US shovels cash into supercomputers hoping to stoke fusion future

Fusion is hot right now - so hot that the US Department of Energy is dumping another $112 million into a dozen supercomputing projects to advance progress on further clean energy breakthroughs. 

The Scientific Discovery through Advanced Computing (SciDAC) program combines the DoE's existing Fusion Energy Sciences (FES) and Advanced Scientific Computing Research (ASCR) programs with the goal of solving complex fusion energy problems using supercomputing resources, including exascale systems. 

"The modeling and simulation work of these partnerships will offer insight into the multitude of physical processes that plasmas experience under extreme conditions and will also guide the design of fusion pilot plants," said DoE Associate Director of Science for FES, Jean Paul Allain. 

The DoE's call for proposals [PDF] on the project focuses on whole-facility modeling as part of the fusion pilot plant program, which meshes with the Biden administration's goal to build such a facility within the next decade.

• Lawrence Livermore lab repeats fusion breakthrough – yep, still kinda works
• Beyond the hype, AI promises leg up for scientific research
• UK's dream of fusion power by 2040s will need GPUs
• Brits and Yanks join forces to make fusion magnets cool again

Projects funded [PDF] through the SciDAC program will include modeling plasmas in fusion reactions, studying turbulence in reactors and using AI to predict and solve energy loss problems. Other programs will simulate the use of stellarators for nuclear fusion, as well as perform more simulations of inertial confinement systems like those used at the US' Lawrence Livermore National Laboratory (LLNL), which late last year made the world's first fusion ignition in a nuclear reactor. 

There has been significant progress with breakthroughs at the National Ignition Facility (NIF) at LLNL in December wherein scientists produced 3.15 megajoules of fusion energy, exceeding the 2.05 megajoules that were required to kick-start a fusion reaction. A week ago, LLNL scientists duplicated their results with an even higher energy yield, but don't get too excited: There's still a long way to go to achieve actual energy positivity. 

At its December ignition experiment, LLNL may have delivered just over 2 megajoules into the fuel capsule, but it took 322 megajoules to fire the lasers themselves - in other words, there's a lot of work needed to get to the point where fusion is actually a net-positive energy producer. 

The DoE will likely be at the forefront of much of that work, at least financially: The $112m awarded yesterday isn't the first fusion computing award to be handed out by the Department. 

Shortly after news broke of the initial fusion breakthrough, the DoE announced $33 million in funds for researchers willing to apply machine learning and AI to the analysis of fusion experiments being conducted by DoE-funded laboratories like LLNL. 

Like this most recent round of funding, the earlier, smaller round was also focused on advancing the fusion pilot plant initiative outlined by the Biden administration last year. ®