Fatigue crack growth prediction method based on machine learning model correction

Fang, Xin; Liu, Guijie; Wang, Honghui; Xie, Yingchun; Tian, Xinliang; Leng, Dingxin; Mu, Weilei; Ma, Penglei; Li, Gongbo

doi:10.1016/j.oceaneng.2022.112996

Cited by 12 publications

(2 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To find the parameters in polynomial regression, the usual optimization method is the least squares method (LSM). 25 Moreover, the gradient descent optimization method can be used 26 ; however, this method is preferred to be used for big data. The objective the optimization methods is to minimize the sum of squared differences between the observed values da dN À Á i and the predicted values f ΔK; θ ð Þ i .…”

Section: Polynomial Regression Modelmentioning

confidence: 99%

Quantifying uncertainty in fatigue crack growth of SLM 316L through advanced predictive modeling

Haselibozchaloee,

Correia,

Braga

et al. 2024

Fatigue Fract Eng Mat Struct

View full text Add to dashboard Cite

Optimizing structural designs is crucial today, with additive manufacturing, particularly selective laser melting, gaining prominence. Thorough mechanical characterization of new materials remains vital. This paper investigates fatigue crack growth behavior in SLM 316L specimens under cyclic loading conditions. The study addresses result uncertainties by using CT specimens aligned along three building directions per ASTM E647 standards and a constant loading ratio (R = 0.1), necessitating mean value and confidence interval predictions. Departing from linear prediction models, innovative Bootstrap Polynomial and Power Regression Models and Bayesian Nonlinear Regression Model updated posterior distribution by Markov Chain Monte Carlo are employed. These approaches leverage bootstrapping to construct confidence intervals, offering robustness and flexibility in handling non‐normal data behavior and limited sample sizes. They provide tailored fits to data curvature, revealing limitations of linear prediction models in capturing observed nonlinear behavior, enhancing reliability in additive manufacturing applications, and advancing material science and engineering.

show abstract

Section: Polynomial Regression Modelmentioning

confidence: 99%

Quantifying uncertainty in fatigue crack growth of SLM 316L through advanced predictive modeling

Haselibozchaloee,

Correia,

Braga

et al. 2024

Fatigue Fract Eng Mat Struct

View full text Add to dashboard Cite

show abstract

“…In recent years, data science, represented by big data and machine learning [18], has become a popular tool in science and engineering in the new era. Some researchers have extended their research tools to machine learning and neural networks and applied new data-driven methods combined with traditional fatigue theory systems to predict fatigue life [19][20][21][22] and predict short crack propagation from a macroscopic perspective [23][24][25]. Some researchers have also applied data-driven methods at the microscopic level to obtain quantitative relationships between microstructure and short crack growth.…”

Section: Introductionmentioning

confidence: 99%

A machine learning study on the fatigue crack path of short crack on an α titanium alloy

Shen,

Lv,

et al. 2023

Phil. Trans. R. Soc. A.

View full text Add to dashboard Cite

In the present study, a physics-informed neural network model based on Bayesian hyperparameter optimization is proposed for the prediction of short crack growth paths. A large number of cyclic loadings at a lower amplitude were applied to an α titanium sample by an ultrasonic fatigue machine to ensure a sufficient amount of data for machine learning. The grain size, grain orientation and grain boundary direction on the path, as well as crack growth direction, were selected as feature data for training the prediction model. The optimizations of the size ratio and the angle operation were conducted to compare different data processing methods, respectively. After evaluation, eventually, a model for predicting crack growth path is obtained with a reliable performance of 10% tolerance on the path angle at each grain boundary. And the prediction effect of the proposed model is better than that of some classic machine learning models and slip trace analysis. This article is part of the theme issue 'Physics-informed machine learning and its structural integrity applications (Part 1)'.

show abstract

Numerical modeling of cyclic softening/hardening behavior of carbon steels from low- to high-cycle fatigue regime

Fincato

Yonezawa

Tsutsumi

2023

Archiv.Civ.Mech.Eng

View full text Add to dashboard Cite

The aim of this study is to characterize the stress–strain behavior of three construction steels (SM490, SM570, and F18B) through both experimental and numerical investigations. The material performance was evaluated by conducting tests on round bar specimens subjected to monotonic, fatigue, and incremental step fully reversed loading conditions. The experimental campaign was conducted to provide valuable information on the mechanical performances of the steels and data for calibrating the material constants required for numerical analyses. The numerical simulations aimed to demonstrate the effectiveness of the proposed unconventional plasticity model, the Fatigue SS model (FSS), in describing the non-linear behavior of the materials under a broad range of loading conditions, including stress states below and beyond the macroscopic yield condition. This aspect is a significant advantage of the FSS model, as conventional elastoplastic theories fail to provide a phenomenological description of inelastic material deformation under stress states within the yield condition. The good agreement between the experimental and numerical results confirms the validity of the calibration of the material constants and the reliability of the computational approach.

show abstract

Fatigue crack growth prediction method based on machine learning model correction

Cited by 12 publications

References 32 publications

Quantifying uncertainty in fatigue crack growth of SLM 316L through advanced predictive modeling

Quantifying uncertainty in fatigue crack growth of SLM 316L through advanced predictive modeling

A machine learning study on the fatigue crack path of short crack on an α titanium alloy

Numerical modeling of cyclic softening/hardening behavior of carbon steels from low- to high-cycle fatigue regime

Contact Info

Product

Resources

About