A survey on machine learning approaches for uncertainty quantification of engineering systems
Обзор подходов машинного обучения к количественной оценке неопределенности инженерных систем
2025-01-30
SCID: 54.1/9epjmp2u
Discuss with AI
Gaussian process regressionengineering systemsmachine learningphysics-informed neural networksuncertainty quantification
Figures from the paper
Abstract (AI)
Abstract Uncertainty quantification (UQ) is essential for understanding and mitigating the impact of pervasive uncertainties in engineering systems, playing a crucial role in modern engineering practice. As engineering products grow increasingly complex and the demand for highly accurate UQ results intensifies, the need for efficient UQ methods has become paramount. Recently, machine learning (ML) techniques, including Gaussian process regression, artificial neural networks, physics-informed neural networks, and many others, have garnered significant attention in both theoretical research and practical applications. The exceptional capability of ML methods to address highly complex problems has positioned them as leading solutions for UQ challenges. This work provides a comprehensive survey of recent advancements in ML-assisted UQ approaches, covering ML techniques for forward UQ analysis, inverse UQ analysis, fault diagnosis, resilience assessment, and management planning. The most prominent and effective ML techniques such as the physics-informed, physics-guided, physics-enhanced, physics-corrected, and physics-constrained approaches, along with advanced strategies and frameworks, are thoroughly reviewed and summarized. Additionally, this paper discusses the challenges facing the field and outlines potential directions for future research. This work serves as a valuable resource for researchers and engineers seeking an in-depth understanding of state-of-the-art ML approaches to UQ tasks of engineering systems.
Key Findings
1
Gaussian process regression, artificial neural networks, and physics-informed neural networks are among the prominent techniques reviewed for engineering UQ.
2
Machine learning methods are increasingly important for efficient uncertainty quantification in increasingly complex engineering systems.
3
Physics-integrated strategies—including physics-informed, physics-guided, physics-enhanced, physics-corrected, and physics-constrained approaches—are comprehensively summarized.
4
The survey covers ML-assisted approaches for forward and inverse uncertainty quantification, fault diagnosis, resilience assessment, and management planning.
5
The survey identifies current challenges and proposes future research directions for advancing machine-learning-based uncertainty quantification.
Research Object
engineering systems
Research Subject
machine-learning-assisted uncertainty quantification, including forward and inverse analysis, fault diagnosis, resilience assessment, and management planning
Publication Details
Publication Date
2025-01-30
Journal
Publisher
ISSN
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest