| In recent years, with the full launch of smart grid construction, and during the operation, maintenance and management of smart grid massive heterogeneous and multi-state data, namely the bigdata, are generated. Load characteristics analysis and load forecasting power grid companies load scheduling and distribution network planning decisions. In order to achieve accurate and efficient analysis of load characteristics and load forecasting, Big Data technology is a very important tool. In big data technology, the data mining technology is an important means of analysis of data mining, showed certain advantages in data analysis, the electric field will be widely used in the future.First it introduced in data mining frequent pattern tree (Frequent Pattern tree) association rules algorithm in load characteristics analysis. By analyzing the load characteristics and Climatic factors, the correlation between the level of economic development, income levels and the time factor, and then based on the application of FP tree association rule algorithm, statistical frequency of all kinds of factors, establish the FP tree. In defining minimum support and confidence of the premise, it can effectively analyze the potential impact of the association rules load factors.Next is the equivalent relation clustering analysis in the application of electricity behavior analysis, this chapter further research power use behavior, through the application of equivalent relation clustering analysis area load curve cluster analysis according to the seasons, and analyze the differences between user habits lead to differences in living in the same season, the behavior of electricity, as well as climatic factors have led to electricity users behavior change, the purpose of the refinement of the analysis at different seasons.Finally combining with the previous two chapters to a city of low pressure area load characteristic analysis and behavior analysis conclusion, electricity by combining with clustering analysis and BP neural network algorithm to forecast the area load, the method will not only load characteristic analysis results fully applied to the prediction algorithm, also greatly improve the efficiency of the accuracy of load forecasting and prediction. |
