How much energy can smart building controls save?

Anyone who quotes a single percentage for every building is guessing. The saving depends on how badly the building is operated today, its operating hours, its plant, and whether anyone acts on what the controls reveal. Published figures are useful for setting expectations, but a credible estimate for a particular building is built up from identified opportunities, each sized from measured data, with overlaps removed. That gives an energy saving; turning it into money and judging the return is a separate step.

What the published figures actually say

Two often-cited US figures show both the potential and the limits of benchmarks.

The US Department of Energy states that, in commercial buildings, high-performance control sequences can deliver, on average, 30% annual HVAC energy savings for a range of building types, with similar reductions in peak HVAC demand. That figure covers HVAC energy only, not the whole building, and it describes the potential of high-performance sequences, not of any particular product.

The Smart Energy Analytics Campaign (2016–2020), which supported 104 organisations and about 6,500 buildings, documented median annual energy savings of 3% for energy information systems and 9% for fault detection and diagnostics.

Both are averages or medians across many buildings. A poorly scheduled building may do much better; a recently commissioned, well-run one may do much worse. Neither figure is a forecast for your building, and neither is a cost saving, since the money depends on tariffs, demand charges and contracts.

Building an estimate you can defend

A defensible estimate is assembled one opportunity at a time. Each begins with evidence of a fault or inefficient sequence, found in trend data, a site survey or a short monitoring period. It is then sized from measured loads and verified hours, not a generic percentage. Next comes a check on the action itself: who will correct it, what that requires, and whether comfort, ventilation or special uses constrain it. The basis of the calculation is recorded too, so that the estimate can be checked later.

Two measures cannot both claim the same kilowatt-hours

Measures that act on the same energy cannot each claim their full stand-alone saving. Scheduling and occupancy control both reduce running hours; a set-point change and a heat-recovery repair both reduce heating demand. Overlapping measures have to be applied in sequence.

Assessed separately, the savings are 150 × 20% = 30 MWh and 150 × 25% = 37.5 MWh, a total of 67.5 MWh. But once the schedule is corrected, only 120 MWh remains for the second measure to act on: 120 × 25% = 30 MWh. The combined estimate is 30 + 30 = 60 MWh, which is 7.5 MWh less than the simple sum. With several overlapping measures, the gap grows.

Decide how to measure before anything changes

The measurement method should be agreed while baseline data can still be collected, before work starts. Lower consumption after installation is not sufficient evidence on its own. The International Performance Measurement and Verification Protocol (IPMVP) explains why: savings cannot be measured directly, because they represent the absence of consumption, so they are calculated from measurements before and after a project with suitable adjustments for changes in conditions. Weather, occupancy and use all change, and the adjustments for them belong in the measurement plan.

Savings also have to last. Those from operational changes erode when overrides become permanent or schedules drift, so the effort of keeping them deserves a place in the plan alongside the effort of making them.