Retrofit & economics

Lighting power density: the compliance number that decides your retrofit before efficacy does

How building energy codes cap installed lighting power per square metre, why an efficacy-led product choice can still fail the check, the two ways LPD is assessed and how they differ, control credits, and the calculation to run before the fittings are selected rather than after.

Written by Yilian Zhikong Technology (Shenzhen) Co., Ltd. · Last reviewed August 2026

Lighting power density is the quantity that decides whether a scheme is legal, and it is frequently the last thing anybody calculates. A specification is written to an illuminance target, fittings are selected on efficacy and price, the scheme is designed, and then somebody asks for the building regulations compliance figure and discovers the installed load is over the cap by a margin that cannot be fixed by swapping a component.

What it is, and what it is not

LPD is installed lighting power divided by floor area, in watts per square metre. Installed power means the total circuit watts — luminaires plus their control gear — present in the space, whether or not they are on. It is a property of what is fixed to the ceiling, not of how the building is operated.

That definition catches people out in two directions. Controls do not reduce LPD, they reduce energy consumption, and most codes account for them separately if at all. And a scheme designed to exactly the required illuminance with an inefficient distribution has a high LPD even if every fitting in it is efficient, because it needed more of them.

Use measured input power, not the nominal wattage on the label. The gap between them is the driver's tolerance and is systematically in the unhelpful direction — the reasons are set out in luminaire efficacy versus source efficacy.

Two assessment methods, two different answers

Codes generally offer a choice, and which one you use materially changes the result.

• Building area method. One allowance for the whole building based on its type — office, retail, warehouse. Simple, quick, and generous to buildings whose spaces are mostly the low-demand kind.
• Space-by-space method. An allowance per space type, summed. More work, and usually the only way to pass a building with a few genuinely demanding spaces, because it lets the demanding space have its allowance without dragging the whole building's average down.

Which allowances apply is jurisdictional and is revised on its own schedule. In mainland China the caps sit in GB 50034 alongside the illuminance requirements, which has the useful property of putting both numbers in the same document. Elsewhere they live in a separate energy code — ASHRAE 90.1 and the IECC in the United States, Part L and its approved documents in England and Wales, the relevant national implementation of the EPBD across the EU. Take the current edition; the caps have tightened repeatedly and a value remembered from a previous project is likely to be generous.

Why an efficacy-led selection can still fail

Efficacy is lumens per watt out of the fitting. LPD depends on how many watts you had to install to hit the illuminance target, which depends on how much of each fitting's output reaches the working plane. The bridge between them is the utilisation factor, and it is a property of the distribution, the room geometry and the surface reflectances — not of the product's efficiency.

The practical consequence: a fitting with slightly lower efficacy but a distribution matched to the room can produce a lower LPD than a more efficient fitting that lights the ceiling. So the order of operations that works is to run the lumen method for each candidate, multiply the resulting count by that candidate's measured input power, and divide by the area — compare the products on that number rather than on the datasheet. It takes ten minutes per candidate and it is the only comparison that answers the question being asked.

The same reasoning applies to glare control. A deep-cell louvre that gets a fitting to a UGR limit costs light output and therefore costs LPD. Both requirements are usually mandatory, and discovering the tension between them after the product is selected is expensive.

Control credits, and their limits

Some codes allow an adjustment to the installed power figure where specified controls are present — occupancy sensing, daylight regulation, or automatic shut-off beyond the mandatory minimum. Where they exist these are worth having, and three cautions apply. The credit is usually conditional on the control actually being installed and commissioned, not merely specified. It is usually capped. And it is not a substitute for a design that fails the base requirement by a wide margin — a credit closes a small gap, not a structural one.

Where the credit is claimed, the commissioning evidence becomes part of the compliance record, which raises the stakes on the settling-in work described in the controls retrofit note. A sensor disabled by facilities after handover has removed a compliance credit, not just a saving.

The order to do it in

1. Establish the applicable code, edition and assessment method, and whether the retrofit is in scope at all, before selecting anything.
2. Get the illuminance and uniformity requirement for each space from the lighting standard, since LPD compliance is only meaningful alongside it — a scheme can always pass LPD by being too dark.
3. Run the lumen method per candidate product and compute LPD from the resulting quantity and its measured input power.
4. Check the demanding spaces individually. One over-lit space with a tight allowance can fail a building whose average is comfortable.
5. Record the calculation with the reflectances, maintenance factor and grid alongside it, because the compliance submission will be checked against it and, unlike a datasheet, it has to be reproducible.