A Review of Physics-Informed Machine Learning Methods with Applications to Condition Monitoring and Anomaly Detection

Yuandi Wu,Brett Sicard,Stephen Andrew Gadsden
2024-01-22
Abstract:This study presents a comprehensive overview of PIML techniques in the context of condition monitoring. The central concept driving PIML is the incorporation of known physical laws and constraints into machine learning algorithms, enabling them to learn from available data while remaining consistent with physical principles. Through fusing domain knowledge with data-driven learning, PIML methods offer enhanced accuracy and interpretability in comparison to purely data-driven approaches. In this comprehensive survey, detailed examinations are performed with regard to the methodology by which known physical principles are integrated within machine learning frameworks, as well as their suitability for specific tasks within condition monitoring. Incorporation of physical knowledge into the ML model may be realized in a variety of methods, with each having its unique advantages and drawbacks. The distinct advantages and limitations of each methodology for the integration of physics within data-driven models are detailed, considering factors such as computational efficiency, model interpretability, and generalizability to different systems in condition monitoring and fault detection. Several case studies and works of literature utilizing this emerging concept are presented to demonstrate the efficacy of PIML in condition monitoring applications. From the literature reviewed, the versatility and potential of PIML in condition monitoring may be demonstrated. Novel PIML methods offer an innovative solution for addressing the complexities of condition monitoring and associated challenges. This comprehensive survey helps form the foundation for future work in the field. As the technology continues to advance, PIML is expected to play a crucial role in enhancing maintenance strategies, system reliability, and overall operational efficiency in engineering systems.
Machine Learning,Artificial Intelligence,Systems and Control
What problem does this paper attempt to address?
This paper reviews the applications of physics-guided machine learning methods in state monitoring and anomaly detection. The study aims to address how to combine physics-based modeling and data-driven approaches to improve the reliability and performance of engineering system state monitoring. Traditional methods rely on physical models and statistical analysis, but they face difficulties when dealing with complex systems and limited accurate physical models. In recent years, Physics-Informed Machine Learning (PIML) has emerged as a new approach that integrates known physical laws and constraints into machine learning algorithms, enabling models to learn from available data while adhering to physical principles. PIML emphasizes the integration of domain knowledge into the framework to enhance accuracy and interpretability, offering advantages over purely data-driven methods. The paper outlines various methods that integrate physical principles into machine learning frameworks, including physical embeddings in feature space, physics-enhanced data, physics-guided regularization, and architecture design. The paper also provides a detailed discussion of the advantages and disadvantages of these methods, considering computational efficiency, model interpretability, and general applicability to different systems. The effectiveness of PIML is demonstrated through case studies and industrial applications, highlighting its potential in condition monitoring and fault detection. In conclusion, this paper addresses the problem of how to leverage physics-informed machine learning to improve strategies for state monitoring. By integrating physics knowledge and data-driven learning, it tackles the challenges of complex systems and enhances maintenance strategies, reliability, and operational efficiency of engineering systems.