Research And Application Of Electric Energy Forecasting And Management System Based On Hadoop Platform

Cloud Computing is a style of computing in which dynamically scalableand common virtualized resources are provided as a service over theInternet.Cloud Computing platform can be used to development highcapability programmes.Not only scholars but also industry takes it as thefuture IT trends. With the same processors and storage installations, CloudComputing can integrate the existing resource in enterprise with intranet toimprove the computation and data security of entire system, and can reducethe investment of system expansion as well.On the basis of studying short-term load forecasting algorithm andHadoop Cloud Computing platform, the thesis has done these works:（1）Aiming to estimate for missing data of the Electric Energy Meteringand Management System,this paper, the theory of optimum weightedcombination which combined GM(1,1)、SVM(Support Vector Machines) andANN are introduced to short-term load forecasting of the system, then wecan get the predicted value.The forecasting results based on the data fromthe distribution electricity center in Eastern Slovakia shows that thealgorithm possesses evident potentiality in the field of short-term loadforecasting.（2）Analyzing the frameworks of Cloud Computing platform and Hadoop,puts forward to build the electric energy forecasting and management system based on Hadoop platform, and provides a roadmap of successfulimplementation, forecasting approach to short-term load of electric energyand analysis of Hadoop platform.