Deep domain adversarial method with central moment discrepancy for intelligent transfer fault diagnosis

Abstract

Big data condition monitoring in the industrial of internet era is indispensable, and intelligent fault diagnosis plays an important role in it. The adversarial learning method is widely used because of its ability to extract domain invariant features to solve the variable speed fault diagnosis problem. However, its training process is often unstable and difficult to converge to the optimal solution, which brings great challenges to the fault detection of equipment. In view of this exasperating problem, a novel model, called deep domain adversarial method with central moment discrepancy, is proposed. The presented model mainly consists of four modules: a shared weight feature extraction network with wide convolution kernel, a supervised classification network, an adversarial domain classification network, and a CMD alignment network. Adversarial domain classification network is employed to extract features that have both category distinction and domain invariance in the process of mutual game learning between features of source domain and target domain. The CMD alignment network can be devoted to align the higher-order moments of two domain features to constrain the instability in adversarial learning. Through the above regularization method, the model shows a relatively stable and higher accuracy of transferring diagnosis in the non-standardized data. The public test data set and the private data set are applied to validate the model. The results show that the proposed model successfully solves the problem of training instability in adversarial learning and has a relatively high diagnostic accuracy.

Big Data Intelligent Collection and Network Failure Analysis Based on Artificial Intelligence

“…After adjusting the sample weight, the sampling training process using the improved LM algorithm is shown in Figure 3. From Figure 4, when the unimproved neural network is trained 100 times, there is still a big gap from the error of 10 -2 , and the neural network combined by the SOM method and the LM method, convergence is reached after 20 trainings [52][53][54][55][56][57][58] . Using the above combination of SOM method and LM method, the training process is shown in Figures…”

Cross Network Representation Matching with Outliers

“…Two categories of matching mechanisms can be identified in these works. First is the statistic moment matching based mechanism such as MMD [17] [18], CMD (Central Moment Discrepancy) [19] and second-order statistic matching [20]. Second is the adversarial matching mechanism which enforces sample representations from different domain to be nondiscriminative by minimizing the adversarial loss.…”

A simulation-data-driven subdomain adaptation adversarial transfer learning network for rolling element bearing fault diagnosis

“…Wen et al [20] used a sparse autoencoder as the general feature extractor of the TL network, and the DA measurement criteria of the training data and the testing data used maximum mean discrepancy (MMD). In addition to MMD, CORrelation ALignment (CORAL) [21] and central moment discrepancy [22] are mathematical statistics for measuring discrepancy in domain distribution. Another DA method is adversarial learning.…”