Generator set humidity derating is the loss of engine output that follows from water vapour displacing dry air in the intake, and Tesla Power calculates it from grams of water per kilogram of dry air rather than from a relative reading. The steps below fix the figure before you request a quotation.
- Water vapour is lighter than air, so a humid intake holds fewer oxygen molecules at the same pressure and temperature.
- The densest tropical air costs about 1.8% of output, while the same air at 1000m costs close to 6%.
- Temperature removes three times as much rating as humidity on the same site, so correct both.
- State the wet-bulb and dry-bulb readings together, and the frames in the export frame range are compared on one basis.
What does generator set humidity derating change in the intake air?
It changes mass of dry air, not the volume. A turbocharged engine moves a fixed volume per second at a given boost, so the power it makes follows the density of that volume. When water vapour takes up part of the total pressure, the dry-air share falls and the cylinder receives fewer oxygen molecules on each stroke.
The effect is smaller than temperature because water vapour reaches only a few kilopascals of the 100 kPa total. At 35°C and 85% relative humidity the vapour pressure is about 4.8 kPa, so dry air supplies 95.2 kPa. That share costs roughly 1.8% of density once the lower molecular weight of water is counted, because the vapour itself still carries mass.
How do tropical and desert sites compare at the same load?
They compare in opposite directions, and that is why a relative humidity figure misleads. A tropical site at 35°C and 85% relative humidity holds about 30 g of water per kilogram of dry air. A desert site at 45°C and 20% holds about 18 g, so the desert air is drier in absolute terms even though it is hotter.
Temperature still dominates. A 500kVA frame at 0.8 power factor carries 400kW at the 40°C reference, and the step to 45°C removes about 5%, or 20kW. The humidity term then removes 1.5% in the desert case and 1.8% in the tropical case, so the two sites finish within 0.3% of each other on the same frame. Yangzhou Tesla Power prints both terms rather than the combined figure, so a buyer can see which one is doing the work.
| Site condition | Vapour pressure | Density against dry air | Share of output lost |
|---|---|---|---|
| 35°C, 85% RH, sea level | 4.8 kPa | 98.2% | 1.8% |
| 45°C, 20% RH, sea level | 1.9 kPa | 99.3% | 0.7% |
| 40°C, 60% RH, sea level | 4.4 kPa | 98.5% | 1.5% |
| 35°C, 85% RH, 1000m | 4.8 kPa | 92.8% | 5.8% |
Why is absolute humidity the figure that matters?
Absolute humidity states the mass of water in a given mass of dry air, in grams per kilogram, and it does not change with temperature. Relative humidity states how close the air is to saturation, so the same 60% reading describes very different water loads at 25°C and at 45°C. A correction table built on relative humidity gives the wrong answer the moment the site temperature moves.
The practical consequence is that a supplier who receives only relative humidity cannot compute the correction. Two sites at 70% relative humidity, one at 28°C and one at 42°C, differ by more than a factor of two in water content. Chinese generator set suppliers often publish a chart against relative humidity anyway, because that is what a site engineer can read from a hygrometer on the wall.
Where should the humidity reading be taken?
Take it at the air cleaner with the set running, at the same time and place as the temperature reading. A hygrometer held at the plant room door reads the corridor. Tesla Power asks for the reading at the air cleaner because a canopy set draws from a pocket that is warmer and wetter than the room, since the radiator plume and the engine both discharge into it. Both readings belong on one line of the site sheet.
Record the barometric pressure as well, because the vapour pressure calculation needs it. A site at 1000m reads about 90 kPa rather than 101 kPa, and the lower total pressure lifts the vapour share of the mixture. That is why the fourth row of the table shows a 5.8% loss rather than the 1.8% the same air produces at sea level.
How much output does water vapour remove on a worked case?
Work the case at 500kVA and 35°C, on the reference conditions in the ISO 8528-1 rating basis. At 85% relative humidity the vapour pressure is 0.85 × 5.63 = 4.79 kPa, so dry air supplies 101.3 - 4.79 = 96.5 kPa. Dry density at 308K reads 96500 ÷ (287 × 308) = 1.092 kg/m³, and the vapour adds 4790 ÷ (461.5 × 308) = 0.034 kg/m³.
The total is 1.126 kg/m³ against 1.146 kg/m³ for dry air at the same pressure, which is 98.2% of density. A 500kVA frame at 400kW therefore holds 400 × 0.982 = 98.2% of the plate. Fuel at that point stays near 0.21 L/kWh, so the set burns about 83 L/h at what is now its full rating. Tesla Power prints the vapour pressure, the dry density and the resulting figure on one line, so the arithmetic can be checked rather than trusted.
Which engines lose more, turbocharged or naturally aspirated?
The naturally aspirated engine loses the full density ratio, because it can only draw what the atmosphere pushes in. The turbocharged engine loses less, because the compressor works against a pressure ratio and the wastegate trims boost to keep the air-fuel ratio inside the smoke limit. A turbocharged set loses about 70% of the density figure in practice, and a naturally aspirated set loses all of it. Most generator set manufacturers state the correction for one reference altitude only.
That difference is why a turbocharged frame is the usual choice above 200kW on a tropical order. The same engine holds 30% more of its rating at the same site, and the exhaust temperature stays inside the limit at full load. Tesla Power quotes the compressor map with any offer above 400kVA, so the boost margin is visible rather than assumed.
What does the factory check before shipment?
Our test bay runs the set at the ambient it will meet in service, or as close as the season allows, and records intake temperature, barometric pressure and humidity beside the load. A set tested on a cold dry morning in Yangzhou has not been proved for a coastal site at 35°C, and the sheet says so rather than implying otherwise.
Where the season does not cooperate, the corrected figure is calculated from the recorded conditions and printed beside the measured one. Two numbers on the sheet show how much of the rating the site conditions will take. The 400kW frame on the WP13 platform leaves the bay with both figures and the engine serial on one page.
How should an enquiry state the humidity?
State four items: the highest dry-bulb temperature, the matching wet-bulb reading, the site altitude, and the load the plant carries at that hour. The wet-bulb figure fixes the absolute humidity without a psychrometric chart, and a site team can read one from a sling hygrometer in ten minutes.
Send those four lines with the load list to the export desk, and the reply carries the corrected figure, the fuel burn at that figure, and the frame that holds the duty. The site planning notes set out how to take the wet-bulb reading, and Tesla Power Equipment sends the same form with every enquiry so two quotations can be laid side by side.
Where do buyers go wrong on tropical orders?
The first error is applying relative humidity to a chart built for absolute humidity, which overstates the loss on a hot afternoon and understates it on a cool wet morning. The second is reading the humidity and forgetting the temperature, when temperature takes three times as much of the rating on the same site. A buyer who corrects for 85% humidity and leaves 35°C uncorrected has corrected the smaller term. A generator set for a coastal site needs both terms on the same line of the enquiry.
The third error is assuming a canopy protects the set from humidity. It does the opposite on a still night, because the engine and radiator add water and heat to a small volume and the louvres cannot clear it. Suppliers who test in a canopy report intake relative humidity 8-12 points above the yard, which is the figure that belongs on the sheet rather than the one from the airport. A generator set in parallel operation doubles that plume, so two units in one room read worse than one.
How is the correction confirmed at handover?
The commissioning engineer takes wet-bulb and dry-bulb readings at the air cleaner with the load applied, then writes the corrected figure beside the measured one. A site commissioned in the dry season carries a note that the figure was not proved at the humid peak, and that note tells the operator what to watch in the wet season. The monthly test schedule in the NFPA 110 test routine is where that watch is recorded.
The same sheet records boost pressure and exhaust temperature at full load, because those two figures show whether the engine had margin left. A set running at the exhaust limit on a mild day has none for the humid season. Tesla Power writes both into the handover pack, and a service visit a year later is read against them.
Does humidity change the alternator rating?
It changes it far less than the engine, and mostly through insulation class rather than air density. An alternator rated for a 40°C ambient carries a small correction above that, but winding temperature rise is the limiting figure and the site ambient sets it. A set in a humid room needs anti-condensation heaters more than a larger alternator. The same frame appears in the 1000kW container build for standby duty, where the plate states the same basis.
That heater matters on a coastal site, because a winding below dew point collects moisture and the insulation resistance falls. Fit the heater and wire it to run when the set is stopped, which is the condition that causes the problem. On the 500kVA silent set with the ATS option the heater sits on the same circuit as the charger, so a site that leaves one live keeps both. A genset for a coastal site that starts weekly without a heater can show a low insulation reading which clears after twenty minutes of running, so the reading is a warning rather than a fault. A generator set for standby duty carries the same heater on a shorter duty cycle, which makes the heater more important rather than less.
Frequently asked questions about generator set humidity derating
Between 0.5% and 2% on most sites, which is smaller than the temperature correction beside it. The figure grows with altitude, reaching about 6% at 1000m in tropical air. A buyer who plans on 2% and checks the altitude covers the range.
It is not useless, but it has to be paired with dry-bulb temperature to give absolute humidity. A chart against relative humidity alone is valid only at the temperature it was drawn for, which is rarely the site temperature. Tesla Power converts the pair before quoting.
It raises the burn per kWh slightly, because the engine works harder for each unit delivered. The change sits inside the tolerance of most fuel curves, so the data sheet still holds for planning. Measure it rather than assume it where the contract carries a penalty.
Yes, and louvre free area is the figure to check first. A canopy that holds the radiator plume also holds the vapour, so intake humidity climbs through the day. A discharge duct that turns the plume away from the intake fixes most of it.
Thirty per cent relative humidity at 25°C and 100 kPa, which is dry air by tropical standards. Every plate figure carries that reference whether or not the site meets it. The correction exists because most sites do not.
It can be re-stated on paper from the recorded conditions, and the plate stays as built. A re-rate that changes the plate needs the engine maker's approval, because the warranty follows the published rating rather than the site figure.
Rain that reaches the filter is a separate fault and a worse one, because it blocks the element and can carry water into the engine. A rain cap and a louvre that turns the air twice handle it. Water in the fuel is a different problem again.
Generator set CE marking requirements sit in the technical file, and the rating basis is recorded there separately from the declaration. Conformity assessment for the assembly names the standards the build was checked against. A buyer reads the two documents together rather than as one, and Tesla Power keeps both in the project folder.