Study On The Probability Model Of Uncertain Demand Response Considering Electricity Consumption Behavior Clustering

The introduction of demand response enables the demand side to “two-way interaction” with the power grid,thereby effectively alleviating the power gap during peak hours,eliminating intermittent renewable energy,and ensuring the safety,stability and economic operation of the power system.In this regard,it is of great significance to study the demand response behavior characteristics of power users,which is a prerequisite for conducting demand response research and practice.In view of the fact that peak-to-valley time-of-use electricity price is currently the most widely used demand response measure,this paper mainly studies this,that is,the demand response model of this paper refers to the user time-of-use electricity price response model.In the past,we find that the user groups with different consumption behaviors have different demand response characteristics by analyzing the user’s electricity behavior.However,the user behavior analysis is not considered in the research on the demand response model.Therefore,this paper combines the two,that is,firstly studies the user’s electricity behavior analysis,and then studies the user uncertainty demand response model.The detailed contents of the research are as follows:(1)The user’s electricity behavior analysis mainly identifies different consumption behaviors by clustering the user’s typical load curve,so the core is the clustering algorithm.An improved density peaks clustering algorithm is proposed.The improved algorithm uses the idea of K-nearest neighbor(KNN)to improve the sample local density and distance calculation criteria of the original algorithm.Firstly,principle components analysis method is used to reduce dimensions of load curves after normalization.Then,the kd tree algorithm is used to carry out the fast k-nearest neighbor search to generate KNN matrix.Finally,the KNN matrix is used to replace the original distance matrix as the input data.Based on the KNN improved local density and distance calculation criterion,the density peaks clustering algorithm is used to cluster the load profiles.Experiments and case analysis show that compared with the traditional clustering algorithm,the improved algorithm has stable and better clustering results,and can effectively reduce the memory consumption and execution time of the original algorithm,which can achieve cluster analysis of high-dimensional massive electricity consumption behavior and provides the basis for modeling uncertain response.(2)In view of some problems existing in the previous literature on the demand response model,this paper improves on this,and establishes a demand response model that takes into account the uncertainty under the peak and valley time-of-use electricity price environment from the perspective of probability.Firstly,cluster analysis is carried out on all users using the improved fast density peak algorithm proposed above to obtain user clusters with different electricity consumption behaviors.Then,diffusion-based kernel density estimation(DKDE)is introduced to estimate the probability distribution.Finally,for each type of user cluster,the uncertain demand response probability model is established based on the DKDE algorithm.The specific processes are: 1)User demand response evaluation indicators are established;2)The Logistic function with nonlinear and continuous characteristics is introduced to replace the previous piecewise linear function of load transfer ratio;3)The load transfer ratio experiment of user cluster with uncertainty is designed.Based on the experimental results and analysis,a more accurate and less error uncertain model of user cluster load transfer rate is established,and the validity of the model is verified;4)Based on the probability distribution of load transfer ratio and baseline load,the monte carlo simulation method and DKDE algorithm are used to obtain the probability distribution of evaluation indexes that can characterize the effect of uncertain demand response and the confidence interval of evaluation indexes at a given confidence level.The case analysis shows that the model in this paper can well simulate the probability distribution and the fluctuation range of users’ response under a given time-of-use electricity price,and provide technical support for power companies to master the characteristics of users’ response behavior,formulate a scientific and reasonable time-of-use electricity price,play a role of peak shaving and valley filling and absorb intermittent new energy.