Research On Dispatching Optimization Of Emergency Generation Vehicles Of Distribution Network Based On Machine-learning

Electric power is closely related to national life and economy,and is one of the indispensable factors in the development of modern society.The reliability of power supply directly affects the people’s quality of life and the level of social and economic development.With the continuous progress of society,the scale of power grid and the load of power supply are growing rapidly.How to ensure the safety and reliability of power system operation in this dual pressure environment is becoming more and more important.In order to cope with the dual challenges of the power grid environment,uninterrupted operation is a major breakthrough to improve the reliability of power supply,while in recent years,most power supply units have gradually shifted their focus of work.As one of the core contents of four types of uninterrupted power supply business(Bypass Operation,Hot-line Operation,Emergency Generation Vehicles,Low-voltage contact)in power supply units,emergency generation vehicles has the characteristics of being relatively safer and more convenient,and the higher frequency of use,which also is one of the key directions of related research to improve the reliability of power supply.The configuration scheme of emergency generation vehicles is an extremely important part in its dispatching optimization.Optimizing the configuration plan of the unit that the emergency generator vehicle belongs to is beneficial to improving the efficiency of its use at the source,reducing unnecessary dispatch times and mileage due to improper configuration,saving time and effort,while greatly reducing the cost of use as well.As the current research rarely focus on emergency power generation vehicles,the formulation of power generation vehicle configuration schemes is usually unreasonable in practice,and its decision-making relies heavily on subjectivity thus lacks a certain degree of scientificity,which will inevitably cause the significant loss of economic,human resources and time.In response to the above problems,in order to obtain the optimal configuration plan of generation vehicles in a period of time,considering the generation vehicles of actual engineering application of the distribution network,the related constraints such as vehicles type,capacity,status and demands are assumed,and the objective function of minimum total distance of vehicles dispatching under multiple factors in a certain dispatching period was designed,while a global optimal allocation model of emergency generator vehicles in distribution network considering cross-regional and cross-time bureaus is proposed.In order to handle characteristics model of the non-deterministic polynomial problem(Nondeterministic Polynominal Hard Problem,also known as NP-Hard Problem,referred to as NP difficult problem)above,considering machine learning,proposed the Improved Genetic Algorithm(IGA),Improved Binary Cirsscross Optimization Algorithm(IBCSO),and Hybrid Binary Crisscross Optimization Algorithm(HBCSO)innovatively which is suitable for distribution network generation vehicles operation mode with the characteristics of discrete variable.Based on the actual operating data form a power supply department in Guangdong Province to carry out simulation research and comparative analysis,combined with the cyclic optimal solution method explains the feasibility,effectiveness and relative superiority of IGA,IBCSO and HBCSO algorithms for solving the discrete and uncertain problem of generating vehicle configuration optimization.At the same time,the all three algorithms have the same advantages of strong convergence ability and convergence speed.Among them,the calculation accuracy of the HBCSO algorithm gets significantly improved with multiple iterations,and the higher overall calculation efficiency as well.Finally,a comparative analysis with the original distribution network power generation vehicle operating data shows that the dispatching status of the distribution network emergency power generation vehicles has been significantly improved under the optimization of each algorithm,which greatly shortens the distance of cross-regional dispatch,reduces the number of dispatches,and more saves the cost of the operation.