Multi-objective Differential Evolution Algorithm For Shipboard Power System Reconfiguration

Modem warfare puts forward higher requirements for the digitization and intelligentization of weapon system,warship is important combat and defense weapon platform in the navy.With the increased capacity of shipboard power system(SPS)and size of network,the operation and protection of shipboard power system becomes complicated.Reconfiguration technology that can restore the load power supply rapidly and to the maximum load after the failure or war loss of the SPS,is an important part of the SPS to ensure stable operation.It can enhance the vitality and combat effectiveness of the warship and provide knowledge of management and decision-making for operators at the same time.SPS reconfiguration optimization problems can be regarded as a multi?objective optimization problems with multiple constraints.Differential evolution algorithm is an intelligent evolutionary algorithm with simple structure,few parameters and strong capability of global search,which is widely used in solving optimization problems.Therefore,in this thesis,differential evolution algorithm is used to study the SPS reconfiguration.The core of the thesis is summarized as follows:In order to ensure the various strategies and control parameters of the differential evolution algorithm adapt to the optimal state and ethance the ability to avoid local optimization in different optimization stages of SPS reconfiguration,a composite adaptive differential evolution algorithm(CADE)was proposed.This method establishes a strategy pooling composed of multiple mutation strategies,and dynamically adjusting the population size under different strategies through the fitness promotion rate of individuals.The control parameters of excellent individuals were used to dynamically adjust mutation factor F and crossover factor CR to achieve a parameter adaptive adjustment.The method is verified by numerical simulations and the single-objective optimization problem of SPS reconfiguration with relative load as the objective function.The results showed that this method could skip the local optimization with better reconfiguration optimization performance and higher computational efficiency.Aiming at the multi-objective optimization problem of SPS reconfiguration,a composite adaptive multi-objective differential evolution algorithm was proposed.Briefly,CADE algorithm is extended and tuned to the optimal framework of NSGA-11,which improves the individual selection method,the calculation method of the promotion rate and the update method of external archives.In order to reduce the computational complexity of the algorithm,an improved non-dominating sorting technique HNDS is introduced.The method is verified by numerical simulations and multi-objective optimization of SPS reconfiguration with the objective function of relative load and switching operation cost.The results showed that the solution set of the reconfiguration scheme obtained by this method has higher convergence accuracy,faster convergence speed,better distribution of solution set and better reconfiguration performance of SPS reconfiguration.In order to obtain a richer and more targeted reconfiguration solution in different working conditions,more reconfiguration objective functions have been considered in the process of SPS reconfiguration.A total of six indicators,including sum of load weight,load capacity,switching operating cost,generator balanced,line rich capacity,load distribution uniformity,are listed as objective functions for SPS reconfiguration,and set up the priority of each objective function under different working conditions,respectively.As the SPS reconfiguration is a many-objective optimization problem,an NSGA-?-CADE algorithm is proposed to solve the optimization problem,which improved the uniformity and distribution performance of the NSGA-? algorithm by using the method of mutation strategy and control parameters adaptive adjustment in the CADE.A phased evolution method,combining with the NSGA-?-CADE algorithm to solve the many-objective optimization problem of SPS reconfiguration,is proposed because the weight method is difficult to reflect the priority level of the objective function under different working conditions.The method is verified and analyzed through two typical fault cases under three different working conditions.The results showed that the reconfiguration solution proposed by this method has better diversity and more targeted compared with the conventional multi-objective optimization method.