Machine learning optimization strategy of shaped charge liner structure based on jet penetration efficiency

Abstract

Shaped charge liner (SCL) has been extensively applied in oil recovery and defense industries. Achieving superior penetration capability through optimizing SCL structures presents a substantial challenge due to intricate rate-dependent processes involving detonation-driven liner collapse, high-speed jet stretching, and penetration. This study introduces an innovative optimization strategy for SCL structures that employs jet penetration efficiency as the primary objective function. The strategy combines experimentally validated finite element method with machine learning (FEM-ML). We propose a novel jet penetration efficiency index derived from enhanced cutoff velocity and shape characteristics of the jet via machine learning. This index effectively evaluates the jet penetration performance. Furthermore, a multi-model fusion based on a machine learning optimization method, called XGBOOST-MFO, is put forward to optimize SCL structure over a large input space. The strategy's feasibility is demonstrated through the optimization of copper SCL implemented via the FEM-ML strategy. Finally, this strategy is extended to optimize the structure of the recently emerging CrMnFeCoNi high-entropy alloy conical liners and hemispherical copper liners. Therefore, the strategy can provide helpful guidance for the engineering design of SCL.

Keywords

• Jet penetration efficiency
• Shaped charge liner
• FEM-ML
• XGBOOST
• MFO
• High-entropy alloy