TY - JOUR A1 - Charmi, Amir A1 - Mustapha, Samir A1 - Yilmaz, Bengisu A1 - Heimann, Jan A1 - Prager, Jens T1 - A Machine Learning Based-Guided Wave Approach for Damage Detection and Assessment in Composite Overwrapped Pressure Vessels N2 - The applications of composite overwrapped pressure vessels (COPVs) in extreme conditions, such as storing hydrogen gases at very high pressure, impose new requirements related to the system's integrity and safety. The development of a structural health monitoring (SHM) system that allows for continuous monitoring of the COPVs provides rich information about the structural integrity of the component. Furthermore, the collected data can be used for different purposes such as increasing the periodic inspection intervals, providing a remaining lifetime prognosis, and also ensuring optimal operating conditions. Ultimately this information can be complementary to the development of the envisioned digital twin of the monitored COPVs. Guided waves (GWs) are preferred to be used in continuous SHM given their ability to travel in complex structures for long distances. However, obtained GW signals are complex and require advanced processing techniques. Machine learning (ML) is increasingly utilized as the main part of the processing pipeline to automatically detect anomalies in the system's integrity. Hence, in this study, we are scrutinizing the potential of using ML to provide continuous monitoring of COPVs based on ultrasonic GW data. Data is collected from a network of sensors consisting of fifteen Piezoelectric (PZT) wafers that were surface mounted on the COPV. Two ML algorithms are used in the automated evaluation procedure (i) a long short-term memory (LSTM) autoencoder for anomaly detection (defects/impact), and (ii) a convolutional neural network (CNN) model for feature extraction and classification of the artificial damage sizes and locations. Additional data augmentation steps are introduced such as modification and addition of random noise to original signals to enhance the model's robustness to uncertainties. Overall, it was shown that the ML algorithms used were able to detect and classify the simulated damage with high accuracy. T2 - 13th European Conference on Non-Destructive Testing (ECNDT) 2023 CY - Lisbon, Portugal DA - 03.07.2023 KW - Machine learning KW - Structural health monitoring KW - COPV KW - Guided waves KW - Damage localization PY - 2023 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-590789 DO - https://doi.org/10.58286/28079 SN - 2941-4989 VL - 1 IS - 1 SP - 1 EP - 6 PB - NDT.net CY - Mayen AN - OPUS4-59078 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - Charmi, Amir A1 - Heimann, Jan A1 - Duffner, Eric A1 - Hashemi, Seyedreza A1 - Prager, Jens T1 - Application of deep learning for structural health monitoring of a composite overwrapped pressure vessel undergoing cyclic loading N2 - Structural health monitoring (SHM) using ultrasonic-guided waves (UGWs) enables continuous monitoring of components with complex geometries and provides extensive information about their structural integrity and their overall condition. Composite overwrapped pressure vessels (COPVs) used for storing hydrogen gases at very high pressures are an example of a critical infrastructure that could benefit significantly from SHM. This can be used to increase the periodic inspection intervals, ensure safe operating conditions by early detection of anomalies, and ultimately estimate the remaining lifetime of COPVs. Therefore, in the digital quality infrastructure initiative (QI-Digital) in Germany, an SHM system is being developed for COPVs used in a hydrogen refueling station. In this study, the results of a lifetime fatigue test on a Type IV COPV subjected to many thousands of load cycles under different temperatures and pressures are presented to demonstrate the strengths and challenges associated with such an SHM system. During the cyclic testing up to the final material failure of the COPV, a sensor network of fifteen surface-mounted piezoelectric (PZT) wafers was used to collect the UGW data. However, the pressure variations, the aging process of the COPV, the environmental parameters, and possible damages simultaneously have an impact on the recorded signals. This issue and the lack of labeled data make signal processing and analysis even more demanding. Thus, in this study, semi-supervised, and unsupervised deep learning approaches are utilized to separate the influence of different variables on the UGW data with the final aim of detecting and localizing the damage before critical failure. T2 - 11th European Workshop on Structural Health Monitoring CY - Potsdam, Germany DA - 10.06.2024 KW - Structural health monitoring KW - Deep learning KW - Ultrasonic-guided waves KW - Anomaly detection KW - Damage localization PY - 2024 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-607396 DO - https://doi.org/10.58286/29751 SP - 1 EP - 8 AN - OPUS4-60739 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -