8 Zerstörungsfreie Prüfung
Filtern
Dokumenttyp
- Beitrag zu einem Tagungsband (12)
- Vortrag (12)
- Zeitschriftenartikel (10)
- Posterpräsentation (6)
- Buchkapitel (4)
Schlagworte
- Structural health monitoring (44) (entfernen)
Organisationseinheit der BAM
- 8 Zerstörungsfreie Prüfung (44)
- 8.6 Faseroptische Sensorik (20)
- 8.2 Zerstörungsfreie Prüfmethoden für das Bauwesen (16)
- 8.4 Akustische und elektromagnetische Verfahren (9)
- 3 Gefahrgutumschließungen; Energiespeicher (3)
- 3.0 Abteilungsleitung und andere (2)
- 8.1 Sensorik, mess- und prüftechnische Verfahren (2)
- 3.5 Sicherheit von Gasspeichern (1)
- VP Vizepräsident (1)
- VP.1 eScience (1)
Eingeladener Vortrag
- nein (12)
The deployment of machine learning (ML) and deep learning (DL) in structural health monitoring (SHM) faces multiple challenges. Foremost among these is the insufficient availability of extensive high-quality data sets essential for robust training. Within SHM, high-quality data is defined by its accuracy, relevance, and fidelity in representing real-world structural scenarios (pristine as well as damaged). Although methods like data augmentation and creating synthetic data can add to datasets, they frequently sacrifice the authenticity and true representation of the data. Sharing real-world data encapsulating true structural and anomalous scenarios offers promise. However, entities are often reluctant to share raw data, given the potential extraction of sensitive information, leading to trust issues among collaborating entities.
Our study introduces a novel methodology leveraging Federated Learning (FL) to navigate these challenges. Within the FL framework, models are trained in a decentralized manner across different entities, preserving data privacy. In our research, we simulated several scenarios and compared them to traditional local training methods. Employing guided wave (GW) datasets, we distributed the data among different parties (clients) using IID (independent, identically distributed or in other words, statistically identical) mini batches of dataset, as well as non-IID configurations. This approach mirrors real-world data distribution among varied entities, such as hydrogen refueling stations.
In our methodology, the initial round involves individualized training for each client using their unique datasets . Subsequently, the model parameters are sent to the FL server, where they are averaged to construct a global model. In the second round, this global model is disseminated back to the clients to aid in predictive tasks. This iterative process continues for several rounds to reach convergence.
Our findings distinctly highlight the advantages of FL over localized training, evidenced by a marked improvement in prediction accuracy . This research underscores the potential of FL in GW-based SHM, offering a remedy to similar challenges tied to data scarcity in other SHM approaches and paving the way for a new era of collaborative, data-centric monitoring systems.
The built infrastructure ages and requires regular inspection and, when in doubt, monitoring. To ensure that older concrete bridges showing signs of deterioration can be used safely, several innovative monitoring tools have been introduced, including but not limited to optical, fiber-optic, or acoustic emission techniques. However, there are gaps in the portfolio. A sensing technique that covers a wide range of damage scenarios and larger volumes, while still being sensitive and specific, would be beneficial.
For about 15 years, research has been conducted on ultrasonic monitoring of concrete structures that goes beyond the traditional ultrasonic pulse velocity test (PV test), mostly using a very sensitive data evaluation technique called coda wave interferometry. At BAM we have developed sensors and instrumentation specifically for this method.
We have instrumented a 70-year-old, severely damaged prestressed concrete bridge in Germany in addition to a commercial monitoring system. We have now collected data for almost 3 years. We can show that we can provide information about the stress distribution in the bridge. We have also been able to confirm that there has been no significant additional damage to the bridge since the installation.
The through-life management of our constantly ageing infrastructure is a basic requirement in order to ensure their structural safety and serviceability. Each structure experiences deterioration processes with time leading to a decrease of structural safety and serviceability. The design of new structures considers the expected deterioration for a defined period, the design service life. However, a frequent survey of structural safety controlling structural condition should be mandatory and a maintenance plan should be an integral part of the design. In addition, many structures have exceeded their design service life already or are very close to it leading to an increasing demand for condition assessment. On the one hand, assumptions made during design are not valid any more due to change of the loads, e.g., increasing traffic loads in terms of number and weights. On the other hand, design codes evolved over time in such a way that existing structures do not comply with today’s standards. In all these cases, the through-life management is an important tool to maintain the accessibility of existing structures with known reliability.
In line with the new Model Code for Concrete Structures, which includes guidance for both – design of new structures and assessment of existing structures, the Task Group 3.3 focused on the compilation of a state-of-the-art guideline for the through-life management of existing concrete structures, including:
Data acquisition by testing and monitoring techniques;
Condition assessment for the evaluation of existing structures;
Performance prediction using advanced methods;
Decision-making procedures to perform a complete assessment of existing structure.
The overall objective of the through-life management is the assessment of the current condition and the estimation of the remaining service life under consideration of all boundary conditions.
The digitalization of quality control processes and the underlying data infrastructures for safety relevant components, such as hydrogen pressure vessels, plays a significant role in the transition towards Industry 4.0. In the current safety regulations for hydrogen pressure vessels, there is no established concept for structural health monitoring. The development of a reliable structural health monitoring methodology for monitoring the structural integrity of pressure vessels enables a fast-forward transition from personnel- and costintensive recurring inspections, a.k.a. periodic maintenance, to predictive maintenance. In the work presented; we investigated the application of ultrasonic guided wave propagation to monitor and assess the condition of Type IV composite overwrapped pressure vessel (COPV). A sensor network of fifteen piezo-electric wafers is placed on the carbon fibre reinforced composite cylinder. Five different artificial damage configurations are created by gluing two different weight blocks on three different locations. The database containing measured guided wave data sets is enriched by two different boundary conditions. We utilized an open-source software, openBIS labnotebook, to store and analyse experimental datasets. The guided wave ultrasonic signals were investigated and analysed by using commonly used ultrasonic features (e.g., amplitude, frequency, time of flight) as well as non-traditional time-series features (kurtosis, skewness, variance). The features were used to calculate damage index and the detection performance for the results has been evaluated. The results suggest that both traditional and non-traditional features assume significant importance in artificial damage detection. The future works will additionally involve the impacts of operational conditions, such as periodic pressure variations temperature loadings as well as material degradations.
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.
We present a prototype of a distributed POF sensing system for strain detection in road construction and civil engineering ready to be used in practical applications. The system is based on a cost-efficient digital incoherent optical frequency domain reflectometry (I-OFDR). In this approach, the strain-induced changes in the Rayleigh backscattering profile along a perfluorinated polymer optical fiber (PF-POF) are determined by measuring the complex transfer function of the sensing fiber using a compact digital data acquisition unit. The digital unit replaces an oversized vector network analyzer (VNA) enabling suitability of the measurement system for out-door use and providing a significant reduction of the total system costs at the same time. The entire sensor con-cept includes the use of geosynthetics with integrated PF-POFs as two-dimensional sensor structures for geotech-nical applications. The robustness and the functionality of the distributed POF I-OFDR sensors were proven in the field. The field test involved embedding of two geomats, each with two polymer optical sensor loops, into the embankment of the federal road B 91 in a section near Leipzig. The measurements conducted after the sensor-based geomats had been installed confirmed both the potential of such smart geosynthetics and the functionali-ty of the digital I-OFDR prototype.