Defense against false data injection attacks on the electric vehicle charging stations data markets

Abstract

Data-driven technology depends on high-quality training data. Although many research institutions advocate for data sharing, private data owners are often reluctant to share their data considering the privacy concerns related to potential data breaches. As a result, the availability of data limits the application of data-driven technologies in energy systems. To enhance the availability of data, we have constructed a data market model for forecasting the power demand of electric vehicle charging stations (EVCSs), enabling data transactions within this market to improve the accuracy of forecasts. Since the data market relies on communication networks to collect data, this makes it vulnerable to malicious false data injection attacks (FDIAs) during transmission, exposing the data market to serious security risks. To ensure the safe operation of this data market, this article proposes a defense method based on Wasserstein Generative Adversarial Network that combines Transformer and Convolutional Neural Network(TCWGAN). This method effectively reduces the impact of FDIAs and has a strong defense against the injection of false data, achieving an accuracy of 95.09 %.