Suite of open-source, HPC-powered physics codes allows engineers to do everything but collect their mail inside a virtual wind power plant
This article is part of the
Fall 2020 R&D Newsletter
In This Issue
Current Research & Development
ExaWind is a confidence builder.
Until now, predicting the performance of entire wind power plants was a near-impossible task—from anticipating the various movements of turbine blades to understanding how varying wind conditions and blade motion can impact wind power plant operational efficiency. Designing and installing wind power plants is an expensive and time-consuming venture, and most current modeling and engineering tools are just not up to the task of tackling the intricacies of wind power plant flow dynamics, particularly in offshore environments.
Driven by the immense power of high-performance computing (HPC), ExaWind is an open-source suite of physics codes and libraries that enables multifidelity simulation of wind turbines and wind power plants.
These groundbreaking simulation capabilities allow engineers to do everything but collect their mail inside a virtual wind power plant, creating a computer-generated environment where they can test their designs in real time and move forward in product development with confidence—minimizing industry risk and ensuring optimized performance down the road.
Prediction Powered by Petaflops and Exaflops
The U.S. Department of Energy (DOE) Exascale Computing Project (ECP), ExaWind, and the Wind Energy Technologies Office (WETO) High-Fidelity Modeling project were launched in 2016, products of a close collaboration between the National Renewable Energy Laboratory (NREL) and Sandia National Laboratories (Sandia). With a multimillion-dollar budget and more than 40 researchers on the task, the project also includes partnerships with Oak Ridge National Laboratory, Lawrence Berkeley National Laboratory, the University of Texas at Austin, and Parallel Geometric Algorithms LLC as part of the broader ECP effort.
ExaWind is a marriage of three primary open-source physics codes: Nalu-Wind, an unstructured-grid computational fluid dynamics (CFD) code; AMR-Wind, a structured-grid CFD code built on the AMReX library; and NREL’s OpenFAST, a whole-turbine simulation code. ExaWind essentially enables researchers to unlock the secrets of flow dynamics, allowing engineers to take a virtual peek inside the operations of a wind power plant and evaluate their designs within this computer-generated environment.
ExaWind aims to run the highest-fidelity wind power plant simulations to date. Employing the Summit supercomputer at the Oak Ridge Leadership Computing Facility, one of the world’s fastest, as well as next-generation exascale-class supercomputers, the team hopes to one day graduate from utilizing petaflop computing power to exaflops (1,000 petaflops). Summit is capable of achieving 200,000 trillion calculations per second (200 petaflops), and the first exascale machines will be at least five times faster. The resulting high-fidelity models will enable a great leap in the understanding of, and ability to predict, complex flows and turbine responses in wind farms.
And better prediction leads to better outcomes.
“If you really understand a wind power plant system and can predict how it will respond to the wind flow, then you’re in a strong position to optimize the design and operation of wind farms, both on land and offshore,” said Michael Sprague, ExaWind principal investigator at NREL.
ExaWind’s reach extends beyond the shoreline.
Offshore wind power plant simulation, particularly floating offshore wind, requires an additional modeling capability from that of land-based wind.
Offshore environments are characterized by complex interactions where the atmosphere meets the ocean surface, such as the proliferation of large, irregularly breaking waves. These interactions intensify the exchange of energy and momentum between the ocean and the atmospheric boundary layer above it.
These environmental manifestations also play a key role in the wind- and wave-related forces that act upon floating offshore wind turbines.
To help capture these fundamental interactions, NREL’s High-Fidelity Modeling team implemented a new moving-wave boundary condition in the ExaWind modeling and simulation environment for offshore atmospheric dynamics.
“With the ability to simulate ocean waves and gain a deeper understanding of how they impact the wind flow at turbine height, ExaWind can ultimately lead to improvements in offshore wind turbine design,” said NREL postdoctoral researcher Georgios Deskos.
The future looks bright for ExaWind, including a recently established partnership with General Electric (GE). GE and NREL are using the Summit supercomputer at the Oak Ridge Leadership Computing Facility to study the impact of winds found in the lower levels of the atmosphere on the performance of Atlantic Coast offshore wind power plants, in a project funded by the New York State Energy Research and Development Authority. GE, NREL, and Sandia are also producing a high-fidelity modeling toolkit for wind resource characterization and wind power plant development in a project funded by the DOE Technology Commercialization Fund. In this project, the team will be adding features to ExaWind to help industry partners better integrate high-fidelity modeling tools into their design workflows. These additions will afford engineers the ability to predict the performance of their designs with confidence.
By fostering a deeper understanding of the dynamics at play in wind energy generation, coupled with its predictive capabilities, ExaWind can help scientists and engineers develop the designs that maximize energy capture, improve turbine life spans, and reduce the cost of wind-generated electricity.