Net Demand energy forecasting is important for competitive market participants, such as the Electric Reliability Council of Texas (ERCOT) and similar markets. For example, accurate predictions can help predict when a supply demand imbalance will create price spikes or crashes, allowing traders and generators to optimize their bidding strategies. It is also important for asset optimization. Generators need to know when and at what price level to commit resources to the market. A lack of forecasting can lead to missed opportunities for profit or operating assets if prices do not cover costs.
The Elcott area in particular has large wind and solar capacity. Net demand forecasts (subtracting renewable generation from aggregate demand) can help predict when traditional power generation is needed to fill the gap from variable renewable resources. Market participants also use forecasting as a risk management tool. Accurate forecasts allow participants to hedge their positions through bilateral contracts or financial instruments, protecting them from unstable market conditions.
Forecasts, on the other hand, can provide insight into operational planning. Making market forecasts for up to 15 days can help managers with unit commitment determination, maintenance scheduling, and resource allocation across a portfolio of generation assets.
In Texas, competitive energy-only market design makes it even more important to forecasting as capacity payments are not available. This will only generate energy and earn you money. The state’s isolated grid, extreme weather events, and high renewable penetration provide more challenging and financially and consequently accurate forecasts than many other markets.
Fortunately, AI (AI) is now able to generate highly accurate predictions from the increased availability of meters and weather data. The complex and robust calculations performed by these machine learning algorithms go far beyond what human analysts can do, making pre-prediction systems essential to utilities. Additionally, they make decisions about their position in the wholesale market for independent electricity producers (IPPs) and other energy traders.
Sean Kelly, co-founder and CEO of Amperon, which provides AI-powered prediction solutions, said that using Excel spreadsheets as a prediction tool was not working properly when he started his business as a power trader in 2005. “Now we’re literally running behind the scenes with Amperon 4-6 models, and there are five different weather vendors who run the ensemble each time,” Kelly said as a guest on the Power Podcast. “So as it gets more confusing, we have to stay above it. That’s where machine learning really starts.”
Wholesale Prices Can Criminal Retail Power Providers
The consequences of not being prepared are disastrous. Having early and accurate predictions means a difference between a business that survives or fails. Effects from Winter Storm URI provide a suitable case.
Ercot wholesale prices typically fluctuate from about $20/mwh to $50/mwh. During Winter Storm URI (February 13-17, 2021), Elcotte set the wholesale electricity price at the $9,000/mW cap due to extreme demand and widespread generation failures caused by storms. This price was valid for approximately 4.5 days (108 hours). This 180x price rise had a devastating economic impact across the Texas electricity market.
The financial fallout was serious. Several retail power providers have gone bankrupt. The most notable energy was to pass the wholesale price directly to the customer, receiving an invoice of over $10,000 for just a few days of electricity. The Brazos Electric Power Cooperative (the largest and oldest electric cooperative in Texas) is covered for Chapter 11 bankruptcy protection after facing a $1.8 billion bill from Ercot. Rayburn Electric Cooperative faced more than $1 billion in energy costs during the storm. CPS Energy (a local government utility at San Antonio) offered Ercot at an overprice, facing $1 billion in storm-related costs.
“Our clients were extremely grateful for their work at Amperon,” recalls Kelly. “We had probably a dozen clients back then, so we told them this was coming on February 2nd,” he said.
With that early warning, Kelly said that Amperon’s client was able to get out ahead of the price swing and buy power at a much lower rate. “Our predictions will come out for 15 days, and Elcotte’s predictions will go out for only seven,” Kelly explained. “So we said to everyone, ‘Alert! Alert! This is coming!” Our in-house meteorologist, Dr. Mark Shipham, was screaming it from the rooftop. That’s why we had many clients who bought $60 in power per megawatt. So, consider purchasing for your 60s. That way, your chances will be 9,000. So a lot of traders made money,” he said.
“Loading all LSES – Serving Entities – The Strils got really bad, but they didn’t get too bad,” Kelly continued. “I remember one client saying, “After buying electricity at age 60, I bought it at 90, I bought it at 130, I bought it at 250. And they said, ‘That’s the best expensive power I’ve ever bought. I was able to maintain my company as a retail energy provider.” And so, these are just some of the ways these predictions can be very useful. ”
Changes have been made, but accurate predictions are still essential
Following Winterstormuli, the Texas Legislature passed the bill, allowing utilities to securitize URI debt through bonds supported by fee payers, spreading costs for decades. It may have saved some companies from bankruptcy, but it did not rule out the financial burden.
Some city-owned utilities received financial support from local governments. Many cooperatives and other utilities have ultimately handed costs to their customers due to growing rate increases over the years. The crisis has revealed a major vulnerability in ERCOT’s market design. In particular, how financial risks were allocated during extreme weather events led to regulatory reforms on weather resistance requirements and market rules.
Still, accurate forecasting remains essential for the electricity industry. With more and more renewable energy added to the grid, Kelly said he sees the market as becoming binary. “It’s going to be zero, it’s going to be one, and that’s going to be $10 or $1,000,” he explained.
“The job is both for software companies, but in reality it’s becoming more and more difficult for those service entities,” Kelly said. “So it’s where we have to adopt new technology and keep ourselves better at all times, we just need to improve our knowledge of the new things that are coming down to the pipes, and work together to make the grid a more stable place.”
Listen to a full interview with Kelly includes details on how the Power Market works. Changes in market dynamics. Other examples of Australia, California and Winterstorm Elliott. The challenges for accurate predictions. How AI improves processes. Plus, listen to the Power Podcast. Click on the SoundCloud player below to listen in your browser, or use the following link to reach the show page for your favorite podcast platform.
Visit us for more Power Podcasts Power Podcast Archive.
—Aaron Larson is Power’s executive editor (@aaronl_power, @powermagazine).
