Gridium launched a coincident demand forecasting service this summer in Texas to help commercial real estate owners manage the “4CP” demand charges that drive a significant portion of their electricity bills. Summers in Texas can be a wild ride, and this year was no exception, featuring high temperatures, heat waves, and some unexpected twists.
A Refresher: How Our 4CP Alerting Service Works
Coincident demand charges are costs imposed on energy users based on their electricity demand at the time of the overall grid peak. In this case, the grid in question is the ERCOT system in Texas, although coincident demand charges are also common in New England and the Midwest. ERCOT levies charges for four coincident peaks every summer (hence “4CP”), one each in June, July, August, and September.
Gridium provides a forecasting and alerting service to help customers manage these 4CP costs. The service continuously analyzes energy demand, weather forecasts, and ERCOT grid conditions to assess the probability of a grid peak. When a peak demand period is predicted, building operators receive timely alerts allowing them to take proactive steps, such as adjusting HVAC loads, reducing lighting, or shifting non-essential operations.
The Summer in Review: Hot…and Spiky
First, let’s review how the weather played out in ERCOT territory this summer. Summers in Texas are always hot, but this year the patterns were somewhat unusual.
- June: an end-of-month heatwave brought high temperatures to Texas unusually early, culminating in an unprecedented weekend peak on Sunday, June 30.
- July: June’s heat wave spilled into July, resulting in a grid peak on July 1 (as we predicted!). The rest of the month was then relatively calm, aside from a massive temporary drop in energy use due to a power outage caused by Hurricane Beryl.
- August: Steadily rising temperatures drove a grid peak on August 20.
- September: Temperatures dropped heading into fall, but the middle of the month was substantially warmer than the beginning of the month, complicating forecasting. Nevertheless, we correctly predicted the peak on September 20.
The red dots represent ERCOT’s four coincident peaks of 2024. These peaks can dictate more than 10% of your energy costs for the year.
Results
Overall, our alerting service predicted 2 out of the 4 peaks in ERCOT this summer. This means that over 50 building operators were able to get advance notice and take action ahead of potentially costly spikes in energy usage. Several of them shared how Gridium’s alerting service made a difference during the 2024 summer peaks:
- A Class A office building in Houston used Gridium’s alerts to preemptively reduce cooling loads during key peak periods, reducing their annual demand charges.
- A corporate office complex outside of Dallas utilized the alerting system to adjust their energy usage across multiple buildings. By responding to the alerts, they successfully avoided major spikes in their demand charges, saving thousands in energy costs.
- A property management firm in Houston shared that Gridium’s alerts allowed them to balance tenant comfort with energy efficiency, as they made strategic adjustments without causing tenant disruptions during peak alerts.
Key Learnings
Our team is constantly working behind the scenes to refine our machine learning algorithms to deliver as much value as possible to our customers. Some things we learned from this season that will enhance the alerting service going forward:
- Peaks can occur on weekends, too. Historically peaks only occur on weekdays. Our forecasting model actually correctly predicted the Sunday peak in June, but we suppressed the alert because we didn’t think it’d be possible. Lesson learned, in the future we will trust the model.
- Forecasting is a balancing act. Our goal for the service is not just to correctly predict peaks, but to also to minimize false warnings on non-peak days. The tension between these two goals hit us in August when once again we held back a warning on a day that turned out to be a peak. We adjusted our thresholds accordingly, but forecasting remains an art as well as a science.
- Coincident demand costs can be managed. Even though coincident demand charges make up a substantial portion of electricity costs in Texas, they can feel about as controllable as the weather. But, forearmed with knowledge of when peaks are most likely to occur, building operators can implement simple strategies to keep these costs down.
Looking Ahead
As we reflect on the 2024 season, it’s clear that data-driven, predictive tools like Gridium’s 4CP alerting service are critical for building operators looking to stay ahead of costly demand charges. Gridium’s algorithms, backed by machine learning, will continue to evolve as we prepare for future summers, and we look forward to using them to help you keep those energy costs down in Texas and beyond.
Want to get on the list for Gridium’s 4CP alerting service for 2025? Contact us.
Contact Us
