Parameter Identification of Oil Paper Insulation Extended Debye Model Based on Spectrum Deconstruction Method

As a classical topological circuit model, the extended Debye model is of great significance for the condition diagnosis of oil paper insulation equipment. Aiming at the uncertain branches and the random identification results of the current extended Debye model. This paper based on the branch characteristics of the model, combining the real and imaginary parts relation of repolarization rate with spectrum differentiation method, proposing a spectrum decomposition method to realize the complete identification of the extended Debye model. Undering the external electric field, the dielectric relaxation process will cover the conductivity loss, and the coupling and superposition of the internal multiple polarization maps make uncertainly to the polarization identification. It obscures the number of branches’s determination of the equivalent model, and makes the identification of model parameters difficult. In order to solve those problems, firstly, the frequency domain dielectric spectrum is decoupled based on Kramers-Kronig transform. Since the resistance and infinite capacitance terms calculated result is 0, the real and imaginary parts of the complex capacitance calculated by K-K relationship only contain relaxation polarization components. Therefore it can realize the separation of the geometric branch and the polarization branch of the extended Debye model, and can also complete the quantitative extraction of the insulation resistance, the geometric capacitance and the relaxation polarization spectrum line. Then, using the spectral characteristics of resistance and capacitance, the unique parameters of insulation resistance and geometric capacitance can be determined by the least square method. Next, according to the spectrum differentiation method, the first-order differentiation of the real part of the polarization complex capacitance is carried out. Because the single relaxation polarization real part is a ladder, the differential spectrum line has a relaxation peak . The number of relaxation mechanisms can be confirmed by the number of peak points of differential spectral lines, and the unique polarization equivalent circuit parameters can be solved by using the uniqueness of peak points. Finally, the feasibility of this method is verified by existing model parameters and examples: the identification error of insulation resistance and geometric capacitance parameters is less than 1%, the polarization equivalent circuit error is less than 7%. In general, the calculated and measured spectral lines are in good agreement and the trend is consistent. The spectral deconstruction method can provide a reliable physical model for the study of the internal relaxation characteristics of oil paper insulation. Through simulation and case analysis, the following conclusions can be drawn: ①the conductance and relaxation processes can be accurately separated by K-K transformation, and the insulation resistance and geometric capacitance parameters can be determined. ②The number of polarization branches of the extended Debye model can be clearly determined by the peak number of the real part differential spectral line of the polarization complex capacitance; The parameters of each polarization branch can be identified by using the spectrum differentiation method to solve the spectrum step by step. ③The spectrum deconstruction method proposed in this paper can completely identify the parameters of oil paper insulation extended Debye model, and the results are unique within a certain accuracy range. © 2023 Chinese Machine Press. All rights reserved.