A new cyclical generative adversarial network based data augmentation method for multiaxial fatigue life prediction

Sun, Xingyue; Zhou, Kun; Shi, Shouwen; Song, Kai; Chen, Xu

doi:10.1016/j.ijfatigue.2022.106996

Cited by 56 publications

(21 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…But there still remain some issues with the existing theoretical system for the prediction of multiaxial fatigue life. Most of the multiaxial life prediction models are based on empirical or semiempirical formulas due to the complexity of fatigue mechanism and the challenge in making complete explanation 3–6 . Especially, the process of fatigue damage accumulation can become more complex when the components are subjected to multiaxial non‐proportional loading 1 .…”

Section: Introductionmentioning

confidence: 99%

“…Yang et al 6 combined long short‐term memory (LSTM) network and fully connected neural network to predict the multiaxial fatigue life of materials. Sun et al 3 put forward a cyclical generative adversarial network model through which Fourier transformation and other semiempirical equations can be integrated to augment data following physical knowledge. Yang et al 13 established a new deep learning method, which incorporated an advanced deep learning mechanism called self‐attention mechanism to predict the multiaxial fatigue life of materials.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multiaxial fatigue life prediction for various metallic materials based on the hybrid CNN‐LSTM neural network

Heng

Gao

et al. 2023

Fatigue Fract Eng Mat Struct

View full text Add to dashboard Cite

A new algorithm optimization‐based hybrid neural network model is proposed in the present study for the multiaxial fatigue life prediction of various metallic materials. Firstly, a convolutional neural network (CNN) is applied to extract the in‐depth features from the loading sequence composed of the critical fatigue loading conditions. Meanwhile, the multiaxial historical loading information with time‐series features is retained. Then, a long short‐term memory (LSTM) network is adopted to capture the time‐series features and in‐depth features of the CNN output. Finally, a full connection layer is used to achieve dimensional transformation, which makes the fatigue life predictable. Herein, the hyperparameters of the LSTM network are automatically determined using the slime mold algorithm (SMA). The test results demonstrate that the proposed model has pleasant prediction performance and extrapolation capability, and it is suitable for the life prediction of various metallic materials under uniaxial, proportional multiaxial, nonproportional multiaxial loading conditions.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Multiaxial fatigue life prediction for various metallic materials based on the hybrid CNN‐LSTM neural network

Heng

Gao

et al. 2023

Fatigue Fract Eng Mat Struct

View full text Add to dashboard Cite

show abstract

“…This requires the study of new learning strategies to expand the sample size, effectively utilize prior knowledge, and achieve data augmentation with fewer labeled samples. Generative Adversarial Networks (GANs) have been used in dataset generation in recent years [12][13][14]. Therefore, in order to construct a reasonable network model, it is necessary to design effective learning strategies and make better use of prior knowledge to achieve effective recognition of human motions.…”

Section: Introductionmentioning

confidence: 99%

Human Motion Recognition in Small Sample Scenarios Based on Gan and CNN Models

Zhong¹,

Li²

2022

PIER M

View full text Add to dashboard Cite

In the research of radar-based human motion classification and recognition, the traditional manual feature extraction is more complicated, and the echo data set is generally smaller. In view of this problem, a method of human motion recognition in small sample scenarios based on Generative Adversarial Network (GAN) and Convolutional Neural Network (CNN) models is proposed. First, a 77 GHz millimeter wave radar data acquisition system is built to obtain echo data. Secondly, the collected human motion echo data is preprocessed, the micro-Doppler features are extracted, and the range Doppler map (RDM) is used to project the velocity dimension and accumulate the two-dimensional micro-Doppler time-frequency map data set of the human motion frame by frame. Finally, the deep convolution generative adversarial network (DCGAN) is constructed to achieve data augmentation of the sample set, and the CNN is constructed to realize automatic feature extraction to complete the classification recognition of different human motions. Experimental studies have shown that the combination of GAN and CNN can achieve effective recognition of daily human motions, and the recognition accuracy can reach 96.5%. Compared with the manual feature extraction, the recognition accuracy of CNNs is improved by 7.3%. Compared with the original data set, the system recognition accuracy based on the sample augmentation data set is improved by 2.17%, which shows that the GAN has an excellent performance in human motion recognition in small sample scenarios.

show abstract

“…Currently, most multiaxial fatigue life prediction models can accurately predict fatigue life under proportional loading (Brown and Miller, 1973; Sun et al. , 2022; Branco et al.…”

Section: Introductionmentioning

confidence: 99%

“…Currently, most multiaxial fatigue life prediction models can accurately predict fatigue life under proportional loading (Brown and Miller, 1973;Sun et al, 2022;Branco et al, 2022;Nourian-Avval and Khonsari, 2021;Guechichi et al, 2011), while the prediction results for fatigue life under non-proportional loading conditions are often not accurate enough (Ma and Liu, 2022). This is due to the constant rotation of the principal strain axis under non-proportional loading conditions, and the material does not easily produce a stable dislocation structure internally, leading to non-proportional additional hardening phenomena (Liu et al, 2020).…”

Section: Introductionmentioning

confidence: 99%

Critical plane-based fatigue life model under multiaxial random loading

Wang

Liu

Hua

et al. 2022

IJSI

View full text Add to dashboard Cite

PurposeEngineering components/structures are usually subjected to complex and variable loads, which result in random multiaxial stress/strain states. However, fatigue analysis methods under constant loads cannot be directly applied to fatigue life prediction analysis under random loads. Therefore, the purpose of this study is how to effectively evaluate fatigue life under multiaxial random loading.Design/methodology/approachFirst, the average phase difference is characterized as the ratio of the number of shear strain cycles to the number of normal strain cycles, and the new non-proportional additional hardening factor is proposed. Then, the determined random typical load spectrum is processed into a simple variable amplitude load spectrum, and the damage in each plane is calculated according to the multiaxial fatigue life prediction model and Miner theory. Meanwhile, the cumulative damage can be calculated separately by projection method. Finally, the maximum projected cumulative damage plane is defined as the critical plane of multiaxial random fatigue.FindingsThe fatigue life prediction capability of the method is verified based on test data of TC4 titanium alloy under random multiaxial loading. Most of the predicting results are within double scatter bands.Originality/valueThe objective of this study is to provide a reference for the determination of critical plane and non-proportional additional hardening factor under multiaxial random loading, and to promote the development of multiaxial fatigue from experimental studies to practical engineering applications.

show abstract

A new cyclical generative adversarial network based data augmentation method for multiaxial fatigue life prediction

Cited by 56 publications

References 51 publications

Multiaxial fatigue life prediction for various metallic materials based on the hybrid CNN‐LSTM neural network

Multiaxial fatigue life prediction for various metallic materials based on the hybrid CNN‐LSTM neural network

Human Motion Recognition in Small Sample Scenarios Based on Gan and CNN Models

Critical plane-based fatigue life model under multiaxial random loading

Contact Info

Product

Resources

About