NERSCPowering Scientific Discovery for 50 Years

Harnessing Star Power: NERSC Aids Decades of Fusion Energy Science

September 25, 2014

Contact: Kathy Kincade, +1 510 495 2124,


2003: Researchers ran the NIMROD code on supercomputers at NERSC to create this 3D numerical simulation of plasma particle drift orbits in a tokamak reactor.

For decades, nuclear fusion has held the promise of a "clean” energy resource that could help wean the world from fossil fuels. However the challenges of harnessing the conditions inside the sun in an Earth-based nuclear fusion reactor have been many, with much of the progress over the last several decades coming in incremental advances.

In 1951, the Atomic Energy Commission – predecessor to the US Department of Energy (DOE) – initiated a fusion energy research program to investigate the feasibility of using a controlled fusion reaction to generate energy. That program began what we now know today as the DOE’s Office of Fusion Energy Sciences (FES), which has supported decades of research designed to produce a safe and sustainable energy resource based on fusion.

In a fusion reaction, energy is released when two hydrogen isotopes are fused to form a heavier nucleus, helium. To achieve high enough reaction rates to make fusion a useful energy source, hydrogen contained inside the reactor core must be heated to extremely high temperatures – more than 100 million degrees Celsius – which transforms it into hot plasma. The most common reactor design at present is the tokamak, which uses toroidal (doughnut-shaped) vessels and extremely strong magnetic fields to confine the plasma as it is heated up.

During the 1950s and 1960s, DOE-supported fusion research focused on gaining new knowledge of radiation and its effects and assessing the environmental impact of nuclear technology. In the early 1970s, these efforts gained a boost from a surprising source: the Arab oil embargo, which prompted a wave of energy-related national policy initiatives in the US and dramatic increases in funding for energy research – fusion and otherwise. It also paved the way for the National Energy Research Scientific Computing Center (NERSC), established in 1974 as the Controlled Thermonuclear Research Computer Center (CTRCC) and later renamed the National Magnetic Fusion Energy Computer Center (NMFECC) before becoming NERSC.

“I believe the ability of the members of the fusion and plasma physics communities to communicate with each other over the network will prove to be more valuable than the ability to work collectively on plasma transport codes in tokamaks,” said Alvin Trivelpiece, a key player in the creation of the center, in a 1999 reflection on the history of NERSC.

Building a Scientific Foundation

With global climate change concerns growing, the ability to produce a reliable, carbon-free energy resource has taken on new urgency. The DOE and FES continue to support research into the fundamental understanding of confined plasma, the Holy Grail of fusion energy, and the key elements needed to achieve a sustainable fusion energy reactor.

Even renowned physicist Steven Hawking is a fan: “I would like nuclear fusion to become a practical power source,” he told TIME magazine in 2010. “It would provide an inexhaustible energy supply, without pollution or global warming.”

Four decades of fusion energy research at NERSC have helped the plasma physics community make significant strides toward this goal. FES-supported research at NERSC ranges from advancing the fundamental science of magnetically confined plasmas and exploring the feasibility of inertial confinement to increasing fundamental understanding of basic plasma science to enhance economic competitiveness and create opportunities for a broader range of science-based applications. In 2013, FES-supported fusion research accounted for more than 360 million compute hours at the center 2013, second only to the Office of Basic Energy Sciences.

2009: Long-time NERSC user Linda Sugiyama of MIT used an extended magnetohydrodynamics code to create this simulation of a new class of nonlinear plasma instability: edge localized modes.

In addition to fusion energy, advances in plasma science have led to applications such as plasma processing of semiconductors and computer chips, material hardening for industrial and biological uses, waste management techniques, lighting and plasma displays, space propulsion, and non-contact infection-free surgical scalpels. Particle accelerators and free electron lasers also rely on plasma science concepts.

Some key findings in plasma research have come out of FES-sponsored research at NERSC over the past decade:

2012: Researchers from Princeton Plasma Physics Laboratory (PPPL) used the SciDAC XGC1 gyrokinetic code on Hopper to model relevant multi-scale physics over the entire plasma volume of the DIII-D fusion reactor. Findings point to the plasma edge as a source of rotation and strongly support the model of turbulence driven intrinsic torque as the origin of intrinsic rotation. (Credit: C.S. Chang, PPPL)

Understanding Plasma Behavior

Stephen Jardin, a physicist at Princeton Plasma Physics Laboratory, has been utilizing the center’s supercomputing resources for fusion energy research since the mid-70s – virtually his entire career.

“Even as I have taken on more teaching and management responsibilities, I still find time to do computational physics research and use cycles at NERSC,” he said. “One of the reasons it is so exciting for me is that I can remember back when we were in the megaflops era and couldn’t even imagine doing the kind of problems we are now routinely computing. Going from megaflops to gigaflops to teraflops to petaflops and beyond has truly revolutionized our research.”

Jardin’s current research focuses on tokamak magnetohydrodynamics (MHD), one of the simplest computer-based representations of plasma. MHD equations are like fluid equations that describe water and air but also include terms for the magnetic and electric fields present in the magnetized plasma in a fusion experiment.

“In the mega and gigaflop days, we would never try and model an actual fusion experiment – it was far beyond our capability,” Jardin said. “We would use a simplified set of equations (reduced MHD) that only had two scalar fields: the magnetic flux and the velocity stream function. We would construct idealized geometries with a particular symmetry so we would only have a 2D problem. We could not use the actual temperatures and densities in the experiment but would use values easier to model with our coarse grids. Our goal was to qualitatively identify instabilities in these idealized configurations and to argue that these same types of instabilities were occurring in the actual experiment.

1996: Stephen Jardin (left) and Horst Simon, Berkeley Lab Deputy Director, at the grand opening of the NERSC/ESnet facility at Berkeley Lab. (Credit: Roy Kaltschmidt, Berkeley Lab)

Fast forward to 2014, with computers that are a million or more times faster and codes that can rapidly run and re-run hundreds and thousands of virtual experiments on plasma behavior in the time it would take to construct just one physical experiment at a fraction of the cost and with far more detailed information about the results.

"We now have full 3D computer models, we can keep eight scalar fields to provide a much more realistic representation of the plasma, and we can often have enough resolution to use the same values of temperatures and densities that are present in the experiment,” Jardin said. “We now do detailed comparisons of our calculations with experimental measurements. We can do long-time simulations of a configuration that is initially stable but then develops an instability and then eventually watch it saturate. We now have the ability to do predictive modeling of experiments that have not been performed or even built.”

Looking ahead, even with new experimental fusion reactors now in place or under construction, fusion is expected to remain heavily dependent on numerical simulations because experimental time on large experiments such as ITER—an international fusion research project being built in Cadarache, France, that is designed to produce magnetically confined burning plasmas for the first time – will be limited. As these larger facilities are developed and constructed, simulations can provide insights into the device’s different modes of operation and enhance the development of optimization and control strategies.

“NERSC has made an enormous impact in our field,” Jardin said. “Not to say that our models are perfect today, but they are much more realistic than they were back in pre-NERSC days.”

About NERSC and Berkeley Lab
The National Energy Research Scientific Computing Center (NERSC) is a U.S. Department of Energy Office of Science User Facility that serves as the primary high performance computing center for scientific research sponsored by the Office of Science. Located at Lawrence Berkeley National Laboratory, NERSC serves almost 10,000 scientists at national laboratories and universities researching a wide range of problems in climate, fusion energy, materials science, physics, chemistry, computational biology, and other disciplines. Berkeley Lab is a DOE national laboratory located in Berkeley, California. It conducts unclassified scientific research and is managed by the University of California for the U.S. Department of Energy. »Learn more about computing sciences at Berkeley Lab.