Data augment method for machine fault diagnosis using conditional generative adversarial networks

Wang, Jinrui; Han, Baokun; Bao, Huaiqian; Wang, Mingyan; Chu, Zhenyun; Shen, Yuwei

doi:10.1177/0954407020923258

Cited by 28 publications

(15 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to reduce the effect of noise on the feature extraction ability of MC1DCNN, AMVMD was combined with MC1DCNN and applied to multi-channel signal fault diagnosis of rolling mill multi-row bearings. Considering the problem that fault data is difficult to obtain and the networks could not achieve good diagnostic accuracy under the condition of unbalanced dataset [ 34 ], a Deep Convolutional Generative Adversarial Network (DCGAN) was embedded in the model training process. Additionally, thanks to the excellent signal processing effect of AMVMD, it can effectively reduce the invalid feature information and noise interference in the signal and improve the dataset enhancement capability of DCGAN.…”

Section: Introductionmentioning

confidence: 99%

Fault Diagnosis Method for Rolling Mill Multi Row Bearings Based on AMVMD-MC1DCNN under Unbalanced Dataset

Zhao

Sun

Lin

et al. 2021

Sensors

View full text Add to dashboard Cite

Rolling mill multi-row bearings are subjected to axial loads, which cause damage of rolling elements and cages, so the axial vibration signal contains rich fault character information. The vertical shock caused by the failure is weakened because multiple rows of bearings are subjected to radial forces together. Considering the special characters of rolling mill bearing vibration signals, a fault diagnosis method combining Adaptive Multivariate Variational Mode Decomposition (AMVMD) and Multi-channel One-dimensional Convolution Neural Network (MC1DCNN) is proposed to improve the diagnosis accuracy. Additionally, Deep Convolutional Generative Adversarial Network (DCGAN) is embedded in models to solve the problem of fault data scarcity. DCGAN is used to generate AMVMD reconstruction data to supplement the unbalanced dataset, and the MC1DCNN model is trained by the dataset to diagnose the real data. The proposed method is compared with a variety of diagnostic models, and the experimental results show that the method can effectively improve the diagnosis accuracy of rolling mill multi-row bearing under unbalanced dataset conditions. It is an important guide to the current problem of insufficient data and low diagnosis accuracy faced in the fault diagnosis of multi-row bearings such as rolling mills.

show abstract

Section: Introductionmentioning

confidence: 99%

Fault Diagnosis Method for Rolling Mill Multi Row Bearings Based on AMVMD-MC1DCNN under Unbalanced Dataset

Zhao

Sun

Lin

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…In the past few years, breakthroughs in machine learning have provided new insight for the development of robust design strategies, especially in the areas of mechanics, materials and structures [5][6][7][8]. Predictive mechanical models are Technical Editor: Monica Carvalho.…”

Section: Introductionmentioning

confidence: 99%

Rapid mechanical evaluation of the engine hood based on machine learning

Chen

Cheng

Zhang

et al. 2021

J Braz. Soc. Mech. Sci. Eng.

View full text Add to dashboard Cite

With the development of machine learning and data mining, rapid design, rapid verification and rapid manufacturing have become the mainstream in the machinery industry. In this paper, the mapping function between the mechanical properties of the hood and its 11 dimensional parameters was mined using machine learning algorithms. By combining XGBoost and LightGBM algorithms with the bagging method, we proposed a hybrid model with hyperparameters optimized by the grey wolf algorithm. Subsequently, several machine learning models were trained and tested on a dataset of 6959 simulation samples, and the proposed hybrid model was found to have excellent predictive performance for the torsional stiffness ( R 2 of 0.9969 and root-mean-square error of 1.32185) and first-order modal frequency (0.9977 and 0.00989) of the hood. Moreover, the SHAP method (Shapley additive explanations) was used as a machine learning interpretation method to explain the predictive process of the mechanical performance. The results show that SHAP has great potential in model interpretation. This paper aims to develop a mathematical model of the mechanical properties of the hood, which can quickly predict the mechanical properties based on each key dimensional parameter. Therefore, engineers and designers can apply this approximate model in their design space exploration algorithms directly without training extra low-dimensional surrogate models.

show abstract

“…[2][3][4] Thus, the data-driven methods that use machine learning have been the research focus. 5,6 The data-driven methods are mainly made up of three stages: data acquisition, feature extraction, and fault recognition. Goyal et al 7 presented a method by training a support vector machine (SVM) model for bearing fault diagnosis.…”

Section: Introductionmentioning

confidence: 99%

Construction of a deep sparse filtering network for rotating machinery fault diagnosis

Cheng

Zou

Wang

et al. 2021

Proceedings of the Institution of Mechanical Engineers, Part D:

View full text Add to dashboard Cite

Deep neural networks (DNNs) have shown potential in intelligent fault diagnosis of rotating machinery. However, traditional DNNs such as the back-propagation neural network are highly sensitive to the initial weights and easily fall into the local optimum, which restricts the feature learning capability and diagnostic performance. To overcome the above problems, a deep sparse filtering network (DSFN) constructed by stacked sparse filtering is developed in this paper and applied to fault diagnosis. The developed DSFN is pre-trained by sparse filtering in an unsupervised way. The back-propagation algorithm is employed to optimize the DSFN after pre-training. Then, the DSFN-based intelligent fault diagnosis method is validated using two experiments. The results show that pre-training with sparse filtering and fine-tuning can help the DSFN search for the optimal network parameters, and the DSFN can learn discriminative features adaptively from rotating machinery datasets. Compared with classical methods, the developed diagnostic method can diagnose rotating machinery faults with higher accuracy using fewer training samples.

show abstract

Data augment method for machine fault diagnosis using conditional generative adversarial networks

Cited by 28 publications

References 21 publications

Fault Diagnosis Method for Rolling Mill Multi Row Bearings Based on AMVMD-MC1DCNN under Unbalanced Dataset

Fault Diagnosis Method for Rolling Mill Multi Row Bearings Based on AMVMD-MC1DCNN under Unbalanced Dataset

Rapid mechanical evaluation of the engine hood based on machine learning

Construction of a deep sparse filtering network for rotating machinery fault diagnosis

Contact Info

Product

Resources

About