Raw water — whether from deep wells, rivers, brackish sources, or the sea — carries two categories of solids that anyone designing a treatment plant has to deal with separately:
- Total Suspended Solids (TSS), particle size 0.01–10 microns
- Total Dissolved Solids (TDS), particle size <0.01 microns
TSS particles float in water because they’re negatively charged and repel each other; settling them on their own is slow and unreliable. The standard approach is coagulation–flocculation–sedimentation–filtration: a coagulant (PAC, alum, or FeCl₃) neutralizes the surface charge, a polymer bridges the particles into flocs, a clarifier drops them out, and a media filter polishes what’s left.
TDS is a different problem. Dissolved salts can’t be settled — they need either an osmotic pressure-driven membrane or a phase change to separate them from water. Indonesian seawater TDS is typically 30,000–40,000 mg/L, well and river water 300–1,000 mg/L. Permenkes No. 492/2010 sets the drinking water TDS limit at 500 mg/L. Three desalination technologies handle the gap between source and target: distillation, reverse osmosis (RO), and ion exchange demineralization or demin plant.
Three desalination paths and when each fits
Distillation lowers TDS by evaporating water — salts stay behind as concentrate and steam condenses as pure water. The energy bill is steep at 10–15 kWh/m³ without heat recovery; multi-stage flash and multi-effect distillation drop that to 3–5 kWh/m³, but RO still wins on energy at 2–4 kWh/m³ for seawater. Distillation now mostly survives where waste heat is essentially free — captive plants with surplus low-pressure steam, for example.
Reverse Osmosis (RO) reduces TDS by reversing the osmotic flow. In natural osmosis, water moves spontaneously from low-TDS to high-TDS through a semipermeable membrane until equilibrium. RO inverts this: a high-pressure pump pushes water from the high-TDS side to the low-TDS side at a pressure exceeding the system’s osmotic pressure. Working pressures run 8–15 bar for brackish water, 55–70 bar for seawater.
Demineralization with ion exchange is still relevant but its position has shifted to polishing downstream of RO rather than as the primary salt-removal step. For low-TDS feed (under 300 mg/L), a standalone demin train can be cost-competitive; above that, RO + mixed bed wins on regenerant consumption by a wide margin.
Membrane materials and rejection
Two membrane chemistries dominate: thin film composite (TFC) polyamide and cellulose acetate. TFC polyamide is the workhorse — higher temperature tolerance (up to 45°C continuous), wider pH range (2–11 for cleaning), and 99.5–99.8% NaCl rejection in standard test conditions. Cellulose acetate survives in niches where chlorine tolerance matters more than high rejection, but it caps out around 95–98% rejection and 30°C operation.
The effective pore size of a TFC polyamide membrane is around 0.0001 microns. Field rejection of 95–99% TDS depends on inlet TDS, temperature (rejection drops about 3% per 10°C rise), pH (CO₂ and boron pass more easily at low pH), and the array design.
Membrane selection follows the source water:
- Seawater RO (SWRO) for seawater above 20,000 mg/L TDS
- Brackish Water RO (BWRO) for brackish water 500–10,000 mg/L
- Tap Water RO (TWRO) for clean water under 500 mg/L
The major suppliers — Filmtec, Toray, Hydranautics — offer high-rejection, high-flux, and low-energy grades. Sizing engineers usually run vendor projection software against the raw water analysis to pick the combination that fits the application.
System layout and key design parameters
A simple RO system runs: feed pump → cartridge filter (typically 5 µm) → high-pressure pump (HPP) → membrane and vessel housing → permeate to storage, concentrate to disposal. Add antiscalant dosing, sometimes sodium bisulfite for de-chlorination, and a Cleaning-In-Place (CIP) module for periodic membrane cleaning.
Three parameters set lifetime and operating cost:
- Permeate flux — product flow per unit membrane area (L/m²/h or LMH). Design flux for SWRO is 12–15 LMH; BWRO 18–25 LMH. Pushing flux higher accelerates fouling and shortens membrane life.
- Recovery rate — permeate flow divided by feed flow. SWRO typically runs 35–50%, BWRO 70–85%. Higher recovery saves feed water but concentrates the reject stream and raises the scaling risk.
- Pretreatment — media filtration, ultrafiltration, and antiscalant dosing. A properly designed pretreatment train holds feed SDI (Silt Density Index) under 5 (ideally under 3) and keeps the reject-side LSI (Langelier Saturation Index) negative.
When flux and recovery sit in the recommended range and pretreatment is sufficient to prevent fouling and scaling, membranes last 5–7 years instead of 2–3.
Failure modes and routine maintenance
Three failure modes dominate: scaling (CaCO₃, CaSO₄, silica precipitation on the reject side), biofouling (biofilm growth), and particulate fouling. The signatures show up in normalized trending data: differential pressure across a stage rising more than 15% from baseline points to particulate fouling. Permeate flow dropping more than 10% at constant pressure points to fouling or scaling. Salt passage rising more than 15% suggests O-ring failure or membrane damage.
Cleaning with acid (citric or dilute HCl for inorganic scale) and alkaline solutions (NaOH with surfactant for biofilm and organics) recovers performance if done before fouling hardens. Frequencies of 1–4 cleanings per year are typical, depending on pretreatment quality.
Choosing a system that lasts
A good RO system isn’t the cheapest one to buy. The honest math has to include operating cost (electricity, antiscalant, CIP chemicals, membrane replacement), tolerance for fluctuating raw water quality (rainy vs dry season for surface water, or seawater intake leaks), and access to local technical service. Energy recovery devices — pressure exchangers or turbochargers — on SWRO save 30–50% of HPP power and pay back in two to three years.
Beta Pramesti designs and operates RO packages from domestic-scale at hundreds of liters per day up to industrial SWRO at thousands of m³ per day, with antiscalant chemistry, pretreatment design, and scheduled CIP service. For your specific case — raw water analysis, production target, footprint constraints — get in touch with the engineering team. See you in the next article on demin plants.