A mathematical expression for air ESD current waveform using BP neural network

A curve-fitting method based on backpropagation (BP) neural network (NN) is proposed for fitting electrostatic discharge (ESD) current waveforms and giving the corresponding expressions in order to analyze the characteristics of ESD and calculate the ESD electromagnetic pulse radiation field. According to IEC61000-4-2, this method is used to fit the ideal contact current waveform of human-metal ESD to obtain the mathematical expression. The waveform parameters of the fitting curve meet the requirements of the IEC61000-4-2. By fitting the measured air ESD current waveform at 4 kV discharge, the mathematical expression is obtained. This paper proposes the mathematical expression for arbitrary curves and analyzes the factors affecting the effects of curve-fitting.

► A new curve-fitting method based on backpropagation (BP) neural network (NN) is proposed for fitting ESD current waveforms.
► This method is used to fit the ideal contact current waveform of human-metal ESD to obtain the mathematical expression. The waveform parameters of the fitting curve meet the requirements of the IEC61000-4-2.
► By fitting the measured air ESD current waveform at 4 kV discharge, the mathematical expression of the measured air ESD current waveform is obtained.