Speaker
Description
Dry friction is commonly present in mechanical structures through joints and interfaces, introducing nonlinear phenomena such as stick–slip transitions and partial locking. While friction is often regarded as a source of damping, recent studies have shown that it can significantly influence vibration-based damage detection, sometimes enhancing damage signatures and sometimes masking them.
Recent research has also explored data-driven strategies to cope with the limited availability of labeled data in structural health monitoring. In particular, nonlinear structural behavior can be exploited to generate richer datasets through multi-excitation testing procedures, enabling physics-informed data augmentation for deep-learning-based fault diagnosis.
Despite these advances, the physical mechanisms by which frictional nonlinearities modify the observability of structural damage signatures are still not fully understood. Experimental investigations on multi-degree-of-freedom systems have shown that increasing the excitation level may either amplify or suppress the spectral signatures associated with structural defects.
This work proposes a complementary interpretation of these phenomena based on a reduced-order dynamic representation of the structure. In particular, the response of the system under different friction regimes is analyzed in terms of dominant modal subspaces and equivalent low-order models. Within this framework, friction-induced fully stuck configurations can be interpreted as a projection of the system dynamics onto reduced modal spaces, which may alter the observability of damage-related features in the frequency response.
The proposed perspective is illustrated through vibration tests performed on a railway pantograph structure, characterized by multiple frictional joints and complex modal interactions. The results provide insight into the mechanisms through which friction modifies the detectability of structural damage and suggest possible directions for improving vibration-based structural health monitoring techniques in nonlinear mechanical systems.
SANTAMATO, Giancarlo, et al. Leveraging systems’ non-linearity to tackle the scarcity of data in the design of intelligent fault diagnosis systems. Nonlinear Dynamics, 2024, 112.18: 16153-16166.