Adaptive Layers Based Atomic Orbital Search for the Co-Optimization of Network Reconfiguration with Battery Storage

Network reconfiguration (NR) is a widely recognized strategy for enhancing the distribution network (DN) performance by minimizing power losses and improving the voltage profile. This study introduces a novel variant of the atomic orbital search (AOS) algorithm that features adaptive layers selection, an improvement over the canonical AOS algorithm, applied for the first time to the NR problem. The number and span of layers in the proposed variant are adaptively calculated at each iteration per the log-normal distribution of the particles in the solution space. It is generic and converges more rapidly to optimal solution because of the adaptive layers selection feature. The proposed variant performance is evaluated by optimizing three benchmark NR cases of the IEEE 33-bus feeder. Moreover, to address the scalability concerns during the day-ahead electricity market operation, a multiperiod co-optimization problem of NR with battery energy storage system and photovoltaic units having 264 decision variables, is solved in an acceptable time. The simulation results outperform the reported results in the literature in terms of standard deviation in achieving the final solution.