传递函数拟合算法及其在变压器绕组变形诊断中的应用

Transfer Function Estimation and Its Application to the Detection of the Transformer Winding Deformation

提出了利用网络的实测数据拟合其传递函数的实用算法．本算法是基于当某极点的实部较小时，该极点对网络传递函数的作用区域集中于其虚部附近，所以利用其虚部附近的测量数据可以较准确的拟合出该极点．对拟合结果使用高斯－赛德尔型迭代法进行修正可以进一步改善所得结果．对几组变压器的实测数据所做的计算证实了所提出的算法。

A utility algorithm to calculate the parameters of the transfer function of a network using the measured data in frequency domain is put forward. Based on the fact that the energy caused by a pole with a small real part is centered on its imaginary part, the pole could be fitted using the measured data close to its imaginary part. While all the poles are attained, the Gauss Seidel type iterations are used to revise the result. Finally, several practical examples are given and the application of the algorithm to detection of transformer winding deformations is studied.