Optimal Distribution Network Expansion and EV Charging Station Allocation Based on Load Forecasted Using Machine Learning

Effective expansion planning is essential for maintaining the stability, reliability, and longevity of distribution networks amidst increasing load demands. This study focuses on strategic expansion planning, incorporating the optimal placement of electric vehicle charging stations (EVCSs) and port numbers in an overlapped road and distribution network. It also addresses the replacement of branch conductors to prevent network overloads, along with optimal scheduling of electric vehicle (EV) loads to reduce its impact during peak loading conditions. Further, the increase in the harmonic distortion due to the increasing EVCSs and other nonlinear loads is mitigated through the optimal allocation of active power filters (APFs). Transformer derating due to harmonics is also evaluated, and the transformer rating is upgraded accordingly. For robust expansion planning, future load demand is predicted using Random Forest (RF) regression, while uncertainties in EV driving patterns are managed using the 2m-point estimation method (2m-PEM). During solution of the expansion planning problem, this paper focuses on minimizing energy losses and costs while maximizing the average happiness index (AHI) of the EV users. A multi-objective optimization framework is proposed and tested on a practical 40-bus distribution network to ensure that the network effectively accommodates future demands and enhances overall system performance.