Share this
by Ben Perry
An Analysis of Hurricane Beryl
Extreme weather events such as hurricanes and blizzards can significantly impact the electrical power grid, causing widespread outages and expensive repairs. Strong winds often knock trees into power lines, damaging the grid infrastructure. Blizzards can wreak havoc, with ice and heavy snowfall impacting power lines, while storm surges from hurricanes can cause flooding, harming substations.
However, the most common issue is trees affecting power lines. Wind, rain, snow, and ice can weaken trees by saturating the soil and adding weight to branches, leading to breaking or toppling. When trees or branches fall, they often land on power lines, causing outages and damage to the electrical grid.
As power producers work to restore and repair the grid, they need detailed demand and weather information to chart an effective response.
The Challenges of Forecasting Energy Demand During Extreme Weather
Although extreme weather events only happen a few times a year, they have a disproportionate effect on the power market data that participants use to keep the grid running smoothly. Power grid participants need to know what to expect during these unpredictable events.
Utilities and asset operators, for example, may struggle to accurately predict demand while lines are being repaired. Seeing demand forecasts at full capacity can help system operators estimate demand more reliably.
Traders, meanwhile, might choose to not participate in the power market during such periods due to greater market risk and uncertainty. Understanding when the system operator expects the system to fully function again can help traders know when to reenter the market.
Yes Energy®’s TESLA Demand Forecasts help power participants navigate the challenges of extreme weather on typically unpredictable days.
ERCOT Load Forecasting During Hurricane Beryl
On July 8, 2024, Hurricane Beryl struck Texas, causing widespread power outages for 2.7 million and leading to 36 fatalities. Although Hurricane Beryl had dropped from a Category 5 to a Category 1 storm by the time it hit Texas, the wind and rain continued to wreak havoc, toppling transmission lines and trees and complicating power restoration efforts.
Hurricane Beryl ERCOT Load and Capacity in Houston
The Houston region bore the brunt of the hurricane, with widespread flooding, downed trees and power lines, and infrastructure damage. With so many Electric Reliability Council of Texas (ERCOT) customers out of power, utilities rushed to restore it.
Yes Energy’s TESLA Demand Forecasts helped link the effects of Beryl’s damage to demand. Utilities knew how many customers were impacted, and what percentage of their total customers were without power. What they didn’t know was how much power customers would have consumed if there hadn’t been outages to the system.
Instead of adjusting our forecast based on estimated outages, we maintained our demand forecast as if the system were at 100% capacity. This approach gave us a clear view of expected demand under regular conditions, which helped utilities make more knowledgeable decisions while they repaired the grid.
In the image below, we use our TESLA’s Demand Forecasts to identify the gap between load (displayed in red) and true demand (shown in orange). The observed load (red) represents metered data where meters were operational and power lines were intact to deliver power through the meters. Lower load data doesn't represent lower demand from the full set of customers; it represents the demand from customers who weren’t impacted by outages. Our estimate of what customers would have consumed had there been no outages is represented by our one day-ahead (DA) bid-close forecast (orange).
Although the hurricane didn’t strike Texas until July 8, we can already see the pre-landfall effect on ERCOT in the days before. In the image below, we use our TESLA Demand Forecasting tool to graph ERCOT-Houston FZ actual usage (in red) against the Independent System Operator’s (ISO) historical forecast (in green) and TESLA ERCOT historical forecast (in orange).
The ISO’s DA forecast tracks the outage-impacted metered load pretty consistently after around July 9. Since ERCOT has a mandate to ensure reliable power delivery to customers, its load forecast should clearly outline its expectations for anticipated repairs to damaged power infrastructure.
Instead of trying to predict outage repairs more accurately than the organization responsible for them, we chose to maintain our demand forecast as if the system were operating at 100% capacity. By comparing our demand forecast to ERCOT’s, traders can track the ISO’s expectation for the megawatt (MW) impact of outages still on the system based on their repair schedule. As you can see below, once the red and green lines approach the orange line, a higher percentage of the grid is in operation.
By July 15 or 16, the grid is nearly fully restored, with demand slightly higher.
Source: Yes Energy’s TESLA Demand Forecasting solutions
Now let’s track the amount of total demand that wasn’t served due to storm-related outages.
The blue-shaded area in the graph below shows the difference between our bid-close forecast for ERCOT (representing demand without outages) and the actual recorded metered demand (affected by widespread outages).
Source: Yes Energy’s TESLA Demand Forecasting solutions
The graph below reflects the blue-shaded area from above as a percentage of our bid-close forecast, which estimates the "true" demand.
Source: Yes Energy’s TESLA Demand Forecasting solutions
Integrate Energy Demand Forecasting into Your Systems
Extreme weather poses a unique challenge for forecasting models. When hurricane season hits, utilities, asset operators, and power traders need to continue to be able to make strong decisions in spite of unexpected events. Our power demand forecasting tool allows you to make better decisions for you and your customers during such events.
Although our Hurricane Beryl example used public data, our energy demand forecasting is equally effective for utilities using proprietary data. In our 30-year history, we have extensive experience in modeling and forecasting for utilities, ensuring that the same high-quality forecasts we produce for the market can be applied to your specific customer base.
To learn how our demand forecasting services can support your needs, reach out to our team or request a demo to see this powerful tool in action.
About the Author: Ben Perry is the senior product manager of forecasting at Yes Energy. Ben brings 10 years of experience as a power demand forecast analyst at TESLA to the Yes Energy product team. He's now focused on applying that experience to steering the Yes Energy forecasting product roadmap to best serve the industry through the energy transition.
