Research on time of use electricity price optimization by integrating clustering algorithm and multi objective dynamic adjustment model

LI Zhen; CHEN Julong; GUO Yuanping; LI Kui; XIAO Xia

Abstract：Aiming at the problems that the high proportion of new energy access leads to the aggravation of grid load fluctuation,the static time division of the existing TOU price mechanism and the simplification of price difference optimization,this paper proposes an optimization method of TOU price based on improved clustering algorithm and multi-objective dynamic adjustment model.By constructing the analysis framework of net load characteristics based on fuzzy membership,combined with the improved k-means clustering algorithm(optimizing the initial clustering center with the load extreme point),the dynamic fine division of peak and valley periods is realized,which solves the problem of strong dependence on experience in the traditional threshold method,and the matching degree of time division and load fluctuation trend is improved by 23.6％.Furthermore,the fusion strategy of Grey Wolf algorithm(GWO)and competitive learning particle swarm optimization(RPCL)is proposed,and a multi-objective dynamic adjustment model is established to minimize the peak valley difference,optimize the power purchase cost and maximize the consumption of new energy.The consistency optimization of cross month time division is achieved through the seasonal CLIQUE clustering algorithm.The example results show that after optimization,the peak valley difference rate is reduced to 28％,the peak valley electricity price ratio is 5.4∶1,and the load peak shaving ratio is increased to 4.93％,which effectively balances the contradiction between supply and demand of source load storage.The research results provide theoretical support and practical tools for the dynamic price adjustment of new energy LED power grid.