2024-07-10 –, While Loop (4.2)
A reference model of airborne wind energy systems will be presented, developed for the International Energy Agency. The requirements included a modular design, good extendability and easy installation without using proprietary software and finally interoperability with larger simulation frameworks, written in Python.
Julia was chosen for performance, readability and maintainability. The solution consists of a set of eight Julia packages that can be used individually or in combination.
Airborne wind energy - harvesting the wind using a tethered wing - is a new technology that promises much lighter, portable wind energy systems that are beautiful to look at. But designing a flying system with low costs and very high reliability is challenging and requires excellent software tools.
The International Energy Agency has a working group on airborne wind energy to provide stakeholders (governments, investors) with tools to investigate the economic feasibility of these new type of energy systems. In this context, and based on the PhD thesis of the author, a set of Julia packages was developed to be able to simulate the loads and the performance of these systems in all kind of sizes and environmental conditions.
In the presentation, the author will first explain the concept and the status of airborne wind energy and then explain how such models are used. Then the package hierarchy will be explained before diving deeper into the implementation of some of the packages. Finally, some of the challenges of using Julia for such a large software and research project are pointed out.
An impressive 3D visualization was implemented using the GLMakie library. The simulation performance using a pure Julia DAE solver is outstanding, a performance of 66 times faster than real-time was achieved. Finally, an outlook on the coming improvements will be presented.
Uwe Fechner, postdoc on wind energy at TU Delft, NL. He did his PhD on the control of airborne wind energy systems also at TU Delft from 2010 to 2016. and is using Julia since 2014.