Digital-Physical Fatigue Testing via Neural Network Crack Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting and tracking fatigue cracks in structural components under non-constant loads, particularly in lightweight structures like aircraft, are time-consuming and costly, with large-scale tests being inefficient and lacking in accuracy due to limitations in simulating complex structure-property relationships around crack tips.
Innovation Solution
A digital-physical testing system that combines a numerical simulation model with a physical testing machine to automate the process of measuring crack tip stress and crack propagation, using a standardized test specimen to simulate the behavior of larger components, and incorporates a neural network for crack detection and tracking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If large-scale or full-scale testing is conducted on entire aircraft structures, then measurement precision and reliability of fatigue crack detection are improved, but testing time and cost increase significantly
Solution Approach 1:
The patent segments the testing process by separating it into two distinct parts: (1) a digital simulation component that models the entire component's crack tip loading, and (2) a physical testing component that examines only a small standardized test specimen. This segmentation allows the system to avoid time-consuming full-scale physical testing while maintaining accuracy through the combination of digital modeling and targeted physical experimentation.
Solution Approach 2:
The patent creates a digital copy (numerical calculation model) of the entire component's crack tip loading conditions. This digital model replicates the complex structure-property relationships and crack tip mechanics that would otherwise require full-scale physical testing, thereby reducing testing time while preserving measurement precision.
2Measurement precision
If large-scale or full-scale testing is conducted on entire aircraft structures, then measurement precision and reliability of fatigue crack detection are improved, but testing cost increases significantly
Solution Approach 1:
The patent segments the testing process into digital simulation and focused physical testing of standardized specimens. This eliminates the need for expensive full-scale physical tests while maintaining crack detection accuracy through the synergistic combination of numerical modeling and targeted experimentation on small, standardized samples.
Solution Approach 2:
The patent uses a digital calculation model to copy and replicate the crack tip loading conditions of the entire component. This digital replica allows for accurate prediction of crack behavior without requiring expensive full-scale physical testing, thereby reducing testing costs while preserving measurement precision.
3Loss of time
If pure simulation using finite element method is used, then testing time is reduced, but reliability and accuracy of crack propagation prediction deteriorate due to inability to fully represent complex structure-property relationships
Solution Approach 1:
The patent merges digital simulation and physical testing into a hybrid digital-physical system. The numerical calculation model provides rapid prediction of crack tip loading, while standardized test specimens undergo actual fatigue testing to validate and calibrate the model. This combination maintains reliability by grounding the simulation in physical reality while achieving speed through the efficiency of numerical methods.
Solution Approach 2:
The patent implements a feedback loop where results from physical testing of standardized specimens are used to validate and refine the numerical calculation model. This feedback mechanism ensures that the simulation accurately represents complex structure-property relationships and crack tip mechanics, thereby maintaining prediction reliability while reducing overall testing time.
Data Source
Figure 1
AI summary
A digital-physical testing system and a digital-physical testing method for characterizing fatigue and/or cracking processes in components or component structures. In the testing system and method according to the invention, the intended connection between the experiment performed on the test specimen level using the testing machine and the numerical simulation at the overall component level, via the mutual exchange of measured and calculated values, establishes a link between the digital/numerical calculation model and the physical test in the testing machine. Through this parameter exchange, an interconnection is achieved between the experiment in the testing machine and the digital/numerical calculation model, whereby they mutually control and regulate each other. The selected interconnection in the testing system according to the invention creates a digital-physical control loop.