Maker Pro

Machine Learning Diagnostics Tool Shows Promise for the Future of Europe’s Energy Management

July 09, 2020 by Sam Holland

Researchers from Rovira i Virgili University (translated from the Catalan title, 'Universitat Rovira i Virgili') have introduced an artificial intelligence system that presents users with an accurate reading of the effectiveness and feasibility of both water and room heating installations in the EU.

Addressing the Problem of EU Energy Consumption

Based in the Spanish region of Catalonia, the Universitat Rovira i Virgili (URV) researchers' chief interest is to act on the high level of energy consumption that is observed in buildings throughout the EU countries. "More than 40% of energy consumption in the European Union is by buildings," write URV staff in their official coverage, "and 63% of this figure is due to residential dwellings."

Moreover, according to a Eurostat report whose data was gathered in 2017, the EU's household energy consumption reaches 78.9% when you account for the 64.1% spent on general room heating (referred to more officially as 'space heating'), plus the 14.8% spent on the production of clean hot water.

Additionally, these high percentages may be a conservative estimate due to the time passed since the years in which the data was researched and published (namely 2017 and mid-2019 respectively).


Infographic of percentage of household energy consumption contributors 2017.

A colour-coded infographic from the European Statistical Office (Eurostat). It shows the percentages of household energy consumption contributors in 2017—the leading example of which is space (i.e. room) heating. Image Credit Eurostat.


The Technology That Requires an AI-Based Monitoring Solution

As touched on, the URV has implemented artificial intelligence (and, more specifically, a machine learning-based) system to address the concerns related to high energy consumption within EU buildings. As discussed later, this is mainly in terms of the extent to which those buildings are environmentally friendly.

But to further understand the importance of the URV's machine learning system, it is first helpful to define the technology that it is designed to enhance: SDHS, i.e., solar district heating systems.

Solar district heating systems are centralised installations that can produce hot water on a large scale. Such installations have large fields of solar thermal collectors: these may be either ground or roof-based, and they provide their gathered solar heat energy to district heating (DH) networks. Alongside the obvious example (namely urban districts), DH networks may cover many other urban areas, from small communities to large cities.

The key difference between typical (and, nowadays, often household) thermal energy systems and SDHS is that the former can store the gathered energy for hours (but potentially up to a week).

The latter technology, on the other hand, uses heavily-insulated underground storage tanks (which are thousands of cubic metres in size): this means that SDHS maintain energy through the very seasons—with the sunniest months, of course, being the ideal time to store energy for the cold days of autumn and winter.


European solar district heating system in Gram, Denmark.

Based in Gram, Denmark: an example of a European solar district heating system, complete with a large water reservoir known as a seasonal put heat storage system (the grey unit surrounded by the field of solar panels). Image Credit: Ramboll.


The Importance of Introducing an AI-Based Monitoring System

The URV's aforementioned machine learning-based system (which the researchers refer to as a 'diagnostic tool') was created in the interest of enhancing the efficiency of SDHS, as well as people's understanding of the extent to which such systems work—again, particularly in terms of their ability to prove environmentally-friendly.

This is a significant function, as the challenges in both rolling out and maintaining solar district heating systems mean that it can be challenging to ascertain just how beneficial they are. Write the researchers in Applied Energy: "SDHS' technical barriers during their design and operation phases, combined with their economic limitation, promote a high variation in quantifying SDHS benefits over their lifetime".

Alongside Catalonia, Denmark (see the above image for an example), Canada, Germany, and Austria are just some of the Northern—and therefore typically cold—countries whose demands for high-energy heating are a testament to the international need for solar district heating systems. And, by extension, of course, the global need for the requisite machine learning systems to be put in place.

Accordingly, the URV researchers went to Madrid to carry out a theoretical pilot study of their diagnostic tool (the technicalities of which can be read in the researcher's journal paper, also linked above).


The URV's Pilot Study and Its Findings

The URV's pilot study in Madrid involved the researchers using the AI diagnostic tool to take stock of Europe's various areas, relative to their climate. This led to their understanding of what constituted, to quote URV's news page, "the optimum design for solar-assisted heating networks with a working life of 40 years".

According to the URV's literature, the study focused on urban districts with communities of 10, 24, 50, and 100 buildings; and, altogether, the diagnostic tool's result was conclusive: solar district heating systems are substantially more eco-friendly than traditional energy systems (i.e. those that are gas heater and boiler-based).

Further showing promising results, the large variation in tested community sizes even led to evidence that the environmental benefits of SDHS become increasingly clear, the more the technology is scaled up. To quote URV's news page:

"In terms of getting a return on the initial investment, the system is more effective the larger the community. For example, in a district of 100 buildings, the investment is recuperated within 13.7 years."


Solar Heat Europe's hopes for zero-CO2 grid.

A vision of the future (the year 2040): a diagram that represents Solar Heat Europe's hopes for a completely zero-CO2  grid, in which a solar district heating system (alongside a wind turbine farm) transmits its gathered renewable energy through overhead power lines and into homes and offices. Image Credit: Solar Heat Europe.


The Conclusions of the URV Research 

Despite the evident benefits of URV's machine learning-based diagnostics tool—particularly in terms of its showing the sustainability of SDHS—the challenges of rolling out solar district heating systems remain marked. To consider the infrastructural challenges alone, "technical, administrative, and financial barriers" remain prevalent by the URV's admission.

Nevertheless, the Universitat Rovira i Virgili's research suggests a promising future for Europe's energy management. As the scientists conclude in their Applied Energy journal: 

"[The research] can assist in proposing the SDHS as a competitive solution instead of the conventional heating systems. Subsequently, it can promote a clear statement regarding the new clean energy for all European packages."

Related Content


You May Also Like