Pondera is seeking a highly motivated Master’s student to develop a machine learning-based weather hindcasting and forecasting model using open-source resources. Currently, numerical weather prediction (NWP) models, such as ERA5 and MERRA-2, are widely used. However, these models operate at coarse resolutions, limiting their effectiveness in accurately modelling wind energy production and wake effects. Computationally intensive techniques, including Computational Fluid Dynamics (CFD), Reynolds-Averaged Navier-Stokes (RANS), and Large Eddy Simulations (LES), provide better accuracy but require significant processing power and long runtimes. This project aims to explore the potential of machine learning (ML) techniques to enhance weather predictions, leveraging state-of-the-art open source ML models such as FourCastNet, Pangu-Weather, and GraphCast.
About Pondera
Since our founding in 2007, we have been at the forefront of renewable energy solutions. As a relatively young company with a wealth of experience, we're dedicated to solving energy, climate and environmental challenges. Our work ranges from small initiatives to global projects in onshore and offshore wind, solar, hydrogen and energy storage. In 2024, Pondera joined forces with Royal HaskoningDHV, an independent international consulting engineering company leading the way in sustainable development and innovation since 1881. Together we combine engineering, design and consultancy with innovative technology to deliver sustainable innovation around the world.
Project overview
This project will focus on developing an ML-based approach for both hindcasting and forecasting weather conditions, with a particular emphasis on offshore wind environments. The developed model will be benchmarked against traditional numerical models such as CFD, LES, and RANS, as well as real-world measurement data.
At Pondera, we encourage equal opportunities and diversity in the workplace, we welcome candidates from all backgrounds. Acquisition in response to this advertisement is not appreciated.
Your role
We are looking for a motivated and team-oriented student to join our team at Pondera, where you will play a key role in advancing renewable energy forecasting. Your responsibilities will include:
- Identify and implement suitable machine learning algorithms for weather forecasting.
- Prepare and preprocess training and testing datasets from meteorological and wind farm data.
- Train, test, and optimise ML models for hindcasting and forecasting offshore wind farm conditions.
- Benchmark ML model performance against CFD, LES, RANS, and real-world measurement data.
We welcome candidates who have..
- A Master’s-level academic background in Engineering, Computing, Physics, or a closely related discipline.
- Programming skills in Python, PyTorch, and JAX or other relevant languages.
- Knowledge of, and ideally experience with, machine learning models and their applications.
- Familiarity with computational modelling techniques for fluid dynamics.
- Strong analytical, problem-solving, and troubleshooting skills.
- Interest in weather modelling and renewable energy technologies, particularly wind and solar farms.
- Good written and verbal communication skills in English.
What do we offer?
- A modern office located next to Arnhem Central Station.
- Collaboration with a team of skilled and motivated professionals.
- Professional development in a growing organization and market
- Excellent employment or internship benefits
Interested?
Submit the following documents via email to Sophie Broekers at s.broekers@ponderaconsult.com:
- Cover letter outlining their motivation and suitability for the project.
- Curriculum Vitae (CV).
- A statement of personal research interests.
