Big data and AI in electrification

Artificial Intelligence (AI) has been around for a while. The first models date back to the 70s but these concepts remained theoretical until we were actually able to teach computers to think for themselves. Today, Artificial Intelligence is everywhere. It allows computers and cloud connected devices to reproduce human-related behaviours such as reasoning, planning and creativity. Artificial Intelligence is primarily dependent on the quantity of data it is given. This is where big data plays an active role. With the increased collection and analysis of digital data, big data and AI are now emerging as rich areas of opportunity for electrification professionals.

Electricity 4.0 : Big data and AI for smarter power management

Big data is a major trend in the energy industry. The Electrical network become smart grid due to Data collected from a variety of sources, such as smart meters, sensors, twin digital. Once stored, this data is an invaluable resource for the industry to make better decisions about energy production and consumption.

Electricity was deployed extensively in the late 19th century, which was the First Wave of Electrification from 1880 to 1920. This period saw the widespread adoption of electrical power in industry and the development of the first electrical grid. Then the Second Wave of Electrification took place between 1920 and 1950 with the expansion of the electrical grid into homes and the development of new electrical appliances such as refrigerators, washing machines… During the third wave of Electrification from 1980 to present, we have seen the growth of the digital revolution and the development of new technologies such as computers, the Internet, and mobile phones.

Today the fourth wave of electrification so-called Electricity 4.0 is characterized by the integration of digital technologies such as Artificial Intelligence (AI), the Internet of Things (IoT), and advanced data analytics into the electricity infrastructure.

The aim of Electricity 4.0 is to create a smarter, more efficient, and more sustainable electricity system that can respond to the fast changing demands (+20% by 2030, +40% by 2040).

Electricity 4.0 is expected to optimize the use of existing assets, integrate renewable energy sources into the grid, increase energy efficiency, reduce greenhouse gas emissions, improve grid stability, reduce costs for customers and provide more reliable and flexible energy services to customers.

Moreover, generative AI and adjacent models are changing the game. Indeed, support technology reaches a new level, application development time is reduced, and powerful capabilities are brought to non-technical users.

Just recently, we saw the buzz around ChatGPT and what it can achieve. For instance, if we ask the question “how do big data and AI impact electrification”, we have to admit ChatGPT answer might not be perfect but is still very impressive.

These technologies will definitely have an impact on the world of electrification. But AI is mainly dependent on the quantity and quality of the data that will be used to learn. Big data provides the storage and processing capabilities necessary to educate the AI by feeding it with a lot of information.

Machine Learning and AI are the winning combo to efficiently exploit big data. This involves identifying patterns via data mining and data science more generally.

Big data: The cloud has won

In the age of big data, the famous wave 2 of “move to cloud” announced by providers is underway and is accelerating. As a reminder, the first phase of a migration to the cloud is a discovery phase that allows to analyze the strengths and weaknesses of an infrastructure and to determine future needs.

The number of detractors is getting smaller every day, privacy and sovereignty issues are both solved by the strategic commitments of clouders and swept away by the ease of use… All sectors – banking, telecoms, insurance, etc. – are rapidly adopting cloud-hosted big data solutions.

The first paradigm shifts are appearing in the world of electrification, driven in particular by operators such as Total Energie or Schneider. We can also note the predominance of the estimated Azure services of Microsoft Vs Aws of Amazon in the field of public cloud related to big data.

Exploring the challenges of generative AI and Big Data in 2023

Generative AI promises to make 2023 one of the most exciting years for AI and, by extension, Big data!
Keep in mind that ChatGPT’s prowess is based on the net recorded in 2021, but, as with any new technology, we must always proceed with pragmatism and measurement, because the technology now presents many challenges:

• Ethics: what sovereignty for data? What protection for personal data? What commitment to transparency and readability by the players?
• Environment: AI and Big Data are a paradox in that they are both a solution for optimizing energy consumption and resource mobilization, but also a cause of this increase;
• Cybersecurity: AI and Big Data in the field of energy is largely based on measuring instruments, therefore on IOT, offering an ever increasing security surface;
• Business Model: if the value of AI in the energy field is no longer to be demonstrated, the business model associated with services is very complex. For example, if we take the residential segment, the Chat GP virtual assistant has made the buzz as has Amazon with the announcement of a massive layoff, including the Alexa division (Amazon’s virtual assistant), in the same week;
• Talents: the development of digital services requires the onboarding of excellent technical skills, but not only. It’s the entire operating model that needs to be rebuilt. The human dimension is one of the biggest challenges brought by AI and Big Data: attractiveness, meaning of work, conditions, etc.

Big Data analysis combined with artificial intelligence also involves various risks. Key concerns include unintended consequences of automated decision-making, increased risk of cyber-attacks due to reliance on technology, inaccurate predictions leading to poor decisions, over-reliance on algorithms instead of human judgement, lack of transparency in the development process, etc…

AI and big data for Nexans

As previously expressed, AI in the energy domain is most often carried by a phygital system, meaning software + hardware.

To this end, an important part of our work in terms of AI and big data concerns the implementation of learning based on neural networks. The latter’s role is to translate images or text from measuring instruments (thermometers, drones, etc.) into numbers. The aim of these approaches is to understand recurrences, date them, predict them and locate them. We are in the AI for grid sensing.

One of the important activities in the field of electrification is the monitoring of networks for all segments: generation, transmission, distribution and use of electricity in buildings and industries. this requires the development and implementation of sensors that measure electrical activity along the value chain.

This is already the case in developed economies at home or in industry with Smartmeters. High voltage transmission lines are also systematically monitored for temperature and voltage. Medium-voltage electricity distribution networks and the connection networks of distributed renewable energies are less frequently monitored.

It is therefore essential to obtain data on the entire electricity deployment chain.

A second important activity is the analysis of data in order to optimise products or systems. this is at the heart of artificial intelligence and Big data.

In technical terms, we mobilize the techniques developed essentially for the field of natural language processing with recurrent neural networks and more precisely convolutional neural networks. In other words, the technology stacks of ChatGPT & DALL-E.

A long-term energy transition

Big Data is a hot topic with huge implications for the energy sector. It is a powerful tool that can be used to improve the efficiency of energy systems, production and consumption. In addition, it can also be used to improve electrical networks and smart technology.

Thanks to Big Data, it is possible to explore various scenarios and objectives related to the energy transition. In particular, this technology makes it possible to analyze how different systems and supply sources are interconnected and how they could be optimized in the long term. Thus, it offers an invaluable perspective to achieve a certain autonomy in a long-term energy transition objective.

The 3S (smart, small & selectivity) are challenges for the years to come. Addressed in a disorganized way today, they will become the real challenges for AI applications tomorrow:

• Smart data: Understanding and monitoring local ecosystems
• Small data: Limit the use of energy-intensive big data
• Selectivity: optimize the resources needed.