| 
|
Caring for your treatment membranes raises uptime reliability
By David Engle
From beverage plants to boilers, sites that polish, consume, and discharge water are loath to take treatment streams offline: Whenever this happens assets and services are running “ down the drain.”
Even scheduled cleaning of critical reverse osmosis (RO) membranes “can get pushed off”at sites straining for high production, notes Tom Fredrickson, filtration group manager of Ecolab of St. Paul, MN, a $5 billion commercial cleaning and sanitizing firm.
There’s value indeed in prolonging maintenance intervals. However, when a shutdown for cleaning can’t be postponed any longer, there’s even more value gained by cleaning (or replacing) for optimized performance and extended service life.
Conversely, there’s money to be lost if RO elements are neglected to the point of impaired performance or failure. As Frederickson observes, “You can run your water system without ever cleaning it, but eventually—and usually at an inopportune time—it’s going to go down; and then you can’t produce.”
In such cases, he adds, you may also have to bear the expense of replacing membranes ruined by neglect. He’s speaking of RO (as are the other experts cited below), but similar cleaning and care principles apply to filtration elements along the entire treatment stream.
Much Can Go Wrong
In the case of RO, multiple issues can crop-up interactively to impair performance. For instance, it’s quite common for system operators to misread or misunderstand the RO performance data that (presumably) they’re monitoring, regarding flow pressures or chemistry. Consequently, not only does the cause of the eventual failure go undetected, but also the solution to fix it to prevent a recurrence remains unclear.
Example: On the day of Fredrickson’s conversation with Onsite Water Treatment, Ecolab happened to be servicing a trouble-call from a beverage plant; investigation showed the root cause to be “a drastic decrease in flow rate that no one was monitoring,” he recalls. Plant workers “were taking the data—but it was staying on a piece of paper. No one looked at it over time.” And the system simply crashed.
Any of several factors may underlie membrane failure, factors which may or may not be related to cleaning. David Scheeter, an RO technician for a Hayward, CA, General Electric (GE) distributor, cites, as an example, the placing of filter elements “in the wrong position” in relation to RO membranes; this, he says, can be a surprisingly frequent cause of fouling and impaired performance in the industrial systems he sees.
Similarly on this erroneous-design theme, Tom Chiara of GE’s Microelectronics Center of Excellence in Phoenix notes that flawed legacy RO engineering throws curves at many plant managers. Some older systems were laid out in cookie-cutter fashion by designers “who are more mechanics than they are chemists or chemical engineers … and are [lacking] a true understanding of what’s going on.” Hence, recurring failure of membranes was sort of built in. Design-related consequences are to be found “ over and over again,” he says. GE, he notes, owns and operates the world’s largest fleet of mobile RO equipment and also operates and maintains customer-owned RO facilities.
As for membranes themselves, directly impinging hazards on these range from ordinary fouling due to neglect to misapplied chemistry and excessive washing, notes Randall Jones, chief executive officer of Wastewater Resources Inc. of Scottsdale, AZ. Wastewater Resources Inc. is an original equipment manufacturer (OEM) water and wastewater skid manufacturer.
Complicating both the care and cleaning and the root-cause failure analysis, he notes, is the fact that “there are probably upwards of 200 firms that make and ship RO membrane systems on a regular basis”—and many more, for occasional orders. And these membrane-makers draw upon scores if not hundreds of materials in their composite membrane designs. Add to this universe the profusion of cleaning-chemical constituents. These—ideally and properly—often need to be custom-mixed and matched to the membrane materials and site condition.
All told, then, hundreds of water-treatment OEMs, component assemblers, and chemical vendors are out there, “all doing things a bit differently,” Jones says. Obviously, there’s ample room for errors, onsite or offsite.
Adding to these challenges, many RO membrane manufacturers—contrary to what you might expect, says Jones—“are loath to tell you what to do” regarding proper care and cleaning. Why so? Because failure, ensuing from following manufacturers’ instruction, “makes it their fault,” he says, adding: “They’ll spend a lot of time telling you what not to do—but very little time on what you should do.”
Due to liability fears, they are far more likely to refer you to outside consultants, he finds. On that score, these, too, may lack much practical or extensive field experience to qualify them to advise you properly.
“So,” he says, “it’s an interesting dance that you’ll do.”
Additionally, manufacturers and system designers “should do extensive water tests on the water that you’re going to use the RO on, before the systems are ever designed,” advises Jones, “so that you know exactly what’s in the water … It’s extremely important that the water be analyzed by the membrane manufacturer to see how high a concentration you can go to without damaging the membranes, and what chemicals, if any, need to be added, [and] what pH adjustments need to be made, in order to run the thing optimally.
“That’s what most people miss,” Jones sums up. “And that’s why they ruin filters.”
Autopsy Reports
One or several causes may lead to anxious help calls to chemical vendors or RO consultants. Corrective efforts require root-cause troubleshooting, followed by appropriate recommendation: e.g., perhaps changes in cleaning chemistry and/or frequency; adjusting pH and/or temperatures; altering operations; better monitoring practices; re-training of maintenance staff; and even significant re-engineering of systems, up to, perhaps, even building an entirely new one.
The first step, as Fredrickson has noted, might consist of an inspection of data logs, along with conducting staff interviews to assess possible human error. Following this, dissection of the underperforming membrane might be needed.
“Basically,” he says, “we open up an element to see what’s going on in there: look at the membrane, look at the surface, and determine what the foulant is.”
Next, perhaps, will come some chemical lab work, “to try to determine a better cleaning approach, at more extreme pH, or high temperatures, to see results … [without] using the plant itself as a ‘guinea pig,’” says Fredrickson.
In the above-noted crisis (which was caused by plugged membrane pores), Ecolab chemical baths succeeded in returning the RO permeate flow rates “back to what were specified,” he says. Then, to prevent a recurrence, changes in pre-treatment regimens (i.e., adding an anti-scalant and changing the injection point) were recommended.
San Marcos, CA–based Avista Technologies’ primary business is selling chemicals for RO systems worldwide, says President David Walker. In that role Avista Technologies, too, is often called-upon to troubleshoot, perform product “autopsies,” and devise better cleaning regimens. Operational data first are gathered on flows, pressures, rejection rates, water supply, and the pre-treatment scheme, leading, he says, to “the right chemistry to clean with and recommend.” In tougher cases, lab-based membrane and water analysis may be called for, and again, cutting membranes open and performing tests with a battery of potentially dozens of anti-fouling or scale-removal chemicals, to determine optimal choices. Walker notes that perhaps 15% or 20% of inbound trouble calls result in needing this kind of customized assessment.
Keeping Closer Watch
Obviously, far better than any post-mortem is preventive medicine i.e., cleaning and operational changes for optimized performance and longevity. The key here is condition monitoring done competently.
In fact, staff failures in keeping adequate tabs on membranes turns out to be not at all infrequent and were mentioned by nearly every cited commentator. RO systems often include automated data logging of permeate flow rates and differential pressures, etc., logged over time. These indices, says Fredrickson, “will help you get to a point where you decide, ‘I need to clean.’” Scheduling this at the right moment not only prevents damage or degraded performance but also avoids over-cleaning, which also wastes operational time and often wears out membranes prematurely.
Monitoring in the larger sense also includes keeping track of the upstream water composition and pre-treatment filter conditions. Changes in either will often impact membranes dramatically, as will be amplified below.
More specifically, good monitoring means:
- Actually scrutinizing numbers conscientiously (which, surprisingly, isn’t always done). As Fredrickson points out, in a busy plant, numbers change day-to-day in only tiny increments. They’re easy to ignore, until “It creeps up on you,” and a shutdown hits. In addition, Fredrickson notes that often operators run the gamut in attitude and analytical skills.
- Surprisingly, the next challenge involves normalizing the data and interpreting it correctly, notes Chuck McCloskey of Siemens Water Technologies. For example, he says, water constituency and quality can be quite significantly altered “if the source [say, a well] changes; heavy rainfall and surface runoff occurs, or you have drought conditions.” In the latter case, “very high TDS [total dissolved solids] levels result from reduced water in the watershed.” Salts begin to concentrate. “Something we see seasonally in coastal areas—we get saltwater intrusion,” he says, causing a “tremendous change in the water quality.”
Thus, he says, “ Water quality should be checked regularly for TSS [total suspended solids] and SDI [silt-density index]. If there’s known changes to the watershed or the aquifer or indications that things have changed by this general test, then more specific tests should be performed.” McCloskey is Siemens’s product manager of mobile and onsite services.
- Another challenge involves determining what to do with the results of all that monitoring. In the case Fredrickson discusses above, system operators were recording raw numbers—but not understanding or doing anything with them. In fact, he says only half-jokingly, it’s fairly common that the RO operator, upon hearing a system shutdown alarm and being the first to arrive in the room “pushes the reset button—then high-tails it out of the room before it starts again.”
| |
 |
|
Making Sure Things Are "Normal" |
Better Monitoring: Two Solutions
To improve matters, a number of service firms (including Ecolab) have offered to provide customers Web-based remote monitoring, giving them another pair of expert eyes. When imbalances or problems arise, systems can dial managers’ and key staffs’ cell phones to alert them.
A second key element: up-to-date operator training. This may sound like an old mantra, but it’s still worth noting in relation to membrane and filter care when plant-production issues are at stake. If training prevents a membrane failure and costly shutdown or lengthens maintenance intervals, the modest investment is easily worth it. And carefully evaluate the competency of training firms, advises Walker.
Fredrickson suggests that appropriate employee instruction might typically consist of two to four hours in a classroom, followed by another hour or two of a walking tour of plant elements. Strive to have “at least one and preferably more [staff] who are knowledgeable,” he adds, “and you will have a much higher likelihood of a successful operation where you’re not fighting fires continuously.
“Filtration system quality control is very operator-driven.”
RO Maintenance, Professional Edition
With that broad backdrop, following next is some collected wisdom on membrane cleaning itself. GE’s Chiara begins by noting that the alternating of fouling and cleaning in the membrane’s work cycle yields “a saw-tooth performance curve.” Brand-new units start out with optimum performance: low water pressure, high flow, and high rejection of the minerals. “Everything looks really good,” he says. Then, as time goes by, fouling occurs, performance degrades, “and the curve starts slanting downward.”
A routine onsite cleaning makes this “shoot straight back up again—but not quite as high as it was the first time.” The cycle repeats, and each time, some measure of performance is lost. Ultimately, membranes reach the point where they’re either candidates for replacement (at, say, $800 each) or—in a stab at returning them as closely as possible to original shape—being removed and shipped somewhere for specialized, thorough restoration.
Typical onsite cleaning regimens consists, explains Sheeter, of “a high pH cleaner to remove the organics and then a low-pH cleaner to remove the calcium and other minerals that might have migrated into the media. Then you flush.” If done correctly, performance can be restored to close to 90% of its original capability.
Neglect of cleaning can easily allow such a buildup of foulants that not only does performance crash, but “it can destroy them,” says Sheeter, who spent 15 years in commercial and industrial RO plant applications in San Jose.
What about cleaning frequency?
Despite the fact that all RO works the same everywhere, “the real-world answer,” several commentators agree, is that the cleaning interval can be as brief as a few weeks, up to “never.”
“It’s all dictated,” explains McCloskey, “by how well the [membrane] is rejecting and removing contaminants from the water and allowing good throughput.”
A rule of thumb, suggests Fredrickson, would be [assuming typical city water] “quarterly; but it really varies,” he adds. “Quarterly may be not be frequent enough in some cases or is too frequent in others.”
Too frequent or too harsh cleaning can easily age sensitive materials like cellulose acetate, while also reducing their effectiveness in solids rejection. Yet, as Fredrickson reports, one customer who suffered a major problem (using a much more durable thin-film composite material) resorted to cleaning them with an aggressive low- then high-pH cleaner, and a sanitizer every four days—“which is fairly extreme,” he notes—and yet the thin-film composite in question “did not show loss of rejection for six months.
“Nonetheless,” he goes on, “you generally don’t want to take a system down or expose it to extreme pHs any more frequently than you have to.”
Again, conscientious monitoring will enable an operator to initiate cleaning when the scaling is still on the casing rather than the membrane so that the cleaner eats at this instead.
Several commentators suggest that a well-engineered RO system “with a good water source and pre-filter could last indefinitely” on as little as “a two-year wash” cycle, says Jones. This is the positive experience reported at the City of Scottsdale’s advanced wastewater treatment plant, which produces 14 million gallons of RO and microfilter pre-treated water daily, yet with only rare semi-annual cleaning. It’s an exceptionally successful water plant, adds Jones.
And what’s the most effective cleaner?
Actually, as noted above, the specific nature of the fouling and the membrane material will determine the correct cleaning chemistry mix. Jones has found, though, as a general rule, that the best cleaner on organically fouled thin-film composite membrane has been trisodium phosphate and Versene 220 from Dow—not pre-mixed but diluted with equal parts of water. The right chemistry—neither too weak nor too caustic—is critical, and so too, he adds, “The pH range alone that the membrane runs into is at least as important as the chemistry that’s used.
“Chemistry, pH, and application technique all have to be taken into account,” he sums up, “and if you don’t, then you’re going to have trouble.”
Clean-in-Place Versus Outsourcing?
Membranes can indeed be either washed where they stand or transported to a specialized cleaning center for a first-class job. Which of the two choices is correct will depend, naturally, on several variables. The clean-in-place (CIP) procedure itself is easy, as Scheeter explains. “In most cases it’s just changing the direction of a three-way valve, hooking up your CIP system, and cleaning the units right there in situ.” The cleaning solution, perhaps being warmed, re-circulates through the RO housings, thus removing foulants and impurities.
CIP’s main virtue, Scheeter adds, is that “it will more likely result in systems being cleaned and properly maintained.”
Other benefits are low cost, convenience, and quick results. As for the latter, Ecolab’s Fredrickson believes that “if you have a proper clean-in-place system and the ability to generate enough cross-flow velocity, enough contact time to get the chemicals on top of the membrane, and you’re doing it within the guidelines of the cleaning parameters, you’re going to have a very successful cleaning.”
Ecolab does provide offsite cleaning, he adds, but clients are typically smaller plants with membranes requiring more customized care.
Jones (whose firm, Wastewater Resources Inc., manufacturers CIP-only skidded treatment systems) regards self-cleaning as indispensable; an RO-lacking CIP will be “difficult to take care of.” Removing and transporting the membranes somewhere else for cleaning is costly and “unrealistic,” he adds. If membranes are so badly fouled that CIP can’t bring them back to life, the money needed for offsite cleaning would be better invested in buying a new set.
Taking a more favorable view of offsite cleaning, McCloskey points out that a cleaning center’s specialized service saves the operator from having to select the often custom-blended chemicals or contend with associated Hazmat handling and disposal and other safety issues. Offsite cleaning also allows a customized membrane foulant analysis and cleaning application. The cleaning itself can be done in a controlled setting that is optimized for pressure, temperature, and velocity. Newly cleaned membranes can then be tested to confirm restored “almost new” flow and rejection capacities.
To avail yourself of the service, advanced scheduling and temporary installation of a replacement set of elements will be necessary, notes McCloskey, whose division runs six cleaning centers in North America. Normal turn-around, he says, will be 10 to 14 days, and cost-wise, the service is “very comparable” to CIP when actual expenses are totaled.
GE’s Tom Chiara sums up: The real solution is not either/or but, rather, “You have to have both” —because an occasional thorough or customized, outsourced cleaning will be justified to dissolve scaling on an aging membrane and restore them to near-new quality. “You certainly do a whole lot better job cleaning it externally than internally,” he says, but this comes at a premium.
What’s Going on Upstream?
Lastly, as already noted by several experts, the importance of correct water filtering and treatment before it reaches the RO cannot be over-stressed. McCloskey explains that proper understanding of raw water quality and preparing it for an RO system typically entails removing colloidal or particulate matter. Filters strain to reach an SDI, which “should be less than five and preferably less than three.”
Multimedia filtration with sand filters, softeners, and cartridge filters are used in various pretreatment schemes to reduce that SDI, he adds. Mineral scaling from calcium carbonate, aluminum, iron, and manganese is another danger for which pre-treatment is sometimes needed. “Remove those foulants before they reach the membrane surface,” he advises. Anti-scalant chemicals should take care of any remaining minerals, which might form. Next, oxidizing agents like chlorine, bromine, chlorine dioxide, and ozone, “ need to be removed before the water reaches the membrane,” he says, using a reducing agent like bisulphate or through carbon absorption. He sums up: “The knowledge base is well established, but if all the details are not paid attention to, there’s potential for problems.”
GE’s Chiara adds that attention should also be paid to where in the system the chlorine is removed. Biological growth is a primary source of fouling in otherwise well-designed systems. On this point he observes that, in an upstream carbon filter, “within 3 inches of the top of that carbon bed there’s no more chlorine in the water. So, from that point all the way through to the RO membrane, you’re not protected from … biological growth.
“Certainly,” he adds, “carbon is a big cause of a lot of these fouling issues,” and so Chiara recommends removal of carbon filters on some systems.
What about monitoring those various filters?
Jim Dobos of Aquatech International offers the tip that, in higher-end systems, once an operational baseline has been established, you can monitor pressure changes on any filter e.g., multimedia or cartridges. “The time to take your elements out and change them is when the pressure differential increases by about 5 psi [pounds per square inch].” But, he adds: “ Nearly nobody does this. Everybody runs it up to about a 15 psi differential.” He cautions, “ They’re not really realizing that, as that differential increases … the particulates work themselves deeper and deeper into the media” until they’re so deep that conventional cleaning won’t remove them.
In the case of cartridge filters under these high-pressure differentials, particles will eventually “slough off the cartridge … and flood into the membranes,” says Dobos, who is director of water management services at Aquatech. Most operators err in running filtration systems “much further than they should … [before doing] some sort of cleaning and some sort of backwash” —indicative, he suggests, of widespread misunderstanding “of how filtration in and of itself works.”
Moving Forward
Over the water-treatment horizon, McCloskey sees “a continuing move toward RO away from ion exchange,” along with usage in more treatment environments. The two will yield continued market growth. Industrial, commercial, and institutional markets for Siemens Water Technologies are all on the upswing, he adds.
Avista Technologies’s Walker concurs that, especially globally, RO usage is on a strong, sustained-growth course. His own company’s sales (which, again, focus on RO-related cleaning chemicals) have increased at an average 12% annual rate for a number of years, he says. Worldwide, RO membranes number about a million, according to industry figures he’s studied.
Chiara’s division at GE develops next-generation high-efficiency RO (HERO) systems capable of much higher water recovery—while needing virtually no membrane cleaning. In some applications, he says, the comparison between the earlier generation and the new “is like comparing a backhoe to a hand-shovel.” HERO, he says, is “the pinnacle of RO technology … Almost every system GE builds for the microelectronics industry now uses it,” he adds, while candidly acknowledging his bias. Nevertheless, GE remains hugely active in RO worldwide too.
Scheeter, from his vantage in the field, observes: “RO is a wonderful invention. It’s been around a long time. But now, with the advent of membranes that operate at about 100 psi, we’re finding that you can produce a great volume of water” at 100 psi instead of 200 psi to 250 psi.
“So, obviously,” he sums up, “there’s a lot of savings—and less wear-and-tear on your system too.”
David Engle is a writer based in La Mesa, CA, specializes in construction-related topics.
OW- September/October 2007
|
