Buildings’ heating systems are among the most important things to monitor in an energy management scheme, not just because they are big users but also because they have so much potential for waste. Poor control can leave them running at times when they are not needed or delivering more heat than is strictly required (or both). Remember that for most of the year they need only a fraction of their maximum output, and that’s without accounting for the fact that they are likely to be oversized in the first place, so there is plenty of headroom for excess consumption. Poor maintenance can cause them to use more fuel than necessary for the heat that they produce, and in buildings with air conditioning, you can end up with simultaneous heating and cooling.

Just to make matters worse, seasonal variation in the weather imposes huge seasonal swings in the quantity of fuel that a building will need, and even during autumn, winter and spring there can be large variations in the weather from week to week. This complicates the management task: how can we work out what consumption should have been each week, so that we can detect exceptions and quantify their impact? As with everything, we need a formula for expected consumption, but what should we use as the independent driving factor?

Average prevailing outside air temperature looks like a good candidate. It meets the three tests for use as a driving factor: (a) its variation causes variation in energy demand; (b) it has a numerical value; and (c) it routinely varies from week to week. But average temperature has two drawbacks, the first being that the expected-consumption formula is non-linear. Consumption will tend to fall proportionately as outside air temperatures rise, but at a certain value (the building’s ‘balance point’) consumption will reach a minimum and there it will stay at progressively higher temperatures. A two-part formula like this is known as a break-point model. The second drawback of average temperature is that its relationship with consumption breaks down when the outside air temperature in a given period straddles the building’s balance point. Take a building with a balance point of 15C and suppose on a particular day the temperature swings between 14C and 10C (an average of 12C). On the next day it swings between 15C and 9C. This is still an average of 12C and the extra fuel used when it is colder is compensated by a reduction when it was less cold, so the outcome would be the same (-ish). But suppose on the third day the temperature swings between 18C and 6C. This is still an average of 12C. But now increased demand at the cold times cannot be compensated by reduced demand at the warmer times, because when it’s above 15C outside the heat demand will have hit the stops at zero and cannot go any lower. This means that there is not after all a determinate link between average air temperature and fuel demand.

To get around these problems we pre-process our outside air temperature measurements for the week (or month) to create a number called the degree-day value for the region we’re interested in. There are a couple of ways to do this but the important thing is that fuel demand can be expected to have a straight-line relationship with the degree-day value. Degree-day figures are readily available free or on paid subscriptions.

Return to A to Z of M&T