Boiler efficiency is the heat delivered in steam divided by the heat supplied in fuel, expressed as a percentage. It is worked out two ways: directly, from output and fuel, and indirectly, by subtracting the losses from a hundred. One of those losses is radiation and convection from the hot surfaces of the boiler, and it is the term that is normally guessed. On the Cochran we surveyed it is 49.8 kW.
Boilers/This page
The direct method compares what came out with what went in. Nothing about the boiler needs to be known beyond the two flows and the state of the water and the steam.
Efficiency = steam flow × (steam enthalpy − feed water enthalpy) ÷ (fuel flow × gross calorific value)
Steam flow in kg/h from the meter, enthalpies from steam tables at the operating pressure and the feed temperature, fuel flow from the gas or oil meter, calorific value from the supplier. Every quantity is read, and that is the strength of the method: it needs no assumptions.
It gives the number and not the place the heat went. where the missing heat went. A boiler at 82 per cent and a boiler at 88 per cent look the same in this calculation.
The indirect method starts at a hundred per cent and subtracts each loss in turn. It takes longer and it tells you which loss to work on.
Efficiency = 100 − dry flue gas loss − loss from hydrogen and moisture in the fuel − blowdown loss − radiation and convection loss
| Loss | How it is normally established | Where we come in |
|---|---|---|
| Dry flue gas | Flue gas analysis: temperature, oxygen, carbon dioxide | Not ours. Measured by the combustion engineer. |
| Hydrogen and moisture in fuel | Fuel composition and calorific value | Not ours. Comes from the fuel. |
| Blowdown | Solids balance against the permitted boiler water figure | Covered on its own page; we measured the flanges, not the water. |
| Radiation and convection | Taken from a chart or assumed as a flat percentage of rating | Measured. FLIR surface temperatures, CAD areas, ISO 12241. On the Cochran, 49.8 kW. |
The last row is why this page is here. The radiation and convection term is the one loss in the list that is habitually assumed rather than read, and it is the one that changes when insulation is fitted or cut away.
The same measured loss, placed against boilers of different output. Arithmetic only: no assumption about the machine beyond its rating.
| Boiler heat output, kW | Surface loss bare, share of output | Panelled, share of output |
|---|---|---|
| 500 | 9.95 % | 0.52 % |
| 1,000 | 4.98 % | 0.26 % |
| 1,500 | 3.32 % | 0.17 % |
| 2,000 | 2.49 % | 0.13 % |
| 3,000 | 1.66 % | 0.09 % |
| 5,000 | 1.00 % | 0.05 % |
| 8,000 | 0.62 % | 0.03 % |
| 10,000 | 0.50 % | 0.03 % |
49.75 kW bare and 2.58 kW panelled are the measured figures for the Cochran we surveyed, divided by the rating in the first column. The loss does not fall as the rating rises: the surface stays the same size and the same temperature, so on a small boiler the same watts are a larger share of the output.
47.2 kW recovered, 7,700 running hours, 90 per cent boiler efficiency on the fuel. Only the price changes.
| Fuel, €/MWh | Saved per year | Over ten years |
|---|---|---|
| 40 | €16,143 | €161,426 |
| 50 | €20,178 | €201,783 |
| 60 | €24,214 | €242,139 |
| 70 | €28,250 | €282,496 |
| 80 | €32,285 | €322,852 |
| 100 | €40,357 | €403,566 |
47.2 kW recovered on the machine we surveyed, 7,700 running hours, 90 per cent boiler efficiency on the fuel side: 363 MWh of heat and 404 MWh of fuel a year. Only the price changes down the column. The row at 60 €/MWh is the figure used everywhere else on this site, and the CO₂ that goes with it is 81.4 t a year at 201.6 kg/MWh.
Bare, the surveyed surfaces of that machine shed 49,750 W. Panelled with removable modules and held at the 45 °C touch-safe design target, assessed to EN ISO 13732-1, the same surfaces shed 2,580 W. The difference is 47,170 W of heat that stays in the boiler.
Over 7,700 running hours that is 363 MWh of heat. At 90 per cent efficiency on the fuel side it is 404 MWh of fuel, and at 60 €/MWh it is €24,214 a year. The carbon that goes with it, at 201.6 kg/MWh, is 81.4 t.
None of those four figures needs a flue gas analysis or a fuel meter. They follow from surface temperatures, areas and hours.
Steam output, fuel input and flue gas temperature of that machine were not measured by us, so we quote no efficiency percentage for it. What we publish is the radiation and convection term and what it is worth.
The direct method needs two meters and a steady boiler. It is the test that gives a number; the indirect method is the one that says where the number went.
Inzonex did not meter steam, fuel or flue gas on any machine here, and quotes no efficiency percentage for one. What we measure is a single term of the indirect calculation: radiation and convection from the hot surfaces, 49,750 W bare and 2,580 W panelled on the surveyed Cochran.
The heat delivered in steam divided by the heat supplied in fuel, as a percentage. It is worked out either directly from output and input, or indirectly by subtracting the losses from a hundred.
Direct method: steam flow times the difference between steam and feed water enthalpy, divided by fuel flow times gross calorific value. Indirect method: 100 minus dry flue gas loss, minus loss from hydrogen and moisture in the fuel, minus blowdown loss, minus radiation and convection loss.
The direct method gives a number and needs two meters. The indirect method gives the same number broken into losses, so it says which one to work on.
From a chart or as a flat percentage of rating. That is why it is the loss most often wrong, because it is assumed.
Surface temperatures with a thermal camera, areas from the CAD model of the machine, and ISO 12241:2022 steady state to turn the two into watts. On the Cochran we surveyed that is 49.8 kW bare.
It changes one term of the indirect calculation. On the machine we surveyed, removable modules take the radiation and convection loss from 49,750 W to 2,580 W, and 47,170 W stays in the boiler.
Enthalpies at the operating pressure and the feed water temperature come from published steam tables. We publish no steam table of our own and quote none for a machine we did not meter.
The surveyed view this page came from, with every element on it.Open
→/ 02The machine the 49.8 kW came from: 28 components, six surveyed views.Open
→/ 03What is examined, by whom, and why insulation that has to be cut is insulation that stays off.Open
→/ 04The other loss term in the indirect method that has a page here.Open
→| Survey | FLIR S62 Pro, emissivity 0.90, reflected 25 °C, 3 m, full radiometric correction. UK commercial boiler house. |
| Geometry | Rebuilt in CAD from the surveyed machine. Area allowances are stated on the page they are used on and are allowances, not measured areas. |
| Calculation | ISO 12241 steady state. Heat loss is calculated from the measured surface temperature and the modelled area — it is not metered. Full method, basis and limits. |