The Six Parameters That Govern Every RO System Design
Industrial RO system design is governed by six interdependent parameters: feedwater SDI <3.0, recovery 75–95% (typically capped at 75% in wastewater service by salt solubility), flux 10–20 LMH, net driving pressure 10–30 bar, salt rejection 96–98%, and array staging sized from feed TDS. These numbers set the pretreatment train, the pump HP, the membrane area, and the daily monitoring KPIs that protect membrane life. None of them is independent — move one and the others shift. Raise recovery and the concentrate TDS rises, the scaling risk climbs, the antiscalant dose changes, and the pump must deliver more NDP to overcome the higher osmotic load. The master rule is simple: feedwater SDI <3.0 is the gating parameter. If the upstream train cannot deliver that, no other design choice — higher pressure, more membrane area, or different staging — will save the membranes from particulate fouling within weeks. The design envelope below is the working set every engineer should commit to memory before specifying a HydropureWater industrial RO system.
| Parameter | Design Envelope (Industrial RO) | Source / Notes |
|---|---|---|
| Feedwater SDI (15-min, 30 psi, 0.45 µm) | < 3.0 target; 3–5 warning band; > 5 unacceptable | Silt Density Index, 15-minute fouling-potential test (per WaterTechOnline) |
| Recovery (permeate ÷ feed) | 75–95% process water; 75% typical cap for wastewater | Capped at ~75% in wastewater service by salt solubility |
| Flux (LMH) | 10–20 LMH for industrial RO on wastewater | Permeate flow per m² of membrane per hour |
| Net Driving Pressure (NDP) | 10–30 bar, function of feed TDS | Feed pressure − osmotic pressure − permeate pressure |
| Salt rejection | 96–98% | Permeate typically has 96–98% fewer dissolved ions than feed |
| Array staging | 2:1 typical (wastewater); sized from feed TDS and target recovery | Two vessels parallel first stage, one vessel second stage |
| Temperature correction | Normalized to 25°C and startup pressure | Compensates for seasonal feedwater temperature swings |
Feedwater Quality and SDI: The Gating Design Parameter
The Silt Density Index is a 15-minute, 30-psi fouling-potential test performed on a 0.45-micron membrane pad; it expresses the rate at which feedwater plugs a fine filter and is the single best field indicator of whether pretreatment is keeping particulate and colloidal loading in check (per WaterTechOnline). The design target is SDI <3.0 at the RO feed — anything in the 3–5 band is a warning, and a sustained SDI >5 will foul a polyamide membrane within weeks. When SDI drifts high, the typical upstream culprits are suspended-solids carryover from a clarifier or DAF system for RO feedwater polishing, cationic coagulant or polymer overdose, and microbiological growth in aged carbon filters. The trade-off is mechanical: tighter SDI means a higher multimedia-filter backwash frequency, a 5-micron cartridge guard filter ahead of the high-pressure pump, and stricter DAF or lamella-clarifier influent controls. SDI should be measured pre-multimedia, post-multimedia, post-carbon, and post-cartridge so the engineer can localize the leak before it kills the membranes (per WaterTechOnline). Cationic filter aids can be used to drop SDI but must be applied with caution — overfeed produces the very fouling matrix they are meant to prevent (per WaterTechOnline).
| Measurement Point | Typical SDI Target | Diagnostic If Exceeded |
|---|---|---|
| Clarifier / DAF effluent | < 6 | Coagulant dose, polymer carryover, or hydraulic surge |
| Post multimedia filter | < 4 | Backwash rate, media condition, or filter aid dose |
| Post activated carbon | < 3.5 | Microbiological growth, channeling, or carbon exhaustion |
| Post 5-µm cartridge (RO feed) | < 3.0 | Cartridge change-out, upstream breakthrough, or pump seal leak |
Recovery, Flux, and Net Driving Pressure: The Hydraulic Triangle
Recovery is the ratio of permeate flow to feed flow; flux is permeate flow per square metre of membrane per hour, expressed in LMH; and net driving pressure is the effective pressure pushing water through the membrane, calculated as feed pressure minus permeate pressure minus the osmotic pressure of the feed-concentrate stream (per WaterTechOnline). For industrial RO on wastewater, the working envelope is flux 10–20 LMH and NDP 10–30 bar, with the higher end of the NDP range reserved for high-TDS feeds. Recovery is capped at about 75% in wastewater service because at higher recovery the concentrate TDS exceeds the solubility of carbonate, sulfate, silica, and fluoride scales — and the membrane starts to scale, not foul (per WaterTechOnline). Temperature is the silent variable: as feedwater warms, permeate flow rises and masks fouling, and as it cools, permeate flow drops and looks like fouling even when the membrane is clean. The fix is normalized permeate flow — a calculation that adjusts the daily data readings to what they would be if the system were operating at startup pressure at 25°C (per WaterTechOnline). Seasonal feedwater temperature variations can make fouling trends difficult to detect if the normalized calculation is not used; for example, surface-water RO feedwater becomes warmer in spring and the apparent permeate flow increase can hide a real fouling problem until the next season, by which point the membrane elements may be severely fouled and the permeate flow rate severely restricted (per WaterTechOnline). The rule of thumb: raising recovery by 10 points raises the concentrate-side osmotic load by roughly 25–35%, which is why high-recovery designs need array staging (typically 2:1) and pump headroom the single-pass design never required. The required inputs for the normalized calculation are temperature, NDP, pressure drop, permeate flow, and feedwater TDS — so those instruments belong in the P&ID from day one.
| Variable | Industrial Wastewater RO Envelope | Constraint Driver |
|---|---|---|
| Flux | 10–20 LMH | Membrane compaction, fouling rate, feed TDS |
| Recovery | 75% typical cap; up to 95% in clean process water with two-pass | Salt solubility (carbonate, sulfate, silica, fluoride) |
| Net driving pressure | 10–30 bar | Feed TDS and temperature; pump selection |
| Salt rejection | 96–98% | Membrane type, age, feed pH, temperature |
| Temperature correction | Normalized to 25°C and startup pressure | Seasonal swings mask or mimic fouling |
Array Staging and Concentrate Management
Array staging is the geometry of vessels in series and parallel; a 2:1 array means two pressure vessels in parallel in the first stage, one vessel in the second stage, and it balances flux and recovery across the train. The concentrate from the first stage becomes the feed to the second, so the second stage always runs on higher TDS and lower flow — the per-vessel recovery must stay inside the design envelope in both stages or scaling breaks out in the tail. For high-recovery designs (>80%) the concentrate is often recycled back to the feed of the first stage, which raises osmotic load further and demands tighter antiscalant control. The diagnostic signal is stage-by-stage differential pressure: if the first (lead) stage shows rising ΔP, the cause is particle, colloidal, organic, microbiological, or coagulant/polymer fouling — chemistry downstream of the membranes is unlikely (per WaterTechOnline). If the last (tail) stage shows rising ΔP, the cause is chemistry — carbonate, sulfate, silica, or fluoride scale — not particulate fouling, and the response is to revisit antiscalant selection, feedwater chemistry, or recovery (per WaterTechOnline). The solubility envelope to design against: CaCO3 LSI/Stiff-Davis index < 0, BaSO4 and SrSO4 ion product < 1.0, and SiO2 < 120 mg/L in the concentrate stream.
Pretreatment Train Sized to the RO Design Envelope
The pretreatment train exists for one reason: to deliver SDI <3.0 to the RO membranes (per WaterTechOnline). Each step is sized to a specific failure mode in the feedwater analysis. Step 1 is a DAF or lamella clarifier for bulk TSS, FOG, and colloidal carryover — multimedia filtration downstream only performs as designed if the upstream suspended-solids load is controlled. Step 2 is a multi-media filter for RO pretreatment (anthracite over sand over garnet) for fine TSS reduction, with SDI tested at the outlet. Step 3 is activated carbon for free chlorine and organics that would oxidize the polyamide thin-film composite membrane. Step 4 is a 5-micron cartridge guard filter as the last mechanical barrier before the high-pressure pump. Step 5 is an automatic antiscalant dosing skid matched to feedwater chemistry and the design recovery curve — antiscalant selection is driven by the dominant scale species (carbonate, sulfate, silica, or fluoride), and the dose is set by the concentrate projection at design recovery. Step 6, an industrial water softener for RO pretreatment, is optional: in many new RO installations a softener is unnecessary and costly, and scale inhibitors are more economical, with softener sizing following feedwater hardness when it is used (per WaterTechOnline). Where the feed carries heavy colloidal or organic load — for example in semiconductor process streams — an ultrafiltration stage can be inserted between the multimedia filter and the cartridge guard to drop SDI well below 3.0 and protect the RO from seasonal fouling events; see the RO design for semiconductor process wastewater reference and the Toronto fab wastewater treatment guide for worked examples.
| Step | Equipment | Function | Design Output to Next Stage |
|---|---|---|---|
| 1 | DAF / lamella clarifier | Bulk TSS, FOG, colloids | TSS < 30 mg/L typically |
| 2 | Multi-media filter (anthracite / sand / garnet) | Fine TSS, turbidity | SDI < 4 post-filter |
| 3 | Activated carbon | Free chlorine, TOC, oxidant removal | Cl₂ < 0.1 mg/L |
| 4 | 5-µm cartridge guard | Final mechanical barrier | SDI < 3.0 to RO feed |
| 5 | Antiscalant dosing skid | Scale inhibition at design recovery | LSI/Stiff-Davis < 0 in concentrate |
| 6 (optional) | Na-form ion-exchange softener | Hardness removal for very hard or sulfate-rich feeds | Hardness < 1 mg/L as CaCO₃ when used |
From Design to Daily KPIs: The Operational Handoff
The six design parameters are not abandoned at startup — they become the daily KPIs that protect the design life of the membrane. SDI is verified pre- and post-pretreatment; recovery is read off the feed and permeate flow meters; flux is tracked as normalized permeate flow; NDP is read off the pump discharge gauge; rejection is calculated from feed and permeate conductivity; and array health is monitored as ΔP per stage (per WaterTechOnline). The single most useful diagnostic is the first-stage-versus-last-stage ΔP rule: first-stage rise points to particle, colloidal, organic, microbiological, or coagulant/polymer fouling, while last-stage rise points to carbonate, sulfate, silica, or fluoride scale (per WaterTechOnline). The clean-in-place (CIP) trigger is not a 25% rise in ΔP — by that point the foulant is forced deeper into the membrane, cleaning efficacy drops, and channeling can develop along paths of least resistance that let cleaning chemicals bypass the rest of the element (per WaterTechOnline). The pressure drop coefficient is more sensitive than absolute ΔP and is the better CIP trigger. The normalized permeate flow calculation needs temperature, NDP, pressure drop, permeate flow, and feedwater TDS as inputs, so all of those instruments must be specified in the P&ID at design time — not retrofitted when the operator notices the first seasonal anomaly. Specifying quality RO membrane elements and pressure vessels and reliable isolation valves and instrument fittings at design stage keeps the daily KPI stream trustworthy.
| Design Parameter | Daily KPI | Instrument / Location | CIP / Investigate Trigger |
|---|---|---|---|
| SDI < 3.0 | SDI test, pre- and post-pretreatment | Online or bench test at each train stage | Sustained SDI > 3.0 at RO feed |
| Recovery 75–95% | Permeate ÷ feed flow | Feed and permeate flow meters | Drift > 2 points from design |
| Flux 10–20 LMH | Normalized permeate flow (25°C, startup NDP) | Permeate flow, temperature, NDP inputs | > 10% decline over 30 days |
| NDP 10–30 bar | Pump discharge pressure − permeate pressure − osmotic Δ | High-pressure pump gauge, permeate backpressure | Pressure rise at constant flux |
| Rejection 96–98% | Permeate vs. feed conductivity | Conductivity probes on both streams | Rejection drop > 2 points |
| Array health | ΔP per stage; pressure drop coefficient (PDC) | Inlet and outlet pressure taps per vessel | First-stage rise = fouling; last-stage rise = scaling |
Frequently Asked Questions
What SDI is required for an industrial RO system?
The design target at the RO feed is SDI < 3.0, measured as a 15-minute, 30-psi fouling-potential test on a 0.45-micron pad. Values in the 3–5 band are a warning; sustained SDI > 5 will foul a polyamide membrane within weeks (per WaterTechOnline).
What is the typical recovery limit for a wastewater RO system?
Recovery is the ratio of permeate to feed flow, and in wastewater RO it is typically capped at 75% because at higher recovery the concentrate TDS exceeds the solubility of carbonate, sulfate, silica, and fluoride scales (per WaterTechOnline). Clean process-water systems can reach 85–95% recovery with two-pass staging and tighter antiscalant control.
How is normalized permeate flow calculated?
Normalized permeate flow is a calculation that adjusts the daily data readings to what they would be if the system were operating at startup pressure at 25°C (per WaterTechOnline). It needs temperature, net driving pressure, pressure drop, permeate flow, and feedwater TDS as inputs, and it removes the seasonal temperature effect that otherwise hides — or mimics — membrane fouling.
How do I tell fouling from scaling on an RO train?
Compare ΔP across the first stage versus the last stage. A first-stage ΔP rise points to particle, colloidal, organic, microbiological, or coagulant/polymer fouling; a last-stage ΔP rise points to carbonate, sulfate, silica, or fluoride scale (per WaterTechOnline). The response is different for each — cleaning chemistry and CIP frequency for fouling, antiscalant selection and recovery review for scaling.