Tarik Sahin M.Sc.

Professur für Computergestützte Simulation im Bauingenieurwesen
Gebäude 41/100, Zimmer 5125
+49 89 6004 3788
tarik.sahin@unibw.de

Tarik Sahin M.Sc.

Werdegang

seit 02/2021 Wissentschaftlicher Mitarbeiter, Institut für Mathematik und Computergestützte Simulationen, Universität der Bundeswehr München
10/2017-06/2020

Master of Science,  Computational Engineering

Masterarbeit: Tetrahedral Mesh Refinement, eine Zusammenarbeit von MTU Aero Engines und dem Institut für Statik und Dynamik an der Ruhr-Universität Bochum

09/2018-10/2019

Wissenschaftliche Hilfskraft, Institut für Stahl-, Leicht- und Verbundbau, Ruhr Universität Bochum

05/2018-10/2019

Wissenschaftliche Hilfskraft, Institut für Statik und Dynamik, Ruhr Universität Bochum

10/2012-06/2017

Bachelor of Science, Luft- und Raumfahrttechnik, Middle East Technical University

Bachelorarbeit: Design, Analyse und Produktion eines senkrecht startenden und landenden Flugzeugs


Forschungsschwerpunkt

  • Hybrid Digital Twins
  • Physics-Informed Neural Networks (PINNs) for solving computational mechanics problems
  • Combination of classical simulation methods with Machine Learning and Deep Learning (Data-driven simulations)
  • Numerical methods, solvers, optimization

 

Use case 1: Hertz contact problem FEM vs PINNs

pinn_fem_hertz.png

Use case 2: Prediction of beam deflection via PINNs with a wall scenario. 

beam_2_y.gif


Preprints and Articles Submitted for Publication

  1. Sahin, T., Wolff, D., von Danwitz, M., & Popp, A. (2024). Towards a Hybrid Digital Twin: Physics-Informed Neural Networks as Surrogate Model of a Reinforced Concrete Beam. Preprint, submitted for publication arXiv web-logo.png

Conference Proceedings (with Peer-Review)

  1. von Danwitz M, Kochmann T, Sahin T, Wimmer J, Braml T, Popp A (2023): Hybrid Digital Twins: A Proof of Concept for Reinforced Concrete Beams. Proceedings in Applied Mathematics and Mechanics, 22(1): e202200146, DOI (Open Access) doi.png
  2. Milani, R., Sahin, T., von Danwitz, M., Moll, M., Popp, A., Pickl, S. (2023). Automatic concrete bridge crack detection from strain measurements: A preliminary study. In: Critical Information Infrastructures Security. CRITIS 2022. Lecture Notes in Computer Science, vol 13723, Hämmerli, B., Helmbrecht, U., Hommel, W., Kunczik, L., Pickl, S. (Eds) . Springer, Cham, Germany, DOI (Open Access) doi.png

Articles in Peer-Reviewed International Journals

  1. Sahin, T., von Danwitz, M.,  Popp A. (2024): Solving forward and inverse problems of contact mechanics using physics-informed neural networks. Adv. Model. and Simul. in Eng. Sci. 11, 11 (2024), DOI (Open Access) doi.png

International Conference Contributions with Abstract

  • Sahin, T.: Solving Forward and Inverse Problems of Contact Mechanics using Physics-Informed Neural Networks, 10th GACM Colloquium on Computational Mechanics 2023, Vienna, Austria, September 11 - 13, 2023.
  • Sahin T.: Physics-Informed Neural Networks With Hard Constraints for Solid and Contact Mechanics, Math 2 Product Conference, Taormina, Italy, May 30 - June 1, 2023. 
  • Sahin, T.: Solving Forward and Inverse Problems of Solid and Contact Mechanics using Physics-Informed Neural Networks, 9th GACM Colloquium on Computational Mechanics 2022, Essen, Germany, September 21 - 23, 2022.
  • Sahin, T., von Danwitz, M., Popp A.: Physics-Informed Neural Networks for Solid and Contact Mechanics, 92nd Annual Meeting of the International Association of Applied Mathematics and Mechanics, Aachen, Germany, August 15 - 19, 2022.
  • Sahin, T., Popp A.: Simulation-assisted deep learning approach for predicting the real contact area and the contact pressure, Mechanistic Machine Learning and Digital Twins for Computational Science, Engineering & Technology (MMLDT-CSET), San Diego, California, September 26 - 29, 2021.

Teaching

Courses & Exercises

Supervised Student Projects / Theses

  • Simulationsgestützte Machine Learning-Ansätze zur Vorhersage der maximalen Durchbiegung eines Trägers (2021)
  • Implementierung von Physik-Informierten Neuronalen Netzen in der Elastizitätstheorie unter Verwendung 2D-Rechengebieten (2022)
  • Optimierung der Hyperparameter für Physik-informierte Neuronale Netze (2023)