Fault diagnosis of dynamic multivariate systems is a challenging problem. In this article, a novel fault diagnosis scheme based on variable-wise stacked temporal autoencoder (VW-STAE) is proposed.

The increasing automation in the design process of electrical machines for vehicles generates huge amounts of data, leading to a growing interest in using machine learning for faster predictions and ...

Here, in order to predict future bearing failures before they occur, we use an autoencoder, a form of neural network, to build a model that identifies abnormal vibrations from continuous sensor ...