Research On Peak Load Shedding And Valley Filling Evaluation Method Of Distribution Network Based On Electricity Consumption Behavior Of Power Users

Smart grid is a new type of power grid which integrates various information communication technologies based on traditional power grid.Smart grid emphasizes deep interconnection between energy system and information system.The high integration of communication and information acquisition technology in physical power grid lays the foundation for the efficient interaction between information and energy.Intelligent meter and other related technologies have obtained a lot of power information data.It is the premise of energy efficient utilization to effectively mine these power data.These power data show the characteristics of large scale,time sequence and low value density.How to analyze and use these data efficiently becomes an urgent problem.The load peak valley difference of the power grid becomes bigger and bigger with the improvement of the national living standard,which affects the stable operation of the power grid.The traditional load cutting and filling method has little effect.The development of energy storage technology provides a new way to solve the load peak valley difference,that is,to let the energy storage system participate in reducing the peak valley difference of load of power users and smoothing the load.Firstly,in view of the characteristics of bigger and bigger load data scale and time sequence,an algorithm model combining wavelet reduction and K-shape clustering is proposed in this thesis.The low frequency information of load data is extracted by multi-resolution analysis in wavelet analysis,and then the low frequency information is clustered by K-shape algorithm.The validity of the algorithm proposed in this thesis is proved by experiments.Secondly,some external factors that affect the power consumption of power users and the short-term load forecasting accuracy are introduced.Based on this,a short-term power load forecasting algorithm is proposed,which combines K-shape clustering and random forest.The K-shape clustering algorithm is used to cluster the load curve data of power users,which can extract and mine the characteristic load pattern,and the random forest algorithm is used to process the external factors data.The algorithm is compared with the random forest regression,BP neural network and support vector regression algorithm,which proves the effectiveness of the proposed algorithm.Finally,through the analysis of the optimization objective function and some constraints of the load curve,a new algorithm model of peak shaving and valley filling with variable power participating in the energy storage system is proposed.The principle logic of the algorithm is analyzed,and the reliability of the algorithm is proved by simulation experiments.