Identifying Chassis Damper Degradation with Physics-Informed Machine Learning (ML) Algorithms

Machine Learning (ML) focuses on developing algorithms that enable computers to learn from data in order to make predictions or decisions. In this context, learning refers to the process of iteratively adjusting the parameters of a deep neural network to satisfy an optimality criterion best possibly. These networks are able to approximate complex, non-linear mappings between input and output data that may not be easily captured by traditional physics-based models.

Degraded chassis dampers strongly influence vehicle safety and ride comfort but often occur after several years of operation. However, in regular workshop checks the slow and continuing degradation is hardly identified. Hence, the replacement of degraded chassis dampers can be significantly delayed. To allow for timely vehicle maintenance, onboard diagnosis methods can be used that rely on the actual degree of degradation extracted from vehicle sensor signals.

To gain knowledge from physical models regarding the degradation effects of chassis dampers, a multi-body dynamics simulation is applied, where chassis damper degradation is mapped from real-world measurements, including detailed test bench measurements of degraded dampers and suspension-related data. Based on this simulation, a fundamental understanding of degradation effects is obtained, and meaningful features are derived to capture these effects in measurement signals. These features are then extracted from measurement data and ML-algorithms are trained to detect the degradation status.

Project goals

• Combine knowledge from physical models and real-world measurements.

• Study on a ML-algorithm to identify the damper degradation status.

Figure 1: Process chain for incorporating physics into machine learning

References

Preparation of paper in progress

Researcher

• Lorenz Ott

Project Partner

• Cariad