Wind
One of the biggest challenges in wind energy meteorology predicting the incoming wind to a turbine. There needs to be improved understanding of the local-scale physics, its interplay with the larger mesoscales characterizing the weather, and the turbine itself.
In order to address this challenge, ENERXICO will do research in novel wind energy design tools. The focus will be on developing and validating methodologies that support the understanding and prediction of the relevant atmospheric scales of motion for the operation and performance of wind turbines and farms in complex wind situations by taking advantage of next generation supercomputers.
ENERXICO´s objectives in renewable energies:
- Classify different wind conditions relevant for wind turbine design by specifying relevant external parameters in complex flow situations focusing on the needs of the meso-scale microscale coupling.
- Develop and implement both dynamic and statistical techniques for the coupling of mean flow and turbulence from the meso-scale via numerical weather prediction.
- Assess the developed meso-scale micro-scale coupling techniques in complex flow.
Specification
| Collaborating institutions | Iberdola, CIEMAT, BSC |
| Software involved | |
| Main mission |
Developing disruptive coupling methodologies between meso-scale and micro-scale models taking into consideration HPC related challenges and not only numerical and modelling aspects |
| Target TRL | TRL 3 |
| Relevant stakeholders | Iberdrola Renovables Energia |
| Achievements up to M24 | Coupling the outputs of WRF and ERA5 atmospheric modelsto Alya |
| Related work and further information | One conference paper is in preparation |
Computational details Alya
| Number of Cores / GPUs | Memory (GB) | Storage (GB) both temporal and permanent | #Files written both temporal and permanent | |
| Minimum | 512 | 1GB RAM per node | ~1 TB | ~50 |
| Average | 2016 | 96 GB RAM per node | ~15 TB | ~200 |
| Maximum | 4032 | 96 GB RAM per node | ~20 TB | ~400 |
Computational details WRF
| Number of Cores / GPUs | Memory (GB) | Storage (GB) both temporal and permanent | #Files written both temporal and permanent | I/O data traffic per hour during job | |
| Minimum | 40 | 32-64 GB RAM per node | ~100 TB | ~100 | ~25000 calls to MPI_Comm_rank() per process on each iteration |
| Average | 240 | 64 GB RAM per node | ~150 TB | ~200 | ~25000 calls to MPI_Comm_rank() per process on each iteration |
| Maximum | 800 | 64 GB RAM per node | ~200 TB | ~400 | ~25000 calls to MPI_Comm_rank() per process on each iteration |
Image
Figure 1: Turbulent flow structures around Alaiz mountain (dark red : low speed structures; light red: high speed structures)
"As one of ENERXICO´s industrial partners, Iberdrola benefits from this as user of ground breaking techniques for wind characterization. This allows Iberdrola to be updated with the state-of-the-art in wind modelling, focusing on the short and mid term advances in the field, in order to guarantee that the best methodologies and techniques are used in the design of new wind farms, or the analysis of wind events that may have a great impact on power production in already operating wind farms, as might be extreme wind events. The tools being developed within ENERXICO help us, therefore, also perform analysis with an increased resolution and accuracy".
Daniel Paredes, Head of Wind resource Innovation at Iberdrola
