Thermal image, or thermogram, becomes a new type of signal for machine condition monitoring and fault diagnosis due to the capability to display real-time temperature distribution and possibility to indicate the machine's operating condition through its temperature. In this paper, an investigation of using the second-order statistical features of thermogram in association with minimum redundancy maximum relevance (mRMR) feature selection and simplified fuzzy ARTMAP (SFAM) classification is conducted for rotating machinery fault diagnosis. The thermograms of different machine conditions are firstly preprocessed for improving the image contrast, removing noise, and cropping to obtain the regions of interest (ROIs). Then, an enhanced algorithm based on bi-dimensional empirical mode decomposition is implemented to further increase the quality of ROIs before the second-order statistical features are extracted from their gray-level co-occurrence matrix (GLCM). The highly relevant features to the machine condition are selected from the total feature set by mRMR and are fed into SFAM to accomplish the fault diagnosis. In order to verify this investigation, the thermograms acquired from different conditions of a fault simulator including normal, misalignment, faulty bearing, and mass unbalance are used. This investigation also provides a comparative study of SFAM and other traditional methods such as back-propagation and probabilistic neural networks. The results show that the second-order statistical features used in this framework can provide a plausible accuracy in fault diagnosis of rotating machinery.

Many industrial structures associated with railway infrastructures rely on a large number of bolted joint connections to ensure safe and reliable operation of the track and trackside furniture. Significant sums of money are spent annually to repair the damage caused by bolt failures and to maintain the integrity of bolted structures. In the U.K., Network Rail (the organization responsible for rail network maintenance and safety) conducts corrective and preventive maintenance manually on 26,000 sets of points (each having approximately 30 bolted joints per set), in order to ensure operational success and safety for the travelling public. Such manual maintenance is costly, disruptive, unreliable and prone to human error. The aim of this work is to provide a means of automatically measuring the clamping force of each individual bolted joint, by means of an instrumented washer. This paper describes the development of a sensor means to be used in the washer, which satisfies the following criteria. 1. Sense changes in the clamping force of the joint and report this fact. 2. Provide compatibility with the large dynamic range of clamping force. 3. Satisfy the limitations in terms of physical size. 4. Provide the means to electronically interface with the washer. 5. Provide a means of powering the washer in situ. 6. Provide a solution at an acceptable cost. Specifically the paper focuses on requirements 1, 2 and 3 and presents the results that support further development of the proposed design and the realization of a pre-prototype system. In the paper, various options for the force sensing element (strain gage, capacitor, piezoresistive) have been compared, using design optimization techniques. As a result of the evaluation, piezo-resistive sensors in concert with a proprietary force attenuation method, have been found to offer the best performance and cost trade-off The performance of the novel clamping force sensor has been evaluated experimentally and the results show that a smart washer can be developed to monitor the condition of bolted joints as found on railway track and points.

Exposure to particulate material (PM) is a major health concern in megacities across the world which use trains as a primary public transport. PM emissions caused by railway traffic have hardly been investigated in the past, due to their obviously minor influence on the atmospheric air quality compared to automotive transport. However, the electrical train releases particles mainly originate from wear of rails track, brakes, wheels and carbon contact stripe which are the main causes of cardio-pulmonary and lung cancer. In previous reports most of the researchers have focused on case studies based PM emission investigation. However, the PM emission measured in this way doesn't show separately the metal PM emission to the environment. In this study a generic PM emission model is developed using rail wheel-track wear model to quantify and characterise the metal emissions. The modelling has based on Archard's wear model. The prediction models estimated the passenger train of one set emits 6.6mg/km-train at 60m/s speed. The effects of train speed on the PM emission has been also investigated and resulted in when the train speed increase the metal PM emission decrease. Using the model the metal PM emission has been studied for the train line between Leeds and Manchester to show potential emissions produced each day. This PM emission characteristics can be used to monitor the brakes, the wheels and the rail tracks conditions in future.

No abstract