Multi-objective Reactive Power Optimization For Power Systems

With the rapid development of the social economy,people’s electricity demand increases,and the requirements for power quality and living environment improve,single-objective optimization can no longer meet the requirements of modern power system operation and management.Multi-objective reactive power optimization has gradually become a hot spot of research nowadays.Scientific and reasonable optimization of multiple objectives can improve the economic efficiency,safety and stability of a power system,which is of great significance for the development of power systems.Multi-objective reactive power optimization of a power system is a complex nonlinear programming problem,which also suffers from the difficulties of high dimensionality and non-convexity.Not only do conflicting objectives have to be optimized simultaneously,but also numerous constraints in the power system have to be considered.There are many problems when solving using conventional optimization methods,such as long computation times and inaccurate optimization results.Therefore,it is necessary to use new optimization methods or to investigate new and improved strategies to deal with such problems more effectively.The decomposition-based multi-objective evolutionary algorithm is a powerful tool for solving complex multi-objective optimization problems such as reactive power optimization,but the algorithm itself has some problems and there is room for improvement.In this paper,based on the original MOEA/D,two improved decomposition-based multi-objective optimization algorithms are proposed.The first improved algorithm introduces three strategies:competitive evolution,adaptive variation and similar selection,while the second improved algorithm introduces an adaptive evolution strategy and an external archiving method.Both algorithms improve the quality of the optimization results and obtain solution sets with uniform distribution and good diversity,providing decision-makers with feasible solutions of high quality and achieving good decision-aiding effects.Finally,the improved algorithm is applied to the multi-objective reactive power optimization calculation for power systems,using the IEEE 30-bus and IEEE 57-bus simulation test systems as examples to solve reactive power allocation schemes that satisfy the minimization of generation costs,emissions,power losses and voltage deviations,with calculated results including the optimal Pareto front and optimal objective function values.The effectiveness of the improved strategy is verified by comparison with the original MOEA/D.And by comparing with other classical multi-objective optimization algorithms,the superiority of the two improved algorithms in dealing with multi-objective reactive power optimization problems in power systems is demonstrated.