Katayoun Eshkofti

I am a postdoctoral fellow in the Division of Decision and Control Systems at KTH Royal Institute of Technology, starting from June 2024. I am working under the supervision of Matthieu Barreau and Henrik Sandberg.

I hold a PhD degree in Industrial Engineering, with a primary focus on physics-informed neural networks (PINNs) and their applications to complex engineering problems, particularly those involving coupled partial differential equations (PDEs). Throughout my doctoral studies, under the supervision of Seyed Mahmoud Hosseini, I successfully applied PINNs and their variations to various industrial and mechanical problems, including thermoelasticity, non-Fickian diffusion-elastic and thermoelastic wave propagation, reliability assessment of complex systems, and the transient response of a porous half-space with strain and thermal relaxation factors.

In my current postdoctoral role, my research direction has shifted toward developing data-driven machine learning algorithms for predictive maintenance. Along this main path, research projects focused on physics-informed machine learning (ML) are being led.

The list of my publications can be found on Google Scholar. Moreover, I gave a talk on my latest research, titled "A gradient-enhanced physics-informed neural network (gPINN) scheme for the coupled non-Fickian/non-Fourierian diffusion-thermoelasticity analysis: A novel gPINN structure," at the CRUNCH seminar at Brown University, which can be found on their YouTube channel and website.