Weathering the storm: AI's role in grid resilience

Australia’s electricity network is struggling against the storm brewing over the energy industry. Aging infrastructure, workforce shortages and increasingly volatile extreme weather events are battering energy infrastructure as it manages the complicated shift to integrating increased renewable generation.

The result? Growing concerns around energy reliability, increased network management complexity, and greater risks to the safety of utility field workers and local communities.

Extreme weather events are causing an uptick in the volume of repairs and maintenance required across the grid. Take Victoria as an example — according to AusNet Services¹, the owner and operator of Victoria’s transmission network, its aging infrastructure is under immense duress. One in seven of Victoria’s 13,000 electricity transmission towers are damaged by extreme or extensive rust, with more than half now set to reach the end of their 70-year service life within the next decade. The impact of volatile weather conditions on this aging network has also caused the number of downed transmission towers to increase over the last 15 years, with more extensive repair work required in dangerous conditions.

These conditions are only getting worse. ‘One-in-100-year’ floods, tropical-strength winds, extreme heatwaves and bushfires are fast becoming regular occurrences. Different methods are urgently needed to enable engineers, maintenance staff and SES workers to manage grid repairs while maintaining energy security for consumers and their own safety. Legacy reactive measures to these events can only provide so much security against the quickly escalating climate threat. We need new tools and emerging technologies, including artificial intelligence (AI), to help reduce the time network engineers spend in the field during increasingly volatile events while preserving the grid.

Granular insights and comprehensive risk evaluation

Large urban network model. Image credit: Neara.

AI-powered tools are unlocking new levels of visibility and predictive capabilities for utilities. By identifying risks sooner, coordinating remote elements of network maintenance and modelling future weather events before they hit, AI capabilities can help form a new foundation of risk-management strategies. The urgency behind this modernised approach is only growing as manual inspection work struggles to keep pace with shifting climate and environmental patterns.

When paired with digital modelling capabilities, AI analysis can provide unparalleled visibility and oversight of electricity networks by creating hyper-accurate digital models of existing infrastructure. The leveraging of data sources including LiDAR, geographic information system (GIS) mapping and severe weather area protection zones is allowing international utilities like Scottish Power Energy Networks to deploy AI-powered weather simulations and asset health indicators. This provides extreme visibility into the state of the grid, allowing engineers to gauge the structural health of equipment and its ability to withstand future severe weather events before they even enter the field.

By aggregating data sources into enterprise-grade 3D modelling technology with advanced AI capabilities, utilities can create hyper-realistic environments for remote inspections and simulation analysis. This approach can transform inspection strategies from convenience-based to risk-based. Field visits will be targeted to the most pressing risks and correct locations, reducing time spent in the field and enabling utilities to perform urgent works faster.

Shifting from a reactive to a pre-emptive approach

Dynamic line rating transmission. Image credit: Neara.

Beyond grid risk analysis and improving infrastructure visibility, AI can help provide the foresight required as extreme weather events continue to hit Australian cities and towns with alarming regularity. Predicting how an oncoming storm, flood or fire might behave and impact the network is critical to optimising response plans. This will ultimately keep communities connected for longer periods, ensure accelerated reconnection post-event and help keep response crews safe. Engineers can leverage AI-powered digital models to model extreme conditions in a safe, virtual environment before or as they arrive. This simulation provides the best insights to direct their responses and prepare for infrastructure damage.

During an extreme flooding event in South Australia, thousands of citizens living and working along a 650 km stretch of the River Murray were severely impacted by the weeks-long event. The rate of rising floodwater was too fast for field crews to manage, with water levels quickly breaching safe powerline clearance zones. SA Power Networks, the state’s largest energy provider, needed a new approach to manage the scale of the event. The company chose to leverage AI and LiDAR data to create a three-dimensional map of the entire River Murray region, including all its assets and their surrounding environment. This was SA Power Networks’ first use of AI-powered modelling capabilities and dramatically shifted its decision-making process throughout the flooding disaster.

Network scenario modelling on different electricity distribution network assets at various flood levels was performed, allowing SA Power Networks to pre-emptively determine, without field analysis, when and where powerline clearances would be breached and require disconnection. During testing, SA Power Networks found that the model was within centimetres of accuracy to reality, and used the map to disconnect and reconnect power to its assets at the precise moments needed. This modelling also allowed customers’ electricity supply to be reconnected ahead of initial timeframes.

In some cases, disconnected regions were re-energised within five days compared to the projected three-week timeframe using traditional extreme weather event processes. This not only impacted the safety of citizens and critical services, with restored powerlines providing critical services like heating, lighting, internet access and telecommunications, it also prevented SES crews and engineers from being sent into the field and dangerous flood zones. In previous events and without the use of modelling, they would have continuously physically measured powerline clearance zones.

The intensifying impacts of climate change are exposing communities and power engineers to increasingly dangerous and unpredictable conditions. Ensuring their collective safety while simultaneously maintaining grid reliability is becoming a formidable challenge. By leveraging advanced technologies like AI to enhance network visibility, simulate potential weather impacts and enable remote operations, we can substantially reduce OHS risks for field crews and build more resilient energy infrastructure for the future.

Jack Curtis, Neara co-founder and CCO.

_{1. AusNet Services, AMS 10-77 Transmission Line Structures: 2023–27 Transmission Revenue Reset.}

Top image caption: Unlocking network capacity. Image credit: Neara.