Novel hybrid market price forecasting method with data clustering techniques for EV charging station application

In addition to provide charging service, Electric Vehicle (EV) charging station equipped with distributed energy storage system can also participate in the deregulate market to optimize the cost of operation. To support this function, it is necessary to achieve sufficient accuracy on the forecasting of energy resources and market prices. The deregulated market price prediction presents challenges since the occurrence and magnitude of the price spikes are difficult to estimate. This paper proposes a hybrid method for very-short term market price forecasting to improve prediction accuracy on both nonspike and spike wholesale market prices. First, Support Vector Classification is carried out to predict spike price occurrence and Support Vector Regression is used to forecast magnitude for both non-spike and spike market prices. Additionally, three clustering techniques including Classification and Regression Trees, K-means, and Stratification methods are introduced to mitigate high error spike magnitude estimation. The performance of the proposed hybrid method is validated with the ERCOT wholesale market price. The results from proposed method show significant improvement over typical approaches.