Evaluation Method for Peak Shaving Capability during Heating Period Based on Data‐Driven and Heating Unit Characteristics

Abstract As the proportion of new energy installed capacity grows, cogeneration units are evolving into regulating power sources. However, the operation mode of ‘using heat to determine electricity’ results in limited peak shaving capability. A comprehensive method to accurately calculate and grasp the peaking shaving capability of cogeneration units is proposed. Firstly, the typical operating conditions of the units are determined through the clustering method of the Gaussian mixture model (GMM) combined with K‐means. Secondly, the simulation model of heating unit operation under variable working conditions is established. Finally, a heat load forecasting method based on long short‐term memory network (LSTM) is proposed, and the peak shaving capability of the unit is calculated according to heat demand and thermoelectric relationship. The superiority of the proposed method is verified by taking a heating unit in Ningxia Hui Autonomous Region, China as an example. © 2025 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.