News Release

AI designs sustainable electricity storage at Graz University of Technology

Based on the vanillin made usable for electricity storage in 2020, an AI-optimised prototype of an environmentally friendly electricity storage system is now being developed in an international project

Business Announcement

Graz University of Technology

An environmentally friendly and efficient electricity storage system based on the storage medium vanillin is being developed.

image: An environmentally friendly and efficient electricity storage system based on the storage medium vanillin is being developed. view more 

Credit: Lunghammer - TU Graz

In 2020, Stefan Spirk from the Institute of Bioproducts and Paper Technology at Graz University of Technology (TU Graz) described the research achievement with which he and his team had succeeded as a “ground-breaking success in the field of sustainable energy storage technologies” to make redox flow batteries more environmentally friendly. They had replaced redox-active components of batteries with conventional vanillin, which meant that critical or environmentally harmful raw materials were no longer needed.

But the new storage medium alone was not enough. Meanwhile, Stefan Spirk is working to design a vanillin electricity storage system that is as sustainable as possible in its overall composition and yet efficient. Areas of application for the fully developed storage unit are primarily the industrial sector and the storage of electricity surpluses from renewable energies. Part of the research project called VanillaFlow are other institutes at TU Graz, Stefan Spirk’s Science Park Graz-based start-up Ecolyte and numerous other project partners.

The project is being funded within the EIC Pathfinder Challenge of the European Research Council and is therefore part of the EU Horizon programme, which is funding research and innovation. The Pathfinder Challenge aims to support the exploration of bold ideas for radically new technologies.

Optimisation of all components and processes

In the VanillaFlow project all components and processes of the storage unit are to be optimised: in addition to vanillin compounds as the storage medium, the membrane, the electrode and the control system. Among other things used for this are the possibilities of Artificial Intelligence and Machine Learning. This allows predictions for models of promising vanillin compounds to be made in a much shorter time than before. The most promising models are then also developed and tested in the laboratory to ultimately find the ideal composition for the storage fluid.

In the case of the membrane and the electrode, the primary aim is to replace the less environmentally friendly materials previously used for this purpose in battery storage systems with sustainable materials as well. For membranes, the Teflon modification Nafion has been used so far. By now, a paper-based membrane has been created, which is constantly being further developed. The patent for this has already been applied for. For the electrode, the project team relies on a carbon felt that offers less resistance through compression and also develops fewer deposits. New coatings and treatments are intended to achieve even better performance here.

Preliminary fine-tuning on digital twin

In order not to have to produce all iterations of the storage medium, membrane and electrode in advance, the project team also resorts to digital support here. By means of a digital twin, the individual components and their interaction can be virtually tested and checked in advance. At the same time, the control system of the storage unit is also being further developed in order to optimise its operation. An underlying artificial intelligence links these virtual results with the VanillaFlow project data. In addition, a techno-economic and ecological review is being conducted to ensure that the storage system is not toxic and complies with current legislation. After all, the finished product should be safe for people and the environment.

As soon as a first prototype of this AI co-designed storage system is ready, it is planned to integrate it into the TU Graz network. The maximum storage performance is intended to be at 10 kW. For future users, however, the performance is scalable according to demand. “When we harnessed vanillin for use in redox flow batteries about three years ago, it was clear to us that this was just the beginning on the road to environmentally friendly and efficient electricity storage for users in industry and power generation,” says Stefan Spirk. “By using AI to design, test and ultimately manufacture a sustainable electricity storage system from A to Z based on this storage medium, we are taking the next important step. Once we have fully developed a storage system without harmful materials, without rare raw materials, but with high efficiency and safety for people and the environment, this is an important piece of the puzzle for the further decarbonisation of the energy system and industry.”

The VanillaFlow Team

The VanillaFlow project is led at TU Graz by Ulrich Hirn, head of the Institute of Bioproducts and Paper Technology. Together with TU Darmstadt, he also supports the Ecolyte team in improving the paper based, proton-conductive membranes. The know-how in the area of Machine Learning is provided by the TU Graz computer scientists Roman Kern from the Institute of Interactive Systems and Data Science and Robert Peharz from the Institute of Theoretical Computer Science. Research into sustainable synthesis methods for vanillin compounds at the Institute of Molecular Biotechnology at TU Graz is being led by Harald Pichler and the carbon felt for the electrode is in further development with Ecolyte and Montanuniversität Leoben. A team surrounding Günter Getzinger, head of the Science, Technology and Society Unit at TU Graz, and the Spanish company Biobide are in charge of the techno-economic and ecological review.


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.