Prediction of Crack Propagation Based on Dynamic Bayesian Network
- Resource Type
- Conference
- Authors
- Chang, Qi; Chen, Lele; Zhao, Heng; Xie, Fangqin
- Source
- 2021 IEEE 15th International Conference on Electronic Measurement & Instruments (ICEMI) Electronic Measurement & Instruments (ICEMI), 2021 IEEE 15th International Conference on. :92-97 Oct, 2021
- Subject
- Bioengineering
Robotics and Control Systems
Signal Processing and Analysis
Predictive models
Prediction algorithms
Fatigue
Mathematical models
Particle filters
Inference algorithms
Data models
crack propagation
Dynamic Bayesian Network
particle filter
ABAQUS
- Language
Aiming at the problem of inaccurate prediction of fatigue crack propagation due to uncertain factors, a method based on Dynamic Bayesian Network (DBN) is proposed in this paper. A crack propagation simulation model is established in the finite element analysis software ABAQUS. The Paris formula is combined with the finite element model(FEM) of the crack propagation to establish the state equation. And the crack propagation prediction model is constructed based on the uncertain parameters defined in the FEM. The strain sensors are adopted to monitor the crack propagation. The strain data and the crack length data are fitted into a function to construct a fatigue crack observation model, and the particle filter algorithm is used to revise the uncertain parameters and to predict the crack propagation. The experimental research shows that the model can be revised continuously through the DBN. The accuracy of prediction for the rest usage life(RUL) of the structure can be improved greatly. The credibility and validity of the method are also proved.