Determining Locational Hosting Capacity of DERs in Distribution System Using Machine Learning

The integration of Distributed Energy Resources (DERs) utilizing renewable sources of energy into distribution system is on rise in due to environmental challenges and advancements in technology. The assessment of Hosting Capacity (HC) enables utilities to determine the maximum allowable installation capacity of DERs into distribution networks without violating operating limits. In this paper, a Machine Learning (ML) based approach is proposed to determine the maximum locational HC of DER and their Limiting Constraints (LC) to be connected at each node considering the limits of system voltage, thermal and back-feed limits. The proposed approach utilizes two types of ML regression models-Feed Forward Neural Network (FFNN) and Random Forest (RF). The assessment of Locational DER HC is investigated on IEEE-33 bus balanced radial system. The high accuracy and fast computation of the proposed method highlight its suitability for online applications.