Research On Sampling Inspection Methods Of Residential Watt-hour Meter In Use And Its Empirical Analysis

Residential watt-hour meters are important tools that are used to measure the electricity of resident customers. Residential watt-hour meters in use (watt-hour meters in use of brief name) are one kind of residential watt-hour meters that have been installed and in the usage. The accuracy of watt-hour meters in use is related to the economic benefits of both customers and electric power manufacturers, so we have to inspect them periodically. Along with the power market reform progressing and the "one family one meter" rule being executed, the quantity of watt-hour meters in use increases quickly. Therefore the shortcomings of current total amount inspection that low efficiency and high cost become more and more obvious. That sampling inspection replaces total amount inspection can consumedly increase the efficiency of inspection and decrease the cost at the meanwhile. Therefore, to design sampling procedures of quality inspection of watt-hour meters in use has great importance.The procedures of watt-hour meters in use sampling inspection could be divided into two parts: survey sampling and qualification judging. The paper concentrates the part of survey sampling making use of the theory of statistic, econometrics and cost analyses. The survey sampling of watt-hour meters in use is to draw some samples from meter collectivity and to estimate the whole situation according to the samples' quality. Its process primarily includes two steps: sample size calculating and stratified sampling designing. In the step of sample size calculating, the paper firstly analyzes two different methods of meter survey sampling in two national standards of China, and compares them in three aspects of statistic theory, discriminating ability and maneuverability. Then we draw a conclusion that one of the two methods is fit to survey sampling of watt-hour meters in use and establish the sample size calculating model of watt-hour meters in use survey sampling after considering the precision and the cost of survey synthetically.According to watt-hour meters in use have the characteristic of large amount and asymmetrical distribution, the paper brings up the step of stratified sampling. Its aim is to divide one collectivity into many smaller subclasses, decreasing the internal difference with stratification, and then increase the total estimating precision. This step mainly demands to resolve two problems: 1. Choice of stratified variable, viz. choosing some index according to which samples should be stratified. For this, the paper establishes multiple linear regression model and single factor variance analysis model, with which finds the stratified variables that influent the quality of meters most obviously from many influential factors. Therefore, the benefit of stratified sampling method becomes enhanced. 2. Allocation of samples in the strata. There are three main methods to allocate samples to every stratum: proportionate allocation, optimum allocation and Neyman allocation. The paper compares the three kinds of methods and, according to the characteristics of watt-hour meter in use stratified sampling, considers Neyman allocation as the appropriate method.Finally, according to the data get from the quality sampling inspection of watt-hour meter in use in C city, the paper makes the empirical analysis for the sample size calculating model of watt-hour meters in use survey sampling, meter quality multiple linear regression model, meter quality single factor variance analysis model and Neyman allocation of watt-hour meters in use stratified sampling. Then the paper uses the Neyman allocation relative benefit formula to evaluate the benefits of the design of stratified sampling.The paper researches into the survey sampling part of the sampling inspection of watt-hour meter in use quality completely and thoroughly, and the results would provide the theory basis to the practical inspection.