Physics-Informed Machine Learning for Industrial Reliability and Safety Engineering: A Review and Perspective

The convergence of physics-informed and machine learning has led to the emergence of Physics-Informed Machine Learning (PIML), a powerful paradigm to enhance the reliability and safety of complex industrial systems. Traditional methods in reliability engineering often rely on physics-based models, which, despite their robustness, face limitations when dealing with highly complex systems involving multiple interacting scales and incomplete knowledge. PIML addresses these challenges by integrating physical laws directly into machine learning models, combining the strengths of data-driven and physics-based approaches. This integration significantly enhances the accuracy and interpretability of predictions, which is crucial for industrial systems’ reliability and safety. Inspired by this motivation, this chapter delves into the state-of-the-art developments in PIML, highlighting its applications, benefits, and challenges within reliability and safety engineering.