Electric Vehicle Charging Load Time-Series Prediction Based on Broad Learning System

Abstract

Accurate Electric Vehicle (EV) charging load time-series prediction is an important prerequisite for enhancing the safe and stable operation of charging stations. However, the EV charging load is strongly nonlinear, highly intermittent and random, which leads to the low accuracy of charging load time-series prediction. To this end, this paper proposes a broad learning system-based EV charging load time-series prediction method. First, the actual data of charging load of EV are analyzed and processed. Further, a charging load time-series prediction model is established using a broad learning system. Simulation experiments based on actual data indicate that the proposed charging load time-series prediction model based on the broad learning system has better prediction performance and also has less computing time compared to prediction models such as back propagation neural network and long-short term memory.