Post-doctoral researcher Digital Twins for the plasma edge in tokamaks (X/F/M)

State-of-the-art simulations of transport effects in the fusion plasma edge has become an import element in the prediction of reactor-scale operational scenarios providing compatibility to both, required heat and particle exhaust constraints and good fusion plasma core performance. Given the multi-scale multi-physics nature of the problem, solutions for a reactor relevant operational regime are hard to achieve given slow numerical convergence rates. An employment of fast numerical tools, that e.g., allow to predict relevant dynamics for plasma control, or allow a full simulation of a discharge by using an integrated approach with suitable fidelity, are not mature yet. In very recent years, the fusion community has started to develop fast surrogate models based on Machine Learning / AI models to speed up significantly the employed tools. Such tools have demonstrated to be generally applicable and to be fast; their predictive capability however is a trade-off balancing speed and level of fidelity.

DIFFER is seeking to hire a post-doctoral researcher for 2 years to work with the Integrated Modeling group at the DIFFER institute in support of EUROfusion projects on AI in fusion exhaust. The work encompasses conduction of plasma edge simulations for reactor design in the EUROfusion DEMO central team for a fusion power plant (FPP) and to develop improvements of fast edge transport model surrogates SOLPS-NN applicable to future fusion reactors such as DEMO. In addition, new hire will collaborate on an additional project related to EUROfusion developments of a digital twin environment (DTE), also in collaboration with EUROfusion public research institutes and DIFFER affiliated private partners.

Responsibilities and tasks:

• Employment of state-of-the-art tools to run plasma edge simulations using SOLPS-ITER to generate data for machine-learning based surrogates such as SOLPS-NN
• Continued development of ML/AI based fast surrogate models at increased fidelity
• Contribution to continued developments of dynamical ML/AI based fast surrogate models
• Deploy the resulting improved and fast models for future fusion power plant design (DEMO, VNS, etc) including uncertainty quantification (UQ). Support integration of fast edge surrogate models in integrated simulations (pulse design simulators)
• Contribution to the development of a Digital Twin Environment (DTE) framework

Required skills:

• Good knowledge of ML/AI based techniques to develop fast surrogates (deep neural networks) and capability to develop own efficient model learning schemes (deep learning techniques, representation learning, active learning, etc)
• Desired: understanding of fusion edge plasmas transport physics
• Desired: Experience with transport solvers of CFD-type
• (optional) Experience with (dynamic) simulations in the plasma edge using edge codes (SOLPS-ITER, EDGE2D, etc)

This position is for 1 FTE, will be for a period of 2 years and is graded in pay scale 10. The position will be based at DIFFER (www.differ.nl) and the working location will be at TU Eindhoven. When fulfilling a position at DIFFER, you will have an employee status at NWO. You can participate in all the employee benefits NWO offers. We have a number of regulations that support employees in finding a good work-life balance. At DIFFER we believe that a workforce diverse in gender, age and cultural background is key to performing excellent research. We therefore strongly encourage everyone to apply. More information on working at NWO can be found at the NWO website (https://www.nwo-i.nl/en/working-at-nwo-i/jobsatnwoi/)

DIFFER: Science for future energy

At the Dutch Institute for Fundamental Energy Research (DIFFER) we work on a future in which clean energy will be available to everybody, anywhere in the world. DIFFER’s mission is to perform leading fundamental research on materials, processes, and systems for a global sustainable energy infrastructure.

Our research focuses on two major energy themes: fusion energy as a clean, safe and sustainable energy source and chemical energy. We work in close partnership with (inter)national academia and industry. DIFFER is one of the ten research institutes of the Dutch Research Council (NWO).

Within our institute physicists, chemists, engineers, and other specialists work together in multidisciplinary teams to accelerate the transition to a sustainable society. DIFFER’s workforce is currently composed of ~160 scientists (of which 60 guests and interns), supported by ~40 technicians and ~4 0 support staff members.

The global nature of the energy challenge is apparent from the international representation of our employees, who originate from over 30 different countries. To strengthen our commitment to diversity, we formed a task force to design, implement, and monitor diversity and gender equality initiatives.