Search

A coupled framework of ballistic simulations and an optimized machine learning (ML) model was developed to accurately predict the kinetic energy reduction of a projectile impacting a target. ML models can require a significant number of data points for proper training, testing, and validation. High-performance computing (HPC) resources can be used to simulate the ballistic impacts of various projectiles against several different target materials using appropriate physics-based hydrocodes. Computational modeling can explore areas where experiments would naturally be cost-prohibitive. These hydrocodes can evaluate large parametric spaces varying the projectile and target variables that are required to train an ML model. In this study a large, generated set of data points was used to develop an optimized artificial neural network (ANN) algorithm to create a fast-running model without prior knowledge of the mathematical relationships between all the input and output variables. The optimized ANN model was developed using Optuna in an HPC environment to tune the hyperparameters needed for the ANN model. This fast-running ML model could then be leveraged to investigate designing optimized targets that could protect against different types of projectiles. The results of this work showed that the optimized ANN model predicted the kinetic energy reduction with a mean absolute percentage error of 2.7% across the validation data. Overall, the optimized ANN model showed excellent agreement across the range of data considered by the computational models. [ABSTRACT FROM AUTHOR]

Machine learning (ML) is well suited for the prediction of high-complexity, high-dimensional problems such as those encountered in terminal ballistics. We evaluate the performance of four popular ML-based regression models, extreme gradient boosting (XGBoost), artificial neural network (ANN), support vector regression (SVR), and Gaussian process regression (GP), on two common terminal ballistics' problems: (a) predicting the V50 ballistic limit of monolithic metallic armour impacted by small and medium calibre projectiles and fragments, and (b) predicting the depth to which a projectile will penetrate a target of semi-infinite thickness. To achieve this we utilise two datasets, each consisting of approximately 1000 samples, collated from public release sources. We demonstrate that all four model types provide similarly excellent agreement when interpolating within the training data and diverge when extrapolating outside this range. Although extrapolation is not advisable for ML-based regression models, for applications such as lethality/survivability analysis, such capability is required. To circumvent this, we implement expert knowledge and physics-based models via enforced monotonicity, as a Gaussian prior mean, and through a modified loss function. The physics-informed models demonstrate improved performance over both classical physics-based models and the basic ML regression models, providing an ability to accurately fit experimental data when it is available and then revert to the physics-based model when not. The resulting models demonstrate high levels of predictive accuracy over a very wide range of projectile types, target materials and thicknesses, and impact conditions significantly more diverse than that achievable from any existing analytical approach. Compared with numerical analysis tools such as finite element solvers the ML models run orders of magnitude faster. We provide some general guidelines throughout for the development, application, and reporting of ML models in terminal ballistics problems. [ABSTRACT FROM AUTHOR]