You can’t track RO performance accurately unless you normalize RO system data. Your flowmeter says permeate is down just 3%. Your membranes may have lost 12%. Cold feed water and extra pump pressure hide the difference, and by the time raw flow drops, a standard cleaning may not fix the problem.
This Daneshline guide shows you how to normalize RO performance data to remove such effects, with the equations, a worked example, cleaning triggers and an Excel template.
Why Normalizing RO Data Matters
Raw operating data moves for reasons that have nothing to do with membrane health. Colder feed water, a throttled feed valve, higher feed salinity and higher recovery all cut permeate flow, and none of them is fouling. The reverse also happens. In warm months an RO can hold the same output while the membranes foul underneath, and in cold months operators can mask fouling by raising feed pressure.
Normalization fixes this by recalculating each day’s data at reference conditions, normally the stabilized performance in the first 48 to 72 hours of operation. ASTM D4516 is the standard practice behind the method, and it covers permeate flow, salt passage and coefficient of performance. The result separates real membrane change from apparent change. That gives you a defensible trigger for CIP and a baseline for any predictive maintenance.
Most plant Human-Machine Interface (HMIs) graph raw data, and those graphs are usually not normalized. LG Chem (a manufacturer of RO membrane elements) goes further and ties its warranty to minimum data logging, so the habit also protects a claim.

What Parameters Should Be Normalized?
Three calculated indicators carry most of the diagnostic weight: normalized permeate flow, normalized salt passage and normalized differential pressure. Feed temperature sits behind all of them. Hydranautics calls an accurate feed temperature reading essential to meaningful normalized data. FilmTec puts the flow effect at about 10% for a 4°C (7.2°F) drop, and a common rule of thumb is 1.5% per °F. A 50 gpm train at 60°F should make roughly 46 gpm after a 5°F drop, with nothing wrong.
Normalized Permeate Flow (NPF)
Normalized permeate flow (NPF) corrects the measured flow for net driving pressure and temperature, relative to your reference. It’s often treated as the most important monitoring parameter. A sustained fall points to fouling or scaling. A rise points the other way, toward membrane damage such as chlorine attack or a failed O-ring.

Normalized Salt Passage (NSP)
Track it as normalized salt passage, the complement of rejection, corrected for permeate flow and temperature. Raw rejection shifts with flux, temperature and feed salinity. When permeate flow drops, less water is left to dilute the salt that passes, so permeate salinity rises even though the membrane hasn’t changed.

Normalized Pressure Drop (NPD)
ΔP is the pressure drop from feed to concentrate. It climbs when feed spacers plug with scale, biofilm or debris, and a plugged spacer also starves downstream elements of driving pressure. Normalize it for feed and concentrate flow, because ΔP rises with flow on its own. Log it per stage where you can. LG Chem’s minimum logging list asks for stage ΔP once a day.

Table 1. Raw versus normalized values (illustrative, calculated with the equations in the next section)
| Parameter | Reference (day 2) | Today, raw | Raw change | Today, normalized | Normalized change |
|---|---|---|---|---|---|
| Feed temperature | 25°C | 20°C | -5°C | Not applicable | Not applicable |
| Feed pressure | 12.0 bar | 14.5 bar | +21% | Not applicable | Not applicable |
| Permeate flow | 100 m³/h | 97 m³/h | -3% | 87.7 m³/h | -12.3% |
| Feed-to-concentrate ΔP | 2.0 bar | 2.2 bar | +10% | 2.30 bar | +15% |
Equations Used in Data Normalization
Every manufacturer’s method follows the logic of ASTM D4516: convert actual data to a fixed set of reference conditions. For permeate flow, the general form published by LG Chem and Hydranautics is:
Qn = Qa × (NDPn / NDPa) × (TCFn / TCFa)
Qa is the measured flow, NDP is net driving pressure and TCF is the temperature correction factor. The subscript n marks the reference condition and a marks today’s actual condition.
Net driving pressure is the feed pressure left after the losses:
NDP = Pf - 0.5 × ΔPfb - Posm - Pp
Pf is feed pressure, ΔPfb is the feed-to-brine pressure drop, Posm is the osmotic pressure at the average feed-brine concentration and Pp is permeate back-pressure. LG Chem also adds the permeate osmotic pressure, which was about 0.02 bar in the example below. The average feed-brine concentration is Cfb = Cf × ln[1/(1 - R)] / R, where R is recovery as a decimal.
The temperature correction factor is:
TCF = exp{ K × [1/298 - 1/(273 + t)] }
Here t is in °C. K depends on the membrane. Published values include 2640 above 25°C and 3020 below it in FilmTec design equations, and 2700 in Hydranautics’ guide. The choice matters in cold water: at 15°C, K = 2640 gives a TCF of 0.735 and K = 3020 gives 0.703, a gap of about 4.5%.
Salt passage and ΔP follow the same pattern. Normalized salt passage is %SPn = (EPFa / EPFn) × (STCFn / STCFa) × %SPa, where EPF is element permeate flow at actual (a) and standard (n) conditions and STCF is the salt transport temperature factor. If the manufacturer gives no STCF, Hydranautics substitutes the TCF. LG Chem normalizes ΔP as ΔPn = ΔPa × [(QFn + QCn) / (QFa + QCa)]^1.5, using feed and concentrate flows.
Worked example. These numbers are illustrative. On day 2 the train runs at 25°C with 100 m³/h permeate, 33.3 m³/h concentrate (75% recovery), 12.0 bar feed, 10.0 bar concentrate, 0.3 bar permeate back-pressure and 1,500 mg/L feed. The feed-brine concentration works out to about 2,770 mg/L, the osmotic pressure to about 2.2 bar and the NDP to 8.51 bar. Later the feed is 20°C, permeate is 97 m³/h and the operator has raised feed pressure to 14.5 bar (concentrate 12.3 bar). NDP is now 10.96 bar and TCF is 0.860.
So Qn = 97 × (8.51 / 10.96) × (1.000 / 0.860) = 87.7 m³/h. Raw flow fell 3%. Normalized flow fell 12.3%, past the 10% cleaning line. The ΔP numbers agree: raw ΔP is up 10%, normalized ΔP is up 15%.
Tools and Templates for Normalizing RO Data
Membrane manufacturers publish normalization programs. Hydranautics’ RODataXL runs in Excel, follows ASTM D4516 and is licensed at no charge. Toray’s TorayTrak is stand-alone software that accepts SCADA (Supervisory Control and Data Acquisition) feeds, manual entry or spreadsheet import. DuPont FilmTec has offered FT-Norm tools for its own elements. WAVE and IMSDesign are design and projection platforms. They help you set a design baseline but aren’t built for daily trending, so their role here is limited.
Excel templates work when you enter data by hand and the inputs are sound. Hydranautics stresses calibrated instruments and accurate reference data over the choice of software. A template needs, at minimum, feed temperature, permeate and concentrate flow, feed, concentrate and permeate pressure, and feed and permeate conductivity. The template offered at the end of this article has daily-log columns for raw and normalized values.
Real-time monitoring is the third route. The catch is that SCADA trends are normally raw, so confirm which equations, reference point and K value a system uses before trusting anything labeled normalized.
Interpreting Normalized Trends to Detect Problems
The rule is simple: clean when any one normalized indicator crosses its line. The lines differ a little between sources, so confirm the numbers with your membrane supplier.
Table 2. Cleaning triggers by source (change from reference)
| Source | Normalized permeate flow | Normalized salt passage or permeate quality | Normalized ΔP |
|---|---|---|---|
| FilmTec cleaning guidance | -10% | +5 to 10% | +10 to 15% |
| Hydranautics RODataXL manual | -10% | Quality decline above 10% | Above +15% |
| Puretec operator guide | -10 to 15% | Not stated as a trigger | +15 to 25% |
| Lenntech troubleshooting guide | -10 to 15% | Permeate conductivity 10 to 15% | 10 to 15% |
The Risk: waiting. One FilmTec case met the flow criterion on day seven, kept running, and was down 45% by day 93. At that point routine cleaning may not restore performance, and extreme cleanings that can damage the membrane may become necessary.
Drop in normalized permeate flow → membrane fouling
A steady fall in NPF with flat ΔP and flat salt passage points to fouling or scaling on the membrane surface. The spacers plug later, so ΔP often lags. If salt passage rises along with the flow loss, think colloidal or metal oxide fouling, or scale. Check the event log before you act. LG Chem asks plants to record every start-up, shutdown and cleaning with its reason, so a step change can be matched to a cause.
Increase in normalized ΔP → scaling or blockage
Rising normalized ΔP means resistance in the feed channel from scale, biofilm or debris. Don’t wait for the absolute limit. FilmTec warns against using 15 psi per element or 50 psi per vessel as the cleaning trigger. A typical normalized vessel ΔP is about 15 psi, so 50 psi means a 333% increase, and that level can push the element scroll out in a telescoping failure that cleaning can’t reverse.
Rejection stability despite real conductivity change → raw data fluctuation
If raw permeate conductivity climbs while normalized salt passage stays flat, the feed changed, not the membrane. A higher feed salinity, higher recovery or warmer feed at constant flow each raise permeate salinity on a healthy element. Cleaning won’t fix that. The opposite pattern is the one to worry about: rising normalized salt passage together with rising normalized flow suggests damage from oxidants such as chlorine, or O-ring leaks.
Common Mistakes in RO Data Analysis
The most expensive mistake is deciding on cleaning from raw data. In the example above, raw flow was down only 3% while normalized flow had already passed the 10% line. A plant waiting for the raw number to move would reach the 30% zone first.
Ignoring temperature and flow swings comes next. Hydranautics notes that normalized data is most accurate when actual conditions stay close to the reference conditions and instruments are calibrated, and it calls feed temperature absolutely essential.
Inconsistent logging breaks the trend. LG Chem’s minimum is flow and pressure every 8 hours, with conductivity, feed temperature and stage ΔP once every 24 hours. FilmTec recommends at least one set of readings a day. Pick an interval and keep it.
Finally, a trend with no context is just a line. Log start-ups, shutdowns, cleanings, chemicals used and the result afterwards, so every dip on the chart has a name. Hydranautics makes the same point from the other side: the value is in the trend, not in any single data point.
Final Notes + Official RO Normalization Software Templates
Use normalized data for every performance decision: when to clean, what to troubleshoot and what to tell a membrane supplier. Raw trends are fine for running the plant, but they can’t judge the membrane. To make this process easier and more accurate, leading membrane manufacturers have introduced their own dedicated normalization software including FT-Norm, RODataXL, and TorayTrak for FilmTec (DuPont), Hydranautics, and for Toray, respectively. Utilizing these official tools or a structured template is the rational way to truly judge the underlying health of your RO membranes.
FAQ
Not necessarily. Cold feed water makes the membrane less permeable, and FilmTec puts the loss at about 10% for a 4°C (7.2°F) drop. A 50 gpm train at 60°F should make roughly 46 gpm after a 5°F fall. Check normalized permeate flow before you call it fouling, because only that number removes the temperature effect.
Clean when any one indicator crosses its line. FilmTec’s criteria are a 10% drop in normalized permeate flow, a 5 to 10% rise in normalized salt passage or a 10 to 15% rise in normalized pressure drop, and Hydranautics uses 10%, 10% and 15%. Many suppliers quote a flat 10 to 15% for all three. Waiting until flow is down 30% makes a standard cleaning much less likely to work.
It restores flow for a while, but it doesn’t remove the cause. Lenntech warns that extra pressure can speed up fouling, scaling and compaction, so the loss returns faster. It also hides the problem in raw data, which is why normalized flow keeps falling while the flowmeter looks fine. Find and fix the cause instead, such as fouling, scaling or weak pretreatment.
LG Chem’s minimum is flow and pressure every 8 hours, plus conductivity, feed temperature and stage ΔP every 24 hours. FilmTec’s guidance calls for at least one full set of readings per day. Keep the intervals consistent so trend points stay comparable.




