Intelligent Analysis Method Of Demand Side User State Under Energy Interconnection

The adaptive access of a large number of flexible loads such as distributed energy,distributed energy storage devices and bidirectional interactive users under the energy internet enhances the randomness of the power-load terminals of the distribution network,significantly changes the structure of the distribution network and the direction of power flow,and also greatly altered the load characteristics.The growth of peak-valley gap and the increasingly sharp contradiction between supply and demand in peak period pose a great threat to the safe and stable operation of the power system.At present,in the process of power system operation,there is a large amount of waste of data resources.At the same time,the gradual development of intelligent power service also provides a basis for the wide participation of users and the independent response to their demand.In order to solve the above problems,this paper develops an intelligent analysis method of demand-side user status under the energy internet.Firstly,the related technologies of intelligent analysis of power consumption state are studied,including the general steps of intelligent analysis of user state,pre-processing technology and feature extraction technology.Secondly,the clustering optimization strategy is proposed to improve traditional unsupervised learning-based clustering.Accuracy index and validity index are introduced to measure the reasonableness of clustering.Thirdly,aiming at the problem that unsupervised learning cannot be classified purposefully according to actual needs,a supervised user power status score based on extreme learning machine is proposed.The structure of ELM network is optimized by comparing the correct rates of training set and test set of activation function and number of nodes in different hidden layers.Finally,because the above classification algorithms are mostly based on small-scale data research,and there are still some shortcomings in the processing of massive user data,a hybrid model of DBN-ELM is proposed to effectively extract Abstract high-level features through in-depth learning to achieve intelligent analysis of user status.Intelligent analysis of user status has certain engineering application and reference value.It can dig into the big data to realize the demand of terminal energy consumption,strengthen the analysis of the customer energy consumption characteristics,subdivide different customer groups,take customer service as the basis,satisfy the deep-seated and diversified comprehensive energy service demand of customers,optimize the mode of user power consumption,and put forward the optimization of user power consumption scientifically and rationally.Suggestions play important roles in the optimal allocation of power resources.