Health Assessment Of Key Equipment Of Smart Grid In The Context Of Big Data

With the rapid development of digital information,the amount of information enjoys an explosive growth.And the deep integration of information and communications technology and electricity generation has contributed substantially to the electric power industry,bring about some qualitative changes.These changes are best manifested by the fact that electric power data has become the core assets of the industry.At present,China's electirc power system is the largest in the world,which makes it all the more important that electric equipment should operate stably and be managed effectively to keep the system safe.Thus,researchers and electric power enterprises are exerting themselves to find out how to rapidly extract information pertaining to a system default from a sea of data.In this process,the enormous amount of real-time data streams generated by sensors in the smart grid poses a serious challenge.Data clustering is an important processing technology in data mining.Researchers have put forward many representative clustering algorithms.They,however,cannot be directly applied because of the emergence of new streaming data.So a new data flow analysis and processing method is needed.Cloud model is a model that combines stochasticity with fuzziness and can achieve qualitative and quantitative uncertainty transformation through specific algorithms.This kind of model claims researchers' attention and is successfully employed in many fields.In light of the above problems,this thesis,based on data stream clustering and cloud model,discusses a new method for electric equipment health assessment.The method includes offline processing and online real-time processing.The former classifies the operating conditions of the equipment based on historical operation data under normal conditions,and calculates the standard level of the combined Gaussian cloud of the equipment in each type of working condition.The latter,using the stream clustering algorithm,identifies the working conditions of those equipment,obtains summary information from streaming data,and computes the combined Gaussian cloud.Then,it calculats a health index for the equipement based on the real-time deviation of Gaussian cloud from the standard level,and classifies the health conditions according to the index.Experiments are carried out under the guidance of this method,with the result that the final result is in accordance with real conditions of the equipment and that this method can be well applied to send a health condition alert.