Setting up GE Water Process Technologies membrane systems without losing your mind
Most people buy into the GE Water Process Technologies brand because the datasheets look clean and the performance guarantees sound reasonable. The reality on the floor is different. I spent three weeks debugging a RO skid that kept tripping on differential pressure alarms. The membrane elements were fine. The problem was a misconfigured conductivity cell that was feeding wrong data into the PLC, which then made the PLC reduce feed pressure, which dropped permeate flow, which made the operators increase pH dosing out of panic, which caused calcium carbonate scaling on the second stage anyway. You do not solve that kind of problem by reading the manual cover to cover. Here is what actually matters when you are integrating these systems into an existing plant. The feed water pretreatment is where 80 percent of failures happen, and it is also where people cut corners. Ge Water Process Technologies recommends 5 micron filtration before the RO train. That is the minimum. If your feed water has high silt density or organic load, you are looking at cartridge filters rated at 1 to 3 microns with automatic backflush capability. I had a site in Texas where the pretreatment was undersized by design because the original engineer thought the SDI would stay below 3. It spiked to 8 every Tuesday when the upstream clarifier did its scheduled dump cycle. The membranes were replaced twice in eight months. The fix was installing a dual-media filter with controlled backwash timing that matched the clarifier cycle, not just running on a timer that someone set two years ago and never updated. For the high-pressure pump selection, do not size for maximum permeate flow at design conditions. Size for the worst-case feed water temperature, which is usually the coldest month in your location. A seawater RO system running at 4 degrees Celsius feed temperature needs roughly 15 to 20 percent more pressure than the same system at 25 degrees to achieve the same flux. If you undersize the pump, you will spend your entire operating year throttling down and never reaching design output. If you oversize it heavily, you will either need a VFD with good low-end torque control or you will be bleeding off pressure through a recycle valve, which wastes energy and causes thermal stress on the membrane elements.
The chemical dosing system deserves more attention than it gets. Antiscalant selection is not one size fits all. The Ge Water Process Technologies antiscalant lineup includes several formulations, and picking the wrong one for your water chemistry will cause scaling problems that look like membrane failure but are really just dosing chemistry mismatch. I had a brackish water system where the feed had high silica and boron levels. The standard antiscalant package handled calcium carbonate and gypsum fine, but the silica stayed in solution until it precipitated on the last few elements of the second stage. Switching to a silica-stabilizing antiscalant formulation solved it immediately. No membrane replacement needed. The first element differential pressure stayed flat for six months after the change instead of climbing steadily.
Operational workflows that actually work
Daily operations on a GE Water Process Technologies installation follow a pretty standard rhythm. You check feed and permeate conductivity, monitor differential pressures across each vessel, record flow rates, and log pH and chlorine levels. The chlorine monitoring point is critical. Free chlorine above 0.1 ppm will degrade standard polyamide membranes within hours. Most systems have a sodium bisulfite dosing point after the cartridge filter specifically for this. The bisulfite dose needs to be high enough to scavenge all residual chlorine but not so high that it drops the pH unpredictably or introduces unnecessary sulfate load into the permeate. CIP procedures are where operators get sloppy. A proper cleaning cycle for TFC membranes involves separate caustic and acid phases. The caustic phase at pH 11 to 12 with sodium dodecyl sulfate surfactant removes organic fouling and biofilm. The acid phase at pH 2 to 3 with citric acid or hydrochloric acid removes inorganic scale. Do not mix the two chemicals. Do not try to shortcut by running a single cleaning solution. The contact time matters too. Each phase needs a minimum of 45 to 60 minutes of circulation at the recommended flow rate, usually 60 percent of normal feed flow per element. After the soak, you flush thoroughly with product water before returning to service. I once saw a plant run a CIP cycle in 20 minutes because the operator thought faster was better. The membranes came out looking fine on inspection but had reduced salt rejection by 3 percent that they never recovered from. That 3 percent loss compounded over the next two years into thousands of dollars in wasted energy and premature element replacement. Element replacement sequencing is another thing people get wrong. When you replace failed elements, do not install them randomly in the vessels. Put the healthiest elements in the first positions where the feed water is most contaminated, and place new elements toward the permeate end. The fouling gradient means the first elements take the hardest hit. If you put new elements in position one and old tired ones in position six, the old ones will fail faster and you will waste the performance of the new ones. Rotate elements between vessels during replacement cycles if you can. It extends overall life by roughly 10 to 15 percent compared to random placement.
Get the Full Details
Things the documentation does not tell you
The troubleshooting guides in the manuals are useful for obvious problems. They are not useful for the subtle ones. Here are a few that I have encountered and that you will not find in the quick-start section. If your permeate flow is declining but differential pressure is also dropping, you likely have a membrane compaction issue or a seal leak, not fouling. Fresh polyamide membranes compact over the first 24 to 48 hours of operation, which causes a natural decline in permeate flow as the membrane structure settles. This is normal. But if it happens repeatedly every few months, check your high-pressure pump seals and the vessel O-rings. I found a cracked O-ring on a second-stage vessel that was letting permeate recirculate back to the feed side. The symptom looked exactly like progressive fouling. The fix was a $4 part and ten minutes of labor. The previous operator had replaced three sets of elements trying to fix a seal problem. Conductivity drift in the permeate stream that correlates with temperature changes is often a sensor issue, not a membrane issue. Temperature compensation algorithms in the PLC can introduce errors if the compensation factor does not match the actual membrane characteristics. Verify your conductivity readings with a handheld meter at least once per shift. If the inline meter and the handheld disagree by more than 5 percent, calibrate or replace the inline sensor. Do not trust the PLC display blindly.
Another counter-intuitive point: higher recovery rates do not always mean better economics. Running at 75 to 80 percent recovery on seawater RO sounds efficient until you calculate the energy cost of the higher operating pressures required and the shortened membrane life from increased scaling tendency. A 65 percent recovery rate with lower pressure and longer element life often has a lower total cost of ownership over three years. I ran the numbers on a project where the client wanted maximum recovery. After factoring in chemical costs for antiscalant at high concentration factors and the projected element replacement schedule, the optimal recovery point was 68 percent, not the 80 percent they were designed for. The sensors and instrumentation from GE Water Process Technologies are generally reliable but not immune to drift. The ORP probes used for chlorine monitoring need weekly calibration with fresh standards. The pH probes drift faster than people expect, especially in low-conductivity permeate water. Replace them every six months regardless of what the calibration curve says. I have seen pH readings off by 0.5 units because someone skipped the replacement schedule and the electrode had degraded past the point of accurate measurement.
When Ge Water Process Technologies systems are not the right choice
These systems work well for desalination, ultrapure water prep, and industrial wastewater reuse. They are not ideal for situations with extremely variable feed water quality where the fluctuations are faster than the control system can respond. I worked on a plant that took water from a river with seasonal turbidity swings from 5 NTU to over 500 NTU. The automated pretreatment could not keep up. The operators ended up manually adjusting coagulant dosing throughout the day, which introduced human error and inconsistency. In that case, a conventional sand filtration train with manual bypass capability would have been more robust than the packaged solution. The GE system was not bad. It was just designed for more stable feed conditions than this site had. For small-scale applications under 10 cubic meters per day, the capital cost of a full Ge Water Process Technologies skid is hard to justify. Modular containerized systems from other manufacturers or even simple gravity-fed RO units with manual control can handle those flows more economically. The automation and remote monitoring features that come with the GE systems are valuable at larger scales but add cost and complexity that smaller plants do not need. If your feed water has high concentrations of hydrogen sulfide or iron, you will need aggressive oxidation and filtration pretreatment before the RO stage. The membrane itself cannot handle these contaminants. Some sites try to run the RO first and use it as a polishing step after biological treatment. That approach fails because the membrane fouls within days. Treat the water properly upstream or use a different separation technology like nanofiltration for the initial pass and reserve the RO for final polishing only.
Practical checklist before you commission
Before you take a new Ge Water Process Technologies installation online, run through these steps. They take about two hours and will prevent most of the problems that show up in the first month of operation. Verify all instrument calibrations against portable reference equipment. Check pH, conductivity, ORP, and pressure transmitters. Document the baseline readings for each point. Take a permeate water sample and send it to a lab for full ion analysis within the first 48 hours of operation. Compare those results to the membrane manufacturer's salt rejection specifications. If rejection is worse than spec, do not assume the membrane is defective. Check the feed water analysis. The most common cause of poor rejection is incorrect feed water chemistry, not membrane damage. Run the system at 50 percent recovery for the first 24 hours before pushing to design recovery. This allows the membranes to fully hydrate and compact gradually. Then incrementally increase recovery in 5 percent steps every 12 hours while monitoring differential pressure and permeate quality. If anything looks wrong at any step, hold there and investigate. Do not push through to maximum recovery hoping the problems will resolve themselves. They will not.
Keep a running log of daily operating parameters. The log does not need to be elaborate. Feed pressure, permeate flow, concentrate flow, conductivity on both sides, pH, temperature, and chlorine residual at the feed point. Once you have three months of data, you can build a performance baseline. Any deviation from that baseline is your early warning system for fouling, scaling, or membrane degradation. Most plants I have worked at do not keep logs consistently. Without the log, you are reacting to problems instead of predicting them. The systems are capable of long service life when operated correctly. The difference between five years and twelve years of membrane life usually comes down to pretreatment quality, chemical dosing accuracy, and how quickly operators respond to early warning signs. Spend time getting those three things right before you worry about pushing for higher flow rates or lower operating costs. The rest follows from there.
