Compressor Water Wash Scheduling with AI for Gas Turbines

Why Wash Timing Matters More Than Wash Frequency

The same wash can make money or lose money, depending on when it happens.

A fouled compressor moves less air. Output falls and heat rate rises. An offline wash brings most of that back, but the unit has to stop. If that outage lands in hours when the market wanted the unit, the lost margin can be larger than everything the wash recovers. If it lands in hours when the unit would have been idle anyway, the outage costs almost nothing. A scheduling approach weighs three things together: how fast the unit is fouling, what the lost output is worth, and when the next low-value window opens. Our performance engineers can review how your washes are timed today.

Online vs Offline Compressor Wash

Two different tools. A good schedule uses both.

An online wash sprays demineralised water, sometimes with detergent, into the compressor inlet while the unit runs. It needs no shutdown and mainly cleans the inlet guide vanes and front stages. In a Turbomachinery Magazine article on compressor degradation, Turbotect's Jean-Pierre Stalder put the gain at about 0.2% to 1% of power per online wash, and recommended washing daily or every two to three days. An offline wash, also called a crank wash, is done with the unit shut down and cooled, turning on the starter at crank speed while wash solution is sprayed in. It reaches more stages and can recover nearly all of the fouling loss. The same article gave 12 to 36 hours for an offline wash. Plan both wash types for your units in a working session.

Summary for orientation. Your OEM wash procedure sets temperatures, flow rates, detergent and limits.

What Drives the Compressor Fouling Rate

Fouling follows the air the unit breathes, not the calendar.

Fouling comes from fine material that passes the inlet filters and sticks to the blades. In the same Turbomachinery Magazine article, Rainer Kurz of Solar Turbines said particles of 1 to 5 microns are the size most responsible for fouling, which is why filter quality matters so much. Oily vapour makes deposits sticky. Salt matters near the coast. These sources change with season and wind, so the fouling rate on one unit varies from month to month. A wash schedule that ignores this will be early in clean months and late in dirty ones. Ask our specialists to read your fouling rate history.

Measuring Fouling with Corrected Performance

Raw output moves with the weather. Corrected output moves with fouling.

Output rises and falls with the weather, so raw megawatts cannot show fouling. Ratings are quoted at ISO reference conditions, which ISO 3977-2 sets at 15 °C, 101.325 kPa and 60% relative humidity. Performance monitoring corrects each reading back to these or to a site reference, using the OEM correction curves for ambient temperature, pressure and humidity. Once output, heat rate and compressor discharge pressure are corrected, a steady downward slope between washes is fouling, not weather. Kurz noted in the article above that compressor discharge pressure is one of the most effective signals for tracking fouling. That slope is the fouling rate, and the schedule is built on it. Set up correction for your units with our performance team.

Worked Example: Lost Output vs Wash Cost

The wash is the same. Only the timing changes the answer.

Take an illustrative 175 MW unit that is expected to run 16 hours a day for the next 28 days, or 448 running hours. Corrected data show that an offline wash now would recover an average of 3.0 MW over that period, allowing for fouling coming back. At an illustrative margin of $30 per MWh, that is 1,344 MWh worth $40,320. Water, detergent, labour and start costs for the wash are taken as $12,000. If the wash fits in a weekend when the unit was not going to be dispatched, the net gain is $28,320. If the same 18-hour wash falls in dispatched hours, it also gives up 18 × 175 = 3,150 MWh, or $94,500 of margin, and the result is a loss of $66,180. All figures are illustrative. Check the numbers for your own unit with our engineers.

Illustrative only. Real margins, heat rate gains, start costs and recovery depend on your unit and market. Heat rate savings are left out for simplicity.

Planning Offline Washes Around Dispatch

The best window is often already in the forecast.

Most plants know days ahead when a unit is likely to be off line or at low load, such as weekends and low-demand nights. The wash agent reads the dispatch forecast, compares each window with the forecast fouling loss and suggests the window with the best net value. It also checks that the window is long enough for cooldown, wash, soak, rinse and drying, and that crew, demineralised water and detergent are ready. If no good window opens before the loss grows too large, it says so, and the team decides whether to take a paid outage or wait. Walk through a real forecast with us in a short session.

Measuring Wash Effectiveness

If a wash is not measured, nobody knows if it worked.

Every wash should leave a before and after record. For an offline wash, that means corrected output, heat rate and compressor discharge pressure at matching load just before shutdown and soon after restart, plus the final rinse water conductivity. For online washes, the trend over the following days shows whether they are slowing the fouling rate. A wash that recovers less than usual is a signal to inspect nozzles, filters or the procedure. Review your wash records with our team.

Compressor Water Wash Scheduling Checklist

Eight checks before the wash plan goes live.

Use this checklist for one unit first. Tick items already in place. Get help with any gaps from our setup team.

Compressor Wash Mistakes to Avoid

A wash in the wrong hour can cost more than the fouling.

These are the patterns that most often weaken wash programs. A short wash review can find them.

How iFactory Runs Compressor Water Wash Scheduling

The Wash Optimization Agent, on your own network.

iFactory's Wash Optimization Agent estimates fouling from performance data, recommends the next online or offline wash around dispatch forecasts and measures recovered MW after each wash. It runs on an on-prem AI server inside your plant.

Wash timing, outage and dispatch decisions remain with your operations team. Figures on this page are worked examples.

Turnkey AI: Delivered, Connected and Live in 6–12 Weeks

You do not build this. It arrives ready.

iFactory ships as a pre-configured NVIDIA AI server with the software pre-loaded, racked and ready. Rack it, plug in power and Ethernet, and the AI is live on your network. Our team handles cabling, network setup, PLC, SCADA, DCS and CMMS integration, historian and dispatch system links, operator training and 24×7 remote monitoring. Data stays on your network. For a scope matched to your plant, request a turnkey quote.

Frequently Asked Questions