Research | Infusion - Nuclear Fusion UGent

At the UGent department of Applied Physics, the research unit Nuclear Fusion (infusion) has been involved in fusion research since over forty years. Growing out of the long-standing expertise of the group in fusion diagnostics and experimentation on various fusion devices across Europe, the research presently concentrates on the rich area of fusion data science. We employ modern techniques from Bayesian inference and machine learning to study the fusion plasma and the technology of fusion devices. This combines two of the most topical and challenging issues of our time: sustainable energy supply and data science.

Research topics

Advanced regression analysis for estimation of fusion scaling laws

Scaling laws explain dependencies of key plasma parameters and contribute to the design of new fusion devices. They are estimated from complex databases by means of sophisticated statistical techniques. Our group plays a leading role in this domain.

Probabilistic characterization of stochastic plasma phenomena

We use probabilistic methods for characterizing stochasticity in fusion plasmas, like plasma instabilities and fluctuations. The challenge is to determine the plasma properties and machine design parameters that influence the corresponding probability distributions, reflecting the underlying physics.

Bayesian modeling of fusion diagnostics

We use Bayesian probabilistic inference for integrated analysis of data from multiple sensors (diagnostics) in fusion devices. This type of "data fusion" allows more reliable measurement of plasma conditions with uncertainty estimates.

Anomaly detection and predictive maintenance

Continuous real-time monitoring of the condition of a fusion power plant will be extremely important. We use machine learning techniques to monitor various components in fusion devices and for predicting possible failures.

Europe is at the forefront of developing one of the most promising long-term energy options: fusion power.

Get involved !