A Deep Learning Approach for Predicting Shockwave Propagation in Compressible Flows

Abstract

This study presents a deep learning–based approach for predicting shockwave propagation in compressible flows. A neural network model is developed to accurately estimate shockwave location, strength, and geometry based on initial flow conditions and governing parameters. The model is trained on an extensive dataset generated from high-fidelity numerical simulations and further validated using available experimental measurements to ensure robustness and generalization. Results demonstrate that the proposed deep learning framework achieves high predictive accuracy, capturing complex shockwave behavior more efficiently than conventional numerical methods. The approach significantly reduces computational cost while maintaining physical consistency, making it suitable for real-time or near–real-time prediction scenarios. This method shows strong potential for applications in aerospace engineering, explosion hazard assessment, and medical procedures involving controlled shockwave interactions.