Understanding Industrial Energy Use through Sliding Regression Analysis

Carl Eger, City of Cleveland
Kelly Kissock, University of Dayton

Keywords

Abstract

Due to rising energy costs and global climate change, many industries seek to improve their energy efficiency. This paper describes a three-step method to analyze utility billing, weather, and production data to understand a company’s energy performance over time. The method uses regression modeling of utility billing data against weather and production data. The regression models are then driven with typical weather and production data to calculate the ‘normal annual consumption’, NAC. These steps are repeated on sequential sets of 12 months of data to generate a series of ‘sliding’ NACs and regression coefficients. The method can quantify successful energy efficiency initiatives and lend insight into the cause of the energy savings. In addition, the method is able to proactively identify energy saving opportunities. The method is demonstrated with a case study. The case study shows that the method is able to disaggregate energy use into weather, production and independent components, accurately measure changes in plant energy efficiency, lend insight into the nature of the those changes, identify savings opportunities and identify changes in overall process control.