Research On The Multi-objective Reactive Power Optimization Based On Modified Imperialist Competition Algorithm

Optimal reactive power dispatch(ORPD)is one of the most important control methods to ensure the economical,safe and stable operation of power systems.Its purpose is to adjust the reactive power flow of the system by adjusting known control equipment in order to seek the most reasonable reactive power distribution.With the continuous expansion of the scale of the power system,the factors that need to be considered for system operation are also increasing.When dealing with the ORPD problem,researchers no longer focus only on the optimization of a single objective,but consider optimizing more objectives simultaneously.Therefore,the research on multi-objective ORPD(MOORPD)has become a hot topic in the field of ORPD problems.In addition,due to the unique advantages of intelligent algorithms in processing optimization problems,the use of intelligent algorithms to solve MOORPD problems has also become the focus of scholars’ research.Therefore,based on the basic characteristics of imperialist competition algorithm(ICA)and MOORPD,a modified ICA is proposed and applied to solve the MOORPD problem.Firstly,the basic characteristics and related concepts of optimal reactive power dispatch problems and multi-objective optimization problems(MOPs)are introduced.Then,a MOORPD mathematical model with three objective functions,two equality constraints and two types of inequality constraints is established.And these objective functions include optimizing total active power losses,optimizing voltage deviation,and optimizing voltage stability index.Secondly,the basic idea and specific process of the original ICA are summarized.And based on the characteristics of ICA,a dynamic balance strategy and distance strategy are proposed to improve ICA.The effectiveness of the improved method is verified by benchmark functions,and then an enhanced ICA(EICA)was obtained.At the same time,in view of the multi-constraint characteristics of the MOORPD problems,a constraintdominated layering mechanism is proposed to solve the problem of state variables violating constraints,and a constraint-dominated EICA(CD-EICA)was obtained.In addition,in order to solve the multi-object solution set ranking problem when using ICA to process MOORPD problems,a quantitative sorting rule is proposed to optimize the screening mechanism of multi-objective solutions,and its performance in processing MOPs is verified by test functions.Then,the final modified ICA: CD quantization sorting-EICA(CDQS-EICA)was obtained.Then,in order to verify the effectiveness of the improved methods when dealing with MOORPD problems,CDQS-EICA and classic MOPSO,NSGA-III,and NSGA-II algorithms are implemented in three different scales of test systems with six cases.In addition,in order to comprehensively evaluate the performance of the improved methods,the convergence and stability analysis of the CDQS-EICA and the three classic algorithms were performed.And the generation distance GD and the hyper-volume HV were used to statistically analyze the distribution,approximation and diversity of the four algorithms.Finally,the experimental results prove that CDQS-ECIA can effectively handle the MOORPD problems.The distribution and approximation of the Pareto front obtained by CDQS-EICA are the best of the four algorithms,and the results of the best compromise solution BCS are also the best.At the same time,according to the results of performance analysis and statistical analysis,it is verified that CDQS-EICA has superiority in terms of convergence speed,stability,and diversity.In addition,compared with other approches reported in recent literatures,the solution of CDQS-EICA also has superiority.In summary,the above experimental results show that the modified imperialist competition algorithm CDQS-EICA proposed in this thesis is a comprehensive solution for solving MOORPD problems,and has certain technological breakthroughs to some extent.