Optimal Operation Of Wind/electric Utility Interconnecte

Optimal Operation of Wind/Electric Utility Interconnected Electrical Power System Using Neural Network

Faculty of Engineering, Elminia University, Elminia, Egypt 1

Object of this paper This paper introduces an application of an artificial neural network on the operation control of the WTG/ utility grid to improve system efficiency and reliability.

2

This paper focus on a hybrid system consists of wind system accompanied with battery storage interconnected with utility grid taking into account the variation of wind speed and load demand during the day. Different feed forward neural network architectures are trained and tested with data containing a variety of operation patterns. A simulation is carried out over one year using the hourly data of the load demand and wind speed at El'Zafranna site, Egypt as a case study. 3

The WES can be either connected to EU without BS or connected to EU with BS. The WES without batteries operates only when wind speeds and the EU supply present but do not supply local loads if the EU fails and if the wind speeds become low. If the WTG's exceed local load and excess power is exported, if they are less than local load the shortage will be supplied from the EU. If the EU fails the inverter disconnect from the EU, local load can be supplied from the battery and from WES. The battery will discharge if WTGs are less than local load and be charged when they exceed local load. WES with BS ensures that all renewable energy generated can be utilized. 4

The WES connected to EU with BS is extremely popular for homeowners and small businesses where backup power is required for critical loads, uninterruptible power supply, cost saving and power quality improvement, telecommunications backup and other necessities [2], [3], [4], [5]. The BS can then be recharged from future excess generation or during offpeak hours. Storage energy can be provided for local load or critical loads when market prices are the highest and consume when they are the lowest.

5

This paper describes a renewable energy system that can provide a total energy of 432*106 kWh to feed the load demand. The system has been designed to supply continuous power of 100 MW load and has the following capabilities: • Maximizes the electric power produced by the WES by detecting and tracking the point of maximum power. • Stores the future excess generation or during offpeak hours electric energy in lead-acid batteries. • Controls the charge and discharge processes of the batteries by using NN. • connects the load demand to the BS when the WES 6 and EU system fail.

WES Connected To EU With BS In this type of intertie system, the load demand has a WES, EU and BS as shown in Fig. 1. If load demand requirements exceed the WES, the shortfall is automatically made up by the EU. If the EU power fails, power will be drawn instantly from the backup batteries to support the load demand. Switching time in case of EU failure is so fast. The connection operation has been done by a neural network. Power flows from the system shown in Fig. 1 must P (t) ± P (t) ± P (t) = P (t) (1) 7 satisfy (1). WES g bat L

Wind Spee

S te p -u p T ra n sfo rm e r

I .G .

G . B .

D C /A C

A C /D C V

F ilte r

d c w

S 4 W b

B u s b a r

B a tte ry S to ra g e (B S )

~ Input

E U O u tp u t

N N fo r W T G /E U a c c o m p a n ie d w ith B S

S te p -d o w n T ra n sfo rm e r

S 3 W b

B u s b a r

S 1 W b

S 2 W b

L o a d

ig. 1. General configuration of the WES/EU/BS interconnected system. 8

The operation of the four switches shown in Fig. 1 can be summarized as shown in following Table .

Mode

1 2

S1Wb OFF ON

S2Wb ON OFF

S3Wb OFF OFF

S4Wb

Generated power vs. Load demand

1

PWES < PL i.e wind speed is very low

1

PWES > PL i.e. wind speed is high

3

ON

OFF

ON

1

4

ON

ON

OFF

0

PWES > PL i.e. wind speed is very high PWES =0 9

Load Characteristic It is assumed here that the load demand varies monthly. This means that each month has daily load curve different from other months. Therefore, there are twelve daily load curves through the year. Fig. 2 shows the daily load curves for January, April, July and October [6].

Fig. 2. The daily load curves for January, April, July and October [6].

10

System Simulation A new computer program has been proposed and written using Matlab software to simulate the system shown in Fig. 1. The flowchart of this program is shown in Fig. 3.

11

S ta rt R e a d H o u r y w in d s p e e d , P a r a m e te r s o f W T G , H o u r ly L o a d d e m a n d F o r i= 1 :1 :1 2

m o n th

F o r tim e = 1 :1 :2 4

B a tte rie s a re c h a r g e d i.e . S 4 W b = 1 a n d th e lo a d is f e d f ro m g rid

y e s

h r

If W in d s p e e d ( t) < L im it o r In v e r te r f a ilu r e

N o F e e d th e lo a d S 1 W b = O N

If S O C ( t) > 0 .8 * s iz e o f b a tte r y

y e s

N o

If w in d s p e e d ( t) > L im it o r P W E S ( t) > P L ( t)

S u rp lu s p o w e to B a tte r S 4 W b = 1 S 1 W b = O

r s e n d y , N

N o

If S O C ( t) > 0 .2 * s iz e o f b a tte r y

y e s

y e s S u rp lu s p o w e r s e n d to E U S 3 W b = O N

N o L o a d f e e d f ro m E U ;S 1 W b = O F F , S 2 W b = O N

L o a d f e e d f ro m b a tte ry S 1 W b = O N , S 4 W b = 0

E n d

Fig. 3. Flowchart of the operational modes of 12

The inputs of this program are: 1- Hourly wind speed. 2- Characteristic of WES module 3- Hourly load demand The outputs of this program are: 1- Generated power from WES. 2- Monthly surplus energy. 3- Monthly deficit energy. 4- Size of battery 5- State of batteries charge. The output of this program is used to be the input of NN. The outputs of NN are four trip signals that sent to switches S1Wb, S2Wb, S3Wb and S4Wb.

13

Proposed Network 

Figure 4 shows the structure of the proposed three layers NN. X1, X2, X3, X4 and t are the five-input training matrices and represent batteries state of charge, electrical power generated Fig. 4. Structure of the from WES, electrical power proposed three layers NN. for EU, load demand and time respectively. The network consists of 5 input layers, 7 nodes in hidden layers and four nodes in output layer which sigmoid transfer function. The network has been found after a series of tests and 14 modifications.

Figure 5 shows the evaluation of the 5+7+4 NN errors. NN have been used to make optimal operation control for power flow between WES, EU and BS

Fig. 5 Relation between error and Epoch for 5+7+4 NN 15

Figures 6 displays the optimal operation of the WES/EU accompanied with BS hour by hour through the day which represents the month of January.

g. 6 Optimal operation of the WES/EU/BS to fe the load demand during January (winter)16

Figure 7 reveals state of charge for BS which corresponding to the optimal operation of the WES/EU accompanied with BS through the months of January and July respectively.

Fig. 7 State of charge of WES/EU/BS during January  17 (winter) and July (summer)

From Fig. 6, Fig. 7 and Fig. 9 (January) it can be noticed that the trip signal which produced from NN sent to switch S1Wb at hours 5, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 and 22. This means that the WES or BS feed the load demand at these hours. On the other hand, switch S2Wb (for example) equal to 1 at hours 1, 2, 4, 5, 6, 7, 8, 9, 10, 14, 20, 21, 23 and 24. This means that the EU should supply the load demand at these hours. On the other hand, the power injected to EU through switch S3Wb at hours 11, 12, 15, 16, 17 and 18.

18

Fig. 9 Outputs of Neural Network for January 19

Finally,  battery  storage  will  be  on  state  of  charge  through switch S4Wb at hours of 1, 2, 3, 4, 6, 7, 8,  9,  10,  11,  12,  14,  15,  20,  21,  23  and  24.  On  the  other hand, the BS will be discharged through the  hours of 5, 12, 13, 16, 17, 18, 19 and 22. Then the  BS  can  feed  the  load  demand  only  during  these  hours accompanied with WES.

20

Conclusions

om this paper the following conclusions can be draw

1- A novel technique based on NN is proposed to achieve the optimal operation of WES/EU accompanied with 2-BS. The 5+7+4 is suitable neural network for accurate operation of WES/EU accompanied with BS at El'Zafranna site. 3- This NN has a very high accuracy as shown in Fig. 9, and achieve the optimal hour by hour operation for WES/EU accompanied with BS as shown in Figs.21

4- Application of operation strategy which proposed in this paper on the WES/EU accompanied with BS enhances the reliability of the system.

22

Thanks for your listening

23