APPLICATION OF NEURAL NETWORK FOR SIMULATION OF ELECTROSTATIC DISCHARGES IN COMPUTER MODELS

Abstract

Modeling of electrostatic discharge (ESD) is critical to ensure the reliability and safety of electronic systems, aerospace components and industrial equipment. However, accurate modeling of ESD events remains a major challenge due to their highly nonlinear, stochastic and multi-scale nature. Traditional numerical methods such as finite element analysis (FEA) and finite difference time domain (FDTD) methods often suffer from computational inefficiency and limited prediction accuracy when modeling complex discharge phenomena [1].

Recent advances in machine learning, particularly deep neural networks (DNNs), offer a promising alternative to improve ESD modeling. Using data-driven approaches, neural networks can learn underlying physical patterns from experimental or high-fidelity simulation data, enabling faster and more accurate prediction of discharge behavior. Methods such as physically informed neural networks (PINNs) and hybrid modeling frameworks can bridge the gap between empirical observations and theoretical models, improving the accuracy of ESD modeling and reducing computational costs [2].

This paper reviews the potential of neural networks in advancing ESD modeling, discussing key methodologies, challenges, and future directions in this emerging field.