How AI in data centres helps drive a more sustainable world

Data centres are the backbone of today’s digital world. But with research suggesting that data centres are now responsible for between 3-4% of global greenhouse gas emissions, innovative solutions will be needed to solve the climate challenge.

With a history spanning more than 170 years, today Schneider Electric provides energy and automation solutions for efficiency and sustainability. Employing more than 140,000 people globally, Schneider combines world-leading technologies with real-time automation software, and services into integrated solutions for homes, buildings, infrastructure, industries and data centres.

As Marc Garner, Senior Vice President of Schneider’s Secure Power Division in Europe, explains, sustainability represents a key topic for the company and for the data centre industry as a whole.

“Over the last 10 years in particular there has been a real focus on sustainability, challenging both ourselves as a business and our customers to become more sustainable,” he comments. “How do we work with our customers, and with the community, to drive sustainability and provide the knowledge and insight that we have and the tools and analytics to ultimately green our planet?”

As Garner describes, there have been significant advancements in terms of how data centres are using energy over the last 10 or 15 years. “Compare where we were in 2007, for example, to where we are today, you see a reduction in power usage effectiveness (PUE) from around 2.5 to about 1.55 PUE,” he comments. “You see some data centres out there now that operate at around 1.1 or 1.15 PUE. That is incredibly efficient in comparison to where we've been.”

As Garner explains, the data centre sits in an interesting area when it comes to the sustainability conversation. “When we talk about sustainability and how bad data centres are within the sustainability agenda, only 1% of the energy consumption worldwide comes from data centre use,” he says. “Around 3-4% of carbon emissions worldwide come from data centre use. 

“It's still a problem,” he clarifies. “The numbers are big enough for it to be significant, and it’s important that we continue to address it and challenge ourselves to move forward. But if 99% of energy is being used elsewhere, how do we start to address those industries and areas to try and reduce carbon output and energy usage?”

This, he describes, is where the data centre market really comes into play. “We talk about Electricity 4.0 as a vector to decarbonise and reduce energy across the world,” Garner says. “To do that you’ll inevitably need more digital infrastructure, more data-driven and software like artificial intelligence (AI), to allow us to use data in a more effective and more efficient way. You also need to look at procuring green renewable energy and to fast-track the transition e away from fossil fuels. 

“However what I think is missed through a lot of this conversation is that if you don't have the data, you can't drive efficiency in that other 99% of industry,” Garner describes. “You can't drive AI, automation and all the great things that come on the back of it. And you need the data centres to hold and process the data to be able to support it.”

AI can help save huge amounts of energy

In data centres, AI solutions can help manage the world’s rising data demands, alongside ever-ambitious sustainability targets. Combined with automation, this technology can be a powerful tool.

“My seven-year-old daughter refuses to turn a light switch off in the house,” Garner laughs. “Human behaviour is fallible. One person does things one way, the other person does things another way. 

“You can't get consistency if you're trying to run a building or a manufacturing plant based on human behaviour. What you can get consistency on, is through automating that system. So in my house, if I put PIR sensors in every room. Now, when my daughter leaves the room, that light automatically goes off. You're less reliant on behaviour, and you're more reliant on technology and automation.

“If we move on to something like an industrial facility where machine learning is introduced, a variable speed drive running at 80% compared to at 100% can save a huge amount of energy in that simple change. AI can really help with that.”

Crucial to this approach’s success is making sure that organisations have access to the right data. “If you've got the right data coming into it to say ‘our capacity to produce is reduced, we don't need to run the machine at 100% anymore,’ your consumption goes down significantly. You start to see optimisation throughout the whole process. And ultimately that will be fed by a data centre into the cloud, stored and managed for analytics to be able to compare and learn from it and do better the next time around.

“I think that's where you start to see this bigger picture of where the data centre plays into driving operational efficiency,” Garner adds. “Around 60% of the reduction will be around energy efficiency, and 40% will come from how we produce or generate energy differently. Once we've got that energy, we've still got to do better with it than what we're doing today. And that's one of the interesting aspects for me where the data centre really comes into it.”

AVEVA acquisition providing a true end-to-end view

In January Schneider Electric announced it had completed the acquisition of global industrial software leader AVEVA, which offers a Digital Twin for the whole asset lifecycle from engineering through to operations and maintenance.

By connecting people with trusted information and AI-enriched insights, AVEVA enables teams to engineer efficiently and optimise operations, driving growth and sustainability.

Its solutions integrate AI technology and machine-learning capabilities, enabling a constant state of self-optimisation through autonomous and semi-autonomous processes and giving users the agility to respond to changing conditions in real time for greater safety, profitability, and sustainability.

As Garner explains, the acquisition of AVEVA is one that helps provide an end-to-end view, enabling more efficient and sustainable operations.. 

“That's really where the AVEVA platform and the other platforms that we have really come into play,” he says. “A facility risk manager is very interested in the facility data, the building management control, the electrical distribution reporting. You have an IT manager that wants to see the IT stack and the IT infrastructure. You have someone who's working in security that is looking at the security stack. 

“But what you don't regularly get is all those systems talking together and working in unison,” Garner adds. “And as you start to see companies and buildings start to work and become more integrated with IoT and the IT and OT integration of a building, these things start to converge. The systems we have aren't always the best to bring them all together and give you a complete holistic view of what the world looks like, and this is where AVEVA really comes into play.

“Now, we're able to take that view, make it relevant for a facility's manager or a CEO or a sustainability manager or maybe for communication externally, and do it all under a single pane of glass. Bringing all those layers of control and software systems and everything underneath it into one view allows you to drive efficiency and operate your business in a much better way.”