Predict Crack Growth Research Articles

A generalized machine learning framework for brittle crack problems using transfer learning and graph neural networks

Despite their recent success, machine learning (ML) models such as graph neural networks (GNNs), suffer from drawbacks such as the need for large training datasets and poor performance for unseen cases. In this work, we use transfer learning (TL) approaches to circumvent the need for retraining with large datasets. We apply TL to an existing ML framework, trained to predict multiple crack propagation and stress evolution in brittle materials under Mode I loading. The new framework, ACCelerated Universal fRAcTure Emulator (ACCURATE), is generalized to a variety of crack problems by using a sequence of TL update steps including (i) arbitrary crack lengths, (ii) arbitrary crack orientations, (iii) square domains, (iv) horizontal domains, and (v) shear loadings. We show that using small training datasets of 20 simulations for each TL update step, ACCURATE achieved high prediction accuracy in Mode-I and Mode-II stress intensity factors, and crack paths for these problems. We demonstrate ACCURATE’s ability to predict crack growth and stress evolution with high accuracy for unseen cases involving the combination of new boundary dimensions with arbitrary crack lengths and crack orientations in both tensile and shear loading. We also demonstrate significantly accelerated simulation times of up to 2 orders of magnitude faster (200x) compared to an XFEM-based fracture model. The ACCURATE framework provides a universal computational fracture mechanics model that can be easily modified or extended in future work.

The Role of Determinism in the Prediction of Corrosion Damage

This paper explores the roles of empiricism and determinism in science and concludes that the intellectual exercise that we call “science” is best described as the transition from empiricism (i.e., observation) to determinism, which is the philosophy that the future can be predicted from the past based on the natural laws that are condensations of all previous scientific knowledge. This transition (i.e., “science”) is accomplished by formulating theories to explain the observations and models that are based on those theories to predict new phenomena. Thus, models are the computational arms of theories, and all models must possess a theoretical basis, but not all theories need to predict. The structure of a deterministic model is reviewed, and it is emphasized that all models must contain an input, a model engine, and an output, together with a feedback loop that permits the continual updating of the model parameters and a means of assessing predictions against new observations. This latter feature facilitates the application of the “scientific method” of cyclical prediction/assessment that continues until the model can no longer account for new observations. At that point, the model (and possibly the theory, too) has been “falsified” and must be discarded and a new theory/model constructed. In this regard, it is important to stress that no amount of successful prediction can prove a theory/model to be “correct”, because theories and models are merely the figments of our imagination as developed through imperfect senses and imperfect intellect and, hence, are invariably wrong at some level of detail. Contrariwise, a single failure of a model to predict an observation invalidates (“falsifies”) the theory/model. The impediment to model building is complexity and its impact on model building is discussed. Thus, we employ instruments such as microscopes and telescopes to extend our senses to examining smaller and larger objects, respectively, just as we now employ computers to extend our intellects as reflected in our computational prowess. The process of model building is illustrated with reference to the deterministic Coupled Environment Fracture Model (CEFM) that has proven to be highly successful in predicting crack growth rate in metals and alloys in contact with high-temperature aqueous environments of the type that exist in water-cooled nuclear power reactor primary coolant circuits.

Computing the Durability of WAAM 18Ni-250 Maraging Steel Specimens with Surface Breaking Porosity

The durability assessment of additively manufactured parts needs to account for both surface-breaking material discontinuities and surface-breaking porosity and how these material discontinuities interact with parts that have been left in the as-built state. Furthermore, to be consistent with the airworthiness standards associated with the certification of metallic parts on military aircraft the durability analysis must be able to predict crack growth, as distinct from using a crack growth analysis in which parameters are adjusted so as to match measured data. To partially address this, the authors recently showed how the durability of wire arc additively manufactured (WAAM) 18Ni-250 maraging steel specimens, where failure was due to the interaction of small surface-breaking cracks with surface roughness, could be predicted using the Hartman–Schijve variant of the NASGRO crack growth equation. This paper illustrates how the same equation, with the same material parameters, can be used to predict the durability of a specimen where failure is due to surface-breaking porosity.

Advantages of the extended finite element method for the analysis of crack propagation in power modules

• XFEM and Level set methods with submodelling has been implemented in this study. • Method easily allows inclusion of crack growth in IGBT wire bond reliability analysis • Bond wire diameter and loop height significantly affect IGBT wire bond reliability. • Wire bond degradation parameter values decrease with increasing crack length. • For similar load density, smaller wire diameter produces longer wire bond lifetime. The techniques of extended finite element method, level set method and the submodelling approach are implemented in this study to model crack and crack growth in ultrasonically bonded thick aluminium wire for the IGBT power electronics modules under different loading conditions for the purpose of lifetime prediction and reliability assessment during design and manufacturing stages. The crack growth and lifetime prediction were performed under cyclic fatigue passive thermal cycling and active power cycling for the bond wire lift-off failure mechanism while the J-integral for different heel crack lengths are predicted under mechanical loads. The analyses showed that the techniques implemented in this paper are effective for modelling such complex geometries and loading conditions and can easily be integrated in a virtual design platform for power electronics. The accuracy of the technique is evaluated by comparing with trends in the published experimental tests and simulation results as well as the standard finite element method which are all in a good agreement. The wire bond crack growth rate under cyclic loading is strongly influenced by the bond thickness and loading conditions.

Fatigue Life Prediction for SiC/Al Materials Based on Path Planning Algorithm Considering Residual Stress

To explore the influence of path deflection on crack propagation, a path planning algorithm is presented to calculate the crack growth length. The fatigue crack growth life of metal matrix composites (MMCs) is estimated based on an improved Paris formula. Considering the different expansion coefficient of different materials, the unequal shrinkage will lead to residual stress when the composite is molded and cooled. The crack growth model is improved by the modified stress ratio based on residual stress. The Dijkstra algorithm is introduced to avoid the cracks passing through the strengthening base and the characteristics of crack steps. This model can be extended to predict crack growth length for other similarly-structured composite materials. The shortest path of crack growth is simulated by using path planning algorithm, and the fatigue life of composites is calculated based on the shortest path and improved model. And the residual stress caused by temperature change is considered to improve the fatigue crack growth model in the material. The improved model can well predict the fatigue life curve of composites. By analyzing the fatigue life of composites, it is found that there is a certain regularity based on metal materials, and the new fatigue prediction model can also reflect this regularity.

Predicting fatigue crack growth through the small and long crack regimes for a military transport aircraft loading spectrum using FASTRAN

Predicting fatigue crack growth across the small and long crack regimes in airframe components is a challenging task. Current fracture mechanics methods require, amongst other things, high fidelity small crack growth rate data and a thorough understanding of the different behaviours of small and long cracks. This paper presents a case study where a fatigue crack growth prediction of Aluminium Alloy 7075-T7351 coupons tested to a military transport aircraft loading spectrum, from nucleation to failure, is attempted using the most recent version of the linear-elastic fracture mechanics-based analysis tool FASTRAN v5.76. The military transport aircraft loading spectrum was applied to different coupon configurations to show the effectiveness of FASTRAN for a range of crack growth scenarios. Crack growth rate data for the prediction in the near-threshold region was experimentally obtained from small cracks using quantitative fractography and this data is shown to be a determining factor in the accuracy of the prediction. This paper highlights the challenges faced by engineers in predicting fatigue crack growth under variable amplitude loading from nucleation to failure and demonstrates the ability of FASTRAN to accurately predict load history effects. Areas for continued improvement are identified with the aim of accurate and reliable fatigue crack growth prediction in a range of structures, conditions, and scenarios.

Influence on structure and fracture mechanics evaluations of a BWR feedwater nozzle following weld overlay repair

Abstract This present paper summarizes applying weld overlay repair on the feedwater nozzle in a boiling water reactor, and evaluates the influence on the structure and fracture mechanics of the component. The requirements for utilizing a full structural weld overlay, including the weld overlay design, effects of welding residual stresses (WRSs), impacts of weld shrinkage, prediction of potential flaw growth, and influence on ASME Code Section III design evaluations, are further defined in ASME Code Case N-504–2. In order to confirm the effectiveness of the weld overlay design of the nozzle, the WRSs following overlay installation were determined using engineering simulation software. The impacts of weld shrinkage from overlay welding are also addressed. The shrinkage stresses were evaluated via a numerical model of the attached piping system to ensure the design margins of the pipe structure. Since the weld overlay technique is well known to be effective in mitigating crack initiation or cracking to leakage, a fatigue crack growth prediction under ASME Code Section XI must be conducted to address the fatigue qualification of the affected component. The ASME Code Section III construction qualification of the whole nozzle region must consider the applicable thermal transient stresses, structural discontinuities, and bimetallic effects that are caused by the weld overlay. This investigation proves that the current stress analyses of the nozzle and the pipe structure are not significantly affected by the added overlay mass, so the structural integrity of the component can be ensured, meeting the requirements of ASME Code Case N-504–2.

Prediction of multiple fatigue crack growth based on modified Paris model with particle filtering framework

This paper presents a novel method based on modified Paris model with particle filtering (PF) framework to prognosis fatigue multi-cracks in metal structures. Since conventional Paris model is only applicable for the single crack, modified Paris model is proposed by introducing the mutual interaction between multiple cracks into stress intensity factor (SIF) expression and utilized to portray multiple fatigue crack growth process. To accurately monitor multi-cracks lengths, Lamb waves are periodically acquired during the fatigue growth of multiple cracks and fed into a deep autoencoder (DAE) network to automatically track response signals variations. Online monitoring model of multi-cracks lengths is then constructed by fitting the mapping relationship between the deep damage feature extracted by the bottleneck layer of DAE network and crack length. PF framework is further adopted to strategically fusing the prediction results of modified Paris model and real-time measurements of online monitoring model to reduce the uncertainties of material parameters and obtain more reliable prognosis results of multi-cracks lengths. The proposed method is demonstrated on center-hole metal specimens with two fatigue cracks. Experimental results show that the proposed method can accurately prognosis two fatigue cracks lengths, in which modified Paris model could more accurately describe multiple fatigue crack growth than conventional Paris model, and online monitoring model based on deep damage feature could more accurately track the fatigue crack length than artificial damage feature.

Suitability Analysis of Machine Learning Algorithms for Crack Growth Prediction Based on Dynamic Response Data.

Machine learning has the potential to enhance damage detection and prediction in materials science. Machine learning also has the ability to produce highly reliable and accurate representations, which can improve the detection and prediction of damage compared to the traditional knowledge-based approaches. These approaches can be used for a wide range of applications, including material design; predicting material properties; identifying hidden relationships; and classifying microstructures, defects, and damage. However, researchers must carefully consider the appropriateness of various machine learning algorithms, based on the available data, material being studied, and desired knowledge outcomes. In addition, the interpretability of certain machine learning models can be a limitation in materials science, as it may be difficult to understand the reasoning behind predictions. This paper aims to make novel contributions to the field of material engineering by analyzing the compatibility of dynamic response data from various material structures with prominent machine learning approaches. The purpose of this is to help researchers choose models that are both effective and understandable, while also enhancing their understanding of the model's predictions. To achieve this, this paper analyzed the requirements and characteristics of commonly used machine learning algorithms for crack propagation in materials. This analysis assisted the authors in selecting machine learning algorithms (K nearest neighbor, Ridge, and Lasso regression) to evaluate the dynamic response of aluminum and ABS materials, using experimental data from previous studies to train the models. The results showed that natural frequency was the most significant predictor for ABS material, while temperature, natural frequency, and amplitude were the most important predictors for aluminum. Crack location along samples had no significant impact on either material. Future work could involve applying the discussed techniques to a wider range of materials under dynamic loading conditions.

Evaluation and optimization of meshless methods to estimation of the 3D-stress intensity factors in mode I–III for fatigue life prediction cracked shaft under uni and multi-axial cyclic loading

In this research, in order to estimate 3D-Stress Intensity Factors in mode I–III, fatigue crack growth and fatigue life prediction of a cracked shaft under various cyclic loading, meshless methods are evaluated and the most appropriate method is selected. For better results, the Base Functions (BFs) are first identified and their performance and efficiency are compared with each other. In addition, while enriching the BFs in all methods and study the effect of increasing the number of sentences of Polynomial-BFs (m = 4, 7,10) and their Linearity or Quadraticity in the accuracy of calculations, two sets of Extend-Enriched-RBFs including the Multi-Quadrics and Gaussian RBFs are used in MQ-RPIM and EXP-RPIM methods. To optimize the shape parameters in the RPIM method and determine the Penalty Factor in the MLPG method, Uni and Multi-objective PSO algorithm was used. Then, a shaft with an initial semi-elliptic surface crack as a 3D-meshless domain for discretizing the weak differential equations was selected and modeled under a fixed latitude tensile, bending, and torsional cyclic loads. Changing the standard values of the shape parameters and estimating their optimal values by the PSO algorithm and comparison of findings with the results of Experimental, MLPG, PIM, FEM, and XFEM methods, has led to the best answer from the applied methods for calculation of; displacement, strain and stress fields, and the SIFs in Mode I–III. Finally, based on the results of uniaxial cyclic load analysis and selection of the MQ-RPIM method, the Multi-axial Cyclic load analysis has been performed on 3D-domain. During this analysis, the Paris Parametric equation along with the Elliptic equation and Liu’s Virtual Strain Energy (VSE) model was used to estimate the fatigue crack growth and fatigue life prediction of a submersible cracked shaft of a pump used in water pumping stations.

Characterization of creep constraint effects on creep crack growth behavior by Q-type parameters

The effect of creep constraint on creep crack growth behavior and the stress field around the creep crack-tip was studied under different specimen geometry conditions. The correlation between the constraint parameters Q and Qm with the stress triaxiality under creep conditions was investigated. The results showed that the Q and Qm depend linearly on the stress triaxiality. Moreover, based on the existing constraint parameters Q∗ and Qm, a new load-independent constraint parameter Qm∗ was proposed to quantify the creep constraint effect. The Qm∗ was superior to Q∗ in quantifying higher out-of-plane creep constraints induced by specimen width and thickness. Furthermore, the Q∗ and Qm∗ were employed to correlate with the normalized creep crack growth rate and initiation time. Finally, the constraint-dependent creep crack growth prediction model based on constraint parameters Q∗ and Qm∗ had been established.

Prediction of fatigue crack growth using convolutional neural network (2nd Report, Prediction of crack propagation on different levels)

Prediction of fatigue crack growth using convolutional neural network (2nd Report, Prediction of crack propagation on different levels)

A framework for automating the parameter determination of crack growth models

Motivated by the need for an efficient fatigue crack growth prediction infrastructure for both legacy and novel materials, we have initiated the development of an automated computational framework capable of determining crack growth model parameters for an equationally defined model. As a first step in addressing this need, the present paper focuses on the Hartman-Schijve crack growth variant of the NASGRO equation by exploring and comparing various global optimization methods for parameter determination and evaluating their performance by using both synthetic and actual data. It also introduces the concept of the total least-squares minimization criterion within the context of crack growth modeling. The development of an open-source software library and an application implementing the approach are also described and are available for distribution to the technical community.

Stress intensity factors of rectangular shape cracks for crack growth prediction

This study aimed at modeling cracks as a rectangular shape to predict crack growth for fitness-for-service assessments. Crack extension driving force was represented by a newly proposed line propagation stress intensity factor (LP-SIF), in which, not a point, but a line on a crack front was assumed to grow in the virtual crack extension for the elastic J-integral calculation. LP-SIF could be derived using the conventional point propagation SIF. It was shown that growth prediction of a rectangular crack using the LP-SIF was more conservative than that obtained by crack growth simulation, in which crack growth was calculated at each node along the crack front. In order to apply the LP-SIF for general use, LP-SIF solutions for various pipe and crack geometries and stress distributions were obtained by finite element analyses. The accuracy of the obtained SIFs was confirmed by comparing with analysis results obtained by two research organizations.

A simple method to determine opening stress for the cracked plate considering thickness effect

In this paper, an approach is proposed to determine crack opening stress (COS) based on the variable quantity of maximum crack opening displacement (VMCOD) considering thickness effect. Based on the three-dimensional elastic-plastic finite element analysis of center cracked tension (CCT) specimens and single-edge cracked tension (SET) specimens, the explicit expressions of relationship between COS and VMCOD are put forward, considering the effect of specimen thickness, crack length, stress ratio, Young's modulus, yield stress and strain hardening. These expressions avoid adopting parameters around crack tip and are applicable to different specimen thicknesses. Further studies show these relationships can be applied to calculate the effective stress intensity factor range and predict crack growth rates for plates of different thicknesses, which correlate well with experimental results.

Kinetic model for prediction of subcritical crack growth, crack tip relaxation, and static fatigue threshold in silicate glass

Prediction of brittle fracture of amorphous oxide glasses continues to be a challenge due to the existence of multiple fracture mechanisms that vary with loading conditions. To address this challenge, we present a model for all three regimes of crack growth in glasses. Regimes I and III are controlled by Arrhenius processes while regime II is transport controlled along with a simple Arrhenius model for viscoelastic stress relaxation. Through dimensional arguments and physical reasoning, we propose a single mechanism which underlies both regime III subcritical crack growth and near-crack-tip viscoelastic relaxation. By combining the subcritical crack growth and viscoelastic models we obtain a prediction for a threshold stress intensity, Kth, below which stresses around the crack relax faster than it propagates. For stress intensity KI<Kth, no subcritical crack growth is predicted to occur, allowing for the design of stable glass systems. The prediction is compared to measured subcritical fracture threshold data for soda-lime silica glasses.

Combined Acoustic Emission and Digital Image Correlation for Early Detection and Measurement of Fatigue Cracks in Rails and Train Parts under Dynamic Loading

Fatigue crack in rails and cyclic-loaded train parts is a contributory factor in multiple railroad accidents. We address the problem of crack detection and measurement at early stages, when total failure has not yet occurred. We propose to combine acoustic emission (AE) testing for prediction of crack growth with digital image correlation (DIC) for its accurate quantitative characterization. In this study, we imitated fatigue crack appearance and growth in samples of railway rail and two train parts by cyclic loading, and applied these two techniques for inspection. Experimental results clearly indicate the efficiency of AE in the early detection of fatigue cracks, and excellent DIC capabilities in terms of geometrical measurements. Combination of these techniques reveals a promising basis for real-time and non-destructive monitoring of rails and train parts.

Numerical prediction of fatigue crack growth based on cumulative plastic strain versus experimental results for AA6082-T6

Numerical prediction of fatigue crack growth based on cumulative plastic strain versus experimental results for AA6082-T6

A numerical study on the correlation of ductile crack extension in Ni-based clad pipe girth welds with circumferential surface cracks and clamped SE(T) specimens

This work describes a numerical investigation on the relationships between ductile crack extension in Ni-based clad pipe girth welds and an API X80 pipeline steel with circumferential surface cracks and SE(T) specimens under fixed grip conditions. A primary objective is to gain further insight and understanding of the effects of constraint on ductile tearing behavior in piping components with dissimilar girth welds and how well crack growth resistance curves measured using single edge notched tension fracture specimens correlate with ductile fracture behavior in circumferentially surface cracked pipes. The numerical description of CTOD−R curves employs a computational cell approach incorporating the Gurson–Tvergaard (GT) constitutive model to simulate ductile crack extension in the crack configurations analyzed. The numerical analyses consider predictions of ductile crack growth in 3-D models of circumferentially cracked pipes with varying crack depth over pipe thickness ratio, a/t, from measured J-resistance curves derived from fracture resistance testing of clamped SE(T) specimens. The analyses also focus on comparing predictions of ductile crack extension using two different definitions for the CTOD.

Peridynamics methodology for elasto-viscoplastic ductile fracture

This paper presents a novel peridynamics (PD) methodology and implementation approach for predicting the ductile fracture of elasto-viscoplastic materials. The proposed analysis model consists of three components: 1) an elasto-viscoplastic constitutive model based on modified Bodner-Partom (BP) constitutive theory; 2) bond damage prediction based on Wang’s unified damage theory; 3) crack growth prediction using the PD critical bond stretch criterion. The PD model for BP materials is realized under the ordinary state-based PD framework to define the bond-wise relationship between deformation state and force state. The use of a unified constitutive theory endows the present model with the characteristic of no yield criterion. In addition, the non-local nature of the PD framework allows discontinuities to naturally occur, thereby ensuring the present model can capture the initiation and propagation of ductile fracture. Simulations of the standard tension tests and mode-I compact tension tests were carried out to demonstrate the use of the PD model for BP materials in practical problems. Particle-wise indicators, representing the intensities of plasticity, damage and fracture, were extracted for post-processing. An in-depth analysis shows that the presented PD model for BP materials can effectively reproduce ductile fractures at the macro-scale, allowing a general PD description of ductile behavior, which brings a new outlook for the application of PD methods.