The ettc2022 is the important international platform for telemetry, test instrumentation, telecontrol and data processing and is taking place for the fifth time in parallel and in cooperation with the SENSOR+TEST in Nuremberg.
Martin Bussas gives a lecture about predictive maintenance with machine learning.
13:30 – 15:30: 9. Machine Learning & Artificial Intelligence, Room 1, Hall 2
9.1 Vehicle Data Analysing System – VeDAS
13:30: 1. Martin Bussas, TROUT GmbH, Kassel (Germany)
Predictive maintenance with machine learning is a new challenge for the automotive industry. Depending on the level of vehicle use and the load on components beyond predefined limits, the detection and elimination of systematic faults and the introduction of demand-oriented maintenance helps increase efficiency and reduce costs.
With this in mind, TROUT has developed VeDAS, a self-sufficient system that can be easily adapted to different vehicles. It is used for the automatic acquisition of vehicle data, which is evaluated with machine learning. Fast and secure communication to the downstream evaluation system is ensured via mobile storage media or wireless communication.
Collected data includes position, acceleration and vibration of the vehicle, as well as speed and distance travelled. In order to take possible environmental influences into account, temperature and humidity are also determined. A structure-borne sound microphone provides information about the operating status of the monitored vehicle. Further data provided by an engine control unit can be accessed via CAN bus interface.
In addition to vehicle data acquisition, VeDAS also provides a logbook function for documenting maintenance activities. Maintenance intervals and deadlines are determined for all entered vehicle assemblies.
The selected method of condition monitoring specifies maintenance intervals and ensures the availability of the vehicle. Expanded functions include determining when an engine oil change is due.