Skip to main content

Unified mesoscale to wind turbine wake downscaling based on an open-source model chain

Final Report Summary - MESOWAKE (Unified mesoscale to wind turbine wake downscaling based on an open-source model chain)

MesoWake is a project sponsored by the European Commission's within an FP7 International Outgoing Marie Curie Fellowship granted to Javier Sanz Rodrigo, Senior Researcher at the National Renewable Energy Center of Spain (CENER). The outgoing phase of this fellowship, from August 2014 to July 2016, was hosted by the National Renewable Energy Laboratory (NREL) of the U.S. Department of Energy. The reintegration phase, back to Europe was hosted by CENER until July 2017. The project counts with the National Center for Atmospheric Research (NCAR) and the Barcelona Supercomputing Center (BSC) as scientific partners.

The objective is to contribute to the development of an open-source model chain that can bridge the gap between mesoscale meteorological processes and microscale wind farm models.

The multi-scale modeling system spans domain sizes of the order of 1e7 to 10 m, at spatial resolutions ranging from 1e4 to 1 m and a temporal range from a few days to fractions of a second. The need for a multi-scale approach is the result of the evolution of wind turbine technology that, over the last decades, has been successful at exploiting wind turbine scaling designs with rotors already spanning more than 150 m diameter and hub heights above 100 m. Furthermore, large wind farm arrays can extend more than 10 km creating their own boundary layer structure with important interaction with the free atmosphere. This increasing range of scales is challenging traditional wind engineering models that consider the wind farm system as an idealized microscale system where surface-layer theories apply. This lack of appropriate physics often leads to wind farm underperformance and high project financing costs.

In effect, wind energy is the fastest growing renewable energy technology to increase the share of global energy demand from 2% in 2010 to 11% in 2030 [1]. To support this prosperous future, unprecedented research programs in the U.S. and Europe have been recently launched to improve our understanding of the complex flow physics around and within wind farms. Better insight into the flow physics has the potential of reducing wind farm energy losses by up to 20% according to the U.S. Department of Energy’s Atmosphere to Electrons (A2e) research initiative. Its European counterpart, the New European Wind Atlas (NEWA) project, leverages national funding from 8 EU Member States to reduce resource characterization uncertainties below 10%. Over the next years these two programs will improve our simulation-based design capabilities. This can accelerate significantly the development of new turbine prototypes, promote innovation and reduce wind farm design uncertainties provided the modeling tools are validated systematically with high-fidelity data.

In MesoWake, special focus is given to the characterization of mesoscale forcing to drive atmospheric boundary layer models using realistic boundary conditions. Turbulence modeling is based on large-eddy simulation (LES) to resolve the turbulent scales that affect turbine and wind farm performance. The open-source unified model is based on a combination of WRF-LES, developed by NCAR, and SOWFA (OpenFOAM-LES coupled to FAST aeroelastic code), developed at NREL. The codes are installed at the MareNostrum high performance computing facility managed by BSC, member of the Partnership for Advance Computing in Europe (PRACE). This will provide a virtual laboratory for high fidelity modeling of atmospheric physics applied to wind energy [2]. This laboratory kicked-off in July 2016 with PRACE-MesoWake, a 3-year Project Access awarded with an initial allocation of 17 million cpu-hours for the first year. Scaling tests and a performance audit were carried out, under the Performance Optimization and Productivity (POP) Centre of Excellence, to determine the suitability of the codes to run in high-performance computing systems.

The MesoWake project is timely at creating a European hub for the development of high-fidelity modeling capabilities based on an open-source framework. This unified model will support ongoing activities in Europe and the U.S. and strengthen the cooperation in this strategic interdisciplinary research area that requires: cross-cutting knowledge across a wide range of atmospheric and engineering sciences, interfacing of models that have been developed separately, large computational resources, and high fidelity experiments for validation. The model evaluation strategy adopted in MesoWake follows the systematic verification and validation framework developed in the context of the International Energy Agency's IEA Task 31 Wakebench [3].

Initial assessment of the meso-micro methodology around the GABLS3 diurnal cycle has shown good consistency of the methodology to couple mesoscale and microscale models asynchronously [4]. The method is relatively straightforward to implement in existing microscale models. A benchmark revisiting GABLS3 for wind energy atmospheric boundary layer models demonstrated the general applicability of the method for LES as well as Reynolds-averaged Navier Stokes (RANS) models [5]. Ongoing work is directed towards demonstrating the general applicability in other sites and wind climate conditions. To this end, the method will be tested within the validation strategy of the New European Wind Atlas (NEWA) model-chain, which will exploit a large database of field experiments in all kinds of conditions. This validation program is leveraged internationally through the IEA-Wind Task 31 Wakebench to produce a validation suite for wind flow models in Windbench repositories following the GABLS3 case as a demo.

The MesoWake research programme was complemented with a training programme focused on large-eddy simulation techniques, high performance computing and a Master in Business Innovation from Deusto Business School. A new design of Windbench as an open science platform was the innovation project for the master’s thesis.

[1] IRENA (2016), REmap: Roadmap for a Renewable Energy Future, 2016 Edition. International Renewable Energy Agency (IRENA), Abu Dhabi,

[2] Sanz Rodrigo J. (2017) Open-Access Virtual Laboratory for Wind Energy in European High Performance Computing and Data Management Infrastructures: The MesoWake Pilot Study. Deliverable of the FP7 MesoWake project, Grant Agreement no 624562, July 2017, doi: 10.5281/zenodo.835300

[3] Sanz Rodrigo, J., Chávez Arroyo, R.A. Moriarty, P., Churchfield, M., Kosovic, B., Réthoré, R.-E. Hansen, K.S. Hahmann, A., Mirocha, J.D. and Rife, D. (2016a) Mesoscale-to-MicroscaleWind Farm Flow Modelling and Evaluation. WIREs Energy Environ. doi: 10.1002/wene.214

[4] Sanz Rodrigo, J., Churchfield M., Kosovic B. (2017) A methodology for the design and testing of atmospheric boundary layer models for wind energy applications. Wind Energ. Sci. 2: 1-20, doi: 10.5194/wes-2-35-2017

[5] Sanz Rodrigo J, Allaerts D, Avila M, Barcons J, Cavar D, Chávez Arroyo R, Churchfield M, Kosović B, Lundquist JK, Meyers J, Muñoz Esparza D, Palma JMLM, Tomaszewski JM, Troldborg N, van der Laan MP, Veiga Rodrigues C (2017) Results of the GABLS3 diurnal cycle benchmark for wind energy applications. Journal of Physics: Conference Series, 854: 012037, doi :10.1088/1742-6596/854/1/012037

Javier Sanz Rodrigo,