Roller bearings are one of the most safety-critical components in many machines. Predicting the vibration-based remaining useful life (RUL) of roller bearings allows operators to make informed maintenance decisions and to guarantee reliability and safety. The health indices (HIs) for degradation assessment are constructed by extracting feature information from the collected data, which significantly influences the prognosis result. Conventional HI construction methods rely heavily on expert knowledge and also have limited capacity for learning health information from the raw data from roller bearings. Furthermore, outlier regions often occur in HIs developed by those methods, and these can easily result in false alarms. To address these problems, a novel HI construction method based on a deep multilayer perceptron (MLP) convolution neural network (DMLPCNN) model, which also considers outlier regions, is proposed in this paper. In the proposed model, a 1-D MLP convolution (Mlpconv) block, consisting of a convolution layer and a micro network, is applied to learn features directly from vibrational data. The learned features are then mapped into an HI using a global average pooling layer and a logistic regression layer. Finally, an outlier region correction method, based on sliding thresholds, is proposed to detect and remove outliers in the HI. The outlier region correction method is able to enhance the interpretability of the constructed HI. The effectiveness of the proposed method is verified using whole-life datasets of 17 bearings. The experimental results demonstrate that the proposed method outperforms conventional methods.

Railway wheelsets consist of three main components; the wheel, axle and axle bearing. Faults can develop on any of the aforementioned components, but the most common are related to wheel and axle bearing damages. The continuous increase in train operating speeds means that failure of an axle bearing can lead to very serious derailments, potentially causing human casualties, severe disruption in the operation of the network, damage to the tracks, unnecessary costs, and loss of confidence in rail transport by the general public. The rail industry has focused on the improvement of maintenance and online condition monitoring of rolling stock to reduce the probability of failure as much as possible. This paper discusses the results of onboard acoustic emission measurements carried out on freight wagons with artificially damaged axle bearings in Long Marston, UK. Acoustic emission signal envelope analysis has been applied as a means of effective tool to detect and evaluate the damage in the bearings considered in this study. From the results obtained it is safe to conclude that acoustic emission signal envelope analysis has the capability of detecting and evaluating faulty axle bearings along with their characteristic defect frequencies in the real-world conditions

Defects in railway axle bearings can affect operational efficiency, or cause in-service failures, damaging the track and train. Healthy bearings produce a certain level of vibration and noise, but a bearing with a defect causes substantial changes in the vibration and noise levels. It is possible to detect the bearing defects at an early stage of their development, allowing an operator to repair the damage before it becomes serious. When a vehicle is scheduled for maintenance, or due for overhaul, knowledge of bearing damage and severity is beneficial, resulting in fewer operational problems and optimised fleet availability. This paper is a review of the state of the art in condition monitoring systems for rolling element bearings, especially the axlebox bearings. This includes exploring the sensing technologies, summarising the main signal processing methods and condition monitoring techniques, i.e. wayside and on-board. Examples of commercially available systems and outputs of current research work are presented. The effectiveness of the current monitoring technologies is assessed and the p–f curve is presented. It is concluded that the research and practical tests on axlebox bearing monitoring are limited compared to the generic bearing applications.