3 ways AI is improving grid resilience and paving the way for a cleaner future

As the world attempts to achieve ambitious renewable energy mandates, outdated grid architecture stands staunchly in the way, writes Jack Curtis of Neara.

The recent International Energy Agency IEA global study determined that the world must add or replace 80 million kilometres of grid infrastructure by 2040 to meet national climate targets and support energy security. For context, this is equivalent to doubling the planet’s current electricity infrastructure footprint.

Natural disasters striking with more impunity and frequency further compound these hurdles — exposing grid weaknesses, requiring costly rebuilds, and endangering people’s lives. As decarboniSation deadlines loom, it is now even more critical to keep grids online so they can play a vital role in integrating renewable energy.

And it is here that artificial intelligence can play a significant role, solving some of the electric utility industry’s most considerable challenges. Specifically, there are three ways we see AI drive grid resilience while accelerating the clean energy transition:

AI is redefining the potential of digital modelling by facilitating hyper-accurate real-life network representation

As digital twins become more entrenched in utilities’ network monitoring and optimisation processes, they must reflect the real-life network as accurately as possible. A table stakes accuracy threshold is contingent on high-quality data, and the key is leveraging a diverse spectrum of data sources that individually and collectively enable unique insights. For example, LiDAR’s accuracy is unrivaled, but satellite imagery is particularly effective in change detection.

The challenge is not only curating the right mix of data sources but, most importantly, codifying these disparate data sources into a single and unified digital representation that unlocks more insight than just a disparate sum of individual parts. Utilizing data sources such as LiDAR, satellite imagery, and GIS in isolation creates an incomplete, fragmented picture of network performance. However, combining them contributes essential substance to a dynamic model ensuring that utilities have the multi-layered context they need to make critical decisions.

Have you read?
The path to grid resilience: Coupling asset fragility and weather
UL Lafayette receives $87M to address grid resiliency

AI can calibrate disparate data sources in a single model that highlights the unique strengths of each one and allows them to compensate for weaknesses in others. This results in a well-rounded, unified portrayal of utility networks sufficiently accurate to drive critical decision-making and seven-figure plus grid hardening investments without verification from manual surveys. AI is enabling traditional digital modelling technology to graduate from a novelty to a core facilitator of network monitoring and optimization operations.

AI augments field surveys with unprecedented context and consistency

Even if utilities had field teams surveying every square inch of their network at all times, there are simply too many cost-and-effect relationships of varying degrees to capture and calibrate with the human eye between network assets and the surrounding environment at any one time. This is where AI takes center stage in providing a new standard of thoroughness and speed for network monitoring in pursuit of grid resilience.

Imagine a field survey playing out in one pocket of a network where a team is assessing pole integrity and marking poles for maintenance and replacement. Several miles away, vegetation contractors are flagging encroachment zones. Now imagine the time it will take for their findings to cross paths.

The team marking poles for replacement might determine that taller poles are better suited to that specific part of the network. Meanwhile, when the time comes to replace the existing poles with taller ones, the new taller poles change the tension on the line. The change in line tensioning ripples through the whole network and causes the cables in the area where the vegetation team just flagged encroachments to sag more than they did during the recent survey. As a result, the vegetation team is now working with siloed, outdated information even though they just did the survey — and that’s just in an as-surveyed context. Things get even more complex when utilities also need to consider how things may change under various environmental conditions such as high winds or high heat.

With the analytical horsepower of AI, a utility’s digital network model becomes a well-oiled machine that can capture all of this at once, conducting non-linear analyses that reveal weaknesses and opportunities that might otherwise take years to surface.

AI’s core pattern-matching strengths represent a critical unlock for the clean energy transition.

One of AI technology’s core strengths is identifying and comparing aberrations from established standards. As such, AI is especially well-suited to identifying risks such as overloaded poles and clearance violations.

In the context of the clean energy transition, AI’s ability to highlight discrepancies at scale is proving especially powerful as utilities can increasingly count on the assets they already have to bring more renewable energy online. Until recently, the high cost and lengthy timelines associated with the clean energy transition presented formidable blockers to moving the needle on clean energy mandates. While new infrastructure continues to be a critical part of the industry’s decarbonization roadmap, AI-based digital modelling technology is helping utilities identify and put to work latent capacity in their existing network via digital line rating.

Traditional line rating methods evaluating capacity are still largely manual, with engineering teams being sent into the field to record, compile, and analyse data from each stretch of the network. This process is both time and labour-intensive, costly to undertake, and cannot provide the level of detail or visibility needed to accurately assess true network capacity. These processes and the overall lack of evolution have caused utilities to act with caution to avoid overloading the network with dangerous levels of current. This has ultimately prevented networks from accessing the benefits of clean energy generation.

With the help of AI, utilities can now digitally re-rate individual spans across their networks at scale based on varying span-by-span clearance nuances. Owing to AI’s ability to process terabytes of ambient rating and clearance data simultaneously and calibrate them across various conditions, AI-enabled digital line rating is emerging as the dark horse of the clean energy transition.

About the author

Jack Curtis is the chief commercial officer and co-founder of electric infrastructure software company Neara.

Curtis is an operations and sales executive with 15+ years of experience in power and utilities across the entire commercial, project and financial value chain in 20+ countries.