Introduction
Our nation’s electric grid is an immensely complex system that is constantly balancing on a knife’s edge. Grid operators must ensure that energy supply meets demand every second: too much or too little supply will lead to massively high energy prices at best and blackouts and billions of dollars in damages at worst (see an aside at the bottom on the physics of why this happens – it’s really interesting!).
Winter Storm Uri, which hit Texas in February 2021, is a textbook example of what happens when the grid falls out of balance. High energy demand for heating during the extreme cold combined with natural gas and renewable energy failuresled to supply and demand falling out of sync. As a result, blackouts left 69% of Texans without power, spiked energy prices in excess of $9,000 / MWh (ERCOT energy prices are typically ~$30-50 / MWh), and led to $80-130B in economic damages.1,2,3
During the extreme conditions of Winter Storm Uri, energy prices in Texas briefly rose to over 150x their average levels.
While Winter Storm Uri was a clear example of a grid failing, major events like these happen rarely. Nonetheless, grid operators must successfully balance the grid every day. This is getting increasingly challenging as the planet heats up and energy demand grows. For instance, high temperatures on an otherwise ordinary day in Texas on August 20, 2024 contributed to the highest ever energy demand in the state, at 85,931 GW, driven mostly by excess energy usage tied to air conditioning.4,5 In response, many grid operators have developed a suite of tools to help balance the grid in times of need. One of these – the focus of this blog post – is demand response.
What is demand response?
Demand response is the generic term for a series of programs implemented by utilities, grid operators, or states with the goal of boosting grid reliability. This takes several forms (see below), but all essentially involve the grid operator alerting participating energy customers that they need to turn off some or all of their energy usage for a set length of time during periods of grid stress. In aggregate, this relieves grid stress and can even help balance supply and demand on the grid.
But why would anyone – for instance, a manufacturing facility or data center that depends on efficient operation (and will lose money if it shuts off operations for several hours) – want to participate in such a program?
Because they can make a profit! When energy consumption is voluntarily reduced during a specified time period, the electrons that these manufacturers and data centers were originally going to consume are freed up for other consumers. In effect, these businesses act as a (virtual) power plant, and are able to sell this newly freed-up energy back to the grid, typically at a very high price. Assuming this energy cost is higher than the revenue these businesses could make through their day-to-day operations, they actually earn additional revenue by participating in demand response. (Note that there are other benefits for participating in demand response programs beyond simple payments for directly reducing energy consumption, such as through coincidental peak management.)
Demand response takes several forms (note that there are others, but these are the most common):
-
Energy: Participants are called on to reduce energy consumption when energy prices get too high, reducing overall demand and costs for other energy users. This can contribute to a “flattening” of the demand curve during peak energy demand hours.
-
Capacity: Participants promise to be available if needed during the highest demand days of the year, and are compensated for this promise (see thecrazy PJM capacity auction results).
-
Ancillary Services: Participants reduce energy demand on short notice for limited periods of time to help maintain the frequency of the electricity on the grid (another critical component of grid stability!)
How does demand response play a role in the climate conversation?
By and large, demand response is used by grid operators in times of high energy demand, and therefore high grid stress. In a world without demand response, these excessive levels of energy demand would be served primarily by what are known as “peaker plants”. Peaker plants are power plants that are able to quickly turn on and off to serve this higher demand when needed. These plants serve as “peakers” primarily because of their high cost of operations: it only makes financial sense to turn them on when demand (and therefore energy prices) are highest. Unfortunately, peaker plants are powered largely by natural gas, and are often located in and around historically disadvantaged communities, contributing not only to carbon emissions, but also to local air pollution.6
Luckily, demand response has emerged as a key tool in grid management. When energy demand is high, grid operators are able to call on demand response assets. By turning off their consumption during these key hours, the demand response virtual power plants are often able to prevent peaker plants from switching on, thus avoiding carbon emissions and air pollution.
Conclusion
Demand-side management is a critical and often overlooked piece of the energy transition. As more and more intermittent renewables are brought online, greater demand flexibility will be required to ensure everything continues to run smoothly. Demand response is a critical tool in the demand-side management toolkit, but more is needed. Beyond demand response, distributed energy resource management (DERMS), supply shifting, time-of-use bill management, and other demand-side treatments will need to be added to our grid’s toolkit as well. As our economy electrifies, driven by the adoption of EVs, the electrification of industrial processes, and changing consumer preferences for household appliances, managing our grid intelligently will be critical in mitigating carbon emissions tied to the energy sector.
An aside: The physics of the grid
“Why does maintaining a balance of supply and demand on the grid matter?” you might ask? I want to give a fast and dirty (and VERY simplified) answer:
One critically important aspect of operating the grid is maintaining a constant frequency. In the US this frequency is 60 Hz, or 60 cycles per second.7 Most machines, systems, and appliances powered by electricity in the United States are designed to run at exactly 60 Hz, or they will break.8 The majority of energy in the US is still produced via fossil fuel generators, which heat steam to spin turbines. These turbines contain powerful magnets positioned next to a set of wires. The magnets’ spinning creates a rapidly changing magnetic field that generates electricity in the wires through a phenomenon called electromagnetic induction. Because of the way the wires are positioned, they generate an electric current that alternates direction. Generators are therefore set to spin at 60 Hz, dictating the frequency across the electric grid.
Turbines are designed to spin at a rate that will generate an alternating electric current (AC) with a frequency of 60 cycles per second, or 60 Hz.
The grid is comprised of thousands of spinning generators. If for some reason (e.g., wear-and-tear, minor mechanical issues) a single generator starts to speed up or slow down, the prevailing frequency of the energy grid will start to force the erroneous generator to correct to 60 Hz. If this faulty generator gets too far out of sync with the other generators, these correction forces can literally tear the turbine apart.
This is also true when supply and demand get out of sync. When supply exceeds demand, there is too little “friction” from the energy users on the grid and that 60 Hz frequency will start to rise. Similarly, when demand exceeds supply, the opposite effect occurs, and grid frequency falls. In both cases, the abnormally high or low frequency, perturbed from the standard 60 Hz level, can damage or even destroy generators. There is not much room for error: our grid’s frequency must stay 59.7-60.3 Hz to avoid serious issues.
Risk of major infrastructure breakdown means that at the first sign of failure, power plant operators will shut off their generators. This typically leads to a cascading problem where shut-offs lead to decreased grid frequency and put additional strain on the remaining power plants, which leads to more shut-offs and even more strain. The cycle will continue until widespread blackouts begin to occur. It often takes days to get all these power plants back up and running. If it’s very hot or cold, that delay can cost lives, as was seen during Winter Storm Uri.
Sources
1https://comptroller.texas.gov/economy/fiscal-notes/archive/2021/oct/wint...
2https://www.ercot.com/content/cdr/contours/rtmLmp.html
3https://www.utilitydive.com/news/ferc-investigates-market-manipulation-w...
4https://www.utilitydive.com/news/electricity-load-growing-twice-as-fast-...
5https://blog.gridstatus.io/a-record-setting-day-in-ercot/
6https://www.gao.gov/products/gao-24-106145