Energy used for a given purpose will often have an element of fixed overhead consumption, sometimes called the ‘base load’, which does not vary from one week to the next. Overlaid on that fixed consumption will be a variable proportion which will change according to how cold the weather is, how much production throughput there was, or whatever variable ‘driving factor’ is relevant for the case in question. Mathematically the expected consumption volume E is related to a driving factor D by a simple straight-line formula:

E = c + m.D

where c and m are constants: m being the gradient of the line on the scatter diagram of consumption against driving factor, and c being where the line **intercepts** the vertical axis. The intercept represents the fixed base-load consumption (per week or whatever interval) while the gradient tells us how consumption changes with variations in the driving factor. In common practice the values of c and m are usually derived by a statistical process, regression analysis, which uses historical values of actual energy consumption and the corresponding driving factor. In a manufacturing process the intercept tells us how much energy is needed to sustain the process without contributing to actual output. Applied to a building’s heating system, the intercept on the scatter diagram of fuel use against degree days can give us an estimate of the non-weather-related consumption which comprises plant standing losses plus, perhaps, domestic hot water or catering (this analysis comes with the caveat that one’s degree-day figures need to have been computed relative to the correct base temperature—a topic that there is not space to explore here).

We can apply the same principle to industrial utility plant as well: objects like air compressors and steam boilers. These use electricity and fuel in proportion to their output of compressed air and steam respectively, but they also have fixed losses which regression analysis will quantify. I have even used the technique to analyse the fuel used by delivery lorries against miles driven: here, the intercept is an indication of idling while stationary.

Focussing on those fixed overhead consumptions can be a profitable route to energy saving, and regression analysis is a quick way to establish how big those base loads are.