Mobile Menu

Summer Seminar Series: Azarakhsh Jalalvand

Summer Seminar Series: Azarakhsh Jalalvand

Date: June 17, 2021

Time: 11:00 a.m. - 12:00 p.m.

Location: Virtual over Zoom

New Light: Rising Stars in Energy and the Environment

Azarakhsh Jalalvand

Real-Time Remote Sensing and Fusion Plasma Control: A Reservoir Computing Approach

Thursday, June 17, 2021

Jalalvand is a senior data scientist at Ghent University-imec in Belgium and has been a visiting postdoc at the Plasma Control Group working with Prof. Egemen Kolemen at Princeton University since 2020.

Jalalvand obtained his doctoral degree in 2015 in artificial intelligence and continued his career as a permanent postdoctoral researcher at IDLab-UGent-imec. Jalalvand has contributed to over 20 national and global projects, leveraging fundamental and applied research in diverse data analysis directions, such as pathological speech/image processing, radar signal processing, bioinformatics, anomaly detection, and predictive maintenance.

In 2020, he was awarded a 3-year special postdoctoral fellowship at UGent-BOF to investigate data-driven models for condition monitoring and plasma control in magnetic confinement devices to produce controlled thermonuclear fusion power.

Nuclear fusion power is a potential source of safe, non-carbon-emitting and virtually limitless energy. The tokamak is a promising approach to fusion based on magnetic plasma confinement, constituting a complex physical system with many control challenges. However, plasma instabilities pose an existential threat to a reactor, which has not yet been solved. Since current physical understanding is not sufficiently advanced to reliably predict instabilities, a way forward is artificial intelligence and data-driven models. In this presentation, we discuss the application of such models in predicting the plasma profiles on confinement time scales using experimental data from DIII-D tokamak. The conducted experiments demonstrate that simple yet effective machin- learning models, namely reservoir computing networks (RCN), achieve comparable results to state-of-the-art deep convolutional neural networks and long short-term memory (LSTM) models, with significantly easier and faster training procedure. This superiority allows for fast and frequent adaptation of the model to new situations, such as changing the environment conditions or predicting plasma profiles on a different fusion device.

About the New Light Series
New Light: Rising Stars in Energy and the Environment is a summer webinar series to spotlight associate research scholars, postdoctoral research fellows, and other early-career researchers affiliated with the Andlinger Center for Energy and the Environment. Weekly webinars feature a diverse range of researchers working on cutting-edge topics across disciplines who seek to solve society’s most pressing problems in energy and the environment. View the full line-up