The key challenge of intelligent fault diagnosis is to develop features that can distinguish different categories. Because of the unique properties of mechanical data, predetermined features based on prior knowledge are usually used as inputs for fault classification. However, proper selection of features often requires expertise knowledge and becomes more difficult and time consuming when volume of data increases. In this paper, a novel deep learning network (LiftingNet) is proposed to learn features adaptively from raw mechanical data without prior knowledge. Inspired by convolutional neural network and second generation wavelet transform, the LiftingNet is constructed to classify mechanical data even though inputs contain considerable noise and randomness. The LiftingNet consists of split layer, predict layer, update layer, pooling layer, and full-connection layer. Different kernel sizes are allowed in convolutional layers to improve learning ability. As a multilayer neural network, deep features are learned from shallow ones to represent complex structures in raw data. Feasibility and effectiveness of the LiftingNet is validated by two motor bearing datasets. Results show that the proposed method could achieve layerwise feature learning and successfully classify mechanical data even with different rotating speed and under the influence of random noise.

Intelligent fault diagnosis of machines has long been a research hotspot and has achieved fruitful results. However, intelligent fault diagnosis is a difficult issue in the case of small sample due to the lack of fault signals. To solve this problem, a small sample focused intelligent fault diagnosis method via multi-modules gradient penalized generative adversarial networks (MGPGAN) is proposed. The proposed method consists of three network modules: Generator, Discriminator, and Classifier. By adversarial training, Generator can generate mechanical signals in different health conditions. Because of the high similarity to the signals obtained in practice, the generated signals can also be used as training data so that the limited training dataset of the proposed method is expanded. Classifier has a strong ability to extract fault features from raw mechanical signals, and then classify different health conditions. The experimental results on two bearing vibration datasets indicate that the proposed method can not only generate bearing vibration signals but also obtain fairly high fault classification accuracy under the small sample condition.