Wide frequency domain modeling method of DCTV based on rational function approximation

摘要

DC voltage transformer (DCTV) is an important device for measuring the voltage of ultra-high voltage direct current (UHVDC) transmission system. Its characteristics of wide frequency transmission directly affect the accuracy of voltage measurement. In order to solve the problem of large harmonic voltage measurement errors caused by the inaccurate acquisition of the broadband transmission characteristics of DCTV, a new broadband modeling method for DCTV is proposed in this paper. This article proposes a new DCTV wide frequency domain modeling method without relying on DCTV parameters. The DCTV transfer function values at different frequencies are obtained by measurement, and the error function is constructed based on the rational approximation fitting principle. Then, the particle swarm optimization algorithm is used to iteration the error function to obtain the optimal solution and to fit the expression of transfer function. The accuracy of the algorithm is studied in different frequency sampling range (low frequency, intermediate frequency, and high frequency) and noise levels. Simulations based on actual DCTV parameters verify that the proposed method has a relative error of only 2.1908e-16% within the 200th harmonic. It has strong anti-noise ability and can control the error within 5% when the SNR of sampling value is greater than 1 dB.