| Nowadays,with the rapid development of electricity data collection system,the electric data is increasing greatly.To obtain the relevant information and patterns of user electricity consumption,the data mining technology is applied to analyze electric data.This information can be used to support the individual and differential services,and meet the demand for building smart power.So an urgent problem needs to be solved is how to effectively mine the electric data and use that information to build the smart grid.In order to solve the above problems,this paper uses the clustering analysis method to mine and identifies the electricity consumption pattern from electricity consumption data and electricity consumption behaviour.And how to use that information to optimize the strategy of power demand response is also studied.Firstly,a selection strategy used to optimize the feature set of electricity consumption is proposed in this paper.It is based on information entropy and feature correlation coefficient.The feature set is realized that is closely associated with pattern and independent of each other in during feature extraction stage.And it can reduce the complexity of clustering computation and improve the effect of cluster analysis.Then a k-mean clustering considering the initial clustering centre is proposed to finish this analysis.The result of the simulation indicates the proposed method can effectively distinguish the users of different consumption pattern.Finally,the strategy of demand response is optimized by using the results of clustering analysis.In order to optimize the effect of demand response and reduce the complexity of calculation,a classification scheduling considering electricity consumption pattern is proposed.Meanwhile,the optimization process is disassembled to the user according to the idea of distributed computing.It can protect the security of user information and make full use of user resources to improve the efficiency of calculation. |
PDF Full Text Request