DEEP LEARNING ELECTRICITY THEFT DETECTION MODEL BASED ON FUNCTIONAL ENCRYPTION STRATEGY

To achieve effective privacy protection and efficient electricity theft detection, this paper proposes a deep learning electricity theft detection model based on functional encryption. Each user encrypted power consumption readings via functional encryption, while operators utilized decryption keys to compute dynamic bills and monitor grid load. A deep learning model was designed to detect electricity theft by performing inner product operations on encrypted data. Simulation results demonstrate the model’s accuracy in identifying fraudulent users while maintaining privacy with acceptable communication and computational overhead.