Dual Consensus ADMM for Node Based DC OPF Considering Communication-censoring Strategy

Abstract

A common belief in power system optimization is that privacy preservation can be enabled by implementing distributed dispatch algorithms. To solve the node based DC optimal power flow (OPF) problem with limited communication resources in a distributed way, this study proposes a communicationcensoring strategy (CCS) to reduce unnecessary communiation cost by using alternating direction method of multipliers (ADMM). Out of further privacy concern, dual consensus is formulated by sharing Lagrangian multipliers associated with coupling constraints between the coupled adjacent buses rather than phase angles that are usually shared. The communication mechanism considered here is censored which judges whether the difference between the latest variables and the previously transmitted counterpart has reached a well-designed censoring function. Though CCS may slow down the optimizaton process, it reduces the communication cost effectively. A simple numerical example illustrates inherent tradeoffs between commnication cost and number of iterations.