The ppb problem: why ultra‑pure steam makes or breaks modern turbines

In high‑pressure steam cycles, “good enough” purity isn’t good enough. One documented case of silica and salt carryover into a 30 MW unit cost more than 5% of generating capacity in a matter of months (watertechnologies.com). Meanwhile, residual chlorides or sulfates—even at sub‑10 ppb (parts per billion)—can trigger stress‑corrosion cracking (SCC, a corrosion mechanism accelerated by tensile stress) in hot turbine alloys (worldwidepower.cn; power‑eng.com).

Standards have caught up. IEC TS 61370 pegs chloride at ≤3 ppb (gasprocessingnews.com), replacing older conductivity heuristics like “0.2 µS/cm,” which corresponds to more than 10 ppb of contaminants (power‑eng.com). In practice, modern plants now aim for low single‑ppb levels across key species.

The operating target is unambiguous: maintain roughly 99.8–99.9% steam dryness and less than 1–2 ppb dissolved solids to avoid blade fouling or SCC. The path there runs through better separation inside the steam drum and continuous, online purity monitoring—plus design tweaks in the turbine’s final stages to tame the unavoidable wetness that forms at low pressure.

Steam purity limits and turbine risk

Impurity consequences show up fast in performance data and metallurgy. Deposit buildup narrows flow passages and raises heat rate (fuel per MWh), while chloride/sulfate ingress—again, even below 10 ppb—raises SCC risk in high‑temperature alloys (worldwidepower.cn; power‑eng.com). Plants increasingly treat anything above low single‑ppb as a deviation from best practice.

Steam drum separators and scrubbers

Multi‑stage separation inside the steam drum makes ppb‑level purity feasible. Primary devices—cyclone separators or baffle arrangements—strip almost all free water by inducing intense swirl; droplets are centrifuged to walls, coalesce, and drain (researchgate.net). Properly sized cyclone trains (internal or external) can drive outlet moisture to roughly 0.1–0.2%—that is, 99.8–99.9% dry (researchgate.net).

Secondary devices—demisters and scrubbers—finish the job. Wire‑mesh pads or corrugated‑plate demisters coalesce fine mist to push carryover below 0.2% by weight (researchgate.net). Plate‑type “steam washers” and perforated scrubbing sheets use a flowing liquid film to strip the last droplets and soluble salts (e.g., silica) before they can vapor‑phase into the steam (worldwidepower.cn; researchgate.net).

Executed well—adequate steam space height, correct cyclone sizing, baffle layout, and upkeep—these internals cut silica, salts, and solids to the ppb regime. Studies report improved separators achieving around ~0.1% moisture and virtually zero suspended solids (researchgate.net).

Online steam quality instrumentation

Real‑time purity data has become standard, replacing slow grab samples. On‑line systems continuously track conductivity, silica, sodium, and chloride so operators can act—blow down, isolate, or flush—before damage occurs. As one analysis puts it, “online steam purity monitoring provides continuous real‑time data, allowing transient steam purity excursions to be identified and actions taken…wherever a purity change could place the turbine at risk” (researchgate.net).

The instrument suite spans ion‑selective probes or micro‑electrochemical cells for sodium/chloride, silica analyzers using UV or molybdate chemistries, and conductivity—often measured as CACE (conductivity after cation exchange, an analytical step in which the sample is passed through a cation exchanger to strip masking ions). For the cation exchange step, plants employ resin media similar to ion‑exchange resin. Newer techniques such as microfluidic capillary electrophoresis can resolve chloride/sulfate down to sub‑ppb levels (power‑eng.com).

Empirical results underscore the payoff. A geothermal plant uncovered previously untended silica carryover after installing online analyzers, avoiding turbine shutdowns. In thermal and nuclear stations, the same approach is “standard practice…to enable rapid identification of out‑of‑spec steam purity that could lead to turbine damage” (researchgate.net). Continuous trends also support forensics: recorded spikes in carryover can later be correlated with deposition or efficiency loss (researchgate.net).

• Faster detection: Transient upsets—boiler foaming, upstream RO upset, condenser leaks—appear in near real time, not hours later (researchgate.net).
• Quantitative trace data: Trending pH, conductivity (CACE), and total silica supports tighter control. Modern wisdom is that even 2–3 ppb chloride/sulfate is risky, so single‑ppb detection matters (power‑eng.com).
• Maintenance targeting: A rise in steam sodium or silica can trigger timely blowdown or chemistry checks to avert deposits (researchgate.net).

Installing and maintaining analyzers involves sample conditioning, sensor protection, and panel‑level utilities—typical water treatment ancillaries that keep instruments stable and responsive.

Final‑stage turbine erosion control

By the low‑pressure (LP) end, some condensation is unavoidable. Tiny droplets—often 5–10 µm—can coalesce and, at blade tip speeds, pit leading edges (intechopen.com). Modern LP design counters this on several fronts.

First, blade and stator geometry: thinning stator trailing edges and widening stator–rotor gaps promotes smaller droplet sizes with less damaging inertia (intechopen.com). Second, moisture extraction slots in casings or guide vanes pull off film moisture before it shears into larger droplets (intechopen.com).

Third, selective heating: a small bleed of superheated steam or electrical heaters on stationary blades evaporates thin water films; tests show heating guide vanes “proved most efficient” in curbing erosion (intechopen.com). Finally, materials and coatings matter: leading edges increasingly use erosion‑resistant hard‑facing such as Stellite 6B or titanium alloys, with surface hardening via laser nitriding or induction flame‑hardening to further boost resistance (intechopen.com).

The net effect is tangible. Erosion rates in modern LP stages are far lower than in pre‑2000 designs, extending maintenance intervals; manufacturer data often cite more than 80% improvement with leading‑edge hardfacing for a given moisture level.

Operational takeaways

Hold the line at roughly 99.8–99.9% dryness and less than 2–3 ppb of soluble impurities, and LP wet‑steam damage becomes manageable. That requires steam drum internals that repeatedly demonstrate around ~0.1–0.2% outlet moisture with virtually zero suspended solids, backed by online monitoring that can see sub‑ppb chloride/sulfate (researchgate.net; power‑eng.com). Design changes to LP stages—geometry, moisture extraction, heating, hard‑facing—then limit the erosion that remains (intechopen.com).

The contrast is stark. One cycle turbine avoided a forced outage after installing improved separators and monitors, while a similar unit without those upgrades suffered a 5% derate within months (watertechnologies.com; researchgate.net). It’s why plants now treat “low single‑ppb” as the baseline for reliability—and back that up with separation hardware and sensors watching every second.