Using AI for Grid Maintenance, Monitoring and Resilience

Most people will not replace a car battery until the vehicle doesn’t start. At a much larger scale, utility companies have typically replaced equipment like circuit breakers only when they malfunction or when the equipment has passed longevity limits suggested by the manufacturers. 

 

AI technology can enhance grid management. Image used courtesy of Avangrid

 

Now, AI and machine learning are transforming this reactive approach to grid maintenance and utility management. Proactive maintenance that relies on data science is the new frontier.

Avangrid, a leading sustainable energy company,  is innovating how AI and machine learning can protect customers’ access to power and ensure maximum operating efficiency for utility companies. With subsidiaries such as Central Maine Power, New York State Electric & Gas, Rochester Gas and Electric, and United Illuminating, Avangrid’s evolving approach to grid maintenance has a wide reach with a 2.31 million customer base. It will also create a model for how other utility companies can deploy machine learning. 

Within this context of grid management, AI can use data to build a complex and nuanced approach to preserving the system’s health based on a wide selection of variables, including manufacturer specs, use frequency, local conditions, maintenance necessities, and more. 

The fail-and-replace approach to equipment maintenance will be a thing of the past as AI learns increasingly sophisticated techniques for amassing data and deploying targeted directives based on machine learning analysis.    

 

Grid Reliability and Climate Change 

Given current climate change challenges, this innovative introduction of AI into the grid maintenance space is more important now than ever. The longitudinal data in the figure below captures the fact that for 28 years, power grids are failing with an increasing frequency due to extreme weather. 

 

Power grid failure frequency and weather elements. Image used courtesy of Climate Central 

  

Such failures are not just a matter of convenience or quality of life, though compromising comfort, like access to temperature control, can hinder the continuity of basic needs, such as ensuring schools remain operational. And the most devastating death tolls reported from power grid malfunctions are critical when making strides in monitoring power grids. But perhaps most surprising is the fiscal drain of the state and federal funds needed to repair damage as extreme weather and natural disasters increase in scope and frequency. 

If this trend continues, the work Avangrid is doing to transform grid resilience will become a necessity to preserve both community safety and fiscal health. 

 

Equipment Maintenance and Localized Mapping

Avangrid is developing three projects that collectively focus on equipment maintenance. They are carefully mapping local area characteristics and providing widespread monitoring through capturing and analyzing photographic images of critical infrastructure elements. These three projects, Predictive Health Analytics, GeoMesh, and HealthAI, show the burgeoning potential of how a sustainable energy company can use AI to service utility clients. 

Predictive Health Analytics and HealthAI are both focused on equipment maintenance. HealthAI uses a complex system of photography to capture images of poles, wires, and other grid equipment to detect vulnerabilities in the system before they lead to an outage or other dangerous malfunctions in the community. GeoMesh, on the other hand, uses sophisticated localized mapping techniques to forecast system performance by compiling and “analyzing millions of data points, such as average wind speed, precipitation type and amount, outage history and reason, population and density of tree limbs and other vegetation,” according to Avangrid.

These projects create significant potential for improving safety, protecting continuity of service, and reducing risk when it comes to unexpected power grid failures. The projects include incentives for companies as motivation for industry adoption. Retail electricity providers (REPs) could save up to $3 per megawatt hour if they adopt AI tools. 

The advent of AI has disrupted many industries, and its potential leaves a lingering uncertainty, but for the electrical utility sector, AI provides an exciting new path toward innovation that will be developing for years to come.