A hybrid ensemble-optimized bidirectional gated recurrent unit method for short-term load forecasting of electric vehicle charging stations

Key Points

• The method significantly improves load forecasting accuracy compared to traditional models, enhancing operational efficiency.
• An average accuracy increase of 15% was noted when employing the hybrid ensemble optimization approach on data from several charging stations.
• Utilizing a bidirectional gated recurrent unit, the analysis captures temporal dependencies, offering superior prediction capabilities.
• Findings suggest that enhanced forecasting could support better grid management and energy distribution, impacting EV infrastructure planning.