Value of Flexibility Management in the Eyes of a Market Trader & Operator
By John Bonnin
(John Bonnin has 25+ years of industry experience, including 15 years at CPS Energy. He has directed the Day Ahead and Real Time activities of Energy Supply and Market Operations, with a focus on managing market risks to native load customers and optimal dispatch of our power generation fleet. Mr. Bonnin has a Master of Science in Management/Computer Resource Management from Webster University and a Bachelor of Science in Chemistry from Louisiana State University)
As a trader and operator in the Electric Reliability Council of Texas (ERCOT) market going into the peak month of August, I’m concerned about two things:
- Coverage of load obligations
- Market opportunities
My worry is moderate if demand on the grid is high and expected to remain so. My worry is even higher if there is a significant amount of thermal resources offline due to forced outages. My worry is at an all-time high if renewable power production is below forecast.
Each day, power traders must estimate the probability of scarcity pricing for that day. To determine the projected scarcity, a calculation of “net load” is commonly done – which starts with a forecast of total load, then subtracts forecasted wind and solar generation. In a market like ERCOT, renewable energy production can vary 5-10 GW from forecast on any given day. In recent years, the highest prices have occurred not on the highest demand days, but on the high demand days when renewable energy underperformed.
On a high-risk day, when scarcity is expected due to any of the reasons above, a quick-response generator provides value to a trader for both of the highest concerns – covering a load position and participating in a market opportunity. A quick-response demand response (DR) resource serves the same role, but its unique attributes may actually present more value opportunities.
In a market like ERCOT, power traders and other operators expect DR deployments across the highest load days. ERCOT publishes its estimate of DR capacity in its Capacity Demand Response (CDR) and Seasonal Assessment of Resource Adequacy (SARA) reports, but the exact amount across a given day has a degree of uncertainty. In recent years approximately 1,500 to 2,000 MW have been deployed across the peak on these days.
Also, within each of these peak days there are timing differences between industrial, residential, and ERCOT peak hours. Many industrial loads peak between 3 and 5 pm, although ERCOT peak hours is usually between 4 and 6 pm. Some residential loads actually peak between 6 and 7 pm, as families turn on appliances to prepare supper, boot up computers for homework, and cool their home. However, if commercial customers, such as hotels and shopping centers, have reduced consumption for the peak hour, this energy use may snap back during the 6-7 pm hour.
In these complex and dynamic cases, a DR resource can cover a load obligation and capture a market opportunity on the same day. If my load obligations are hedged for the market peak (4-6 pm), deployment of DR across these time periods decreases my actual load, and the energy represented by the DR is sold back to the market (at high prices). Continuance of the DR deployment through 7 pm, especially selected residential DR programs, can reap market prices and lower the coincident peak of my load. A generation resource cannot provide coincident peak reduction.
These strategies are enabled by systems to communicate and relay instructions to the loads that participate in demand response. The ability to target and bring online selected customers, staged through a high-demand period, gives grid operators flexibility in responding to the dynamics of prices across the four to five-hour peak period. More resources can be called when prices are higher, and released as prices soften, rather than an “all or nothing” deployment. For example, when ERCOT prices reach a cap of $9000/MWh for a few days a year, having a 100MW of flexible capacity for a few hours can yield as much as $5MM in savings. If those hours are also co-incident peaks, the savings can double.
Flexibility Management platforms such as AutoGrid Flex™ offer a compelling value proposition for market operators and traders, giving them real-time situational awareness supported by sophisticated data analytics. Intuitive AI-based controls with optionality for various programs allow granular and targeted dispatch based on dynamic market conditions. Finally, accurate measurement and validation of performance of individual resources allow a deeper customer engagement after the event and for fine-tuning operations.