Jelila Yohanis Dabesa, Pamuła Wiesław
Department of Transport Systems, Traffic Engineering and Logistics, Faculty of Transport and Aviation Engineering, Silesian University of Technology, 40-019 Katowice, Poland.
Faculty of Mechanical Engineering, Jimma Institute of Technology, Jimma University, Jimma P.O. Box 378, Ethiopia.
Sensors (Basel). 2022 Aug 24;22(17):6373. doi: 10.3390/s22176373.
Micro-electromechanical-systems (MEMS) based sensors are used for monitoring the state of machines in condition-based maintenance tasks. This approach is applied at tram depots for the purpose of identifying faulty wheels on trams in order to eliminate defective trams at the entry or dispatch gates. The application of MEMS-based sensors for the detection of wheel faults is the focus of this study. A method for processing of the collected sensor data is developed. It is based on assessing the energy of vibrations at different frequency bands. Maximal Overlap Discrete Wavelet Packet Transform (MODWPT) is used for obtaining a description of the sensor data. The task of finding the energy threshold for detecting faulty wheels, frequency band and parameters of MODWPT which most distinctly distinguish the wheels is the goal of the method. The weighted difference (DW) between the extreme values of energy in a frequency band for normal and faulty wheels is proposed as the measure of the ability to distinguish the wheels. The search for the solution is formulated as a discrete optimisation problem of maximising this measure. Both the simulation and experimental results indicate that faulty wheels have greater vibration energy than normal wheels. The properties of this approach are discussed and evaluated.
基于微机电系统（MEMS）的传感器用于在基于状态的维护任务中监测机器状态。这种方法应用于电车车厂，目的是识别电车上的故障车轮，以便在电车进出大门时消除有缺陷的电车。基于MEMS的传感器在车轮故障检测中的应用是本研究的重点。开发了一种处理收集到的传感器数据的方法。该方法基于评估不同频带的振动能量。最大重叠离散小波包变换（MODWPT）用于获取传感器数据的描述。找到检测故障车轮的能量阈值、最能明显区分车轮的频带和MODWPT参数是该方法的目标。提出了正常车轮和故障车轮在一个频带内能量极值之间的加权差（DW）作为区分车轮能力的度量。将寻找解决方案的问题表述为最大化该度量的离散优化问题。仿真和实验结果均表明，故障车轮比正常车轮具有更大的振动能量。对该方法的特性进行了讨论和评估。
