Research on Data Mining of Shaped Charge Jet Device Design Based on Shapley Additive Explanations Theory

Abstract

ABSTRACT The formation process of the shaped charge jet has typical fluid characteristics. In order to clarify how parameters of the shaped charge affect jet velocity, knowledge mining is performed on the nine design variables of a shaped charge using the Shapley Additive exPlanations (SHAP) data mining method. The ABAQUS software is employed to perform 108 sets of simulations with varying structural parameters. An intelligent machine learning algorithm is utilized to establish the mapping relationship between the design parameters and the jet velocity. Based on the high‐precision velocity regression, SHAP's partial and global interpretability are fully leveraged to conduct attribution analysis of samples and examine the role each design parameter plays in jet formation. The study identifies the key design parameters influencing maximum velocity and penetration velocity as the liner's thickness, angle, length, and the charge amount, with respective critical values of 0.95 mm, 30.9°, 26.76 mm, and 125 mm 3 . The design criteria for shaped charge jet devices derived from data mining are as follows: reduce the liner's thickness and angle while maintaining a larger charge quantity; this requires increasing the device radius, the liner's length and height, and enlarging both the inclination angle and length of the case. The research results can provide some inspiration for the applied research on the shaped charge jet.