Reverse osmosis desalination uses selective membranes to separate lower-salinity water from a concentrated feed. The membrane matters, but quality, recovery, electrical energy and downstream conditions all determine whether the total supply is useful.
The gold standard links an auditable water-and-salt balance to normalized membrane performance and independent water-quality verification. The calculations teach what is measured—and what cannot be inferred from one appealing number.
eduKateSG · Water Desalination Membrane Performance
Find your next learning step
Start with flow and chemistry, calculate recovery and flux, and check the evidence for deterioration and useful output.
Full chapter index · Practice and transfer · Education root · How X Works
Full chapter index
Define desalination output · Chapters 1–2
Calculate membrane performance · Chapters 3–4
Diagnose membrane change · Chapters 5–7
Verify whole-plant claims · Chapters 8–10
Follow connected sources · Chapters 11–12
CHAPTER 1 OF 12 · Define desalination output
1. A reverse-osmosis membrane is one part of the water supply
Seawater desalination uses a pressure-driven membrane process to produce a lower-salinity permeate stream and a concentrated residual stream. The technique can be effective, but a good membrane-separation result is not the same as proven safe drinking water at the consumer.
A complete seawater treatment plant may include intake screening, clarification, ultrafiltration, high-pressure reverse osmosis, energy recovery and post-treatment. Singapore’s PUB documents actual facilities using several of these stages because source-water solids, microorganisms and dissolved salts require different barriers.
Begin with a diagram showing feed, permeate and concentrate streams, the pressure boundary and where each sample is taken. A membrane element, a pressure vessel, an RO train and an entire plant have distinct measurement boundaries.
The gold standard reports feedwater source and chemistry, membrane area, pressure, temperature, product quality, recovery, electrical load, time period and uncertainty. A favorable number from one stage must not be presented as performance of all stages.
This is an educational explanation of measurement and modelling; it does not provide safe operating pressures, membrane-cleaning instructions or a certification of product water for consumption.
CHAPTER 2 OF 12 · Define desalination output
2. Flux, recovery, salt passage and rejection answer different questions
Recovery is the fraction of incoming water that becomes permeate. Flux describes produced water per active membrane area per unit time. Salt passage and rejection compare solute concentrations between feed and permeate, under clearly defined sampling conditions.
An RO train could produce more water after a pressure increase while simultaneously using more electricity or facing stronger scaling risk. Thus a higher flux is not automatically an efficiency or sustainability improvement.
Likewise, a high rejection percentage does not establish that every possible contaminant has been controlled. Water chemistry, microorganism removal and post-treatment must be assessed against applicable specifications.
Recovery and membrane rejection interact with concentrate management. Higher recovery can increase the dissolved-solid concentration in remaining water, potentially intensifying scaling and limiting the stable operating envelope.
A useful performance table labels each variable’s units and instrument. Conductivity may help estimate ionic content, but a conductivity reading is not a complete chemical assay and depends on composition and temperature.
CHAPTER 3 OF 12 · Calculate membrane performance
3. Worked example: a simple three-stream water balance
Assume a fictional steady feedwater flow Qf = 100 cubic metres per hour and permeate Qp = 45 cubic metres per hour, with no other water inflow or leakage across the defined train boundary.
Conservation gives concentrate flow Qc = Qf − Qp = 100 − 45 = 55 cubic metres per hour. Recovery is Y = Qp/Qf = 45/100 = 0.45, or 45%.
Across one hour, the 100 m³ feed equals 45 m³ permeate plus 55 m³ concentrate. This verifies the arithmetic of the model; it does not independently establish that the flow meters were calibrated.
If a second day has 80 m³/h feed and 40 m³/h permeate, product flow is lower but recovery rises to 50%. The distinction demonstrates why production and recovery should never be substituted for one another.
These values are invented. They do not identify an appropriate recovery target for any feedwater, membrane or plant. Real balances account for sampling times, flushes, bypass lines, storage changes and instrument error.
CHAPTER 4 OF 12 · Calculate membrane performance
4. Worked example: salt rejection and concentrate concentration
Imagine the same train receives a feed dissolved-solids concentration Cf = 35,000 mg/L and produces permeate concentration Cp = 350 mg/L. The measured apparent rejection fraction is 1 − Cp/Cf = 1 − 350/35,000 = 0.99, or 99%.
Observed salt passage under this convention is Cp/Cf = 1%. The apparent percentages refer to the defined concentration metric and operating sample; they are not guarantees about every individual dissolved species.
For a hypothetical conservative solute with well-mixed steady flows, QfCf = QpCp + QcCc. Substitute known values to estimate concentrate concentration Cc = (100×35,000 − 45×350)/55 ≈ 63,350 mg/L.
This result excludes precipitation, adsorption, concentration polarization, chemical reaction and unmeasured side streams. It is not a prediction of the exact local concentration at a membrane surface, where scaling risk may differ.
All concentrations are teaching inputs rather than observations of a particular source or product. Real monitoring should preserve analysis method, temperature, sampling position and uncertainty.
CHAPTER 5 OF 12 · Diagnose membrane change
5. Worked example: membrane water flux and energy per cubic metre
A hypothetical train with active membrane area A = 3,000 square metres produces Qp = 45 cubic metres of permeate per hour. The area-normalised flux is J = Qp/A = 0.015 m/h, or 15 litres per square metre per hour (15 LMH).
Now imagine a clearly defined process boundary consumes electrical power of 135 kilowatts during that same steady hour. Specific energy consumption at this boundary is SEC = 135 kW/(45 m³/h) = 3.0 kWh/m³.
The result is not necessarily a whole-plant energy intensity. Intake, pretreatment, high-pressure pumps, energy recovery, chemical dosing, post-treatment and other auxiliaries may be inside or outside the chosen boundary.
PUB describes desalination’s energy demands and actual integrated Singapore plants. The context demonstrates why a number measured at one RO membrane stage cannot be compared uncritically with a full-site number.
For annual production, time-weight electrical metering and product-water metering over a common period. Rated power multiplied by all calendar hours or nominal membrane flux multiplied by year-end installed area need not reflect actual operation or downtime.
CHAPTER 6 OF 12 · Diagnose membrane change
6. Osmotic pressure changes apparent membrane productivity
A simplified water-permeation relation is Jw ≈ Aw(ΔP − Δπ), where Aw is an appropriate water permeability, ΔP is the hydraulic pressure difference and Δπ is the effective osmotic pressure difference.
Take invented conditions with ΔP = 58 bar and Δπ = 25 bar. The simplified net driving pressure is 33 bar. For illustrative measured flux 15 LMH, the apparent permeability ratio would be 15/33 ≈ 0.455 LMH/bar.
Actual RO membrane vessels have changing feed and concentrate composition and pressure along their flow paths. Concentration polarization and permeate-side pressure complicate the use of one uniform driving-force value.
Water temperature can alter viscosity and flux, while feed salinity changes osmotic opposition. A cold day or salinity increase can reduce raw permeate output even when the membrane itself has not acquired new fouling.
Before diagnosing membrane degradation, compare equivalent operating conditions or apply a suitable standardization model with documented assumptions and valid instrument readings.
CHAPTER 7 OF 12 · Diagnose membrane change
7. Fouling, scaling and damage require distinct explanations
Fouling may involve particles, colloids, organic materials or biofilms. Inorganic scaling arises when poorly soluble compounds precipitate or deposit under suitable chemistry. Seal leakage, membrane degradation and sensor defects are different additional mechanisms.
These problems can affect water production, pressure drop and salt passage in different combinations. No one trend proves which foulant is present or which maintenance action is required.
Pretreatment helps reduce the burden on RO membranes. PUB’s explanation of the Tuas Desalination Plant describes flotation and ultrafiltration as mechanisms for reducing membrane fouling and supporting efficiency.
Investigating a problematic train may require membrane condition records, online sensors, laboratory chemistry, deposit characterization and qualified membrane autopsy methods where appropriate. A trend plot alone cannot identify a deposit’s chemistry reliably.
Cleaning choices, inspection of pressurized vessels and membrane replacement are safety- and plant-specific activities. A general education article should teach how evidence supports decisions, not invent chemical procedures or maintenance thresholds.
CHAPTER 8 OF 12 · Verify whole-plant claims
8. Normalization distinguishes environmental changes from deterioration
ASTM D4516-19a describes standardizing RO performance observations so variable pressure, temperature, feed concentration and recovery can be compared against selected reference conditions. That is a more defensible method than comparing raw flow on two different days.
Suppose the fictional normalized permeate rate was 50 m³/h at baseline and later 42 m³/h using the same appropriate reference model. Percentage difference is (42 − 50)/50 = −16%.
The decrease supports a claim about change in normalized production only if data and model are reliable. It does not constitute a universal trigger for membrane cleaning, replacement or emergency plant response.
By contrast, a raw-flow decline accompanied by approximately stable normalized production can suggest that operating differences explain much of the change. It still does not certify product-water chemistry or eliminate all membrane faults.
A scientific study of corrected normalized permeate flux describes how osmotic-pressure correlations and temperature correction models affect trends. Record the normalization equation, software version, reference conditions and uncertainty in every serious comparison.
CHAPTER 9 OF 12 · Verify whole-plant claims
9. Energy recovery and concentrate impacts belong in the account
Seawater RO consumes pumping energy to overcome osmotic and hydraulic resistance. Some pressure energy remains in concentrate and can be recovered, reducing the net electrical burden under suitable equipment arrangements.
Imagine a constructed energy ledger with 180 kW gross pumping input, 45 kW usable energy recovered as a distinct non-overlapping credit, and 30 kW other loads. Net consumption is 180 − 45 + 30 = 165 kW.
For the fictional 45 m³/h permeate rate, this gives 165/45 ≈ 3.67 kWh/m³ for the stated boundary. It is an arithmetic example, not an observed efficiency or safe engineering configuration.
Concentrate disposal affects the receiving environment, particularly where salinity, residual process chemicals and discharge mixing matter. The most attractive energy or recovery figure may not represent the best environmental balance.
Report electrical SEC, usable potable water, availability, concentrate management, maintenance and lifecycle costs separately. Their combination supports system choices better than one unqualified notion of high performance.
CHAPTER 10 OF 12 · Verify whole-plant claims
10. Independent quality checks complete desalination performance
An RO membrane is a selective barrier, not a universal microbial or chemical safety certificate. Final water may require stabilization, mineral adjustment, additional disinfection or other site-specific steps before distribution.
Online conductivity and flow measurements are useful, but independently calibrated instruments, contaminant-specific laboratory tests and the applicable water-safety process still determine whether the complete supply meets requirements.
Measurement data can be distorted by a sampling-line problem, temperature mismatch, instrument drift, different blend ratios or faulty time pairing. Archive raw and corrected data with quality flags so later interpretation can be audited.
Independent verification should not use the same biased readings that adjusted the model as if they were separate proof of accuracy. Where possible, use suitable held-out samples or reference methods.
A responsible technical summary says what was actually measured, where, over what interval and with what uncertainty. Only authorized operators and regulators can judge safety and operational decisions for a real water supply.
CHAPTER 11 OF 12 · Follow connected sources
11. Practice, FAQs and connected Gold Standard learning routes
Practice A: At 80 m³/h feed and 40 m³/h permeate, recovery is 50%. Explain why an increase in recovery could be unfavorable for scaling or energy.
Practice B: With Cf = 35,000 mg/L and Cp = 350 mg/L, salt rejection is 99%. Why does that not certify all water-quality hazards?
Practice C: For 45 m³/h product over 3,000 m² membrane, compute flux of 15 LMH. What happens to raw flow when operating temperature changes?
Practice D: Compare normalized flow 50 to 42 m³/h and obtain a 16% decrease. Name the calibration, feed chemistry and pressure records needed to interpret it.
FAQ — Is rejection the same as recovery? No. One concerns solute concentration and the other concerns water flow.
FAQ — Does falling raw production prove fouling? No. Changing temperature, pressure, salinity and recovery can change permeate flow.
FAQ — Can good RO performance alone guarantee safe drinking water? No. The complete treatment, monitoring and distribution system must satisfy applicable safety requirements.
Transfer the balance from seawater to brackish groundwater. Re-evaluate osmotic driving force, source contaminants, feasible recovery and concentrate handling before assuming the same membrane design works.
These verified connected routes explain complementary systems and measurement methods:
- The Gold Standard of Water Treatment — the wider water-treatment system
- The Gold Standard Of Water Quality Monitoring — measurements and compliance context
- The Gold Standard of Water Quality Sensor Calibration — traceable instrument readings
- The Gold Standard of Water Balance Modelling — conservation and flows
- The Gold Standard of Water Treatment Process Control — responsible control and barrier monitoring
- The Gold Standard of Hydrology — source-water pathways
- The Gold Standard of Statistics — uncertainty and change detection
CHAPTER 12 OF 12 · Follow connected sources
12. Primary sources, scope and the next learning step
ASTM’s standardization guidance and PUB’s public desalination descriptions provide two different kinds of reference: a comparison procedure and examples of integrated real-world plants. Peer-reviewed work gives additional context about fouling and normalized flux.
Every water flow, salinity, pressure, flux, energy and performance value above is an invented teaching number. The examples do not describe a Singapore plant’s current readings or authorize an operating configuration.
An excellent next project is an audit sheet listing feed, permeate and concentrate flow, composition, pressure, temperature, quality, flux, energy, uncertainty, instrument calibration, normalization and verification checkpoints.
Primary and specialist references for source-water, RO and standardization:
- PUB Singapore — Desalinated Water — RO, pretreatment and energy
- ASTM D4516-19a — Standardizing RO Performance Data — reference conditions for comparable performance
- Fouling and Scaling in RO Plants — Review — foulants and diagnostic indicators
- Corrected Normalized Permeate Flux — Research — limits of temperature and osmotic corrections
The immutable Clementi Mathematics teaching reference illustrates instructional sequence and is not an operating procedure for desalination.