Study On Parameter Optimization Method Considering The Influence Of Multiple Random Variables In Electrical Measurement

With the increase of power system scale,transmission line voltage level and non-linear load,the impact of power quality and power system accidents on national production and life has become prominent.Improving the accuracy of electrical measurements and promoting the development of power grid detection has become an important issue to be solved urgently in the power system.For some special measurement systems such as capacitive voltage transformers,there are many components with uncertain values or the parameter values of these components are easily affected by the external environment and cannot be accurately estimated.We can call these parameters more Random Variables.The systematic errors caused by these multi-random variables often have a non-negligible impact on the measurement results.How to accurately value the multi-random variables is the key factor for accurate correction of the measurement results.Therefore,this paper proposes a parameter optimization method that takes into account the effects of multiple random variables in electrical measurement.Before studying the multi-random parameter optimization method,this paper first studies the multi-random variable parameter processing method.This method starts from the probability density of each component of multiple random variables,and determines the independence and correlation between multiple random variables.According to the independence of multiple random variables,the multiplicative relationship can be used to determine the joint probability density and probability distribution of multiple random variables.However,the joint probability distribution cannot be determined based on the correlation between multiple random variables.Therefore,this paper estimates the joint probability density function and joint probability distribution of multiple random variables based on asymmetric kernel density estimation,so that the following multiple random variables Prepare for parameter optimization methods.The multi-random variable parameter optimization method proposed in this paper is studied from two angles.The first is a multi-random variable parameter optimization method based on minimum error.This method uses different measurement results as reference values in turn,and calculates the maximum error value of the remaining measurement results relative to the reference value.The measurement result is taken as the optimal result,and the corresponding multiple random variable combination value is taken as the optimal parameter combination.The second is the multi-random variable parameter optimization method based on maximum probability.This method takes the error tolerance of the measurement result as the constraint condition,takes the probability that the relative error falls within the error tolerance as the optimization goal,and calculates the different measurement results as the reference value.At that time,the error of other measurement results relative to the reference value is calculated as the total probability that the relative error at a certain reference value meets the error tolerance requirement,and the combination of multiple random variables corresponding to the maximum probability is used as the optimal parameter combination in the maximum probability matrix.Finally,this article takes a certain type of capacitive voltage transformer as an example.Through theoretical calculation and simulation analysis,it is found that the multi-random variables that affect the measurement of CVT harmonics are the compensation reactor stray capacitance C_cand the intermediate transformer stray capacitance C_p1.The proposed method optimizes the accurate values of stray capacitances C_cand C_p1.The CVT field harmonic correction experiment verified the accuracy of the method in this article.