Distributed Economic Dispatch of Microgrids Based on ADMM Algorithms With Encryption-Decryption Rules

Distributed economic dispatch (ED) has emerged as a critical issue in microgrid operations due mainly to the wide application of various clean energy as well as energy storage units. The openness of communication networks in microgrids can lead to privacy breaches, which pose a serious threat to the entire electricity market. As such, this paper presents a distributed ED algorithm based on the alternating direction method of multipliers (ADMM), where a quantization-based encryption and decryption rule is integrated to avoid privacy leakage while iteratively acquiring the optimal ED scheme. By resorting to the property of monotonically convergent sequences, a sufficient condition about the learning rate is profoundly revealed to guarantee the algorithm convergence. Two extended results are presented, respectively, to enhance the convergence rate and meet the requirement of plug-and-play scenarios. Finally, the validity (both privacy and optimality) of the proposed algorithm is verified by using the dual-source trolleybus system in Beijing. Note to Practitioners - This paper develops an engineering-oriented ED algorithm that optimizes the total generation costs of smart grids online while guaranteeing system constraints. Shared network communication undoubtedly plays a significant role in achieving iteratively the optimal solution of distributed algorithms. However, some crucial and sensitive information exchanged via an open and shared network could be eavesdropped by malicious attackers, which could result in a serious security threat affecting the reliability and stability of the smart grid. To overcome such a shortage, an encryption-decryption rule is constructed via a dynamic quantizer. In light of such a rule, the presented algorithm based on ADMM can iteratively acquire the optimal ED solution in a distributed way, realizing the requirements of optimality and privacy. The desired range of the learning rate is disclosed to guide the parameter selection, and two improved versions are proposed to meet more general engineering practice involving plug-and-play scenarios.