Accurate Fatigue Life Prediction Research Articles

Enhanced fatigue SN curve generation for gear bending fatigue life prediction using multi-objective optimization algorithms

This study presents a comprehensive methodology for fatigue life prediction and SN curve generation in gear systems, integrating specimen and product-level testing, simulations, and multi-objective optimization. The approach addresses the discrepancy between specimen-based testing and actual product performance through a novel multi-fidelity framework. To solve the proposed multi-objective optimization problem, several optimization algorithms were compared and analyzed, demonstrating the effectiveness of each approach in balancing different objectives. A key innovation is the development of an empirical formulation-based surrogate model, enabling efficient optimization while significantly reducing computational demands. The methodology covers both specimen and product-level testing along with simulations, offering a holistic solution for accurate fatigue life prediction. Extensive validation through experiments and comparison with commercial software confirms the reliability and practical applicability of the proposed method. This approach provides a sustainable and efficient solution for SN curve generation and fatigue life prediction in automotive gear train development, allowing for continuous improvement as new data becomes available. Its practicality and accuracy make it particularly valuable for early-stage design optimization and large-scale applications not only in the automotive industry but also in many other industries where gear fatigue analysis is critical.

Prediction of multiaxial fatigue life with a data-driven knowledge transfer model

A data-driven model is presented for accurate prediction of multiaxial fatigue life based upon the principle of transfer learning (TL). The Tradaboost framework is explored to adjust the weights of training data from different sources, actuating information transfer from domain knowledge to the data-driven modeling of multiaxial fatigue life. Subsequently, extensive experimental results tested under the proportional and non-proportional circle loadings are collected for model evaluation. The results demonstrate that the proposed model is more accurate than domain knowledge-based, conventional data-driven, and comparable TL-based models, with a low data requirement, showcasing good applicability for multiaxial fatigue life assessment.

A frequency domain enhanced multi-view neural network approach to multiaxial fatigue life prediction for various metal materials

Fatigue damages and failure widely exist in engineering structures, particularly in industries such as aerospace, automotive, and construction, where components are often subjected to complex multiaxial loading conditions. Accurate prediction of fatigue life is critical for ensuring the safety and longevity of these structures. In this study, a novel multi-view deep learning model incorporating frequency domain analysis for fatigue life prediction is proposed. The proposed model integrates a Convolutional Neural Network (CNN), a Long Short-Term Memory Network (LSTM), and FNet that combines frequency domain analysis, in a parallel structure to extract features from the loading paths of materials. These extracted features are then connected to a fully connected neural network to predict fatigue life. The model was validated using fatigue data collected from 6 different materials, encompassing 17 loading paths and 336 samples. Additionally, ablation experiments were conducted, and the extrapolation capabilities were evaluated using specifically designed test sets. The results demonstrate that the proposed model exhibits excellent predictive performance and extrapolation capabilities. We anticipate that the multi-view approach, along with its accuracy and applicability, can offer potential applications in engineering fields that require reliable, data-driven models to assess material durability under complex loading scenarios.

Probabilistic hydrogen-assisted fatigue crack growth under random pressure fluctuations in pipeline steels

The increasing demand for energy and the global threat of climate change have driven the search for alternative energy sources, with hydrogen emerging as a prominent substitute for fossil fuels. The fatigue behavior of pipeline steels under gaseous hydrogen is a critical problem that is impeding the industry's adoption of hydrogen into the current natural gas infrastructure. A brief review of existing hydrogen-assisted fatigue crack growth (HA-FCG) studies, which reveal several key gaps, is given first. Existing HA-FCG models predominantly address constant amplitude loading, while the realistic driving force is random loading in gas pipelines. Also, the current uncertainty quantification studies for HA-FCG focus on material randomness and overlook the large uncertainties associated with random pressure fluctuations. To address these issues, this study proposes a HA-FCG model that utilizes a time-based subcycle approach, allowing for direct application to random spectrum loads without the need for cycle counting. A model parameter as a function of hydrogen operating conditions is introduced to capture the different regimes in HA-FCG, and the model predictions are compared with ASME B31.12 code. Following this, statistical analysis of random pressure fluctuation data collected from natural gas pipelines at multiple locations is performed. The realistic industry pressure data shows distinct statistical features, and it is observed that the high-fidelity data (high sampling frequency) is beneficial for accurate fatigue life predictions. Uncertainty quantification and load reconstruction are performed by the Karhunen–Loève expansion with a post-clipping procedure, leading to a probabilistic HA-FCG analysis. The paper concludes with key findings and suggests directions for future research.

Multiaxial low-cycle fatigue life model for notched specimens considering small sample characteristics

PurposeSufficient sample data are the necessary condition to ensure high reliability; however, there are relatively poor fatigue test data in the engineering, which affects fatigue life's prediction accuracy. Based on this, this research intends to analyze the fatigue data with small sample characteristics, and then realize the life assessment under different stress levels.Design/methodology/approachFirstly, the Bootstrap method and the principle of fatigue life percentile consistency are used to realize sample aggregation and information fusion. Secondly, the classical outlier detection algorithm (DBSCAN) is used to check the sample data. Then, based on the stress field intensity method, the influence of the non-uniform stress field near the notch root on the fatigue life is analyzed, and the calculation methods of the fatigue damage zone radius and the weighting function are revised. Finally, combined with Weibull distribution, a framework for assessing multiaxial low-cycle fatigue life has been developed.FindingsThe experimental data of Q355(D) material verified the model and compared it with the Yao’s stress field intensity method. The results show that the predictions of the model put forward in this research are all located within the double dispersion zone, with better prediction accuracies than the Yao’s stress field intensity method.Originality/valueAiming at the fatigue test data with small sample characteristics, this research has presented a new method of notch fatigue analysis based on the stress field intensity method, which is combined with the Weibull distribution to construct a low-cycle fatigue life analysis framework, to promote the development of multiaxial fatigue from experimental studies to practical engineering applications.

Investigating the relationship between natural frequency and residual strength and stiffness of cross‐ply laminate under cyclic loading

AbstractPredicting the fatigue life of wind turbine rotor blades is a challenging and crucial engineering task. This study investigates the correlation between modal parameters and the degradation of residual strength and tensile modulus in the cross‐ply glass epoxy laminate [0/90]7 used in wind turbine blades under fatigue loading. The tensile and vibration characteristics were assessed, followed by constant amplitude fatigue tests at 35%, 43% and 55% of the ultimate tensile strength, with R = 0.1 and a frequency of 8 Hz. The modal analysis was performed on the cycled specimens at life fractions from 0.05 to 0.70 and residual modulus and strength were obtained. The results establish a well‐defined correlation between these residual mechanical properties and the natural frequency. Normalized residual strength, tensile modulus and natural frequency demonstrated similar behaviors during the fatigue life. An initial rapid decrease in the first tenth of the life fraction was observed, followed by minimal changes up to a life fraction of 0.7. The strong correlation between the first mode natural frequency and both the residual strength and the tensile E‐modulus provides a promising basis for developing accurate fatigue life prediction models for fiber‐reinforced composite structures. © 2024 Society of Chemical Industry.

An enhanced multiaxial low-cycle fatigue life model

To improve the accuracy of predicting multiaxial fatigue life, an enhanced multiaxial low-cycle fatigue life model is proposed based on the FS model. This model introduces a correction parameter for the stress-related term to consider the influence of normal stress and shear stress on the critical plane. The feasibility of the model is investigated, and it is validated using fatigue test data from eight different materials. The results indicate that the proposed model is applicable for both symmetric and asymmetric loading conditions under constant amplitude loading, with highly accurate fatigue life prediction results.

A physics-informed neural network approach for predicting fatigue life of SLM 316L stainless steel based on defect features

Defects in additively manufactured materials severely limit the performance of parts in practical applications, often exposing them to the risk of fatigue failure. In order to improve the reliability and performance of additively manufactured parts, it becomes crucial to accurately predict the fatigue life of the material. Although traditional semi-empirical formulas can assess the effect of defects on the fatigue performance of parts, they still lack detailed research and consideration of defect morphology features. Therefore, this study proposes a method based on Physics-Informed Neural Networks (PINN). This method improves the predictive capability of the model and enhances its interpretability by extracting the sensitive features of critical defects and embedding known physical knowledge or fracture mechanics methods as loss functions into the training process of the neural network. Additionally, the method effectively captures the complex relationship between defect features and fatigue life, providing a deeper understanding of the model prediction results. The results show that the PINN model considering feature-related knowledge has higher prediction accuracy and reliability, and all predicted fatigue life are narrowed within 2-factor bands, enabling more accurate prediction of fatigue life for SLM 316L stainless steel under different processing conditions.

Nonlinear fatigue damage constrained topology optimization

In engineering applications, plenty of components are subjected to variable-amplitude cyclic loading, resulting in fatigue damage, which is one of the main forms of structural damage. While the linear damage rule has long served as a fundamental approach, its limitations necessitate advancements for more accurate fatigue life predictions. Hence, this paper introduces a pioneering method termed nonlinear fatigue damage constrained topology optimization (NFDCTO). This method integrates several key components: the rainflow counting method to evaluate the non-proportional cyclic load levels, Basquin's equation to describe the S-N curve, and Morrow's plastic work interaction rule to calculate the nonlinear cumulative damage of the structure. Consequently, we establish a mathematical model for the NFDCTO method based on these components. The incorporation of penalized damage aggregation with the P-Normal function is used to address significant constraints and singularity challenges. Furthermore, by employing the adjoint method, we derive sensitivity equations for both the objective function and fatigue constraint function concerning the design variables. Subsequently, the superiority of the NFDCTO method over traditional linear fatigue damage method were verified applying 2D l-shaped beam and simply supported structures as examples. Concurrently, bridge structures were employed to investigate the effect of the sensitivity index of the material to the history of variable amplitude stresses on the optimization results. In addition, the influence of the fatigue penalty factor on the topology optimization results was assessed using a 2D cantilever beam. Finaly, we verify the effectiveness and feasibility of the NFDCTO method in addressing fatigue optimization challenges for 3D structures utilizing two 3D examples.

Accuracy analysis of fatigue life prediction in asphalt binders under multiple aging conditions based on simplified viscoelastic continuum damage (S-VECD) approach

Fatigue damage of asphalt binder is a notable distress at the intermediate temperature. Several empirical and mechanistic models can be utilized to predict the fatigue life under different loading modes based on limited laboratory data. The prediction accuracy of utilizing different models remains a question. This work utilized the S-VECD based approaches to analyze the effect of failure definitions on prediction accuracy. Different fatigue failure criteria are applied to linear amplitude sweep (LAS) and time sweep (TS) results, inclusive of highly-aged asphalt binders. The candidate methods include original S-VECD approach of S at failure (Sf), GR, and DR failure criteria. The GR failure criterion is found to be effective in predicting fatigue failure within three candidate models by comparing measured and predicted results. Various factors influencing the prediction accuracy of fatigue life prediction are discussed. The purpose of this work is to compare the currently available prediction models based on S-VECD approaches and further investigate the potential effect of the different factors for determining the variables within the models.

Minimum life point determination and fatigue life prediction of thermal barrier coatings based on extended stress and strain method

The study develops the extended stress and strain method (ESSM) capable of extracting the stress-strain perpendicular to the 3D thermal barrier coatings (TBCs) interface and shear stress-strain along the interface. Then, the stresses perpendicular to the interface and the shear stresses along the interface are analyzed in comparison with the conventional stresses. Finally, fatigue life prediction of TBCs is carried out based on the ESSM. The results show that, the cycle maximum value locations of shear stress along the interface are located at 5.1 μm along the circumferential direction to the peak. The previous experimental study of TBCs proves that it is reasonable to use the shear stress to determine the location of the minimum life point, and the minimum life point location is the cycle maximum value location of the shear stress. The maximum error in fatigue life prediction of TBCs using the ESSM 45.6 %, which is 20.5 % less than that using equivalent strain. The ESSM can directly determine the location of minimum life point and improve the fatigue life prediction accuracy of TBCs, which has great potential in damage analysis and fatigue life prediction.

Fatigue life prediction method for reinforced concrete deck slabs subjected to repeated moving wheel loads and failing in shear

AbstractAs bridge deck slabs experience the direct impact of repeated concentrated wheel loads, they are more susceptible to fatigue‐related damage than other bridge components. In this context, a fatigue life prediction method for reinforced concrete (RC) bridge decks has been developed based on failure mechanisms of RC slabs tested under moving wheel loads. Initially, the formulas encompassed shear strength and S–N equations, accommodating variations in environmental and support conditions, are proposed to facilitate the fatigue life prediction of slabs exposed to constant moving wheel load. Subsequently, a model for shear strength degradation formulated using experimental findings on beam‐like elements is utilized in predicting the fatigue life of slabs subjected to stepwise moving wheel loads. The validity of the proposed method was established by comparing its results with previous experimental data. Notably, compared with the existing prediction equation, the proposed method exhibited more accurate fatigue life prediction results.

A Novel Physics-Guided Neural Network for Predicting Fatigue Life of Materials

A physics-guided neural network (PGNN) is proposed to predict the fatigue life of materials. In order to reduce the complexity of fatigue life prediction and reduce the data required for network training, the PGNN only predicts the fatigue performance parameters under a specific loading environment, and calculates the fatigue life by substituting the load into the fatigue performance parameters. The advantage of this is that the network does not need to evaluate the effect of numerical changes in the load on fatigue life. The load only needs to participate in the error verification, which reduces the dimension of the function that the neural network needs to approximate. The performance of the PGNN is verified using published data. Due to the reduction in the complexity of the problem, the PGNN can use fewer training samples to obtain more accurate fatigue life prediction results and has a certain extrapolation ability for the changes in trained loading environment parameters. The prediction process of the PGNN for fatigue life is not completely a black box, and the prediction results are helpful for scholars to further study the fatigue phenomenon.

A P–S–N curve fitting method based on mixed Weibull distribution and expectation-maximization algorithm

In practical engineering, there are many factors affecting the fatigue life of actual components, and a single Weibull distribution is insufficient to accurately describe the impact of various factors on fatigue life. In order to solve the above problems, 316L stainless steel is taken as the experimental object to study the effects of phase difference and stress levels on fatigue life based on the mixed Weibull distribution model. Firstly, the mixed Weibull distribution model is converted into the mixed extreme value distribution model, and used Euler's constant to calculate relevant parameters of extreme value distribution and Expectation-Maximization (EM) algorithm to determine the probability density function of the mixed Weibull distribution. Secondly, the probability density function is integrated to obtain the distribution function and the fatigue life under different failure probabilities is calculated. Finally, the influence of non-proportional additional reinforcement and stress amplitude on fatigue life is studied by taking them as hidden variables of mixed Weibull distribution, which is used to obtain P-S-N curves for various failure probabilities. Through the proposed method, the influence of different factors on fatigue life can be considered from the perspective of statistics, and the accuracy of fatigue life prediction can be promoted.

Multi-Stage Hybrid Loading Fatigue Life Prediction Study

Abstract Fatigue is a damage accumulation process in which material properties are continuously deteriorated and degraded under cyclic loading. Damage accumulation analysis plays a key role in preventing the occurrence of fatigue failure, and the damage evolution mechanism is an important parameter of interest for fatigue behavior. The fatigue cumulative damage life prediction model is established by focusing on the damage evolution mechanism, and an improved Manson–Harford cumulative damage nonlinear model is developed to calculate fatigue damage life by considering the effect of load ratio of multi-stage loading on fatigue life. A fatigue life prediction method considering high-low cycles and multi-stage mixed loads is proposed. The fatigue life prediction accuracy of the model is verified by fatigue tests on structural samples. The accuracy of the fatigue life prediction error is the highest among similar models under the same test, the model proposed in this paper.

PINN-AFP: A novel C-S curve estimation method for asphalt mixtures fatigue prediction based on physics-informed neural network

The accurate prediction of fatigue life of asphalt mixture is the key to the design of long-lasting durable pavement. Currently, a critical aspect influencing the accuracy of life prediction methods for asphalt mixtures, based on Viscoelastic Continuum Damage Mechanics (VECD), is the precision of the damage characteristic (C-S) curve. Within the VECD framework, the C-S curve of asphalt mixture is regarded as an intrinsic material property. However, it has been observed in engineering applications that the C-S curve of the material exhibits notable sensitivity to varying loading conditions. The nonlinear fitting method based on a small number of experiments is difficult to fully characterize the fatigue performance of the material, while a large number of complete material fatigue tests are expensive in time and money. Based on the above problems, a physics-informed neural network embedded in VECD, PINN-AFP, is proposed. It can accurately predict the complete material C-S curve based on a small amount of pre-fatigue data of the material, thus achieving accurate prediction of the fatigue life of asphalt mixture. The case study uses the fatigue test data of AC-25 as an example, and the results demonstrate that the proposed PINN-AFP has strong generalization ability and prediction accuracy, achieving the state-of-the-art in the mainstream machine learning and deep learning methods with an average 5.2% fatigue life prediction error.

Research on fatigue curve fitting methods based on the notch stress approach

PurposeThis study investigates the fitting techniques for notch fatigue curves, seeking a more reliable method to predict the lifespan of welded structures.Design/methodology/approachBuilding on the fatigue test results of butt and cruciform joints, this research delves into the selection of fitting methods for the notch fatigue curve of welded joints. Both empirical formula and finite element methods (FEMs) were employed to assess the notch stress concentration factor at the toe and root of the two types of welded joints. Considering the mean stress correction and weld misalignment coefficients, the notch fatigue life curves were established using both direct and indirect methods.FindingsAn engineering example was employed to discern the differences between the direct and indirect approaches. The findings highlight the enhanced reliability of the indirect method for fitting the fatigue life curve.Originality/valueWhile the notch stress approach is extensively adopted due to its accurate prediction of component fatigue life, most scholars have overlooked the importance of its curve fitting methods. Existing literature scantily addresses the establishment of these curves. This paper offers a focused examination of fatigue curve fitting techniques, delivering valuable perspectives on method selection.

Axial and bending very high cycle fatigue of completion string

Preventing or mitigating very high cycle fatigue (VHCF) damage and failure of completion string is essential for ensuring their long-term stable operation. Therefore, a comprehensive understanding of the string's fatigue performance in tension, compression, and bending over an extended service life is necessary. This study investigated the evolution, microscopic mechanisms, fracture behavior, and life prediction of VHCF damage in completion string under axial and bending conditions. It was found that the intragranular twin structures on the critical interface of VHCF source directly participate in lattice reconstruction, enhancing the grain's strain coordination ability, and increasing the proportion of facet area on the fatigue fracture surface. Based on the fracture characteristics of VHCF, the Mayer empirical formula was improved, and a novel convolutional neural network (CNN) based VHCF life prediction model was proposed. Utilizing the CNN's powerful fracture feature extraction and nonlinear modeling capabilities, the VHCF fatigue life prediction accuracy was significantly enhanced, with prediction errors all within 2 times the experimental results' error range. Based on the axial and bending VHCF fracture mechanisms of completion string and CNN, a more advanced method for predicting axial and bending VHCF life is provided, offering theoretical supports for the engineering safety service and life assessment of completion string.

A novel multiaxial fatigue life prediction model based on the critical plane theory and machine-learning method

In order to characterize the fatigue failure and damage mechanism under complex multiaxial loads, several multiaxial semi-empirical fatigue models, such as Fatemi-Socie (FS), Smith-Watson-Topper (SWT) and Wang-Brown (WB) models, were proposed to explain the relationship between fatigue life and stress/strain based on experimental analysis or observation. Although the semi-empirical model is widely used in practice because of its simplicity, but it is difficult to uniformly model the mean stress effect of a wide range of materials and loading conditions. To address this issue, a multiaxial fatigue life prediction model based on critical plane theory and machine learning is proposed in this work. Through the multi-layer stacking mechanism, the model comprehensively utilizes domain knowledge and original data information, and integrates the advantages of different models in capturing data and utilizing features. The experimental results showed that the proposed model achieves stable and highly accurate fatigue life prediction of the GH4169, wrought Ti-6Al-4V and TC4 materials with complex working conditions.

Recent developments and future trends in fatigue life assessment of additively manufactured metals with particular emphasis on machine learning modeling

AbstractAdditive manufacturing (AM) has emerged as a very promising technology for producing complex metallic components with enhanced design flexibility. However, the mechanical properties and fatigue behavior of AM metals differ significantly from conventionally manufactured materials, thereby presenting challenges in accurately predicting their fatigue life. This study provides a comprehensive overview of recent developments and future trends in fatigue life prediction of AM metals, with a particular emphasis on machine learning (ML) modeling techniques. This review recalls recent developments and achievements in fatigue characteristics of AM metals, ML‐based approaches for fatigue life prediction of AM metals, and non‐ML‐based methodologies for the same purpose. In particular, some commonly used regression and classification techniques for fatigue evaluation of AM metals are summarized and elaborated. The study intends to furnish researchers, engineers, and practitioners in the field of AM with a guidance for the accurate and efficient prediction of fatigue life in AM metal components.