Cold Air, Hot Debate: Inside the Two Ways CCGTs Stop Icing Before It Stops Them

In humid, cold weather, gas turbine inlets are vulnerable to hoarfrost, sleet, snow, freezing fog — and to a less obvious risk: icing that begins inside the ductwork even when ambient air is above 0°C. As moist intake air accelerates through filters and ducts, adiabatic expansion can drop the temperature below the dew point, triggering condensation or ice that spikes pressure drop and can force a shutdown or damage the machine (Turbomachinery Magazine; Nederman MikroPul).

Industry guidance converges on a simple target: add about 8°C of heat (ΔT, temperature rise) to the incoming air to keep relative humidity (RH) under ~68% and prevent condensation. That 8°C lift (≈Δ12°F) typically translates to a 10–15% heat input for moist air; as a rule of thumb, ~0.2 g H₂O/m³ adds ~10% heating load (Nederman MikroPul).

Two anti‑icing architectures

Engineers generally pick between two approaches: heat-exchanger coils in the inlet plenum, or compressor-bleed hot air injected upstream. Both aim to raise inlet temperature, but they pay for it differently — one with added pressure drop (ΔP), the other by diverting mass flow.

Heated coil systems (steam, glycol, electric)

In coil systems, ambient air flows across a heat exchanger supplied by steam, hot water, glycol, or electric resistance. Coils add a finite inlet pressure drop, typically a few hundred pascals (Pa; the pascal is the SI unit of pressure). Data show ~180 Pa of inlet ΔP costs ~0.27% of turbine power; a full coil train can impose 200–500 Pa, a few tenths of a percent of output (ResearchGate; Nederman MikroPul).

Heat duty scales with mass flow. For 100 kg/s of air and an 8°C rise, Q̇ is roughly 0.1 MW per °C, or ≈0.8 MW of thermal energy. Large plants often supply this with bleed-exhaust steam, waste heat, or a dedicated boiler. Coils require pumps or piping for the heating fluid and add maintenance touchpoints (piping, coil fouling, potential leaks). The upside: no combustion air is diverted, so the performance hit is mainly the frictional ΔP and modest pump power. Vendors report over 1,300 inlet-heating coils (serving 30–790 MW plants) installed worldwide since 2000, and systems are designed to heat from far below freezing (≈−50 to −60 °C) to safe temperatures (Super Radiator Coils).

Because heat is applied before filters, coils protect both the filter bank and compressor inlet from icing (Turbomachinery Magazine). Designs prioritize preventing ice formation rather than melting ingested snow, since melting requires high latent heat and can be uneconomic (Nederman MikroPul).

Compressor‑bleed hot air systems

Bleed systems tap a portion of hot, compressed air from the compressor discharge (typically ~250–350 °C at ~10–20 bar for large frames) and inject it into the inlet via a manifold and nozzles. That recirculated heat lifts inlet temperature without an external boiler. Many large frames (e.g., Siemens W501F, GE 7FA, Alstom 11N2) offer bleed anti-icing as an option (Nederman MikroPul).

The advantages are straightforward: negligible added inlet ΔP (no coil obstruction) and simple mechanics (pipes and valves), with logic integrated in controls (Nederman MikroPul). The downside is power: any air diverted for heating cannot do work in the combustor. Industry sources estimate up to ~2% output reduction for typical bleed anti-icing; diverting ~3–5% of compressor flow might deliver an ~8°C rise and a roughly 2% net output loss (Camfil). Nozzle jets also generate noise (≈85 dBA at 1 m), so silencers may be required (Nederman MikroPul).

Control and safety are central. The system must shut off or bypass in warm weather to avoid overheating, and OEMs limit the maximum bleed fraction — often 5–10% — to avoid surge or performance issues. Notably, these systems are not exhaust-gas recirculation; combustion chemistry is unchanged. If ventilation air were ever used as a hot source instead (not typical for bleed), hydrocarbons could be entrained (Nederman MikroPul).

Pressure drop versus power diversion

It’s ultimately a trade: [coils impose a permanent inlet ΔP penalty](https://beta.co.id/en/blog/the-coldweather-dilemma-at-ccgts-heat-the-inlet-or-bleed-the-compressor), while bleed systems divert mass flow. Measured results show ~180 Pa of added ΔP can cost ~0.27% of power; a coil’s 200–500 Pa costs a few tenths of a percent. By comparison, bleed anti-icing often costs ~1–2% of power. Coils need an external heat source; bleed uses “free” compressor heat, but with direct derate (ResearchGate; Camfil).

A 100 MW case in numbers

For a 100 MW gas turbine in snowy conditions, lifting the inlet by 8°C at full flow — say 300 kg/s — needs ≈2.4 MW of heat (300 kg/s × 1.005 kJ/kg·K × 8 K). A bleed arrangement might do this by diverting ~5% of compressor flow, at a roughly 2–3% output cost. A steam-coil system would deliver 2.4 MW with a ~0.3–0.5% loss largely from inlet ΔP (ResearchGate; Nederman MikroPul).

The business call hinges on operating profile: if the machine frequently runs in cold, humid air, the hour-by-hour derating from bleed heating may outweigh coil capex and utilities.

Market adoption and hybrid use

Coil-based inlet heating is common in new builds or retrofits where external or waste heat is available — with more than 1,300 installed across 30–790 MW units since 2000 (Super Radiator Coils). Compressor-bleed options are standard offerings for many large baseload frames in cold regions of Europe, North America, and at high altitudes (Nederman MikroPul). Some operators employ hybrid tactics: low bleed in mild conditions, coils in severe icing.

Design guidelines and controls

• Heat duty: design for about an 8°C lift in the worst-case moisture to keep RH ≈68% and avoid condensation. Higher lift is generally only required if intentionally melting ingested snow, which is usually not economical (Nederman MikroPul).
• Bleed sizing: compute the required bleed fraction via energy balance. Tapping 3–5% of flow from the last compressor stage (roughly 500–800 °F, ~100–200 psi air) typically yields a few °C of inlet heating; OEM limits, often 5–10%, set the ceiling (Nederman MikroPul).
• Coil sizing: select coil face area and fin configuration for target ΔT at design flow; quantify ΔP and pumping power. Keeping ΔP in the ~180–500 Pa range typically limits output loss to <1%. Use corrosion-resistant materials suited to cold, moist inlet locations (ResearchGate; Turbomachinery Magazine; Nederman MikroPul).
• Controls: automate anti-icing based on both ambient temperature and humidity/dew point; modern systems modulate to maintain ambient+8°C (±3°C) for performance (Nederman MikroPul).
• Safety and standards: Indonesia’s tropical climate seldom needs anti-ice, but high-altitude or maritime sites should check local temperature/humidity records. API 614/616 guidance notes inlet heating for sub‑5°C conditions. No specific Indonesian regulation applies, so IEC/ISO or OEM guidelines are used.

Bottom line for CCGT operators

Inlet heating coils deliver precise temperature control with a minor (<1%) power penalty dominated by added ΔP, but they need auxiliary heat and space. Bleeding hot compressor air is mechanically simple and avoids inlet ΔP, yet typically costs ~1–2% of turbine power and adds noise. Site weather data, available heat sources, and acceptable derating determine the choice — with the common goal of 8–10°C inlet uplift to prevent filter icing, avert unscheduled outages, and avoid efficiency loss from ice-induced pressure spikes (ResearchGate; Camfil; Turbomachinery Magazine; Nederman MikroPul).

Sources: turbomachinery industry guides and case studies (Turbomachinery Magazine; Nederman MikroPul; Nederman MikroPul; Camfil), vendor data (Super Radiator Coils; Camfil), and technical papers on inlet pressure versus power (ResearchGate).