28 July 2020 in IEEE/ASME Transactions on Mechatronics
It confronts great difficulty to apply traditional artificial intelligence (AI) techniques to machinery prognostics and health management (PHM) in manufacturing systems due to the lack of abnormal samples corresponding to different fault conditions. This paper explores an unsupervised feature learning method for machinery health monitoring by proposing a generative adversarial networks (GAN) model that exploits the merits of the autoencoder (AE) and the traditional GAN. The major contribution is that the data distribution of the normal samples is accurately learned by the GAN model within both the signal spectrum and latent representation spaces. Specifically, the discriminative feature for machinery health monitoring is learned in an unsupervised manner by the proposed method in three steps. First, the proposed GAN model is trained by the normal samples of the inspected machine with the aim to correctly reconstruct the signal spectrum and its latent representation. Then, the trained model is applied to test the online samples of the same machine with unknown health conditions. Finally, the dissimilarity between the tested samples and their reconstructed ones in the latent representation space is taken as the discriminative feature. The feature value will increase significantly if a fault occurs in the inspected machine because the abnormal samples are never trained in the proposed GAN model. Experimental studies on three different machines are conducted to validate the proposed method and its superiority over the traditional methods in detecting abnormal points and characterizing fault propagation.
1 June 2022 in IEEE Transactions on Instrumentation and Measurement
Confronting with the problems of inconsistent data distributions, noise interference, and class imbalance in fault diagnosis of train bearings under variable working conditions, this article proposes a cross-domain fault diagnosis model called dynamic balanced domain-adversarial networks (DBDANs). The major contributions of the proposed DBDAN are that a novel frequency band attention module (FBAM) with a clear physical interpretation is embedded in the feature extraction of the DBDAN to alleviate the noise interference, and an adaptive class-level weighting strategy is designed for the imbalanced dataset to tackle the class imbalance problem. Specifically, the FBAM highlights the time–frequency contents in the frequency bands containing state information by adaptively learning the time–frequency characteristics of bearing vibration signals. Then, high-level representations are extracted by the following feature extractor, which are used for fault recognition by the label classifier and alignments of marginal and conditional distributions across domains by the domain discriminator. The adaptive class-level weighting strategy is used to dynamically balance the distribution alignments between the imbalanced classes. The DBDAN model employs Wasserstein distance to measure the distribution discrepancy and is trained in an adversarial way. The experimental results of cross-domain fault diagnosis of a common rolling bearing and a train axle-box bearing under variable working conditions verify the superiority of the proposed method over other domain-adversarial methods in terms of diagnostic accuracy, convergence speed, and stability.
24 July 2023 in IEEE Transactions on Instrumentation and Measurement
The problem of imbalanced data amounts between the samples of healthy and faulty conditions degrades the performance of traditional intelligent diagnosis methods for rotating machinery. This paper explores a data generation approach to improve the diagnostic performance of key components in the rotating machinery. A new generative adversarial networks (GAN) model, called categorical feature GAN (CFGAN), is proposed to generate sufficient synthetic samples for the minority faulty conditions so as to rebalance the datasets. The main contribution is that the quality of the synthetic samples is ensured by considering both consistency and diversity with the real samples, which conduces to machinery fault diagnosis. Specifically, latent categorical features are firstly learned from the real fault samples via the proposed CFGAN, which is an integration model of autoencoder (AE) and auxiliary classifier GAN (ACGAN). Then, synthetic categorical features are produced by interpolation among the learned categorical features followed by slight noise addition. Finally, synthetic samples are generated using the synthetic categorical features through the generator of the CFGAN. In addition, Wasserstein distance (WD) is applied to the CFGAN to avoid mode collapse in training. Because categorical features rather than random noise are used to generate samples, and the interpolation is performed in the feature space instead of the noisy data space, the consistency of the synthetic samples is improved as compared to the traditional methods. The diversity is ensured by the use of diverse synthetic categorical features and the WD. Two machinery datasets under varying imbalance ratios confirmed that the proposed CFGAN model can synthesize high-quality samples and has better generalization capability over the existing data generation methods in machinery fault diagnosis.
18 January 2026 in Expert Systems with Applications