Creating industry segment load profiles for energy system analysis from real company data

Industrial energy demand becomes increasingly important for energy system modeling and planning as forecasts expect industry to become the single largest energy consumer in a de-carbonized world. Load profiles for industrial electricity and heat demand, however, are rare and often specific or difficult to use. Confidentiality prohibits using real measured load data, especially for research. Heterogeneity and external dependencies, in particular, complicate modeling industrial energy demand. Previous research identified top-down modeling using linear regression as an established method to create representative load profiles. In this article, we present a load profile generation tool for individual industry segments. We developed a two-stage multiple linear regression to generate anonymized representative load profiles from real measured load data. The tool was implemented in MATLAB with a confidentiality-by-design approach. For validation, we used standard load profiles from the literature to analyze each part of our load profile tool. Additionally, we present load profiles for three industry segments created from real data collected in Germany. Results indicate that the anonymization process through two regression stages reliably removes confidential characteristics from load data even with as little as two input data sets.

Duke Scholars

Author Stefan M Goetz Psychiatry & Behavioral Sciences, Behavioral Medicine & Neur ...