Privacy-Preserving Optimal Battery Control for Energy Storage Systems Using Adaptive Dynamic Programming

This paper investigates the problems of high electricity costs due to price fluctuations and data transmission security risks in energy management systems. To mitigate these issues, first, a privacy-preserving communication framework based on encryption-decryption mechanisms is designed to enhance data security and prevent malicious attacks. Then, a battery management approach based on an adaptive dynamic programming algorithm with privacy protection is designed to optimize energy scheduling, reduce electricity costs, and extend battery lifetime. The algorithm is implemented using an actor-critic neural network architecture, and it is proven that the weight estimation errors are uniformly ultimately bounded. Finally, simulation results validate the designed approach’s capability in reducing electricity costs and enhancing data security.