Optimal Power Flow Based On Improved Differential Evolution

With the large-scale grid integration of distributed energy systems,distributed power systems and distributed smart grids,as well as the rapid development of new energy sources and diverse loads,power system optimization has become an extremely challenging and significant task.Optimal power flow,as one of the main tools for power system operation and planning,helps to optimize the control configuration of the grid so that the power system operates in an economically safe and reliable environment.In general,the optimal power flow optimization problem is considered as a non-linear,non-convex,large-scale numerical optimization problem that relies on line and bus data for its complexity.Therefore,how to design an effective optimization method directly determines the performance of the optimal power flow configuration power system.As a simple and efficient intelligent optimization algorithm,the differential evolution(DE)has received great attention and been widely used in different areas due to its good convergence,robustness,and ability to find the global optimum.However,there are still some problems that need to be solved when the traditional DE and its variants are used for optimal power flow,such as the parameters of the DE itself need to be manually pre-determined according to different problems,how to effectively handle the constraints of a large number of equations and inequalities in the optimal power flow,how to consider balancing multiple optimization objectives and whether multiple optimal power flow optimization problems can be solved simultaneously.In this thesis,the optimal power flow based on DE is investigated as a further improvement of DE to provide new ideas and methods to help optimize the optimal power flow in power systems with a single objective,multiple objectives,and multiple optimal power flow simultaneously,thus providing a more effective tool for power system optimization.The main research contents of this thesis are as follows.(1)Adaptive constrained DE for solving conventional optimal power flowThe DE has good global search capability,but its performance is easily affected by its algorithm parameters,which are often difficult to give according to different optimal power flow.Therefore,an adaptive constrained DE is designed to solve the single-objective optimization problem in traditional optimal power flow.The algorithm mainly addresses the shortcomings in the adaptive DE(JADE)by firstly designing a crossover rate sorting mechanism based on the value of the objective function;secondly designing a strategy to reuse the successful evolution direction;and finally adopting a feasible solution priority rule to effectively deal with the constraints in the optimal power flow to ensure the feasibility of the solution.Its performance is first verified on13 benchmark functions,followed by experiments on the optimal power flow in the standard IEEE-30 bus system,and the results show that the proposed method has significant performance improvement on different cases.(2)Adaptive DE for solving optimal power flow integrating renewable energy sourcesAs most conventional power systems use non-renewable energy as the raw material for thermal power generation,this tends to lead to depletion of non-renewable energy and environmental pollution.To mitigate this power system model,the IEEE-30 bus system that integrates renewable energy sources is proposed,and then an improved DE is designed to solve the single-objective optimal power flow optimization under this model.In the method,the problem of random allocation of the crossover rate and artificially given population size in DE is addressed.Firstly,a crossover rate allocation strategy based on the value of the objective function is proposed to enable adaptive allocation of crossover rate to the corresponding individuals in the population;secondly,a dynamic population reduction method is designed to address the problem of user-defined population sizes in DE;finally,an adaptive penalty function is used to handle the constraints in the optimal power flow.The experimental results show that the method is more competitive in optimizing the optimal power flow for this integrated renewable energy source.(3)Multi-objective DE for solving optimal power flow integrating renewable energy sourcesOptimal power flow for integrating renewable energy sources have received considerable attention due to the ability of power generation systems that integrate renewable energy sources to meet daily demand while effectively reducing harmful emissions.However,most of the work is based on single-objective optimal power flow,which means that only one objective is often considered when carrying out optimization,while other important optimization objectives are neglected.For this reason,a DE based on multi-objective optimization is proposed to solve the multi-objective optimal power flow for integrating renewable energy sources.In this method,firstly,the classical multi-objective optimization algorithm NSGA-Ⅱ is investigated and a multi-objective DE based on non-dominated sorting is designed;secondly,a crossover rate adaptive mechanism is designed to ensure the diversity of gene interactions in the population;finally,a feasible solution priority rule is adopted to ensure that the solution provided by the improved adaptive multi-objective DE is feasible.Simulation results show that the designed method can effectively consider multiple optimization objectives simultaneously,and the resulting optimal configuration can help improve the stability of this power system as well as reduce the corresponding generation costs and real power losses.(4)Evolutionary multitasking DE for solving multiple optimal power flow simultaneouslyIn the case of the previous method for optimal power flow optimization in power systems,only one system optimization problem can be solved in a single run.For example,when solving for the IEEE-30 bus system,the IEEE-57 bus system cannot be solved at the same time.This results in the algorithm needing to start the optimization search all over again when solving other systems,ignoring the similarity that may exist between different power systems.In other words,the knowledge of solving one optimal power flow optimization is useful for solving another optimal power flow.For this reason,an evolutionary multitasking based DE is designed to solve multiple optimal power flow optimization problems simultaneously.In the designed method,a 0-1unified search space is used to initialize the whole population to establish potential connections between multiple optimal power flow optimization problems;then a knowledge transfer strategy based on population information sharing is proposed to efficiently perform knowledge transfer,thereby improving the solution accuracy and convergence of the target task.The experimental results show that the proposed method is not only able to solve multiple optimal power flow optimization problems simultaneously,but also to obtain more accurate solutions.