Scalable Distributed Non-Convex ADMM-based Active Distribution System Service Restoration

TL;DR

This paper introduces a scalable distributed optimization framework using a non-convex ADMM approach for active distribution system restoration, enabling efficient DER scheduling and network reconfiguration in large-scale networks.

Contribution

It develops a novel non-convex ADMM-based method with relax-drive-polish phases for distributed distribution service restoration, addressing binary decision challenges.

Findings

Effective restoration in large-scale networks demonstrated on IEEE test feeders

The method achieves high-quality solutions with good scalability

Distributed reconfiguration and load pickup are successfully implemented

Abstract

Distributed restoration can harness distributed energy resources (DER) to enhance the resilience of active distribution networks. However, the large number of decision variables, especially the binary decision variables of reconfiguration, bring challenges on developing effective distributed distribution service restoration (DDSR) strategies. This paper proposes a scalable distributed optimization method based on the alternating direction method of multipliers (ADMM) for non-convex mixed-integer optimization problems and applies to develop the DDSR framework. The non-convex ADMM method consists of relax-drive-polish phases, 1) relaxing binary variables and applying the convex ADMM as a warm start; 2) driving the solutions toward Boolean values through a proximal operator; 3) fixing the obtained binary variables to polish continuous variables for a high-quality solution. Then, an autonomous clustering strategy together with consensus ADMM is developed to realize the distributed cluster-based framework of restoration. The nonconvex ADMM-based DDSR can determine DER scheduling and switch status for reconfiguration and load pickup in a distributed manner, energizing the out-of-service area from local faults or total blackouts in large-scale distribution networks. The effectiveness and scalability of the proposed DDSR framework are demonstrated through testing on the IEEE 123-node and IEEE 8500-node test feeders.