Decentralized Optimal Power Flow for Multi-Agent Active Distribution Networks: A Differentially Private Consensus ADMM Algorithm
Chao Lei, Siqi Bu, Qifan Chen, Qianggang Wang, Qin Wang, Dipti Srinivasan
Abstract
In multi-agent active distribution networks, the information exchanges in the ADMM algorithm for the decentralized distribution-level optimal power flow (D-OPF) may expose sensitive load flows of tie-lines across adjacent agents. This may be overheard by adversarial agents for business competition. To preserve this privacy, this paper proposes a differentially private consensus ADMM (DP-C-ADMM) algorithm, which can offer a mixture solution of both realistically optimal generator outputs and obfuscated-but-feasible load flows of tie-lines. And <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\epsilon -$ </tex-math></inline-formula>differential privacy holds for load flows of tie-lines across agents over iterations. Case study justifies the theoretical properties of this algorithm up to specified privacy parameters.