Industrial reverse osmosis (RO) procurement is decided by four numbers before price: feed total dissolved solids (TDS), silt density index (SDI) at the membrane inlet, target recovery rate, and daily permeate demand in GPD or m3/day [S3][S4]. Spec sheets that quote rejection rate and capacity without naming the feed TDS envelope routinely underperform from day one, because operating pressure, flux, and element life all scale with feed chemistry [S4].
For facilities above 80 GPH (≈300 L/h), system sizing follows the relation Required feed capacity = (Daily permeate ÷ Recovery) × 1.15–1.25 safety factor; brackish water (BWRO) systems typically run 50–85% recovery, seawater (SWRO) 35–45%, and low-TDS municipal or well water 75–80% [S4][S1]. A facility needing 10,000 GPD permeate at 75% recovery thus requires roughly 13,500 GPD of feed capacity, before margin [S4].
Feedwater envelope and element geometry
TDS bands dictate membrane class: under 1,000 ppm uses low-pressure BWRO at 100–150 psi, 1,000–10,000 ppm uses standard BWRO at 150–400 psi, and 10,000–45,000 ppm requires SWRO at 800–1,200 psi, with 97–99.8% salt rejection across the family [S4]. Every current industrial-grade element is a polyamide thin-film composite (TFC); cellulose acetate (CA) elements from the 1970s are obsolete and should be rejected at the RFQ stage [S4].
For high-flow industrial skids, the 8-inch × 40-inch 8040 element is the de facto standard; 4040 elements fit smaller commercial packages below roughly 5,000 GPD permeate [S4]. Operating temperature also matters: every 1°C below the 25°C reference drops membrane permeate output by 1–2%, so cold-feed installations must be derated or staged [S4].
Free chlorine and other oxidants destroy polyamide TFC membranes; feed must be reduced to below the element's continuous chlorine tolerance, normally handled with activated carbon or sodium metabisulfite dosing before the high-pressure pump [S4]. Iron and manganese above 0.1 ppm trigger oxidation plus media filtration; silica above saturation drives recovery limits and antiscalant selection [S4].
Pre-treatment, SDI, and recovery trade-offs
SDI15 at the membrane inlet must be ≤3, with lower values expected on fouling-prone feed; values above 5 force multimedia filtration, cartridge filtration, or ultrafiltration upstream and frequently void membrane warranties [S4]. The economics are well known to operators: every dollar saved on pre-treatment typically costs three dollars in lost membrane life within the first 24 months [S4].
Recovery is not a free parameter. Pushing BWRO past 80% on hard feed raises calcium carbonate and calcium sulfate scaling risk; pushing SWRO past 45% raises boron and silica passage. Antiscalant dosing plus pH adjustment (often sulfuric acid) is the standard hedge, with lime or sodium hydroxide for post-treatment pH correction depending on the downstream use [S5][S4]. Energy-recovery devices (ERDs) are now standard on SWRO above about 100 m3/day permeate, cutting specific energy from 6–8 kWh/m3 to 2–3 kWh/m3 [S6].
Waste brine volumes of 15–25% of feed are typical for BWRO; sites without a sewer or evaporation pond permit must size for zero-liquid-discharge (ZLD) with a brine concentrator and crystallizer, which usually doubles the skid's footprint and capex [S5][S6].
Membrane selection criteria: a side-by-side comparison
The decision matrix for the RFQ stage fits on one page. Using public 2024–2026 buyer guidance, the three operating classes line up as follows against the criteria that drive cost and risk [S4][S6]:
<strong>Low-pressure BWRO (TFC, <1,000 ppm feed)</strong>: 100–150 psi operating pressure, 97–99% rejection, 75–80% recovery, lowest energy and pump capex, suited to tap, well, and light industrial process water. Limitations: not valid above ~1,500 ppm; sensitive to feed hardness spikes [S4].
<strong>Standard BWRO (TFC, 1,000–10,000 ppm feed)</strong>: 150–400 psi operating pressure, 97–99.5% rejection, 50–85% recovery, the workhorse for municipal reuse, food and beverage, boiler feed, and microelectronics rinse. Limitations: pump and energy cost roughly 2–3× the low-pressure class; requires reliable antiscalant dosing [S4][S3].
<strong>SWRO (TFC, 10,000–45,000 ppm feed)</strong>: 800–1,200 psi operating pressure, 99–99.8% rejection, 35–45% recovery, mandatory for seawater intake and high-salinity produced water. Limitations: high-pressure pump, ERD, and corrosion-resistant piping (typically 2507 super duplex or FRP) drive capex 3–5× a comparable BWRO skid of the same permeate capacity [S4][S6].
Who RO is for, and where it is the wrong tool
RO is the right answer when permeate total dissolved solids must drop below 50–200 ppm and the feed is a known, characterized stream: boiler feed make-up, microelectronics rinse, pharmaceutical water pretreatment, food and beverage ingredient water, and seawater or brackish desalination [S1][S3]. It is the wrong answer for low-TDS polishing where a deionized water system or simple carbon filter handles the load at lower capex, and it is the wrong primary tool for streams dominated by organics, oil, or non-ionic silica at saturation, where activated carbon, ion exchange, or a thermal brine concentrator is a better fit [S3][S5].
For procurement teams, the disqualifying RFQ answer is a vendor that quotes capacity without a feed analysis in hand.
Cost drivers, lifecycle, and CIP discipline
Industrial RO pricing depends on far more than the skid itself: pretreatment, monitoring, automation, materials of construction, membrane replacement interval, and CIP-versus-offsite cleaning all swing total cost of ownership by 30–60% over a 10-year horizon [S3]. Membrane life of 3–5 years is normal on well-pretreated BWRO, dropping to 2–3 years on SWRO or fouling-prone feed, with element replacement often representing 10–20% of annual operating cost [S4][S6].
Clean-in-place frequency is the leading indicator of pre-treatment health: a well-tuned BWRO runs CIP every 3–6 months; a poorly tuned one runs it monthly and still loses permeate flow between cycles [S3]. Offsite membrane cleaning by a service vendor is the cheaper option for skids below 50 m3/day, while larger plants justify a dedicated CIP skid with heated RO permeate flush and vendor-supplied detergent recipes [S3].
For new builds, lifecycle cost comparisons between vendor quotes should normalize to the same feed assumption, the same recovery target, the same flux (LMH or GFD), and the same energy unit cost, otherwise the comparison is meaningless [S1][S6].
Standards, monitoring, and what to write into the PO
Spec language should pin operating pressure in psi or bar, recovery percentage, maximum feed TDS, maximum feed SDI15, free chlorine tolerance, permeate flow at reference temperature (25°C), and element model code with quantity and pressure-vessel layout [S4]. Monitoring expectations follow industry practice: feed conductivity, permeate conductivity, interstage pressure, and differential pressure per vessel, with a 4–20 mA plus HART or Modbus output to the plant DCS or pressure transmitter gateway [S1][S4].
For seawater or high-recovery builds, ASTM D4516-style element integrity testing and vendor FAT data on rejection at design point should be referenced; for sanitary or pharmaceutical use, element material certificates and sanitization compatibility (typically pH 1–13 short-term, 2–11 long-term for polyamide TFC) belong in the quality plan [S4][S6].
Trackable signals over the next procurement cycle: whether vendors offer brackish elements rated for 600+ psi (energy-saving BWRO), and whether feedwater shortages are pushing more municipal tenders toward 80%+ recovery with two-pass designs [S6]. Site decisions for buyers evaluating industrial water treatment skids should watch both moves, because the membrane spec on a 2026 RFQ will differ materially from one written before feedwater costs doubled in their region.
Spec-level background on the components involved: linear guide.