Case Study: Inlet Fogging Retrofit for a Coastal Peaking Plant

Two peaking units on a humid coast were handing back capacity every hot afternoon. Published ratings put the output shortfall of a gas turbine, at a hot inlet against the ISO 15 °C reference, at up to 40 percent of nameplate. The site already ran media evaporative coolers ahead of the inlet filters, and those pads were holding only about 80 percent approach to the wet bulb. On an August peak the gap was tens of megawatts across a two-unit site, surrendered at the moment power was dearest. The answer was not more pad area. It was switching the cooling method to high-pressure inlet fogging, where the water cools the air by evaporating inside the inlet duct.

This is the record of that class of retrofit rather than of one plant. The figures below describe the generic shape of a coastal two-unit simple-cycle site, and every number sits inside published ranges for turbine inlet air cooling. The arithmetic is shown so a reader can trace the sizing from a droplet to the feed pump.

Case Snapshot

Figures are the generic shape of this class of coastal two-unit peaking retrofit, not a plant record.

Item	Value
Industry	Simple-cycle peaking power, coastal hot humid climate (anonymized)
Plant class	Two frame-class gas turbine units, nominal 100 MW class each
Original cooling	Media evaporative cooler bank ahead of the inlet filters, about 80 percent approach
Peak inlet condition	Dry bulb about 35 °C, wet bulb about 29 °C, depression about 6 K
Retrofit	High-pressure inlet fogging manifold downstream of the final filters
Injection pressure	About 2000 psi, about 138 bar
Nozzle set	316L stainless fog nozzles, 0.1 to 0.3 mm orifices
Water supply	Reverse osmosis skid, conductivity-controlled demineralized water
Final filtration	5 to 20 micron filter downstream of the RO unit
Droplet target	About 90 percent of flow at 20 microns or smaller
Result band	Humid-climate dry-bulb drop 5 to 6 °C, approach 85 to 95 percent

The four numbers that carry the case:

• Inlet dry-bulb drop: 5 to 6 °C in a humid climate.
• Wet-bulb approach: 85 to 95 percent, against 80 percent for the media pads.
• Injection pressure: about 138 bar through 316L stainless fog nozzles.
• Final filtration: 5 to 20 micron downstream of the RO skid.

The Challenge

The plant was not short of cooling hardware. It was short of wet-bulb depression to work with. A coastal humid site sits at a smaller gap between dry bulb and wet bulb than a desert site, and every evaporative method is bounded by that gap. The installed media pads had been sized against a mild summer and a fresh pad. In service they were holding about 80 percent approach, the ordinary figure for a media evaporative cooler.

The second problem was parasitic load. A media pad bank sits in the inlet stream and pushes the whole air flow through a wetted matrix. That matrix carries a pressure drop, and the compressor pays it every operating hour. The plant was spending that fan penalty on the pads while collecting only 80 percent of a small depression. In the hottest hours the inlet temperature still sat above where the units wanted to run.

The third problem was condition. Media pads foul with dust, salt and biological growth, and approach falls as the matrix loads. A coastal site pulls airborne salt into the duct, which feeds scaling on the pad face. Cleaning and replacement took the inlet out of its design condition on a shrinking cycle. The problem was not one failed part. It was a gas turbine inlet cooling method with no reserve.

Why Hot Inlet Air Costs Output

A gas turbine is a mass-flow machine, and hot air is thin air. As the inlet temperature rises, air density falls, and so does the mass flow the compressor can pass at a given speed. Less mass flow means less power from the same hardware. That is why published ratings express output against an ISO 15 °C reference and show a derate as the inlet climbs.

Published figures put that derate at up to 40 percent of nameplate in hot inlet conditions. The coastal site lived at the hot end of that band through most of the cooling season. Two units at nominal 100 MW class each, every one short by a fraction of the band, is a plant-scale figure on a peak afternoon.

The penalty reaches wider than output alone. At a hot inlet the compressor works harder for the same delivered power, so heat rate rises while power falls. A peak-hour megawatt is the most valuable a peaking plant sells. The hot inlet removed supply when demand and price were highest, the economic reason turbine inlet air cooling exists.

Why the Media Coolers Stalled at Eighty Percent

Approach is the fraction of the available wet-bulb depression that a cooling method actually captures. The dry-bulb drop divided by the dry-bulb depression, times 100, gives the approach in percent. A method that reaches the wet bulb has an approach near 100 percent.

Media evaporative coolers typically deliver about 80 percent approach. That is respectable in a desert where the depression is large, and it is weak in a humid climate where the depression itself is small. Published performance for the media class sits at that figure because the air does not reach full saturation across a wetted matrix in the depth available.

At the coastal site the peak dry bulb sat near 35 °C and the wet bulb near 29 °C, a depression of about 6 K. An approach of 80 percent turns a 6 K depression into about a 4.8 K drop. That left the units still warm on the days that mattered, and the same pad bank in desert air with a 30 K depression would recover far more.

The fan penalty compounded the limitation. The pads bought their 4.8 K at the cost of a pressure drop across the whole inlet stream, and the compressor paid that drop every hour. A cooling method that removes heat without crossing a matrix spends nothing on inlet pressure. That distinction became the core of the turbine inlet air cooling argument.

The Wet-Bulb Ceiling

Every evaporative method is bounded by the wet bulb, and no amount of water beats that limit. The wet-bulb temperature is the floor the air can reach through evaporation alone at constant enthalpy. The distance from dry bulb to wet bulb is the depression, and the depression is set by the weather rather than by the equipment.

This ceiling is why an evaporating cooling method cannot be judged on hardware quality alone. A coastal site with a 6 K depression gives any fog or media system a small budget, and the physics does not care how much water is available.

The gap between good and ordinary methods is approach, not principle. Published performance for a well tuned inlet fogging system puts the approach at 85 to 95 percent of the depression. Media evaporative coolers sit near 80 percent. On a 6 K depression that spread is real but modest in degrees: about 4.8 K for the media and 5.1 to 5.7 K for a fog system at 85 to 95 percent.

The conclusion the plant drew was straightforward. The ceiling was low, so the goal could not be a large temperature drop. The goal was to reach closer to a low ceiling while removing the pressure penalty, and that narrowed the field to one evaporating cooling method.

The Solution

The retrofit changed the cooling method rather than enlarging it. A high-pressure inlet fogging manifold went into the inlet duct downstream of the final filters, on both units. The manifold runs about 2000 psi, about 138 bar, through 316L stainless fog nozzles with 0.1 to 0.3 mm orifices. A reverse osmosis skid feeds the manifold with conductivity-controlled demineralized water, and a 5 to 20 micron filter sits downstream of the RO unit to protect the small orifices.

The reason for high pressure is droplet size. At 138 bar a fog nozzle shears the water into droplets small enough to evaporate within the few metres of duct between the manifold and the compressor bellmouth. Published figures put about 90 percent of flow in droplets at or below 20 microns, finishing evaporation inside the duct so no free water reaches the compressor.

The family choice follows the duty. The manifold is a high-pressure water spray whose only job is evaporation into a moving air stream, which is exactly the class of misting nozzle built for pressure-driven fine atomization. The manifold count, the orifice size and the filtration band then follow from the size of the duct and the water the air can absorb.

Water quality is a hard requirement, not an upgrade. Minerals carried into the compressor deposit on the blades and foul the gas path, so the conductivity of the injected water is the specification that matters. Demineralized water from an RO skid keeps the dissolved load low, so cooling-water blade fouling stays off the plan.

Manifold Placement and Droplet Size

Placement and droplet size are coupled, and neither can be chosen alone. The manifold sits downstream of the final filter so that any particle large enough to reach an orifice has already been stopped. Orifice sizes of 0.1 to 0.3 mm demand filtration in the 5 to 20 micron band, and biofilm and mineral creep are the two failure modes that put a fog manifold out of service.

Evaporation distance is the constraint that sets the placement. Droplet evaporation time scales with the square of the droplet diameter. A 100 micron droplet takes about 100 times as long to evaporate as a 10 micron droplet. That square law is why a coarse spray that looks wet can cross the whole duct and land on the compressor while a fine fog disappears into the same distance.

Residence time has to be read against air velocity. The duct between the manifold and the bellmouth gives the droplets a known window, and the nozzle count is set so the injected water can finish evaporating inside it. Published guidances treat droplet size and spray quantity together because spraying past saturation carries an erosion risk on the blades.

The layout also fixes the pressure drop, and this is the structural advantage. Fog adds near-zero pressure drop to the inlet stream, because the droplets travel with the air instead of crossing a matrix. A media bank and a chiller both add parasitic loss, and a series counterflow chiller arrangement recovers only about 8 percent of chiller system work by published review.

Sizing the Orifice Flow, the Water and the Pump

Three calculations carry the sizing: the flow from one orifice, the water the air can absorb, and the pump that drives the manifold. Each one is shown with its substitution so the numbers can be checked rather than trusted.

Orifice flow leads. Water flow through a plain orifice follows Q in litres per minute equals 0.65 times the square of the diameter in millimetres times the square root of the pressure in bar, for a discharge coefficient near 0.95. The square root of 138 is 11.75. Substituting a 0.2 mm orifice: Q = 0.65 * 0.04 * 11.75 = 0.306 litres per minute. A 0.3 mm orifice gives Q = 0.65 * 0.09 * 11.75 = 0.687 litres per minute. That per-nozzle flow builds the manifold count.

Nozzle count follows from the target flow. At about 0.3 litres per minute per nozzle, a manifold delivering 60 litres per minute on one unit needs roughly 200 fog nozzles in that unit. Across two units the manifold set passes about 120 litres per minute.

The water the air can take up follows the latent-heat balance. The evaporated mass is m equals air flow in kilograms per second times specific heat 1.005 times the temperature drop in kelvin times 3600 divided by the latent heat 2257 kilojoules per kilogram. Substituting 1000 kilograms per hour of air, which is 0.2778 kilograms per second, at a 10 K drop: m = 0.2778 * 1.005 * 10 * 3600 / 2257 = 4.45 kilograms per hour. Published design figures for a working fog system run higher, near 7 to 10 kilograms of evaporated water per 1000 kilograms of air per 10 K, because the practical band carries approach margin and drift above the pure enthalpy floor.

Scaled to the site, a frame-class inlet passes air in the hundreds of kilograms per second, and 400 kilograms per second is taken here for the arithmetic. A 5.5 K drop then evaporates about 400 * 1.005 * 5.5 / 2257 = 0.98 kilograms per second, near 3500 kilograms per hour on a single unit. That is the make-up water the RO skid delivers at peak.

Approach is checked with its own formula. Approach equals the dry-bulb drop divided by the dry-bulb depression times 100. Worked on a desert case, an inlet at 35 °C with a 12 °C wet bulb cooling to 14 °C gives (35 - 14) / (35 - 12) = 21 / 23 = 91.3 percent. At the coastal site the depression is about 6 K, so an approach of 85 to 95 percent yields a drop of 5.1 to 5.7 °C, against about 4.8 °C for the media pads at 80 percent.

The pump closes the loop. Pump hydraulic power is flow times pressure over 612. At 138 bar and 60 litres per minute per unit: kW = 60 * 138 / 612 = 13.5 kW of hydraulic power per unit. Published overall pump efficiency for this class runs 60 to 75 percent, so the shaft demand is 13.5 / 0.70 = 19.3 kW per unit, about 39 kW across two units.

Wear deserves its own line because it drifts a working manifold. A diameter that grows by x percent passes about (1 + x/100) squared minus 1 more flow. Five percent wear gives (1.05) squared minus 1 = 0.1025, about 10 percent over-delivery at the same pressure. That is why the plant tracks manifold flow as the wear signal and not nozzle appearance alone.

The Results

No site audit figures appear in this table. The values are the generic shape of this class of retrofit, anchored to the published ranges below.

Metric	Before: media evaporative coolers	After: high-pressure inlet fogging	Why it moved
Inlet approach	About 80 percent	85 to 95 percent of the depression	Droplets evaporate in the duct rather than across a wetted matrix
Dry-bulb drop at peak	About 4.8 K on a 6 K depression	5.1 to 5.7 K	Higher approach captures more of the same small depression
Cooling pressure drop	Media matrix drop plus fan penalty	Near zero, the droplets travel with the air	No matrix in the inlet stream
Injection pressure	None, wet matrix	About 2000 psi, about 138 bar	Fine atomization needs high pressure
Water quality	Filtered make-up	Demineralized, conductivity-controlled	Minerals carried into the compressor foul blades
Droplet size	Coarse, pad-borne	About 90 percent at 20 microns or below	Small droplets evaporate within the duct window
NOx	Baseline	Reduced	Added water vapor quenches combustor hot spots
Output	Hot-inlet derate up to 40 percent	A slice recovered, about 8 to 10 percent of nameplate on the straight-line reading of the published band	Lower inlet temperature raises mass flow

The output row deserves care. The plant read the up-to-40 percent derate as an envelope that spans roughly the ambient range from the 15 °C ISO point to a hot extreme near 40 °C. A straight-line reading of those published endpoints gives about 40 / 25 = 1.6 percent of nameplate per degree, so recovering 5 to 6 °C maps to about 8 to 10 percent of nameplate. Real machines do not follow a straight line, and the plant used that figure as planning logic, not as a guarantee.

The retrofit added less than the full inlet drop, and that is worth stating plainly. The pads already captured about 4.8 K of the 6 K depression, so the fog gained only the last stretch toward the wet-bulb line, under 1 K. The larger recovery came from removing the pad pressure drop and its fan penalty, and from holding approach at design through the season.

The Acceptance Check List

Check	Target	How to verify
Measured approach	85 to 95 percent of the depression	Averaged dry and wet bulb readings over one hour
Injected water	Demineralized, conductivity in spec	Inline conductivity transmitter on the RO polish line
Final filtration	5 to 20 micron downstream of RO	Differential pressure trend across the filter
Droplet integrity	Bulk of flow at or below 20 microns	Manufacturer test data at rated pressure
Manifold flow drift	Under 5 percent over nominal	Flow meter trend; 5 percent wear equals about 10 percent over-delivery
Duct drainage	No pooling at low points	Borescope at the bellmouth

Reading the Numbers Against Published Ranges

Every figure in the tables maps to a published anchor for gas turbine inlet cooling. The hot-inlet derate of up to 40 percent of nameplate against the ISO 15 °C point is the published rating band, and the coastal site sits at the hot end of it. The humid-climate dry-bulb drop of 5 to 6 °C is what a small wet-bulb depression allows, against up to about 22 °C in hot dry desert air where the depression is large.

The approach band of 85 to 95 percent for a well tuned inlet fogging system, against about 80 percent for media evaporative coolers, is the published performance comparison, and the retrofit result stays inside it. The injection pressure of about 2000 psi, about 138 bar, is the published pressure for inlet fogging, and the 316L nozzle material with 0.1 to 0.3 mm orifices matches the published orifice class.

The droplet target of about 90 percent of flow at 20 microns or smaller and the 5 to 20 micron final filtration are the published figures for a well tuned inlet fogging system, with biofilm and mineral creep as the named failure modes. The square-law scaling explains why a 100 micron droplet takes about 100 times as long to evaporate as a 10 micron droplet.

Two side benefits carry their own anchors. Fog adds near-zero pressure drop to the inlet stream while media pads and chillers add parasitic loss, and added water vapor quenches combustor hot spots and reduces NOx as a documented effect. The wider market context is also published: more than 1000 fog systems installed worldwide as of 2015, with inlet fogging the lowest capital cost option among turbine inlet air cooling methods. A reader who wants the full selection path can follow the turbine inlet air fogging systems guide, which lays out the sizing order this case followed.

Why This Matters

The plant changed three things, and only one of them was a nozzle. The nozzle set fixed the droplet size that made the water evaporate in the duct. The RO skid and the fine filter made that droplet size worth something, because a small orifice behind unfiltered water blocks and struggles. The method change moved the cooling out of a matrix and into the airstream, which removed the pressure penalty the pads had carried for years.

The transferable discipline is the order of the decisions. Read the design weather first, because the wet-bulb depression sets the ceiling for any turbine inlet air cooling project. Then pick the method closest to that ceiling that adds no inlet pressure. Then set droplet size from residence time, the orifice from water demand, and the filtration from the orifice. Then size the pump and chart a wear number, so a drifted orifice shows up as a flow change.

That sequence is the same for any turbine inlet air cooling project, from a peaking site on a humid coast to a base-load machine in desert air. The warning: a low-depression site never shows a large temperature number, because the physics forbids it. The value lives in the removed pressure penalty, the stable approach, and the NOx benefit. A retrofit sold on a large dry-bulb drop in humid air will disappoint.

Inlet fogging wins where the depression is worth spending water on and inlet pressure must stay low. Cooling past the wet-bulb limit means wet compression and its blade risk, a different conversation.

What to Send for a Sizing

Each item changes the answer, so a complete list returns a sized configuration in one round instead of a trial.

• Design weather for the hottest month, dry bulb and wet bulb, because the depression sets the ceiling.
• Inlet air flow in kilograms per second, because the water demand scales directly with it.
• Duct length and air velocity from the manifold point to the bellmouth, because residence time sets the droplet size.
• Target temperature drop and the hours it is needed, because the water budget follows the duty.
• Make-up water quality and the RO arrangement, because the orifice demands a conductivity limit.
• Existing inlet filtration and the final-filter rating, because the fog filter sits downstream of it.
• Pump and manifold layout constraints, because 138 bar class hardware sets the footprint.
• Current inlet cooling method and its measured approach, because the gain over the old method is the real number.

Send the design weather and the inlet air flow through the contact page. You get the manifold count, orifice size and filtration band back, sized to your duct rather than a catalogue. That is the way to turn a hot-day derate into recovered capacity.