Analysis And Research On The Influence Of Single Factor And Multiple Factors On The Verification Error Of Electric Energy Meter

In China Southern Power Grid Corporation,the energy metering equipment of Yunnan Power Grid often controls the quality of these metering equipment through annual inspection.And every year,the number of electric energy meters that need to be verified is up to several million.If there is a problem in the verification process,it will bring about a large social impact and huge economic losses.The quality of the energy meter is mainly measured by the error data of the verification.Therefore,it is very important to choose a high-precision electric energy meter supplier in actual production.Therefore,this paper analyzes the algorithm of laboratory electric energy meter error verification data of a measurement center in Yunnan Province,and performs kmeans clustering,error curve fitting and GM(1,1)single factor on the verification error data of each electric energy meter supplier.And PCA-MLP multi-factor prediction and other processing.The purpose is to purchase power meters from different suppliers in the future,and provide strategic support and reference.As well as the electric energy meter error detection,the predicted curve can be used as a measure of the abnormality verification data.In response to the problems described above,the following research work was done:(1)A large number of verification error data generated by the electronic energy meter in the laboratory verification process.Faced with a large amount of systematic error data,the characteristics of these error data are analyzed in detail,and the characteristic values of the error sequence are selected as sample data to establish The electric energy meter error verification analysis prediction model.(2)Analyze the eigenvalues of different factors.For the calibration error of the same batch of electric energy meter of the same supplier,after the data pre-processing,firstly use the kmeans clustering algorithm to classify the error data,and then select the error accuracy most.The good supplier data is added to the prediction model,and the gray prediction GM(1,1)model is used to analyze the single factor verification error.Then,the influence of the internal and external factors of the electric energy meter in the actual verification process on the energy meter error verification data is used.Principal component analysis method screens out the main factors affecting the calibration error of electric energy meter,and uses PCA-MLP combined algorithm to construct multi-factor electric energy meter verification error prediction model.(3)Using the energy meter calibration error data of a metrology center in Yunnan Province,the energy meter error verification analysis and prediction model is applied to the analysis and prediction calculation of error data.After the analysis of the example experiment,the model established in this paper is significantly improved in prediction accuracy.Based on the electric energy meter error verification analysis and prediction model is applicable to the actual electric energy meter error prediction.