WaternaCall us+44 7450 201927

RO data normalization explained: what it is and why it matters

Why raw reverse osmosis data misleads, how normalization (ASTM D4516) corrects permeate flow and salt passage for temperature, pressure and salinity, and how to use it.

Mohamed Oulabi, Principal Process Engineer, Waterna6 min read

Ask an operator how their reverse osmosis plant is performing and the answer is usually the permeate flow on the screen. The trouble is that permeate flow changes every day for reasons that have nothing to do with the membranes. Cold water, saltier feed or a lower feed pressure all reduce it. A warm spell can hide weeks of fouling.

Normalization removes those effects, so that what is left is the condition of the membranes themselves. It is the single most useful thing a plant can do with its RO data, and it is the basis of every serious performance assessment.

What changes raw RO performance

Three operating conditions move permeate flow and salt passage independently of fouling or damage:

  • Temperature. Water viscosity falls as it warms, so membranes pass more water. As a rule of thumb, permeate flow changes by roughly 3% per °C. Salt passage rises with temperature too.
  • Net driving pressure (NDP). Permeate flow is roughly proportional to the feed pressure minus the osmotic pressure, permeate pressure and half the pressure drop. More pressure, more water.
  • Feed concentration. Saltier feed raises osmotic pressure, which reduces the net driving pressure and so the flow, and it changes salt passage.

Recovery matters as well, because it sets the average concentration along the membranes.

What normalization does

Normalization recalculates today’s performance as if the plant were running at a fixed set of reference conditions, usually those at start-up or after the last membrane replacement. The method most plants use is set out in ASTM D4516, Standard Practice for Standardizing Reverse Osmosis Performance Data.

In outline:

  1. Normalised permeate flow = measured permeate flow × (reference NDP ÷ actual NDP) × (temperature correction factor at reference ÷ at actual).
  2. Normalised salt passage = measured salt passage, adjusted for the change in average feed-concentrate concentration and temperature, relative to reference.
  3. ΔP is usually trended directly, ideally corrected to a reference flow, because it depends strongly on how much water passes through the feed channels.

The temperature correction factor comes from the membrane manufacturer, as it differs between membrane types. Using a generic one is acceptable for trending but not for warranty discussions.

What good normalised data looks like

On a healthy plant, normalised permeate flow and salt passage are almost flat lines, with small steps after each clean. Real problems then stand out:

  • A steady downward slope in normalised flow — gradual fouling.
  • A step — an event: a pretreatment upset, a dosing failure, a change of feed source.
  • Rising salt passage with falling flow in the last stage — likely scaling.
  • Rising salt passage with rising flow — possible membrane damage, an O-ring or interconnector leak, or oxidation by chlorine.

See Membrane fouling in RO for how to read these patterns stage by stage.

Common mistakes

  • Normalising the whole train only. Fouling and scaling happen in different stages. Without stage pressures and flows, the diagnosis is a guess.
  • Poor instruments. Normalization amplifies errors in flow, conductivity and pressure. A drifting conductivity meter produces a convincing — and false — trend. Calibrate the instruments that feed the calculation.
  • Changing the reference. Resetting the baseline after every clean hides the permanent loss that builds up over the years.
  • Too little data. One reading a day, taken at different times, adds noise. Hourly averages from the control system are far better.

From spreadsheet to software

Many plants normalise in a spreadsheet, at least for a while. It works, but it depends on someone keeping it up to date, and it rarely tracks each stage. RO Insight™ normalises the data automatically from the plant’s historian or exports, trends every stage, and highlights fouling, scaling and instrument faults as they develop.

If you would rather have an engineer review your data first, a membrane performance assessment normalises your history and tells you what it shows.