Since 1986 - Covering the Fastest Computers in the World and the People Who Run Them

Language Flags
June 4, 2013

hpc4energy Incubator Touts Successes with Energy Companies

Alex Woodie

HPC can save energy companies enormous amounts of time in the development of new products and technologies compared to the traditional methods, according to hpc4energy.org, an incubator project at the Lawrence Livermore National Laboratory.

According to the infographic below, HPC analysis is saving companies like General Electric, Robert Bosch, and the New England Independent System Operator (IS) thousands of hours by using supercomputers to develop new products and technologies.

By parallelizing the positive sequence load flow (PSLF) software calculations and running it on a supercomputer instead of a desktop PC, the GE Energy Consulting group was able to drop the time required to run the analysis from 23.5 days to 23 minutes. That’s a savings of 33,817 minutes on the calculation that’s used to model how large blocks of electricity can safely and securely flow across a power grid.

Bosch also saw big time savings on calculations it was running to model how a gasoline engine can switch from traditional spark ignition (SI) to homogeneous charge compression ignition (HCCI), which car makers are considering adopting as a way to deliver Diesel-engine like efficiencies in gasoline engines. It used to take Bosch 14 days to model 10 cycles in the engine as it transitions SI to HCCI modes. Running on Sierra, Lawrence Livermore’s 165-terraflop supercomputer, the company was able to model the transition in just 4.5 days.

Meanwhile, the New England ISO used the lab’s HPC resources on a project that is comparing two different ways of performing unit commitment (UC) calculations to match expected demand with generation sources on an electric grid. As the New England ISO seeks more energy from renewable sources, it’s grappling with how to meet the demands of its 6.5 million customers while dealing with the lower reliability rates of renewable energy sources. By parallelizing its UC modeling software and running it on a 1,600 core system, the company was able complete modeling in 90 minutes, compared to 33.5 days using a desktop.

Lawrence Livermore and its HPCInnovationCenter group started the hpc4energy project in 2011 with the goal of pairing national labs with a select group of energy companies to demonstrate how HPC resources can help the energy companies with their technology and product development, and to boost American competitiveness in the energy field. The lab put out a request for proposals, and in 2012 picked six of the 30 submitted proposals to execute, at no cost to the participants.

Related Articles

NASA Builds ‘Climate in a Box’

HPC and the True Cost of Cloud

Exascale Advocates Stand on Nuclear Stockpiles