Machine learning, IoT and big data for energy efficiency: a use case

Although these disruptive technologies come with challenges, the opportunity is there for the energy market to leverage them to solve key issues such as energy efficiency in buildings.
Published: Mon 16 Jul 2018

Cutting past the hype to the real use cases of the internet of things (IoT), big data and machine learning for the energy sector is of increasing importance.

By now, stakeholders and energy market players should know the technologies are coming - re-imagining their uses to solve crucial energy challenges is the next step.

This is what Filippo Ferraris, Chief Product Officer at Enerbrain discussed in his presentation, “Reducing energy consumption with machine learning” at the energy efficiency Engerati meet, specifically the use case for energy efficiency.

Ferraris cites the first device connected to the internet of things - a live-feed camera enabling office-workers to monitor when their coffee was ready, as a key example of why and how IoT technologies have developed: they are problem solving and people oriented.

The technology and challenges

The IoT is already well underway, and directly impacting how players in the energy market operate. Ferraris says: “The current status is that we have around 34bn products that are connected online right now, which is an insane number.”

This, in turn, enables the development of smart cities. Ferraris explains: “Cities are starting to become smarter; they produce data, they have the ability to sense what is happening in real time. This enables a ton of applications, from improving security to improving the daily life of the citizens.”

However, the development of these cities will not come from the top-down, Ferraris says. The implementation of sensors and developments in large infrastructures will inevitably happen from the bottom-up, and the onus is for the energy sector to get on-board.

However, IoT technologies are not without their problems. Ferraris explains: “One of the key elements is that we need more and more open standards. Every company is developing their own type of database, their own way of saving data, their own protocols. We’re getting towards standards, but it’s a very slow process, and unfortunately this can only be done from a top-down level.”

Also, both devices and infrastructures consume energy all the way from the antennae to the storage database, as does the computing power needed to process everything, creating further issues for energy efficiency the world-over.

Then, there is the reality of device obsolescence, with Ferraris citing that every device we use today will be thrown away in maximum 4-5 years.

Additionally, Ferraris says: “Other questions include how do we protect the privacy and security of the infrastructure.”

In the presentation, Filippo Ferraris explains the structure and conception of Enerbrain’s IoT machine learning solution.

Disruptive technology for energy efficiency

The first question Ferraris addressed upon the conception of Enerbrain was how the energy sector can use these technologies to improve our planet.

He says: “When we started this project, we began with a very large issue that stuck in my mind, which is that 40% of the energy of the entire planet is used by buildings.”

So where does this huge consumption level come from, and is it justified? Ferraris asks: “Most of that energy is used to keep us warm in the winter and cool in the summer, but how many times have you still felt uncomfortable?”

This is where the opportunity for leveraging machine learning comes in. Ferraris says: “In our vision, we imagine a future where buildings will learn from the users and dynamically improve, where they can change all of the settings in real time. We can do that thanks to machine learning - we can learn when are the peak hours, when are buildings used the most. We can leverage that so that we can improve the system.”

Filippo Ferraris explains how Enerbrain achieved energy savings of 35% year-round for an Italian shopping mall, not by shutting down devices, but rather by fine-tuning them.

An IoT solution for building energy management

Enerbrain has leveraged IoT to control, manage and make buildings efficient.

Ferraris says: “We have developed environmental sensors that are able to monitor, for example, CO2 levels, temperature, air quality. All that data is very valuable for machine learning and artificial intelligence algorithms. That technology is then able to process the data in the cloud and tell the building in real time how it should behave.”

Then, he explains, thanks to heating, ventilation, and air conditioning (HVAC) actuators, devices which are connected to the cloud and installed on HVAC systems, operators can teach the building how to save energy and improve comfort.

Ferraris details the implementation of the solution: “We install our HVAC IoT solution in the heart of the building, where the energy is consumed the most, and it is able to control all of the valves in the building, thanks to the data coming from the cloud and from the sensors.”

From there, Ferraris explains how end-users see the results: “Thanks to IoT and big data, we can share all of the data with our clients, but more importantly these buildings become more sustainable, more comfortable and more efficient.”

The results, Ferraris explains, are significant beyond in single buildings: “On average, we save 30% of all energy used. If you make that global, there’s a huge impact on the efficiency of the planet.”

For more insight to the use of machine learning, IoT and big data in energy efficiency, explore the Engerati Meets coverage of Fillipo Ferris’ talk, “Reducing energy consumption with machine learning” and more.