Predicting ballistic resistance based on the mechanical properties of armored ceramics

The relationship between the mechanical and ballistic properties of armored ceramics was studied using a numerical method that combines finite element simulation and machine learning. A dataset containing the physical properties, mechanical properties, and ballistic performance of ceramic was built through numerical simulation with the JH-2 model. The machine learning method was employed to establish a relationship between residual penetration depth and ceramic’ properties, including its density, elastic modulus, Poisson's ratio, Hugoniot elastic limit (HEL), dynamic compressive strength, quasi-static compressive strength, and bending strength. Through finite element simulation and machine learning, univariate analysis of the physico-mechanical properties of ceramics was achieved by predictive modeling. It was found that density, dynamic compressive strength, quasi-static compressive strength, and HEL had the most significant impact on the residual penetration depth, making them the key controlling parameters for the ballistic performance of ceramics. By combining the prediction results of the machine learning model with the influence patterns of the parameters, a predictive formula for the ceramic protection coefficient based on the key parameters was established.