Multiple industries nationwide (petrochemical/coal chemical/power/printing and d
Pilot to full-scale engineering (tens to tens of thousands of m³/d)Membranes Aren't Worn Out, They're Washed Out: RO/UF Fouling Diagnosis and Cleaning Protocol Correction
The polyamide membrane housed in a sealed membrane shell only passes water and involves no friction, so it appears that it "won't wear out." However, in field statistics, the actual service life of membrane elements is often only two to three years, far shorter than the 3–5 年 claimed by manufacturers. A closer look at the accounts reveals that what truly "uses up" the membrane is not operating hours, but the cycle of fouling and cleaning—especially cleaning when it should not be done, not cleaning when it should be, using the wrong chemicals, and escalating parameters. This article brings fouling diagnosis and cleaning decisions back to calculable data.
I. The physicochemical nature of four types of fouling and their respective "fingerprints"
Membrane fouling is not a single phenomenon, but the superposition of four processes with different mechanisms, each with a different way of "striking" membrane performance. Confusing the types and using the wrong chemicals is the most common cause of cleaning failure.
① Particulate and Colloidal Fouling
Suspended solids, clay flakes, silica micropowder, and iron oxide flocs deposit on the membrane surface to form a cake layer, characterized by "intercepted but flushable": particles accumulate on the membrane mainly through physical stacking with weak binding forces, and most can be removed by conventional backwashing or cross-flow shear. The fingerprint is simultaneously elevated feedwater SDI and turbidity, with inter-stage pressure differential rising first while flux declines slowly. Submicron colloids in the range of 1–3 μm are the culprit—conventional sand filtration only intercepts down to 10–20 μm, and these colloids remain stably suspended due to their negative ζ potential, directly plugging the membrane surface (Reynolds & Bauhm public information).
② Organic Fouling
Natural organic matter (NOM), proteins, polysaccharides, and oils adhere to the membrane surface through adsorption. Its difference from particulate fouling lies in strong binding forces and resistance to physical detachment. The fingerprint is elevated feedwater COD/TOC, slowly but continuously rising pressure differential, gradual flux decline while salt rejection remains relatively stable. Industry recommendations suggest controlling feedwater TOC within 2 mg/L to inhibit organic matter and bacterial growth (Dalton Membrane Technology public information).
③ Scaling
Sparingly soluble salts on the concentrate side crystallize and precipitate on the membrane surface after exceeding their solubility product, mainly calcium carbonate, calcium/barium sulfate, silicates, and iron-manganese oxides. It is driven by concentration factor and occurs in the terminal stage. The fingerprint is typical: when LSI/SDSI exceeds limits, terminal-stage pressure differential rises, and salt passage increases along micro-cracks in the scale layer—this is the key feature distinguishing scaling from other fouling types (Spans Envirotech public information). Among these, silica scale is the most difficult to tackle: silicon dioxide has chemical properties similar to glass, and conventional acid-alkali cleaning is essentially ineffective. Its solubility is approximately 100 mg/L at 25 ℃ and approximately 160 mg/L at 40 ℃ (public physicochemical data), and as little as 50 ppb of aluminum in water can trigger performance degradation at silica concentrations of only 10 ppm (China Membrane Industry Association Operations & Maintenance Branch public information).
④ Biofouling
Bacteria proliferate on the membrane surface and secrete extracellular polymeric substances (EPS), forming a slime layer. This is the most "expensive" category: EPS enmeshes organic matter, inorganic scale, and particles together to form a composite fouling layer that is extremely difficult to remove completely. The fingerprint is rapidly rising pressure differential (normal monthly increase rate should be <5%, >10% requires immediate intervention—Kebei Environmental Protection public information), slimy and odorous membrane surface, and declining salt rejection. Membrane surface dissection of reclaimed water reuse systems shows that chlorine-resistant bacteria are highly enriched on the membrane surface (Mycobacterium relative abundance 9.72%–73.42%, Methylobacterium 5.65%–24.71%), indicating that conventional chlorine disinfection is nearly ineffective against biofilms and instead selects for chlorine-resistant bacterial communities ("Optimization Study on Reverse Osmosis Membrane Cleaning Technology for Reclaimed Water Treatment Based on Membrane Fouling Component Identification," journal paper).
II. Criteria Analysis: What Each of the Five Indicators "Governs" and What It Does Not
Each indicator for determining "whether to clean" has a clear applicable boundary. Using them as universal thresholds will lead to misjudgment. The following breaks them down one by one.
| Criterion | What It Actually Measures | Applicable Boundaries / Common Misuse | Basis |
|---|---|---|---|
| SDI₁₅ | The flow rate decay rate of 0.45 μm membrane under 207 kPa, reflecting the particulate and colloidal fouling tendency | Does not reflect biological, organic, or scaling fouling tendencies; the validity upper limit of SDI₁₅ is 75% flow loss, so its theoretical maximum is only 5.0; results from different test durations are not comparable | ASTM D4189; AWC public information |
| LSI / SDSI | The degree of calcium carbonate saturation (LSI=pH−pHs) | Only valid for calcium carbonate, excluding calcium sulfate, silica, and phosphate; originally designed for ambient-temperature drinking water, it has accuracy deficiencies when used for cooling water/boiler water and should be used as a trend value | ASHRAE Handbook Ch.50 |
| Turbidity | Light scattering intensity, reflecting suspended particles | Insensitive to submicron colloids and dissolved pollutants; "looking clear" does not equal SDI compliance | Reynolds & Bauhm public information |
| Inter-stage Differential Pressure ΔP | The degree of blockage of the membrane surface/feed channel | Varies with feed flow rate and temperature, and must be normalized before comparison; judging by absolute value alone easily leads to misjudgment | ASTM D4516; GB/T 23954-2009 |
| Normalized Flux / Salt Passage | The true water production capacity and desalination capacity after conversion to baseline temperature and pressure | Requires TCF correction; this is the only indicator that can directly determine "whether to clean or replace" | DuPont FilmTec Technical Manual; ASTM D4516 |
Why "Normalization" Is Essential Before Judgment
Membranes are temperature-sensitive: for every 1 ℃ increase in water temperature, permeate flow increases by approximately 2%–3%; conversely, for every 3 ℃ decrease in temperature, permeate flow drops by approximately 10% (Vontron Technology public data). Looking at absolute values alone, a winter "permeate flow drop" is easily misjudged as severe fouling, when in fact the membrane is not dirty. The commonly used engineering temperature correction factor (TCF) is:
TCF ≈ 1.03(25−T) or TCF = exp[K×(1/Tabs−1/298)], K≈2640–3000
Where T is the actual water temperature (°C), with a reference of 25 ℃. Normalized permeate flow = measured permeate flow × TCF; normalized salt passage = measured salt passage ÷ TCF (approximate). The baseline should be established after commissioning or cleaning, within 48 h of stable operation, and thereafter verified at least once per week with trend lines plotted (ASTM D4516, AWWA M61). A deviation of the normalized value from baseline of 10%–15% triggers investigation and usually scheduling of cleaning; a deviation of 30% may indicate permanent damage has already occurred.
III. Cleaning Decision: When to Clean, In What Order, and How to Set Parameters
3.1 When to Clean — Three Red Lines, with "Normalization" as an Indispensable Prerequisite
GB/T 23954-2009 "Technical Specification for Cleaning of Reverse Osmosis System Membrane Elements" and the DuPont FilmTec technical manual have essentially consistent trigger conditions: normalized permeate flow declining by 10%–15% from the initial value, normalized salt passage increasing by 5%–10% from the initial value, and normalized inter-stage differential pressure rising by 10%–15% from the initial value; meeting any one of these should trigger scheduling of cleaning. Vontron Technology likewise gives three thresholds of "permeate flow drop of 15%, salt passage increase of 10%, differential pressure increase of 15%," with consistent logic: normalize first, then compare to baseline, then examine the trend.
It is necessary to distinguish between the two errors of "cleaning too early" and "cleaning too late." Cleaning too late: membrane surface fouling compacts and biofilm matures, cleaning recovery rate drops significantly, and energy consumption and water quality risks rise simultaneously during the delay. Cleaning too early: if the system needs cleaning every few weeks, that is not "valuing maintenance" but rather the bill for deferred pretreatment — at this point, the pretreatment should be modified, not the cleaning intensity increased.
3.2 Alkali First Then Acid, or Acid First Then Alkali — Default is "Alkali First, Acid Second," but There Are Exceptions
The engineering default sequence is alkali cleaning first, acid cleaning second, with sufficient flushing with permeate to neutrality (pH 6.5–7.5) in between before proceeding to the next step. The logic is: alkali cleaning saponifies organics, disperses colloids, and strips biofilm, clearing the way for subsequent steps; if the order is reversed, inorganic salt residues from acid cleaning will cross-link with organics from subsequent alkali cleaning, forming a composite fouling layer that is harder to remove; and when alkali cleaning residue carries over into the acid cleaning step, calcium carbonate instantly generates CO₂ bubbles upon contact with acid, causing scouring damage to the membrane surface (Changhai Environmental Protection public data).
There are three types of exceptions: ① When carbonate scaling is the absolute dominant cause with no obvious organic fouling, acid cleaning may be performed first; ② Calcium sulfate/barium sulfate scaling is not amenable to ordinary acid cleaning, and an alkaline + chelating agent (STPP, EDTA) system should be used instead; ③ Silica scaling requires a fluoride-containing system (see Fact-Check Box III). Some power plants actually adopt a three-stage "acid cleaning + alkali cleaning + acid cleaning" process to treat mixed calcium sulfate scale (Huaneng Shang'an Power Plant case).
3.3 Parameter Window: pH, Temperature, Concentration, Duration
Cleaning is not about "the stronger the chemical, the hotter the water, and the longer the soak, the better." The table below shows parameter windows that are consistent with each other in publicly available sources.
| Parameter | Alkaline Cleaning (Organic/Biological) | Acid Cleaning (Inorganic Scaling) | Key Constraints |
|---|---|---|---|
| Chemicals and Concentration | 0.1%–0.5% NaOH; typically formulated with 0.025%–0.1% surfactant (SDS/SDBS) | 2% citric acid (pH adjusted with ammonia); or 0.5%–1% hydrochloric acid; for combined fouling, use a mixture of hydrochloric acid + citric acid | GB/T 23954-2009: acid cleaning pH 2–4, alkaline cleaning pH 10–12 |
| pH | 11–12 (some cases 11.5) | 2–3 (some references 2.5–3.5) | pH change >0.5 requires additional chemicals; 30 min unchanged is considered reaction complete |
| Temperature | 30–35 ℃ | 25–30 ℃ (some cases 30–35 ℃) | Polyamide membrane cleaning should not exceed 45 ℃; pH tolerance range narrows at high temperatures |
| Circulation Flow/Pressure | Low flow (1/3 of normal cleaning flow or half of design flow) fed and circulated; circulation pressure typically 0.2–0.4 MPa, low enough to produce no permeate | Flow per membrane housing: low flow approx. 5 m³/h, high flow 9–10 m³/h (power plant case) | |
| Duration | Circulation 30–60 min + soaking 30 min + recirculation | Circulation 30–60 min | Single-stage circulation for alkaline/acid cleaning recommended ≤2 h; chlorine-containing oxidative cleaning strictly controlled within 30 min |
| Make-up Water | Must be prepared with RO permeate or deionized water; tap water is strictly prohibited (residual chlorine will oxidize the membrane) | GB/T 23954-2009 / FilmTec Technical Manual | |
3.4 Differences in Cleaning Systems Between UF and RO
Ultrafiltration is a low-pressure membrane with better oxidation resistance, and its cleaning regime is more "aggressive" and more frequent than that of RO, divided into three levels: physical backwash (high frequency) + maintenance chemical cleaning CEB (medium frequency) + restorative deep CIP (low frequency). Physical backwash is performed once every 20–120 min of operation, with a flow rate of 1.5–2 times the permeate flow and a pressure of 0.15–0.25 MPa; CEB adds 50–500 mg/L sodium hypochlorite to the backwash water or adjusts the pH to 2 with acid solution, 1–2 times per day; deep CIP is triggered at 60–120 天 or by differential pressure. TMP red lines: daily stable 0.04–0.10 MPa, rising to 0.12 MPa triggers CEB, rising to 0.15 MPa shuts down for CIP, exceeding 0.15 MPa for a long time can very easily cause batch fiber breakage of hollow fiber membranes (public information from Qiukanghong, Changhai Environmental Protection, AQUChem, etc., commercial sources). Material differences must be kept firmly in mind: PVDF is chlorine-resistant and can be cleaned with sodium hypochlorite; PES has poor chlorine resistance, so the membrane manufacturer's tolerance table must be checked before use.
IV. Real Ledger: Cleaning Data from Public Projects and Peer Review
The table below summarizes publicly verifiable engineering cases and research results. The "recovery rate" in manufacturer cases is mostly expressed as the degree of permeate increase, which differs from the laboratory "flux recovery rate (FRR)" basis and should not be directly compared across sources.
| Project / Scale | Fouling Type and Characterization | Cleaning Protocol | Recovery Performance | Source |
|---|---|---|---|---|
| Huaneng Shang'an Power Plant Circulating Water Treatment RO | Calcium sulfate + calcium carbonate + microbial combined fouling; permeate flow decline, desalination rate decrease, differential pressure increase | Three-stage "acid wash + alkaline wash + acid wash"; acid wash pH 2–3 hydrochloric acid, alkaline wash 2.5% sodium carbonate + 2% EDTA-4Na | Permeate flow increased by 35.4%, first-stage differential pressure decreased by 35.7%, second-stage differential pressure decreased by 50% | Zhongjin Environmental Protection public engineering data (case source) |
| Shenmu Electric Chemical Thermal Power Workshop RO (two trains) | Long-term scaling and foulant attachment; excessive inter-stage differential pressure, elevated permeate conductivity, declining permeate flow | "Alkaline wash + acid wash" staged cleaning; alkaline wash pH 11–11.5, 30–35 ℃, acid wash pH 2–2.5; circulation–soaking–recirculation | Permeate flow 75 → 96 m³/h, inter-stage differential pressure decreased, parameters returned to normal | Shenmu Electric Chemical public report (case source) |
| Changzhi Thermal Power Boiler Make-up Water RO | Organic and microbial combined fouling; continuously rising permeate conductivity, sharply increased inter-stage differential pressure, declining permeate flow | Alkaline wash (organic stripping) + acid wash (inorganic scale dissolution), with strict control of three parameters: cleaning solution temperature, circulation flow rate, and pH | Permeate conductivity fell back to the excellent range, permeate flow increased by nearly 20% compared with before cleaning, differential pressure dropped significantly | Jinneng Holding Changzhi Thermal Power public report (case source) |
| CHN Energy Fujian Fuzhou Company RO | After heating retrofit, first-stage differential pressure rate surged and permeate production rate plummeted; cleaning required every 35 天, security filter cartridges replaced every 30 天 | Instead of increasing chemical dosing, optimized pretreatment + adjusted dosing frequency + flushing time 600 min + non-oxidizing biocide 5 天 cycle + strictly controlled UF influent residual chlorine to 0.1–0.3 mg/L | Cleaning cycle 35 → 100 天, security filter cartridges 30 → 120 天, chemical dosage reduced by 1/3, annual savings of approximately 300 ten-thousand yuan | China Electric Power Technology Market Association 2025 年 Power Generation Operation Water Treatment Technology Innovation Achievement Award report (case source) |
| Shaanxi Coal Coal-to-Olefins Reuse Water RO / 10 trains, permeate 1600 m³/h, recovery 73% | High salinity and high COD, complex and variable water quality; high differential pressure, frequent cleaning (once every 15 天) | High-efficiency concentrated scale inhibitor (0.3 ppm) + alternating dosing of two non-oxidizing biocides (to eliminate resistance) | Recovery 65% → 73%, differential pressure significantly reduced, cleaning cycle 15 → 50–60 天 | OMEX New Materials public engineering performance (vendor source) |
| A Large Thermal Power Plant in Northern China Boiler Make-up Water RO (domestic membranes, surface water/Yellow River water) | Organic, colloidal, and microbial fouling; after 6 months of operation, first-stage differential pressure increased and permeate flow declined | NaOH + sodium dodecylbenzenesulfonate + sodium tripolyphosphate, pH≈11.5, wash 8 h; followed by hydrochloric acid + citric acid pH 2–3, 30 ℃, wash 1 h; final stage non-oxidizing biocide treatment; staged, low flow → soaking → high flow | First-stage differential pressure returned to normal, permeate flow recovered to 45 m³/h, good recovery | China Environmental Protection Network "Application of Domestic Reverse Osmosis Membranes in Thermal Power Plant Boiler Make-up Water Projects" (case source) |
| Reclaimed Water Treatment RO (pilot study) | Predominantly organic–biological fouling; inorganic elements on membrane surface mainly Al, Ca, Mg, with Al heavily deposited despite low influent concentration | NaOH alkaline wash followed by HCl acid wash, 35 ℃, 9.6 m³/h; alkaline wash 0.1% NaOH + 0.025% SDBS + 1% EDTA-4Na + 2% NaCl (pH≈12); acid wash 0.2% hydrochloric acid + 2% citric acid (pH≈2); repeated short soaking–circulation | Flux increased by 31.98% (total circulation 2 h + soaking 4 h, equivalent to circulation 2 h + soaking 15 h); optimized flux recovery rate 45.74%; improved removal of Ca/Mg/Fe, but Al removal rate remained low | "Optimization Study on Reverse Osmosis Membrane Cleaning Technology for Reclaimed Water Treatment Based on Membrane Fouling Component Identification" (peer-reviewed) |
| Organically Fouled RO Membrane (laboratory, DOE optimization) | Organic fouling | 0.68 wt% EDTA, 20 min, 20 ℃, flow rate 409 mL/min (statistically optimized results, superior to manufacturer-recommended 60 min/30 ℃) | Flux recovery rate approximately 86.6% (predicted); measured cleaning efficiency approximately 2.3 percentage points higher than manufacturer recommendation | Park et al., Environmental Engineering Research, 2018 (peer-reviewed) |
| Discarded RO Membrane Regeneration (laboratory) | Mixed fouling | High temperature 45–50 ℃; acid and alkaline equal duration 3 h each (pH 3.0 and 12.0 optimal); 0.005% sodium hypochlorite | High-temperature flux recovery 71%, salt rejection >80%; acid-alkaline equal duration recovery 72.43%; 0.005% sodium hypochlorite reached 85.2%, but increased concentration caused oxidation and decreased desalination rate | Renewal (Techno-Press) paper (peer-reviewed) |
| A Food Factory UF System in Shandong | Operation at maximum flux caused TMP to rise by 40% over 3 months and filtrate turbidity to increase | Reduced design flux by 15% + added weekly CEB + installed 100 μm pre-screening | Cleaning interval extended from weekly to monthly, membrane element lifespan increased from 2 年 to over 4 年 | Loyal Membranes public technical guide (commercial source) |
V. Process Comparison: Trade-offs of Four "Fouling Control" Approaches
| Route | Applicable Scenarios | Advantages | Costs and Boundaries |
|---|---|---|---|
| Physical Flushing / Backwash (including air scouring) | Loose filter cake dominated by particles and colloids; UF routine maintenance | No chemicals, no chemical damage, lowest cost, can be performed at high frequency | Basically ineffective against adsorbed organics, biofilm, and crystalline scale; excessive backwash pressure (>0.25 MPa) can damage hollow fibers |
| Online Chemical Cleaning CIP (without removing membranes) | Mild to moderate fouling; clear fouling type | Short downtime (typically 2–6 h/stage), controllable operation | System piping and cleaning solution must be compatible; during staged cleaning, "cleaning this section fouls that section" easily occurs |
| Offline Cleaning (removing membrane elements) | Online cleaning cannot restore performance; uneven fouling distribution | Individual membrane elements can be treated in a targeted manner; autopsy analysis can be combined to locate fouling | Hoisting and reinstallation carry mechanical damage risks (O-ring crushing, reverse installation causing salt rejection to drop to zero); downtime is significantly extended |
| Membrane Element Replacement | After cleaning, standardized flux remains below 70%–80% of the initial value, salt passage doubles without recovery, cleaning intervals collapse | One-time restoration of performance and water quality assurance | High cost (the replacement volume of membrane element sets is considerable); if pretreatment is not improved, new membranes will be consumed at the same rate |
VI. Fact-Check Box: Three Most Commonly Misused Definitions
Many sites treat "feedwater SDI<5" as a universal threshold. In fact, SDI only reflects the particulate and colloidal fouling tendency, and does not indicate biological, organic, or scaling tendencies (public information from American Water Chemicals). Moreover, SDI₁₅ is valid only on the premise of a flow rate loss <75%, so its theoretical maximum is exactly 5.0; the upper limit of SDI₁₀ is 7.5, and the upper limit of SDI₅ is 15. Results from different durations are not comparable (ASTM D4189). The "passing line" also varies by tier: most membrane manufacturers' guaranteed value is SDI₁₅≤5, DuPont recommends ≤3, and semiconductor/pharmaceutical ultrapure water typically requires <1. SDI only addresses particulates and colloids; biological and scaling issues require other indicators.
LSI is similarly misused: LSI = pH − pHs, where pHs = (9.3 + A + B) − (C + D), A = [log(TDS) − 1]/10, B = −13.12 − log(T + 273) + 34.55, C = log[Ca²⁺] − 0.4, D = log[alkalinity]. The ASHRAE Handbook clearly states that such indices only indicate the degree of calcium carbonate saturation, are not applicable to other calcium compounds such as calcium sulfate, and certainly do not cover silica; moreover, LSI was originally designed for ambient-temperature drinking water, and its accuracy is seriously deficient when used for cooling water and boiler water, making it more suitable as a trend criterion (ASHRAE Handbook—Fundamentals Ch.50; Langelier 1936). Calcium sulfate/barium scaling requires SDSI or a dedicated sulfate scaling model, while silica scaling requires a separate SiO₂ calculation on the concentrate side.
Polyamide composite membranes (BW30 series) have a cleaning pH tolerance of 1–12, but the higher the temperature, the narrower the pH tolerance range: at about 30 ℃, pH 1–12 can be tolerated; at 50 ℃, it narrows to pH 2–10; SW30 seawater desalination membranes have a cleaning pH of 2–11. The cleaning temperature for conventional industrial RO membranes generally should not exceed 45 ℃ (some sources give a limit of 40 ℃); cellulose acetate membranes can tolerate 50 ℃; heat-sanitizable membrane elements for pharmaceutical/food applications can tolerate 85 ℃ after special design (public information from Dalton Membrane Technology). Conclusion: The upper temperature limit is not a constant; it must be checked against the specific membrane model manual and judged together with the pH at that time. Silica scaling is the only exception: conventional acid and alkali are ineffective, requiring a 0.1% HF + 0.4% HCl or 0.1% NaF + 0.4% HCl system, and hydrofluoric acid is highly toxic and causes extremely severe injuries, so it must be handled by professionals with waste liquid neutralization (public information from the Operation and Maintenance Branch of the China Membrane Industry Association)—the safety threshold is extremely high, and it must not be attempted on one's own.
Each cleaning is a chemical and mechanical shock: high/low pH stress, oxidation risk, and flow impact. The quantitative statement given in public O&M information is that membranes cleaned monthly age at about three times the rate of those cleaned quarterly (Spans Envirotech, commercial source pending verification); long-term high-concentration cleaning can shorten the effective membrane lifespan by 20%–30% (public information from Dalton Membrane Technology). Moreover, there are too many points where cleaning operations can go wrong—incorrect concentration preparation, excessive soaking time, uncontrolled temperature, incomplete rinsing—each one increasing the probability of human error. Therefore, the correct mindset is not "clean frequently," but rather "clean once at the right time, with the right chemicals, and at the right parameters."
VII. Engineering Truth: Five Field Boundaries
References
- GB/T 23954-2009 "Technical Specification for Cleaning of Membrane Elements in Reverse Osmosis Systems", Water Treatment Chemicals Branch of the National Technical Committee for Chemical Standardization, issued on 2009, implemented on 2010.
- ASTM D4189 (SDI Test Method); ASTM D4516 (Standardization of RO Performance Data).
- DuPont / FilmTec, FilmTec Reverse Osmosis Membranes Technical Manual (chapters on cleaning trigger conditions and cleaning solution parameters).
- Vontron Technology (Vontron), "Guidelines for Cleaning of Reverse Osmosis Membrane Systems", publicly available technical documentation.
- ASHRAE Handbook—Fundamentals, Chapter 50: Water Treatment (LSI and RSI), ASHRAE.
- Langelier W. F., 1936, Journal AWWA; Ryznar J. W., 1944, Journal AWWA.
- Park K-B et al., 2018, "Optimization of chemical cleaning for reverse osmosis membranes with organic fouling using statistical design tools", Environmental Engineering Research, 23(4).
- "Optimization of chemical cleaning of discarded reverse osmosis membranes for reuse" (Renewal / Membrane Water Treatment) and Mahidol University "Statistical optimization of chemical cleaning conditions…" publicly available research.
- "Research on Optimization of Reverse Osmosis Membrane Cleaning Technology for Reclaimed Water Treatment Based on Identification of Membrane Fouling Components", journal paper (including community sequencing and pilot-scale cleaning optimization).
- Operation and Maintenance Branch of China Membrane Industry Association, "Common Cleaning Methods and Chemical Formulations for Reverse Osmosis Membranes" and "Challenges in Membrane Concentration for Zero Liquid Discharge—Silicate Scaling", publicly available technical documentation.
- Case study on reverse osmosis membrane fouling and cleaning in the circulating water treatment system of Huaneng Shang'an Power Plant, publicly available engineering documentation from Zhongjin Environmental Protection.
- Public report on specialized reverse osmosis membrane cleaning at the thermal power workshop of Shenmu Electrochemical Co. (alkaline cleaning pH 11–11.5, acid cleaning pH 2–2.5, water production 75→96 m³/h).
- Public report on emergency reverse osmosis membrane cleaning at Jineng Holding Changzhi Thermal Power (organic + microbial composite fouling, water production increased by nearly 20%).
- CHN Energy Fujian Fuzhou Company "Reverse Osmosis Membrane Life Management and Early Warning" QC project, report on the 2025 年 Power Generation Operation Water Treatment Technology Innovation Achievement Award from the China Electric Power Technology Market Association.
- Shaanxi Coal Coal-to-Olefins Reclaimed Water Project (water production 1600 m³/h, recovery rate 73%), publicly available performance documentation from Oumei New Materials; "Application of Domestic Reverse Osmosis Membranes in Boiler Make-up Water Projects of Thermal Power Plants", public report from China Environmental Protection Network.
- American Water Chemicals "Silt Density Index (SDI)"; Reynolds & Bauhm "Colloidal Turbidity & SDI Control"; Spans Envirotech "RO System Troubleshooting"/"RO Membrane Life and Replacement"; AXEON Water "How Do Membrane Cleaning Systems Extend RO System Lifespan?"; Loyal Membranes "Ultrafiltration Membrane for Water Treatment" (all publicly available technical documentation, some from commercial sources).
Supplementary Notes (Data Gaps)
- The quantitative relationship between "cleaning frequency and membrane lifespan" currently only has commercial O&M data as its basis, lacking support from authoritative standards or large-sample peer-reviewed studies. Please rely on the regression of this plant's membrane element replacement records as the standard. [To be supplemented]
- There is a lack of domestic peer-reviewed quantitative reports on the membrane flux recovery rate of UF membrane deep CIP; publicly available information on engineering safety specifications and waste liquid disposal standards for cleaning in fluoride-containing systems is limited. [To be supplemented]
