Energy benchmarking

In the UK there is a tradition of benchmarking buildings in terms of their annual energy consumption, either against each other, or against published yardstick figures. The purpose is to identify poorly-performing buildings and infer the scope for improvement.

The measure usually used for comparison is the normalised performance indicator (NPI). This is the annual consumption per square metre of floor area -- normalised for weather, exposure, and hours of use.

The adjustment for weather is necessary because the buildings being compared could be in different regions, or could have had their annual consumptions measured at different times; either way, they would have been exposed to different weather and this will skew the results. The adjustment procedure corrects the measured consumption to that which would have occurred if the building had experienced a standard weather year, which is taken as 2,462 degree days (if the actual weather had been colder, for example, the consumption would need to be reduced to make the comparison fair).

To do the adjustment we need two figures:

The adjusted annual consumption A is then given by F + (V x 2462 / D)

F can be easily estimated if monthly consumption data are available: it is merely 12 times the consumption in a summer month when no heating is used. V is the difference between F and the total annual energy use.

In the official government advice, an additional adjustment factor is allowed, with some buildings deemed to be "sheltered" or "exposed". In this model a sheltered building's weather-related variable consumption should be inflated by factor 1.1, while an exposed building's consumption should be deflated by factor 0.9. Actually the issue is not whether the building is exposed or not: it is whether the local climate differs systematically from the published regional values.

Published yardstick values are for typical occupancy, and buildings with different occupancy regimes need further correction. A simple pro rata adjustment is sometimes suggested, but this is inappropriate for heating fuel, since the heat loss rate from a building responds only slowly as the heating is started and stopped.

If for example one were dealing with a building occupied during standard office hours, five days a week, I would suggest applying the following corrections to buildings occupied for longer hours.