| Team Name |
Nemosys |
| Track |
Industry (Public Sector) |
| Use Case |
AI for tires of rubber-tired metro maintenance performance support |
| Technology |
Data Mining and Machine Learning |
| Region |
AP (South Korea) |
| Team lead |
@alfon |
| Team members |
@youngjinchoi @jyyoo @Jongwoon_Kim |
Our team wants to utilize our capability in RAMS and to improve the prediction through SAS Data Mining and Machine Learning. We would like to consider the sensor data from the tire pressure monitoring system (TPMS) in the tires of the rubber-tired metros.
Metro maintenance is performed periodically based on the predefined schedule since it is important to consider it simultaneously with the metro service operation schedule and with the other metro units. Thus, it is different from private cars which can be repaired at any time. Also, if the tire failure does not affect the service operation or safety, it is better to delay the maintenance at failure until the daily operation finished or, if possible, until the next predefined maintenance schedule.
The model will consider the sensor data from TPMS, such as the tire pressure, temperature, and vibration, plus the train speed, the total train operating time (or distance in kilometer), and several others as the input. The target measures are the tire tread depth and crack/puncture presence. For every round-trip operation, the model is analyzed and the metro maintenance necessity is predicted. Remember that one tire failure may not affect the service failure or safety and all tires in one train must be considered simultaneously in the corrective maintenance performance decision-making. Thus, our team use case is the round-trip predictive maintenance of tires of the rubber-tired metro.
