FaultFace: Deep Convolutional Generative Adversarial Network (DCGAN) based Ball-Bearing failure detection method

Failure detection is employed in the industry to improve system performance and reduce costs due unexpected malfunction events. So, a good dataset of desirable for designing an automated failure system. However, industrial process datasets are unbalanced contain little information about behavior uniqueness these events high cost running just get undesired behaviors. For this reason, performing correct training validation methods challenging. This paper proposes methodology called FaultFace on Ball-Bearing joints rotational shafts using deep learning techniques create balanced datasets. The uses 2D representations vibration signals denominated faceportraits obtained by time–frequency transformation techniques. From faceportraits, Deep Convolutional Generative Adversarial Network produce new nominal behaviors dataset. A Neural trained fault employing compared with other evaluate its Obtained results show that has