How Artificial Intelligence Can Make More Of Energy

by Thorsten Koch, MA, PgDip

Smartphone assistants already use artificial intelligence (AI). But AI is equally relevant in the energy sector. Through analyses and predictions, through secure and smart networks of the future. Aside from that, AI could soon help consumers put money aside.

AI can save energy. Apple already uses 100 percent renewable energy in its data centers, and Amazon uses 50 percent for its infrastructure. Smaller companies, however, are not yet ready. That’s why the worldwide energy share of cloud computing is 3 percent. The irony is that without AI, that share would be much higher. Second downer: the hardware itself is energy-hungry. Even if a company decides to use shared server parks.

Making Use of AI

One sector in which energy can be saved is the energy sector itself. Offshore operations of oil and gas extraction, processing and distribution may well benefit from machine learning and data analysis, based on model simulations. Not only because of artificial, so-called neural networks, which are capable of preinterpreting data, especially big data – which they have to because the amount of data would simply be unmanageable.

In the UK, the UK National Data Repository (NDR) started operations in early 2019 – with more than 130 terabytes of capacity today. In the private sector, for example, Shell is working with Microsoft to make better use of AI. Data integration and visual analysis are usually combined. AI synthesizes the available data – more or less in real time. With the right measuring instruments, companies can better assess the technical situation, as well as that of the markets. Even with increasing data volume.

Future Energy

AI will certainly help to harness clean energy, but also render it affordable for consumers, without completely calling into question coal and gas. Better use of infrastructure may even make energy more financially advantageous. Consumer dissatisfaction can also be identified by AI – this which promises a better remedy for households.

Predictions can reduce dependence on energy storage. However, this cannot incorporate completely unpredictable events. Renewable and alternative energy sources are still going to be affected by adversity, such as heat waves and storms. A cloudy day with covered skies can, if one relied solely on sustainable energy sources, lead to bottlenecks in supply. On a sunny or windy day, on the other hand, it can lead to an oversupply. AI is already being used today to coordinate supply and demand, which equals production and consumption, faster and better. This includes oil production, but also solar and wind energy.

Nnergix in the Spanish city of Barcelona, for example, can already coordinate and control renewable energy reserves, so that they are available as soon as needed, thanks to load management, through analysis of sensor data. Repairing and replacing overused, damaged parts, or predictive maintenance, is used by General Electric and others, whereas AI is used for machine condition analysis of wind turbines and hydro generators, thus contributing to the occupational safety and health of workers. The application of AI and robotics have limitations of cost, storage and computing capacity, but also of the environment in which they are to be used.

Household Energy

Energy production is not the only field in which AI can be of advantage. In the future, AI will not only try to measure what consumers want to buy, but also help them save. AI as part of the Internet of Things, i.e. smart household appliances, AI can analyze, for example, how the energy requirements of a household are dependent on the time of year and the time of day, which has come to be known as Home Energy Management.

Also, energy sharing between households through the usage of microgrids, are allowed by AI, which allows for small energy networks to be created. Intelligent analysis units to control how energy is distributed require security measures such as anti-virus and firewalls which evaluate energy distributors in subnets to ensure information security and smooth operations. This is all the more important for the Smart Cities of the future. At the moment, developers invent partly self-healing systems, by identifying weak points, and allowing for quick responses.

Reducing Costs

Companies enjoy financial advantages when they voluntarily step on the brakes in times of high energy demand. AI allows to identify and identify when this ought to be. Demand-side flexibility can, hence, reduce costs, but it also bears certain question marks, because, as we know, not everything is fully predictable. It is also problematic when data is sold that has not been sufficiently anonymized. Or, as clinical experiments show, where human oversight is lacking and AI, which can be extremely adaptive, learns in directions that are unwarranted.

Data analysis, data interpretation and monitoring can be handled graphically, for example by vector-based tensor units. By entering keys or human speech, an AI unit receives commands, and through visual representation it passes on its findings. Once a unit based on a neural network is trained, it acts in a more and more structured and reliable manner. The question, however, is to whether the structure of a system is too restrictive, or else too flexible. And to what extent, for example, new developments are sufficiently reflected. At the end of the day, technology is necessarily linked to integrity.

The market for artificial intelligence in the energy sector is growing. Larger companies such as Siemens and IBM or smaller AI-oriented companies already make big profits. The global market is expected to grow by USD 6 billion from 2019 to 2023. The North American market currently accounts for half of this.


Whether in energy production, in energy retail or distribution: above all, the actual use of AI is a question of cost, but also of feasibility and acceptance. Development is rapid, and no one will ultimately be able to avoid it. Soon, AI will most likely ensure that both the economy and consumers are better prepared for the future – at least for foreseeable, medium-term problems. So it should be possible to use energy in an intelligent mix even more efficiently than it is done today.

Artificial intelligence recognizes patterns and finds solutions. In many cases, faster, sometimes even better than humans, however humans should remain at the helm. Overall, the trend towards AI is forward-looking, as companies are players in the race for technical edge, market share, and revenue.

AI could also help to discover and promote energy-saving techniques in other areas, such as vehicle construction. But also in the transport sector, which in 40 years has more than doubled its emission levels. To give an example: when electric-powered trucks shall fill up with electricity as needed, and still every one of them will start, the next morning, with a full battery.

December 2019

This site can be reached via:

Author: author

Leave a Reply

Your email address will not be published. Required fields are marked *