Renewable energy transition: examining the impacts of wind energy through simulation

Renewable energy transition: examining the impacts of wind energy through simulationThis post was contributed by Remco Verzijlbergh, co-founder and CEO, Peter Baas, R&D Specialist at Whiffle, and Ilan Gleiser Principal ML Specialist, Milo Oostergo, Principal Startup SA, and Ross Pivovar Senior SA, Simulations at AWS

The world is currently facing a crucial turning point in the fight against climate change, and renewable energy sources like wind energy are playing a major role. As we continue to move towards a more sustainable future, it is important to understand the impacts of the anticipated large-scale roll-out of wind energy. This is where atmospheric model simulation comes into play.

By performing large scale simulations on HPC clusters in the cloud, built using  Amazon EC2 P4d instances and powered by Nvidia A100 GPUs, we can get a better understanding of how wind energy affects our atmosphere and the environment as a whole.

In this blog post, we’ll delve into the realm of simulations and explore the impact of wind energy on our planet.

The Rising Power of Wind Energy

In the quest for a sustainable future, wind energy has emerged as a powerful player in the fight against climate change. The rising power of wind energy has captured the attention of scientists, policymakers, and the general public alike. It has quickly become a driving force in the transition towards renewable sources of energy, and for good reason.

Harnessing the power of wind is not a new concept. Humans have been utilizing wind energy for centuries, from simple windmills for grinding grains to more advanced wind turbines that generate electricity. However, it is only in recent decades that wind energy has gained significant momentum as a viable alternative to fossil fuels.

One of the main reasons for the increasing popularity of wind energy is its environmental benefits. Unlike traditional energy sources such as coal or natural gas, wind energy does not produce harmful emissions or contribute to air pollution. This is a crucial advantage in the battle against climate change, as reducing greenhouse gas emissions is essential to mitigating its impact.

Furthermore, wind energy is a renewable resource, meaning that it will never run out. As long as the wind keeps blowing, we can continue to harness its power. This is in stark contrast to fossil fuels, which are finite resources that will eventually be depleted.

Another key factor driving the rise of wind energy is its economic potential. As technology has advanced, the cost of producing wind energy has steadily decreased. In fact, wind energy is now one of the cheapest sources of electricity in many parts of the world. This affordability, coupled with the growing demand for clean energy, has led to a boom in wind energy installations globally.

In addition to its environmental and economic benefits, wind energy also has the potential to enhance energy security. Unlike fossil fuels, which are often imported from other countries, wind energy can be produced locally. This reduces dependence on foreign sources of energy and strengthens national energy independence.

The increasing power of wind energy is evident in the rapid growth of wind farms around the world. These vast fields of turbines are capable of generating large amounts of electricity, powering homes, businesses, and even entire cities. As technology continues to advance, wind turbines are becoming more efficient and capable of harnessing even more energy from the wind.

Harnessing wind energy has significant positive effects on the atmosphere. By replacing traditional energy sources with wind energy, we can reduce greenhouse gas emissions and combat climate change. Wind energy does not release carbon dioxide or other pollutants that contribute to air pollution and global warming.

Moreover, wind energy can contribute to improved air quality. By reducing reliance on fossil fuels, which emit harmful pollutants such as sulfur dioxide and nitrogen oxide, wind energy helps to alleviate respiratory issues and promote healthier living conditions.

Unraveling the Data: Using Simulations to Predict Future Scenarios

As we strive for a greener future, simulations play a vital role in understanding the impacts of wind energy on our planet. Using superior computing power to solve the complexities of the weather with unparalleled accuracy, we can unravel the data and predict future scenarios. By simulating large-scale wind farms and analyzing the interaction with the atmosphere, we can gain valuable insights into the sustainability and environmental effects of renewable energy. These simulations offer a powerful tool for decision-making and policy development, ensuring a more informed and efficient transition to a sustainable energy future.

Who is Whiffle?

Whiffle services are designed to empower renewable energy industries and optimize operations by providing accurate meteorological forecasts, precise wind and solar power production forecasts, comprehensive wind resource modelling and detailed energy yield assessments. They are the driving force behind the revolutionary Whiffle Wind web app, which is significant in renewable energy development. With their cutting-edge Large-Eddy Simulation (LES)-powered weather Computational Fluid Dynamics (CFD) modelling technology, Whiffle is helping to pave the way for a greener and more sustainable future.

What sets Whiffle Wind apart is its user-friendly interface, in contrast to traditional weather-modeling software which is often complicated and cumbersome. Whiffle Wind makes it easier to access and interpret the data with relative ease, making it a valuable tool for professionals and enthusiasts alike.

Whiffle’s ultra-high resolution atmospheric model is also important to renewable energy development and operations. While traditional weather models work with grid blocks of 10 by 10 kilometers, Whiffle works on a resolution of 100 x 100 meters – a hundred times more information density on all weather-defining factors. With its hyper-local forecasts and simulations, Whiffle can help businesses plan for adverse weather events, ensuring the safety and efficiency of renewable energy operations.

Whiffle’s stack can also deliver highly accurate predictions, allowing businesses to optimize their operational performance. Whiffle’s model includes elements like buildings and trees. High-resolution digital terrain models are used to allow for detailed simulation in complex terrain. Users can upload wind turbines at any desired location to assess their energy production. As a full-blown weather model, Whiffle’s LES also calculates cloud coverage, which allows for detailed solar radiation predictions.  This means that whether you need to consider wind farms or solar installations, Whiffle can help you to get the insights you need for making informed decisions.

Core to Whiffle’s technology is Large-Eddy Simulation (LES). This advanced approach to weather modelling follows the laws of atmospheric physics very closely, allowing for detailed simulations of turbulent atmospheric flows. By capturing the small-scale processes that traditional numerical weather prediction models miss, Whiffle can provide a more comprehensive understanding of weather conditions.

Fig 1. Achieve accuracy tailored to your specific site requirements, irrespective of terrain complexity, location or the intricacies of wind turbine clusters.

Fig 1. Achieve accuracy tailored to your specific site requirements, irrespective of terrain complexity, location or the intricacies of wind turbine clusters.

Delving into large-scale wind farms: impact and sustainability

As the renewable energy transition gains momentum, the installation of large-scale wind farms is becoming increasingly common. For example, the installed capacity of offshore wind energy in the Dutch part of the North Sea is expected to grow dramatically in the next few decades. Several research institutes and the Dutch government are examining scenarios with up to 100 GW (around 5,000 to 10,000 large wind turbines) of offshore wind in 2050. For comparison: the currently installed capacity is around 5 GW, so this means a 20-fold increase.

However, the impact of such a large-scale roll-out of wind energy on the atmosphere and vice versa is highly uncertain. Understanding these interactions is crucial for the wind energy sector and policy makers as they strive to develop sustainable and environmentally friendly solutions.

This is where the WINS50 project comes into play. The project, carried out by Whiffle, TU Delft and KNMI, aims to reduce uncertainties surrounding the interaction between a large-scale roll-out of offshore wind energy and the atmosphere. To achieve this, the project uses Whiffle’s high-resolution turbine-resolving weather model, GRASP. With this advanced model, researchers from Whiffle are simulating the entire Dutch North Sea with the currently-installed wind power capacity and with the projected 2050 capacity. The data they produce will be made available to the wind energy sector and will be used for in-depth studies.

Within the WINS50 project, a multi-GPU LES model has been developed to enable computational domains of hundreds by hundreds of kilometers. This model will provide even more detailed and accurate information about the atmospheric interactions of large-scale wind farms, helping to inform future decision-making and policy development.

With the combination of advanced simulation models and the collaboration between research institutes and the wind energy sector, we can gain a deeper understanding of the impact and sustainability of large-scale wind farms. By reducing uncertainties and improving our knowledge, we can continue to drive the renewable energy transition forward and create a greener and more sustainable future for all.

Reference Architecture

 Fig 2. AWS High Performance Compute reference architecture comprised of multi GPU AWS Parallel Cluster open-source tool that builds a complete HPC environment with all the benefits of the cloud built-in.

Fig 2. AWS High Performance Compute reference architecture comprised of multi GPU AWS Parallel Cluster open-source tool that builds a complete HPC environment with all the benefits of the cloud built-in.

To run the simulations, Whiffle used AWS ParallelCluster – a fully-supported open-source tool that builds a complete HPC environment with all the benefits of the cloud built-in. That includes the elasticity to scale horizontally (and contract) based on need, instant access to the latest technologies (CPUs, GPUs, and accelerators), and the flexibility to iterate resource selection in minutes to optimize costs or squeeze out better performance.

For the cluster configuration, Whiffle used a single queue with four P4d instances that are each powered by eight GPUs. The NVSwitch GPU interconnect enables each GPU to communicate with every other GPU in the same instance with 600 GB/s bidirectional throughput and single-hop latency.

Whiffle’s testing showed that the LES model scaled extremely well when run with 8 GPUs in parallel. It supports their philosophy of keeping the model architecture as simple as possible so they can quickly adopt advances in computing and cloud infrastructures. As the performance of parallel GPU compute resources grow, we can all start planning for continental or global scale weather forecasting on a 100m resolution in the future.

The data set and simulation results are stored on Amazon FSx for Lustre which is mounted on the head node and compute nodes, and provides sub-millisecond latencies, high throughput, and millions of IOPS.

The final results from the simulation are stored in Amazon Simple Storage Service (Amazon S3) after post-processing.

The dataset

The dataset is the result from the WINS50 LES described above and it’s available in the Registry of Open Data on AWS. It’s approximately 40TB in size, and consists of three types of data:

  1. Three-dimensional meteorological data over the entire domain at an hourly resolution for selected heights above the surface. Variables included here are wind speed, wind direction, temperature, etc. In addition, surface variables such as the surface pressure of downward solar radiation are available in two-dimensional format at hourly resolution.
  2. For 600 locations distributed over the domain many meteorological variables are provided over the entire atmospheric column with a 10-minute resolution. Turbine data gives energy production, wind speed and thrust for all the simulated turbines, i.e. more than 10,000 for the 2050 scenario. Turbine data is available at 5-minute resolution.

There are many possible use cases for the data, including scientific fields from the atmospheric sciences and energy system sciences, too. Within the energy sector, the data can inform grid operators, wind farm owners, developers of new wind farms and energy traders. Finally, the data can be used to train AI based models, opening more use cases where computationally-fast models are required.


As we forge ahead in the renewable energy transition, there are bound to be both opportunities and challenges on the horizon. The growing popularity of wind energy presents a unique opportunity to transition towards a more sustainable future. By harnessing the power of the wind, we can significantly reduce our carbon footprint and combat climate change.

Advancements in technology, like Whiffle’s simulation tools coupled with the use of HPC clusters on AWS, offer new avenues for improving the efficiency and accuracy of wind energy simulations.

But there are challenges. Wildlife impact and visual concerns must be carefully addressed to ensure the widespread adoption of wind energy is safe and beneficial. Acknowledging these challenges and leveraging the power of simulation and HPC, we can work to achieve that.

The content and opinions in this blog are those of the third-party author and AWS is not responsible for the content or accuracy of this blog.

Remco Verzijlbergh

Remco Verzijlbergh

Remco Verzijlbergh is co-founder and CEO at Whiffle, a TU Delft spin-off specialised in high-resolution weather forecasting for the renewable energy sector. He is also an associate professor at TU Delft, working on numerical models to support the integration of renewable energy sources in the energy system. As a CEO for Whiffle his mission is to bring the world’s most accurate weather model, the fastest computational resources and the best machine learning algorithms together to reduce weather risks in the renewable energy sector.

Ilan Gleiser

Ilan Gleiser

Ilan Gleiser is a Principal Emerging Technologies Specialist at AWS WWSO Advanced Computing team focusing on Circular Economy, Agent-Based Simulation and Climate Risk. He is an Expert Advisor of Digital Technologies for Circular Economy with United Nations Environmental Programme. Ilan’s background is in Quant Finance and Machine Learning.

Milo Oostergo

Milo Oostergo

Milo Oostergo is Principal Startup Solution Architect based in Amsterdam. He enjoys working with Climate Tech startups to help solve the most challenging sustainability challenges and leads the Climate Tech initiative for the EMEA startup team.

Ross Pivovar

Ross Pivovar

Ross has more than 15 years of experience in a combination of numerical and statistical method development for both physics simulations and machine-learning. Ross is a Senior Solutions Architect at AWS focusing on development of self-learning digital twins.

Peter Baas

Peter Baas

Peter Baas is an R&D Specialist at Whiffle. He has a background in atmospheric research and is a expert in wind energy. In recent years, he has been working with Whiffle’s high-resolution weather model. He wrote publications on the impact of large offshore wind farms on the atmosphere and vice versa. Peter served as project leader of the WINS50 project.