A simplified methodology for baseline model evaluation and comparison

The accurate quantification of energy efficiency savings is critically important for applications such as Eskom DSM projects and Section 12L tax incentives. However, these savings cannot be measured directly and are generally calculated using a baseline as point of reference. The quality of a baseline will therefore have a significant impact on the final result. A review of several publications determined that regression models are often used as baselines. This article consequently gives a brief overview of how regression models are typically developed and evaluated. The research also highlights several misconceptions relating to the statistical evaluation of regression models. A new methodology is presented to simplify the evaluation of potential baseline models. The approach avoids the complexity of statistical evaluation by graphically presenting results. This novel approach enables the objective comparison of several potential models. The benefit is demonstrated with the application of three practical case studies.