Inlet air cooling is the family of techniques that claws back some of that lost output. This guide explains how each technique works, what it costs in water and complexity, and how designers account for it when sizing a plant for a compute campus.
Why hot air costs output
A gas turbine compressor takes in a roughly fixed volume of air each second. Cold air is denser, so that volume contains more mass, more oxygen and therefore room for more fuel. Warm air contains less. As a rule of thumb used in preliminary studies, a heavy-duty machine can lose a few percent of output for every ten degrees Celsius of ambient rise, and aeroderivative machines are often more sensitive. Always take the real curve from the manufacturer’s data for the exact model.
The turbine’s rating is quoted at standard reference conditions, usually 15 degrees Celsius at sea level. Sites that are hot, high, or both will see less than the headline figure on their worst days. That is why plant designers start from the site’s design temperature rather than from the catalogue number. Our note on heat rate at part load covers the efficiency side of the same behaviour.
Evaporative coolers and fogging
The simplest approach is evaporative cooling. Air passes through a wetted media pad, water evaporates, and the air temperature falls toward the wet-bulb temperature. Effectiveness is typically stated as a percentage of the gap between the dry-bulb and wet-bulb temperatures. It works best in dry climates, where that gap is wide, and does very little on a humid day.
Fogging sprays fine droplets into the inlet duct, cooling the air in the same way. If more water is sprayed than the air can absorb, the surplus droplets enter the compressor and evaporate there, which is called wet compression and gives a further small boost. It demands high-quality demineralised water, careful nozzle design and checks on compressor blade condition. Both approaches have low capital cost and a small pressure loss, but they depend on weather and consume water.
Mechanical and absorption chillers
A chiller removes heat through a cooling coil in the inlet duct, so it can bring air below the wet-bulb temperature and work in humid conditions where evaporative methods struggle. Mechanical chillers are driven by electric motors and take auxiliary power from the plant, which offsets part of the gain. Absorption chillers use heat, often from turbine exhaust, and add a different set of equipment and controls.
The cooling coil adds a pressure drop to the inlet, which costs a little output all year, including the cool days when the chiller is off. The balance of gain and loss must be modelled hour by hour against site weather data. Chilled water or ice storage is sometimes added so the chiller can run at night and release cooling in the afternoon peak, shifting its power use away from the hours when the plant is working hardest.
Water, treatment and permits
Evaporative and fogging systems consume water. The amount depends on the airflow, the temperature drop sought and the hours of use, and it should be estimated at the design stage and compared with what the site can supply. Demineralised water for fogging needs its own treatment plant, storage and quality monitoring, and any blowdown must be routed to an approved discharge.
Chillers with cooling towers also use water, while air-cooled chillers use more electrical power instead. Where water is tight, that trade drives the choice. Some campuses have also found that cooling the inlet has an effect on emissions, because fuel flow changes with output; the air permit and monitoring plan should be reviewed with this in mind, alongside the broader guidance on permits for data center gas turbine plants.
Feeding the result into plant sizing
Designers do not assume that inlet cooling is always available. Equipment can be out for maintenance, water supplies can fail and a humid heat wave can erase the benefit of evaporative systems. A sound approach is to size the plant so that it can meet the campus load at the design temperature with the cooling system in service, and then test what happens if it is not, using the redundancy rules chosen for the site.
A short worked example shows the idea. If a plant needs 90 MW delivered on the hottest design day and each turbine gives 28 MW at that temperature, four machines cover 112 MW and leave one out of service with 84 MW available, which falls short. With inlet cooling adding perhaps a few percent per machine, the spare margin improves, but the owner must decide whether to rely on it. Our step by step guide to sizing generation for a data center sets out how these margins combine with redundancy and load growth. The figures above are illustrations of the method, not a design.
Controls, maintenance and operating practice
Inlet cooling systems need to be integrated with turbine controls so that they switch on and off at defined ambient conditions, protect against over-spray and shut down safely when a fault occurs. Operators should be able to see the temperature drop achieved, the water used and the inlet pressure drop on a single screen.
Maintenance matters. Media pads need cleaning or replacement, nozzles clog, chilled water pumps need attention and demineralised water quality must be tracked. A maintenance plan written alongside the plant design keeps these systems working on the days they are needed most.
About the authors
This guide comes from the USP&E gas turbine performance engineers who model site output across the year for campus power plants. USP&E is an EPC and O&M contractor, not a broker, with 150+ power projects in 45+ countries, 25,000 MW contracted since 2002 and 24+ years in the field. The wider picture on power for compute campuses is on our guide to gas turbine plants for data centers.
Offices: Colorado Springs (1755 Telstar Dr, Suite 300), Knoxville (800 S Gay St, Suite 700), London (162-168 Regent Street), Johannesburg (22 Kildoon Rd, Bryanston), Cape Town (31 Brickfield Road, Woodstock) and China (No. 1 Ningzhou Road, Wuxi). Regions: USA | AFRICA | UK | UAE | CHINA. ISO 9001 and ISO 45001 certified.
Inlet air cooling questions
Why does a gas turbine lose output on hot days?
Warm air is less dense, so the compressor draws in less mass of air per second and the turbine can burn less fuel, which lowers power output.
What is the difference between evaporative cooling and fogging?
Evaporative coolers pass air through wetted media, while fogging sprays fine droplets into the inlet duct. Fogging can also push droplets into the compressor, called wet compression.
Do chillers work in humid weather?
Yes. A chiller can cool air below the wet-bulb temperature, so it still helps when evaporative systems have little effect, though it uses auxiliary power.
Does inlet cooling use a lot of water?
Evaporative and fogging systems consume water in proportion to airflow and temperature drop, so the supply should be checked early in design.
Can a plant rely on inlet cooling to meet the load?
Designers usually test the plant both with and without it, because cooling equipment can be out of service or ineffective on a humid day.
Is inlet cooling an upgrade or part of the original plant?
Either. It can be specified in a new plant or retrofitted, although a retrofit must check duct space, controls and compressor suitability.
Check your hot-day output
Send the site design temperature, load profile and turbine data for your campus plant, and our performance engineers will review how inlet cooling affects the unit count and redundancy.
If a document set is easier than a conversation, send it to info@uspeglobal.com or upload a summary through the performance review enquiry; to talk it over, dial +27 10 822 2324. Turbine generators for campus power plants are listed in our gas turbine inventory.