AC Equivalent Parameters Identification of Nonlinear Dielectric Properties of PE/SiC Composites Based on Genetic Algorithm

Insulation materials with field-enhanced non-linear conductivity and / or non-linear permittivity can achieve adaptive matching between material performance parameters and the value of the space electric field, and intelligently improve the uniformity of the spacial distribution of electric field in the insulation structure. The commonly used models to characterize the nonlinear conductivity and nonlinear polarizability of the dielectric include polynomial and hyperbolic sine form models and so on. There are several parameters in each model, the parameter identification method based on genetic algorithm can accurately obtain the parameter values in the model. Silicon carbide polyethylene composite (PE/SiC) materials with different SiC filler concentrations were prepared to verify the effectiveness of the identification method. A test system with a program-controlled signal generator and a linear high-voltage amplifier as the core was set up. Samples with different filler concentrations were tested at different excitation amplitudes. Time domain data of excitation and response under various conditions were obtained. The parameters of different samples were extracted according to the method proposed above, and the inferred response according to the model and parameters had a good agreement with the measured response, which proved the effectiveness of the method. The parameters of different samples were analyzed and found that the parameters in the model increased with the degree of nonlinearity of the dielectric.