Suspension dampers degrade over time. They operate under a wide range of load conditions and environmental influences, causing changes in damping characteristics that may affect vehicle safety and ride comfort. Accurate simulation models that capture these changing characteristics are essential for degradation effect analysis, yet corresponding research remains limited in the literature.This article proposes a simulation model for degraded suspension dampers based on the Universal Differential Equations framework. The modelling approach starts with an equivalent mechanical model, covering known physical effects, which is then enhanced through the integration of Neural Networks into its system dynamics. Physical consistency constraints are enforced by auxiliary losses throughout training.Test bench measurements of functional and degraded dampers reveal that oil loss introduces strongly nonlinear, transient and asymmetric changes to damper dynamics. These effects become increasingly pronounced at higher excitation frequencies and smaller stroke amplitudes. The newly developed Neural Equivalent Mechanical Model was trained and validated using the test bench data and was shown to effectively capture the dynamics induced by degradation. Further interpretation of the learned neural functions reveals insights into how oil loss influences the dynamics captured by the developed damper model.
CC BY 4.0
