In Part 1 of this two-part series (see MPS, March 2009, pp 15-19), Dan Sampson, now with WorleyParsons, Sacramento office, discussed the general design options available for zero liquid discharge (ZLD) systems. This second article focuses on the many challenges associated with the operation and maintenance of such systems.
The large variety of ZLD designs precludes an exhaustive examination of every problem, but several common problems occur across the broad range of ZLD designs. These common problems occur in several different categories. Some are general in nature and apply regardless of the ZLD design. Others are more specific. These include problems with preconcentration systems (wastewater reverse osmosis and pretreatment) and problems with thermal systems (evaporators, brine concentrators, and brine crystallisers).
General problems and solutions
All ZLD systems, regardless of design, face challenges associated with manpower, degradation and redundancy, and inconsistent operating approach. Addressing these concerns, regardless of the ZLD equipment, will improve reliability and capacity while reducing maintenance.
Most ZLD owners begin with the mistaken assumption that ZLD systems can be operated with the same staff and resources as a typical water treatment plant. Part 1 of this two-part series explained in detail that ZLD systems are entirely different. They’re much more complex and increased manpower is required. The additional manpower depends on complexity. Relatively “simple” ZLD designs require 7 – 9 additional and experienced operators and maintenance personnel. A simple design might consist of thermal concentration followed by solid waste production.
Even simple ZLD systems must be monitored continuously – 24 hours per day, 7 days per week – by a dedicated operator with no other duties. In addition, even simple designs require mechanical work (at least four hours per operating day) and electrical/instrumentation support (at least four hours per operating day). Finally, management is essential. In simple designs the ZLD system manager can carry other duties. Plant chemists, for example, may also act as ZLD system managers to co-ordinate operations and maintenance activities.
“Complex” ZLD designs require 12-14 extra people. Complex designs include multiple processes. Preconcentration followed by thermal concentration and solid waste production, for example, represent a “complex” design. Complex systems require at least two operators, again 24 hours per day and 7 days per week. One operator typically monitors the preconcentration equipment (wastewater reverse osmosis pretreatment, the wastewater RO system, etc), while the second operator monitors and controls the thermal and solid waste production portions of the ZLD plant. Increased complexity also requires increased maintenance. Complex systems typically require a dedicated mechanic and a dedicated electrician or instrument technician. In addition, a full-time manager is required.
ZLD manufacturers sometimes create the wrong impression. No ZLD plant “runs itself”. These systems require at least one experienced person, two for complex systems, to constantly monitor the process and run chemistry. It’s important that these operators have no other duties. The ZLD process changes constantly and requires constant oversight.
Degradation and redundancy
Most ZLD engineering work focuses on nameplate capacity and capital cost. Risk analysis receives little attention. Owners tend to purchase the least expensive system available with a nameplate capacity to treat the required flow. The ZLD system feed chemistry is often uncertain, but any deviation from a design feed chemistry results in a loss of ZLD system capacity. The result is that ZLD systems almost never meet their nameplate capacities.
In the best case, system owners can expect 75%-95% mechanical reliability of the ZLD system components. Further, expect degradation of at least 20% between cleanings (80% of nameplate capacity). These two performance challenges mean that a typical ZLD system operates, on average, at around 60%-70% of its nameplate capacity. For example, a ZLD system sized to treat 500 gallons per minute can reliably treat only 300–350 gallons per minute on average.
It’s important to remember that this represents the best case. Effective system capacity decreases as ZLD system complexity increases because much more can go wrong. Additional capacity must be purchased and installed. For “simple” designs the system should consist of at least two independent trains, each train sized to treat 60% of the peak ZLD influent flow. That provides an effective capacity of 90%-120% of the nameplate for each train. The redundancy requirement increases for complex designs. Complex systems should also include two independent trains, but each train must be sized to treat 100% of the peak influent flow. Increasing complexity results in decreased reliability, so more redundancy (additional design margin) must be included.
System reliability can be improved through better management of spare parts. A single-point failure analysis should be performed. Purchase additional shelf or installed spares (regardless of design) if the failure of that component will result in a ZLD system shutdown. This risk assessment should be part of initial plant design and the cost analysis of ZLD options should include redundancy.
Inconsistent operating approach
The ZLD system manufacturer typically supplies an “operating philosophy”, but no detailed operating procedures. Individual operators often address both normal control and upsets differently. These varied approaches may all be correct, but they add variability to an inherently unstable process. This increasing variability results in a further loss of effective ZLD treatment capacity.
Detailed operating procedures are essential, but they do take a long time to develop (at least a year). The development of detailed operating procedures requires a dedicated, full-time, senior operator. The procedure developer must review vendor documentation and also watch each of the ZLD system operators as they conduct major evolutions (startup, shutdown, cleanings, upsets, etc). The resulting procedure provides consistency that’s crucial to maximize ZLD system reliability and controllability.
Dedicated staff is also important. Fewer “cooks in the kitchen” minimises variation in process control. ZLD system operation is at least as complex as the operation of most larger industrial or power production processes. Plants don’t put junior operators in the control room and the same concept applies to the ZLD plant.
Plants that dedicate staff to ZLD operation experience fewer and less frequent upsets. As stated earlier, most plants begin with the assumption that the ZLD system can be operated like any other water treatment system. In broad terms, that usually means one operator is responsible for all of the main plant’s water treatment systems as well as the operation of the ZLD system. Most plants abandon their initial manpower plan within weeks and hire or contract additional manpower dedicated to operating the ZLD plant.
Preconcentration systems: problems and solutions
As noted in Part 1 of this two-part series, preconcentration of the waste stream minimises the size of final concentration equipment. Preconcentration, also called “intermediate concentration”, raises the total dissolved solids in the waste stream using relatively low-energy processes such as wastewater reverse osmosis (WWRO). The WWRO permeate recycles to the front end of the plant to supplement raw water. The WWRO reject passes to the final concentration equipment (brine concentrators or crystallisers, in most cases).
Sparingly-soluble minerals (calcium, magnesium, and silica) must be removed or stabilised first to prevent WWRO membrane scaling, but the removal process isn’t perfect. These remaining minerals concentrate in the reverse osmosis reject. This section describes the common problems and solutions associated with preconcentration systems. Problems include WWRO membrane fouling and WWRO scale formation.
WWRO membrane fouling
Suspended solids can cause membrane fouling in both pretreatment RO and WWRO systems. The problem is especially severe in WWRO systems, however, because WWRO feedwater often contains very large numbers of very small particles. Figure 1 provides a particle analysis for the influent feed to a WWRO system.
Experience indicates that these particles are extremely difficult to remove using traditional total suspended solids (TSS) removal technologies. The particles don’t coagulate well, so they tend to pass through clarifiers. Media filters can generally remove particles larger than about 10 microns in size, but the majority of these particles are much smaller than that. As Figure 1 demonstrates most particles are less than 1 micron in size. Thus, filters and clarifiers don’t remove these small particles. Silt density index (SDI) may be high and the WWRO membranes may foul rapidly so cleaning frequency and membrane degradation increase.
The cleaning frequency of reverse osmosis elements used to purify raw water varies, but typically runs from cleaning once per month to cleaning once per six months. It’s not uncommon for WWRO elements to foul and require cleaning every few days.
A particle analysis profile of entire system should be performed to determine particle origin. It is important to know where the small particles come from, where they concentrate, and where they tend to decrease. The answer may not be straightforward. In some cases the particles come primarily from the makeup water, while in others the cooling tower produces the majority of these small particles.
The small-particle fouling issue is especially severe when the plant makeup water comes from recycled or surface water sources. Both recycled and surface waters tend to contain large numbers of small particles with little shape-charge. The lack of a shape-charge makes both coagulation and flocculation challenging, hindering the formation of larger particles that can be removed by traditional filtration technologies.
Cooling water chemical treatment creates further challenges. Most cooling water chemical treatment programmes include the use of both mineral and general dispersants. These products are intended to prevent fouling and scale formation in the cooling tower and cooling system heat exchangers. Unfortunately the benefits provided by dispersants in cooling water become problems when attempting to treat cooling tower blowdown in a ZLD system. The cooling water dispersant tends to reinforce the shape-charges that keep particles apart or interferes with the formation of desired precipitates.
The right amount of cooling water dispersant is absolutely essential, but dispersant feed should be minimised to the maximum extent possible. Automatic control of cooling tower dispersant feed is essential. Automatic control should include the capability to control dispersant feed using direct, online, automatic measurement of the dispersant residual. Control of the active dispersant component establishes a relatively constant threshold which must be overcome in cooling tower blowdown treatment systems.
Particulate source control and cooling tower dispersant control can help minimise the number of small particles and their tendency to disperse, but some chemical treatment to remove suspended solids will be required. Jar testing must be performed to determine the most effective suspended solids removal chemistries. Jar testing should be performed any time there’s a significant change in the chemistry of any of the ZLD system feed streams.
Chemical feed points, especially coagulant chemical feed points, are often poorly located. Coagulant effectiveness increases as contact time with the treated fluid increases. Many ZLD system designs incorporate a coagulant feed point literally a few feet before the ZLD feedwater enters the filtration system. The coagulant has virtually no time to work when fed in this way and filtration effectiveness suffers. Relocate chemical injection points to maximise efficacy.
Many plants with WWRO systems take all of these actions and still have problems with membrane fouling. When traditional filtration systems fail, then ultra-, nano-, or other fine filtration technologies may be required. Plants report good success with these technologies, but it is important to remember that these technologies produce an additional waste stream that must be addressed in the ZLD design.
Regardless of the mitigation measures taken, WWRO users should maintain a complete set of replacement membranes onsite or readily available. Upsets in pretreatment equipment can result in WWRO membrane pluggage in just a few hours. Lacking replacement membranes, many plants have either shutdown or curtailed production because of a badly fouled ZLD WWRO incapable of processing the flow required to sustain plant operation.
WWRO membrane scale formation
WWRO systems receive feedwater from cooling and other systems. In general, this water has already been concentrated nearly to saturation. Left untreated, further concentration in the WWRO reject would result in scale formation. Removal or stabilisation of scale-forming minerals must occur and is usually accomplished in pretreatment systems prior to the WWRO itself. These upstream units must be operated properly to prevent WWRO scale formation.
Cold-lime softening followed by ion exchange often provided hardness (calcium and magnesium) removal. The removal of alkalinity usually requires a degassifier. As discussed earlier, suspended solids removal often requires the addition and tuning of additional filtration equipment.
Monitoring of WWRO performance provides the best opportunity to minimise adverse impacts from upstream treatment system upsets. All users of WWRO systems should monitor normalised data. Spare membranes, mentioned earlier, should also be maintained ready for use. Frequent membrane autopsies help determine the causes and, therefore, corrective actions required to maximize membrane life.
Finally, anti-scalant feed may help minimise scale formation and prolong the time between membrane cleanings. Anti-scalant feed may provide benefits, but it can also impact the performance of thermal equipment used to further concentrate the WWRO reject. Like cooling water dispersant feed, care must be taken to control anti-scalant feed and minimise its impact on downstream equipment.
Thermal systems: problems and solutions
ZLD treatment requires mineral precipitation in thermal equipment. Thermal equipment uses energy to increase the temperature of feedwater. Boiling begins. A portion of the feedwater is recondensed and reused as distillate. The remaining liquor concentrates until minerals begin to precipitate. Calcium sulphate typically forms at lower temperatures, followed by sodium chloride as temperature increases. Some designs drive the precipitation of calcium sulphate and sodium chloride in the same unit (brine crystallisers, for example) while other designs drive the formation of calcium sulphate first (in a brine concentrator or evaporator) followed by the formation of sodium chloride in a crystalliser.
Total solids increase as these precipitates form. Brine concentrators and evaporators typically see total solids concentrations in the 20%-25% range (200000-250000 parts per million). Crystallisers see higher concentrations, typically in the 50% total solids range. The total solids formed consists of total suspended solids (TSS) and total dissolved solids (TDS).
The ratio of TSS to TDS is system specific, but critical to the proper operation of the thermal equipment. If TSS and TDS are out of balance, then rapid corrosion and/or rapid fouling of the thermal equipment can occur.
This ratio is especially critical to prevent calcium sulphate precipitation on brine concentrator walls, for example. Brine concentrators (also called “evaporators”) drive the formation of calcium sulphate. Calcium sulphate preferentially precipitates on other calcium sulphate crystals. TSS increases as this precipitation occurs, providing “seed” for further precipitation. If TSS is too low, then calcium sulphate precipitation occurs on the metal walls of the brine concentrator and rapid fouling can occur. Figure 2 shows such an occurrence. The brine concentrator flood box receives heated liquor and provides distribution into the brine concentrator tube bundle. The photo shows severe fouling of the flood box caused by calcium sulphate precipitation.
In many designs brine concentrator blowdown is simply an automatic valve that bleeds off a portion of the recirculating liquor. This blowdown method removes both TSS and TDS at the same rate. There is no way to easily recover and reuse calcium sulphate “seed”.
A better blowdown design includes hydrocyclones. Hydrocyclones use centrifugal force to preferentially remove dissolved solids while returning suspended solids to the brine concentrator. They help maintain TSS in the brine concentrator, maximising the inventory and surface area of calcium sulphate seed crystals to drive precipitation to the crystal surface rather than to the walls of the brine concentrator tubes.
Hydrocyclones wear over time. They should be checked periodically. They can be retrofitted to existing systems if the TSS/TDS balance is difficult to achieve.
Even a well-controlled TSS/TDS balance cannot prevent all scale formation on the walls of thermal equipment. Regardless of the thermal equipment used (brine concentrators or crystallisers), scale formation will occur and periodic cleaning will be required. Many plants rely on chemical cleaning to remove these accumulated deposits, but chemical cleaning is only effective if a very good mechanical cleaning is performed first. Heavy deposits must be removed to allow good chemical contact with remaining scale.
Figure 3 shows another brine concentrator flood box after mechanical cleaning. Good mechanical cleaning results in what appears to be mostly bare metal with small amounts of deposit remaining. Attention to this important activity results in cleaner surfaces and a longer service interval before the next cleaning.
Scale build-up can cause corrosion as well as loss of heat transfer. Figure 4 shows titanium corrosion in a brine concentrator tube. This particular type of corrosion appears near the upper end of the tube. It is caused by the superconcentration of the circulating liquor in an oxygen-starved environment between the flow distribution nozzle (Figure 5) and the titanium tube itself. Chlorides leach into and embrittle the titanium.
Some plants installed plastic sleeves at the upper end of the tubes. The distribution nozzle is inserted within the plastic sleeve. A void space still exists between the distribution nozzle and the plastic sleeve, but the plastic doesn’t corrode. It’s important that the sleeve fit tightly enough within the upper tube end to prevent any void space between the sleeve and the tube wall. The sleeve prevents superconcentration of liquor near the tube wall.
Share information, learn from othersZLD systems present a host of problems in addition to those mentioned here. Online instrumentation, plugging of lines, chemical feed systems, unusual contaminants, and many other problems can and do occur. Joining a ZLD users’ group can help. There’s plenty of ZLD operating experience, but ZLD users have been slow to share information. Problems faced by one plant have probably been faced and solved by other plants. Share information willingly, in great detail, and often.
Just about everyone hates the ZLD system they have. All ZLD systems have problems. ZLD system vendors endeavour to improve, but every improvement brings new challenges. Problems still exist in new installations. The willingness to listen and share information provides one of the best ways to address ZLD problems. ZLD systems present challenges enough. Learn from the experiences of others, and let others learn from yours.