Research Overview
Ballistic penetration problems involve complex, multi-parameter interactions between projectile geometry, impact velocity, target thickness, and obliquity. Classical analytical models (e.g., Recht-Ipson) capture single-variable trends but fail to generalise across parameter combinations. This research develops data-driven surrogate models for ballistic penetration — using ANN combined with SVD/CP tensor decomposition (the same methodology established in constitutive modeling) to predict ballistic limit velocity (BLV) and residual velocity across the full multi-parameter impact space. The surrogate models are trained on experimental datasets augmented by validated FEM simulations, enabling accurate, physics-grounded predictions for protective structure design.
Surrogate Model · Corresponding Author · 2024
Yunfei Deng, Xiaoyue Yang, Xianglin Huang* · Thin-Walled Structures, Vol. 203, 112161 (2024)
Key Contributions
- General ANN-based ballistic limit velocity (BLV) model accounting simultaneously for target thickness, impact angle, and projectile nose shape
- Combines experimental data with validated FEM simulations to generate training datasets for sparse ballistic data scenarios
- k-Fold cross-validation implemented for reliable model assessment with small sample sizes
- SVD/CP decomposition applied to reveal decoupled relationships between BLV and independent parameters
- Rank-1 decomposition achieves <6% MAPE while capturing fully decoupled effects of all input variables
- Equivalent analytical models derived for direct engineering application
ANN-predicted ballistic limit velocity surface as a function of target thickness and impact angle.
Yunfei Deng, Xiaoyue Yang, Xianglin Huang*, "Determination of ballistic resistance model of finite thickness material based on artificial neural network," Thin-Walled Structures, Vol. 203, 112161 (2024).
DOI: 10.1016/j.tws.2024.112161
Corresponding Author · 2025
Yunfei Deng, Yixu Lv, Xiaoyue Yang, Chunzhi Du, Xianglin Huang* · Thin-Walled Structures, Vol. 216, 113685 (2025)
Key Contributions
- Comprehensive ANN-based residual velocity surrogate model covering multiple impact variables: initial velocity, target thickness, impact angle, projectile nose shape
- Extended dataset generated by combining experimental results with validated numerical simulations
- ANN + SVD/CP decomposition framework extended from BLV surrogate to residual velocity prediction
- "Safe region" and "perforating region" identified and mapped in multi-parameter impact space
- Decoupled CP components physically correlated with material failure mechanisms
- Outperforms classical Recht-Ipson model for multi-parameter residual velocity prediction
Residual velocity surrogate model prediction surface, showing the transition from safe to perforating regime.
Yunfei Deng, Yixu Lv, Xiaoyue Yang, Chunzhi Du, Xianglin Huang*, "Determination of residual velocity model of finite thickness material based on Artificial Neural Network," Thin-Walled Structures, Vol. 216, 113685 (2025).
DOI: 10.1016/j.tws.2025.113685
First Author · 2018
Xianglin Huang, Wei Zhang, Yunfei Deng, Xiongwen Jiang · International Journal of Impact Engineering, Vol. 113, pp. 212–221 (2018)
Key Contributions
- Systematic experimental study of polymer-aluminium bi-layer protective plates against projectile impact
- Comparison of polycarbonate (PC) vs. polymethyl methacrylate (PMMA) polymer layers with AA2024-T4 aluminium
- Effect of projectile nose shape (blunt vs. ogival) on penetration and deformation mechanisms
- Effect of polymer layer placement (impact side vs. back side) on ballistic performance
- PC layers provide superior ballistic resistance improvement over PMMA; nose geometry significantly affects failure mode
- Establishes the experimental dataset and design guidelines that underpin later surrogate model development
Ballistic impact test setup and failure mode characterization of polymer-aluminium layered protective plates.
Xianglin Huang, Wei Zhang, Yunfei Deng, Xiongwen Jiang, "Experimental investigation on the ballistic resistance of polymer-aluminium layered plates," International Journal of Impact Engineering, Vol. 113, pp. 212–221 (2018).
DOI: 10.1016/j.ijimpeng.2017.12.002
Additional Co-Authored Works
Related Impact Mechanics Publications
2024
International Journal of Mechanical Sciences, Vol. 276, 109382
2022
International Journal of Impact Engineering, Vol. 160, 104083
2016–2017
Sandwich structures and underwater impulsive loading (3 papers)
International Journal of Impact Engineering, Vol. 94, 109; AIP Conference Proceedings, 1793