Abnormality Detection and Privacy Protection Str ategies for CPSS Power Marketing Inspection Bu siness with the Collaboration of Big Data and AI

Kai Li^*Pingyan MoYongjiao YangHanyang XieZhixiong Shen

• Guangzhou Power Supply Bureau, Guangdong Power Grid Co., LTD., Guangdong, China

With the widespread adoption of Cyber-Physical-Social Systems (CPSS) in the power industry, power marketing inspection has become a critical component in ensuring the secure and reliable operation of power systems. However, it faces significant challenges due to the massive volume of inspection data and the complexity of electricity usage scenarios. Traditional inspection methods often fail to effectively identify potential risks and abnormal behaviors. To address this issue, this paper proposes an intelligent security identification model for power marketing inspection, integrating advanced artificial intelligence techniques to enhance security defense and risk management in CPSS-based power marketing systems. The proposed model incorporates a work order correlation matching algorithm, fault interval detection algorithm, electricity consumption prediction algorithm, and a business anomaly identification algorithm. First, users are categorized using multi-source data to precisely detect abnormal electricity usage. Then, the model employs a correlation algorithm to uncover intrinsic links between fault handling and electricity refund work orders for the same user, revealing potential security vulnerabilities. Subsequently, the fault interval detection algorithm is used to locate fault periods, and electricity consumption within these intervals is dynamically estimated using a prediction algorithm. Finally, an intelligent classification model based on Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM) networks is developed, leveraging key security features to accurately identify abnormal business behaviors. Experimental results on three publicly available power industry datasets demonstrate that the proposed model significantly outperforms traditional methods in terms of accuracy, recall, and F1-score for security event detection. This approach effectively enhances the safety and reliability of power marketing inspection in CPSS environments and offers a novel technical framework for power system protection and privacy preservation.