This internship will be a part of StructMonitor project, which aims to perform non-destructive testing of concrete structures using full wave inversion. FWI involves solving the wave equation many times for each sensor location; and updating the velocity model. To make it more efficient, we are interested in solving this using PINNs. They have emerged as an attractive mesh free solver for PDEs; however they have their own limitations such as spectral bias; time error propagation and slow to train. We look to investigate these problems and make them more efficient.
Funding:
Erasmus + grant available depending on eligibility criteria of your home university
Keywords:
PINNs, NDT, Operator learning, sciml, scientific machine learning, pde, applied math, civil engineering, strucutural health monitoring, fwi, full wave inversion.
About us:
The chair is focusing on fundamental and applied research in fields of geomechanics and geotechnics. In particular the following fields are under active Research: Numerical Method Development in Geomechanics, Energy Geotechnics & Climate Change Adaption and Soil Design / Soil and Geostructure Improvement
Training PINNs, reading and understanding state-of-the-art papers; writing code for those methods; performing analysis of failure modes; and producing peer-reviewed papers.
Writing efficient code in python. Solving real world problems related to SHM and NDT. Have access to experimental data, which will enable them to clean and use them for training models. Understand different PINN variants; and their limitations.
Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Education and Culture Executive Agency (EACEA). Neither the European Union nor EACEA can be held responsible for them.
