Reconfiguration of Distribution Networks using a Genetic Algorithm with Rank-Weight Selection and Constrains

摘要

Distribution network reconfiguration is categorised as a complex optimisation problem. The main goal for these problems is to determine the optimal path with the least power losses. Once this goal is achieved, a second goal is to achieve the best computation time possible. Reconfiguration of a distribution network is a cost-effective solution to improve the overall efficiency compared to other optimisation techniques. Multiple different algorithms may converge at the same optimal solution; however, not with the same computation time. Therefore, a Genetic Algorithm with an improved selection function, Rank-Weight selection, and multiple constraints are proposed. The proposed optimisation technique is used to search a population of switches for the optimum path in the distribution power network. The population of the GA contains several chromosomes, represented by a string of binary values. Because only radial networks are feasible solutions, the fundamental loop approach is implemented to improve the search and eliminate nearly all non-feasible solutions. The chromosomes will only consist of open switches in the network, reducing the computation time. By combining all these techniques, the computation time to find the optimal solution was improved.