Making infrastructure more sustainable with AI

The views expressed are those of the author and do not necessarily reflect the position of FORESIGHT Climate & Energy

Adopting AI can help improve sustainability without costly reconstruction of infrastructure

The sustainability discussion needs to broaden from simply being about the environment to one that includes how all aspects of humanity can adapt to embrace sustainability. One surefire means of approaching this is by using technology to identify and address the issue—in particular, artificial intelligence (AI).

Indeed, it has been suggested that AI could enable 93% of the environmental targets defined in the United Nations’ Sustainable Development Goals (SDG). The figures speak for themselves: A building can last anywhere from 50 to 150 years.

In contrast, grid infrastructure such as transformers or switchgear can last for about 50 years, notwithstanding the time it takes to move a power grid project from planning and design to construction.

Unfortunately, today’s buildings do not always support a CO2-neutral footprint. Depending on where a building is located, it will require cooling or heating (or both).

ELECTRIFICATION GAINS

In recent years huge strides have been made to ensure that these buildings consume only electricity for operation. If we consider the heating of commercial buildings, for example, most of the required energy is electricity to drive ventilation.

To bring this into context, buildings globally account for 33% of primary energy consumption and 40% of energy resources, presenting a massive opportunity in terms of energy reduction potential. Added to this, around 75% of commercial buildings are not energy efficient, with energy usage being beyond the needs of the occupants.

AI-based energy management systems could save up to 30% of the energy lost through inefficient energy usage. One concrete example is that of office buildings in Malaysia, where a study concluded that 57% of office building energy is consumed for ventilation. The potential here for savings is massive and AI can be the driving solution.

The costs associated with transforming buildings into carbon-neutral entities are often debilitating and even new projects are not designed with reduction, reuse or recycling in mind. All of this comes at a time when regulations in most regions of the world are defining efficiency goals for buildings.

But luckily help is at hand in digital technologies. Using digital technologies to increase the efficiency of heating and cooling power within a building allows existing infrastructure to operate more sustainably. It is also much quicker to deploy than replacing an entire building.

STEP FORWARD AI

While many digital technologies are already being employed to enable sustainable infrastructure, we are now also seeing the increased use of AI. There is no denying that some AI technologies and applications still need to mature but others are now ready for scalable roll-out.

Thanks to the increased adoption of digital technologies in the form of a net of sensors, control systems and Internet of Things devices, we are seeing widespread use of AI for anomaly detection, predictive maintenance, quality inspection, condition monitoring and process optimisation. All of which can have a significant impact on a building’s performance.

One practical example that resonates far and wide is that of predictive maintenance. Indeed, a key example of the effective use of AI in existing infrastructure is its ability to, through continual monitoring, identify patterns that could lead to a breakdown in infrastructure. If you identify an emerging fault, teams can act before an issue arises. This avoids the need for costly remediation skills.

Real-time information from device and sensor data showcase trends that, when altered, can create the backdrop for a more sustainable building. While data is a blessing, it can also be a curse. The more digital one becomes, the more data you typically collect.

While organisations should be focusing on limiting the collection of data to that which is most beneficial to driving efficiency this often is not the case. The more data they collect, the more difficult it becomes to unpack meaningful trends and insights. That is unless they make use of AI to extract the detail needed in real-time and extrapolate the patterns and insights hidden in the data.   

AI AND SUSTAINABILITY

The goal is to move from fossil fuel electricity generation to a grid that supports renewable energies. While some countries can operate their entire energy and heat grid in theory on renewable power, many others are still centralised and only support the legacy way of managing power.

What is needed is an interconnected grid system that can switch between different energy supplies and intelligently respond to changing levels of energy supply. One key tool in ensuring that this is being done effectively is AI.

As the infrastructure we build becomes more interconnected, the need for AI-infused digital systems will increase. There will come a time when the complexity of these systems will defy human-mediated monitoring and control.

In the short term, building AI systems to support human decision-making will help make infrastructure more sustainable, but in the long term, it should be built to support a closed-loop system where eventually the AI will be able to act (within defined parameters) on its own and only be supervised by humans.

One such use case is that of overhead line inspections of an energy grid, where the challenge is to ensure high reliability of energy transmission. For this, quick and reliable detection and fault assessment are required. By employing deep learning on multi-sensor systems, you can generate digital models of the infrastructure.

The resulting model provides a reliable, efficient and zero-harm infrastructure tool for line monitoring and the savings should not be downplayed. Condition monitoring not only reduces maintenance costs but can also increase productivity by 15% and reduce the number of technical outages by up to 70%.

Making the transition to a low-carbon economy is one of the most difficult challenges on corporate agendas. While AI alone might not be the silver bullet, its role in addressing infrastructure sustainability cannot be underestimated.

Consequently, we need to create AI that is responsible, understandable and trustworthy. We must now foster an AI-ready culture, in terms of training and lifelong learning, so that we can fully harness its ability to augment the information people have, empower them to take better decisions and ultimately help to make our infrastructure more sustainable.

If you have a thoughtful response to the opinions expressed here or if you have an idea for a thought leadership article regarding an aspect of the global energy transition, please send a short pitch of 200 words outlining your thoughts and credentials to: opinion@foresightdk.com.