Physics-informed neural networks framework for functionally graded cylinder: Forward analysis, inverse material identification, and stress-driven optimization

Key Points

• The framework enables effective inverse material identification for functionally graded materials, enhancing engineering applications.
• Key evidence shows improved accuracy with a reduction of computational costs in inverse analysis methods.
• Analysis using physics-informed neural networks offers a novel approach to forward analysis in complex material systems.
• The findings highlight potential advancements in material optimization, stressing the need for further validation in real-world scenarios.