4

Scale-Up, Stability, and Reliability

PROCESS SCALE-UP

Although significant research and development have been focused on using SCWO for the destruction of organic compounds (Shanableh and Gloyna, 1991; Tester et al., 1993; Modell, 1989), a number of technical challenges must still be overcome before widespread, full-scale implementation will be feasible. These include better management of salts/solids, better understanding of the phase behavior and kinetics/flow dynamics for scale-up, a better understanding of the corrosion characteristics of potential materials of construction, and the development of process monitoring and control regimes.

Because a fundamental understanding of the fluid dynamics, mixing processes, and reaction kinetics occurring in a SCWO reactor does not exist, process scale-up has been based on reactor residence time, engineering judgment, and significant pilot-scale testing. This empirical approach to scale-up is especially important for complex systems, like VX hydrolysate, that contain large quantities of salt and may have several phases. For example, mixing of materials in the reactor is important; cold feed must mix with hot product to bring the material to reaction temperature; solids condensed from the supercritical fluid will impinge on, and may stick to, the reactor walls or other surfaces. There is no good way to extrapolate these complex flow effects from smaller to larger scale.

Fundamental information on the number of phases and the phase behavior of VX hydrolysate, as well as the kinetics of destruction and related oxidant effects, mixing and fluid dynamics, salt nucleation and precipitation characteristics, and the handling of salts within and downstream of the reactor must be determined before equipment can be scaled up from bench-scale or limited pilot-scale testing. Unfortunately, the Army's schedule for VX destruction does not allow sufficient time for thorough development of a fundamental understanding. Therefore, significant pilot testing and development will be needed for designing the full-scale treatment unit. In addition, the full-scale design and testing protocol should be flexible enough to incorporate findings, as well as engineering and equipment changes to accommodate potential problems that may be encountered.

The key areas of concern for scale-up application of SCWO to VX hydrolysate are: (1) salts management and solids handling; (2) the impact of the oxidant (e.g., air versus oxygen) on mixing and heat balance; (3) corrosion of the materials of construction; and (4) reliability of pressure let-down systems. A better understanding of the design, engineering, and operational implications of these issues is also critical for demonstrating the stability and reliability of the full-scale process. The pilot testing on VX hydrolysate described in the previous chapter focused on demonstrating the destruction of VX and did not specifically address these issues, although it did reveal potential problems in all of the aforementioned areas. The design issues are discussed in more detail below. Table 4-1 provides a comparison of the Army's full-scale design (described in detail in Chapter 5 ) with the pilot-scale designs.

SALT MANAGEMENT AND REACTOR DESIGN

Given the large quantities of salt generated by SCWO treatment of VX hydrolysate, it is clear that solids management, both within and downstream of the SCWO reactor, is a critical issue. In general, the separation of inorganic salts generated during SCWO can be problematic. At the low densities and solvent dielectric of typical SCWO systems (200 to 275 bar, 500 to 650°C), most inorganic salts dissociate very little and are essentially insoluble. These solids may adhere to the reactor walls or associated plumbing hardware (valves, inlets, feed and exit lines, etc.).

Precipitated salts can also clog pressure let-down systems. It may be possible to “unclog” reactor parts via cooling and flushing with water, which should lead to redissolution. However, this fix would require



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 35
Using Supercritical Water Oxidation to Treat Hydrolysate from VX Neutralization 4 Scale-Up, Stability, and Reliability PROCESS SCALE-UP Although significant research and development have been focused on using SCWO for the destruction of organic compounds (Shanableh and Gloyna, 1991; Tester et al., 1993; Modell, 1989), a number of technical challenges must still be overcome before widespread, full-scale implementation will be feasible. These include better management of salts/solids, better understanding of the phase behavior and kinetics/flow dynamics for scale-up, a better understanding of the corrosion characteristics of potential materials of construction, and the development of process monitoring and control regimes. Because a fundamental understanding of the fluid dynamics, mixing processes, and reaction kinetics occurring in a SCWO reactor does not exist, process scale-up has been based on reactor residence time, engineering judgment, and significant pilot-scale testing. This empirical approach to scale-up is especially important for complex systems, like VX hydrolysate, that contain large quantities of salt and may have several phases. For example, mixing of materials in the reactor is important; cold feed must mix with hot product to bring the material to reaction temperature; solids condensed from the supercritical fluid will impinge on, and may stick to, the reactor walls or other surfaces. There is no good way to extrapolate these complex flow effects from smaller to larger scale. Fundamental information on the number of phases and the phase behavior of VX hydrolysate, as well as the kinetics of destruction and related oxidant effects, mixing and fluid dynamics, salt nucleation and precipitation characteristics, and the handling of salts within and downstream of the reactor must be determined before equipment can be scaled up from bench-scale or limited pilot-scale testing. Unfortunately, the Army's schedule for VX destruction does not allow sufficient time for thorough development of a fundamental understanding. Therefore, significant pilot testing and development will be needed for designing the full-scale treatment unit. In addition, the full-scale design and testing protocol should be flexible enough to incorporate findings, as well as engineering and equipment changes to accommodate potential problems that may be encountered. The key areas of concern for scale-up application of SCWO to VX hydrolysate are: (1) salts management and solids handling; (2) the impact of the oxidant (e.g., air versus oxygen) on mixing and heat balance; (3) corrosion of the materials of construction; and (4) reliability of pressure let-down systems. A better understanding of the design, engineering, and operational implications of these issues is also critical for demonstrating the stability and reliability of the full-scale process. The pilot testing on VX hydrolysate described in the previous chapter focused on demonstrating the destruction of VX and did not specifically address these issues, although it did reveal potential problems in all of the aforementioned areas. The design issues are discussed in more detail below. Table 4-1 provides a comparison of the Army's full-scale design (described in detail in Chapter 5 ) with the pilot-scale designs. SALT MANAGEMENT AND REACTOR DESIGN Given the large quantities of salt generated by SCWO treatment of VX hydrolysate, it is clear that solids management, both within and downstream of the SCWO reactor, is a critical issue. In general, the separation of inorganic salts generated during SCWO can be problematic. At the low densities and solvent dielectric of typical SCWO systems (200 to 275 bar, 500 to 650°C), most inorganic salts dissociate very little and are essentially insoluble. These solids may adhere to the reactor walls or associated plumbing hardware (valves, inlets, feed and exit lines, etc.). Precipitated salts can also clog pressure let-down systems. It may be possible to “unclog” reactor parts via cooling and flushing with water, which should lead to redissolution. However, this fix would require

OCR for page 35
Using Supercritical Water Oxidation to Treat Hydrolysate from VX Neutralization TABLE 4-1 Comparison of Full-Scale Design with Pilot-Scale Design Tested with VX Hydrolysate (February 1997) Reactor Property Pilot Scale Full Scale Reactor size 4.25 in ID × 6 ft long 10 in ID × 15 ft long Operating conditions (T, P) 650°C, 4,000 psi 650°C, 4,000 psi Oxidant and oxidant stoichiometry air (100 percent excess) oxygen (50 percent excess) Maximum operating interval 8 hrs undefined, TBD a Feed composition VX hydrolysate/H2O; salt simulant VX hydrolysate, ton container cleanout effluent, decontamination fluids Mass flow rate 63.7 lb/min•ft2 105.7 lb/min•ft2 Residence time b 40 sec 60 sec Heating flow rate 0.19 × 106 Btu/hr 5 × 106 Btu/hr Linear velocity b 8.25 ft/min 13.7 ft/min Pressure let-down system valves TBD Materials of construction titanium (corroded metals observed in effluent) TBD (platinum suggested; more testing needed) aTBD = to be determined bEstimated based on exit condition Source: Adapted from General Atomics, 1997b; Stone and Webster, 1997a. considerable shutdown and start-up time and, depending on frequency, could result in additional thermal and pressure stresses on system components. Large quantifies of precipitated salts left in a reactor could significantly reduce reactor volume, with commensurate shortening of residence times, thereby leading to reduced DREs (destruction removal efficiencies). These possibilities suggest that either the management of solids must be better understood or stringent monitoring of destruction removal efficiencies may be necessary. In addition, the effect of precipitated salt on mixing and fluid flow characteristics could be significant. Scale formation by carbonate generated from organic carbon oxidation may also present a problem and must be properly managed. Despite these potential difficulties, it appears, as described below, that engineering and mechanical equipment that would mitigate some of these problems are becoming available; however, further testing at appropriate scales to define the full-scale design is necessary. Solid salts that could be expected to separate out in the reactor from the oxidation of VX hydrolysate in the presence of NaOH include sodium phosphates, sodium sulfate, and possibly sodium carbonate (depending on pH). Several research studies have focused on understanding the behavior of salts in subcritical and supercritical water, including the phase behavior of NaCl, NaSO4, and phosphates (Broadbent et al., 1997; Tester et al., 1993; Martynova, 1976; Martynova and Smirnov, 1964; Ravich and Yastrebova, 1959). Although considerable progress has been made in understanding and measuring the solubility and phase behavior for simple systems, the behavior of salts in SCWO systems is a complex and poorly understood phenomenon that nevertheless must be managed for effective SCWO implementation. Recently, a number of approaches have been investigated for managing salts more effectively. These include higher system pressures, solubilization with added NaOH, lower temperature catalyzed reactions at higher densities, and new reactor concepts (i.e., transpiring wall reactors). Although these approaches may be promising, none has been developed enough for full-scale treatment of VX hydrolysate. All of these approaches will require considerably more development and pilot-scale work than appears to be necessary for a vertical

OCR for page 35
Using Supercritical Water Oxidation to Treat Hydrolysate from VX Neutralization cylindrical reactor operated at approximately 650°C and 275 bar. Another alternative would be to increase the dilution of hydrolysate to reduce salt concentration. However, this approach would require continuous feed of supplemental fuel, and a proportionate scale-up of all process components after the neutralization process step; it would also potentially increase process complexity and substantially increase capital and operating costs. Therefore, this option should be considered only if additional pilot-scale testing indicates that SCWO operation at the currently specified hydrolysate concentration is not practical. Higher Pressure Systems At significantly higher pressures, the density of water can increase to the point that many salts become soluble again in supercritical water (Foy et al., 1994). This higher pressure SCWO system looks very promising; however, a good deal more development work would be needed to assess its potential for VX hydrolysate. Higher pressure systems could also increase the cost of equipment considerably, which would have to be weighed against other salt management issues. Dissolution in Molten Sodium Hydroxide Several researchers have investigated the concept of using high concentrations of NaOH to prevent carbonates and, presumably, a wide range of other salts from precipitating in SCWO reactors (Borovaya and Ravich, 1968). High loadings of NaOH can result in a separate molten NaOH phase under SCWO conditions, which has been demonstrated to dissolve significant quantities of carbonate and, presumably, a number of other salts. Unfortunately, strongly caustic high temperature water is very corrosive, and materials of construction then become a considerable concern. Catalyzed Oxidation Catalyzed oxidations at decreased temperatures result in higher densities and salt solubilities. At lower temperatures, (= 350°C), salts can be considerably more soluble and thus kept in the reaction medium. The penalty of lower temperatures is significantly lower DREs. Several studies (Ding et al., 1995) have reported the use of homogeneous or heterogeneous catalysts, which flow through the reactor to enhance oxidation kinetics and increase DREs (e.g., metal systems, solid carbonate). Disadvantages of catalyzed reactions include that the catalyst activity and concomitant DRE must be closely monitored and that the catalyst most likely will have to be separated from SCWO effluent. Little is known about catalyst lifetime or about poisoning and attrition mechanisms. Transpiring Wall Reactor The recently developed transpiring wall platelet is claimed to overcome corrosion of the reactor wall and salt deposition by deposition of a thin film of clean (often much lower temperature) water uniformly along the reactor wall. Testing to date, however, indicates salt buildup in the reaction entrance region. Fabrication of the platelet may be complex if noble metals (e.g., platinum) are required for corrosion resistance (Shoenmann et al., 1997). A liquid rocket engine type injection system has also been described, which can be used for rapid mixing at the reactor inlet. Although this technology is currently slated for full-scale treatment of Navy shipboard wastes and Army smokes and dyes, most, if not all, testing to date has been on bench-scale systems with tube diameters of about 7.6 cm. Engineering Control Several vendors have claimed that, through considerable testing, they should be able to design engineering controls (often proprietary) to manage salts in SCWO effluents more effectively. Although it may be possible to remove inorganic salts directly as dry solids or concentrated salt solutions or slurries, this would require pressure let-down at elevated temperatures. Instead, most scenarios involve pressure let-down after cooling (via the addition of extra water and/or heat exchange) to allow for redissolution of most, if not all, salts. Conclusion The first four options, although they may seem promising, are somewhat immature or underdeveloped, and

OCR for page 35
Using Supercritical Water Oxidation to Treat Hydrolysate from VX Neutralization the fifth involves vendor-specific proprietary technology that may require larger scale testing. The Army has opted for a full-scale design that incorporates cooling of the SCWO effluent by adding cold water and heat exchange prior to depressurization. Salts are then recovered through evaporation. CHOICE OF OXIDANT, MIXING, AND HEAT BALANCE The pilot tests on VX hydrolysate used air as the oxidant at 100 percent stoichiometric excess. The full-scale design calls for pure oxygen at 50 percent stoichiometric excess. The substitution of pure oxygen for air could have significant implications for mixing and for the heat capacity of the waste stream. The reduction in excess oxygen will probably not affect the destruction of hydrolysate constituents. The reaction has been demonstrated to be approximately zero order with respect to oxygen when sufficient excess oxygen is present. Bench-scale data on MPA oxidation showed essentially the same conversion at 50 percent and 100 percent excess oxygen, as would be expected for a zero order reaction (Gloyna and Li, 1997). A much smaller amount of excess oxygen (10 percent excess) resulted in reduced MPA conversion (92.1 percent versus 99.9). The residence time for the full-scale reactor will be longer than that of the laboratory unit. Pilot-scale testing completed to date has demonstrated greater conversion of hydrolysate constituents than anticipated based on laboratory testing with MPA. The volumetric flow rate of fluid through the reactor is reduced substantially by eliminating nitrogen (from air) and by using less excess oxidant. The molar flow rate is also reduced by about 50 percent. As a consequence, the residence time for the full-scale unit is calculated to be twofold greater than for the pilot-scale unit. The use of air as an oxidant results in the addition of considerable quantities of nitrogen gas (up to about 50 percent by volume for the VX hydrolysate pilot test), which could result in different fluid phase characteristics within the reactor than when pure oxygen is the oxidant. Flow dynamics and mixing could be significantly different in a larger scale system if the phase behavior is significantly different. The most critical impact of changes in fluid dynamics could be on the transport of salts out of the reactor at reduced shear stresses, which may result in increased salt adhesion, reactor plugging, or lower conversion efficiencies. In the pilot-scale reactor, the solid salts were carried out by the flowing fluid intermittently with reasonable frequency (e.g., less than 30 seconds between peaks). The physical situation in the reactor is not well enough understood to predict whether added nitrogen and potentially added turbulence would be beneficial or detrimental to conversion efficiency and salt removal from the reactor. The effects of substituting oxygen for air cannot be determined confidently without some pilot-scale testing. The presence or lack of a considerable amount of nitrogen can significantly affect the heat capacity for the process stream, as well as the heat balance for the full-scale design. By eliminating the nitrogen, the heat capacity of the material flowing through the reactor would be reduced more than twofold. High levels of hydrolysate conversion in the reactor depend on rapid mixing to bring the relatively cold feed up to reaction temperature very quickly. Again, it is not clear how the reduced heat capacity and mixing of the fluid will affect the ability of the reactor to increase the temperature of the inlet feed stream. Another effect of changing the oxidant from air to oxygen is that the exit gas from the SCWO reactor will be oxygen-rich. This characteristic raises the possibility of combustion of the carbon filter on the SCWO vent unless the exit gas is cooled or diluted. This question should be addressed in the final process design and quantitative risk assessment for the SCWO process. Another change in the feed to the full-scale facility relative to pilot testing is that it will include the treatment of spent decontamination fluid/NaOH solution, with a small amount of organic material (presumably the same products found in the hydrolysate). The spent decontamination fluid is planned to be treated in combination with the effluent from ton container cleanout. The presence of these fluids will reduce the heat released in the reactor. Some auxiliary fuel (e.g., diesel) may be needed to maintain the reactor temperature at 650°C. CORROSION AND THE MATERIALS OF CONSTRUCTION Materials of construction that can withstand the extreme pressures, temperatures, and often corrosive conditions within a SCWO reactor remain one of the key challenges for the broad implementation of SCWO technology. This is particularly true for the treatment of wastes that contain heteroatoms and generate significant

OCR for page 35
Using Supercritical Water Oxidation to Treat Hydrolysate from VX Neutralization quantities of oxyacids or their salts depending on the cations present in the system. Numerous studies have demonstrated that streams that contain chloride are especially corrosive. In general, it is believed (and some test results have corroborated) that salt solutions at intermediate temperatures (e.g., during pressure letdown or during heat-up, or near inlet or exit areas) can be much more corrosive than the supercritical fluid in the reactor. Erosion from precipitated salts at high velocities could also be an important issue. Materials of construction are an important consideration for the reactor inlet system, the reactor itself, and all hardware and plumbing downstream of the reactor (e.g., pressure let-down equipment and reactor exit plumbing). Recent studies have demonstrated that corrosion is a critical issue for SCWO hazardous waste disposal systems. Test system materials have exhibited pitting, stress corrosion cracking, and accelerated general corrosion (Latanision, 1995; Latanision et al., 1997). The operational performance and reliability of SCWO will depend on mitigating corrosion under operating conditions (Latanision et al., 1997). Highly caustic VX/NaOH hydrolysate (pH~14), combined with the relatively high temperature and high pressure of the SCWO process, produces highly corrosive and erosive mixtures of reactants, reaction intermediates, and products. Pilot-scale tests in a titanium reactor produced significant amounts of solid TiO2 and other metal oxide corrosion products after only a few hours of operation (General Atomics, 1997a,b). Careful selection of internal materials of construction will be essential, or reactor lifetime will be short, and system efficiency will be compromised. Welds and other joints are often sites of severe corrosion, so the reactor should be designed with the minimum number of joints and welds necessary for structural integrity. Furthermore, the failure of reactor liner joints should be anticipated as a possible process failure mode. Metal coupon tests sponsored by the Defense Advanced Research Projects Agency at General Atomics (General Atomics, 1997c), as well as analyses of the survivability of metallic thermocouples in Armysponsored tests (General Atomics, 1997b), indicate that noble metals, specifically platinum and gold, withstand the corrosive assault best, particularly under the caustic (high pH) conditions near the reactor inlet. Although noble metal reactors will clearly be expensive, they will probably still represent a small fraction of the overall cost of a SCWO system. System reliability may be enhanced and down time reduced by reactor designs that include replaceable internal sleeves in the reactor entrance section, which is exposed to caustic; parallel reactors; or systems in which spare reactors can easily be substituted. Additional data on materials lifetimes and failure modes under SCWO conditions representative of those with VX/NaOH hydrolysate will be necessary for making informed selections of reactor construction materials and operating parameters. Current information is insufficient for the confident selection of materials. The Army recognizes that selecting the best materials of construction is a key issue and has embarked upon a materials testing program. PRESSURE LET-DOWN SYSTEMS Reliable and stable operation of the full-scale SCWO process will require a robust system for depressurization that is not subject to clogging by precipitated salts and is not vulnerable to significant corrosion. The pressure let-down system may be the weak link in the full-scale process chain. The pressure let-down system will probably have to handle some precipitated salts because all salts may not redissolve prior to depressurization. Furthermore, as described above, the depressurization system will be exposed to a fairly corrosive environment. Thus, materials of construction will have to be tested for corrosion and reliability. The full-scale design calls for the addition of cold water and for further cooling by a heat exchanger prior to pressure let-down. The pilot-scale test used back-pressure valves for heat exchange, which were not very reliable. The final fullscale design will be specified by the system contractor. SUMMARY The operating and reactor parameters for the fullscale design, as well as for the pilot-scale test, are summarized in Table 4-1 . The scale-up factor from pilot-scale test to full-scale design is about 25, in terms of the amount of VX hydrolysate to be treated (13,600 kg/day versus 545 kg/day). The operating temperature and pressure are designed to be the same (650°C, 275 bar). The pilot plant ran on air at 100 percent stoichiometric excess, while the full-scale design calls for pressurized oxygen at 50 percent excess. To date, pilot-scale testing has demonstrated a continuous operational time of only eight hours; the design time for

OCR for page 35
Using Supercritical Water Oxidation to Treat Hydrolysate from VX Neutralization continuous on-line operations for the full-scale plant has not been established. The residence time for the full-scale reactor will be increased by a factor of about two relative to the pilot-scale test if oxygen rather than air is used. The feed to the full-scale reactor will include fluid from ton container cleanout and spent decontamination fluid in addition to VX hydrolysate. Additional fuel will be needed to make up for the decreased heating value of these two streams and maintain the reactor temperature at 650°C. The pressure let-down system and the materials of construction still must be determined through testing. There is no strong fundamental basis for the scale-up of SCWO. Given that the Army's schedule does not permit substantial fundamental work on VX hydrolysate to identify and measure fundamental data (e.g., kinetics, reaction order, fluid dynamics and mixing, corrosion, and salt nucleation and precipitation), a substantial amount of pilot-scale work should be done to define the engineering parameters more confidently and to resolve the issues of scale-up and materials of construction discussed above. Furthermore, given that there will probably still be uncertainties for the full-scale plant, substantial start-up time may be necessary, and the plant may have to operate in a developmental mode for some time. Facility planning should provide for sufficient flexibility to take this into account, as well as to implement key lessons learned from full-scale or larger testing.