Detection of Tram Wheel Faults Using MEMS-Based Sensors.

Area of Science:

• Engineering
• Mechanical Engineering
• Condition Monitoring

Background:

• Condition-based maintenance (CBM) relies on sensor data to monitor machine health.
• Micro-electromechanical-systems (MEMS) sensors offer a viable solution for CBM applications.
• Identifying faulty tram wheels is crucial for operational safety and efficiency in tram depots.

Purpose of the Study:

• To develop and evaluate a method for detecting faulty tram wheels using MEMS-based vibration sensors.
• To establish a data processing technique for identifying defective wheels at entry or dispatch gates.
• To optimize sensor data analysis for accurate fault detection.

Main Methods:

• Utilizing Maximal Overlap Discrete Wavelet Packet Transform (MODWPT) for vibration signal decomposition.
• Assessing vibration energy within specific frequency bands to characterize wheel condition.
• Formulating a discrete optimization problem to maximize the distinction between normal and faulty wheel signatures.

Main Results:

• Faulty tram wheels demonstrate significantly higher vibration energy compared to normal wheels.
• The developed method effectively distinguishes between normal and faulty wheel vibrations.
• Simulation and experimental data validate the proposed approach.

Conclusions:

• MEMS-based sensors combined with MODWPT analysis provide an effective means for tram wheel fault detection.
• The proposed method enhances condition-based maintenance strategies for tram fleets.
• Early identification of faulty wheels improves safety and reduces operational disruptions.