A novel frequency-dependent-model for electric conductors

This paper deals with the inclusion of frequency effects (skin effect) in an enhanced transport delay model for electric cables of arbitrary length connected to a PWM-IGBT-Variable Frequency Drive (VFD), to better predict overvoltage peaks in the line during transient. The problem and the importance of predicting peak overvoltages to determine the appropriate insulation to be used in cables is addressed. From the step response of the cable, and using Genetic Algorithms (GA), an impulse function for cables is characterized. The novelty of this characterization is the emulation of an infinite long cable. Such a function is used to predict frequency-dependent voltage and current in longer cables. Simulation and experimental results are provided. Finally, an extension of this model is made for the n-conductor cable.