A process for detecting statistically significant changes in energy consumption patterns by monitoring for changes in the parameters to a parametric energy model. Two parametric models of energy consumption are created: the first being a model providing an initial base line of energy consumption, the second being a test model to be compared to the initial base model. Statistically significant changes are detected by using a difference score that compares the parameters of two models along with the uncertainties of each parameter to determine whether the differences in the parameters of each model indicate a statistically significant deviation in the energy consumption pattern.