and renewable technologies, consideration was also given to solar, tidal and wave power although in depth analyses was not undertaken due to uncertainties regarding technical specifications and capabilities.

5.6

To recall, the most technically feasible ancillary services outlined in Section 3 were: •

TSO Frequency response;

•

TSO Regulating and standing reserve; and,

•

DNO Security of Supply contributions.

5.7

As the operating regimes for the different listed types of distributed generation vary considerably, the list of technologies has been segregated into renewable and non-renewable sources. It is assumed that the incentives for renewable generators will remain consistent over the coming years, i.e. to maximise outputs in the pursuit of ROC revenues unless more attractive revenue opportunities can be identified.

5.8

This section focuses on the new services which could be provided by each technology type. The capabilities and opportunities for both renewable and non-renewable distributed technologies are summarised in Table 16 and Table 17 at the end of this Section.

76

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Renewable Distributed Generation Technologies Wind Power TSO Frequency Response 5.9

Wind generation technology has historically been based on simple fixed speed induction generators with very little control over the dynamic performance of the generator (passive stall turbines). Over recent years, significant progress has been made regarding the development of active stall and pitch regulated variable speed wind turbines. This development is important, both in terms of increased turbine efficiency but also control capabilities.

5.10

The latest wind turbine technology is in principle capable of providing frequency response by controlling the electrical power output. Using blade angle control wind turbines should be able to easily provide high frequency response, by reducing its output in response to an increase in frequency.

5.11

Regarding the provision of low frequency response, a part-loaded wind farm should be able to rapidly increase its power output to the maximum when system frequency falls and then back-off as system frequency recovers. Although this does not appear to be technically challenging, this service is expected to be required only very occasionally. Clearly, in order for a wind farm to provide low frequency response it would need to run part-loaded. This implies that some of the available wind energy would not be utilized.

5.12

However, with large penetrations of wind generation, there will be occasions (generally during low demand days over summer) when the number of conventional units needed to supply the remaining load will be so few, that adequate levels of response and reserve may not be possible to maintain37. In extreme situations available renewable generation may even exceed the demand during some periods. These conditions would generally occur during the periods of low demand coinciding with (very) high output of wind generation. When all available options are exhausted and the amount of conventional plant on the system is still insufficient to provide adequate system support services, wind generation could be constrained off (de-loaded) and then instructed to take part in frequency regulation tasks, as the available wind power cannot be absorbed by the system. The opportunity cost of this operation will need to include losses of ROC revenue.

37

C Chen, G Strbac, X P Zhang, "Evaluating the impact of plant mix on frequency regulation requirements”, UPEC 2000, Belfast, Sept 2000.

77

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 5.13

Under NETA (and subsequently BETTA) if National Grid requires plant to operate at a lower level in order to provide frequency response, the change in output is instructed by accepting the bid in the BM (the bid price is declared by the generator). This market mechanism should ensure that the economics of renewable energy production are reflected in the scheduling of frequency response.

5.14

Although wind generation is not expected to be regularly instructed to provide low frequency regulation service, the costs associated with the necessary wind turbine upgrades appear minimal and hence may be justified.

Large wind farms and Grid Code reviews 5.15

Although the penetration of intermittent renewable resources and other forms of distributed generation may displace significant amount of energy produced by large conventional plant, there are concerns regarding system security relating to the ability of such new generation technologies to withstand various network disturbances and to provide adequate system support services.

5.16

Wind generation uses different technology to conventional plant and generally (at the moment), is not able to provide a similar spectrum of support services to the system. At the relatively small levels of penetration this can usually be tolerated. However, operating the system with large amount of such plant could pose major challenges in terms of sustaining system integrity.

5.17

Given the anticipated levels of wind generation deployment in the UK, TSOs have recently set out a proposal that specifies requirements for connecting of wind generation to the transmission network and these are detailed in the Grid Code38. In a number of other countries, Grid Codes are being reviewed to reflect the trend of increased levels of wind generation.

5.18

The proposed Grid Code requires wind generation to be capable of operating continuously over the full frequency range between 47.5 Hz and 52 Hz, while maintaining constant active power output between 49.5 Hz and 50.5 Hz. In addition, between 47 Hz and 49.5 Hz, their active power output should not reduce more than pro rata with system frequency.

5.19

The provision of primary, secondary and high frequency response capability is also required on large wind farms. Furthermore, wind

38

Grid Code consultation document, June 2004 (http://www.nationalgridinfo.co.uk/grid_code).

78

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION farms are required to have a steady state reactive power capability of 0.95 power factor leading to 0.95 power factor lagging at rated MW output at the point of connection. Wind farms will be subject to reactive power dispatch across a range of loading conditions. This will facilitate wind generation taking an active role in network voltage control. 5.20

In addition to frequency and voltage control requirements, one other key topic relates to the ability of wind generation to continue operating during transmission network faults, to avoid the widespread tripping of wind farms and losses of substantial amounts of active power generation. This is known as the fault ride-through (FRT) capability.

5.21

For the duration of a fault on the transmission network, the voltages on the faulted phases are assumed to be zero at the point of fault. Considering the relatively low transmission circuit impedances, such fault conditions can cause a large transient voltage depression across wide network areas. Conventional synchronous generators are expected to trip only if a permanent fault occurs on the circuit they are directly connected to. However, other electrically nearby generators that are connected to healthy circuits will remain connected and stable after the faulted circuits are disconnected. At present the transmission system is operated to withstand a maximum instantaneous in-feed loss of 1320 MW (Sizewell B).

5.22

However, if generation connected elsewhere on the transmission network would not remain stable (or would potentially disconnect) during or after the fault is cleared, such generation could exacerbate the original fault conditions and more than 1320MW of generation could be lost from the system.

5.23

A study has recently been carried out recently to estimate the additional system cost that would need to be incurred in order to accommodate wind generation with various degrees of robustness to withstand faults on the UK transmission network39. The findings suggest that the system cost of holding additional response could increase considerably for wind penetrations above 10GW. Furthermore, given the indicative cost of providing fault ride through, it would be cost efficient to invest in equipment and solutions necessary to enable wind generators to ride-through faults.

39

T Bopp, G Strbac, Value of fault ride through capability of wind farms in the UK, a report to the DTI, August 2004.

79

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Contribution of wind generation to network security 5.24

For wind generation to provide security, the output must remain at or above a certain required level for a minimum period of time, designated as Tm. This persistence time has a considerable impact on the capability that can be associated with wind generation and is related to the duration of the system conditions for which such generation may be able to avoid or reduce customer disconnections. There are three distinct system conditions, each of which can be associated with different minimum persistence times. Table 15 contains recommended values for Tm associated with the critical switching, repair and maintenance activities detailed in ER P2/5.

5.25

Wind generation connected to a demand group could be used to support loads during periods whilst transfer capacity is being switched. In such circumstances, Tm can be conceived to be the period between the maximum restoration time specified in ER P2/5 for the Class of Supply being considered and the point in time when switching of the transfer capacity is completed. The generation should be considered as contributing to security whilst this switching takes place. The associated value of Tm is likely to be relatively short. Manual switching could take up to 3 hours at 11 kV, while for 33kV systems with remote switching this may take up to 15 minutes.

80

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Table 15 − Recommended40 values for Minimum Persistence Times (Tm) for Switching Actions

ER P2/6 Group Demand

Switching

Repair

Maintenance

A

n/a

n/a

n/a

B

24 hours

24 hours

2 hours

C

3 hours

5 days

18 hours

D

3 hours

15 days

24 hours

E

3 hours

90 days

24 hours

5.26

Consequently, the capability of wind generation to contribute towards system security will be greater in systems with automated switching compared with those in which manual switching is utilised. In a similar way to the above, wind generation may be sufficient to provide immediate support to a group following a non-damaged first circuit outage, given that the likely restoration time will be of the order of 15 - 30 minutes in most circumstances. In this case, the minimum persistence time Tm is equivalent to the restoration time.

5.27

This situation would typically occur in instances where the group demand increased above 12MW. Instead of automating, manual switching could be retained and intermittent DG relied on for the time period between 15 min and up to 3 hr. Therefore, the worst-case scenario is to adopt a value of 3 hours for Tm.

Minimum persistence times for repair activities 5.28

For situations where, after a first circuit outage (caused by a damaged circuit), sustained support is required until the faulted circuit is repaired, the contribution of intermittent wind generation may be limited. In such cases, the duration of repairs could last for a number of days, and this would then be the period for which security support is required. However, contribution of wind generation rapidly reduces with increasing persistence times and hence this is not likely to be a significant opportunity for wind.

40

These times were recommended by Work Stream 3 of the Technical Steering Group of DGCG.

81

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Minimum persistence times for maintenance outages 5.29

ER P2/5 considers that a circuit outage may be due to planned maintenance. In emergencies, these activities can frequently be closed down within a short period of time. Consider a circuit that is on a scheduled planned maintenance outage. This circuit is assumed to be in a group of circuits that, together with one or more intermittent sources, supply a group demand. The situation to be considered is the support of this group demand following a damaged fault on one of the remaining circuits. Under this situation, any intermittent generation connected to the group could contribute to the supply until the circuit on maintenance is put back into service, assuming this restoration time is less than the likely repair or restoration time of the failed circuit. Consequently the minimum persistence time Tm for which generation may need to provide support is related to the time it takes to complete an urgent return to service on the circuit undergoing planned maintenance.

Wind Power summary 5.30

Depending upon the size, specification and operating regime of wind turbines, it may be possible to source a limited number of ancillary services from wind farms in the future. These include: •

TSO Frequency response;

•

TSO reactive power; and

•

DNO network security services.

5.31

As operational uncertainties apply regarding the ability of wind farms to deliver these services, it must be assumed that opportunities, especially in the short-term, will be limited.

5.32

In order to determine the ancillary service capabilities of a wind farm, a thorough technical and operational analysis should be undertaken for each individual site.

Biomass 5.33

Biomass is an example of a base-load, predictable, renewable generation technologies, of medium size. Compared to intermittent wind, integration of biomass plant is likely to reduce system operation and development cost. Even with high penetrations, biomass would be widely distributed around the country and hence it is likely to reduce flows in high voltage distribution and transmission

82

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION networks, potentially postponing investments in these networks and reduce overall network losses41. 5.34

Biomass can comprise a wide range of technologies and fuel sources. One of the most likely forms of biomass generation would be based on energy crop incineration. From an electricity system perspective, these are all expected to exhibit similar characteristics.

5.35

Exceptions to this will be biomass based CHP schemes, as their operation will be heat rather than electricity driven. The ability of CHP plant to contribute to ancillary services is discussed in more detail later in the report.

5.36

Size of biomass plant will be limited by the ability to transport and store large volumes of low energy-density feedstock. It is not expected that the biomass plant will exceed 30MW-50MW sizes. Hence, smaller schemes are likely to be connected to 11kV while larger to 33kV.

5.37

Biomass generation will use traditional, well established steam turbine- synchronous generation technology. The plant is likely to operate as a base-load generator running at full output. In some cases, biomass generation could be used as a cogeneration plant (biomass based CHP) for steam production in which case the electricity output would depend on the demand for steam (CHP is discussed extensively later in the report, from paragraph 5.63).

TSO services 5.38

In principle, large biomass schemes will be able to contribute to system frequency regulation, similarly to conventional large steam sets. However, in order for the plant to provide low frequency response the generator would need to run part-loaded. Given that the opportunity cost of operating part-loaded in frequency sensitive mode would be relatively high, due to the losses of ROC revenue (in addition to reduced energy sales and losses in efficiency of operation), this is unlikely to be attractive to either plant or the system operator42. It is even less likely that this plant would provide synchronous reserve due to smaller value of this service.

41

SCAR, ILEX/UMIST, October 2002.

42

In order to position the plant in frequency sensitive mode, the system operator will need to de-load it and compensate the plant operator opportunity cost.

83

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 5.39

On the other hand, there is scope for the provision of high frequency response. As indicated earlier however, the value of this service is not very significant43.

5.40

Provision of standing reserve would not be feasible, as it would be unlikely to bring the plant on the system (from cold) within 20 minutes.

DNO services 5.41

Existing biomass plant has a relatively high technical availability and hence the capability to contribute to provision of security related services would be significant, provided that there are no restrictions in fuel availability. However, the need for such services is likely to be location specific and depend on the strength of local network. The evaluation of security services is discussed in Section 6.

5.42

Given that biomass technologies use synchronous generators and may be equipped with AVRs, such plant could contribute to voltage control in the local network. The ability to control reactive power may enable the scheme developer to connect the plant to weak distribution and avoid network reinforcement cost. Furthermore, biomass plant could supply some local reactive needs and hence increase the transport capabilities of the existing circuits. The availability of this support would need to be relatively high for DNO to consider this contribution.

Landfill gas 5.43

Landfill gas generation would use synchronous generation technology. The plant is likely to operate as a base load generator running at full output, provided that there are not limits of gas availability.

5.44

Typically, the size of the majority of landfill gas sites in the range 0.5 – 1.5MW. Some larger sites could be above 1.5MW, but usually these comprise multiple sets. Landfill gas would primarily be connected to 11kV.

TSO services 5.45

Landfill gas units are too small for provision of TSO frequency regulation services while provision of standing reserve is not feasible

43

With large penetration of wind however, the need for downwards regulation is likely to increase and this increase in volume of the service may also lead to increase in value of the service. However, it is unlikely that the provision of this service would lead to significant financial benefits.

84

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION if already operating. In instances where the gas supply is becoming depleted, there may be opportunities to provide some reserve services.

DNO services 5.46

Availability of landfill gas generation is relatively high, although the performance of individual sites may vary and will be primarily driven by gas availability. Landfill gas plant will generally be able to contribute significantly to provision of network security service. The importance of this contribution will be driven by the strength of the local network.

5.47

When evaluating contribution of landfill gas generation potential concerns may be associated with the uncertainty in future operation. In this context, it is interesting to note that the majority of existing sites have NFFO 3,4 and 5 contracts, although some sites that were developed under the NFFO 1 and 2 arrangements are still in operation.

Hydro generation 5.48

Small and medium sized hydro schemes without significant storage capacity are likely to be characterised by considerable variations in available water flow and hence output, particularly if the catchment area is on rocky or shallow soil. Therefore, uneven rainfall will lead to a variable resource. This generation would be normally classified as intermittent, similar to wind.

5.49

On the other hand, larger schemes would normally have some storage capabilities.

5.50

Small hydro schemes may use induction or synchronous generators, while larger schemes are based on synchronous generation.

TSO Services 5.51

Although in principle, medium size hydro scheme based on synchronous generation would be able to provide low frequency response, this is unlikely to be sufficiently attractive to schemes without storage, due to relatively high opportunity cost associated with spilling water and cost of lost ROC revenue. In addition to having installed governor control equipment, remote real time monitoring system will be required, which would contribute to increase of service provision cost.

5.52

Small schemes would not normally be adequately equipped to provide frequency regulation services.

85

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 5.53

Provision of standing reserve would be feasible for schemes with reservoirs providing that the minimum requirements can be met.

DNO services 5.54

Similar to wind generation, intermittent hydro generation could contribute to security provided that the output can remain at or above a certain required level for some minimum period of time. For schemes examined44 the persistence of small hydro plant was found to be more significant than that of wind and hence there is more scope for hydro to contribute to contribution to network security. This implies that hydro generation should be able to contribute to security not only during the period in which network transfer capacity is being switched in following a circuit outage, but also during repair and maintenance activities, as the contribution of hydro plant is likely to last for days.

5.55

Larger hydro with storage could make a contribution not only to network security, but also to service quality, as much of this plant is equipped with AVR and governor facilities necessary for islanding operation, including re-synchronisation equipment45. Such services would be generally of high value to DNOs, particularly in cases if interruptions to large number of customers may be avoided.

Non-Renewable Distributed Generation technologies Large Combined Cycle Gas Turbine power stations Frequency response, reserve and reactive power services 5.56

Large CCGT power stations, connected into 132 kV distribution networks, already participate in both the TSO’s frequency response and regulating reserve markets. Such power stations are few in number.

5.57

Regular CCGT provision of frequency response is comparatively rare due to the technical limitations of many gas turbines. Some modern gas turbines capable of rapid changes of output, whilst in frequency sensitive mode, are well placed to continue participation in this market. The TSO’s existing market arrangements are adequate for this purpose.

44

P2/6 Data Collection, Power Planning Associates, April 2004.

45

This is particularly the case with medium size hydro schemes developed prior to privatisation.

86

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 5.58

As the technical and infrastructure requirements for regulating reserve services are less onerous than those for frequency response, this could facilitate wider CCGT participation in the BM. The extent to which CCGT plant will participate in regulating reserve markets will depend upon the relative economics of generation from different fuel sources. Should gas fired plant represent the marginal generators of the future, it is highly likely that CCGT plant will continue to provide regulating reserve. As such plant is required to have a direct interface with BM through dedicated BMUs, the submission of bids and offers is simplified.

5.59

Overall the provision of frequency response and regulating reserve from CCGT plant is likely remain significant for the future. The anticipated increase in gas-fired generation may result in increased numbers of CCGTs providing such services.

5.60

The ability to control reactive power and the proximity to the transmission network is likely to be useful to the TSO for transmission voltage control purposes. CCGT aleady participate in this market.

DNO Network Security Services 5.61

CCGT plant has a high technical availability and hence the capability to contribute to provision of security related services could be significant. The need for such service is likely to be location specific and depend on the strength of local network. Given that such CCGT power stations are usually connected at 132 kV, the opportunities to provide security contributions are likely to be few due to the smaller number of installations and increased redundancy on 132 kV networks. However, opportunities could emerge to defer 400(275)/132 kV transformer reinforcement.

5.62

Given that CCGT utilise synchronous generators and will be equipped with automatic voltage regulators, such plant could contribute to voltage control in the local network.

Large CHP Installations 5.63

Figure 16 illustrates a typical schematic for a large scale CHP installation utilising a gas turbine and a Heat Recovery Steam Generator (HRSG) to raise high-grade steam, which can be used to power a steam turbine and feed an industrial process. Within this cycle, electricity is generated from both gas and steam turbines.

87

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Figure 16 − Typical schematic for a large scale CHP installation Stack Combustion Chambers Turbine

S

Compressor

T Air in Heat Recovery Gas Supply

Steam Generator

S

Supplementary Burners

Steam Turbi ne

Process Steam

5.64

A critical design consideration for CHP schemes is the balance between electrical and heat outputs. Usually the latter dictates the former, i.e. the gas turbine is sized according to the heat profile requirements of the industrial process.

5.65

Often CHP plants are specified such that the gas turbine runs at full load and thus provides a base load of heat input into the HRSG and supplemental burners are used to meet peak process steam requirements. Such arrangements optimise the efficiency of the CHP scheme and simplify the control arrangements as many installations are controlled with reference to HRSG inlet temperatures.

5.66

A consequence of such arrangements is that the operating envelope of the gas turbine, and hence the electrical generator, is severely constrained by the demands of the industrial process requiring heat. Normally the operational requirements of the industrial process will take precedence over any electrical output considerations.

5.67

During the 1990s there was a trend towards ‘over-sizing’ the electricity generating infrastructure on some large CHP schemes in response to advantageous trading arrangements in wholesale electricity markets. Such initiatives sacrifice overall thermal efficiency in return for an increased ability to export power.

88

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 5.68

Such schemes do not usually qualify as ‘Good Quality’ installations according to current DEFRA classifications although these can be more flexible in terms of the controllability of electrical outputs whilst still ensuring that process steam requirements can always be satisfied.

5.69

Another feature of some large CHP schemes is the ability to continue providing heat to the industrial process in situation where the gas turbine has been taken out of service. Again, supplementary burners must be utilised to meet the process heat requirement in such circumstances whilst electrical loads can be supplied from the local distribution network.

5.70

The availability of supplementary burners can effectively decouple electrical and heat outputs to the extent that the gas turbine can be operated independently of the industrial plant.

5.71

One consideration common to both optimised and over-size CHP schemes is security of supply. Whilst the primary motive for industrial customers installing CHP is to manage down energy costs, a secondary motive can be to improve security of supply. In the event of a supply interruption some schemes are configured to disconnect from the local network such that electrical supplies can be maintained to on-site equipment, effectively requiring the generators to operate in island ‘site’ mode.

5.72

In islanded operation, the on-site electrical generators must be capable of maintaining site frequency and voltages so must be equipped with governor control and Automatic Voltage Regulation (AVR) infrastructure in order to match the varying on-site electrical load. Again the requirement to independently deliver the required process heat output is achieved through the use of supplementary burners.

5.73

Sites with supplementary burners (thereby de-coupling heat and electrical outputs), governor control and AVRs can be regarded as distributed generators with highly flexible electrical capabilities and should thus be able to provide a wide range of ancillary services to both the TSO and DNO.

5.74

Whether CHP operators choose to provide such services will depend upon commercial priorities. Some operators with high value heat and electrical loads may be unwilling to provide network support services during times of system stress (e.g. post fault), preferring to operate in an islanded mode until any network constraints have been removed from the system, thereby limiting their risk exposure.

89

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Frequency response services from large CHP installations 5.75

A definitive statement regarding the capabilities of large CHP to provide frequency response services is not possible as there are a number of site-specific dependencies.

5.76

Frequency response services require part-loaded yet flexible and responsive generator operation through governor control. Essential to the provision of frequency response services from such providers will be the ability to de-couple electrical and heat loads, which invariably requires supplementary burners. It should be noted that such operation would immediately lead to reduced thermal efficiencies and thus increase costs.

5.77

Without such infrastructure, it is unlikely that a CHP operator could provide such services. Also, the technical specification of the gas turbine would also need to be assessed to ensure that the desired rate of change of output was achievable. It is reported that many gas turbines deployed within CHP schemes will not be sufficiently flexible to provide frequency response services.

5.78

Those CHP schemes equipped to operate in islanded mode may be able to provide frequency response services as these are likely to be fitted with the necessary infrastructure (albeit minus TSO monitoring hardware). Such operation would represent a new mode of operation for many CHP operators and the attractiveness of service provision would need to be determined on an individual basis.

5.79

The largest CHP schemes, i.e. those over 100 MW, will be required to comply with Grid Code requirements so should already be capable of providing frequency response services to the TSO.

5.80

Overall, the ability of large CHP schemes in general to provide frequency response services is not clear. Undoubtedly, some installations will be able to provide such services. Plant configurations and infrastructure requirements represent the key dependencies although there is a high probability that the largest schemes will have the broadest capabilities.

Reserve Services from large CHP installations 5.81

As the economics supporting investment in large-scale CHP installations are linked to high utilisation, energy intensive processes, there is a high probability that such CHP will operate continuously. This immediately discounts any opportunities in standing reserve markets. Indeed, even if a CHP plant was in standby mode, it is unlikely to be sufficiently flexible to satisfy TSO requirements.

90

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Consequently any opportunities regarding reserve will relate to regulating reserve markets. 5.82

Similar to frequency response, regulating reserve requires partloaded, flexible plant. Again there will be a requirement to de-couple electrical and heat loads through the use of supplementary burners. This will reduce thermal efficiencies and thus increase operating costs.

5.83

Without the ability to decouple, it is unlikely that a CHP operator could provide reserve services. Also, the technical specification of gas turbines would also need to be evaluated to ensure that the desired ramp rates could be achieved.

5.84

The largest CHP schemes are likely to have direct interfaces with the BM via dedicated BMUs so the submission of bids and offers should be straightforward. For smaller schemes, registered in settlement through a supplier, there would be a requirement for the supplier to submit bids and offers on behalf of the CHP operator, inevitably complicating service delivery. In order that smaller operators could participate in the BM, it may be necessary for suppliers to aggregate the capabilities of many CHP operators to offer meaningful reserve volumes to the TSO.

5.85

Overall, large CHP does not appear a natural candidate to provide reserve services. The complexity of service delivery from such highly specialised plant, coupled with settlement complications mean that regulating reserve will continue to be provided by large flexible power station units with dedicated BM interfaces. Any opportunities which arise are likely to be linked to the largest and most flexible CHP schemes.

Network security services 5.86

Currently, the commercial operation of different CHP plants is extremely variable due to energy market conditions. CHP should be capable of a relatively high technical availability and hence should be capable of contributing to the provision of security related services. However, the ability of each CHP provider would need to be assessed individually to determine whether contributions could be sourced.

5.87

Given that large CHP may be equipped with AVRs, such plant could contribute to voltage control on the local network. The ability to control reactive power may enable scheme developers to connect the plant to weak distribution networks and avoid network reinforcement costs. Furthermore, CHP could potentially supply some local reactive requirements, thus increasing the transport capabilities of the existing circuits.

91

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 5.88

The availability of this CHP support would need to be relatively high for DNO to consider contributions. As has already been discussed, the CHP operators own concerns regarding site security of supply may take prudence to the extent the generation will be disconnected from distribution networks during network disturbances.

Micro-CHP capabilities 5.89

This section discusses the potential for micro CHP to provide ancillary services and outlines the infrastructure that might also be required. High-level estimates are provided regarding the costs of the infrastructure.

Potential for Micro CHP to supply ancillary services 5.90

The micro CHP schemes considered are those used in domestic applications. They range in size between 1 and 5kW and are normally connected at 230V. Most schemes employ Sterling engine technology. In common with large schemes these devices are heat led. The generators in typical micro CHP devices are of asynchronous design with no governor or AVR.

5.91

Taken individually, micro CHP cannot meaningfully provide any of the ancillary services evaluated as part of this work. However when aggregated together, micro CHP could theoretically provide some service capabilities. Therefore, a condition precedent for micro CHP to provide services to either the TSO or DNOs, is that the outputs from units would need to be aggregated together.

5.92

Should infrastructure be available to coordinate aggregate responses from micro CHP, it may be feasible to facilitate the provision of standing reserve and DNO network security services. It would not be possible for aggregated micro-CHP to provide any low frequency related services because inductive current inrushes would exacerbate any low frequency problems and the gradual transition into power generation would be too slow.

DNO security services 5.93

On LV networks, DNOs could utilise the aggregate output of microCHP generators to offset network demands and support local demand groups during peak periods. This could ensure compliance with planning standards without necessitating investments in network infrastructure.

5.94

In order to deliver the desired contributions to particular Group Demands, the DNO would need to ensure that the remote aggregation and operation arrangements were able to target particular areas with sufficiently high densities of micro CHP. This 92

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION would enable security contributions to be initiated at any time, although the DNO would probably need to assume the role of aggregator and controller in these circumstances46. 5.95

However, if DNOs were only likely to require micro-CHP security contributions during winter peaks, probabilistic approaches could be adopted whereby the DNO would accurately predict how much micro CHP contribution could be relied upon. During winter peak demand periods, there would be a high likelihood that many micro CHP installations would be operating.

5.96

Group Demand support from micro-CHP could be relevant to Supply Classes A, B and C from Table 1 of ER P2/5. This is mainly because of the favourable timescales for restoration of group demand following a single circuit outage. The most onerous timescale is 15 minutes to restore group demand minus 12 MW applicable to class C. This time scale is comparable to the 20 minutes for generation contracted for standing reserve in the reserve market to be up and running.

Standing Reserve 5.97

Assuming generation from many micro CHP installations could be initiated simultaneously, it should be possible for micro CHP to also contribute towards standing reserve. The infrastructure required to control large numbers of micro CHP installations would be very similar to that required for aggregating outputs for DNO security contributions.

5.98

Assuming both the DNO and TSO required aggregate services from micro CHP installations, any infrastructure costs could be shared. As the DNO is likely to require more locational control than the TSO, it would make sense for the DNO to act as aggregator (however this would have other implications as explored in Section 6).

5.99

The infrastructure would basically need to comprise communications to switch the micro generation on or off as required, an interface with the homeowner’s local control and metering arrangements. The key issues and options related to these infrastructure components are discussed below.

Communication 5.100 The objective of remote aggregation is to enable a group of micro CHP units to be controlled centrally making communication the

46

Contribution of aggregated (controlled) micro-CHP is not considered in the proposed ER P2/6

93

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION central consideration. When the generating units involved are micro CHP units some form of broadcast based communication would appear to be a sensible means of achieving the required levels of aggregated response within the required timescales. For example to achieve a 3MW response from 1-5kW requires simultaneous communication with 2 to 3 thousand units. 5.101 The units themselves must be equipped with transceivers and be able to carry out specific instructions as directed by the aggregator (or central controller). Local control issues and options are discussed in the next section. Potential communication options are outlined below. 5.102 Figure 17 below shows a schematic diagram of the communication process from the TSO to the micro CHP. Due to the large number of small CHP units it is not practical for the TSO to instruct the units directly and monitor proof of delivery from each unit. Therefore a single instruction would be sent out to the aggregator requesting a specified level of response. The infrastructure required to send this instruction can take one of several forms. 5.103 The simplest form would be via the telephone using the PSTN or mobile services. Figure 17 − Communication process from TSO to micro CHP via the aggregator TSO Single instruction to aggregator multiple instructions to micro CHP units

Aggregator

Micro CHP units

5.104 The communication from the aggregator to the micro CHP units is more challenging. For this application two-way communication is essential for:

94

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION •

on/off status reporting and control;

•

despatch of CHP generators; and

•

metering data communication

5.105 The communications options currently available include: •

Radio tele-switching: (RTS) via radio frequency broadcast;

•

PSTN Dial-up: Communication by public telephone;

•

ADSL: Asynchronous digital subscriber line allows permanent connection to the Internet in the home;

•

mains signalling (e.g. ripple control); and

•

GSM/UMTS: Mobile phone technology, General System for Mobile (GSM) and its successor Universal Mobile Telecommunications System (UMTS).

5.106 The RTS system in the UK uses BBC 198kHz transmitters (Radio 4 long wave) to regulate the charging of storage heaters. Because RTS is unidirectional and tends to have a very low bandwidth it is not suitable for control and monitoring of micro CHP. 5.107 While offering two-way communication with adequate bandwidth and wide coverage countrywide the Public Switched Telephone Network (PSTN) is also not suitable because it does not allow broadcasting from a central control point. There are also problems of system integration in the home between a phone line used for domestic calls and Internet access, and a micro generator control unit. Provision of a dedicated line would be uneconomic. 5.108 In recent times the introduction of Asymmetric Digital Subscriber Line (ADSL) services by BT and other operators, allows permanent connection to the Internet in the home, independent of the telephone. Broadcasting over the Internet is possible, but the main limitation of this option is that ADSL coverage does not extend to rural areas, and the high cost will limit its take up. Broadband cable offers similar Internet access, with the same limitations. 5.109 Ripple control is the best-known form of mains signalling and it has been deployed for many years to control water heaters as part of demand side management. The high cost associated with these schemes prevents their use on a large scale. However, transmission across the mains wiring within the home works well as a Local Area Network (LAN) technology and various proprietary techniques are in use. 5.110 The current mobile phone technology, General System for Mobile (GSM) and its successor Universal Mobile Telecommunications

95

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION System (UMTS), are capable of supporting both bi-directional and broadcast data communications over most of the UK. This technology is the most promising option for application in monitoring and control of aggregated micro CHP participating in a reserve market. 5.111 The particular features of GSM that are well suited to this application are the Short Message Services (SMS). A detailed description of these services has been discussed by Boait 200247. In summary SMS services comprise SMS Point to Point (SMS-PP), text messaging service, and Cell Broadcast (SMS-CB). Message rates of 2,500 per second can be processed by current high performance SMS systems. This would readily support, for example, a daily status report from millions of generators to the central database. 5.112 SMS-CB allows a data message of between 1 and 15 pages, each of 82 octets, to be transmitted to all GSM receivers in a specific geographical area comprising one or many cells. Because each receiver has an association with only one cell, it is possible to broadcast different messages on adjacent cells with no ambiguity in the overlap zone. The size of cells varies from a minimum diameter of about 200 metres in urban areas to 15 km in rural areas. The combination of an efficient broadcast mode and fine geographic granularity would allow the Radio Tele-switch concept to be extended to create sophisticated control and data exchange protocols. 5.113 A further advantage of SMS services is that data carried by these protocols is protected by the forward error correction and encryption included in the GSM standard. This will allow a direct interface to the application layer software performing the required control functions.

Local control for micro CHP and Metering 5.114 To allow the various system components within the home to be procured or replaced separately, but integrated successfully, a standardised home network is required. At present there are several competing protocol standards of which the European Home System (EHS), BACnet, and LONTalk/LONWorks are particularly well developed for heating and electrical control applications. 5.115 The most convenient communications medium within the home is the mains wiring, for which suitable LONTalk modules are already available.

47

P. J. Boait, ‘Telecommunications Architectures And Protocols For Control And Despatch Of Embedded Micro Generators’, Fifth International Conference on Power System Management and Control, 2002. (Conf. Publ. No. 488), 17-19 April 2002, Pages:329 334

96

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Data transfer and control protocols 5.116 As discussed above, protocols are required for generator status reporting and control, despatch of CHP generators and communication of metering data.

Status reporting and control 5.117 To allow accurate aggregation each generator should report that it has been enabled for normal operation. CHP units should also report the heating control timer settings that will determine when they can generate. 5.118 The central controller should be able to disable generators singly or in an area (defined by cell broadcast).

Micro CHP generator despatch 5.119 The use of this SMS messaging to despatch domestic CHP generators can be achieved by dispatching a single instruction to several units to increase or reduce output. The primary communication from the aggregator could also be accomplished through a smart meter, which would in turn communicate with micro CHP unit via the mains. Since the smart meter has to communicate its reading to the central point, there is a lot of merit in using a single point for all communication between the central point and a particular site. Again this could favour the DNO as aggregator. 5.120 In normal operation the heating output will be controlled by a wall thermostat which triggers the heating when the room temperature falls below a threshold set by the consumer, and cuts it off when the temperature rises past that threshold. Where water heating is also provided by the micro CHP a similar control regime would also be available. 5.121 The heating operates on a duty cycle determined by the heat loss rate of the house, which in turn depends on the difference between the thermostat setting and the external air temperature. 5.122 Dual heat supply functions (space and water heating) enhance the capability and duration of the micro CHP unit. Remember that for standing reserve the TSO’s instructed level of output should be provided within 20 minutes and last for at least 20 minutes with the recovery period being no more than 20 hours.

Cost estimates 5.123 It is reported by Boait, that depending on the extent to which cell

broadcast is taken up by PTOs for other applications, it should be

97

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION possible with large scale deployment (>100,000 units) to obtain a cost of 2p per unit or below for overnight message exchange, giving an annual cost of £7.30 per unit. The cost for dispatching SMS messaging as and when required would not be very different depending on how frequently the service is called upon. It is also reported that the cost of the GSM transceiver if integrated with the domestic control unit should cost no more than £50. These costs would have to be compared with income earned from providing the service to decide whether it is worth making the investment or not.

Other technical and commercial issues 5.124 There are number of technical and commercial issues that must be considered in the application of micro CHP to supply reserve to TSOs and security services to DNOs. These issues are discussed briefly below.

Technical 5.125 It is envisaged that the collective output of a group of micro CHP units requested to generate power (for reserve or security services) from cold will start high (but less 100%) and decay as shown in Figure 18. Figure 18 − Response scheduling of groups of micro CHP units

% MW Response 100%

Average response over time

95%

T1

T2

T3

Time (minutes)

5.126 The initial power output will in general fall below 100% as some units may fail to start or may already be running. In general units will run for periods ranging 30 to 60 minutes and switch off depending on maximum temperature settings for water heaters and space heating. After a period of cooling off, the units would restart until the temperature constraints again force them to switch off. In order to ensure that a certain level of response is sustained the DNO, acting as

98

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION aggregator, would have to organise the micro CHP units into groups with staggered starting times T1, T2 and T3 as shown in Figure 18. 5.127 The response of micro CHP units will vary seasonally and with time of day. In winter for example it would be expected that most units would already be running during the day to keep homes warm whereas at 3:00 AM in the morning most units would be expected to be off as people will be asleep. In summer, all space heating may be switched off leaving only water heating. The duration the unit is likely to be able to run will be significantly shorter on a hot summer afternoon than on cold winter morning. In addition, customers will retain overall control of the heating equipment within the home and may override DNO instructions, thus reducing the availability of the service. 5.128 The data needed to derive typical micro CHP group response curves shown in Figure 18 would be gathered over time.

Commercial 5.129 On the commercial side, suppliers have contracts with final customers and with DNOs. DNOs also have contracts with TSOs. In operation of the system the DNO who has no direct contracts with the final customers will effectively issue instructions to operate their CHP units. As there are commercial implications each time a DNO (or TSO via the DNO) calls upon a customer to provide either standing reserve or security related services, the DNO would have to notify suppliers as well. Therefore suppliers would also have to be equipped with SIM cards to enable them to receive notification via SMS each time their customers are called upon to supply a service. 5.130 The cost implications to the customer and hence the compensation they would expect to receive for service delivery would vary seasonally and with time of day. The cost to the customer of providing the service is composed of the fuel cost (gas) incurred less the income earned from energy sales for the duration of service delivery. The economic justification for micro CHP is based upon customers’ requirements for heat and electricity is a by-product of the heating process. Therefore if the customer has no use for the heat generated during the time the service is being delivered they would seek full cost recovery (assuming the micro CHP unit could somehow dump heat). Conversely, if the heat could actually be used within the house, then the cost would be lower. It may not be straightforward to determine whether or not the heat was needed at the time the service was called upon. It might be simpler to simply compensate the customer assuming the heat was surplus to their requirements. Clearly this is a subject for further debate.

99

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 5.131 The other interesting issue relates to performance when called upon to provide the service. It is to be expected that the performance of different customers when called upon to provide a service will vary for a whole host of reasons. For example a customer may not respond because their micro CHP unit may already be running or there could be failure in communication.

Standby Generation 5.132 Diesel generators are frequently used to provide standby generation for industrial and commercial operations. They vary in size but tend to be less than 5MVA and are normally connected at 11kV or 33kV.

Range of services 5.133 Most diesel sets use synchronous generators fitted with a droop governor and an AVR as they are required to operate as stand alone units. In theory therefore diesel generator sets could provide frequency response and reactive power services. However because the machines are normally on standby, they cannot provide frequency response services as these services can only be provided by part-loaded machines that are already running and synchronised to the grid. 5.134 Standby diesel is more likely to provide standing reserve to TSOs and security (and perhaps niche service quality related services to DNOs). Depending on size, some aggregation may be required to realise these services. The infrastructure requirements to facilitate participation of standby diesels in standing reserve, security and service quality markets are discussed below.

Infrastructure requirements remote start-up and communication 5.135 Unlike the micro CHP case where there are potentially many thousands of units, the number of standby diesels is likely to be a lot fewer, most likely in the region of hundreds for a given area. The infrastructure requirements for an aggregated service from a group of diesel generator sets would mainly comprise metering and the communication infrastructure to switch the generation on or off as required. This infrastructure could be to a large extent similar to that required to facilitate micro CHP participation in standing reserve and security services markets for individual stand by operators. 5.136 As discussed earlier, a consequence of the 3MW participation threshold has been that aggregated standing reserve services have already developed. These aggregation arrangements are much simpler than those proposed for micro-CHP as they apply to customers with a relatively small number of standby generators

100

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION (typically <20) or customers with a ‘large’ standby generator. The infrastructure approach adopted utilises a dedicated personal computer provided by the TSO and an iSDN communication link between the TSO and the aggregator/customer, which is used to issue service initiation instructions. Dedicated communication and control arrangements, installed by the customer, remotely start the standby plant and monitor ongoing operation. Such data is also fed back to the TSO. 5.137 The costs of the existing arrangements are high and not necessarily scaleable. The economics of such service provision is justified by the size of standby generators utilised. To see a step change in the extent of aggregation undertaken, lower cost control infrastructure would need to be developed.

Infrastructure to facilitate operation in parallel with the grid 5.138 In order for standby units to provide reserve and security services, they would have to operate in parallel with the grid. Diesel units purchased specifically for standby duty may not have the required facilities and infrastructure to facilitate this mode of operation. At the very least the following hardware and facilities would be needed: •

synchronising facilities; and

•

enhanced protection.

Synchronising facilities 5.139 Synchronising two separate alternating current systems is a wellestablished practice. Three conditions must be fulfilled before the two systems can be connected; 1) the phase angle, 2) frequency and 3) voltage magnitude differences between the two systems must be within acceptable tolerances. This is to reduce the amount of current flow when the two systems are connected, hence avoiding any damage to transformers and other equipment such as distributed generation. 5.140 There are several existing devices that are used to check synchronism before two ac systems are connected to together. Some of the common ones are synchroscope, phase angle voltmeter, synchronising check relay and synchronising relay that checks the difference in phase angle, frequency and voltage between two systems before closing the circuit breaker to connect the two systems. 5.141 The synchroscope and phase angle voltmeter are used in the situations where circuit breaker is closed manually. The synchronising check relay is used to provide an electrical interlock in

101

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION order to prevent closing of the circuit breaker when the frequency, voltage and phase angle differences between the two systems are not within the acceptable limits. 5.142 Automatic synchronization equipment would be preferred for reserve applications as the time it takes to call out an operator to carry out manual synchronisation may exceed the 20 minutes time stipulated for standing reserve to be up and running. Where automatic synchronization equipment is fitted, excitation and governor control of the diesel unit will regulate the terminal voltage and speed of the generator to match the grid frequency and voltage automatically. 5.143 Synchronising equipment is standard and machines that do not have it could be retrofitted if required.

Protection 5.144 When a standby diesel unit is to be operated in parallel with the grid it is important to ensure the machine is equipped with appropriate protection to protect it from the grid, which is a very large source of fault current. Therefore basic generator protection including in particular reverse power, negative phase sequence, over-current and earth fault would have to be installed. Again these protection devices are standard and can be installed at relatively low cost.

Summary of theoretical DG capabilities 5.145 The ancillary service capabilities of renewable and non-renewable technologies are summarised in Table 16 and Table 17. Table 16 − Summary of renewable technology capabilities DG DG Technology TechnologyType Type Ancillary Ancillary Service Service

Wind Wind non-DFIG non-DFIG

Wind Wind DFIG* DFIG*

Size Size

<<50 50MW MW

>50 >50MW MW

Frequency Frequency

HF HFonly only

Reserve Reserve

Possible Possible

Possible Possible

Biomass Biomass

Land LandFill Fill Gas Gas

Solar SolarPV PV

Hydro Hydro

<100 <100kW kW

>>1MW 1MW

Possible Possible

Possible Possible

1–100MW 1–100MW 11––10MW 10MW HF HFonly only

HF HFonly only

Possible Possible

Possible Possible

Future Future islanding? islanding?

Future Future islanding? islanding?

Reactive Reactive Network Network Support Support

Limited Limited

Black BlackStart Start

Future Future islanding? islanding?

* Wind Farms <50 MW may employ DFIG machines in future Table 17 − Summary of non-renewable technology capabilities DG DG Technology TechnologyType Type Ancillary Ancillary Service Service

CCGT CCGT

Large LargeCHP CHP

Micro MicroCHP CHP

Diesel Diesel&& OCGT OCGT (Standby) (Standby)

Size Size

>100 >100MW MW

1-100 1-100MW MW

11––55kW kW

<<50 50MW MW

102

Frequency Frequency

Limited Limited

Reserve Reserve

Possible Possible

Limited Limited Possible: Possible:High High penetrations penetrations

Reactive Reactive Network

Possible: High

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

[This page is intentionally blank]

103

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

6. 6.1

COMMERCIAL & TECHNICAL FRAMEWORK This section explores potential arrangements for the provision of different ancillary services from distributed generation in terms of: •

service recipients, providers and intermediaries;

•

consistency with the current commercial arrangements; and

•

new commercial arrangements.

Service recipients, providers and intermediaries Service recipients 6.2

The primary recipients for the selected ancillary services will be either the TSO or individual DNOs.

6.3

For the purpose of this work, TSO frequency response and TSO reserve have been regarded as non-locational national services. By contrast, TSO reactive power and all the DNO-centric services have been treated as being highly locational, i.e. the position of the service provider on the recipients network has been regarded as critical.

Service providers 6.4

Consistent with the scope of this work, the parties evaluated as potential providers of the above services have been the various forms of distributed generation rather than demand side participation (which have been specifically excluded from this study). As distributed generation is a broad term encompassing many generating types, the work has focused on existing and near commercial DG technologies.

6.5

Although most forms of distributed generation will have a theoretical capability to provide many of these ancillary services, it will be not possible to make absolute statements regarding the service capabilities of different generation types due to the wide variety of other considerations influencing service delivery. Such other considerations include: •

size of generator;

•

generator availability: Technical and commercial;

•

intermittency of generation;

•

location and voltage of Connection;

•

extent of Active Management on the local DNO network;

104

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION •

presence of network constraints impacting upon generation operation;

•

interactions with other generator services;

•

sophistication of generator control infrastructure; and

•

staff presence.

6.6

Whilst absolute statements confirming technology capabilities are not appropriate, it is possible to outline the generic capabilities for the different generation types. These generic capabilities are discussed further in Section 4.

6.7

Other non-technological considerations will impact upon service delivery, reinforcing the requirement that most DG installations will need to be evaluated on a site-specific basis to determine their actual ancillary service capabilities.

Scope for intermediaries 6.8

It is envisaged that in many cases, it will be possible for a direct relationship to be established between service provider and recipient. This will be especially true for the larger distributed generators. This does not necessarily preclude alternative arrangements whereby the provision of ancillary services is facilitated by an intermediary positioned between the recipient and provider.

6.9

The primary driver for indirect arrangements relates to generator size. In many instances, individual small generators will be unable to provide sufficient output to satisfy recipient requirements. In such cases, it may be possible for small generators to contribute collectively by aggregating their individual outputs to create a useful of level service for a recipient. Aggregation activities may be suited to the following range of different market participants:

6.10

•

suppliers;

•

DNOs;

•

customers (owning a number of distributed generators); and

•

3rd party operators.

In the case of services being provided to the TSO, especially response and reserve, aggregation could remove TSO operational constraints regarding the initiation of a service from multiple generators. From a system operation perspective, it is unlikely that the TSO will want to maintain potentially thousands of interfaces with small individual generators, especially in cases where service delivery cannot be guaranteed.

105

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 6.11

In circumstances where delivery uncertainty will undermine a provider’s ability to offer an attractive service, aggregators could become involved and employ statistical techniques across a number of similar providers to reduce recipient uncertainty and hence increase service attractiveness.

6.12

It appears that aggregation activities could be particularly well suited to non-locational, national services, as the number of potential service providers can be maximised. This does not preclude aggregation services also evolving for locational services.

6.13

The characteristics of different aggregation activities are developed further later in this Section.

Consistency with the current commercial arrangements Energy trading and settlement 6.14

It should be noted that the market structures for electricity currently differ between England & Wales and Scotland although it is intended that these arrangements will be harmonised across Britain from April 2005 following BETTA implementation. For the purposes of this report, a post BETTA outlook has been adopted.

6.15

In the absence of exemptions, all large and grid connected generators (as outlined in paragraph 2.5) are required to comply with the Grid Code, the Connection and Use of System Code (CUSC) and the Balancing and Settlement Code (BSC). Operationally, such generators are required to trade and settle their outputs through a dedicated generation BMU account under NETA.

6.16

By definition, distributed generation is not connected to the transmission system and therefore most DG is not required to comply with the Grid Code, CUSC or BSC. Instead, the energy from distributed generation is traded and settled as negative consumption through a supplier’s demand BMU, i.e. supplier aggregate consumption under each BMU is effectively reduced by the output from distributed generation. Consequently, it is important for Suppliers to accurately forecast generation output in order to manage their imbalance exposures in the BM.

6.17

These settlement arrangements have advantages and disadvantages for distributed generators. The avoidance of a direct interface with the BM has costs benefits for the distributed generator in terms of avoided operational interfaces (hardware, software and staff), simplified information flows and simplified contracting arrangements. Also, as distributed generation metering points are

106

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION registered on DNO Metering Point Administration Systems (MPAS), this facilitates straightforward changes of supplier (purchaser in the case of a generator) through the processes defined in the Master Registration Agreement (MRA). The Supplier-Hub principle also reduces the administrative burden on generators by delegating responsibilities for metering, data collection and data aggregation to others. 6.18

The main disadvantages of this indirect approach to trading and settlement are reduced flexibility in terms of BM participation and the requirement to share any trading benefits with suppliers.

6.19

In the context of ancillary service provision, where distributed generators are potentially required to change output levels at short notice, supplier exposures to energy imbalances may increase, especially where instructions are issued within NETA Gate Closure timescales. These exposures are discussed in greater detail later in this Section.

6.20

Similarly, TSO instructions requiring distributed generators to change output within Gate Closure timescales may require Bid/Offer ladders to be submitted by the supplier to ensure payment. In such circumstances, the supplier would receive corresponding payments from the TSO, the benefits of which would need to be factored into any commercial agreement between the supplier and the distributed generator.

6.21

Whilst such indirect trading mechanisms, via supplier BMUs, are perhaps not ideal for ancillary service provision, these arrangements do not preclude the development of such services in the future. Also it should be recognised that these arrangements do provide generators with a cost effective means of trading their main energy outputs.

Supplier Renewable Obligation 6.22

The Renewable Obligation requires suppliers to source an increasing proportion of their electricity requirements (by volume) from renewable sources until 2015. Suppliers comply with this obligation by procuring Renewable Obligation Certificates (ROCs). Should Suppliers fail to procure sufficient renewable generation to meet their targets, they are required to buy-out any shortfalls by paying into the RO Buy-out fund at an indexed rate of £31.39/MWh.

6.23

For each Renewables Obligation period, total buy-out payments received by Ofgem, together with any interest earned on them, is distributed amongst those suppliers who have correctly presented

107

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION ROCs in proportion to the total number of correctly produced ROCs for the period. 6.24

The net effect of this mechanism is to enhance the value of renewable generation to a current value of approximately £52/MWh. The high value of renewable electricity exports has a significant impact on the revenues of renewable generators. These cash flows are evaluated in Section 7.

6.25

As a high proportion of renewable generation is connected to distribution networks and thus settled through suppliers’ NETA demand accounts, the administrative complexity of administering Renewable Obligation compliance is therefore simplified.

6.26

Overall, as there is currently a shortfall of renewable generation to the extent that not all suppliers are able meet their targets, ROC prices are forecast to remain high. One effect of high ROC prices is that renewable generators are strongly incentivised to maximise outputs to avoid foregoing any ROC income. Similarly, suppliers are incentivised to comply with the Renewable Obligation and will often require renewable generators to maximise their outputs.

6.27

The impact of these incentives on renewable generators to maximise output will have significant implications regarding the attractiveness of any ancillary services requiring part-loaded operation.

Connection and Operational agreements 6.28

All DNOs require connection agreements with individual distributed generators. In the supply market, the arrangements for demand customers are simplified through the use of standardised connection terms, applicable to most customers. In non-standard situations (e.g. industrial sites), DNOs often adopt a more detailed, model-form agreement, which was developed nationally, and can be tailored to the requirements of each customer’s connection. As yet, no such arrangements have been developed for distributed generators. Indeed, considerable regional variations exist in relation to the terms of connection for distributed generators.

6.29

In instances where the connection of a distributed generator to a distribution network requires significant and costly system changes48, DNOs can sometimes offer less expensive alternative arrangements where the generator is prepared to accept operational constraints, e.g. requirements to reduce output or cease generation under certain

48

This situation may be somewhat ameliorated by the move to ‘shallowish’ DG connection charging regime as proposed from 1 April 2005.

108

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION pre-defined network conditions. Such arrangements often appear attractive to distributed generators, as the cost savings can be significant and the operational constraints are manageable, although this may then preclude the provision of certain ancillary services. 6.30

Both connection and operational agreements provide an existing contractual relationship between DNOs and distributed generators, which could be extended to incorporate terms for ancillary service provision to DNOs.

Use of system 6.31

Distribution Use of System Agreements (UoSAs) are bilateral contracts predominantly between DNOs and suppliers (the primary users distribution systems). These agreements are important to DNOs as the bulk of their regulatory income is derived through Distribution Use of System (DUoS) charges.

6.32

Historically, distributed generator connections have been paid for according to a deep connection charging policy, so it has not been necessary for DNOs to apply use of system charges upon distributed generators. From April 2005, the connection charging arrangements for new distributed generation will move from deep onto a shallower basis, i.e. a lower initial capital contributions will be sought from connectees although ongoing use of system charges will be introduced where network reinforcement is required.

6.33

During the next regulatory price control period, there is an expectation that a more cost reflective pricing mechanism will be developed for users of distribution networks, including generators. Under such arrangements, it will be possible to accommodate services provided to and services from generators in DUoS charges. Taken to the extreme, it could be feasible for generators, providing significant levels of ancillary services to DNOs, to receive negative DUoS charges.

6.34

Whilst cost reflective DUoS charging represents an interesting medium term development with respect to ancillary service provision, it should be recognised that existing UoSAs do not as yet fully accommodate distributed generation and amendments will be required to introduce direct Use of System charging.

Commercial arrangements for DG ancillary services 6.35

Dependent upon the nature of the ancillary service provided and the different parties involved, there are a number of different commercial

109

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION arrangements that could be adopted to reward generators for service provision. 6.36

The selection of the most appropriate commercial arrangement for each service will largely depend upon the number of service providers able to offer the required service and the financial materiality of the market, relative to the costs of operating the commercial arrangement.

Market Opportunities 6.37

A pre-requisite for markets or market mechanisms to develop, is the requirement for multiple providers to be capable of providing each service or multiple buyers requiring particular services.

6.38

Another factor facilitating the development of competitive market mechanisms is the standardisation of product offerings to minimise differences in provider value propositions.

6.39

Whilst it is possible that market based arrangements could emerge with respect to national, non-locational services (e.g. frequency response and reserve), such approaches will be less suitable where services are bespoke or there is lack of alternative service providers. Such considerations will be particularly relevant to the highly locational DNO services, especially whilst DG penetrations remain low.

6.40

In recent years, the TSO arrangements for the procurement of frequency response and reserve (both regulating and standing) have become increasingly market based and this trend appears set to continue. One possibility for increased DG participation in these markets will be to extend the current arrangements such that smaller providers can participate. However, it should be recognised that the TSO’s current arrangements do not actually preclude participation from smaller generators, albeit through aggregators.

6.41

One means of increasing small generator participation in TSO ancillary service markets would be through an expansion of aggregation services. To date, the development of aggregation services has been slow with only one independent aggregator currently active in these markets. As aggregation is a commercial activity, its attractiveness depends upon a number of wider market considerations such as market structure, the levels of generation overcapacity and prices. The attractiveness of expanded aggregation services is further discussed later in this Section.

6.42

Market-based mechanisms that are particularly well suited to competitive service provision, especially in monopsony situations,

110

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION include tender based approaches whereby the service recipient defines the service requirement and requests potential providers to submit proposals for the service. The extent to which service recipients define their service requirements depends upon the recipient’s desired outcome. In instances where recipient service requirements are prescriptive, a price focussed tendering process will result. Alternatively, where service specifications are less prescriptively defined, providers have more latitude to influence the service offering, which can encourage innovation. 6.43

Where a service recipient is primarily focussed upon price, standardised approaches can be appropriate although recipients must ensure that service requirements are not over specified as this only serves to increase the cost of all provider offerings. This issue is particularly relevant to ancillary service provision at the TSO level, as the mandatory capabilities of certain generator types, as defined in the Grid Code, are currently under review.

6.44

It should be noted that, whilst mandatory generator capabilities can facilitate market-based mechanisms, such approaches do not guarantee the delivery of least cost solutions.

6.45

Other, more localised, ancillary services for which market-based mechanisms could be developed (at a distribution level) include TSO reactive power displacement and DG contributions to network security. In both instances, it would only be appropriate to consider competitive approaches to service procurement once sufficient DG was located within a particular area.

Non-market based approaches 6.46

A range of alternative commercial arrangements are available for situations where insufficient generators are available to merit market based approaches. These could include: •

bilateral arrangements;

•

cost reflective approaches; and

•

administered solutions.

6.47

Bilateral arrangements represent a highly flexible approach whereby two willing parties agree mutually acceptable terms for service provision and delivery. Such arrangements are usually formalised through legal contracts, which can be of fixed duration or ‘evergreen’ in nature.

6.48

Difficulties can arise with bilateral arrangements where one party, either the buyer or the seller, has disproportionate power in

111

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION negotiations. This can be especially relevant in monopoly or monopsony situations where regulatory guidance can be required. 6.49

Bilateral arrangements could be particularly relevant for ancillary services, especially regarding those services provided to DNOs, e.g. generator security contributions, IIP services and voltage support. To ensure negotiations are equitable, it could be necessary for the regulatory guidance to be issued regarding valuation methodologies and procurement principles.

6.50

Also, in order to ensure that a consistent approach is adopted nationally, it could be beneficial to develop standardised bilateral contractual forms, which could be localised by DNOs to reflect specific situations. The benefits of such an arrangement would be simplified and reduced cost contracting for both parties and consistency of approach for distributed generators, irrespective of where they chose to operate in the country.

6.51

Cost reflective approaches could also be adopted whereby recipients reward providers according to the costs incurred in delivering the service. Similar cost pass through arrangements can also be developed. Difficulties can arise where parties fail to agree the associated costs of service provision. Other disadvantages include the absence of any strong incentives on providers to reduce the cost of services.

6.52

Administered solutions often require the intervention of an empowered third party, such as a regulator, to determine the commercial arrangements for a service, including reward mechanisms and prices. It should be noted that administered, centrally managed, solutions are the least favoured by regulators and market participants alike.

6.53

Reasons for the unpopularity of such approaches include the burden upon administrators, the requirement for third party intervention in market operations (which can influence impartiality), a lack of flexibility in terms of service provision and the inability to capture and reward enhanced services. Consequently, such approaches are often only adopted as a last resort.

Contractual options 6.54

For each potential distribution level ancillary service, there will be a variety of possible contracting options. As has already been outlined in the review of TSO ancillary services, a range of contractual structures already exist. These include: •

mandatory and commercial arrangements for frequency response;

112

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION •

for commercial frequency response services, there is also a distinction between firm and optional contracts; and

•

for standing reserve services, providers can select whether they provide services on a flexible or committed basis.

6.55

It is not envisaged that the emergence of new distribution level providers of frequency response and reserve will necessitate significant changes to these arrangements. As arrangements for both of these services can already accommodate distributed generation, it is likely that any changes should relate to the minimum size of participants.

6.56

For services to DNOs, similar contractual forms could also be adopted although it appears likely that for network support related services, DNOs will be keen to ensure service delivery which would thus favour firm and committed approaches.

6.57

There is an ongoing discussion regarding Grid Code requirements for distributed generation. Whilst mandatory arrangements will feature for the largest distributed generators connected at or above 33 kV, many DG ancillary services are likely to be commercial in nature.

6.58

As reliable service delivery could be critical to DNOs, there will be a high probability that contracts will need to address the issues surrounding non-delivery of service.

6.59

Similarly, in situations where service recipients seek to influence generator-operating regimes, contractual provisions regarding enhanced commercial availability may be required.

Payment arrangements 6.60

There are a variety of different remuneration arrangements, which could be utilised to reward generators for the provision of services to DNOs and TSOs. These include: •

Flat fees: These represent a simple arrangement whereby the generator would receive payments irrespective of season, time of day, running status, delivery requirement and quantities delivered.

•

Time based charges: These are effectively a variant of the flat fee approach but time capped according to windows within which the recipient requires the service.

•

Capacity based charges: Capacity based arrangements can be used to secure a proportion of a generator’s output where the provider does not necessarily require the full generation capability.

113

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

6.61

•

Option fees based on availability: Similar to capacity charges, although these can incorporate time based elements in addition to capacity in order that the recipient pays to secure capacity as required.

•

Exercise fees based on utilisation: Exercise fees can be useful in that they provide a linkage to actual usage of a service and output. In an energy context, such arrangements are often used to recover operating and fuel costs.

•

Reduced DUoS payments: Whilst each of the above options can be used as an incremental means of remunerating generators for new services provided, incorporating the value of such service within use of system charges represent a more holistic approach.

It should be noted that such approaches are not mutually exclusive, i.e. combinations of the above approaches could be implemented. The advantages and disadvantages of each approach are outlined in greater detail in Section 7.

Procurement Arrangements for Different Ancillary Services 6.62

6.63

From Sections 4 and 5, it can be concluded that the most likely ancillary services to provided by distributed generators will include: •

Frequency response and Standing Reserve; and,

•

Network Security Contributions.

Potential procurement arrangements for each of these services are discussed in greater detail in this section.

Frequency Response and Standing Reserve Services 6.64

DG already has the ability to provide frequency response and standing reserve services to the TSO through either direct contractual relationships or indirectly through aggregators. It is envisaged that the current arrangements for commercial frequency response services and non-BM standing reserve will continue.

6.65

It should be noted that the majority of distributed generators seeking to provide these services would be non-renewable due to the strong incentives to maximise generation outputs, i.e. avoidance of partloaded and standby operation.

6.66

The main constraints regarding the participation of the smaller generators in each of these markets relate to the technical capabilities of generating equipment (as discussed in Section 5), the commercial attractiveness of providing the service to the TSO (revenue available),

114

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION the infrastructure required for participation and particularly the availability of aggregation services. 6.67

It is anticipated that the TSO will continue to require contributions from providers of a scale that is useful to the management of the transmission system. As response and standing reserve both require service delivery in quantities of over 1 MW, any direct participation opportunities will only apply to the larger distributed generators, i.e. those already participating.

6.68

In order for smaller generators (<1MW) to participate in either of these markets, it will be essential for new aggregation arrangements to evolve. The current aggregation arrangements require expensive dedicated communication and control infrastructure, which would be cost prohibitive for smaller generators. Alternative arrangements will be required potentially based upon GSM type technologies as discussed in Section 5.

6.69

Assuming that infrastructure barriers can be overcome, distributed generators with either frequency response or standing reserve capabilities will probably require the services of aggregators.

6.70

A range of different parties could provide aggregation services. These include: •

dedicated 3rd Party Aggregators;

•

suppliers within large vertically integrated groups;

•

stand alone suppliers; and,

•

DNOs.

6.71

It is interesting to note that aggregation services have been slow to develop, only being offered by stand alone suppliers and dedicated 3rd party aggregators, i.e. the largest market participants do not offer such services to distributed generators.

6.72

For DNOs this is perhaps not surprising, as aggregation requires active energy trading capabilities and market interfaces (e.g. BM), which DNOs do not have. Whilst the aggregation of energy outputs is not a natural DNO activity, this does not preclude DNO involvement in instances where the DNO requires delivery of other services, e.g. network security contributions from micro-CHP.

6.73

The reluctance of the large vertically integrated groups to provide these aggregation services can perhaps be explained by the potential impact on their own transmission connected generation portfolios. As transmission connected generators currently provide the bulk of

115

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION frequency response and standing reserve, any increased participation from DG would effectively displace their own plant. 6.74

Another feature of standing reserve, which undermines the attractiveness of independent aggregation services, is where the aggregator must assume responsibility for the delivered energy. Unless the aggregator can sell the aggregate DG output in the wholesale market, this energy will be spilled in the BM at potentially low prices. It would not make economic sense for any supplier to pay a premium for such energy as the supplier was equally well placed to procure the energy directly. Consequently, such aggregation only appears attractive for aggregators with associated supply businesses.

6.75

Independent aggregation services (where the aggregator does not assume any responsibility for the delivered energy) rely upon a supplier’s registration of the distributed generator in settlement to offset the energy produced. In such instances, aggregator instructions to distributed generators to provide standing reserve will impact upon the associated supplier’s BM imbalance exposure. This is not an ideal situation for suppliers, although the financial implications associated with the delivery of unpredicted energy volumes into a demand account are not currently onerous. The low volumes of standing reserve energy, relative to the overall size of a supplier’s NETA demand account, may not be significant. Also any energy delivered would cause the supplier demand account to go ‘long’ such that the supplier would receive a payment from the BM.

6.76

Whilst the latter option is currently a minor feature of the ancillary services market for standing reserve, and suppliers tolerate the related imbalance implications for their demand accounts, it is not clear that this arrangement is robust to large-scale expansion.

6.77

The contractual options for aggregation are illustrated in Figure 19. As can be seen, a variety of aggregation contractual arrangements are possible depending upon which parties undertake the aggregation activity. The right hand arrangement could be utilised in situations where the DNO was procuring other services from the same portfolio of distributed generators, e.g. security from microCHP. Also similar arrangements could evolve if the DNO (or any other party) had invested in a communication infrastructure between a central point and customer premises, e.g. smart metering initiatives.

6.78

It can be concluded that increased participation from smaller distributed generators in the frequency response market and especially the standing reserve markets, will require new forms of aggregation to evolve, utilising lower cost communication infrastructure.

116

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Figure 19 − Contracting options for Aggregation

TSO Supplier

Aggregator

DNO Supplier

Generator Contract Network Security Services 6.79

The high-level valuation methodology adopted for all network support services relate to the deferred or avoided costs of network investment. This assumption would apply to network security contributions, voltage support and power flow management.

6.80

Distributed generation can also substitute for installation of remote control and switching facilities, and as explained in section 4, this is an area that is particularly relevant when considering security contribution of intermittent generation, such as wind.

6.81

As such network services are all highly locational in nature, it is likely that bilateral contractual arrangements will evolve before there is sufficient DG penetration to enable more market-based approaches.

6.82

Also, as network security services appear to be the most promising DNO service, especially in the medium term, this section focuses on the procurement, valuation and funding arrangements, which could apply to these.

Historic Procurement of DNO Network Security 6.83

Before evaluating the potential procurement process for DG sourced network security, it is worth reviewing the arrangements which have historically been adopted for network asset based security contributions.

117

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 6.84

By monitoring network usage on a local basis, DNOs are able to determine whether any areas of their network are likely to become non-compliant with network security standards. Consequently DNOs will have records indicating the margin by which demand groups exceed ER P2/5 planning requirements.

6.85

Should a situation arise where a particular demand group is drifting out of compliance (e.g. through load growth), the DNO will typically develop options to resolve any security shortfalls. Typically this will involve network based solutions due to the low penetration of distributed generation and the restricted recognition of DG security contributions as outlined in ER P2/5.

6.86

When developing network solutions, DNOs must estimate the magnitude of a particular security requirement over a wide timeframe due to the long operational lifecycles associated with network assets, e.g. DNOs would not wish to replace a new transformer as a consequence of inaccurate load growth forecasting.

6.87

After designing the preferred network solution to a particular security problem, the DNO is then able to fund the purchase and installation of the related assets through CAPEX budgets.

6.88

A consequence of such funding is that the DNO is able to secure a long and guaranteed rate of return on the assets employed, irrespective of whether the initial network usage assumptions proved to be accurate. The DNO is thus protected from risks associated with stranded assets.

6.89

An interesting feature of this procurement process is that the DNO identifies the initial problem, develops potential solutions and implements the necessary upgrades, i.e. the DNO undertakes the role of purchaser and provider of the service. Also, it should be recognised that only the DNO understands the network security requirements. These considerations will be relevant when considering competing security arrangements from non-DNO sources.

Proposed network security procurement process including DG 6.90

In the same way as at present, DNOs would monitor their networks to ensure compliance with security standards. In the event of a demand group drifting outside planning standard compliance, the DNO would determine nature of any security shortfall in terms of timescale and duration. The DNO would also need to identify any constraints regarding the application of network solutions, e.g. environmental or planning obstacles.

118

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 6.91

Upon determining the network security requirements to ensure planning standard compliance, the DNO would then be required to determine whether an existing or a proposed (firm) distributed generator could provide the necessary contribution within the demand group.

6.92

The contribution of different generation technologies would be determined by factors outlined in the ER P2/6 proposal based upon availability. In the event that DG could provide the required security contribution, the DNO would then be required to seek the permission of the generator to utilise this contribution before committing to network investment.

6.93

In the event that, under normal operating conditions, the generator would not be able to provide sufficient contribution, the DNO may wish to explore whether the generator’s contribution could be enhanced to meet the desired requirement.

6.94

In both instances, the DNO and the generator would enter into a commercial contract regarding the provision of a security contribution. It should be noted that the existing regulatory treatment for procuring network security does not accommodate such arrangements.

6.95

A consequence of such an approach would be that existing DG sources of network security would be utilised (if appropriate) before initiating network investment.

Valuing DG security contributions 6.96

The valuation methodology for the provision of network security contributions would most likely be based upon the avoided costs of network reinforcement. The reason for this is comparative costs of network and generation solutions. In the majority of cases, network solutions will be of lower cost than a generation solution so the DNO would not ever be incentivised to pay more for network security than could achieved through a network solution.

6.97

The valuation methodology would also need to address the following critical questions:

6.98

•

How much security contribution would the DNO procure?

•

For what duration will the DNO require the contribution?

If network and generator based solutions to network security shortfalls were to be treated equally, the DNO would be required to procure the generator’s entire security contribution (even if this took

119

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION the DNO beyond network planning requirements), over the full operating life of the generator. 6.99

An alternative arrangement would be for the DNO to only procure sufficient generator security to meet the planning standard. The duration of security contracts with generators could also be shortened from the full operating life of the generator. An example of an alternative timescale could be the duration of a price control period.

6.100 The arrangement equivalent to a network solution would be advantageous to the generator as it would provide certainty regarding income streams and could also improve the economics of new distributed generation projects. The above alternative arrangement might be more attractive to DNOs in that more flexibility is offered. 6.101 In future, as the level of DG deployment increases and the density of DG within a particular load groups also increases, it may be possible for the DNO to procure security contributions through a competitive tender process from a range of providers. 6.102 Alternatively, it could be possible for DNOs to influence the location of new distributed generators through the availability of security contracts.

Contractual structures for DG network security provision 6.103 The nature of contracts between DNOs and distributed generators is likely to depend on the degree of control over generator operating regime required by the DNO. 6.104 In instances where a generators’ normal operating regime had historically provided the DNO desired security contribution, it is unlikely that the DNO would need to influence generator operation. In such a case, it may be appropriate to reward the generator with a flat fee arrangement based upon £/MW contribution or £/month. Obviously, the DNO would need to be informed of any proposed changes in the generator’s operating regime. 6.105 Where DNOs sought a generator security contribution in excess of that delivered through normal generator operation, it is likely that the generator would be required to respond to DNO operational instructions to enhance security contributions. In such cases, service delivery becomes a more important consideration and contracts may reflect this. By structuring rewards according to availability (£/MW) and utilisation (£/MWh), the DNO would effectively incentivise the generator to deliver when required.

120

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 6.106 Taken one step further, DNOs will, no doubt, want to evaluate the implications of non-delivery, which could lead to the prospect of penalty clauses within generator contracts.

Expected Level of Delivery 6.107 When setting penalty arrangements for under delivery (or rewards for over delivery) it must be recognised that the contributions prescribed by ER P2/5 are delivered with a probability that is less than one. Assuming unit availabilities of 86% (as in ER P2/5), and the probabilities of delivering and not delivering the allocated contribution are shown in Table 18.

121

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION Table 18 - F factors of identical non-intermittent units of 86% availability with the probabilities of delivering these contributions49.

Number of Units

F Factor (%)

Probability of Delivering Contribution

Probability of not Delivering Contribution

1

49.9

86.0

14.0

2

60.7

74.0

26.0

3

65.6

94.7

5.3

4

68.0

90.3

9.7

5

69.6

85.3

14.7

6

71.1

80.0

20.0

7

72.5

74.4

25.6

8

73.5

91.1

8.9

6.108 We observe that the probability of delivering the allocated contributions varies between about 75% and 95%. This means that a generating system comprising 7 units is expected to deliver its contribution of 72.5% in only three out of four occasions (more precisely in 74.4% of all situations). In other words, it is expected that the generators will not deliver their allocated contribution for 25% of the time. This concept is carried forward and is embedded in the proposed update ER P2/6. This feature of the security standard should be adequately reflected in the penalty/reward arrangement and the generators should not automatically be expected to deliver on every occasion.

Basis for non-delivery payments 6.109 When setting non-delivery mechanisms for DG in the context of the contribution they could make to network security, it is important to consider the basis on which the value of the penalty payment should be determined. 6.110 A penalty could be expressed in terms of a reducing generator security contribution following non-delivery of the expected

49

For wind generation, the probability of delivery of the allocated contribution is about 40%

122

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION performance. This would be consistent with the principles of P2/5. The generator could be allocated a lower contribution value (F factor in P2/6) and this would lead to corresponding reductions in option fees for subsequent periods. The same principle can be applied to generators offering enhanced security contribution. 6.111 The penalty for non-delivery should not be linked to DNO exposures associated with increased CMLs and CIs caused by generator unavailability. This is because there is no direct link between the contribution that the generator makes to network security in the context of ER P2/5 and actual network performance as measured by CMLs and CIs50. Clearly the contribution to network security and the contribution to network performance are two different matters51. 6.112 In the longer term, therefore, there may be a requirement to develop security standards that are consistent with DNO output performance measures.

Safeguarding the procurement process 6.113 As already discussed, the DNO as service procurer and potential service provider occupies a powerful position in any negotiations with a generator. Indeed, the current regulatory arrangements incentivise DNOs to pursue network solutions through CAPEX. Also there is an information asymmetry between service procurer and provider. Consequently, there may need to be safeguards built into the procurement process to protect the interests of generators. 6.114 Taken to the extreme, a DNO with a network security shortfall may be aware of a generator on their network with an operating regime that would resolve the problem. In such a circumstance, the DNO might avoid declaring their security requirement whilst utilising the generator’s contribution for compliance purposes, and the generator would remain unaware. 6.115 Examples of potential safeguards to avoid this happening might be: •

A requirement on DNOs to declare the level of network security headroom for the different demand groups on their network.

50

It should be remembered that ER P2/5 does not deal with the question of outage frequency but is only concerned with the maximum restoration times.

51

The recent update to ER P2/5 was developed to allow a simple and straightforward extension to the existing ER P2/5 Table 2, such that contributions from modern generation technologies could be recognised in the short term. The approach adopted was consistent with the original ER P2/5 methodology and hence unrelated to CIs & CMLs.

123

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION •

A requirement on DNOs to evaluate the availability and suitability of DG security contributions before committing network investment.

•

A requirement on DNOs to seek the approval of generator before being allowed to use the generator’s security contribution for compliance purposes. Such an obligation could benefit DNOs as it could incentivise generators to enter formal contracts with a DNO.

6.116 There are a number of issues here concerning the procurement of security contributions from DG, which may require further work to develop more comprehensive arrangements. It may be beneficial to develop a consistent framework, which could be applied nationally. Ofgem’s newly formed Distribution Commercial Forum could take such arrangements forward.

Interactions between different distribution level ancillary services 6.117 In an era of significantly increased levels of DG operating on active distribution networks, the opportunities for DG to provide ancillary services may increase and also situations could arise in which individual generators were contracted to provide multiple services. 6.118 This section evaluates the implications of increased levels of active management on ancillary service provision and also investigates whether conflicts could arise regarding the provision of different services.

Compatibility of ancillary service provision with active management 6.119 As discussed earlier, passive networks are designed to accommodate the full range of envisaged operating parameters (including generator outputs) with a minimal requirement for network monitoring and reconfiguration. Such approaches are sometimes referred to as ‘fit and forget’. Features of passive networks tend to be a strong reliance on primary infrastructure such as transformers, switchgear, OHLs and cables, which results in electrically robust systems. Passive networks tend to be less well developed in terms of secondary infrastructure, which includes communications, monitoring, and control equipment. 6.120 By contrast active networks are less reliant on primary infrastructure although more reliant on secondary infrastructure. A definitive statement regarding which approach to network managements is most advantageous to ancillary services from DG is not possible, as the requirements of the TSO services differ from those of DNO services.

124

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 6.121 Ancillary services to the TSO from distributed resources will most easily be accommodated from passive networks. This is because, irrespective of operating status, the distributed generator should be able to respond to TSO instructions if it is within the generators’ capability. However, on active networks there is an increased likelihood that due to distribution network constraints, certain modes of operation may not be permitted by the DNO. Consequently there could be increased delivery uncertainty regarding TSO ancillary services originating from distributed generators on active networks. 6.122 Conversely, there will be most scope for ancillary services to DNOs from distributed generators connected to active networks. This relates to the reliance on fewer primary assets. In the case of a passive network, it is unlikely that the DNO will require as many services from distributed generators, as will be the case in an actively managed scenario. Also the availability of more sophisticated secondary infrastructure on active networks will facilitate more generator participation. 6.123 It is clear however that increased levels of DNO secondary infrastructure will facilitate more DG services to both the TSO and DNOs.

Hierarchy of services 6.124 In instances where a distributed generator is providing multiple services to both the TSO and DNO, it is useful to determine which service takes precedence in the event of conflicting instructions. Examples of such services could be: •

Local Voltage Support;

•

DNO Network Security; and

•

frequency response and reserve services.

6.125 Determining the relative importance of such services can be accomplished by considering the extent to which a service is localised or national and the timescales for compliance with the relevant network performance requirement. 6.126 Applying such tests reveals that voltage, as an absolute localised requirement, represents the highest priority due to the absolute nature of statutory limits. The requirements to provide security contributions will be slightly less onerous due to the time based nature of the service requirement (as it might only be required at peak) and the time constants associated with thermal overloads. Frequency Response and reserve, as national services, represent the

125

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION lowest priority for such generators, as it should be possible for the TSO to instruct alternative providers of the service. 6.127 Such hierarchies will influence the types of contract that the distributed generator selects to enter. In the above example, the generator would be unlikely to enter a firm frequency response or reserve contract.

Information flows for service provision 6.128 When a distributed generator receives an instruction to provide a particular service, there could be implications for other parties involved along the supply chain. This section evaluates which parties would need to be informed of service initiation and also focuses on supplier implications.

Party notification requirements 6.129 In the event that the TSO issues an instruction for a distributed generator to provide a frequency response or a reserve service, there will inevitably be a change in generator output. A key question relates to whether the DNO would need to be informed of such an instruction. 6.130 As already discussed in the active management section above, it is highly likely that a passive network would accommodate the full operating range of a distributed generator without the requirement to reconfigure. This implies that the DNO would be indifferent as to the service requirements of the TSO. 6.131 On an actively managed network, the DNO would require details of the generators full operating range and the nature of any services to be provided. The DNO would then design the network to prevent unsafe situations arising. This would merit careful DNO consideration during the design phase. In the event of a TSO instruction to change generator output, the active management control arrangements would automatically react to prevent undesirable outcomes (e.g. voltage rise) so again the DNO would not need to be informed of each TSO instruction. The DNO’s initial information requirements may become more detailed in such situations. 6.132 A complication for the generator in such situations is that the active network could intervene and prevent service delivery to the TSO. Such instances would need to be addressed within contracts. Inevitably, any delivery uncertainty would undermine service value.

126

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

Supplier implications of ancillary service provision 6.133 A more general consideration is the effect that service delivery instructions (from either the TSO or DNO) could have on suppliers. As most distributed generators output will be settled through a suppliers NETA demand account, any changes in operation will impact upon the supplier’s imbalance exposure within the BM. 6.134 Critical to the impact on supplier imbalance are the timescales within which service delivery is required following initiation (the key timing consideration being NETA Gate Closure). In the event that a supplier could be provided with sufficient notification of an ancillary service delivery instruction, the supplier could accommodate the proposed change in output before Gate Closure. Should ancillary service instructions be issued within an hour of Gate Closure, the supplier will inevitably have an increased imbalance exposure. 6.135 Whilst such supplier exposures may be insignificant and thus tolerable with relatively low levels of DG ancillary service provision, this cannot be guaranteed should DG ancillary service provision become more commonplace. In such situations suppliers would seek notification of service instructions to amend their commercial positions. For instructions implemented within Gate Closure timescales, suppliers may choose to factor such risks into contract prices for energy. This would probably be most significant for instructions to de-load as the supplier demand accounts would go ‘short’, thus exposing supplier to system buy prices.

127

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

7.

IMPACT ASSESSMENT

7.1

This section considers the impacts upon different market participants involved in the supply and delivery of ancillary service from distributed generators.

7.2

The attractiveness of different ancillary services will depend on the materiality of each service to the service provider relative to other income streams. Consequently, it is important to estimate indicative values for each service.

Impact on Distributed Generation 7.3

In order to consider the impact of ancillary service provision on distributed generation, it is necessary to evaluate the services provided by different types of distributed generator.

7.4

By adopting the capabilities of the different DG technologies as discussed in Section 5, the impact on specific types of DG can be considered.

Frequency response 7.5

As the only new distributed technology with a consistent capability to provide low frequency response services is wind power utilising Doubly Fed Induction Generator (DFIG) technology, it is most appropriate to consider the impact of frequency response in this context.

7.6

As has already been discussed, the requirement for wind power, or indeed any renewable source, to provide frequency response will be comparatively rare due to the part loading requirements and the costs associated with losses of ROC revenue.

7.7

The only time that such generators are likely to enter frequency responsive mode is during periods of low national demand (summer nights) when renewable output may need to be curtailed.

7.8

In such circumstances, the wind generator will be de-loaded according to TSO instructions by exercising a high bid price to recompense the generator for the loss of ROC revenue.

7.9

Upon entering frequency responsive mode, the generator might receive a payment of £4/MW/h (assuming the generator was capable of both primary and secondary response at current prices). So assuming a 100 MW wind farm was required to provide this service during summer weekends (26 occasions) for approximately 4 hours

128

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION per night, the addition revenue earned would equate to £4 x 26 days x 4 hours x 100 MW = £41,200 per annum, i.e. £0.40/kW. In the context of a 100 MW wind farm with 30% utilisation factor, the annual ROC revenue would equate to approximately £14m, i.e. payments for low frequency response services would add less than half of one percent to the wind farm’s revenues. 7.10

With the level of frequency response income being so low, it is questionable whether the wind developer would recover the costs of the required infrastructure.

7.11

By contrast a 400 MW flexible CCGT earning approximately £50m per annum from energy sales, could earn up to an additional £1m per annum from frequency response services (£2.50/kW), which represents a 2% increase in revenues.

Standing reserve 7.12

In the standing reserve market at present, the most flexible plant can earn approximately £23/kW52 per annum from standing reserve services. It should be recognised that the costs of entry for the lowest cost OCGT plant are in excess of £45/kW53 per annum. Consequently, the standing reserve market is not attracting new entry at present.

7.13

However, this does not imply that the value of standing reserve will remain these levels in the future. With the increased deployment of intermittent generation, the value of standing reserve could increase significantly over the next decade although would not exceed the cost of new entry but this does not imply that the provision of such services would be unattractive to existing standby generation.

7.14

Applying such data to a 10 MW distribution connected standby generator, which would require periodic testing, an additional revenue stream of approximately £70,000 per annum could be generated for an activity that the operator would potentially have undertaken anyway. Such an arrangement would appear to be an attractive option and perhaps explains the increasing number of nonBM providers.

7.15

By contrast, should the infrastructure become available to facilitate the remote aggregate switching of many thousands of micro CHP installations, standing reserve could be provided from such sources. Indicative infrastructure capital and operating costs have been estimated to be between £10 and £15 per annum for each 1 kW

52

Market participant figure.

53

ILEX data.

129

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION device. The fact that not all such devices would be able to respond to TSO instructions (some would be switched off, already generating or heat limited) would result in a considerably lower value for the service. Should the most effective provider currently be able to earn £23/kW per annum, the uncertainties associated with the delivery and the duration of service from micro-CHP could reduce this figure potentially below £7/kW. This figure is gross of any fee paid to the aggregator. 7.16

At such levels, the service would not cover the costs of the infrastructure unless the communication infrastructure could be used to facilitate other services such as smart metering. Even if the value of the service were to triple, it is difficult to envisage an income of an extra £20 per annum (before infrastructure costs) influencing a customer’s selection of heating system.

DNO network security services 7.17

To investigate the materiality of DNO network security services, consider the different options for resolving a network security shortfall as illustrated in Figure 20. In this 33 kV, 20 MVA example, an additional 5 MW security contribution is sought which can be provided by three options. These investments are as follows: Figure 20 − Options to address a 5MW security shortfall at a 33/11 kV, 20 MVA substation. Reinforcement Options: Additional 5 MW Requirement

Option A X 20

20

G

Option C X

X

X

Option B

Option A: Extend/Reinforce Substation

Option B: Adopt a Generation Solution

• Additional 33/11kV transformer

•New generator

• Two additional circuit breakers

•Utilise existing generator

• Reinforcement of 33 kV OHL to primary sunstation

130

Option C: Invest in Transfer Capacity

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

7.18

•

Option A: Reinforce the existing substation. The costs of this option were calculated for a range of different infrastructure requirements. The simplest reinforcement was to add a single 20 MVA transformer. The most involved reinforcement assumed the additional requirement to upgrade the 18 km circuits at 33 kV supplying the substation with underground cables. Other reinforcement options were also considered.

•

Option B: Investment in a generation solution. By utilising the least cost OCGT technology with a security F factor of 50%, a 10 MW generator could be used to provide an additional 5 MW security contribution.

•

Option C: Invest in 5 MW of transfer capacity at 11 kV involving a new 10km OHL or cable.

The range of annualised cost estimates54 for each option is included in Table 19 below. As can been seen, all of the network solutions are of lower cost than the generation solution. Whilst the least cost option is to invest in 5 MW of transfer capacity (assuming the circuit to be OHL), the simple substation reinforcement delivers 20 MW of security contribution for less than twice the cost of the transfer capacity option. The annualised cost of the generation solution is more than 20 times that of the cheapest network solution. It can also be seen that the most complicated network reinforcement (involving a new 33 kV cable) approaches the cost of a generator solution although it delivers 20 MW rather than the 5MW from the generator. Table 19 − Comparison of costs of delivering additional security contributions55 Cost hiera rchy & Security Contribution Transfer Capacity OHL

£21k pa

5 MW

Reinforce substation only

£38k pa

20 MW

Transfer Capacity Cable

£60k pa

5 MW

Reinforce substation & one new 33 kV OHL

£107k pa

20 MW

Reinforce substation & reconductor 2 x 33 kV OHLs

£111k pa

20 MW

Reinforce substation & one new 33 kV Cable

£467k pa

20 MW

54

A 6.5% cost of capital and a 20 year depreciation period was assumed for network assets. OCGT Generator (inc O&M) £545k pa 5 MW A 13% cost of capital and a 10 year depreciation period was assumed for the generator.

55

Network cost data from EDF Energy. Generation costs from ILEX market reports.

131

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

7.19

Assuming the DNO would pay the generator its avoided costs of network reinforcement, such payments would range between £10k and £120k (£1/kW - £12/kW) per annum depending upon the complexity of the network solution. It is anticipated that most reinforcements would be at the lower end of this range.

7.20

Assuming the average costs of network reinforcement to be £30k per annum, this could represent an attractive service for a standby generator in the same way that standing reserve services are also attractive. It appears from this analysis that network security payments and standing reserve payments could be of a similar order of magnitude.

7.21

This analysis also suggests that in rare instances, where the costs of network solutions are high the value of generation contributions could increase significantly.

7.22

In the context of a biomass generator with 70 % availability, the ROC income would equate to approximately £4m per annum whereas a similar non-renewable generator could expect energy revenues of approximately £1.3m per annum56. The magnitude of network security payments relative to these energy revenues is less than 1 per cent for the renewable generator and less than 3% for the nonrenewable generators. In both cases the value of security contributions is less than the forecasting errors for energy revenues.

7.23

In instances where DNOs are prevented from investing in network solutions and there is a suitable distributed generator located within the demand group, it could be advantageous for the generator to provide security to the local network. Such applications will represent niche opportunities where, environmental, planning or terrain related issues constrain DNO activities.

7.24

It is unlikely that a decision to invest in a distributed generator would be influenced greatly by the prospect of security contribution payments. However, a generator with a choice of locations, might select the one with security opportunities.

56

ILEX projections.

132

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 7.25

Overall, it appears as if the opportunities for DG to provide ancillary services to the TSO and DNOs will be technology, service and location specific. Where a distributed generator is able to provide such services, it appears likely that the incremental revenue streams will be regarded as useful supplements to the original business case justifying the initial investment. It is unlikely that any individual service would merit investment in generation plant alone.

7.26

Inevitably, niche applications for the provision of services by distributed generators will emerge. These will normally relate to situations in which network constraints apply, preventing the network operator from pursuing a network solution or where the network solution is prohibitively expensive. Such opportunities will usually relate to distribution, rather than transmission system constraints.

7.27

In the longer term (post 2015), with sufficient distributed generators connected within individual load groups, there may be increased opportunities for DG to provide voltage and flow management services, owing to increase in aggregate availability of generators. Further work could be undertaken to evaluate such opportunities.

Impact on DNOs 7.28

In the short to medium term, the ability of DNOs to source network security services from distributed generators will not fundamentally change the way in which DNOs operate their networks.

7.29

The availability of network security contributions from distributed generators will broaden the scope of potential solutions for DNOs when accommodating any network security shortfalls. Such contributions will be particularly useful to DNOs in situations where constraints restrict the adoption of conventional network solutions. These situations could include environmental, planning or terrain related constraints.

7.30

A potential problem for DNOs relates to the current regulatory framework in terms of CAPEX and OPEX funding distinctions. At present, network security is procured through CAPEX, which does not currently accommodate generation contributions. The current arrangements for Operational Expenditure (OPEX) are not ideal either in that the DNO could be financially penalised for funding DG contributions through this route. More work will be required to establish a suitable funding mechanism for network security and support.

7.31

In order that a consistent and transparent set of arrangements can emerge to facilitate increased security contributions from DG, it will be necessary to establish principles (and potentially standardised

133

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION commercial arrangements) for procurement processes and valuation methodologies. Indeed, the Distribution Commercial Forum might consider the development of a standardised ‘model form’ contract suitable for localisation by individual DNOs. 7.32

A major concern of DNOs regarding the reliance on distributed generators to provide network support services is likely to relate to non-delivery risk exposures. Such exposures could be financial, regulatory or legal in nature. Consequently, the issues associated with service non-delivery require further exploration.

7.33

The extent of opportunities on DNO networks will largely relate to load growth and asset replacement profiles. Whilst it has not been possible to determine which of these two will provide the greatest number of opportunities, it may be useful to evaluate this distinction further in future.

Impact on the TSO 7.34

The provision of ancillary services from DG to the TSO will effectively broaden the choice of operators available to the TSO and potentially enhance competition.

7.35

Similarly, the emergence of new providers of existing services may also be advantageous to the TSO facing increased balancing service requirements. In particular, the increased availability of standing reserve from distributed players could be particularly useful with increased volumes of intermittent generation on the system. It should be noted that the TSO currently secures less than 800MW of standing reserve from non-BM participants whereas the actual opportunity may be in the order of 5 GW.

7.36

With respect to intermittent generation, the TSO’s concerns regarding system stability have been widely documented elsewhere. It remains unclear as to whether intermittent generation, as the source of many of these concerns, will actually be required to resolve any related network stability issues. The revenues associated with periodic frequency regulation from wind farms may not be sufficiently attractive to developers.

7.37

Increased numbers of aggregators and new aggregation services may evolve to facilitate increased levels of participation from DG in TSO ancillary service markets. Such developments could be beneficial to the TSO compared with a non-aggregated alternative, requiring the TSO to manage interfaces with many small distributed generators.

7.38

Whilst the current aggregation arrangements have been successful in encouraging non-BM participants into the standing reserve market,

134

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION the costs of the associated infrastructure could deter wider participation. In order to extend aggregation opportunities further, new low cost communication and monitoring arrangements should be evaluated. Such infrastructure could facilitate wider participation by smaller generators.

Impact on Suppliers 7.39

As the output from distributed generation is largely purchased by suppliers and settled through supplier demand accounts within NETA, suppliers must ensure they understand generator operating regimes and whether generator operation is likely to be affected by instructions to provide ancillary services. It should be noted that both NETA and the Supplier-Hub arrangements can accommodate TSO and DNO ancillary service provision from DG.

7.40

Supplier concerns will relate primarily to imbalance exposure in the BM. Unless suppliers are notified of scheduled ancillary service provision, it will not be possible for suppliers to revise demand forecasts accordingly. As has already been examined, instructions issued post Gate Closure will inevitably impact upon a supplier’s imbalance exposure.

7.41

Whilst the numbers of distributed generators remain low, supplier exposures are likely to be manageable. Indeed, instructions to increase generator output may not actually penalise suppliers as they could be cashed out favourably. Note that increased generator outputs will tend to cause supplier demand accounts to be ‘long’.

7.42

Obviously, a simple way for such exposures to be managed would for the supplier to act as the aggregator between the TSO (or DNO) and the generator. Alternatively, suppliers would need to ensure that they were notified in the event of ancillary service instructions being issued to generators with whom they had contracts. It will be interesting to observe whether large suppliers, owned by vertically integrated parents, choose to offer such aggregation services.

7.43

One other impact upon suppliers relates to the supplier’s renewable obligation targets. It is conceivable that a conflict could arise between a distributed generator willing to provide ancillary services through part load operation and a supplier incentivised to maximise renewable purchases. Any such conflicts would need to be managed by the affected parties.

Impact on Aggregators 7.44

New aggregation opportunities could emerge with increased numbers of distributed generators. In addition to the existing

135

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION opportunities relating to TSO ancillary services, it may be possible for aggregators to co-ordinate action from DG to provide network security contributions to DNOs. 7.45

At the TSO level, whilst the current infrastructure arrangements may not facilitate the participation of small individual distributed generators, new infrastructure may become available, thus extending participation. It might be beneficial for the TSO and aggregators to explore how such arrangements could be structured.

7.46

For those aggregators without a supply business (or a BM interface), it may become necessary for the aggregator to provide information to suppliers in future regarding their operational impact on contracted generators. This would enable suppliers to manage imbalance. Alternatively, it may be more attractive for potential aggregators to also operate supply businesses.

Overall impact on the British electricity sector 7.47

The provision of a broader range of ancillary services from a wider range of distributed generators could be beneficial in terms of the costs of operating the transmission and distribution networks.

7.48

The provision of ancillary services from existing generators potentially increases asset utilisation both for generation equipment and networks. Such increased utilisation could avoid or defer investments in network infrastructure, and would thus represent an overall cost saving.

7.49

In addition, the provision of ancillary services from distributed generation could in niche situations, avoid the negative impact of network investments in environmentally sensitive areas.

7.50

Increased DG participation in TSO standing reserve markets may increase the levels of competition in the future. This could be especially relevant if, through increases in intermittent generation, the scope for such services is set to expand.

7.51

The increased opportunity for DG to provide services may also increase the scope for aggregation and thus catalyse the emergence of new aggregators.

7.52

Specifically, it should be recognised that the provision of both DNO and TSO ancillary services from distributed generation should not undermine security of electricity supplies.

136

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 7.53

Similarly, it should also be recognised that the provision of ancillary services from distributed generation should not impact negatively upon Government climate change targets.

137

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

[This page is intentionally blank]

138

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

8. 8.1

CONCLUSIONS AND RECOMMENDATIONS This work evaluated opportunities for distributed generation to provide ancillary services to the TSO and DNOs. The ancillary services evaluated were: •

TSO Frequency Response;

•

TSO Regulating and Standing Reserve;

•

TSO Reactive Power;

•

DNO Security of Supply contributions;

•

DNO Quality of Supply services; and

•

DNO Voltage and Power Flow management services.

8.2

Although all of these services were explored in detail, only TSO frequency response, TSO regulating and standing reserve and DNO security of supply contributions represent realistic opportunities for distributed generators in the short or medium term.

8.3

CCGT and DFIG wind generators were the most promising technologies for the provision of TSO frequency response services whereas CCGTs, diesel standby generators and perhaps micro CHP were best placed to provide reserve services.

8.4

It was found that, to varying degrees, DNO security of supply services could be provided by most existing distributed generation technologies.

8.5

As the majority of existing DG has been installed for electricity supply purposes, very few generators are equipped with the infrastructure necessary to provide ancillary services. Such infrastructure includes governors, automatic voltage regulators, resynchronisation facilities and appropriate protection, monitoring and communication facilities.

8.6

The most appropriate commercial arrangements for response and reserve services appear to be market-based mechanisms. Ideally the TSO’s current arrangements could be extended. Expanded aggregation arrangements, utilising lower cost infrastructure, would facilitate increased participation from small generators.

8.7

The most appropriate commercial arrangements for DNO security of supply services appear to be bilateral contracts due to the local and site-specific nature of security requirements.

8.8

Opportunities for DG to provide ancillary services will undoubtedly increase as DG penetrations and availabilities increase.

139

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 8.9

The analysis undertaken suggests that value of the most feasible ancillary services will be relatively low. Consequently, such services will represent incremental revenue opportunities for DG. In general, it would not be possible to develop business cases for investing in DG solely on the basis of ancillary service income.

8.10

Niche opportunities will emerge for DG to provide ancillary services, usually in circumstances where constraints restrict network development, e.g. environmental, planning and terrain related constraints.

8.11

In an era with significantly increased levels of DG operating on active distribution networks, the opportunities for DG to provide ancillary services may increase. However, on active networks there is an increased likelihood that due to distribution network constraints, certain modes of operation may not be permitted by the DNO. Consequently there could be increased delivery uncertainty regarding the provision of TSO ancillary services from distributed generators connected to active networks.

8.12

In circumstances where a distributed generator receives conflicting instructions regarding the provision of different ancillary services, local services should take precedence over national services.

8.13

Higher penetrations of DG will increase DNO options regarding network operation and development decisions, which could (in certain situations) lead to lower overall costs.

8.14

Increased penetration of DG could also enhance competition in TSO markets for frequency response and reserve. This could be particularly relevant should demand for these services increase with intermittent generation.

8.15

As the output from distributed generation is largely purchased by suppliers and settled through supplier demand accounts within NETA, suppliers must ensure they are aware of generator operating regimes and also whether generator operation is likely to be influenced by ancillary service provision. Supplier concerns will relate to imbalance exposures in the BM and the fulfilment of ROC targets. Suppliers will require notification of ancillary service provision, in order to suitably revise demand forecasts. Ancillary service instructions issued post Gate Closure will inevitably impact upon a supplier’s imbalance exposure and potentially reduce the value of the energy supplied.

8.16

The introduction of Register Power Zones (RPZs), created to encourage DNOs to develop and demonstrate new, more cost effective ways of connecting and operating generation, could provide

140

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION an initial platform for the development of appropriate ancillary services contracts. In order for new technical solutions to become widely accepted, an appropriate contractual framework will need to be established, and RPZs could be used to develop such arrangements. 8.17

It is important to stress that provision of ancillary services from DG should not jeopardise or degrade security of supply and may even contribute to its enhancement in future.

8.18

It should be recognised that the provision of ancillary services from distributed generation should not impact negatively upon Government climate change targets

8.19

In addition, the provision of ancillary services from distributed generation could, in niche situations, avoid any negative impact of network investment in environmentally sensitive areas.

Recommendations 8.20

The extent of opportunities on DNO networks will largely relate to load growth and asset replacement profiles. Whilst it has not been possible to quantify the relative magnitudes of these opportunities within this project, such information will be critical to evaluating the market potential for network security services under alternative future development scenarios and should be explored further.

8.21

In order that a consistent and transparent set of arrangements can emerge to facilitate increased security contributions from DG, it will be necessary to establish principles (and potentially standardise commercial arrangements) for procurement processes and valuation methodologies. Indeed, the Distribution Commercial Forum might consider the development of a standardised ‘model form’ contract suitable for localisation by individual DNOs.

8.22

A major concern of DNOs regarding the reliance on distributed generators to provide network support services will relate to nondelivery risk exposures. Such exposures could be financial, regulatory or legal in nature. Consequently, the issues associated with service non-delivery require further exploration.

8.23

Whilst the current aggregation arrangements have been successful in encouraging non-BM participants into the standing reserve market, the costs of the associated infrastructure could deter wider participation. In order to extend aggregation opportunities further, new low cost communication and monitoring arrangements should be evaluated.

141

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION 8.24

A potential problem for DNOs relates to the current regulatory framework in terms of CAPEX and OPEX funding distinctions. At present, network security is procured through CAPEX, which does not currently accommodate generation contributions. The current arrangements for OPEX are not ideal as DNOs could be financially penalised for funding ancillary services through this route. More work will be required to establish a suitable funding mechanism for network security and support.

8.25

As this work explored ancillary services opportunities for DG in the short to medium term, alternatives for the long term need to be investigated given that there may be fundamental changes to the structure of power systems, particularly in the context of the recent Technical Architecture initiative.

142

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

[This page is left intentionally blank]

143

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

ANNEX A – TABLE 1 OF ER P2/5

Minimum demand to be met after Class of supply

Range of group demand

First circuit outage

Second circuit outage

A

Up to 1 MW

In repair time (Group Demand)

NIL

B

Over 1 MW to 12 MW

Within 3 hours (Group Demand minus 1 MW)

NIL

In repair time (Group Demand)

C

Over 12 MW to 60 MW

Within 15 minutes (Smaller of Group Demand minus 12 MW and 2/3 Group Demand)

NIL

Within 3 hours (Group Demand)

D

Over 60 to 300 MW

Immediately (Group Demand minus up to 20 MW (Automatically disconnected)) Within 3 hours (Group Demand)

E

Over 300 to 1500 MW

Immediately (Group Demand)

Within 3 hours (For Group Demands greater than 100 MW, smaller of Group Demand minus 100 MW and 1/3 Group Demand) Within time to restore arranged outage (Group Demand) Immediately (All customers at 2/3 Group Demand) Within time to restore arranged outage (Group Demand)

CEGB Planning Memorandum PLM-SP2 F

Over 1500 MW

Scottish Board Security Standard NSP 366

144

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

[This page is left intentionally blank]

145

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

ANNEX B – CONTRIBUTORS TO THE STUDY The authors would like to thank the following individuals for their input to this study:

Anthony Musgrave

British Gas Power Generation

David Tolley

RWE Innogy

Simon Cowdroy

Econnect

Ham Hamzah

RWE Innogy

Tony Woods

EDF Energy

Raoul Thulin

RWE Innogy

Paul Wynne

E.ON UK Renewables

Alan Laird

Scottish Power

Peter Turner

E.ON UK

Peter Thomas

Scottish Power

Mark Lewitt

Future Energy Solutions

Ron Allan

UMIST

Russell Reading

Gaz de France

Thomas Bopp

UMIST

Mark Bailey

Gaz de France

Nick Jenkins

UMIST

Andrew Malins

National Grid Transco

Derek Lumb

UMIST

John Greasley

National Grid Transco

Danny Pudjianto

UMIST

Lewis Dale

National Grid Transco

David Andrews

Wessex Water

Arthur Cooke

Ofgem

Christian Hjelm

Western Power Distribution

Gareth Evans

Ofgem

Nigel Turvey

Western Power Distribution

146

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

[This page is intentionally blank]

147

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

ILEX Quality Control Check Sheet ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENENERATION Report Unique Serial No: 2004/087

Director Stephen Andrews Date 20 September 2004

Project Manager David Porter Date 20 September 2004

Author David Porter Date 20 September 2004

Quality Control Beverly King Date 20 September 2004

ILEX is a member of Electrowatt-Ekono, part of the Jaakko Pöyry Group

148

ANCILLARY SERVICE PROVISION FROM DISTRIBUTED GENERATION

149

ILEX Energy Consulting King Charles House Park End Street Oxford, OX1 1JD UK

Tel: +44 (0)1865 722660 Fax: +44 (0)1865 722988 E-mail: [email protected] www.ilexenergy.com www.ewe.ch

ILEX Energy Consulting Ltd, King Charles House, Park End Street, Oxford OX1 1JD. Registered in England No. 2573801 ILEX is a member of Electrowatt-Ekono, part of the Jaakko Pöyry Group