Exponential fitting algorithms for digitally recorded HV impulse parameter evaluation

Abstract

Standard analysis of high voltage impulses comprises determination of parameters describing the recorded impulses. These include peak value, front time, and time to half-value. This paper describes several fitting algorithms intended to facilitate extraction of these parameters from recorded data. These algorithms were used to fit bi-exponential models to smooth lightning and switching impulses, to lightning impulses having oscillatory fronts or peaks where the evaluation of the waveform parameters is based on the construction of a smooth curve, and to tail-chopped lightning impulses. The overall effectiveness of these algorithms and their influence on parameter estimates are presented and discussed. >