Boiler efficiency monitoring means tracking how much of the heat in your fuel ends up in the steam, and why the rest is lost. You can calculate it directly, from steam output and fuel input, or indirectly, by adding up the heat losses. This guide explains both methods with a worked example, the readings to watch and what drift in each usually means.
What boiler efficiency means
Boiler efficiency is the share of the fuel's energy that is transferred into the water and steam. If a boiler is 85% efficient, 85 parts of every 100 parts of fuel energy go into the steam, and 15 parts leave through the stack, the blowdown, the casing and unburnt fuel.
Fuel is usually the largest running cost of a boiler house, so a small drop in efficiency is paid for every hour the boiler runs. A falling efficiency is also often the first sign of soot, scale, a burner out of tune or a drifting instrument.
Efficiency is quoted on the fuel's gross calorific value (GCV), common in India, or net calorific value (NCV). They give different numbers for the same boiler, so state which you use.
The direct method: output divided by input
The direct method, also called the input-output method, compares the heat added to the steam with the heat supplied in the fuel.
Efficiency = Steam flow × (Steam enthalpy − Feedwater enthalpy) ÷ (Fuel flow × GCV) × 100
- Steam flow: kg/h of steam leaving the boiler.
- Steam enthalpy: heat content of the steam in kJ/kg, read from steam tables at the steam pressure (and temperature, if superheated).
- Feedwater enthalpy: heat content of the water entering the boiler, which depends mainly on its temperature.
- Fuel flow and GCV: kg/h of fuel (or m³/h for gas) and its gross calorific value in kJ/kg.
The direct method needs only a few readings, which suits continuous monitoring. But it gives one number and no explanation, and any error in the steam or fuel flow meter goes straight into the result.
Worked example of the direct method
A fire-tube boiler burns furnace oil and produces saturated steam at about 10 bar(g). Over a steady hour at normal load the readings are:
| Reading | Value |
|---|---|
| Steam flow | 10,000 kg/h |
| Steam enthalpy (assumed, saturated steam at about 10 bar(g)) | 2,780 kJ/kg |
| Feedwater temperature | 105 °C |
| Feedwater enthalpy (assumed, water at 105 °C) | 440 kJ/kg |
| Fuel flow | 650 kg/h |
| GCV of fuel (assumed) | 42,000 kJ/kg |
- Heat added per kg of steam = 2,780 − 440 = 2,340 kJ/kg.
- Heat into the steam = 10,000 × 2,340 = 23,400,000 kJ/h (23.4 GJ/h).
- Heat in the fuel = 650 × 42,000 = 27,300,000 kJ/h (27.3 GJ/h).
- Efficiency = 23.4 ÷ 27.3 × 100 = 85.7%.
The same readings give the evaporation ratio: 10,000 ÷ 650 = 15.4 kg of steam per kg of fuel. It is easy to track day to day, but only comparable when steam pressure, feedwater temperature and fuel quality stay similar.
Now suppose the fuel flow meter is under-reading by 5%, so the boiler is really burning 682.5 kg/h. The true heat input is 28.665 GJ/h and the true efficiency is 23.4 ÷ 28.665 × 100 = 81.6%, not 85.7%. That gap of about four points comes entirely from one instrument, which is why calibration matters as much as the formula.
The indirect method: adding up the heat losses
The indirect method, or heat-loss method, works the other way round. It calculates each way heat escapes and subtracts the total from 100%.
Efficiency = 100 − (sum of all heat losses, as % of fuel input)
The main losses are:
- Dry flue gas loss: heat carried away by the hot combustion gases. It rises with stack temperature and with excess air, and is usually the largest single loss.
- Moisture and hydrogen losses: hydrogen in the fuel burns to form water, and that water, plus any moisture in the fuel and the combustion air, leaves as vapour carrying latent heat.
- Unburnt losses: carbon monoxide in the flue gas and, for solid fuels, unburnt carbon in fly ash and bottom ash.
- Radiation and convection loss: heat lost from the casing and openings. It is fairly constant in absolute terms, so it is a larger share at low load.
- Blowdown loss: hot water deliberately drained to control dissolved solids. Some test methods treat it separately from boiler efficiency, but in practice it is fuel you have paid for, so many plants account for it.
You need a fuel analysis (carbon, hydrogen, moisture, ash), flue gas O2 or CO2, flue gas and ambient temperatures, and for solid fuels the unburnt carbon in ash. It takes more work, but it shows where the heat is going and does not depend on accurate steam and fuel flow meters.
Formal acceptance and performance tests follow recognised standards, such as BS 845, ASME PTC 4 and the relevant Indian BIS standards. For day-to-day monitoring the aim is simpler: one consistent method, applied the same way every time, so that changes show up.
Readings to monitor continuously and what drift means
A spot test shows how the boiler performed on the day. Continuous monitoring shows when something changes. These readings matter most:
- Steam flow: the load. Needed to calculate efficiency and to compare like with like.
- Steam pressure: frequent swings point to load changes the controls cannot follow, or a burner cycling on and off.
- Feedwater temperature: a fall means more fuel per kg of steam. It often signals less condensate coming back, a deaerator problem or an economiser that is bypassed or fouled.
- Fuel flow: rising fuel at the same steam flow is the clearest sign that efficiency is falling.
- Flue gas O2 (or CO2): the measure of excess air. High O2 means heat is being wasted heating unnecessary air; very low O2 risks incomplete combustion.
- Flue gas CO, where measured: rising CO means fuel is not burning completely, from too little air, a worn burner or poor atomisation.
- Stack temperature: a slow rise at the same load usually means fouled heat transfer surfaces, soot on the fire side or scale on the water side.
- Drum level: mainly a safety reading, but unstable level can signal foaming, poor water treatment or a control valve problem.
- Blowdown: flow or frequency, plus boiler water TDS or conductivity. Too much wastes hot water; too little lets scale form.
A common rule of thumb is that every rise of about 20 °C in stack temperature costs roughly one percentage point of efficiency. Treat that as a rough guide for spotting trends, not a figure for your boiler.
Symptom, likely cause and what to check
| Symptom | Likely cause | What to check |
|---|---|---|
| Stack temperature rising at the same load | Soot on fire-side tubes, scale on water side | Soot blowing records, tube inspection, water treatment and TDS history |
| High flue gas O2 | Too much excess air, air leaks into the furnace or ducts | Burner air-fuel settings, damper linkages, casing and duct leaks, O2 analyser calibration |
| CO rising, or black smoke | Too little air, poor atomisation, worn burner parts | Burner nozzle and atomising steam or air, fuel oil temperature, air-fuel ratio across the load range |
| Feedwater temperature falling | Less condensate return, deaerator or economiser issue | Condensate return flow, steam traps, deaerator pressure, economiser inlet and outlet temperatures |
| Fuel per tonne of steam rising | Real efficiency loss, or a drifting flow meter | Compare with stack temperature and O2 trends; check fuel and steam meter calibration |
| Drum level swinging | Load swings, foaming, level control problem | Boiler water chemistry, level transmitter, feed control valve, load pattern |
| Frequent burner starts and stops | Boiler oversized for the load, controls poorly set | Pressure set points and differentials, turndown, how many boilers are online |
Practical ways to improve boiler efficiency
Tune excess air
Some excess air is needed for complete combustion, but any extra air is heated and sent up the stack. Tune the air-fuel ratio across the full load range, not at one firing rate, and use O2 readings to hold it. O2 trim control can keep the setting as conditions change.
Recover heat from flue gas
An economiser uses flue gas to preheat feedwater, and an air preheater uses it to warm combustion air. Both lower stack temperature. Check fuel sulphur first: cooling flue gas too far can cause acid corrosion.
Control blowdown
Base blowdown on measured boiler water TDS or conductivity rather than a fixed routine. Where continuous blowdown is used, flash steam and heat can often be recovered from it.
Keep heat transfer surfaces clean
Soot and scale both act as insulation. Good water treatment prevents scale; soot blowing or tube cleaning deals with soot. The stack temperature trend shows when cleaning is due.
Insulate and return condensate
Repair damaged insulation on the boiler, valves, flanges and steam lines. Return as much clean condensate as you can, because it is hot, already treated water, and it raises feedwater temperature.
Manage load sensibly
Boilers at very low load, or cycling on and off, lose more heat in purges and from the casing. Where several boilers serve one header, run the mix that keeps each at a steady, sensible load.
Many of these fixes are maintenance tasks that can be triggered by data rather than by the calendar. Our guide to predictive vs preventive maintenance covers how to decide which approach suits each asset.
Common boiler monitoring pitfalls
- Uncalibrated instruments. As the worked example shows, a small flow meter error can move the result by several points. O2 analysers and thermocouples drift too. Calibrate the key instruments on a schedule.
- Averaging over mixed loads. A monthly average that mixes full-load days, low-load nights and start-ups hides real changes. Compare efficiency at similar loads, and treat start-ups and shutdowns separately.
- Stale fuel data. GCV varies between lots, especially for coal and biomass. A default value used for months can make the boiler look better or worse than it is.
- Alarms nobody sees. A high stack temperature alarm on a control room screen does little if nobody is watching it on the night shift. Alarms need an owner, a way to reach that person and a record of who acknowledged them.
Boiler trips also cost production; our guide on how to reduce unplanned downtime covers tracking and cutting such stops.
Bringing boiler readings together with MIE
MIE, the ThinklytixAI Manufacturing Intelligence Engine, is built to watch readings like these. Its demo plant is a power and steam plant, with boiler steam pressure and drum level on the overview and trend screens. MIE takes readings you already have from existing PLCs, gateways and historians over HTTPS, MQTT, OPC-UA or Modbus, with no new sensors, and only accepts data from machines and tags you have registered. It shows a plant overview, trends and a page per machine, and sends alerts on Slack, WhatsApp or webhook that escalate through timed tiers until someone acknowledges them. Teams can also ask plain-English questions about their plant data.
Because fuel burned in boilers is a direct source of emissions, the same readings matter for carbon reporting. MIE Carbon, for emissions tracking and estimated carbon credits, is in development and will build on the plant data MIE already collects.
Summary
Use the direct method for a quick, continuous efficiency figure and the indirect method to find where the heat is going. Watch steam flow, pressure, feedwater temperature, fuel flow, O2, stack temperature, drum level and blowdown together, compare like loads, keep instruments calibrated and make sure every alarm reaches someone who will act on it.