The boiler makeup water blueprint: how high‑pressure plants hit ultra‑pure specs

In power generation, the makeup water spec sets the tone. High‑pressure boilers require extremely pure feed; even minor turbidity or hardness can trigger costly scale and corrosion. The resulting treatment train is disciplined: clarification to strip solids and organics, dual‑media filters to polish to low NTU, a primary demineralizer — either traditional ion exchange (IX) or reverse osmosis (RO) — and a final polisher to push conductivity into single‑digit micro‑siemens and silica into trace parts‑per‑billion.

The stakes are quantified in vendor and industry guidance. Typical coagulation can eliminate >80–90% of turbidity and color (ionexchangeglobal.com), with clarified water often reduced to <5 NTU (NTU: turbidity index) before filtration. Properly designed dual‑media filters routinely produce effluent turbidity <0.5–1 NTU and extend downstream RO run‑times (ionexchangeglobal.com). Most RO vendors specify feed turbidity <1 NTU and iron <0.1 mg/L to prevent membrane fouling.

Raw water clarification and media filtration

Raw water — whether river, well, or process effluent — first goes through clarification (coagulation/flocculation and sedimentation). That front end typically lives in a clarifier, and plants often standardize chemical addition using a dosing pump. Typical coagulation can eliminate >80–90% of turbidity and color, yielding turbidity often reduced to <5 NTU before filtration (ionexchangeglobal.com).

Downstream, dual‑media filtration (anthracite over sand) polishes clarified water by capturing fine particulates and organics without rapid clogging; properly designed multimedia filters routinely produce <0.5–1 NTU effluent (ionexchangeglobal.com). Plants commonly specify sand media for the lower bed and anthracite as the upper layer to maintain low pressure drop and long filter runs.

Where raw hardness or alkalinity is high, chemical softening or cation‑exchange softeners are added before RO or ion exchange to reduce Ca²⁺/Mg²⁺ to near zero (www.ewt-wasser.de). Plants that standardize on softening often integrate a dedicated softener upstream to protect the primary demineralizer.

Clarification hardware varies by site, but a conventional clarifier remains the workhorse for suspended solids removal in the pretreatment step described above.

Deaeration and decarbonation targets

Dissolved gas control is essential: deaeration (steam stripping or vacuum/membrane degassing) cuts dissolved O₂ and CO₂ to very low levels to prevent corrosion and to reduce acid load in demineralizers. Targets cited include ~0.02–0.04 mg/L O₂, with boiler‑feed guidelines calling for <0.007 mg/L O₂ at ≥20 bar (www.lenntech.com) (www.lenntech.com) (www.lenntech.com). A degasser between cation and anion stages is standard: it strips ~90–99% of free CO₂ (to ~10–20 µg/L) and nearly all O₂ (www.power-eng.com) (www.ewt-wasser.de). Absent degassing, even ~10 mg/L CO₂ is enough to prevent meeting boiler makeup specs in an RO/EDI train (www.power-eng.com).

Two‑bed ion exchange train

Traditional ion exchange (IX: swapping ions on resins) uses a two‑column train: a strong‑acid cation in H⁺ form followed by a strong‑base anion in OH⁻ form (www.ewt-wasser.de). The cation stage softens and lowers alkalinity; a degasser vents the resultant CO₂; then the anion tank yields water of ~1 µS/cm or better (roughly 0.1–2 MΩ·cm resistivity). Plants standardizing on IX often specify a packaged demineralizer built around a ion exchange core.

High‑pressure service demands essentially zero hardness (<0.05 mg/L) and tight silica control (<0.2 mg/L) at this stage (www.lenntech.com) (www.engineeringtoolbox.com) (www.engineeringtoolbox.com), with a final mixed‑bed polisher combining cation and anion resins to finish the water. Mixed beds routinely deliver 10–15 MΩ·cm or >0.07 µS/cm; resin management often includes dedicated ion exchange resins suited to strong/weak functional groups.

Reverse osmosis as primary demineralizer

RO (reverse osmosis: pressure‑driven membrane separation) can also serve as the primary demineralizer. Modern power plants increasingly deploy double‑pass RO or RO+EDI to reduce chemical use (www.power-eng.com) (www.ewt-wasser.de). One‑pass RO typically removes ~95–99% of dissolved salts. A two‑stage RO system (with antiscalant and pH adjustment) produced boiler‑makeup water with conductivity <0.07 µS/cm and silica <10 mg/L at a total flow of 24 m³/h (www.eurowater.com).

RO trains are often delivered as integrated membrane systems, with antiscalant dosing handled upstream by a dosing pump. For inland sources, plants commonly standardize on brackish‑water RO as the core unit; after RO, a mixed‑bed or electrodeionization (EDI) polisher is still required to hit the final boiler‑feed spec (www.power-eng.com).

Polishing to ultra‑pure conductivity

Regardless of primary demineralization, a polishing step closes the gap to ultra‑pure water. In chemical‑based plants this is a regenerateable mixed‑bed ion exchanger (combined cation+anion resins in one vessel), producing 10–15 MΩ·cm or >0.07 µS/cm. Typical operating cycles last a few days to a week before the bed is split and regenerated with acid and caustic. Many sites deploy a dedicated mixed‑bed to secure silica and trace ions to ppb levels.

Some modern designs use continuous EDI for polishing, eliminating batch regenerants; a commercial EDI system recently delivered ~14 MΩ·cm water at 24 m³/h capacity (www.eurowater.com). In all cases, the goal is boiler‑feed quality: at 60–70 bar, ASME/ABMA limits list TDS ≤625 ppm and silica ≤0.20 ppm (www.engineeringtoolbox.com) (www.engineeringtoolbox.com). Most coal and combined‑cycle plants actually target much lower levels (often ≪50 µS and silica ≪0.05 mg/L) to avoid any risk of carryover. Ongoing monitoring of conductivity and cation conductivity guards against resin exhaustion or breakthroughs.

Performance and cost trade‑offs

Economics split along familiar lines. RO systems entail higher energy and capital costs but drastically cut chemical use. One study comparing IX to RO (with clarifier–filter pretreatment in both) reported RO demanded ~17.5% more raw water and produced ~150% more concentrate by volume — while that concentrate carried 20–30× less salt (www.researchgate.net). In practice, RO plants usually recover ~70–85% of feedwater; the rest becomes a brine waste. Residual scaling and membrane care are handled by design (softening, antiscalants) and periodic flushing.

IX systems have lower upfront cost but high operating costs: 70–80% of IX expenses go to regenerant chemicals (salt, acid/caustic) (www.researchgate.net). RO plants are less chemically intensive but their O&M is dominated (70–80%) by electricity for high‑pressure pumps and membranes (www.researchgate.net). Membrane replacement is periodic but infrequent. Large plants with very hard or mineralized water often favor RO/EDI, while smaller plants or softer feeds may stay with conventional IX; energy, chemical prices, and water/waste constraints drive the call (www.researchgate.net) (www.researchgate.net).

Design sizing and operations

A well‑designed train includes redundancy and automation. Key metrics include flow capacity, recovery, and final conductivity. A midsize power plant might need 100–1000 m³/day of makeup; each section (filter, demineralizer, polisher) is sized with safety margins. Typical removal efficiencies: >90–99% of turbidity and iron in pretreatment, >90% of hardness/alkalinity in softening, and 95–99% of dissolved solids in RO. Backwash intervals for filters range from hours to days depending on loading; IX regeneration frequency depends on load, with larger two‑bed sets often running weeks between regenerations and mixed beds 1–7 days.

Industry trends favor membrane‑integrated systems to reduce chemicals and footprint — many new plants use RO+EDI with minimal chemical dosing (www.power-eng.com) (www.ewt-wasser.de). In one reported case, RO adoption significantly cut regenerant brine disposal (by ~20–30× less salt in effluent) (www.researchgate.net). The catch: pretreatment must be robust because membranes are sensitive to particulates, iron, and biofouling. Dissolved gases remain critical — membrane vacuum or packed‑tower degassers routinely reduce O₂ to <0.02 mg/L and CO₂ to a few ppm before polishing (www.lenntech.com).

Standardized ancillaries simplify upkeep across the treatment train, from instruments and valves to filter internals; many operators bundle these as water treatment ancillaries and maintain spares through parts and consumables programs.

Standards and verification

Regulations and standards lock the targets. Indonesian practice, mirroring international norms, mandates boiler‑feed water treatment — pretreatment, deaeration, and demineralization — to remove elements causing corrosion and scaling, such that feedwater meets ABMA specifications (id.scribd.com). In practice, finished makeup water must have virtually zero hardness, carryover‑free steam purity, and trace alkalinity. Compliance is verified by periodic testing (silica, sodium, conductivity) to international limits, with design decisions (RO vs IX) tied to measurable targets — e.g., no more than ~0.02–0.05 mg/L silica and ~0.007 mg/L O₂ in the final feedwater (www.lenntech.com) (www.engineeringtoolbox.com).

Sources and references

Published engineering guidelines and industry reports frame the numbers cited here. Boiler feedwater chemistry tables and limits: www.engineeringtoolbox.com, www.engineeringtoolbox.com. Case studies and system designs: www.eurowater.com, www.power-eng.com, www.researchgate.net, www.researchgate.net. Indonesian boiler requirements: id.scribd.com. Pretreatment and dual‑media filtration performance: ionexchangeglobal.com. Boiler feedwater characteristics and dissolved gas targets: www.lenntech.com, www.lenntech.com. Additional IX/RO design context: www.ewt-wasser.de.

Details on specific units (flow rates, conductivity, silica) come from recent case reports and supplier data (www.eurowater.com), while operational sensitivities around degassing and pH adjustment before RO are discussed in www.power-eng.com. The comparative IX vs RO performance and cost splits, including water and brine flows, are detailed in www.researchgate.net and www.researchgate.net. All numerical guidelines and outcomes above are drawn directly from these sources.