Click for next page ( 6


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 5
The Blue Grass Chemical Agent Destruction Pilot Plant’s Water Recovery System BACKGROUND The design for the Blue Grass Chemical Agent Destruction Pilot Plant (BGCAPP) at the Blue Grass Army Depot near Richmond, Kentucky, is complete, and as of this writing the facility is under construction and about 50 percent complete. The planned operational life of BGCAPP is 3 to 5 years, beginning in 2017. 1 A detailed description of the design is beyond the scope of this report. A general description of the interim design can be found in the National Research Council (NRC) report Interim Design Assessment for the Blue Grass Chemical Agent Destruction Pilot Plant (NRC, 2005). Although the description of the unit operations in the 2005 report is accurate, the present design is somewhat different, due mostly to reductions in the number of various types of process equipment. In the demilitarization process planned for BGCAPP, the chemical agents—GB, VX, and mustard agent H—will be neutralized with hot caustic (for GB and VX) or hot water (for mustard agent H) after being removed from the munitions. Under the terms of the Chemical Weapons Convention, the products of this neutralization, called hydrolysates, must be further treated before they can be released for final disposal. Some of the energetics from munitions will also be neutralized on-site, including energetics from projectile bursters, rocket fuzes, and rocket propellant that has been contaminated with chemical agent. This energetics hydrolysate will be blended with agent hydrolysates prior to being processed by supercritical water oxidation (SCWO). In the SCWO process to be used at BGCAPP, water will be heated to 650ºC (1200ºF) and pressurized to 230 atmospheres (3,400 pounds per square inch gauge [psig]), well above its critical point of 374°C (705ºF) and 218 atmospheres (3,204 psig), whereupon it becomes supercritical. This occurs in a reactor vessel, which at BGCAPP will be a Hastelloy C-276 tube that is 7.625 inches in diameter and 120 inches in length. Oxygen and the blended hydrolysates are introduced into the reactor along with the supercritical water. The hydrolysate blend will have a residence time of about 15 seconds in the reactor. 2 The SCWO process can best be understood by thinking of supercritical water as a highly pressurized gas. Under these conditions, oxygen is highly reactive and will oxidize the elements in the hydrolysates into their most stable oxidized forms—carbon will be oxidized into carbon dioxide, hydrogen into water, and the sulfur and phosphorus into sulfates and phosphates. Organic materials, normally insoluble in water, volatilize at SCWO temperatures and are miscible in the supercritical water. Inorganic salts, which 1 Neil D. Frenzl, Surajit Amrit, P.E., and John W. Barton, Bechtel Parsons Blue Grass, “Blue Grass Chemical Agent-Destruction Pilot Plant Water Recovery System (WRS), RO: Addendum,” briefing to the committee, July 20, 2011. 2 Dan Jensen and Kevin Downey, General Atomics, “SCWO: Overview of Design and Review of Prior Test Results,” briefing to the NRC standing Committee on Chemical Demilitarization, September 14, 2011. -5-

OCR for page 5
normally dissolve and ionize in liquid water, do not volatilize and remain suspended as solids in the supercritical water. To prevent these salts from adhering to the wall of the reactor and eventually blocking the flow through the reactor, other salts will be added to form eutectics, which will keep the salts molten so that they flow through the reactor. At the end of the reactor, the pressure and temperature will be reduced and the water will return to a subcritical state. At this point, the salts will re-dissolve and the gases will separate from the liquid. The SCWO environment is highly reactive and corrosive. Sacrificial titanium liners will be inserted into the reactor to protect the reactor walls. These will be replaced periodically as they corrode, depending on the agent being processed. The corrosion products from the liner will exit the reactor as titanium dioxide (TiO2) particulates, which must be removed prior to treatment of the SCWO effluent in the water recovery system (WRS). The exact titanium content in the SCWO effluent will depend on the agent being processed. The SCWO process uses large quantities of water, which are continuously flowing through the reactor. To conserve water, the SCWO effluent will have the salts filtered out by the WRS by means of reverse osmosis (RO), and the recovered water will be recycled for use as quench water for the SCWO reactor. In the RO process, water is forced through a membrane by pressurizing it above the membrane’s osmotic pressure. The membrane is designed to reject salts and to pass water only. The water passed through the membrane is called permeate, and that left on the input side of the membrane is called RO reject. As the salt concentration increases in the RO reject, the osmotic pressure increases, and greater pressure is required to force the water through the RO membrane and separate it from the salts, driving up the amount of energy used by the WRS. The BGCAPP design anticipates that about 70 percent of the water can be recovered by this technique. An examination of this process is the main focus of this letter report. Until now, RO technology has not been employed in chemical demilitarization operations to recover water from a plant effluent. Because RO membranes and equipment are susceptible to failure from chemical attack, fouling, and other mechanisms, the Program Manager for Assembled Chemical Weapons Alternatives (PMACWA) requested that the National Research Council review the BGCAPP WRS design to identify possible issues related to the operability and reliability of the planned WRS. SCOPE OF THE REPORT This report focuses solely on the BGCAPP WRS. The scope of this study is limited to WRS operations, which begin when SCWO effluent and steam and cooling blowdown water enter the pretreatment system and end when the purified water exits the RO units. The input stream from the SCWO is assumed to be as described in earlier BGCAPP reports and in this report. The characteristics of the SCWO effluent flowing into the WRS are vitally important to the operation of the WRS, but any assessment of the SCWO process itself is beyond the scope of this study. Similarly, any additional treatment of the WRS effluent after exiting the RO units and prior to disposal is beyond the scope of this study. -6-

OCR for page 5
The statement of task for this study (presented in full in Attachment A) initially required the committee to:  Obtain information from the equipment vendor on water recovery system (WRS) installations that treat comparatively similar effluents to those at BGCAPP.  Contact a representative industrial installation to review its reverse osmosis (RO) system operational and maintenance history, and determine the degree to which operability has been acceptable.  Ascertain the likelihood that the quality of the recycled water will meet requirements for its re-use as quench water in the plant.  Review materials of construction to determine whether adequate performance can be expected over the anticipated operational life of BGCAPP, specifically addressing potential concerns for corrosion, fouling, and stress cracking.  Produce a letter report on determinations resulting from the above examinations. The committee is composed of members with decades of experience and broad knowledge of the use of RO systems to treat waters in a wide variety of settings around the world. Several committee members have long been involved in industrial applications of RO systems. Regarding the first two items in the statement of task: based on members’ extensive experience and knowledge, the committee judged that there are no representative industrial applications that could be used as any meaningful basis of comparison to the RO application planned for BGCAPP. The committee did query the RO vendor chosen by BGCAPP as to whether it had ever treated effluents similar to those that will be treated at BGCAPP. The vendor had never treated anything similar to the unique compositions of the BGCAPP SCWO effluents. In the course of performing its work, the committee obtained details about the SCWO process only insofar as that process affects the process stream that the WRS will treat. In reviewing and assessing the BGCAPP WRS, the committee recognized the following:  The footprint for the RO system in the building is limited by the present design;  The BGCAPP design is complete and construction is underway, making significant changes to the design challenging;  This RO system will only be operational for 3 to 5 years, until all the munitions are destroyed and the resulting hydrolysate has been treated; and  Any modifications to the design will necessitate amendments to the present Resource Conservation and Recovery Act permits, which govern plant operations, and will require negotiations with the Kentucky Department for Environmental Protection. The remainder of this report describes and reviews the design of the RO pretreatment system, the RO system, and the materials of construction (MOC) selected for the WRS. The committee’s findings and recommendations are incorporated in the text near the discussion that supports them. -7-

OCR for page 5
EFFLUENTS EXPECTED FROM THE SCWO SYSTEM The SCWO effluents are expected to be salt solutions with a range of 1 to 3 percent dissolved solids content, consisting primarily of sodium chloride, sodium sulfate, and sodium dihydrogen phosphate. The SCWO effluents are also expected to contain suspended solids primarily consisting of the following:  Titanium dioxide from the nerve agent campaigns,  Iron oxide transported along from the mustard agent-filled projectiles during the mustard campaign, and  Precipitates that form from waste constituents such as calcium, aluminum, and phosphate. Tables 1 through 3 show the results of analyses of SCWO effluent performed in 2004. The hydrolysates for these analyses were produced by the neutralization of actual agent at an Army laboratory, were blended with energetics hydrolysate, and were then treated with an SCWO unit at a General Atomics site in its test SCWO unit. Table 1 Liquid Effluent Analyses for GB SCWO Performance Tests 9/15/2004 9/16/2004 9/16/2004 9/16/2004 06:40 00:30 12:00 14:30 Tap water Analyte Units Result Result Result Result Result TOC(1) mg/L 1.5 1.7 1.8 1.4 NA TOC(2) mg/L 1.4 1.7 1.6 1.4 NA TOC(3) mg/L 1.5 1.7 1.7 1.3 NA TOC(4) mg/L 1.4 1.5 1.6 1.4 NA Chloride mg/L 2,820 2,680 2,530 2,110 NA Fluoride mg/L 144 140 177 116 NA Aluminum μg/L 6,040 7,130 16,500 6,110 30 U Calcium μg/L 22,000 27,300 26,300 17,100 52,400 Chromium μg/L 161 62.7 296 292 3.0 U Iron μg/L 615 220 1,110 1,120 10 U Magnesium μg/L 8,330 10,600 10,200 6,580 21,200 Molybdenum μg/L 25.0 U 25.0 U 25.0 U 25.0 U 5.0 U 52.4 B 114 206 4.6 B Nickel μg/L 103 Phosphorus μg/L 349,000 326,000 339,000 321,000 20 U Potassium μg/L 4,970 B 4,110 B 4,070 B 3,500 B 3,870 Sodium μg/L 5,480,000 4,970,000 5,110,000 4,860,000 76,400 Sulfur μg/L 2,330,000 2,080,000 2,150,000 2,050,000 54,200 Titanium μg/L 5,560 4,840 5,080 5,080 5.0 U NOTE: Analytical codes are U, Undetected; B, analyte found in method blank, result not valid; NA, not analyzed; TOC, total organic carbon. SOURCE: Adapted from BPBG, 2005. Table 2 Liquid Effluent Analyses for VX SCWO Performance Tests 10/14/2004 10/15/2004 10/16/2004 10/17/2004 08:45 09:30 07:30 07:00 Analyte Units Result Result Result Result TOC(1) mg/L 0.16 J 0.19 J 0.23 J 0.13 J TOC(2) mg/L ND 0.17 J 0.24 J 0.12 J Continued -8-

OCR for page 5
10/14/2004 10/15/2004 10/16/2004 10/17/2004 08:45 09:30 07:30 07:00 Analyte Units Result Result Result Result TOC(3) mg/L 0.08 J 0.18 J 0.22 J 0.12 J TOC(4) mg/L 0.10 J 0.18 J 0.22 J 0.13 J Chloride mg/L 4,020 3,570 3,940 2,280 Sulfate mg/L 12,200 10,400 11,800 6,880 Phosphate mg/L 743 X 689 X 712 X 680 X Aluminum μg/L 17,200 15,600 15,700 15,700 Calcium μg/L 174 186 662 32,600 Chromium μg/L 10.0 11.3 6.8 3.8 J Iron μg/L 100 244 74.9 64.4 Magnesium μg/L 60.4 96.3 18,300 21,800 Molybdenum μg/L 5.0 U 5.0 U 5.0 U 59.7 Nickel μg/L 71.3 81.7 64.0 83.1 803,000 708,000 761,000 736,000 Phosphorus μg/L Potassium μg/L 4,990 5,260 18,500 7,070 Sodium μg/L 9,540,000 8,620,000 9,210,000 8,560,000 Sulfur μg/L 4,380,000 3,870,000 4,090,000 3,870,000 Titanium μg/L 22,000 22,000 23,100 23,200 NOTE: Analytical codes are J, analyte positively identified but result is approximate; U, undetected; B, analyte found in method blank, result not valid; X, estimated maximum possible concentration; TOC, total organic carbon. SOURCE: Adapted from BPBG, 2005. Table 3 Liquid Effluent Analyses for Mustard SCWO Performance Tests 9/25/2004 9/26/2004 9/27/2004 12:30 10:00 06:15 Analyte Units Result Result Result TOC(1) mg/L 0.17 J 0.20 J 1.4 TOC(2) mg/L 0.19 J 0.19 J 1.3 TOC(3) mg/L 0.15 J 0.20 J 1.4 TOC(4) mg/L 0.19 J 0.18 J 1.3 Chloride mg/L 4,040 4,390 4,020 Fluoride mg/L <2 U <2 U <2 U Sulfate mg/L 10,900 11,800 10,900 Aluminum μg/L 8,360 7,450 9,730 Calcium μg/L 170 B 231 B 38,100 Chromium μg/L 176 147 209 Iron μg/L 633,000 542,000 790,000 Magnesium μg/L 400 375 15,500 Molybdenum μg/L 25.0 U 25.0 U 36.0 B Nickel μg/L 363 281 234 1,790 1,750 1,940 Phosphorus μg/L Potassium μg/L 7,520 B 6,220 B 10,100 Sodium μg/L 8,040,000 8,140,000 8,150,000 Sulfur μg/L 3,930,000 3,930,000 3,990,000 Titanium μg/L 192 133 136 NOTE: Analytical codes are J, analyte positively identified but result is approximate; U, undetected; B, analyte found in method blank, result not valid; TOC, total organic carbon. SOURCE: Adapted from BPBG, 2005. -9-

OCR for page 5
The elements that will likely impact the RO system are indicated in Table 4. TABLE 4 Concentrations of Elements Present in the Three Hydrolysates (in mg/L) and the Possible Forms of Solids That May Be Present Mustard Agent H Element GB VX Possible Solids Al 6-16 25-27 7.4-9.7 AlPO4 Ca 17-27 0.1-33 0.2-38 (Ca)2(PO4)3 Fe 0.2-1.0 540-790 Fe2O3 P 320-350 708-803 1.7-1.9 Ma PO 4 Ti 4.8-5.5 22-23 <0.2 TiO2 S CaSO4 2,050-2,330 3,870-4,380 3,930-3,990 a M refers to “metal” and can be Al, Ca, Mg, etc. The elemental concentrations in the hydrolysates will exceed the solubility product for minerals such as AlPO4, (Ca)2(PO4)3, and Fe2O3, which are the forms likely to be found when the hydrolysate is oxidized. If insufficient phosphate is available in the hydrolysates, then the precipitates formed are likely to be hydroxides. The mustard hydrolysate is supersaturated with CaSO4, so precipitation and scaling of the RO membrane with this solid is of concern when this solution is processed by RO as well. Precipitation of calcium by phosphate in GB and VX hydrolysates may reduce the concentration of calcium that may prevent CaSO4 precipitation from being a problem when these hydrolysates are processed by RO. If these substances are present as particulates in the SCWO effluent, they will be removed if the coagulation and filtration processes prior to effluents arriving at the RO unit are functioning properly. How much of the iron, calcium, and aluminum solids will settle out in the hydrolysate storage tank and the SCWO effluent tanks prior to effluents arriving at the WRS pretreatment system is not predictable. This issue is discussed in detail below, in the “Pretreatment System” section of this report. The SCWO effluents will have overall salt concentrations similar to those of brackish water, but the specific compositions of the effluents will be unique, coming as they do from the processing of chemical agent and energetics hydrolysates. Therefore, although experience from RO plants that treat brackish water and seawater can provide guidance about the challenges that might be expected in the BGCAPP WRS, such experience is not necessarily directly applicable to the planned BGCAPP WRS. The SCWO effluent to be treated by the WRS will be at a higher temperature (38°C/100°F) than the water treated at seawater desalination plants. This elevated temperature can be expected to increase water flux and influence the rejection of salts by the RO membranes. This, again, demonstrates that the effluents to be treated by the BGCAPP WRS are unlike any other process stream that has been treated by RO in other commercial and industrial settings. In the course of its data gathering, the committee queried the vendor personnel about whether they had ever treated an RO influent similar to the expected SCWO effluents; they replied that they had not. Further, in the committee’s knowledge and experience, no water recovery system, industrial or otherwise, has ever treated effluent -10-

OCR for page 5
streams like those that will be treated at BGCAPP. The committee’s judgment, based on the individual members’ expertise and data gathering, is that the SCWO effluents expected at BGCAPP will be unlike any other influent previously treated by an RO system to date. Finding. The compositions of the expected supercritical water oxidation effluents to be treated at BGCAPP are unique, and similar effluents have never before been treated by reverse osmosis. Finding. Whatever testing could be conducted of effluents similar to those expected from the supercritical water oxidation system would be beneficial to ensuring that the water recovery system operates as expected, or to uncovering problems prior to systemization. Identifying problems as early as possible reduces the risk of significant disruptions to the overall project schedule. Recommendation. It should be investigated whether precipitates might possibly form as the supercritical water oxidation effluents are being processed by reverse osmosis (RO), and whether steps, such as suitable inhibitor addition, can be taken to prevent the development of RO membrane scaling problems. DESCRIPTION OF THE WATER RECOVERY SYSTEM The WRS will desalinate SCWO effluents, cooling tower blowdown, and steam boiler blowdown for reuse as quench water in the SCWO process. The system was designed:  To operate with an efficiency of 70 percent water recovery with a maximum of 500 mg/L total dissolved solids (TDS) in the permeate, and  To ensure one full day’s storage of RO permeate to permit SCWO operation in case the WRS is not operating. To accomplish these operations, the WRS includes:  Three SCWO effluent storage tanks where the effluent will be analyzed to ensure that the total organic carbon concentration is less than 2 parts per million (ppm);  A conventional pretreatment system consisting of coagulant and antiscalant addition (dual pumps on each unit), media filtration (six units), and canister filters (three) prior to the RO units;  Three spiral wound reverse osmosis units (two operational, one spare); and  Storage tanks used to hold RO permeate to clean the RO membranes periodically. Figure 1 shows the flow of material from hydrolysis, through the SCWO process, up to the pretreatment step in the WRS. It also indicates where the cooling tower and steam blowdown is blended with the SCWO effluent. The dashed arrows indicate changes recommended by the committee (discussed in more detail below): namely, two -11-

OCR for page 5
RO bypasses should be added, one to redirect blowdown water directly to the blowdown- water holding tanks or the RO reject tank if the water softener fails, and the other to divert softened water directly to RO permeate if water quality allows. Figure 2 shows the flow of material through the WRS. The arrow showing the addition of coagulant was added by the committee for clarity. Chemical agent Agent To RO Water SCWO Hydrolysis hydrolysate pre - SCWO effluent process storage treatment tanks (3) tank Emergency bypass of RO system Cooling tower and Water steam softener blowdown To RO permeate Spent regenerant to RO reject tank FIGURE 1 The flow of material from hydrolysis, through supercritical water oxidation (SCWO), up until the pretreatment step in the water recovery system (WRS). The dashed lines show changes recommended by the committee, as discussed in this report. NOTE: RO, reverse osmosis. Antiscalant Coagulant RO units (3) RO permeate From SCWO effluent tanks To RO permeate tanks SCWO Multimedia Canister High- effluent filters filters pressure RO reject tank (6) (3) pump pumps Recycle To RO reject tanks pump FIGURE 2 Process flow diagram for the BGCAPP water recovery system (WRS) including the pretreatment and reverse osmosis (RO) system (modified by the committee to add the coagulant insertion point). SOURCE: Neil D. Frenzl, Engineering Manager, Bechtel Parsons Blue Grass, and Surajit Amrit, P.E., Mechanical Engineering Lead, Bechtel Parsons Blue Grass, “Blue Grass Chemical Agent-Destruction Pilot Plant Water Recovery System (WRS), RO System Overview and Material of Construction and Related Issues,” presentation to the committee, July 19, 2011. The overall operation of the WRS will be monitored by the facility control system. Items monitored will include the following: -12-

OCR for page 5
 Temperature indication for the feed to the RO units,  Flow indication for the feed to and discharge from the RO units,  TDS concentration (through conductivity monitoring) of RO permeate,  Differential pressure across the RO unit (feed versus reject),  Differential pressure across the multimedia filters and canister filters, and  Proportional flow ratio and total flow rate indication for the caustic injection system (BPBG, 2009). PRETREATMENT SYSTEM Water Softening Two streams will be blended to form the WRS influent—that is, (1) SCWO effluent and (2) cooling tower and steam blowdown water. Removing the calcium in the RO influent stream is necessary before it arrives at the RO system because calcium could have a significant negative impact on the operation of the RO units. Water softening will be used to remove the calcium (a water hardness component) from the cooling tower and steam blowdown water. According to BPBG (2007), ion exchange columns will be used to soften this stream prior to its being blended with the SCWO effluent (see Figure 1). Ion exchange softeners can produce water with a very low level of calcium. However, the level of residual calcium will be determined by the operating procedure of the ion exchange process—particularly (1) the amount of calcium leakage allowed in the softener effluent before regeneration and (2) the concentration and quantity of regenerant applied. The spent softener regenerant 3 will be combined with RO reject water for final disposal. The total calcium concentration in the SCWO effluent and softener effluent blend will have an important effect on water recovery from the RO process. It is also possible, however, that the effluent from the softener will have sufficiently low calcium content, making it suitable for blending directly with the RO effluent, bypassing the RO unit. This would reduce greatly the amount of water that must be processed by the RO system. The hydraulic design of the RO process targets 70 percent recovery, but a higher recovery might possibly be achieved if the calcium in the ion exchange-treated cooling tower and steam blowdown water effluent is sufficiently low and there is minimal calcium in the SCWO effluent after pretreatment. The design recovery of 70 percent may lead to scaling of the membranes if substantial calcium remains in the softener effluent and/or if there is substantial calcium in the SCWO effluent after pretreatment. The SCWO effluent from the processing of mustard agent hydrolysate had a high concentration of calcium, although it should be noted that only one valid analysis is presented in Table 3. This level of calcium combined with the high sulfate concentration shown in Table 3 indicates that CaSO4 precipitation may take place in the RO system 3 The regenerant is the waste solution resulting when a high-concentration solution of NaCl is used to renew the hardness removal capacity of the ion exchange resin. Typically the regenerant contains high concentrations of NaCl, calcium salts, and magnesium salts. -13-

OCR for page 5
when mustard agent hydrolysate is being processed. The SCWO effluent from the processing of GB hydrolysate contained calcium, and one sample of SCWO effluent from the processing of VX hydrolysate also showed a high level of calcium (see Tables 1 through 3). It is possible that the concentrations of calcium may be substantially reduced by precipitation with phosphate before filtration, thus reducing the possibility of calcium scaling when the GB and VX SCWO effluents are being processed. To prevent scaling from calcium in the SCWO effluent, a polyphosphate chemical has been selected as an antiscalant for RO pretreatment at BGCAPP. Adequate pH control is also necessary for the effective control of scaling. For example, reducing pH to 5 or less would protect against RO membrane fouling from the precipitation of CaSO4/CaCO3. Finding. Water softening of the cooling tower and steam boiler blowdown is essential. Otherwise, there could be catastrophic scaling and failure of the reverse osmosis (RO) system. If the ion exchange system is not functional, the calcium-laden blowdown water must not be blended into the RO feedwater without further modifications to the pretreatment and RO units. Finding. The use of pH control is also a necessary component to protect against scaling. Using pH control in addition to the use of chemical antiscalants would provide the best scaling control. Recommendation. It should be confirmed during systemization that calcium removal is complete and that softener monitoring and regeneration procedures have been established to avoid calcium mineral scaling in the reverse osmosis units. Recommendation. The BGCAPP design should include a water bypass from the water softener around the reverse osmosis system in the event that the softener fails. The water should be returned to the blowdown-water holding tanks, or sent to the RO reject if the holding tank capacity is insufficient. Finding. The reverse osmosis (RO) system hydraulic design target of 70 percent recovery assumes that there is no calcium in the RO influent. The actual level of calcium in the RO influent will depend on how the ion exchange softening process is operated and on the level of calcium in the SCWO effluent after pretreatment. The SCWO effluent does not pass through the softener, and its level of calcium could be substantial. Recommendation. Additional reverse osmosis hydraulic design simulations (e.g., ROSA) should be considered using different levels of calcium in the feed, taking into account the calcium levels anticipated in the SCWO effluents, in order to establish the target level of recovery for each campaign. Finding. If the quality of the softened water meets the requirements for use as quench water, then the softened water could bypass the reverse osmosis (RO) system and be injected directly into the RO permeate. Recommendation. Regularly bypassing part of the softened water around the reverse osmosis (RO) system directly to the RO permeate should be considered, rather than -14-

OCR for page 5
adding it to the RO feedwater. If this is done, the blended product water could still meet the total-dissolved-solids requirement of 500 mg/L, and the load on the RO system will be reduced. This may become important if other problems accelerate membrane fouling. Coagulation of the Suspended Solids in the SCWO Effluent Overview of Coagulation The use of RO membranes requires some form of pretreatment in order to remove suspended solids and to reduce turbidity and the silt density index (SDI). Solids found in water are broadly grouped into two categories: suspended solids and colloids. Suspended solids are operationally defined as those that can be removed from water by filtration through a membrane with a 0.45 micron pore size, and colloids are solids that pass through such a membrane. Many suspended solids and colloids can be stable suspensions that do not settle rapidly because of their small size, surface charge, and other factors. These suspensions must be destabilized by treatment with coagulants that aggregate the particles so that they can be removed by sedimentation and/or filtration. Chemical coagulants include aluminum and iron salts and both inorganic and organic polymers. The addition of coagulant may have to be followed with a gentle mixing step called flocculation to aggregate the solids and ensure that subsequent sedimentation and filtration processes work effectively. The interactions between coagulant chemicals, the water constituents, and the suspended solids are often difficult to predict. Thus, determining the best coagulant and dose usually requires laboratory testing to ensure that the system will function properly. It might also be necessary to add an organic chemical, called a filter aid, to improve filtration performance. Two approaches are used for particle removal by coagulation. (1) The traditional method is flash-mixing–flocculation–sedimentation–filtration. In this method, the coagulant is added in a short-retention-time reactor (30 seconds to several minutes) with mechanical mixing. This flash-mixing step is followed by flocculation and sedimentation prior to filtration. Typical hydraulic retention times for flocculation are between 2 and 30 minutes, and for sedimentation they are 1 to 4 hours. (2) The second approach, direct filtration, includes in-line—that is, static—mixing followed by filtration. Direct filtration is less expensive to install but often is less efficient than flash-mixing–flocculation– sedimentation–filtration. However, direct filtration is often adequate for many applications, especially those that have low total suspended-solids concentrations and particles that aggregate well. The current BGCAPP design uses direct filtration with an in-line static mixer. The committee has identified two challenges with this approach. The first challenge at BGCAPP is the large mass of suspended solids to be removed by the media filters. It appears from the water quality data in Tables 1 through 3 that the solids loading to the filters will be within proper design values for the GB and VX campaigns if aluminum and calcium do not precipitate. However, aluminum and calcium precipitation is likely, and the media filters could be rapidly overloaded with solids. During the mustard agent campaign, the media filters could also be overloaded by the iron particles in the mustard hydrolysate. How much iron, aluminum, or calcium solids will settle out in the various holding tanks and how much will be carried over into -15-

OCR for page 5
membranes to meet the water quality objectives, provided the RO pretreatment (coagulation and media filtration) removes suspended solids to avoid fouling and operational problems with the RO units. Finding. Given the required total dissolved solids limit of 500 mg/L and the use of seawater reverse osmosis membranes, the water recovery system effluent should be suitable for reuse as quench water in the supercritical water oxidation system. MATERIALS OF CONSTRUCTION MOC Planned for Use in the BGCAPP WRS The MOC planned for the BGCAPP WRS are shown in Tables 5 through 9. TABLE 5 Materials Planned for Use in the Coagulant Skid RO Equipment Tag Number RO Equipment Name Original MOC Revised MOC MT-RO-0103 Coagulant tote bin CPVC/PVC No change MF-RO-0103 Coagulant tank agitator CPVC/PVC No change MP-RO-0115 A/B Coagulant injection pump and spare CPVC/PVC No change Piping CPVC/PVC No change NOTE: CPVC, chlorinated polyvinyl chloride; MOC, materials of construction; PVC, polyvinyl chloride; RO, reverse osmosis. SOURCE: Neil D. Frenzl, Engineering Manager, Bechtel Parsons Blue Grass, and Surajit Amrit, P.E., Mechanical Engineering Lead, Bechtel Parsons Blue Grass, “Blue Grass Chemical Agent Destruction Pilot Plant Water Recovery System (WRS), RO System Overview and Material of Construction and Related Issues,” presentation to the committee, July 19, 2011. TABLE 6 Materials Planned for Use in the Antiscalant Skid RO Equipment Tag Number RO Equipment Name Original MOC Revised MOC MT-RO-0105 Antiscalant tote bin CPVC/PVC No change MT-RO-0105 Antiscalant tank agitator CPVC/PVC No change MP-RO-01108A/B Antiscalant injection pump and spare CPVC/PVC No change Piping CPVC/PVC No change NOTE: CPVC, chlorinated polyvinyl chloride; MOC, materials of construction; PVC, polyvinyl chloride; RO, reverse osmosis. SOURCE: Neil D. Frenzl, Engineering Manager, Bechtel Parsons Blue Grass, and Surajit Amrit, P.E., Mechanical Engineering Lead, Bechtel Parsons Blue Grass, “Blue Grass Chemical Agent Destruction Pilot Plant Water Recovery System (WRS), RO System Overview and Material of Construction and Related Issues,” presentation to the committee, July 19, 2011. TABLE 7 Materials Planned for Use in the Multimedia Filter Skid RO Equipment Tag Number RO Equipment Name Original MOC Revised MOC MK-RO- Multimedia filters Carbon steel 316L stainless steel 0101A/B/C/D/E/F Piping CPVC/PVC No change NOTE: CPVC, chlorinated polyvinyl chloride; MOC, materials of construction; PVC, polyvinyl chloride; RO, reverse osmosis. SOURCE: Neil D. Frenzl, Engineering Manager, Bechtel Parsons Blue Grass, and Surajit Amrit, P.E., Mechanical Engineering Lead, Bechtel Parsons Blue Grass, “Blue Grass Chemical Agent Destruction Pilot -33-

OCR for page 5
Plant Water Recovery System (WRS), RO System Overview and Material of Construction and Related Issues,” presentation to the committee, July 19, 2011. TABLE 8 Materials Planned for Use in the Canister Filter Skid RO Equipment Tag Number RO Equipment Name Original MOC Revised MOC MK-RO- Canister filters Carbon steel 316L stainless 0102/0202/0302 steel Piping CPVC/PVC No change NOTE: CPVC, chlorinated polyvinyl chloride; MOC, materials of construction; PVC, polyvinyl chloride; RO, reverse osmosis. SOURCE: Neil D. Frenzl, Engineering Manager, Bechtel Parsons Blue Grass, and Surajit Amrit, P.E., Mechanical Engineering Lead, Bechtel Parsons Blue Grass, “Blue Grass Chemical Agent Destruction Pilot Plant Water Recovery System (WRS), RO System Overview and Material of Construction and Related Issues,” presentation to the committee, July 19, 2011. TABLE 9 Materials Planned for Use in the Reverse Osmosis Skid RO Equipment Tag Number RO Equipment Name Original MOC Revised MOC MP-RO-0105/0205/0305 RO high-pressure pumps 316 stainless steel AISI 904 duplex stainless ML-RO-0101/0201/0301 RO units Polyamide thin-film No change composite MP-RO-0114/0214/0314 RO recycle pumps 316 stainless steel AISI 904 duplex stainless 316 stainless steel 316L stainless steel NOTE: MOC, materials of construction; RO, reverse osmosis; alloy 904 is normally identified as 904L. It is not a duplex alloy. It is an austenitic stainless steel with a composition of 19.0-23.0 percent chromium, 1.0-2.0 percent copper, 4.0-5.0 percent molybdenum, 23.0-28.0 percent nickel, and 0.02 percent carbon; AISI, American Iron and Steel Institute. SOURCE: Neil D. Frenzl, Engineering Manager, Bechtel Parsons Blue Grass, and Surajit Amrit, P.E., Mechanical Engineering Lead, Bechtel Parsons Blue Grass, “Blue Grass Chemical Agent Destruction Pilot Plant Water Recovery System (WRS), RO System Overview and Material of Construction and Related Issues,” presentation to the committee, July 19, 2011. Challenges Posed by Planned Operational Conditions at BGCAPP The BGCAPP WRS will operate at about 38°C (100°F). The SCWO effluent (called the process stream) that it will treat is projected to contain high levels of NaCl, Na2HPO4, Na2SO4 and NaHCO3 (BPBG, 2010). The corrosiveness of this water is not known, but the committee is concerned that the MOC selected as of this writing may experience excessive localized corrosion and stress corrosion cracking (SCC). The total salt concentration in this water will range from about 1.2 percent in the RO inlet stream to about 4 percent in the RO reject stream. The NaCl portion will range from about 0.3 percent in the inlet stream to over 1 percent in the reject stream. Thus, both streams are considered to be brackish. The total expected salt molar content of the RO reject stream will be similar to that of seawater, although the elemental distribution of salts in the process stream will be very different. This unique distribution makes it difficult to predict with any degree of confidence the corrosion behavior of the alloys planned for use. Normal industry practice in a situation such as this is to conduct corrosion tests with the candidate materials to evaluate their resistance to corrosion in the expected operating environment. The most -34-

OCR for page 5
reliable test is a long-term immersion test using material coupons, some with artificial crevices and others that are stressed. These tests typically take several months to yield valid data. Short-term laboratory corrosion tests are available to provide useful predictions of pitting tendencies and SCC resistance. These are discussed in more detail below. The committee’s comments are based on the expected short operational lifetime of the plant. If the planned BGCAPP operational life were to be extended, a more conservative approach might be warranted. Erosion and erosion-corrosion are not discussed because the flow rates through the WRS are expected to be low enough that these should not pose a challenge to the system. Corrosion and Stress Corrosion Cracking Corrosion Pitting attack is the most probable form of corrosion in the expected WRS operational environment. In a pitting attack, the total amount of metal loss may be relatively small, but the attack is localized, and deep penetration can occur rapidly once the attack has started. One possible place where pitting can occur would be under an accumulation of solids in the system. A closely related form of corrosion is crevice corrosion, in which the attack is located at a crevice or under deposits on the metal surface. It, too, can produce a deep penetrating attack with relatively small amounts of total metal loss. In both pitting and crevice corrosion, there is typically an incubation period during which the attack is not evident, followed by a rapid penetrating attack. Both forms of attack are most prevalent when the fluid flow rate is slow or where the fluid flow is obstructed, such as in flanged joints, under gaskets, and in dead-end passages such as instrument lines. Continuous fluid flow reduces, but does not eliminate, the risk of these forms of attack. Process shutdowns without immediately draining and flushing the system with deionized water create an environment in which the danger of these attacks is greatly increased. Intergranular corrosion of the weld heat-affected zone can result if the alloy has a high carbon content. In this heat-affected zone, the carbon combines with the chromium and molybdenum in an alloy, such as type 316 stainless steel, to produce on either side of a weld a band in which the alloy content is seriously reduced, especially at the grain boundaries. The formation of grain-boundary chromium carbide, in particular, leads to a depletion of chromium in the alloy solid solution, thereby reducing the corrosion resistance in this band. This band is thus susceptible to selective attack. Not only is the corrosion resistance reduced by the carbide production, but the band also becomes anodic relative to the rest of the material, resulting in a very small anode (where corrosion occurs) driven by a very large cathode (the rest of the material). To overcome this problem, alloys are designed with either low carbon content or constituents such as niobium or titanium that bind the free carbon by forming carbides. Another form of corrosion that may occur during process shutdowns is MIC. In this case, colonies of microbes form and either directly attack the metal or produce conditions in which pitting and crevice corrosion are encouraged. Even small pockets of liquid, such as at low spots in the piping system (e.g., elbows) and dead legs, can result in a serious MIC attack. To prevent moisture from condensing and collecting, a system is -35-

OCR for page 5
often continually flushed with dry air or nitrogen when not operating. In the BGCAPP WRS, the cooling water blowdown is a possible source of microbes. If the system is maintained in wet standby condition, appropriate measures must be taken to avoid the growth of microbial colonies. If the system can be stored without the membranes in place, especially during the long interval between installation and systemization, the complete draining and drying of the system during system shutdowns can prevent MIC attack during shutdown. At BGCAPP, stainless steel types 316 and 316L have been considered for use in the WRS. The primary difference between type 316 stainless steel and type 316L is the difference in maximum carbon content permitted (0.08 percent in 316 and 0.03 percent in 316L). The primary reason to specify type 316L rather than type 316 stainless steel is to avoid intergranular corrosion in weld heat-affected zones. The overall corrosion resistance of the two alloys is essentially the same except in weld heat-affected zones. As previously mentioned, it is not possible to determine the corrosion behavior of type 316 stainless steel in the BGCAPP WRS without test data from representative environments. The nearest archetype systems with which the committee members have experience are seawaters and brackish waters. In these cases, both 316 and 316L have marginal corrosion resistance, and their use has been largely replaced by duplex stainless steels such as 2205. The compositions of the alloys being discussed are shown in Table 10. Alloy 2205 has considerably greater crevice corrosion resistance than that of type 316 (or 316L) stainless steel, as shown in Table 11. TABLE 10 Percentage Composition of Several Alloys Alloy C Cr Cu Mo N Ni 316L 0.03 max. 16.0-18.0 — 2.0-3.0 — 10.0-14.0 2205 0.03 max. 21.0-23.0 — 2.5-3.5 0.08-0.20 4.5-6.5 904L 0.02 max. 19.0-23.0 1.0-2.0 4.0-5.0 — 23.0-28.0 NOTE: C, carbon; Cr, chromium; Cu, copper; Mo, molybdenum; N, nitrogen; Ni, nickel. TABLE 11 Critical Crevice Temperature Alloy Temperature (°C/°F) 2205 40/104 316L <20/<67 NOTE: This is at 3,000 ppm Chlorine, -300 mV versus saturated calomel electrode (SCE). SOURCE: Arnvig et al., 1996. Arnvig et al. (1996) found that the critical pitting temperature 8 for 2205 in 1 molar NaCl is 46°C (115°F). They did not report comparable data for type 316 stainless steel. Other researchers, using ASTM [American Society for Testing and Materials] International’s ASTM G48, Practice C, found the critical pitting temperature of type 316 stainless steel to be 20°C (67°F) and that of type 2205 to be 36°C (97°F). A more highly alloyed duplex alloy, 2507 (a super duplex alloy) had a critical pitting temperature of over 70°C (158°F) (Crum and Shoemaker, 2009). There is a difference in the corrosion behavior of cast and wrought stainless steels in seawater and brackish waters. Malik et al. (2011) report that the corrosion rates of cast 8 The critical pitting temperature is the temperature above which localized corrosion begins to occur. -36-

OCR for page 5
duplex steels are at least one order of magnitude higher than those of forged alloys. Similar behavior would be expected with type 316/316L stainless steels vis-à-vis casting versus forging. Malik et al. (2011) also report that the crevice corrosion rate is strongly dependent on the pitting resistance equivalent (PRE) number of the alloys. The PRE number of an alloy is equal to x % chromium + 3.3x % molybdenum + 16x % nitrogen, where x is a given percentage of chromium. In general, the crevice corrosion occurrence rate in cast duplex stainless steels decreases linearly with increased PRE number. 9 Malik et al. (2011) found that the super duplex alloys S39274 and S32750 have outstanding corrosion resistance against general and localized corrosion in Arabian Gulf and Red Sea open seawater. Francis et al. (2011) report that high-alloy stainless steels have a wide passive range and, depending on the cathodic reaction, they can adopt a wide range of electrochemical potentials. These potentials are reported to range from -450 millivolts (mV) saturated calomel electrode (SCE) for de-aerated seawater to +600 mV SCE for chlorinated seawater. The potential for natural seawater is about +325 mV SCE when a biofilm has formed, or about +100 mV SCE without a biofilm. If the oxygen content of the seawater is reduced, the potential further decreases and can be around -199 mV SCE (Francis et al., 2011). It is reasonable to speculate that the corrosion potential of the anticipated process fluids from SCWO will also be somewhat dependent on the cathodic reaction. Francis et al. (2011) also reported that the high-pressure section of the RO unit that is the subject of their paper operates at a redox potential of +250 to +350 mV Ag/AgCl saturation and that most plants operate near the maximum potentials. This redox potential range corresponds to an open-circuit potential of +100 mV to +200 mV SCE for stainless steel (Francis et al., 2011). Francis et al. (2011) report tests conducted by Byrne et al. (2009) to determine the critical crevice corrosion temperature for a range of stainless steels as a function of potential. These results, presented here in Figure 3, show that 316L stainless steel would be totally unsuitable for seawater RO plants because of the low temperature at which crevice corrosion initiates. Alloys such as 2205 and 904L showed good crevice corrosion resistance at +100 mV SCE, but much-reduced resistance (crevice corrosion temperatures of 20°C to 25°C/68°F to 77°F) at +200 mV SCE. As most seawater RO units operate close to the maximum redox potential, stainless steel will have a potential close to +200 mV SCE. This explains the service failures of both 904L and 2205. Super duplex alloy Z100 has better crevice corrosion resistance and has given good service in both the high-pressure and reject brine sections of seawater RO plants. The discussion above indicates how important it is to know what the corrosion potential of the candidate alloys will be in the BGCAPP RO unit. Since there are no actual process fluids available to use in conducting tests, the next best approach would be to prepare a synthetic process fluid that chemically matches the composition of the expected process fluid and to use this fluid to conduct electrochemical tests in order to characterize the corrosion behavior of these alloys. For example, the anions present in the process fluids will have an effect on the corrosion behavior of the chosen MOC. In this 9 Although there is a general trend of corrosion resistance increasing with increased PRE, and although the PRE can be used to estimate the pitting corrosion resistance of one material relative to another material, the PRE alone cannot predict whether a given material will perform satisfactorily without the representative environment’s being known. -37-

OCR for page 5
vein, Pohjanne et al. (2007) found that the sulfate ion increases the pitting resistance of type 304 stainless steel in a chloride-sulfate solution. Since the process fluid is 700 600 500 Z100 904L 400 Potential (mV SCE) 22%Cr Duplex 300 200 100 0 316L –100 –200 0 10 20 30 40 50 60 70 80 90 100 Relative critical crevice temperature (°C) FIGURE 3 Relative crevice critical temperatures of some stainless steels in seawater as a function of potential. SOURCE: Byrne et al., 2009. Previously published proceedings of the IDA World Congress on Desalination and Water Reuse in Dubai, UA - October 2009. expected to have a relatively high sulfate-to-chloride ratio, it is possible that pitting will be less of a problem than it would be without the sulfate. This further demonstrates the desirability of conducting corrosion tests with the process fluid. There are ASTM International and NACE [National Association of Corrosion Engineers] International standard tests for evaluating the resistance of alloys to pitting and crevice corrosion. For example, ASTM G48 (ASTM, 2009a) refers to test methods and procedures to determine pitting and crevice corrosion resistance of stainless steels and related alloys when they are exposed to oxidizing chloride environments. These tests are designed to cause localized corrosion more quickly than in most natural environments. Consequently, corrosion damage during testing will generally be more severe than in a natural environment for a similar period of time. Procedures are described and identified for determining critical pitting temperatures for stainless steels and for nickel-base alloys as well as for developing a relative ranking of the susceptibility of such alloys to crevice and pitting corrosion. Test results can be used to rank the resistance of alloys to pitting and crevice corrosion. ASTM G48 uses ferric chloride solutions as the test environment because this test chemistry is related to the chemistry in pit or crevice sites on ferrous alloys in chloride-bearing environments. Relative alloy performance in these tests has been correlated to performance in certain real -38-

OCR for page 5
environments, such as natural, ambient-temperature, seawater, and strongly oxidizing, low-pH, chloride-containing environments. Cyclic polarization is another accelerated test that may be used to assess the relative susceptibility of a series of alloys to pitting and crevice corrosion. ASTM Standard G61 (ASTM, 2009b) describes this approach. In this test, one could use as the test environment a synthetic process fluid that chemically matches the composition of the expected process fluid. In either approach, the objective would be to produce an accelerated evaluation of the susceptibility of alloys to pitting and crevice corrosion. In response to corrosion concerns in the BGCAPP design, Battelle prepared a white paper on the corrosion of 316 stainless steel in the RO unit (Battelle, 2011). In this paper, type 316L is recommended for the filters and pumps, and duplex 2205 stainless steel is recommended for the piping throughout the process. The PMACWA found that 316L pumps are not available. (Note: 316 is actually a wrought alloy designation. The cast alloy designations are CF3M and CF8M for 316L and 316, respectively.) The reason for using the “L” grade is to avoid weld heat-affected zone corrosion, and so the use of cast pumps made of the equivalent to regular grade 316 stainless steel (CF8M) may be appropriate since there would be little or no welding on the pump body. The overall corrosion resistance of CF3M and CF8M would be expected to be similar. This assumes that subsequent testing confirms the suitability of type 316 alloy for this application. Finding. If the testing recommended in this report is not conducted, the use of a duplex alloy such as 2205 for the piping lines may be appropriate, based on the preceding discussion. Stress Corrosion Cracking Stress corrosion cracking occurs when a susceptible alloy is exposed to an environment that is capable of initiating cracking events when sufficient operational tensile stresses are present. Thus, knowledge of the anticipated stress field in any component of interest is paramount, as is the chemistry of the service environment. The 60°C (140°F) threshold for SCC mentioned in the attachments to the Battelle white paper (Battelle, 2011) is not completely borne out in practice. Experience with SCC of austenitic stainless steels (e.g., 304 and 316) suggests that there is little danger of cracking below about 45°C (113°F), as shown in Figure 4. This temperature, from Freedman et al. (2004), is close to the planned operating temperature of the BGCAPP RO unit. While the risk may be small, it would be prudent to confirm the absence of stress corrosion cracking tendencies with laboratory tests. The slow strain rate test, ASTM G129, is used to test resistance to SCC in metallic materials, in a variety of environmental conditions, in an accelerated manner (ASTM, 2006a). It is used for the rapid screening and/or comparative evaluation of the resistance of materials to SCC in relation to environmental, processing, and metallurgical variables. ASTM G129 has been used to evaluate materials, heat treatments, chemical constituents in the environment, temperature, and chemical inhibitors. Slow strain rate testing can be used to evaluate a wide variety of metallic materials in environments that could cause SCC, covering a broad range of temperatures and pressures. Since this is an accelerated test, the results are not meant to necessarily indicate the performance of -39-

OCR for page 5
10000 1000 Chlorides (ppm) 100 10 1 0.1 0 50 100 150 200 250 300 350 Temperature (°C) FIGURE 4 Stress corrosion cracking limits for 304/316 stainless steels in a variety of mixed process waters containing chlorides. SOURCE: Freedman et al., 2004. © MATERIALS TECHNOLOGY INSTITUTE, INC. [2004]. Reprinted with permission of the Copyright Owner. materials in a given operational environment. Rather, it provides a basis for material screening. It can be used to detect environmental interactions with materials. It can also be used to conduct a comparative evaluation of the effects of various metallurgical and environmental variables on the sensitivity of materials to environmental cracking problems. Constant-load or -strain SCC tests should also be conducted in environments that simulate the expected operational environment. If possible, actual operational experience should be gained so that a correlation between the test results and anticipated operational performance can be developed. ASTM G36 is an accelerated test for SCC of various stainless alloys; it uses boiling magnesium chloride as a test environment, but could be modified to serve as a guide to constant-load testing in a simulated service environment (ASTM, 2006b). ISO [International Organization for Standardization] Standard 7539 is a comprehensive stress corrosion testing standard that includes constant-load as well as slow strain rate testing details (ISO, 1989). Saithala et al. (2010) found that the resistance to SCC increased as the PRE number increased, but for all of the alloys tested there was a critical cracking potential above which SCC occurred. This result suggests that using an alloy with a higher PRE number than that of 316L stainless steel would be beneficial if it is determined that type 316L would be risky in this application. 10 MOC Findings and Recommendations Finding. Given that the committee is not aware of any other water recovery systems that treat effluents similar to those expected from the BGCAPP supercritical water oxidation system, the bases for the present material selections are not well supported. Finding. Given the limited information available, it cannot be determined with confidence whether the materials of construction currently selected will be adequate for their planned applications. It is possible that the corrosion resistance of type 316L stainless steel may be adequate for the piping, filters, and pumps (CF3M or CF8M) in 10 See footnote 9. -40-

OCR for page 5
this system for the planned operational lifetime of the system. However, it is also possible that it would be inadequate. Laboratory testing would reduce this uncertainty. Finding. Testing at elevated temperatures may be required for the piping between the supercritical water oxidation reactor and the first pretreatment unit. Recommendation. Laboratory qualification testing should be conducted to validate the selection of materials from a corrosion engineering perspective. The modes of corrosion that should be investigated are pitting, crevice corrosion, and stress corrosion cracking (SCC). Accelerated potentiodynamic polarization testing is a useful approach to evaluation of pitting and crevice corrosion. With the knowledge of the stress state anticipated on relevant components, SCC testing can be pursued. The slow strain rate test is recommended for evaluating SCC. These corrosion tests should be conducted in a chemical environment similar to that expected in the supercritical water oxidation effluents and the reverse osmosis reject. Candidate alloys for evaluation include type 316 stainless steel, a duplex alloy such as 2205, a super duplex alloy, and perhaps a super austenitic stainless steel. Recommendation. Unless the corrosion testing recommended in this report is performed, to be conservative at the least an alloy such as 2205 SS should be used where appropriate. REFERENCES Arnvig, P-E., B. Leffler, E. Alfonsson, and A. Brorson. 1996. Machinability, corrosion resistance and weldability of an inclusion modified 2205 duplex stainless steel. acom 2-1996. Available at http://www.avestapolarit.com/upload/documents/technical/ acom/Acom96_2a.PDF. Accessed November 2, 2011. ASTM (ASTM International). 2006a. Standard Practice for Slow Strain Rate Testing to Evaluate the Susceptibility of Metallic Materials to Environmentally Assisted Cracking, ASTM G129-00(2006). West Conshohocken, Pa.: ASTM International. ASTM. 2006b. Standard Practice for Evaluating Stress-Corrosion-Cracking Resistance of Metals and Alloys in a Boiling Magnesium Chloride Solution, ASTM G36-94(2006). West Conshohocken, Pa.: ASTM International. ASTM. 2009a. Standard Test Methods for Pitting and Crevice Corrosion Resistance of Stainless Steels and Related Alloys by Use of Ferric Chloride Solution, ASTM G48- 03(2009). West Conshohocken, Pa.: ASTM International. ASTM. 2009b. Standard Test Method for Conducting Cyclic Potentiodynamic Polarization Measurements for Localized Corrosion Susceptibility of Iron-, Nickel-, or Cobalt-Based Alloys, ASTM G61-86(2009). West Conshohocken, Pa.: ASTM International. Avista (Avista Technologies, Inc.). 2005. Compatibility Study: Vitec 3000 Versus Roquest 3000, May. San Marcos, Calif.: Avista Technologies, Inc. -41-

OCR for page 5
Battelle. 2011. WP-112: Corrosion of 316 Stainless Steel in Reverse Osmosis Reject Water Lines, Rev. B, February 22. Richmond, Ky.: Blue Grass Chemical Agent Destruction Pilot Plant Project. Bhattacharyya, D., W.C. Mangum, and M.E. Williams. 1999. Reverse osmosis. Pp. 1757- 1759 in Kirk-Othmer Encyclopedia of Chemical Technology, 4th Ed. J.I. Kroschwitz, ed. New York, N.Y.: Wiley. Bechtel. 2007. Mechanical Systems Data Sheet: Reverse Osmosis Unit, Rev. E, October 15. Richmond, Ky.: Blue Grass Chemical Agent Destruction Pilot Plant Project. BPBG (Bechtel Parsons Blue Grass). 2005. Technical Risk Reduction Project (TRRP) 07 and 09 Report on Supercritical Water Oxidation Blended Feed Performance Tests, Rev. 0, May 5. Richmond, Ky.: Blue Grass Chemical Agent Destruction Pilot Plant Project. BPBG. 2006. Trend Notice, SPB Deletion of Clarifier and Clear Well Tank, Trend Number TN-24915-06-00126, September 13. Richmond, Ky.: Blue Grass Chemical Agent Destruction Pilot Plant Project. BPBG. 2007. SCWO Building Water Recovery—R.O. Unit Process Flow Diagram, Rev. 6, Sheet 1/2, October 2. Richmond, Ky.: Blue Grass Chemical Agent Destruction Pilot Plant Project. BPBG. 2009. System Design Description for Water Recovery System, Rev. 2, July 16. Richmond, Ky.: Blue Grass Chemical Agent Destruction Pilot Plant Project. BPBG. 2010. SCWO Building Water Recovery—Tanks Process Flow Diagram, Rev. 8, Sheet 2 of 2, June 11. Richmond, Ky.: Blue Grass Chemical Agent Destruction Pilot Plant Project. Brehant, A., V. Bonnelye, and M. Perez. 2002. Comparison of MF/UF pretreatment with conventional filtration prior to RO membranes for surface seawater desalination. Desalination 144(1-3):353-360. Byrne, G., R. Francis, G. Warburton, and J. Wilson. 2009. Electrochemical potential and the corrosion resistance of stainless steels in SWRO applications. Paper presented at the International Desalination Association World Congress 2009, October, Dubai, UAE. Available at http://www.rolledalloys.com/technical-resources/corrosion- resistant-alloys. Accessed November 2, 2011. Crum, J.R., and L.E. Shoemaker. 2009. Defining acceptable environmental ranges and welding procedures for corrosion resistant alloys, Paper 09381. NACE International Corrosion 2009 Conference Proceedings. Houston, Tex.: NACE International. Francis, R., G. Byrne, and G. Warburton. 2011. The corrosion of superduplex stainless steel in different types of seawater, Paper 11351. NACE International Corrosion 2011 Conference Proceedings. Houston, Tex.: NACE International. Freedman, A.J., A.S. Krisher, and D. Steinmeyer. 2004. Guidelines for Troubleshooting Water Cooled Heat Exchangers. St. Louis, Mo.: Materials Technology Institute. Ho, W.S.W., and K.K. Sirkar. 1992. Membrane Handbook. New York. N.Y.: Van Nostrand Reinhold. -42-

OCR for page 5
ISO (International Organization for Standardization). 1989. Corrosion of Metals and Alloys—Stress Corrosion Testing—Part 4: Preparation and Use of Uniaxially Loaded Tension Specimens, ISO 7539-4:1989. Geneva, Switzerland: ISO. Li, N.N., A.G. Fane, W.S.W. Ho, and T. Matsuura. 2008. Advanced Membrane Technology and Applications. Hoboken, N.J.: Wiley. Malik, A.U., M. Mobin, F. Al-Muaili, and S. Al-Fozan. 2011. Corrosion behavior of duplex stainless steels in Arabian seawater, Paper 11171. NACE International Corrosion 2011 Conference Proceedings. Houston, Tex.: NACE International. NRC (National Research Council). 2005. Interim Design Assessment for the Blue Grass Chemical Agent Destruction Pilot Plant. Washington, D.C.: The National Academies Press. Pendergast, M.T.M., and E.M.V. Hoek. 2011. A review of water treatment membrane nanotechnologies. Energy and Environmental Science 6(4):1946-1971. Pohjanne, P., A. Sarpola, L. Carpen, M. Riihimaki, P. Kinnunen, T. Hakkarainen, and J. Ramo. 2007. Stainless steel pitting in chloridesulfate solutions: The role of cations, Paper 07198. NACE International Corrosion 2007 Conference Proceedings. Houston, Tex.: NACE International. Saithala, J.R., S. McCoy, J.D. Atkinson, and H.S. Ubhi. 2010. Critical stress corrosion cracking potentials of stainless steels in dilute chloride solutions, Paper 10287. NACE International Corrosion 2010 Conference Proceedings. Houston, Tex.: NACE International. Siavash, M.S., T. Mohamamdi, and M.K. Moghadam. 2001. Chemical cleaning of reverse osmosis membranes. Desalination 134(1):77-82. Wilf, M., and M.K. Schierach. 2001. Improved performance and cost reduction of RO seawater systems using UF pretreatment. Desalination 135(1-3):61-68. Williams, M.E. 2003. A Brief Review of Reverse Osmosis Membrane Technology. Available at http://www.wescinc.com/files/RO_Review.pdf. Accessed November 23, 2011. -43-