Skip to main content

Computer simulations in inertial confinement fusion - my part in their downfall

Dr. Christopher Brady

School of Engineering, Warwick

Inertial confinement fusion (ICF) offers a possible route to fusion power by using high power lasers to collapse pellets of deuterium-tritium fusion fuel to densities fifty times greater than lead. If everything goes according to plan this compressed fuel will spontaneously undergo thermonuclear ignition, where the initiation of a thermonuclear reaction at one point in the fuel will lead to the propagation of a fusion burn wave that will consume the remaining fuel. The main experiment working in the area is the US National Ignition Facility (NIF), a $4 billion facility at the Lawrence Livermore National Laboratory, which was created as part of the National Ignition Campaign : a mixed program of experiments, theory and simulations.

The simulation requirements for ICF work are stringent : collapsing geometry with a 50 fold decrease in radius, Mach 15 shocks, highly non-linear equations where normally tolerable numerical errors can lead to large changes in the final solution and finally the requirement that the code scale to thousands of processors. There is a well accepted "gold standard" in simulation techniques used by most researchers in the field, including most of those involved in the NIC: the Arbitrary Lagrangian Eulerian (ALE) radiation hydrodynamic code. The only problem is that these codes are large, complex and generally unavailable to the academic community due to their military origins. At Warwick I have been involved in the UK's first major attempt to develop one of these codes intended for civilian use in the ICF community : the ODIN project. In this talk I will detail ICF, the challenges involved in simulating it and how ODIN and other ALE codes attempt to address these challenges.