Dokumen.tips_novel-antiscalant-dosing-control.pdf

DESALINATION Desalination 157 (2003) 209-216

ELSEVIER

www.elsevier.com/locate/desal

Novel antiscalant dosing control E.H. Kelle Zeiher, Bosco Ho, Kevin D. Williams* Ondeo Nalco Company, POB 11, Winnington Avenue, Northwich, Cheshire CW8 4DX, UK Tel. +44 (1606) 74488; email: [email protected]

Received 3 February 2003; accepted 10 February 2003

Abstract

Ii1this paper we identify a new technology to provide online, accurate, in-situ, reliable antiscalant dosing monitoring for any reverse osmosis system. Keywords: Antiscalant dosing control; Novel technology; Reverse osmosis

1. Background In the past decade, reverse osmosis (RO) systems have moved from being an exotic and vulnerable technology to being a reliable and robust conventional technology. As such, the use of reverse osmosis systems has become integral to the production o f pure water and ultrapure water in industry. In light of this, the reliability of the RO membrane becomes critical within the operation of industrial systems. Experts routinely tout the benefits of RO system monitoring and numerous technical papers extol the benefits of scale and fouling control. A *Corresponding author.

number of detailed texts and technical short courses are available to those interested in learning details of the RO operation [ 1]. In addition to the standard texts, the American Society for Testing and Materials (ASTM) offers various standards for membrane operation [2]. All sources agree that a consistent factor in the deployment of RO technology has been the need to prevent the fouling and scaling o f the membrane [3-5]. Fouling occurs when organic or biological material collects on the membrane surface, causing increased system AP and/or decreased permeate production. Most fouling is addressed by using pre-treatment techniques such as filtration. Scaling, which is the primary focus

Presented at the European Conjerence on Desalination and the Environment: Fresh Water for All, Malta, 4-8 May' 2003. European Desalination Society, International Water Association.

0011-9164/03/$- See fi'ont matter © 2003 Elsevier Science B.V. All rights reserved PII: $001 l - 9 1 6 4 ( 0 3 ) 0 0 4 0 0 - 4

210

E.tf. Kelle Zeiher et al. / Desalination 157 (2003) 209 216

of this paper, occurs when sparingly soluble minerals present in the feedwater precipitate on the membrane surface, usually at the tail end of an RO system. Scaling can be avoided or minimised using a variety of methods, including pre-treatment and addition of scale inhibitors.

2. Scale tbrmation and prevention Scaling occurs when sparingly soluble salts, present in the feedwater, are concentrated by virtue of the RO process. Table 1 shows how the water chemistry changes from the feed to the reject stream in a typical RO system. Common scales include silicates, sulphates, phosphates, and carbonates. Calcium carbonate is the most common potential scalant in brackish water industrial RO systems. Although lmmerous techniques are known to avoid scaling, it remains a serious problem in RO operation. A significant number of membranes sent for destructive (autopsy) analysis still have extensive carbonate scale on them [6]. Many RO operators consider scale formation to be unavoidable. Thus, a need exists for advanced monitoring and control techniques to prevent scale formation and the problems it creates. To review, the most common scale control techniques are Table I Scaling potential changes as water passes through a RO system Feedwater Rejectwater (75% recovery) Calcium, ppm CaCO3 150 558 Bicarbonate, ppm CaCO3 100 389 pH 7.9 8.4 Temperature, °C 20 20 LSI -0.08 1.69 Scaling No Yes Antiscalant required No Yes LSI, LangellierSaturation Index is used to predictcalcium carbonate scale formation accountingfor the contributory chemical factors, calculatedusing ROl 2 program [15].

• Lime softening • Ion exchange softening • Acid dosing • Speciality antiscalants Although well understood, hot and cold lime softening has largely been supplanted by ion exchange softening in recent years. Sodium regenerated ion exchange softeners technically offer a good method for scale control. However, the significant capital and operational expense involved with using these softeners make other alternatives more attractive. In addition, brine regenerate discharge can be an issue [7]. Acid dosing is often used to prevent CaCO 3 scale because acid is inexpensive on a weight basis and the dosage can be monitored easily with the use of a pH probe. This makes it ideal for use in systems in which remote monitoring is employed. The disadvantages of acid are high cost in high alkalinity waters, the handling and safety issues, and CO 2 production that increases permeate conductivity (thus making a degassifier necessary in certain applications). In addition, many commercially available acids are contaminated with metals that can accelerate lnembrane fouling. Antiscalant dosing is known to be very effective at preventing scale formation [8]. It is an attractive option because it has the demonstrated advantages o f operational and capital cost reduction, environmental acceptability and safety when compared to the alternative technologies. The disadvantage ofantiscalant dosing in the past has been the fear of product overdosing and the inability to easily confirm that product is present in the system. Unlike acid dosing, which is easily monitored by pH, or softening, which can be monitored by use of ion selective electrodes, antiscalant monitoring involved cumbersome laboratory techniques that were poorly suited to rapid, on-site evaluation [9]. The key to the effective performance of a n y scale prevention technology is that it has to function 100% of the time due to the dynamic nature of the water condition within the RO system. If the

211

E.H. Kelle Zeiher et al. / Desalination 157 (2003) 209-216

technology is off-line, because, for example, the salt has not been replenished for an IX softener or the antiscalant dosing pump has failed, then there are no mechanisms in place to prevent the formation of scale crystals within the operating RO system. Scaling conditions form in the RO system as permeate water is removed from the bulk water stream, leaving behind the dissolved solids. As the volume of the bulk water continues to be reduced, the concentration of solids begins to exceed the solubility limit of certain salts. The result is scale formation and precipitation. A simplified explanation of scale formation (crystallisation) is shown in Fig. 1. This process is continuous within the membranes and is driven towards the crystal formation with greater intensity close to the membrane surface (the boundary layer) and in the final membrane of the RO system's final stage. The failure to provide the inhibiting mechanism will allow this process to continue without restraint. Antisealant technologies work by delaying the growth phase of crystallisation, thus inhibiting the formation of crystals (Fig. 2). In order to be effective, the proper dose of antisealant must be present in the system at all times. Underdosing (or dosing interruptions) can lead to immediate scaling in waters with high salt concentrations. It is generally recognised by industry experts that a proper monitoring program is essential to operating any water treatment unit process, including a reverse osmosis system. Diligent monitoring improves the reliability and performance of RO systems and will significantly improve the life cycle cost of operating this unit process. Lira itations in technology, combined with insufficient plant resources have often left RO systems under monitored and under protected. One area which has not been adequately monitored is antiscalant concentration. The industry standard for antiscalant dosage monitoring, timed draw down, only gives an average concentration over time. Since the hydraulic holding times of

!011

Protonuel

Crvst

Oro

Fig. 1. Stages of crystallisation. Threshold Agents (Anti-precipitants)

tons

Protonuclei

•

Ordering

Crystals

Growth ~ : ~ ~

Delay the Ordering Process

Fig. 2. Effect of antiscalants (threshold agents) on crystallisation. RO systems are short compared to other unit processes, the variability inherent in timed draw down measurements can be unacceptably high. If, in addition to standard system performance monitoring, the antiscalant dosing could be monitored and controlled, it would provide confidence that the system was being treated and would alert the system operators to treatment interruptions or system flow dynamics that could lead to increased cleaning and maintenance costs. By employing fluorescent molecule monitoring [10], which has been successfully used in boiler and cooling water systems, a new technology has recently been developed for monitoring and control of the RO antiscalant, and it is now possible to realise these advantages. Patent applications have been filed for this novel technology. An overview of this exciting new technology and its applications in field systems is given below.

212

E.H. Kelle Zeiher et al. /Desalination 157 (2003) 209-216

operation ensured that the chemical concerns were fully satisfied and that the TRASAR molecule is compatible with the antiscalant and the membrane materials of construction. Extensive use of the TRASAR molecules (over the past l 5 years) in boiler and cooling water systems have demonstrated that it is environmentally acceptable and very cost effective.

3. Fluorescent molecule (TRASAR ®) technology Fluorescence is a property of certain organic molecules in which the molecule adsorbs light at one frequency and emits light at a different frequency. When such molecules are combined with antiscalant chemicals in a known proportion, they act as a "bar code" on the antiscalant, allowing the concentration ofantiscalant to be monitored via simple, on-line equipment. The molecules typically chosen for fluorescent tracing can be detected at very low concentrations, typically on the order of ppb or ppt in systems. In addition, they are chosen to fluoresce at wavelengths not visible to the human eye. TRASAR technology is based on the use of a fluorescent molecule that provides a clear and identifiable fluorescence that can be separated from the natural background. Fig. 3 shows a simplified diagram of the fluorometer used to detect the presence of the TRASAR molecule on line. When adding anything to a membrane antiscalant, several factors must be considered. The material must be: • Compatible with the membrane surface • Compatible with the antiscalant chemistry • Environmentally acceptable • Cost effective

4. Trial site 1 report This is an industrial system producing ultrapure water for high-pressure steam boilers. Additional treatment is mixed bed ion exchange (IX). The system (Fig. 4) is configured with either one of the RO units or both RO units operating to meet the demand for treated water. Each unit produces approximately 20 m~/h of permeate water. There had been reliability problems due to fouling and scaling, despite use of proprietary antiscalants. The initial troubleshooting investigation began with substituting the current antiscalant with TRASAR antiscalant. Timed drawn down measurements of antiscalant concentration indicated that the dose was constant and correct for the system. The continuous real-time results, however, showed a different story and provided the key diagnostic aid. Tracking of the feedwater antiscalant dose showed a much more consistent dose

The laboratory studies based on industry standard tests [11-13] such as soak tests and field

t

Detector

Z 2 ....... , d Excitation Filter Emission Filter

~tbe g

vXJ

Flow Cell

Light Sollrce

tt

Water

Sample In

Fig. 3. Fluorescence monitoring equipment.

E.H. Kelle Zeiher et al. /Desalination 157 (2003) 209-216

~

213

Secondary RO Permeate to boiler Cartridge H.P. filter Pump

Well water source

9,'.

~

i

~

feedwater tank

O

n

Feedwater sample point

Primary > RO centrate q/water sample point

Chemical injection point

Fig. 4. System overview trial site 1.

50

45

_.~ 4 0

o~ ¢o 35

30 E ~_ 25 =.o_ 20 o~15

810 5 0 o

o

o

o

o

~:~

o

o

Time, minutes

ofantiscalant when two RO units were operating then when only one unit was operating (Fig. 5). The feedwater for the RO system was pumped from a borehole. The borehole pump was found to be stopping and starting very frequently when only one RO was running. The hydraulic shock from pump starting and stopping caused erratic feeding of the antiscalant from the dosing line, giving intermittent mini slug-doses and the wide fluctuation. The problem was uncovered with the use of traced antiscalant and corrected by recommending that both RO units operate simultaneously to more effectively use the water produced by the borehole pump. Using the preferred operating protocol, chemical dose still varied by about 4-2 ppm of the target

o

Fig. 5. TRASAR'~' monitoring demonstrates best operating conditions.

antiscalant dose. Further improvements were made by using automatic control to dose the antiscalant. An automatic controller, installed on the TRASAR detector, was activated and used to automatically change the chemical feed pump stroking rate so that the appropriate amount of chemical was delivered at all times. Using the automatic control, the antiscalant dose was within 4-0.5 ppm of the target dose at all times (Fig. 6). The results show that the extremes of underdosing and overdosing were removed, which means that the RO membranes were more effectively protected against scaling and fouling. The long-term results of this trial were: • Operating costs reduced • RO system now reliable

214

E.H. Kelle Zeiher et al. /Desalination 157 (2003) 209 216

25

20 ¸ o

15;

[3_ B-

5

0

200

400 600 800 1000 Consecutive minutes of operation

1200

Fig. 6. Effect ofTRASAR"' control. • The maintenance supervisor reported that"The system has been running very smoothly which we have not experienced for quite a while." Overall, the expected cost savings for this system is about 6 140,000-170,000 per year depending on labour costs. This is t~ 780-950/1000 m 3. 5. Trial site 2 report

The second trial site also used RO product water for feed to high-pressure steam boiler (Fig. 7). Following the primary RO, the water was further treated by passing the permeate through a second RO to produce ultrapure water for the boilers.

The customer was seeking to improve control of the RO system and supported the r e m o t e monitoring possibilities offered by T R A S A R Technology. Traced antiscalant was used to treat the system. The product was monitored by means o f a fluorometer that measured the dosage of antiscalant in the concentrate line o f the RO. A controller automatically adjusted the dosing pump stroke frequency to assure a continuous dose of 12 ppm of antiscalant in the concentrate stream. All RO system parameters, including the antiscalant concentration, calculated percent recovery, percent rejection and normalised permeate flow, were logged and recorded using advanced monitoring capabilities [14]. During the early stages of the trial, a data download from the performance monitor indicated that the dosage level o f antisealant had fallen below the 12 ppm set point in the concentrate stream. Prolonged operation without antiscalant, or with an inadequate dose o f antiscalant, could lead to severe fouling conditions. The customer was alerted, and inspection of the chemical feed system revealed a crimped feed line. Although the pump was working at full speed, very little chemical was able to reach the RO feedwater due to the blocked line. The situation was corrected and chemical treatment reached the appropriate dosage within minutes (Fig. 8). The system has now been fully automated to provide alarms at a local control room for the direct attention o f the site engineering staff.

Antiscalant Cv~tYer~ ~ ~ - ~

Permeate to Post, Treatment and Boilers

©

Bisulfite ~, Sample Point Concentrate to Cooling Towers

Fig. 7. Systemoverviewtrial site 2.

215

E.H. Kelle Zeiher et al. /Desalination 157 (2003) 209-216 5O .....

45

Conc

~Set

Point

+ Control Limits 4o

_

35

_

3. Crimped Feed Line Repaired, - Control Reestablished

1. ChemicalFeed Falls Below Set Point, Indicating Problems

30

,<

/

25

ft.

2. ControllerTries to Bring Feed Back to Set Point Value

15

10

..........

t

ii:

i

5

0

• i

25

30

35

r

40

i:i ....

--

45

50

55

60

65

70

75

Time

Fig. 8. TRASAR" monitoring identifies crimped chemical feed line.

6. Conclusions

References

•

[1] A variety of courses are offered by David H. Paul, Inc, including (a) Reverse Osmosis Water Treatment, Short Course, July 24, 2002, Chicago; (b) Scaling Fouling & Chemical Cleaning, Short Course, July 24, 2002, Chicago; (c) Membrane Training Camp, June 15-19, 1998, San Diego. [2] Relevant standard from ASTM International include: (a) ASTM D4194-95, 01-Ju1-1995 Standard Test Methods for Operating Characteristics of Reverse Osmosis Devices; (b) ASTM D4516-00, 01-Aug-2000 Standard Practice for Standardizing Reverse Osmosis Performance Data; (c) ASTM D4692-87, 01-Aug1987 Standard Practice for Calculation and Adjustment of Sulfate Scaling Salts (CaSO4, SrSO4, and BaSO4) for Reverse Osmosis; (d) ASTM D447289, 01-Nov-1989 Standard Guide for Recordkeeping for Reverse Osmosis Systems; (e) ASTM D4195-88, 01-Nov-1988 Standard Guide for Water Analysis for Reverse Osmosis Application; (f) ASTM D4993-89, 01-Nov-1989 Standard Practice for Calculation and Adjustment of Silica (SiO2) Scaling for Reverse

• °

Effective monitoring and control is provided by the T R A S A R technology. Costs can be minimised with timely and accurate alarms to warn o f low antiscalant dosing The use o f R O T R A S A R h a s demonstrated that the automated monitoring and control o f antiscalant dosing works effectively and efficiently.

Antiscalant dosing is cost effective, environmentally acceptable, and safe. Automated antiscalant monitoring and control eliminates a major drawback to antiscalant use: the fear o f overdosing or underdosing. In addition, remote m o n i t o r i n g and control o f antiscalant dosages is now possible. Taken together, these advantages make antiscalant use a clear choice for consistent, safe, effective scale control in RO plants.

216

[3]

[4]

[5]

E.H. Kelle Zeiher et al. /Desalination 157 (2003) 209-216

Osmosis; (g) ASTM D4692-01, 01-Nov-2001 Standard Practice for Calculation and Adjustment of Sulfate Scaling Salts (CaSO4, SrSO4, and BaSO4) for Reverse Osmosis and Nanofiltration; (h) ASTM D3739-94, 01 -Jul-1994 Standard Practice for Calculation and Adjustment of the Langelier Saturation Index for Reverse Osmosis; (i) ASTM D4582-91, 01-Jul1991 Standard Practice for Calculation and Adjustment of the Stiffand Davis Stability Index for Reverse Osmosis; (j) ASTM D3923-94, 01-Ju1-1994 Standard Practices for Detecting Leaks in Reverse Osmosis Devices; (k) ASTM D4516-00, 0l-Aug-2000 Standard Practice for Standardizing Reverse Osmosis Performance Data; (1) ASTM D4189-95 01,-Nov1995 Standard Test Method for Silt Density Index (SDI) of Water; (m) ASTM D4194-95 01,-Ju1-1995 Standard Test Methods for Operating Characteristics of Reverse Osmosis Devices. D.C. 13randt, G.F. Leitner and W.E. Leitner, Reverse osmosis membranes: state &the art. Reverse Osmosis Membrane Technology, Water Chemistry, and Industrial Applications, Z. Amjad, (Ed.), Chapman & Hall, New York, 1993, p. 18. T. Matsuura, Future trends in reverse osmosis membrane research and technology. Reverse Osmosis Membrane Technology, Water Chemistry, and Industrial Applications, Z. Amjad, (Ed.), Chapman & Hall, New York 1993, p. 58. A. Buckley, C.J. Brouckaert and C.A. Kerr, RO applications in brackish water desalination and in the treatment of industrial effluents. Reverse Osmosis Membrane Technology, Water Chemistry, and industrial Applications, Z. Amjad, (Ed.), Chapman & Hall, New York 1993, p. 281.

[6]

[7]

[8]

[9]

[10]

[11]

[12] [13] [14] [15]

F. Max and E.G Darton, Statistical review of 150 membrane autopsies, 62nd Annual International Water Conference, Pittsburgh, Pennsylvania, USA, October 21-25, 2001. 13. Andrews and J. Mazur, The impact of eliminating softeners as pretreatment for reverse osmosis systems. Proe. 61st Annual International Water Conference, Pittsburgh, Pennsylvania, USA. October 22-26, 2000. W. Byrne, Reverse Osmosis. A Practical Guide for Industrial Users, 2nd ed., Tall Oaks Publishing Inc., 2002, p. 179. W. Byme, p. 358 cites several examples of analytical methods for determining scale inhibitor concentrations. Most of these methods involve multi-step laboratory manipulations, making them poorly suited for field applications. M.J. Chmelovski, Fluorescent-traced chemical programs not just a feed and control program, but a problem-solving tool, 58th Annual International Water Conference, November 3-5, 1997. C.A. Buckley, C.J. Brouckaert and C.A. Kerr, RO applications in brackish water desalination and in the treatment of industrial effluents. Reverse Osmosis Membrane Technology, Water Chemistry, and Industrial Applications, Z. Amjad, (Ed.), Chapman & Hall, New York, 1993, p. 285-290. FilmTec Corporation memorandum, Compatibility Approval of Antiscalant. Hydranautics Corporation, Technical Service Bulletin, TSBII5.03, May 1998. The RO-Eye TM Performance Monitor, available from Ondeo Nalco, was used to track RO perforlnance. PermaCare RO 12 software, Ondeo Nalco, May 2002.