Electric Power Data Processing Based On The Cloud Computing Platform

A lot of data is produced with the development of electric power systems, which brings a great challenge to the data processing system in electric power industry. In recent years, the cloud computing and big data technology have achieved rapid development. Many fields, such as internet,telecommunication, education, government administration, and so on have adopted cloud computing to processing business. This thesis is about the electric power data processing based on the cloud computing platform. The main contents include two aspects. Firstly, an electric power business processing scheme is designed and implemented which uses resources of the cloud computing platform to process massive data. The scheme improves the efficiency of electric power business processing. Secondly, a classification algorithm of electric power consumption data is put forward.The algorithm distinguishes normal electric power consumption data and abnormal electric power consumption records, providing supports for reasonable decision of electric power enterprises.In this thesis, the electric power businesses such as electric power data acquisition, electric power data computing, electric power data storage and query are firstly analyzed in detail. These businesses are very important to the electric power industry, and the businesses data amount has a trend of rapid growth. So the traditional electric power business processing schemes face great challenges. Then, after summarizing the characteristics of these businesses, an electric power business processing scheme based on the cloud computing platform is proposed. The scheme includes five modules: data collection and distribution, online data processing, offline data processing, data storage and query, Web interaction. The scheme handles the electric power data in its full life cycle by utilizing the storage and computing resources of the cloud computing platform and the big data processing tools integrated in the platform. The scheme can effectively deal with the rapid growth of electric power data, and complete the electric power business.An improved classification algorithm for electric power consumption data is proposed. This algorithm introduces the concept of "information quantity" to classification algorithm based on Logistic Regression. The features of electric power consumption data are selected according to the information quantity, and the features with bigger information quantity are selected as the prediction features of Logistic Regression. This algorithm reduces the complexity and insures the accuracy rate. At the end, three experiments are designed to test the performance of the cloud computing platform, the electric power business processing scheme and the classification algorithm. The experimental results show that the cloud computing platform is stable and reliable, the electric power business processing scheme can complete the electric power business under the condition of rapid growth of electric power data and the classification algorithm of electric power consumption data is efficient.