Abstract:
Background High-altitude electromagnetic pulse (HEMP), generated by nuclear explosions at high altitudes, is characterized by an extremely high amplitude, broad pulse width, and extensive geographic coverage. It poses a severe threat to modern electronic systems, communication infrastructures, and power grids. Accurate and efficient prediction of the HEMP environment is essential for evaluating its potential impact and formulating protective measures.
Purpose Traditional numerical methods for HEMP prediction are often computationally intensive and time-consuming. This paper aims to develop a fast and accurate prediction model based on an artificial neural network (ANN) to overcome these limitations and enhance computational efficiency while maintaining prediction accuracy.
Methods The proposed model integrates the Karzas–Latter high-frequency approximation model with the World Magnetic Model to establish a physical basis for HEMP simulation. A deep neural network architecture is constructed, comprising one input layer, eight hidden layers, and one output layer. The Sigmoid function is adopted as the activation function, and the mean squared error is used as the loss function during training.
Results Experimental results demonstrate that the ANN-based model can accurately predict the peak electric field intensity of HEMP across a wide area within a very short computation time. Compared with conventional numerical methods, the model significantly reduces the required calculation time while achieving high predictive accuracy, making it suitable for rapid environment estimation and scenario analysis.
Conclusions The developed ANN model provides an efficient and reliable tool for fast prediction of the HEMP environment. It offers substantial practical value for HEMP risk assessment, emergency response planning, and design of protection strategies for critical infrastructure. The research outcomes can serve as a valuable reference for both academic and applied disciplines concerned with electromagnetic pulse effects.
下载: