Study On Multiobjective Security Constrained Unit Commitment Based On An Efficient Evolutionary Framework And Ⅰ-ANSGA-Ⅲ Algorithm

Solving the security constrained unit commitment problem(SCUCP)is an important aspect of optimal power system operation and is one of the most important daily tasks that an independent system operator or regional transmission organisation must perform in the daily electricity market,providing significant economic benefits.Security constraints have long been recognised as difficult constraints in SCUCP and SCUCP can be greatly simplified if redundant security constraints can be identified and eliminated.With global warming and climate change,increased public awareness of environmental protection and the rapid development of wind power generation,SCUCP will evolve into a multiobjective optimization problem when considering the combined benefits of economic,environmental protection and energy conservation,with the task of finding a well converged and uniformly distributed non-dominated front.Due to the high dimensionality of variables,large number of constraints and complex problem structure,multiobjective SCUCP(MOSCUCP)is a large-scale constrained multiobjective optimization problem for which no efficient algorithm exists to date.In this paper,based on the summary of existing research results,we propose a heuristic algorithm(HA)for fast identification of redundant security constraints in SCUCP,construct a mathematical model for minimizing the combination of operating costs,emissions,and wind power curtailment with three objectives of security constraints,and propose an evolutionary framework for efficiently solving MOSCUCP and an improved adaptive approach to nondominated sorting genetic algorithm III(I-ANSGA-III)incorporated into the framework.The algorithm and framework were implemented using Matlab programming and tested on several systems.The computational results and analysis show that HA can identify and remove most of the redundant security constraints and has a significant computational efficiency advantage over the comparative algorithms in the literature.The proposed evolutionary framework can effectively improve the convergence performance of multiobjective evolutionary algorithms and improve the distribution of non-dominated fronts by comparing the computational results with the most representative and popular algorithms used in the literature.Furthermore,as the number of objectives increases,MOSCUCP may have complex non-dominated fronts,in which case and without prior information about the shape,orientation,continuity,convexity,etc.of the non-dominated fronts,I-ANSGA-III is still able to find trade-off solutions that are uniformly distributed on a global and local scale.