https://www.mdu.se/

In today's district heating (DH) energy market, it is common to use user functional categories in price models to determine the heat price. However, users in the same category do not necessarily have the same energy consumption patterns, which potentially leads to unfair prices and many other practical issues. Taking into account heat usage characteristics, this work proposes two data-driven methods to cluster DH users to identify similar usage patterns, using practical energy consumption data. Efforts are focused on extracting representative features of users from their daily usage profiles and duration curves, respectively. Employing clustering based on these features, the resulting typical usage patterns and user category distributions are discussed. Our results can serve as potential inputs for future energy price models, demand-side management, and load reshaping strategies.