Multi-objective service restoration of distribution systems using user-centered methodology

Abstract

A multi-objective problem formulation of service restoration and a three-stage methodology tailored to user-centered service restoration for large-scale, unbalanced distribution systems are presented. One feature is that it allows the experience and engineering judgment of users to be integrated into the three-stage methodology. Since solutions suggested by a method need to be compromised with engineering judgment, this methodology involves system users (at Stage II), meta-heuristics (at Stage I) and local heuristics in combination with restoration strategies (at Stage III) for efficient and robust performance in service restoration. An adaptive K-means method is incorporated in the proposed group-based PSO (Stage I) for global search. It also assists users in their decision making. The evaluation results of the proposed user-centered methodology on the IEEE 123-node test feeder and a 1011-node practical unbalanced network show its effectiveness and flexibility in restoring service after faults on these two test systems.

A modified backward/forward sweep-based method for reconfiguration of unbalanced distribution networks

“…In this work, we have selected the smallest test feeder (13-node test feeder) and largest test feeder (123-node test feeder) and modified them by adding switches connected to some nodes. Figure 2 and Figure 3 show the modifications applied to the two test cases based on [32] and [33]. Figure 2 and Figure 3, it can be observed that there are tie and sectionalizing switches.…”

Cost-Effective Outage Management in Smart Grid under Single, Multiple, and Critical Fault Conditions through Teaching–Learning Algorithm

“…Service restoration for large-scale unbalanced distribution through multistage modified adaptive K-means method was proposed in [14]. The optimal DGs' placement based on spanning tree search strategy for service restoration was proposed in [15].…”

An AC Mathematical Model for Solving Complex Restoration Problems in Radial Distribution Systems in a Treatable Runtime

“…Among the main heuristic methods that have been investigated for restoration in radial EDSs are heuristic rules based on expert knowledge [1], [2], a strategy of depth-first search on a decision binary tree [3], strategies that find a radial restoration topology through a constructive process of opening of branches on a meshed topology [4], [5], a heuristic that reconnects the de-energized areas in large-scale EDSs by minimal paths to the branches on the boundary with these areas [6], strategies of branches exchange besides the boundary of the de-energized area in large-scale EDSs [7], [8], a multi-agent approach [9], multi-objective optimization methods using an evolutionary approach and fuzzy sets [10] and using an evolutionary algorithm that incorporates specialized genetic operators and solves the problem in largescale EDSs [11], and a NSGA-II proposal that optimizes four objective functions in a hierarchical structure [12]. Other studied heuristic approaches have been a four-stage method that combines reconfiguration with intentional islanding of distributed generators, allowing the restored EDS to operate in an islanded way with these generators [13], a method based on multi-agent systems using expert systems rules for autonomous restoration in active EDSs [14], a three-stage heuristic strategy, whereby the system operator participates in the second stage, filtering a set of feasible solutions generated in the first stage to be improved in the third stage through a local search process [15], a two-stage proposal that does the optimal island partitioning in a smart EDS for feasible operation with distributed generators [16], a strategy based on the weighted ideal point method that solves the problem with multiple objectives [17], and a proposal using simulated annealing and a local improvement strategy, where the search space is represented by a set of permutation vectors composed by the switches, and the time for the switching operations is estimated using a scheduling approach considering multiple dispatch teams [18]. Other important topics on restoration have been investigated in [19]- [21], whose proposals are an iterative method that generates alternative energizing path schemes for restoration after a blackout [19], an alternating direction method of multipliers that separates the restoration problem into two sub-problems to solve it iteratively [20], and a method based on the reinforcement learning technique that can address the self-healing a...…”