Machine Learning Gets the Grid Ready to Weather Storms

This content was originally presented at Clarion Energy Enlit Africa. View the Storm Readiness segment from this virtual event.

Extreme weather is on the rise. The number of named storms in North America has grown from 17 in 2017 to 30 just three years later in 2020. Those storms bring injuries and fatalities, business and travel disruptions, flooding and property damage, and power outages. Utilities looking to reduce the impact of severe weather events must shift from restoration to preparation. GE Digital’s state-of-the-art Storm Readiness technology makes that possible.

Transitioning from restoration to preparedness

Think of all you could do with the ability to more accurately predict severe weather events and what this might mean for your customers and network assets. Utilities with data-driven capabilities to improve storm readiness can reduce:

• CMI/CAIDI *
• Avoid costs from canceling planned work & unnecessary moving of outage response resources
• Provide customers and the public with more accurate information that affect their day-to-day lives
• Avoid excessive outage response overtime labor costs and ramp down storm resources faster

Instead of storm responsiveness, Storm Readiness technology allows the utility to shift from restoration to preparation.

What Storm Readiness at the edge looks like

It all begins with the ability to more accurately predict severe weather events. With access to GE Digital’s Storm Readiness weather layer forecasts, utilities can prepare in advance.

The power outage prediction modeling also draws on the utility’s historical data, such as what caused previous outages at particular locations. Differentiating between a storm-related event, for instance, and a vegetation-related incident in the Outage Management System (OMS) can help the system to work optimally to predict with the most accuracy. Offering the ability to see up to 72 hours in advance, the application enables the operations team to make critical decisions ahead of the anticipated weather event. Instead of reacting, the utility can anticipate the weather event, where the grid will be impacted and what types of service disruptions could happen.

This offers several benefits:

• Proactive cost control by avoiding excessive outage response and overtime labor costs
• Improved restoration times via more accurate staging and dispatch of first responders field crews
• Improved customer satisfaction

This data-driven knowledge also improves health and safety. With the awareness offered by the application, field crews can be kept off the roads when winds, rain or other conditions are at their worst.

Always improving storm readiness

An added benefit of digital technology in Storm Readiness is the ability to always be improving. The platform itself is configured for continuous learning from each weather event to increase the accuracy of prediction.

The machine learning approach sees the storm readiness engine ingesting the weather inputs along with data from asset details from the utility’s Geographic Information System (GIS) and historical outage data from the Outage Management System (OMS). Then, accounting also for temporal factors such as the season the analytic engine maps out a storm readiness response.

The utility’s predictive outage reports include useful insights regarding the location of the impact, assets involved, and anticipated duration. Repair recommendations from the engine can detail the number and type of crews needed, and the recovery time. Better still, the more information the software ingests, the more power and accuracy it has to offer. Integrating data from other areas of the utility can further strengthen algorithmic success. For example, GE Digital’s AI-enabled Visual Intelligence (VI) software can quickly ingest data from either field or aerial inspections, identify assets and recognize changes that require attention.

Service disruptions are, regrettably, inevitable. GE Digital’s software offerings help the utility to train its people and prepare for the worst. The machine learning tools offer access to a playback feature allowing users to go back in time and leverage past experiences to continuously improve and make more informed decisions going forward. Some utilities today, based on their geography, don’t expect a lot of weather-related incidents. Yet, factoring in the changing global environment, having a weather-related outage response plan in place makes sense.

Additionally, with growing momentum for distributed energy resources (DERs), the utility that may not have had to consider the impact of storms in the past may become more attuned to weather conditions. GE Digital’s data management platform helps the utility to know and understand its network as the number of devices connected to the grid grows, notably thanks to the extensive use of the most advanced protocols such as IEEE 2030.5, which enables to affordably collect data from and send controls to the field over the internet in a very secure manner, connecting even the smallest smart device. Instead of waiting to harness the power of the available data when the worst happens, it is better to build out the visibility in advance.

Collecting data has grown more affordable and efficient for utilities. The important thing is to make the data meaningful while being able to share it across the organization to understand whether an outage was directly or indirectly weather related. This means thinking beyond a tree being blown down by extreme winds. Utilities must also consider service disruptions caused by an asset being extremely hot for too long or some other indirect weather factor.

Certainly, five or ten years of accurate historical data points with context will help. But the machine learning draws not only on what your utility knows already, but also on GE's Digital experience.

Preparing the grid for the future is founded on digital transformation. This article has focused on Storm Readiness, but view the Clarion Energy Enlit Africa webinar to learn more about this offering as well as GE Digital’s Visual Intelligence inspection data management and DER orchestration tools. Ultimately, whatever your challenge in improving asset management and fortifying the grid, GE Digital can help.

Watch the full virtual presentation Grid at the Edge of the Storm from Clarion Energy Enlit Africa.

*CMI/CAI