French AI and Renewables Energies, current state, and prospects
IA & Renewable Energie Systems
José Sayritupac, an Ingelib Engineer , works as a data scientist oriented to photovoltaic systems. He has 4 years of experience on photovoltaic systems modeling and simulation, he has gradually shifted to the exploitation of large quantities of data and dedicate to machine learning algorithms for the increase of performance of photovoltaic plants.
In 2021 Artificial intelligence has stop being an unknown word.
We all have heard about this new technology that involves futuristic results only with the collection of a massive quantity of data and statistics.
A consensus on the artificial intelligence concept defines it as: “a group of technologies that allows to simulate intelligence and to accomplish task of perception, comprehension and decision making”1. Together with AI we have heard about Machine learning techniques, it is also defined as: “… subfield of artificial intelligence, which enables machines to learn from past data or experiences without being explicitly programmed” [2]. Both concepts are the tools expected to be exploited by a wide range of companies in the world to boost the confidence on the data and to push for a more lucrative business.
France has collaborated with the development of many mathematical supports for the AI. In particular, I can mention that one of the first tools used for the quick exploitation of AI models is Scikit-Learn and it was developed initially by a French person (David Cournapeau) and since then it has received several upgrades from many collaborators and research institutes such as INRIA. But the AI in France is not only focus on packages development. As a matter of fact, in March 2018 the mathematician and member of the national assembly Cédric Villani made official a report with more than 200 pages with possible strategies to be adopted by France to increase the impact of the AI on France and Europe [3]. Few months later, in November 2018 the minister of the higher education, research and innovation Fredérique Vidal, has presented the national strategy on the research of the artificial intelligence, as well as the allocation of 665 million of euros as a budget until 2022 [4], some other sources mention that the budget reach up to 1.5 billion euros until 2023 [5]. Another example on the interest of the French government on the AI was the establishment of an agreement with Germany on the bilateral consortia on AI matters [6].
Great! then France is working on a deeper introduction of the AI as a research topic on the society. I must mention that the penetration of the AI within the French industry was described extensively in a report named “Prospective, Intelligence artificial – Etat de l’art et perspectives pour la la France” [1] release by the General Direction of Enterprise (DGE). Such report provides a technology state of the art, a sectorial analysis, the position of France in AI topics and the territorial strategies for the development of futures opportunities in this domain. Many articles review this extensive report of more than 300 pages (I recommend [7] and [8]). After all, we can say that France wants to have AI everywhere. But, what about the renewable energies (RE) sector ? At first, the current trend for the use of AI seems low as mentioned by [9], it’s stated that only 6% of French companies are using AI, although this analysis is only focus on already deliver products (such as chatbots). In fact, the use of AI on the RE still in development and many products and services are meant to exist on the coming years. The media provides several used-case for the AI-RE mix, and I will like to list a few of these :
– Energy management optimization of RE thanks to AI, is expected to have a tool to manage complex energy systems, due to the introduction of solar and wind to the energy mix., as well as a more capable system response against potential risk from interconnected device failures [10].
– AI and the prediction of wind, a model from Google expected to forecast the wind production with up to 36 hours in advance, it will allow wind parks to better plan their production and increase their efficiency [11].
– AI for the optimal photovoltaic modules cleaning, a French society (Sunibrain) develop a tool to predict the best period of times on which pv modules should be cleaned [12]
The list of use-cases is obviously larger and could be extended but the goal of this article to is not oriented to the discussion of the solutions of AI on the renewable sector. I want to focus on the future impact of this technology within the energy sector. It is just a matter of time for the AI to fully involve the whole energy sector, smart demand and production are probably a few years ahead of this article. I present a prediction on the future operation of the energy market with the AI as the main tool as :
– A decentralized and self-organized environment of energy production sources with a day-to-day connection to weather conditions.
– Evaluation of Energy consumption and future consumption according to weather forecast through AI models.
– Energy market and disturbance automatic correction with AI models and wide exploitation of non-structured data.
In my opinion the future of the energy market is in fact part of the current market already, because of this, I rather to do an extrapolation of the current uses instead of a prediction of the future appliances linked to the AI. Additional sources on the integration and common uses of the AI in the current and future market are available at [13] and [14] on which the authors briefly describe the operation of AI in the energy sector. In addition, the European union provides list of minimal requirements that an AI should have to be deploy on a large market (check the ethics guidelines for trustworthy AI [15]). I found that the AI will be completely integrated in the common application of the energy market, in such a way that it will be impossible to define the marked without mentioning the AI as the main tool. But what challenge me the most is the impact of these application on the behavior of the final users, and we are the final users.
I recently read a book on a philosophical point of view of the AI and the human as an individuum [16] and it highlights the use and abuse on the AI by many countries as a tool of manipulation of the individua. What we have seen from the current used of AI is that the user becomes the product, in consequence a part of the algorithm that could be optimize –manipulated–, the term used by the author is nudge. Multiples examples of these nudge exist, one is the Netflix case, on which users are proposed more films based on their rate of interest, same case for amazon products on which the algorithm will display related articles, and Youtube videos will constantly display title related videos to keep the user within their platform.
To reduce all these nudges, a system of laws should prevent companies to make profit from human behavior manipulated by a machine. With this idea, the European union release a set of rules to avoid that AI system can manipulated the consumers [17]. In the case of the AI and renewable energies (AI-RE). My main concern on the uses of AI-RE mix is the risk of the final users to be manipulated to produce or consume energy at a specific time interval to optimize the energy AI algorithm. This behavior is expected if the energy market can have an influence on any of the other activities from the user (work environment, social media, food consumption), so we will become a part of the algorithm and not the outcome from it. From the other side, it could be an interesting application to dodge extreme energy consumption, and it could allow for a better use of the planet’s resources. Only time will let us know what the outcome was.
Written by
José Sayritupac
Sources:
[1] Prospective, Intelligence artificielle: état de l’art et perspective pour la France,
[2] Web article. Definition on AI and machine learning
https://www.javatpoint.com/difference-between-artificial-intelligence-and-machine-learning
[3] Report from Cédric Villani
https://www.aiforhumanity.fr/pdfs/MissionVillani_Report_ENG-VF.pdf
[5] https://www.entreprises.gouv.fr/fr/numerique/enjeux/france-terre-d-intelligence-artificielle
[6] https://www.bmbf.de/files/191016_Roadmap_FRA_GER_AI_Network.pdf
[8] https://www.thinksmartgrids.fr/actualites/france-leader-ia-domaine-energie
[9] Only 6% of french companies ues AI according to Eurostat.
[10] Tout le potentiel de l’IA au service de l’énergie,
https://www.lemondedelenergie.com/ia-energie/2020/02/26/
[11] Ai et énergies renouvelables l’avenir de notre planète,
[13] What is Artificial Intelligence in the Energy Industry, article on the integration of IA in the energy market as solution on smart grid configuration, and description of the future integration in the energy market as autonomous algorithms.
https://www.next-kraftwerke.com/knowledge/artificial-intelligence
[14] Artificial Intelligence Market In Energy & Utilities – Forecast(2021 – 2026)
https://www.industryarc.com/Report/17916/artificial-intelligence-market-in-energy-utilities.html
[15] List of trustworthy AI https://digital-strategy.ec.europa.eu/en/library/ethics-guidelines-trustworthy-ai
[16] Koenig, G. (2019). La fin de l’individu. Voyage d’un philosophe au pays de l’intelligence artificielle. Éditions de l’Observatoire.
[17] data protection, Europe: annonce officielle du projet de réglementation en matière d’intelligence artificielle