Ctg sit23 hub banner.jpg

The role of AI in sustainable offshore energy production

By accelerating the adoption of AI in the offshore renewable space, energy companies can dramatically reduce CO2 emissions and crucially reduce costs

Overreliance on fossil fuels throughout the last century has drastically increased the concentration of atmospheric carbon dioxide (CO2) leading to rising temperatures and extreme weather across the globe. Transitioning towards cleaner renewable energy sources is essential to the decarbonisation agenda and efforts to prevent a global climate crisis.

Winds of change

Offshore wind is one of the sustainable sources of energy that governments and climate activists are pinning their hopes on. In the UK, Boris Johnson has stated his ambition for the country to become the ‘Saudi Arabia of wind power'.

Rapid construction of new wind turbines and expansion of existing wind farms is a key pillar of his government's energy strategy. However, despite the fast growth of offshore wind sector - generation capacity is predicted to soar from 35 GW to 234 GW over the next 10 years - the energy transition business is not without its downsides.

Connecting the UK's wind farms to the mainland is an extensive network - sometimes up to 500 km - of underwater cables that, along with the foundations and all the other components of the turbines, need to be monitored, maintained and repaired occasionally. Doing so requires immense resources, both in terms of finances and manpower. Large vessels used by the energy companies to maintain wind farms can be crewed by up to 60 people, from engineers and specialist equipment operators to cooks and cleaners. They can also cost anywhere between £1 million and £10 million a month to operate all while using a huge amount of fuel - meaning that over their lifetimes carbon emissions can be upwards of 275,000 tonnes.

The work that the crews of these vessels undertake is crucial to the upkeep of offshore wind turbines, but innovation is desperately needed to reduce their environmental footprint. Fortunately, advances in technologies such as artificial intelligence (AI), machine learning and robotics are enabling energy companies to make increasingly important strides towards sustainability.

Never send a human to do a machine's job

AI has long been heralded as a transformative technology in many industries. In offshore energy production, there are signs that it could be game-changing. Marine robotics are already improving efficiencies and opening up new, more sustainable ways of working in offshore environments. However, advanced technologies such as simultaneous localisation and mapping (SLAM) and increasingly autonomous underwater vehicles (AUVs), still not widely adopted, represent the next step for a more sustainable and efficient industry.

In order to maintain wind farm assets, energy companies need to conduct surveys of the seabed to assess any structural damage - or erosion that could ultimately lead to failure if left unattended - and to identify any objects that could pose a threat to the integrity of the cables if they were to wash up against them. Typically, these surveys are conducted using sonars from manned vessels that map the sea floor. If something is identified as a potential risk, manually operated ROVs are deployed to collect video footage.

This is where lack of efficiency and requirement for manpower starts to become restrictive. Each ROV requires at least two pilots, and once the data has been collected, an additional team is required to manually inspect the video footage. The more people needed, the bigger the ships and the greater the carbon emissions. It is a situation that would benefit enormously from more autonomous technology.

Autonomy allows the information collected by SLAM vision systems to be analysed in real-time, presenting alternative options to the operator so that they can navigate obstacles or correct course to account for tidal currents. Without it, the ROV would simply run a predefined path that the operator would follow. With autonomous technology, pilots can also be located on shore in a purely supervisory role.

Greater autonomy means fewer pilots, and fewer pilots mean smaller vessels and less CO2 pollution.

AI and data

A typical inspection of subsea assets generates a vast amount of data. Expeditions can last for up to three months and video footage is captured in high, 4K definition from multiple cameras. On top of this, there is data from the multibeam sonars to process, as well as every other data stream on the vessel and the robots, of which there can be more than 30, updating hundreds of times per second. Analysis of video footage can take hundreds of hours, and manual interpretation of potential risks requires tens of people to be onboard for the duration of the voyage.

The sheer volume and size of the data collected by the ROVs is also problematic for energy companies as it is too large to be transmitted via satellites. As such, data analysis is conducted onboard vessels to determine whether there are any urgent issues that need to be addressed. The longer the analysis takes, the longer vessels must remain at sea. Through machine learning, video footage can be analysed for abnormalities and then categorised, enabling human operators to quickly make an assessment of any potential risk to the turbine infrastructure. It helps to significantly reduce the time needed to analyse the footage and eliminates the need for additional crew members to be on the vessels at all. Instead, the work to be done from the shore, lowering the costs significantly and preventing needless carbon emissions.

Lessons learned from under the sea

By accelerating the adoption of AI in the renewable energy transition space, energy companies can dramatically reduce CO2 emissions and crucially reduce costs, making it easier, cleaner, and cheaper to move closer towards a net zero economy.

The work being carried out on offshore wind facilities is certainly something that other industries can take stock of. In time, autonomous technology, along with machine learning analysis and SLAM systems, can be applied to almost any robotics. It can be utilised in any environment, from the deepest sea trenches to hostile and unreachable environments such as nuclear facilities, in the air using drones, or even on distant planets. This technology has huge potential, but before we deploy it on Mars, we must first tackle the most pressing issues facing us and our own planet.

Brian Allen is CEO of Vaarst