Title Participants Abstract "Adaptive Health Monitoring and Fault Prognostics of Rotating Machinery: Statistical Model-based Remaining Useful Life Prediction" "Junyu Qi" "Effective health monitoring and fault prognostics are crucial for ensuring the reliable operation of rotating machinery. Challenges arise in detecting faults under varying operating conditions and accurately predicting Remaining Useful Life (RUL). Statistical Model (SM)-based fault prognostics is advantageous in the application due to the fast computation and less complexity of implementation. However, several challenges still exist in the state-of-the-art studies: (i) how to construct high-quality prognostics HI, (ii) how to select the proper statistical model, (iii) how to choose the effective state estimator, (iv) how to decide the start prediction time (SPT) and (v) how to set the proper failure threshold. Moreover, (vi) the good quality endurance data are very limited in the real-world, either in the academic or industrial areas. The data scarcity has hindered the development of fault prognostics and the innovation of how to solve the abovementioned obstacles. The primary goal of this PhD research is to develop an advanced SM-based fault prognostics methodology, providing solutions to the abovementioned open questions. In this research, the following objectives are expected to achieve: 1. Developing high-quality prognostics HIs that outperform the classic HIs such as the RMS, kurtosis, spectral kurtosis, peak to peak. The developed HIs are easy to implement without requiring explicit information of the operating condition and of the rotating machinery. Thus, they can be transferrable to different components of rotating machinery such as bearings and gears, etc. 2. Developing the intelligent selection of the SPT for fault prognostics instead of the manual selection. The intelligent SPT detection method is robust to random spikes of the data series. 3. Choosing a statistical model that can be suitable for a set of machineries. The developed strategy can treat failure patterns for not only one bearing but also for other bearings with complex degradation trajectories. 4. Finding the most effective estimator, which guarantees the robustness, high accuracy, and computation efficiency in the practical implementation. 5. Setting the failure threshold as close as possible to the real failure amplitude of the testing HI. More accurately set of the failure threshold can improve the accuracy of the estimated RUL. 6. Designing an experimental setup for degrading bearings. The customized setup allows constant & varying speeds and loads and degradation based on overload and overspeed." "Anomaly detection and multi-step estimation based remaining useful life prediction for rolling element bearings" "Junyu Qi, Rui Zhu, Alexandre Miguel Ricardo Mauricio, Konstantinos Gryllias" "Domain Knowledge Regularised Fault Detection" "Douw Marx, Konstantinos Gryllias" "Rotating machinery speed extraction through smartphone video acquisition from a radial viewpoint" "Toby Verwimp, Alexandre Miguel Ricardo Mauricio, Konstantinos Gryllias" "This article investigates the viability of using a smartphone as a cost-effective tool for rotational speed extraction. The proposed method exploits the geometrical image deformations induced by a smartphone’s rolling shutter camera, which reads pixel lines sequentially with a delay equal to the rolling shutter period. This characteristic period is the basis for the developed methodology and allows measuring constant and varying speeds above the Nyquist limit related to the camera’s frame rate. With respect to the rotating shaft, the smartphone is pointed to the side surface of the shaft (radial viewpoint). A method is proposed to measure the speed from the deformation of a zebra pattern as a consequence of the sequential readout of a rolling shutter camera. A mathematical model, which also takes into account a possible misalignment between the shaft and the camera’s rolling shutter reading direction, is developed and validated on an in-house test rig. It is shown that the proposed method is robust against changing lighting conditions and a misalignment between the shaft and the smartphone camera. Normalized Root-Mean-Square Errors (NRMSE) of about 3% (or less) are reached for the smartphone aligned with the shaft with peaks in the percentage error of about 5% to 8%. Nevertheless, even for a misalignment of 60°, the NRMSE is only 4% with variations in the percentage error to about 10%. The proposed method shows promising results with room for improvements, making it possible to develop a smartphone application to measure a shaft’s rotational speed in the future." "An anomalous frequency band identification method utilising available healthy historical data for gearbox fault detection" "Konstantinos Gryllias" "GLDAN: Global and Local Domain Adaptation Network for Cross-Wind Turbine Fault Diagnosis" "Dandan Peng, Wim Desmet, Konstantinos Gryllias" "Predicting pitting severity in gearboxes under unseen operating conditions and fault severities using convolutional neural networks with power spectral density inputs" "Rik Vaerenberg, Ali Hosseinli, Rui Zhu, Konstantinos Gryllias" "The PHM North America 2023 Data Challenge tasked participants to diagnose the pitting fault severity of a gearbox from a three-channel vibration signal. This work summarizes the authors’ proposed diagnostics solution which consists of a convolutional neural network with an ordinal loss criterion, trained on the power spectral density of the signal. This method is selected based on a rigorous evaluation using three dedicated validation sets, designed to evaluate the model’s ability to generalize to unseen operation conditions and fault severities. Ultimately, the proposed approach achieved a competition validation score of 282.2 and a test score of 213.3." "Differentiable Short-Time Fourier Transform Window Length Selection Driven by Cyclo-Stationarity" "Douw Marx, Konstantinos Gryllias" "Automated operational modal analysis for the monitoring of a wind turbine blade" "André Guilherme Massano Tavares, Wim Desmet, Konstantinos Gryllias" "Cavitation detection based on instantaneous angular speed" "Toby Verwimp, George Mousmoulis, Konstantinos Gryllias" "The study and detection of cavitation in hydraulic turbo-machinery has gained recently particular attention from both manufacturers and end users, due to the damage caused on machines’ surfaces that may result in costly sudden shutdowns. The traditional cavitation detection approaches exploit the noise and vibration signals that sense the vapor bubble implosion pressure waves, while advanced demodulation techniques are used in order to locate the resonances excited from these waves. This work aims to investigate an alternative methodology, where detection is achieved by processing the speed signal of rotary encoders. The motivation is that, contrary to microphones and accelerometers, the encoder is mounted directly on the shaft, where the rotor is connected and thus, is in direct contact with the cavitation waves. More specifically, in this study a centrifugal pump is tested under healthy and cavitating conditions, while an incremental encoder is rigidly installed on the pump’s shaft. The encoder’s signal is processed in order to accurately estimate the raw speed signal. In addition, the order spectrum is used so as to extract the diagnostic indicators based on integer order amplitudes. From the results it can be concluded that particular components can give useful insights allowing the detection of both heavy and light intensity cavities."