ESTIMATING THE CHARGING CONSUMPTION FOR EVs USING A NOVEL NEURAL NETWORK TECHNIQUE

Abstract

In this study, we introduce the water wave optimized bidirectional long short-term memory (WWO-BLSTM) model for predicting the charging usage of electric vehicles (EVs).WWO can be utilized to optimize the charging schedules of EVs, enabling the flexible change of charging patterns. The estimation of EV charging use implements BLSTM, a model that analyzes sequential data in forward and backward directions. Initially, we collected a dataset that includes 10,595 unregulated charging operations from workplace charging. This dataset represents the diversity of EV charging .A comprehensive data cleansing procedure was performed. To ensure the suggested method is effective, we employ MATLAB software to conduct simulations. This model was able to obtain a recall of 96%, F1 score of 93%, accuracy of 88% and precision of 95%. We offer outstanding outcomes for the charging consumption of EVs using our innovative WWO-BLSTM methodology.

Keywords

• electric vehicle (evs)
• energy consumption
• prediction
• charging
• neural network
• water wave optimized bidirectional long short-term memory (wwo-blstm)