Condition monitoring of rotating shafts is essential for ensuring the reliability and optimal performance of machinery in diverse industries. In this context, as industrial systems become increasingly complex, the need for efficient data processing techniques is paramount. Deep learning has emerged as a dominant approach due to its capacity to capture intricate data patterns and relationships. However, a prevalent challenge lies in the black-box nature of many deep learning algorithms, which often operate without adhering to the underlying physical characteristics intrinsic to the studied phenomena. To address this limitation and enhance the fusion of data-driven methodologies with the fundamental physics of the system under study, this paper leverages physics-informed neural networks (PINNs). Specifically, a simple but realistic numerical case study of an extended Jeffcott rotor model, encompassing damping effects and anisotropic supports for a more comprehensive modelling, is considered. PINNs are used for the estimation of five parameters that characterize the health state of the system. These parameters encompass the radial and angular position of the static unbalance due to the disk installed on the shaft, the stiffness along the principal axes of elasticity, and the non-rotating damping coefficient. The estimation is conducted solely by exploiting the displacement signals from the centre of the disk and, to showcase the efficacy and precision provided by this novel methodology, various scenarios involving different constant rotational speeds are examined. Additionally, the impact of noisy input data is also taken into account within the analysis and the performance is compared to that of traditional optimization algorithms used for parameters estimation.

Damage diagnosis of thin-walled structures has been successfully performed through methods based on tomography and machine learning-driven methods. According to traditional approaches, diagnostic signals are excited and sensed on the structure through a permanently installed network of sensors and are processed to obtain information about the damage. Good performance characterizes methods that process Lamb waves, which are described by long propagation distances and high sensitivity to anomalies. Most of the methods require extracting damage-sensitive features from the diagnostic signals to drive the damage diagnosis task. However, this process can lead to loss of information, and the choice of the specific feature to extract may introduce biases that hamper damage diagnosis. Furthermore, traditional approaches do not perform well when composites are considered, due to the anisotropy, inhomogeneity, and complex damage mechanisms shown by this type of material. To boost the performance of methods for damage diagnosis of composite plates, this work proposes a convolutional neural network (CNN)-based algorithm that localizes damage by processing Lamb waves. Different from other methods, the proposed method does not require extracting features from the acquired signals and allows localizing damage through the regression approach. The method was tested against experimental observations of Lamb waves propagating in two composite panels and in a hybrid panel, and the performance of two different sensor arrays was investigated. The pseudo-damage approach was used to generate large enough datasets for training the CNNs, and the performance of the framework was evaluated by localizing pseudo-damage and real damage determined by low-velocity impacts. The CNN-driven method accurately localized damage in all the considered scenarios, and it also outperformed traditional damage indices-based approaches, such as the reconstruction algorithm for probabilistic inspection of defects.