Cloud-Based Frequency Control for Multiarea Power Systems: Privacy Preserving via Homomorphic Encryption
Kun Xu, Yugang Niu, Zhina Zhang
Abstract
This article proposes a cloud-based load frequency control (LFC) scheme with privacy preserving for multiarea power systems. To prevent sensitive information of power systems from disclosure in the cloud service, Paillier cryptosystem is adopted to encrypt the sampling data. An encrypted LFC scheme is constructed by means of the homomorphism of Paillier cryptosystem, that is, the encrypted output does not need to be decrypted in cloud service. Considering that a bigger key size provides better privacy level, but will result in longer time delay, the tradeoff between privacy preserving level and control performance is analyzed by utilizing deterministic network calculus to evaluate the delay bound. Meanwhile, the decentralized characteristic of designed LFC scheme can avoid possible privacy leakage during the controller design procedure. Moreover, an optimization method is proposed for searching the suboptimal control gains. Finally, a simulation on a three-area power system is presented, in which 4 different key sizes from 256 to 2048 bits are tested to verify the effectiveness of the proposed scheme.