Building energy controls systems all claim to improve the energy efficiency of a building, after all that’s the system’s primary job. How do we know that energy is being saved or that efficiency has improved? For example, if I have an air conditioning unit that ran for 12 hours total on one day then then only 2 hours on another day, does that actually represent a “savings” of energy? Perhaps the space temperature set point (the thermostat) was changed between those days. Maybe the outside temperature was different. Maybe one day was sunny and the other cloudy.

The point here is that just looking at run time or metered energy usage is not enough to really know much about efficiency. What is needed is a way to measure energy usage against energy required. For heating and cooling systems we can use a rudimentary unit of measure called a “degree day.”

A degree day is difference between the outside temperature and some reference temperature for an entire day. An industry standard is to use 68F as the reference and then to use the average outside temperature for the day, so if a day’s average outside temperature was 65F, then then you have 3 *heating* degree days. These are “heating” degree days because the actual temperature must be heated to reach the reference. Similarly, if the average outside temperature were 71F, then we’d have 3 *cooling* degree days. The basic premise is that it gives a measure of the amount of heating or cooling that would be required to adjust a space temperature from the outside temperature to a given set point.

The standard model of using 68F and the daily average has some problems, however. The 68F temperature is used because it assumes that the people, equipment and appliances would create some heat that would result in a 2-3 degree increase above the baseline temperature in the space. Often times we have a set point that isn’t 68F or that changes throughout the day. If it’s 71F outside and my thermostat is set at 71F then my HVAC system isn’t going to run. The standard model, however, would say that I had 3 cooling degree days. Zero run time for three cooling degree days would look like great efficiency, when the reality is that we just had a set point that differed from the standard model. Using the average daily temperature has similar problems because the temperature throughout the day is very rarely constant.

A better calculation of a degree day is to use the actual set point the heating or cooling system is running against and to calculate and sum partial degree days periodically. The more frequently you do the calculation, the better your approximation of degree days will be. A constant integration would be perfect but in reality, since outside temperatures don’t swing too quickly, calculating every 15 to 20 minutes or so gives a good approximation. Using the actual setpoint allows the model to reflect the demand we are placing on the building. For example, if we have a night set back of 60 degrees for heating then we do not add these extra heating degree days during this time and have a model that more accurately reflects the energy demands we are placing on the building.

As an example, let’s look at the outside temperature profile in Lansing, IL on April 17, 2017.

The average temperature for the day was approximately 54.2F. Now let’s look at that in terms or heating degree days using three models: A) The standard model with a set point of 68F B) A model that calculates the degree days every 20 minutes based on actual temperature and a fixed set point of 68F and C) A model that calculates degree days every 20 minutes, but using a set point that is 68F during the day and 66F at night.

Model | Heating Degree Days | % Diff from Standard |

Standard | 13.81 | – |

20-minute Calculation, 68F set point | 13.70 | 0.75% |

20-minute Calculation, 67F set point | 12.7 | 7.9% |

20-minute calc, moving set point | 12.7 | 7.9% |

As you can see, the standard model estimates higher energy usage than either of the other two models. Where the actual set point differs from the standard model, the difference can be significant; a nearly 8% difference for a one degree difference on the thermostat.

Once you have the data to calculate degree days, whether you use the standard model or an improved one, you can then look at the ratio of either run time or metered energy usage (therms, kilowatt hours, etc.) per degree day to get a feel for overall efficiency of the building system. The change in *this ratio* is what is of interest. If energy usage per degree day has decreased, then you’re more efficient and saving money.

This isn’t a perfect model. It can’t account for a wide array of factors like solar load, heat from occupants or open windows, but it is a directly measurable assessment of a building’s performance and when looked at over time it can be used to show a variety of things, from improvements due to an energy control system to even gradual degradation of the heating or cooling system itself.

The key is to have a building model that helps score energy efficiency and identify equipment failures or degradation with limited information. The periodic calculation of heating and cooling degree days against actual set point works to provide an acceptable model that accomplishes these goals.