Enhancing the performance of power distribution systems through integrated network reconfiguration and distributed generation design

Abstract

Reconfiguration of networks and distributed generation (DG) together leads to better performance of a network. To ensure system enactment, it is therefore necessary to determine appropriate size and placement of DG. However, there is a huge solution search space for sizing and situating of demand generation with Network Reconfiguration (NR), which makes it a complicated problem. Throughout the optimization process, removing these non-radial choices adds computational burden and lead to a local optimal solution. To reduce complexity of searching, Modified Chaotic Particle Swarm Optimization (MCPSO) algorithm is adopted to obtain a near optimal solution of designing, sizing, and placing the network with improved voltage profiles and minimized power loss. It introduces a combination of chaotic inertia adaptation, uniform initialization and a stochastic personal learning strategy contributing to improved search diversity and convergence stability. For the purpose of demonstrating efficacy of a simultaneous approach taking changeable power factor, the proposed approach is assessed using IEEE-33 and 69 bus using MATLAB. The findings demonstrate that discretizing reconfiguration search space implemented by encoding the network configuration as a discrete set of switching states prevents MCPSO from getting trapped in local optimums. On contrasting with conventional Particle Swarm Optimization (PSO), the proposed MCPSO algorithm results in active and reactive power loss reduction of $27.78\%$ and $76.36\%$ respectively for 33 bus system and $6.67\%$ and $25.5\%$ respectively for 69 bus system. The outcomes reveal that suggested algorithm provides optimal solution contrasted to state of art approaches.