Development and Comparison of Model-Based and Data-Driven Approaches for the Prediction of the Mechanical Properties of Lattice Structures

2024-10-08

Journal of materials engineering and performance (Print)

AI Summary

Accurately and efficiently predicting the mechanical properties of lattice structures in additive manufacturing remains challenging due to inherent trade-offs between prediction fidelity and computational speed. Method: This study establishes a unified experimental benchmark and conducts the first systematic comparison of analytical models, reduced-order finite element models, and data-driven approaches—including random forests and graph neural networks (GNNs)—in terms of generalizability and interpretability. We propose a physics-informed hybrid modeling framework that integrates model-driven and data-driven components. Contribution/Results: The data-driven submodel achieves high accuracy in compressive strength prediction (MAE = 3.1 MPa; error < 5.2%), while the model-driven submodel delivers >1000× computational speedup. The integrated framework simultaneously ensures high predictive accuracy and real-time inference capability, offering an interpretable, deployable paradigm for intelligent lattice structure design.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Problem

• 3D printing
• lattice structures
• mechanical properties

Innovation

• 3D printing
• lattice structure optimization
• predictive modeling