Vulcan Logo
Heating sizing check

Heating sizing check

Last updated: 14/11/2026

The heating sizing check answers three questions about the heating of a model:

  • Is the heat pump or boiler big enough for the coldest hour of the year?
  • Is it too big, so that it switches on and off?
  • What is the lowest flow temperature at which the radiators still give enough heat?

It does not run a simulation. It does not tell you how efficient the system is over the year: the full run does that. It does not check single rooms, and it does not include hot water. For the rest of the system inputs, see the Geometry Systems Guide.

Where it shows

The check appears in the element panel in three places. Each block is called Sizing:

  • A System element of type heat source (heat pump or boiler). It shows a range bar.
  • A System element of type space heat system. It shows a chart.
  • A drawn wet emitter (radiator). It shows as one line.

A heat source that is too small or too big gets a warning flag in the Elements panel and in the inspector. So does a radiator system that is too small, and each of its radiators. A flag says the verdict only. For example: The radiators are too small at 55 °C flow. The numbers are in the Sizing block. The flags are warnings only. They never block save, run or export.

If one System element holds more than one heat source, the Sizing block shows each one under its name, and a flag names the heat source it is about.

Where the heat loss comes from

Vulcan asks the HEM engine for the heat transfer coefficient of each zone. The heat transfer coefficient is the heat the zone loses for each degree of difference between inside and outside. The engine finds it with its own static calculation, from the fabric, the thermal bridges and the ventilation. There is no simulation.

Vulcan then finds the design heat loss:

Design heat loss (kW) = heat transfer coefficient (W/K) × (set-point − design outside temperature) ÷ 1000

  • The design outside temperature is the coldest hour in the weather file that the model uses. The default weather, for FHS and for HEM:EPC, is the Bedford file.
  • For FHS, the set-point is 21 °C. This is the higher of the living area (21 °C) and the rest of the house (20 °C).
  • For HEM:EPC, the set-point is the static internal temperature of the engine (21 °C).
  • The block names the file and both temperatures. For example: Coldest hour in the Bedford weather file: −6.5 °C outside, 21 °C inside.
  • In a model with more than one zone, the text says zone where it otherwise says house.

Worked example, from the FHS demo model. The heat transfer coefficient is 325.7 W/K. The coldest hour in the Bedford file is −6.5 °C. So the design heat loss is 325.7 × (21 − (−6.5)) ÷ 1000 = 325.7 × 27.5 ÷ 1000 = 9.0 kW.

This is a design-day check. It is not a performance prediction.

Heat pump or boiler: the range bar

This check covers air-to-water heat pumps and boilers. The bar shows the range in which the unit can run on the design day. A marker on the bar shows what the house needs.

  • Right end: the most heat the unit gives on the design day. For a heat pump, this comes from its declared test data: the capacity at test point A (−7 °C), at the design flow temperature. For a boiler, it is the rated power.
  • Left end: the least heat the unit gives while it runs. Below this, it must switch off and on again. For a heat pump, it is the minimum modulation rate × the capacity. For a boiler, it is the minimum modulation load × the rated power.

Both ends matter. If the need is above the right end, the unit is too small: the house gets cold on the coldest days. If the need is below the left end, the unit is too big: it switches on and off, which wears it and costs efficiency.

The bar label gives the size first. For example: 10.4 kW on the design day · min 3.5 kW. If the minimum is not known, the label is 10.4 kW on the design day.

  • Between declared design flow temperatures, Vulcan interpolates the capacity, as the engine does. Outside the declared range, it uses the nearest declared one and says so.
  • If the design day is colder than test point A, Vulcan uses the coldest declared point (A or F) as declared. If the design day is colder than that point too, the block says so. Vulcan does not derate the capacity.
  • A heat source that feeds several systems is compared with the sum of the heat loss of their zones.
  • A zone with more than one heating system is not sized. Vulcan cannot tell how the heat loss divides between them. This applies to the systems that Vulcan could otherwise size. The block says More than one heating system serves this zone.

Radiators: the chart

This check covers radiators. The chart shows how much heat the radiators give at each flow temperature, from 30 °C to 70 °C. The chart goes past 70 °C when the design flow temperature is higher.

  • A dashed line shows what the house needs.
  • Where the curve crosses the dashed line is the lowest flow temperature at which the radiators give enough heat. A ring marks it.
  • The zone left of the crossing is marked Not enough heat. The zone right of it is marked Enough heat.
  • A dot shows the design flow temperature. If the controller uses weather compensation, the dot is at the flow temperature for the design day.
  • The sentence above the chart gives the crossing. For example: Radiators give enough heat from 38.9 °C flow. Vulcan searches up to 100 °C, or up to the design flow temperature when that is higher. If no flow temperature in that range works, the block says so.
  • The line can bend where the pump reaches its maximum flow rate.

How Vulcan works out radiator output

Vulcan uses the formula that the HEM engine uses for radiators. The Rust engine (FHS) and the Python engine (HEM:EPC) are the same here.

In words: a radiator gives heat according to how much warmer its water is than the room. The water is not at the flow temperature all the way through. It cools as it gives off heat, so it comes back at the lower return temperature. The engine uses the mean of the two.

As a formula, for each radiator:

Output (kW) = c × (T_mean − T_room)^n

  • c and n come from the radiator's test data (c can be c per metre × length).
  • T_mean is the mean water temperature: (flow temperature + return temperature) ÷ 2.
  • T_room is the room temperature. Vulcan uses the set-point (21 °C). The engine uses the zone air temperature.
  • The radiators' outputs add up.

The return temperature depends on how fast the water flows:

  • Variable flow. The water cools by the design temperature difference across the emitters (for example 10 K). The pump changes its flow rate to make this true. If the flow rate it needs is above the maximum flow rate, the pump stays at the maximum and the water cools more. Below the minimum flow rate, it stays at the minimum and the water cools less.
  • Fixed flow. The pump runs at the design flow rate. The water cools by as much as the heat it gives off needs.

Why the mean water temperature matters: it is lower than the flow temperature, so radiators give less heat than the flow temperature alone suggests. A simpler rule, flow temperature − set-point, overstates the output. The error is largest at low flow temperatures, which is where a heat pump runs best.

Worked example, from the FHS demo model. It has four radiators, variable flow, a design temperature difference of 10 K and a flow rate of 3 to 18 litres a minute. The house needs 9.0 kW.

  • At 38.9 °C flow, 9.0 kW needs 12.8 litres a minute at 10 K. That is inside the range. So the return is 28.9 °C and the mean is 33.9 °C, which is 12.9 K above the room. At that difference the four radiators give 9.0 kW. So 38.9 °C is the lowest flow temperature that works.
  • At 55 °C flow, the radiators could give more. But at 18 litres a minute, the most the pump allows, the water cools by 17 K. The mean is then 46.5 °C, and the radiators give 21.3 kW.
  • Above about 43 °C flow the pump is at its maximum, so the line bends there.

Suggested lower flow temperature

For a heat pump, the block can suggest a lower design flow temperature. The suggestion then replaces the sentence above the chart. For example: You could lower the design flow temperature to 39 °C. The radiators still give enough heat, and the heat pump runs more efficiently.

The suggested temperature is the lowest flow temperature at which the radiators give enough heat, rounded up to a whole degree. Vulcan suggests it only when all of these are true:

  • Every heat source that feeds the radiators is a heat pump, and it heats only this system. If a heat pump also heats another system, that system can keep the higher flow temperature.
  • The radiators are Big enough at the design flow temperature.
  • At the suggested temperature, each heat pump is still big enough and not too big. Vulcan works out its range again at that flow temperature, from the same declared test data as the range bar.
  • The suggested temperature is at least 1 °C below the flow temperature on the design day.
  • The suggested temperature is not below 30 °C, the lowest flow temperature of the weather compensation curve (the minimum flow temperature), or the lowest flow temperature in the heat pump's test data. The controller never runs below its minimum flow temperature, and below the test data the heat pump's output is not known.
  • The design day runs the design flow temperature that you set. With weather compensation, this is true only when the design day is cold enough for the full design flow temperature: for example, at or below the minimum outdoor temperature. On a milder design day the controller runs a lower flow temperature than the setting, so Vulcan suggests nothing.

In the FHS demo model, the radiators give enough heat from 38.9 °C. But at 39 °C the heat pump gives 8.5 kW, which is below the 9.0 kW the house needs. So the block suggests nothing and keeps the sentence Radiators give enough heat from 38.9 °C flow.

How to read the result

Result Where Meaning
Big enough Heat source The need is inside the range of the unit: below its most and above its least output.
Too small Heat source The most the unit gives is below the need. The block gives the outside temperature below which the gap appears, and what happens then (see below).
Too big Heat source The least the unit gives is above the need. On the design day, it makes more heat at its minimum than the house needs, so it switches on and off.
Big enough at 55 °C Radiators At the design flow temperature, the radiators give at least what the house needs.
Too small at 55 °C Radiators At the design flow temperature, the radiators give less than the house needs. The sentence gives the flow temperature that would work, if one up to 100 °C does.

On/off (non-modulating) heat pumps are never flagged Too big.

For a heat source that is Too small, the block says what covers the gap below that outside temperature:

  • The heat pump has a backup heater: Below −6 °C outside, the backup heater covers the gap.
  • The heat pump has no backup heater, but another heating system that Vulcan does not size (for example a direct electric heater) serves the same zone: Below −6 °C outside, another heating system in the zone must cover the gap.
  • Neither: Below −6 °C outside, rooms run below set-point.

Example of a result for a heat pump that is too small:

House needs 9.0 kW. Heat pump gives 8.8 kW on the design day and turns down to 3.5 kW. Below −6 °C outside, the backup heater covers the gap.

No flag means "nothing to say". It does not mean "fine".

What is not checked

When Vulcan cannot size a system, the block says Not sized: and the reason. It shows no verdict. Vulcan does not size these:

  • Underfloor heating
  • Fan coils
  • Storage heaters
  • Warm air
  • Direct electric heaters
  • Heat batteries
  • Heat networks
  • Ground-source, water-source and exhaust-air heat pumps
  • Heat pumps that feed a primary thermal store
  • Load-compensated controls
  • A bypass that sends return water back into the flow
  • Zones with more than one heating system
  • Models with no heat source or no emitters

Limits

  • The heat loss is for the coldest hour only. It is exact for the model as prepared. It is only as right as the model's inputs.
  • The heat pump size is the declared capacity at test point A (−7 °C), or at the coldest declared point when the design day is colder. Vulcan does not derate it. A design day colder than the point used is a caveat, and the block says so.
  • Radiator output is the steady state: the radiators are warm and the room is at the set-point. The engine uses the zone air temperature. Vulcan takes it as the set-point.
  • The check covers space heating only. It does not include hot water. There is no room-by-room check, because the model has no rooms.

Targets and timing

The check works on Rust HEM:FHS targets and on HEM:EPC targets, when the target has a design heat loss export. No released target has it yet, so released targets show nothing.

  • On a target that cannot size, the Sizing block shows nothing and no flags appear.
  • Vulcan gives a reason (Not sized: and why) only on a target that has sized a model since you selected it. On any other target, the block also shows nothing when the target is not downloaded or the model has errors.
  • For HEM:EPC, the check uses the Python engine. It takes about 7.5 seconds to start and about 6 seconds for a model. Later results for an unchanged model are instant.
  • For a Rust HEM:FHS target, the engine starts in about 1 second.
  • Vulcan stops the engine after 5 minutes without use. The next result then starts it again.
  • The check never downloads anything. If the target is not downloaded, the block shows nothing. Run the model once to download it.
  • The result updates about 1.5 seconds after you stop editing.

Command line

Vulcan Community includes a command that runs the same check. It ships with the next release of the Vulcan Community command line, bundled in the npm package. It needs no WASM.

vulcan-community sizing <model.json> --design-heat-loss <file> [--json]
  • <model.json> is the preprocessed core input of the target. It is not a prepared model.
  • <file> is a design-heat-loss JSON file: { coldestHourC, setpointC, weatherName, zones }. The command does not run a HEM target.
  • --json prints the result as JSON: { status: "sized", coldestHourC, setpointC, weatherName, zoneCount, systems[], heatSources[], notes[] }, or { status: "unavailable", reason } when there is no usable design-heat-loss file.

A script in the Vulcan repository makes both files from a Rust HEM:FHS target: node scripts/hem-targets/design-heat-loss.mjs. There is no command-line route for HEM:EPC yet. HEM:EPC sizing runs only in the editor.

Exit codes:

Code Meaning
0 Vulcan computed a result. This includes Too small.
1 A zone in the design-heat-loss file is not an object, or has a missing or non-finite number. Or another failure: a usage error, or a model file that is missing or not JSON.
3 --design-heat-loss is not given, or the file is missing, cannot be read as JSON, or lacks coldestHourC, setpointC, weatherName or a zones object.

Example output, for the FHS demo model:

Heating sizing
Coldest hour in the Bedford weather file: −6.5 °C outside, 21 °C inside.

Heat source "hp": Too small
  Below −6 °C outside, the backup heater covers the gap.
  House needs 9.0 kW. Heat pump gives 8.8 kW on the design day and turns down to 3.5 kW.

System "radiators" (whole dwelling): Big enough at 55 °C
  Radiators give enough heat from 38.9 °C flow.
  Radiators give 21.3 kW at 55 °C flow. The house needs 9.0 kW.

System "backup" (whole dwelling): Not sized: Direct electric heaters are not sized yet.

Space heating only. Hot water is not included.

With a heat pump 1.5 times as big, the heat pump is still big enough at 39 °C, so the system lines read:

System "radiators" (whole dwelling): Big enough at 55 °C
  You could lower the design flow temperature to 39 °C. The radiators still give enough heat, and the heat pump runs more efficiently.
  Radiators give 21.3 kW at 55 °C flow. The house needs 9.0 kW.

Still stuck?

← Back to Documentation