The cold‑weather dilemma at CCGTs: heat the inlet or bleed the compressor?

When cold, humid air meets a gas turbine intake, physics takes over. Anti‑icing is generally needed when the inlet sits around –5 °C to +5 °C with more than 70% relative humidity, according to Camfil (www.camfil.com). In practice many operators trigger anti‑ice when ambient drops below about 4 °C (39 °F) under moist conditions (www.turbomachinerymag.com).

The risk is not just frost on a grille. Airflow through filters can cool adiabatically (temperature drops as pressure falls without added heat) and condense or freeze. Ice accumulation raises filter pressure drop (Δp, the static pressure difference across a component) and “inlet suction,” cutting turbine output (www.turbomachinerymag.com; www.camfil.com). Camfil pegs the penalty starkly: a mere 100 Pa extra Δp costs roughly 0.2% of full‑load power and about 0.1% extra fuel at part‑load (www.camfil.com).

Left unchecked, icing can escalate into forced outages. Filters or nozzles can clog, prompting emergency trips or damage (www.turbomachinerymag.com; www.nedermanmikropul.com). The fix is simple in concept: keep intake surfaces above freezing or warm the incoming air—typically by ≥7–8 °C over ambient (www.camfil.com; www.researchgate.net).

Heating‑coil intake systems (hot fluid or resistive)

One mainstream approach is to heat the airstream or critical surfaces with hot‑fluid coils in the filter house. OEMs often supply hot‑water/glycol coils upstream of filters and inlet guide vanes (IGVs, adjustable vanes that set the angle of air entering the compressor) (www.turbomachinerymag.com). When anti‑ice is demanded, warm glycol—sourced from a steam condenser, a process boiler or an electric boiler—circulates through finned coils to preheat incoming air. The advantages are simplicity and controllability: the system heats only the inlet flow or surfaces, and it raises air temperature without adding humidity (www.turbomachinerymag.com; www.researchgate.net). In combined‑cycle plants (CCGTs, which pair a gas turbine with a steam cycle), available waste heat—[warm condenser cooling water](https://beta.co.id/en/blog/the-dirty-secret-of-clean-megawatts-why-ccgt-cooling-water-chemistry-makes-or-breaks) or even high‑pressure turbine inlet steam—can often feed the coil loop nearly for “free” (www.researchgate.net; www.researchgate.net).

The trade‑offs are tangible. Adding coils upstream increases pressure loss and flow resistance; Turbomachinery Int’l (TMI) notes coils “will add a pressure loss to the filter house” and are exposed to dirty air, so fouling and maintenance rise (www.turbomachinerymag.com). Coils also call for ancillary hardware—pump, heat exchanger, glycol tank, plumbing—and footprint, adding capital to a system used only during cold spells (typically a few percent of operating hours) (www.turbomachinerymag.com). In practice, designers size to lift the entire intake flow by about 8 °C or more (www.camfil.com). A back‑of‑the‑envelope estimate for required heat power is mass flow × c_p × ΔT (c_p is air’s specific heat at constant pressure), and Sammak (GE) reported typical loads: a Siemens SGT‑800 needed roughly 683 kW to prevent icing versus about 491 kW for an SGT‑700 (www.researchgate.net).

• Operation: Heat intake air via finned hot‑fluid coils (often glycol‑water) or embedded resistive mats (www.turbomachinerymag.com; www.turbomachinerymag.com).
• Advantage: Simple design; any hot‑water source (waste heat) can drive it; only warms air (no added moisture) (www.turbomachinerymag.com).
• Drawback: Adds static pressure loss to inlet (e.g., 0.5–2% additional drop) (www.turbomachinerymag.com); coils foul and require pump, fluid‑glycol system; adds footprint and cost for seldom use.

Compressor‑bleed air heating (gas‑film heating)

The alternative is to route hot compressor‑discharge air back to the inlet. A small bleed from early stages or a discharge tap is injected into the intake tract, or fed through internal heating galleries, warming filters, IGVs, and duct walls. That bleed is hot and dry (typically ~300–400 °C after the compressor) and can mix with intake air or pass through drilled passages—“gas‑film heating,” essentially the reverse of turbine blade cooling (www.turbomachinerymag.com). Hardware can be as simple as a control valve and a distribution manifold to inject bleed evenly into the filter house or plenum (www.turbomachinerymag.com; www.researchgate.net).

The appeal is targeted heating: only ice‑critical surfaces are warmed, avoiding broad filter heating and extra miscellaneous pressure drop (www.turbomachinerymag.com). The hot air is inherently clean and nearly saturated (little extra moisture), and response is fast. Upfront costs can be low. The drawback is material: any bled mass flow can’t reach the combustor. Sammak reports a 2–5% net power loss with compressor‑bleed anti‑ice; a 400 MW gas turbine could give up 8–20 MW when active, depending on ambient and bleed fraction (www.researchgate.net). For context, 100 Pa of extra filter Δp costs about 0.2% power (www.camfil.com); bleed impact can be an order of magnitude larger.

There are practical snags, too. Modifying IGVs or ducts to accept bleed passages is complex and expensive, and spares become custom (www.turbomachinerymag.com). Downstream filters can be blown with hot dust, shortening life (www.turbomachinerymag.com). Bleed valves can be noisy and may require silencing (www.researchgate.net). Unsurprisingly, TMI notes “gas‑film” systems are “rarely used in stationary gas turbine installations” (www.turbomachinerymag.com). Some combined‑cycle plants instead recirculate exhaust flue gas through a heat‑exchanger or an AIMS‑like separator system, but that is relatively uncommon in simple‑cycle format.

• Operation: Divert compressor discharge air (usually a few percent of flow) into inlet/manifold or via drilled IGV passages, heating surfaces before compression (www.researchgate.net; www.turbomachinerymag.com).
• Advantage: Only heats ice‑sensitive surfaces; minimal added hardware beyond valve/manifold (www.turbomachinerymag.com); no reliance on boilers or pumps.
• Drawback: Sacrifices combustion mass flow, causing a 2–5% net power loss (www.researchgate.net); complex modifications increase cost; risk of filter clogging from “puffing” dust (www.turbomachinerymag.com).

Performance trade‑offs and outcomes

Choosing coils versus bleed is ultimately an engineering and economic trade‑off. Published analyses, including Sammak’s GE thesis, confirm that bleed‑air designs inflict a larger performance penalty than coil/heater solutions (www.researchgate.net). For example, Sammak found the bleed‑system on an SGT‑700 incurred more thermal loss than a water‑coil system on an SGT‑800 (www.researchgate.net). In practical terms, a CCGT that heats its inlet with waste heat pays only a small parasitic pump or boiler load, while a bleed design directly reduces megawatt output. TMI’s review echoes this: bleed anti‑ice is “least expensive” to install but its “biggest disadvantage” is the 2–5% output loss (www.researchgate.net), whereas coil systems simply lose a few kilowatts in pump/blower power.

Figure‑like comparison: raising intake air 8 °C (say –2 °C to +6 °C) might require only a few hundred kW of heating—Sammak’s 491–683 kW examples make the point (www.researchgate.net)—which, for a 100 MW turbine, is less than 1%. In contrast, bleeding 3–5% of inlet flow can cost several megawatts. Thus in cold climates or where waste heat is free, hot‑fluid coils tend to yield higher net plant output. By contrast, isolated peaking units or very fast‑start machines might lean toward bleed air for simplicity if a derate is acceptable (www.researchgate.net; www.researchgate.net; www.turbomachinerymag.com).

International practice reflects those choices. In combined cycles, it is common to utilize low‑pressure steam or warm condensate to heat inlet water loops—the “most suitable and economical” choice, per Sammak (www.researchgate.net; www.researchgate.net). In simpler open‑cycle engines in frigid climates, OEMs often supply inlet heaters or electrical mats. Camfil’s guide suggests actively heating to +7–8 °C to fully avoid filter icing (www.camfil.com), whereas bleed air must be carefully controlled and may not suffice alone if extreme icing loads occur.

Implementation and regional context

On implementation, coil systems require fluid loops (glycol brines or steam) and fail‑safe controls; bleed systems integrate with compressor control. Both use sensors (temperature/humidity) and logic—typically an anti‑ice valve opens (coil pump starts or bleed opens) when humidity and temperature cross thresholds, such as ambient below about 4 °C combined with a high dew point (www.turbomachinerymag.com; www.camfil.com).

In tropical regions like Indonesia, icing events are rare; however, international‑standard plants are often specified for 0 °C and may include anti‑ice hardware “as‑built.” There are no Indonesian regulations specifically mandating anti‑icing, so designers default to manufacturers’ guidelines or international standards (e.g., API and IEC guidance for cold climates). In practice, Indonesian CCGTs usually omit elaborate anti‑ice if ambient never requires it; OEMs may still install basic measures if the unit could operate in seasonal highlands or be relocated.

Bottom line

Heating coils and compressor bleed each prevent ice, but not at equal cost. Raising the intake by about 8 °C with coils can preserve nearly full generation for less than 1% equivalent power, with only small parasitic loads (e.g., 491–683 kW in reported cases) (www.camfil.com; www.researchgate.net). Bleed air is simpler to install but typically “costs” 2–5% in lost turbine power—often tens of megawatts on utility‑scale units (www.researchgate.net). Climate, plant configuration, and waste‑heat availability decide the winner.

Sources: Industry and research publications (Turbomachinery Int’l, turbine OEM papers) report specific power impacts and design pros/cons (www.turbomachinerymag.com; www.researchgate.net; www.camfil.com). Academic studies (Sammak 2006) quantify heating loads and output loss (www.researchgate.net; www.researchgate.net). Safety knowledge‑base articles (Camfil, Nederman) give operating thresholds and performance effects (www.camfil.com; www.nedermanmikropul.com).