Condensate polishers in HRSG cycles are being run to ultra‑pure specs with real‑time cation conductivity and sodium alarms, backed by lab checks and EPRI‑style action thresholds. The goal: keep CPU outlet water at ≤0.2 μS/cm CACE and ≤2–3 ppb Na — or act fast.
The condensate polishing unit (CPU) exists to scrub ionic and particulate contaminants before returning condensate to the boiler/HRSG (heat recovery steam generator) feedwater. Industry guidelines all target ultra‑pure polisher effluent, with EPRI “Normal” limits at sodium <2–3 ppb Na and cation (acid) conductivity (CACE, a conductivity reading after passage through a strong acid cation exchanger) <0.15–0.2 μS/cm (pdfcoffee.com). ChemTreat highlights the same control points for CPU or post–RO/Mixed‑bed water — ≤0.2 μS/cm CACE and ≤2 ppb Na (chemengonline.com) (chemengonline.com).
Silica is controlled, too — typically <10 ppb to protect turbines (pdfcoffee.com) (chemengonline.com). Leading plants set even tighter targets; Eskom, for example, specifies ~0.5 ppb Na (pdfcoffee.com). Maintaining these targets supports efficiency and corrosion control; sodium traces in steam — even at µg/L (ppb) levels — can trigger stress‑corrosion in high‑pressure turbines (mena.hach.com).
In practice, typical guideline values are CACE ≤0.2 μS/cm, sodium ≤2 ppb, silica ≤10 ppb (pdfcoffee.com) (chemengonline.com). Excursions beyond Action Levels trigger corrective action: plan regeneration or shutdown if sodium exceeds ~4–6 ppb (Action Level 1) with a fix within one week, and take action within 24 hours if >8–12 ppb (Action Level 2) (pdfcoffee.com) (pdfcoffee.com).
At the equipment level, a modern condensate polisher deploys ion exchange to “polish” steam condensate after heat exchange cooling. Mixed resin quality matters to hit these low ppb and sub‑µS targets; many operators specify high‑purity ion exchange resins and, when applicable, mixed beds that deliver very low TDS and silica. For context, mixed resins that separate during regeneration provide less than 20 ppb silica and very low TDS in mixed‑bed service.
Outlet purity targets and action levels
EPRI’s “Normal” CPU outlet limits — sodium <2–3 ppb, CACE <0.15–0.2 μS/cm — anchor the control plan (pdfcoffee.com). ChemTreat’s operating advice aligns on ≤0.2 μS/cm CACE and ≤2 ppb Na (chemengonline.com) (chemengonline.com). Silica is typically held at <10 ppb (pdfcoffee.com) (chemengonline.com). Plants that pursue tighter controls cite examples such as Eskom’s ~0.5 ppb Na specification (pdfcoffee.com).
Action Levels organize responses: schedule regeneration within one week at ~4–6 ppb Na (Action Level 1) and intervene within 24 hours at >8–12 ppb (Action Level 2) (pdfcoffee.com). The corrosion stakes are explicit; µg/L sodium in steam is enough to initiate stress‑corrosion in high‑pressure turbines (mena.hach.com).
Online analyzers at CPU outlet
Cation conductivity (CACE) is the most sensitive indicator of ionic ingress. A low‑range conductivity meter paired with a strong‑acid cation column removes ammonia and amplifies acid contaminants (pdfcoffee.com) (wjf.ca). Typical practice uses an online CACE meter with automatic temperature compensation; a degas unit strips CO₂. Instrument range covers 0–1 μS with ~0.001 μS resolution, with alarms near the target (e.g., 0.2 μS). By comparison, passive specific conductivity ≤0.1 μS is used for ultra‑pure RO polishers (chemengonline.com). Ancillary sample conditioning hardware (e.g., degassing, holders, panels) is typically supplied as water‑treatment ancillaries.
Because trace sodium is a critical contaminant, plants deploy an online Na⁺ analyzer. Options include ion‑selective electrode or colorimetric units with on‑line pH adjustment — for example, diisopropylamine dosing acted upon by a dedicated dosing pump — to accelerate the sodium response. Multiple sources stress sodium’s importance: “the only way to monitor sodium correctly is by using online instruments” (wjf.ca), and leaks show up “hours earlier” in Na than in conductivity (wjf.ca). Na alarms are typically set just above the target (e.g., 2–3 ppb), since even µg/L in steam is hazardous (mena.hach.com).
Online silica analyzers (e.g., UV/molybdate methods) add coverage for carryover leaks; if not installed, silica is sampled offline frequently. Because silica distills with steam, targets remain <10 ppb (pdfcoffee.com) (chemengonline.com). Dissolved oxygen (DO) is tracked continuously; a DO probe flags condenser air ingress, with sudden DO >20 ppb suggesting a leak. Continuous ORP/pH sensors are also used, especially where copper heaters are present. Recommended feedwater DO is <5–10 ppb (chemengonline.com).
All analyzers report to DCS/SCADA (distributed control/supervisory control and data acquisition). Alarms are configured at target and action levels — for example, Na >6 ppb or CACE >0.2 μS — and trends are logged continuously. Rising CACE or Na over days often foretell resin exhaustion. Instruments are calibrated regularly (daily or per shift) with standards. Vendors such as SWAN/Hach/WJF supply factory‑calibrated modules built for 24/7 power‑plant service (wjf.ca) (wjf.ca). In many plants, CPU monitors sit on the feedwater quality display; any spike prompts an inspection for regeneration need. Resin performance is tracked alongside quality, with attention to remaining capacity and breakthrough behavior in systems supplied with high‑grade ion‑exchange resin.
Offline validation and lab checks
Online alerts are validated by periodic lab work. Ion chromatography (IC) — an analytical technique that resolves ions in the low‑ppb range — is run monthly (or weekly if issues) on polisher effluent for anions (Cl⁻, SO₄²⁻) and cations (Na⁺, K⁺, NH₄⁺), yielding precise sodium and chloride confirmations. ICP/OES (inductively coupled plasma/optical emission spectroscopy) is used monthly or quarterly for metals (Fe, Cu, Mg, Ca) at low ppb, checking for resin breakthrough or corrosion products.
Silica (SiO₂) is analyzed offline weekly (e.g., molybdate method). Potentiometric titration covers alkalinity/hardness, which should be at “zero” in an effective CPU. TOC (total organic carbon, UV‑persulfate) is run periodically if organic ingress is suspected. A handheld or benchtop conductivity cross‑check (15 mL cell) validates the online sensor.
Lab results are compared to targets; for instance, offline Na from IC >3 ppb or silica >10 ppb — both outside target ranges (pdfcoffee.com) (chemengonline.com) — triggers a chemistry review and potentially regeneration. Lab data also calibrates the analyzers; if IC Na runs higher than the online unit, the analyzer’s calibration factor is adjusted.
QC thresholds, alarms, and data review
All online and offline data feed a QC database with weekly trend charts for Na and CACE. Approaching Action Level 1 — sodium trending above ~4 ppb — leads to scheduling resin regeneration within one week (pdfcoffee.com). Exceeding 8–12 ppb Na (Action Level 2) requires action within 24 hours (pdfcoffee.com). Sharp jumps in conductivity or DO prompt immediate investigation for a possible condenser leak.
The CPU quality plan documents thresholds explicitly — for example, a CACE alarm at 0.2 μS/cm and a Na alarm at 5 ppb — with clear response procedures. Over time, a well‑run CPU shows ≥95% of Na readings <3 ppb and CACE <0.2 μS (pdfcoffee.com) (chemengonline.com). Deviations — multiple high‑Na events, for instance — prompt root‑cause analysis, including resin fouling checks or condenser leak inspection, and may lead to performance tests that measure remaining resin capacity or “breakthrough” curves. Where resin separation and regeneration are required, plants often specify dedicated mixed‑bed trains for ultra‑low silica and TDS.
Measured outcomes and operating flexibility
The payoff is quantifiable purity and reduced corrosion risk. If makeup water sodium is ~500 ppb (typical raw RO feed), achieving 0.5 ppb at the CPU equates to >99.9% Na removal — a target cited by plants (pdfcoffee.com). Holding ≤0.2 μS CACE (versus, say, 2.0 μS without polishing) is a tenfold reduction in ionic load. EPRI notes that condensate polishers can enable operation through minor condenser leaks that might otherwise force shutdown (powermag.com).
Key metrics include fraction of time in spec — for example, “Na <3 ppb in 98% of samples” — regeneration frequency (days between resin changes), and correlations to boiler feedwater iron or hardness. A mature trendline might show that after instituting the QC plan, polisher outlet Na never exceeded 2 ppb and renewal intervals extended by 30%. The overarching aim is consistent in cited guidance: keep polished condensate as pure as or purer than makeup water. One source notes that with appropriate controls, steam cation conductivity can be safely held at <0.15 μS/cm even in high‑pressure units (powermag.com).
Reference basis for thresholds
Modern guidelines and case studies underpin the numeric targets: EPRI and allied guidance via pdfcoffee.com, ChemTreat’s monitoring advice via chemengonline.com, and instrument practice via wjf.ca. For example, ChemTreat summarizes core monitoring targets — CACE ≤0.2, Na ≤2 ppb, SiO₂ ≤10 ppb — in chemengonline.com and chemengonline.com, and EPRI explicitly defines target and action levels for Na and conductivity at the polisher outlet (pdfcoffee.com) (pdfcoffee.com). These form the foundation of the QC plan’s thresholds and alarm strategy and are implemented through analyzers and ancillary gear often packaged as supporting equipment for water treatment.
