Fatigue reliability analysis of aeroengine blade-disc systems using physics-informed ensemble learning.

Area of Science:

• Aerospace Engineering
• Mechanical Engineering
• Computational Science

Background:

• Traditional methods for aeroengine blade-disc fatigue reliability analysis suffer from low computational efficiency or accuracy.
• Complex multi-component systems pose significant challenges for existing reliability assessment techniques.

Purpose of the Study:

• To propose a novel physics-informed ensemble learning (PIEL) method for efficient and accurate fatigue reliability analysis of aeroengine blade-disc systems.
• To establish a PIEL-based system reliability framework incorporating physical characteristics and failure correlations.

Main Methods:

• Decomposing complex system reliability analysis into single-component analyses based on physical characteristics.
• Developing a PIEL model by integrating constitutive responses and multi-material physical properties into ensemble learning.
• Quantifying component failure correlations using the Copula function to build a system reliability framework.

Main Results:

• The proposed PIEL method demonstrated superior computational accuracy and efficiency compared to direct Monte Carlo, support vector regression, neural networks, ensemble learning, and physics-informed neural networks.
• Validation on a typical aeroengine high-pressure turbine blade-disc system confirmed the method's effectiveness.
• The PIEL approach significantly enhances the reliability analysis of complex engineering systems.

Conclusions:

• The PIEL method offers a significant advancement in fatigue reliability analysis for aeroengine blade-disc systems.
• This work provides novel insights into physics-informed modeling for structural integrity applications.
• The developed framework is validated as an efficient and accurate tool for complex system reliability.