Peter Jansson: used nuclear fuel decay heat

The storage of used nuclear fuel is a multifaceted problem. One issue is that spent nuclear fuel can heat itself up from the inside. A normal heat sensor only measures the temperature of the outside of the fuel, and could miss a critical situation developing in the interior. A fast and accurate method for measuring the temperature on the inside of a used nuclear fuel element has been developed, and computer simulations were a crucial element of this method.

Peter Jansson is currently a researcher in the Nuclear Fission Diagnostics and Safeguards group in the division of Applied Nuclear Physics in the department of Physics and Astronomy. Besides their work on used nuclear fuel, the members of this group develop, among other things, monitoring systems for 4th generation nuclear reactors and techniques for ensuring that nuclear material is used in accordance with international non-proliferation treaties.

Experimental setup with a gamma radiation source
at the top left and detector on bottom right.

The new used nuclear fuel heat measurement technique is an indirect method that relies on measurements of gamma radiation at a detector placed outside the fuel (see image to the right). A large set of Monte Carlo simulations were performed to build a model tying internally generated gamma radiation to gamma ray measurements. The Monte Carlo approach captures a more complete physics than alternative computational methods, but require a vast amount of CPU power in order to provide a clear solution. With the simulations completed in December 2016, it is now possible use gamma-ray sensors to quickly make accurate temperature measurements of used nuclear fuel.

Peter Jansson had a SNIC Medium project at UPPMAX and NSC, and the calculations were divided between the system Tintin and Triolith. Due to excellent use of the bonus queue, the total runtime was far greater than the allocation — 5.5 million core-hours over a period of one year, 4 million of which were at UPPMAX. Two articles are expected to be published, as well as practical implementation of the method by SKB.