What the study found
Explainable machine learning models were able to predict high-strain-rate mechanical responses of laser powder bed fusion (LPBF)-fabricated A286 steel lattice structures tested under split Hopkinson pressure bar (SHPB) loading. The study reports that prediction performance varied by response metric and model, and that the results showed nonlinear interactions between impact pressure and lattice type.
Why the authors say this matters
The authors say the framework can provide transparent, high-fidelity predictions and support data-driven lattice design strategies for impact-resistant applications in aerospace and defence. They also conclude that the explainable AI (XAI) approach can help identify interpretable design rules and reduce reliance on exhaustive physical prototyping.
What the researchers tested
Three lattice topologies were additively manufactured: body-centred cubic (BCC), honeycomb and gyroid. They were tested under dynamic pressures from 2–7 bar, and key response variables were extracted: peak stress, maximum strain, maximum strain rate and energy absorbed. Multi-output regression models using CatBoost, XGBoost, Random Forest, Gradient Boosting and Extra Trees were trained with impact pressure and lattice type as input features.
What worked and what didn't
CatBoost gave the highest accuracy for peak stress prediction (R² = 0.9848), XGBoost for maximum strain (R² = 0.9877), Gradient Boosting for strain rate (R² = 0.9659) and Random Forest for energy absorption (R² = 0.9839). Shapley additive explanations showed nonlinear interactions between impact pressure and lattice type, especially beyond 6 bar. Surrogate trees and local interpretable model-agnostic explanations identified design rules, including BCC lattices at 6 bar or above being associated with optimal energy absorption.
What to keep in mind
The study notes that the dataset was relatively small because SHPB testing and LPBF fabrication cycles were experimentally constrained. The authors say generalizability to other alloys, lattice topologies or loading regimes still needs validation, and the framework does not include temperature effects, anisotropy or microstructural evolution during impact.
Key points
- The study predicted dynamic mechanical responses of LPBF-fabricated A286 lattice structures using explainable machine learning.
- Different models performed best for different outputs, with R² values reported as high as 0.9877.
- Impact pressure and lattice topology showed nonlinear interactions, especially above 6 bar.
- Interpretability tools identified a rule linking BCC lattices at 6 bar or above with optimal energy absorption.
- The authors note limits from a small dataset and say broader validation is still needed.
Disclosure
- Research title:
- Explainable models predicted dynamic responses of LPBF A286 lattices
- Authors:
- Veera Siva Reddy Bobbili, Chandrasekara Sastry C, Hafeezur Rahman A.
- Institutions:
- Kurnool Medical College
- Publication date:
- 2026-01-21
- OpenAlex record:
- View
- Image credit:
- Photo by Muhammad Mahedi Hasan on Pexels · Pexels License
Get the weekly research newsletter
Stay current with peer-reviewed research without reading academic papers — one filtered digest, every Friday.