CONTACT

Background Articles and Features

The Untapped Opportunity of Machine Learning (ML) and Artificial Intelligence (AI) in the German/European Energy Storage Industry

It took some time since the beginning of the mobility revolution and the “Energiewende”, until in 2019 the demand for energy storage is constantly rising. Current market numbers indicate about 37 GW[1] installed in Germany alone, with huge capacities added each year as well as the capacity of several gigawatt expected to roll out with new electronic vehicles.

At the same time the global trend towards electrification leads to more and more devices demanding energy storage. New product niches are created in a very fast pace like E-scooters or E-bikes. While this is good news in general, it leads to a shift in global value chains, which constitutes a significant challenge for Germany/Europe. Thus, the on a global scale rather small energy storage industry in Germany and Europe should be alert in order to not miss out on a very important new development, namely, the untapped opportunity of Artificial Intelligence for energy storage value chains.


The economic shift needs a new approach – Artificial Intelligence as untapped opportunity

In order to grasp the importance of the argument and to comprehend why the focus on Artificial Intelligence is of crucial importance for Germany and Europe, the economic consequences of the above described scenario must be understood. In simple words, it brings about a global shift of value chains. A prime example is the replacement of the importance of engines, the core competency of German carmakers and suppliers, with batteries, which will turn it into the most expensive and valuable part of the car. Also, the disruption in energy production is to name. For Germany, but also for other countries in Europe, the question is therefore who controls the economically benefits resulting from this shift? And if this will enable the dominant player to take out larger chunks of the value chain? In fact, current fears are that dominant Asian players will be able to easily take out more margin of the car value chain due to the dependency on their supply and their dominance in Lithium-Ion[2] batteries.


Production alone is a too narrow focus

In the light of these facts, the author of this article argues that competencies in Artificial Intelligence for the battery industry should be the first priority. This is in sharp contrast to the current debate on subsidizing the Gigafactory establishment in Germany as a political response, which is the preferred way in which German politics approaches the structural industry issue, namely with the initiative by the minister of economic affairs Peter Altmaier. As much as initiatives in the field are welcomed, the macroeconomic and industry economic analysis of industry structure and industry life cycles models reveal that it might be the wrong tool at the right time[3]. While this is not the place to discuss the reasons in detail, some arguments are quickly listed: The market for Lithium-Ion is already completely split-up, it is not the basket case for jumpstarting a new industry, and its potential job contribution is very low[4].

 

Artificial Intelligence in context of the battery industry – making use of the data layer

In contrast, it is worthwhile considering the untapped opportunity of Artificial Intelligence. Some background first: For the first time in history data collection of real time operations of machines and devices becomes more and more available at almost no cost. Sensors and cheap transmission technologies allow for a systematically and automatically gathering of operational data that until recently required repetitive on-site intervention by highly paid and skilled engineers. At the same time, a new class of analytical business tools has become available, which benefit every year from the general research efforts and improvements of Artificial Intelligence. The application to name is the subfield Machine Learning which shifts the task of reproducing a model of the real world by a programmer to the computer that defines and programs the model itself, fed by data and searching for connections[5]. Although this technology is emergent and has its downside, it is the upcoming application layer we are looking at.

But what are the implications for businesses, in particular the energy storage industry? Basically, it means that the competition of having the critical mass of the best engineers in a given field move to a new field. In case of the battery industry this means to invest in AI talent, instead trying to match e.g. the comparably high number of electrochemical specialists and engineers of LG or others with own personnel, research and universities in Germany.
In this field fewer, but highly specialized programmers, data analysts, and engineers of the field can make the difference. Tapping this kind of skill set and knowledge, therefore, totally levels the playing field of the battery industry but also of many other industries.

And this is not pure theory. While currently preparing an assessment study to identify and quantify the economic impact of Artificial Intelligence on the value chain of the battery industry[6] some of the benefits, a number of successful case studies and several young companies already applying these technologies could be identified.

Among the German examples the author wants to highlight is the advanced battery analytics start up TWAICE. Knowing the project since its early stage, the founders have been interviewed to provide a data point. The current research of the company backed by data shows that by using these technologies in combination with standard models an improvement of up to 20-30% seems to be possible[7]. As a second example a recently presented research[8] by Prof. Birke and Max Weeber is carefully optimistic for the area of quality control.
 

Outlook of the future

The interesting fact is that these numbers do, however, only indicate the improvement of storage solutions already produced or the production process. But thanks to the growing data consisting of both models and test data which then in addition are in near future backed by real data, the next progress will be the acceleration of research around battery materials.

In fact, first research shows that combining data sources to identify promising research streams is possible.[9] But the possibilities are not limited to this. Another opportunity is adding data to the analysis that has remained unused so far. Two possibilities are audio data of batteries or pictures. Who knows what the systematic analysis of these will bring to light? Or what other form of data one could also gather and analyze?

In conclusion, the data layer and the knowledge to properly use new analytic tools and methods to analyze them is significantly changing the competitive field and opens up an opportunity for Germany and Europe to gain a competitive advantage in the battery field and upcoming technologies built on patents and knowledge. Closing it is to say that these kinds of tools are closely connected to German/European core technologies like measurement devices or testing devices that more and more compete with the software layer.

The opportunity is there, it just needs to be tapped.

 

Arnbjörn Eggerz, Founder
Iceventure

 


[1] Renewables Market insight 06 March 2019; Inspiratia - www.inspiratia.com/welcome/research

[2] Reportedly at the 16th Batteriestammtisch by the speaker, Asian players now demand to deliver not only basic cells but advanced solutions around it, significantly increasing their value chain participation - 10.01.2019

[3] Analysis by Iceventure – www.iceventure.de

[4] Presentation of results from „Roadmap Integrierte Zell- und Batterieproduktion Deutschland AG 2 - Batterietechnologie der NPE“; Dr. Kai-Christian Möller, Stellv. Sprecher Fraunhofer-Allianz Batterien, Fraunhofer Gesellschaft; https://www.iceseminars.eu/events/batteriestammtisch/10-runde-batteriezellfertigung-auch-in-deutschland.html - last accessed 062019

[5] Compare for discussion “Notes on AI bias”, Benedict Evans, April 2019; https://www.ben-evans.com/benedictevans/2019/4/15/notes-on-ai-bias - last accessed 062019

[6] “The economic impact of AI and ML on the battery industry value chain” [working title]; forthcoming study by Iceventure 2019; https://www.iceventure.de/studien/the-economic-impact-of-ai-and-ml-on-the-battery-industry-value-chain.html - last accessed 062019

[7] Internal study by TWAICE Technologies GmbH; 2019 - www.twaice.com

[8] Program title “Artificial Intelligence as a future part of battery design and production”; Prof. Dr. Kai Peter Birke, University Stuttgart; Battery Expert Forum, Frankfurt, 11.04.2019; https://www.battery-experts-forum.com/images/PDFs/Program.pdf - last accessed 062019

[9] Results shown and discussed in a personal talk with a professor of a leading German university.


The ees International Magazine is specialized on the future-oriented market of electrical energy storage systems, not only from a technological-, but also a financial and application-oriented point-of-view. In cooperation with ees Global, the ees International Magazine informs the energy industry about current progress and the latest market innovations.
Contact: Xenia Zoller - zoller(at)ees-magazine.com