Hybrid Approaches for Deep Learning-Based Structural Health Monitoring

Description of the project

We are currently soliciting applications for the 2024 MSCA Postdoctoral Fellowship scheme in the field of deep learning-based structural health monitoring.

Structural health monitoring relies on sensor data, which can be interpreted either through model-based methods that compare measured responses with predictions from a physical structural model (see e.g. Rocchetta et al., 2018), or through data-driven methods that analyze patterns and anomalies in the sensor data using statistical or machine learning techniques (see e.g. Manzini et al. 2022).

In this work, the focus will be on hybrid approaches that combine physical models and data analysis, particularly by leveraging the structure’s vibration characteristics to extract relevant features from the sensor data before further processing

The study will focus on bridges instrumented with accelerometers, capturing vibrations induced by vehicles crossing the structure to assess its dynamic response (Raedersdorff et al., 2024). Since the bridge is monitored under real-world traffic conditions, we have little control over the ‘experimental’ parameters, such as vehicle weight, speed, or spacing, which are largely unknown and cannot be prescribed. These uncontrolled and variable conditions highlight the need for hybrid approaches, where vibration-based features extracted from the sensor data are combined with structural models to reliably interpret the bridge’s response.

References

Manzini, N. et al. (2022). An Automated Machine Learning-Based Approach for Structural Novelty Detection Based on SHM. In: Pellegrino, C., Faleschini, F., Zanini, M.A., Matos, J.C., Casas, J.R., Strauss, A. (eds) Proceedings of the 1st Conference of the European Association on Quality Control of Bridges and Structures. EUROSTRUCT 2021. Lecture Notes in Civil Engineering, vol 200. Springer, Cham. https://doi.org/10.1007/978-3-030-91877-4_134

Raedersdorff, J., Chateauneuf, A., Clair, D., & Bouchaïr, A. (2024). Field assessment of a vibration-based damage identification procedure. In J. S. Jensen, D. M. Frangopol, & J. W. Schmidt (Eds.), Bridge Maintenance, Safety, Management, Digitalization and Sustainability (1st ed.). CRC Press/Balkema. https://doi.org/10.1201/9781003483755

Rocchetta, R., Broggi, M., Huchet, Q., & Patelli, E. (2018). On-line Bayesian model updating for structural health monitoring. Mechanical Systems and Signal Processing, 103, 174–195. https://doi.org/10.1016/j.ymssp.2017.10.015

About the university and the unit

The Université Clermont Auvergne (UCA) is a higher education institution located in Clermont-Ferrand, France. It offers a wide range of programs across various academic fields, from fundamental sciences to humanities and social sciences, as well as law, economics, engineering, and health sciences. UCA is home to approximately 38,000 students and 1,300 academic staff.

The postdoctoral fellow will be hosted at the Institut Pascal, an institute of engineering and applied science, working with the Mechanics, Materials, and Structures team. The team includes about 50 permanent staff and 50 non-permanent members (PhD students, post-docs, etc.). A 3,200 square-meter experimental facility is available and might be used for experimental validation of the methods developed during the project.

Application

Expressions of interest must be sent to Pierre Beaurepaire (pierre.beaurepaire@sigma-clermont.fr by February 28th.

If you are interested in pursuing this project, please send your CV and a short motivation letter (maximum 1 page) detailing your research area, motivation for the project, project ideas, and other relevant information.

The application will need to be prepared collaboratively between the candidate and the host institution. Ensure that your schedule allows sufficient time to work on it during the summer months.

Eligibility

Applicants must:

·         have a PhD degree at the time of the deadline for applications (11th September 2024)

·         have a maximum of 8-year experience in research

·         not have resided or carried out their main activity (work, studies, etc.) in France for more than 12 months in the 3 years before September 2026.