Study Of Algorithm Based On Cloud Computing On Bad Data Identification Of Power System

Cloud computing is an computing model developed in commercial areas after evolution of the concept which mixed with grid computing, parallel computing and distributed computing. It is widely used in the analysis and processing of massive dataset. Cloud computing applications makes the idea that improving the ability of data processing while reducing terminal equipment performance to be realized. The development of smart grid and the growth of information in power system put forward higher requirements on future data processing and storage capabilities of power grid. After study of the mechanism operation of cloud computing and power system we found the similarity they own. So the cloud computing can be used in power system to improve the capability of information processing and data storage.This article examines the computing processes of existing cloud computing technology and the open source hadoop platform model, proposing a method of Power System identification of bad data cloud computing which based on hadoop cloud computing platform and mapreduce software framework and combined with bad data identification of gap statistic algorithm. In this method, we optimized the GSA by using Elbow criterion and Maximum-minimum distance method to resolve the impact on accuracy of GSA that caused by reference distribution and selecting clustering initial value at random. To verify the effect of optimization, we made the IRIS dataset of UCI standard database had a identificational simulation by old and new GSA. The thesis designed cloud solution of optimized GSA base on mapreduce framework, and realize it with Java language. In order to prove the advantage of optimized GSA with cloud computing proposed by the thesis, using IEEE-14nodes trend data to generate simulation of power system data sets. And made identificational simulation of optimized GSA under stand-alone model or cloud computing model respectively. Comparing two models under different amount of data in different dimensions and accuracy also time of identification. The result proved in dimension and large amounts of data in the data set, optimized GSA with cloud computing model faster than stand-alone mode. This provided a basis of practical application for cloud computing in power system. Also demonstrated the fesasibility of application of cloud computing technology in electric power systems.