Petri Rauhakallio, VP Customer Operations at Sharper Shape, shares lessons that could be learned from the US and how technology could supercharge the testing and inspection of power lines.
2021 was a turbulent year for the US power grid, where, in February of 2021, Texas suffered statewide power outages due to inadequately winterised natural gas equipment, and in Northern California, the Dixie fire caused one fatality, burned 963,309 acres and destroyed at least 1,300 buildings while expelling high volumes of carbon dioxide into the atmosphere. So, it stands to reason that 2022 will see North America’s power grid radically upgraded to shoulder more responsibilities and achieve greater resilience than ever before.
A major upgrade and extension of power lines will be needed to achieve the US target of decarbonising the grid by 2035. The US Infrastructure Bill provides $7 billion in funding and regulatory reforms to help enable flexible long-distance, high-voltage interstate power lines. The aim is to transform the US power grid into an interconnected ‘system of systems’, where electricity can be seamlessly moved between states in the event of local wind or solar power shortages. Long-distance transmission lines will also connect the grid to many more remote, rural green energy sources. This will also vastly expand the array of powerline infrastructure that requires inspection and maintenance.
At the same time, 2022 brings growing challenges to grid resilience. Last year, the average American home endured more than eight hours without power, according to the US Energy Information Administration – more than double the outage time five years ago. So, the pressure is on for utilities.
An extended challenge
The Infrastructure Bill’s funding to allow the upgrade of and extend transmission lines will require a parallel transformation of powerline inspection to monitor and manage a larger and more interdependent network. Reliance on intermittent energy sources will mean that energy providers will need to respond rapidly to provide power at great distances. What happens in one part of the grid will increasingly matter in other locations. An outage in one particular area could now affect not only that, but also its ability to play a role in the interconnection of other territories in the event of a solar or wind power shortage. Cumulatively, power lines must be inspected over greater range and with greater accuracy and frequency since each grid takes part in an even wider ecosystem.
Despite considerable advancements in technology, today’s powerline inspection methods are insufficient to meet the looming challenge.
Though new technologies are now emerging which could deliver an insurmountable boost to the scale, efficiency and accuracy of inspections, adoption remains a challenge for a traditionally conservative industry mired in old habits and a lagging pace of change.
From the ground up
Manual inspectors have natural blind spots and relying on ground-level human teams can create corresponding blind spots across the network. For example, something like a planned maintenance on part of the powerline system could have much wider repercussions across a future interconnected grid ecosystem. With advancing technology, and automated aerial inspections capturing a richer and wider scope of visual data than teams restricted to ground-based perspective, it is possible to understand the effects of these small impacts on the wider network and react more quickly.
This too is applicable to vegetation management. A human inspection of tree species and associated powerline hazards are limited. Whereas artificial intelligence (AI) can be fed vast vegetation species databases alongside tailored hyperspectral data to detect and prioritise maintenance of potential network-wide hazards at far greater scale and speed. Combined with thermal inspection and UV, this technology can outperform professional arborists and Qualified Electrical Workers (QEWs).
Going the distance
Current inspection systems are also unable to keep up with the vastly expanded scale of networks envisaged by the recent Infrastructure Bill. For example, ground crews following the conventional inspection method of ‘walking the line’ can only cover around 10 miles per day. Meanwhile, a helicopter could cover 100-250 miles a day, and BVLOS capable drones can offer similar coverage but with the added bonus of safety as there is no pilot in the cockpit and a better endurance level.
To put this into perspective, as a partner to utilities, Sharper Shape can inspect 25,000 miles of power lines in one month. Achieving this with conventional methods would take 24 full-time ground staff.
Conventional manual network inspection is also sequential and siloed with findings compiled after completion, rendering it cumbersome and slow to inspect vast networks. Live-feeding data from all inspection assets to a centralised Living Digital Twin would instead enable multiple features and factors to be examined in parallel, enabling inspections of vast networks and monitoring of the network risk model in real time. For example, when it comes to LiDAR technology, it takes just three minutes per mile of line corridor for AI inference.
AI also enables multiple factors to be considered in concert, from temperature and wind speed to the type, location and behaviour of vegetation near power lines. For example, data on high winds and warm temperatures could be matched with vegetation data to predict and prevent fires by pinpointing areas and locations of concern. Data from multiple sources such as drone sensors and weather satellites could also be fed into 4D living digital twins to create rich multi-layered maps of networks modelling the impact of future scenarios such as floods or droughts.
Dissolving data silos
With utilities faced with managing an exponentially expanding grid against a more diverse and fluctuating array of climate hazards, they need to integrate data from multiple sources. Just as the power grid will need to be more interconnected to prevent outages, utility data will also need to be more integrated.
Future technologies will facilitate sensor fusion where grid data from drones, helicopters or satellites are automatically integrated into a 4D living digital twin, providing rich, real-time oversight. Using this data-rich approach, network risk assessments and AI can be then used to better prioritise activities such as inspections and vegetation clearance. While many network maps are outdated and do not record recent upgrades or maintenance, digital twins would form a ‘living document’, recording risks, repairs or upgrades in real-time.
By helping to accurately pinpoint or predict hazards, damage or degradation, digital twins could enable inspection resources to be more efficiently targeted and drive ‘predictive maintenance’ of the grid. Algorithms can then be applied to inspection data to help prioritise more critical issues and grid locations supporting critical services or key customers, thereby enabling the allocation of the maintenance budget to focus on optimally removing prioritised risks.
Crucially, in the event of extreme weather events or blackouts, digital twins can identify priority sites for emergency services or repair crews. AI can also institutionalise insights from previous events, learning to associate specific ‘red flags’ with specific mitigation strategies. When this data is brought together by sensor fusion, AI algorithms will ultimately learn to predict and prevent blackouts, creating a virtuous circle where the grid grows more resilient from each hazard or fault.
New year, new techniques
With the continuous expansion of critical infrastructure, people and places will require inspections to cover a greater array of infrastructure and hazards at higher speed than ever before. This requires inspection capabilities to be dramatically scaled up and improved through automated aerial inspection and the harnessing of AI platforms to deliver faster, smarter, data-driven decisions. AI can inspect networks at greater scale and speed than humans, remove human bias and error from critical network data and enable proactive and ultimately predictive maintenance and resilience.