When dealing with billing or daily data, building energy usage can be organized into three categories: heating load, cooling load and base load. During cold weather conditions, a building needs to be heated to achieve a comfortable temperature; similarly, hot conditions require cooling. Other elements, such as refrigeration and lighting, remain more consistent over the course of the year and can be represented in the base load. The relationship between energy use and temperature is reflected in the building consumption profile shown below. Heavier use occurs during periods of high and low temperatures (summer and winter) while usage during more temperate periods (spring/fall) is more moderate. The OpenEEMeter fits a linear model to this data and can use this model to estimate the building’s energy consumption, given outdoor weather conditions. Depending on the fuel a building uses for heating or cooling, models with a single temperature coefficient (i.e., heating or cooling) may be more appropriate.
For each meter that is being analyzed, the CalTRACK methods require that a set of candidate models with known parameters be fit to the consumption and weather data. The OpenEEmeter fits four types of models to the data, as seen below. Each model type is fit multiple times to the same data, but with different temperature balance points ( within a known range between 30-90 F). Thus, a particular meter may end up with hundreds of potential candidate models. Some of these models are dropped due to insufficient readings below or above the temperature balance points (see CalTRACK guidelines for details). The actual model that is selected for that meter is the one with the highest adjusted R-squared (indicating that it does the best job at explaining the variance in the data). Since these are linear models, each model can be fully described by up to five coefficients (depending on model type): an intercept indicating the baseload energy consumption per day, a heating balance point and a heating slope, and a cooling balance point and a cooling slope. These model coefficients are then used to generate derivative outputs depending on the use case.
ASHRAE Guideline 14
UMP Chapter 8