EBF 483
Introduction to Electricity Markets

9.1.2 Frequency Regulation


9.1.2 Frequency Regulation

We have mentioned the system frequency a couple of times so far in this lesson. Maintaining the frequency throughout the system at something very close to 60 Hertz is critically important. Every single generator in a big power grid must be spinning at this same speed or the system itself may become unstable. As an analogy here, imagine a car going in a straight line. All the wheels have to be turning at the same speed. If one wheel suddenly started to go faster than the others, what would happen? Well, if that one wheel started spinning just a little bit faster, then the rest of the car might exert a force on that one wheel to get it to slow back down. The car would keep going straight. If that one wheel started to spin a lot faster, the car might veer out of control.

The same logic applies to power grids. If the system frequency deviates a little bit from 60 Hertz, then the spinning generators will naturally exert more force on one another to bring that frequency back to 60 Hertz. If the deviation is really large, then by itself the grid will become unstable. Frequency regulation (or just "regulation" for short) is a tool employed by power grid operators in those cases when the system frequency gets too high or too low.

To understand how frequency could get too high or too low, we'll use yet another analogy that is shown in the figures below. Think of a power grid as being kind of like a bathtub with a faucet and a drain. The level of water in the bathtub is like the frequency of the power grid. If the faucet is much larger than the drain, then the level of water in the bathtub will rise. Similarly in a power grid if supply all of a sudden becomes much larger than demand, then the frequency will rise above 60 Hertz. This might happen if there is a sudden surge in supply (if the wind suddenly picks up, rapidly increasing wind power output, for example) or if there is a sudden drop in demand (such as everyone in the US turning their televisions off at the end of the Super Bowl).

See long description below image and caption
Figure 9.2: Power system frequency is like water flowing in and out of the bathtub. For the water level to stay constant, inflows must exactly equal outflows.
Click here for a long description of the figure

There are three beaker shaped diagrams with inflow (generation) and outflow (load) ports:

  • In the first diagram, the in-flow equals the out-flow and the frequency is stable at 60Hz.

  • In the second diagram, the in-flow is greater than the out-flow and the frequency rises about 60Hz.<

  • the third diagram, the in-flow is less than the out-flow and the frequency falls below 60Hz.

Source: Joe Eto, Lawrence Berkeley National Laboratory
Shows the normal frequency of a system is between 60.02-59.89. Outside this, frequency control actions are taken. described in text below
Figure 9.3: Deviations from the 60 Hz frequency can lead to corrective actions that may threaten the reliability of the power grid.
Source: Joe Eto, Lawrence Berkeley National Laboratory

If the drain is larger than the faucet, then the reverse happens - the level of the water in the bathtub will fall. In the power grid, if demand starts to exceed supply then the system frequency will fall below 60 Hertz. This happens most often if there is a sudden drop in supply, like a large generator becoming suddenly disconnected from the grid.

Typically, over-frequency events are easier for grid operators to handle than under-frequency events. If frequency starts to exceed 60 Hertz, this typically happens slowly and grid operators can respond by reducing output from some generators. Under-frequency events, however, can be more serious because they are often unexpected and involve the loss of a large source of electricity. When this happens, recovery of the system frequency to 60 Hertz involves three phases, which are collectively known as "frequency control." These three phases are illustrated in the figure below and can be summarized as follows:

  • Primary frequency control is triggered automatically, without any human intervention, right after the under-frequency event. Generators that are equipped with frequency sensors will adjust their output automatically.
  • Secondary frequency control is triggered within tens of seconds, also automatically, if the under-frequency event does not correct itself. Secondary frequency regulation is sometimes called Automatic Generation Control (AGC).
  • Tertiary frequency control is triggered within a few minutes if the under-frequency event does not correct itself through primary or secondary frequency control mechanisms. Tertiary frequency control typically involves the power grid operator manually adjusting the dispatch of some power plants.
see long description below
Figure 9.4: A sudden drop in system frequency triggers an automated response to correct the frequency, followed by manual interventions from power system operators. Ancillary services provide these responses.
Click here for a long description of the figure

The diagram shows two pictures:

  • The first is a graph showing a sudden drop in frequency on a graph. Frequency is on the y-axis and time is on the x-axis. On the graph, the frequency drops to 59.90Hz between 0 and 8 seconds. This is called the arresting period. In between 8 and 21 seconds the rebound period occurs where the system raises to about 59.94Hz on the graph. After 21 seconds the system is in the recovery period. The system frequency starts to gradually increase from 59.94Hz after 30 seconds but it takes about ten minutes from the initial drop to return to the starting frequency of 60Hz.

  • The second is a graph with power on the y-axis and time on the x-axis showing what controls are used when a power dip occurs. The primary frequency control happens immediately through about 10 minutes and uses power in a bell-shaped curve. Primary Frequency Control is a governor response and frequency-responsive demand response. Then the secondary frequency control kicks in. It starts at about 10 seconds but doesn’t start using much power until about 30 seconds. It then increases in power until about 10 minutes where it then starts to decline. Secondary Frequency Control is generators on automatic generation control. The final control is the tertiary frequency control which gradually increases in power after ten minutes and levels off at an arbitrary power around 25 minutes. The tertiary frequency control is generators through operator dispatch.

Source: Joe Eto, Lawrence Berkeley National Laboratory

The service that we call "frequency regulation" is typically triggered a few minutes after a frequency deviation event, after secondary frequency regulation has kicked in. In areas that have restructured the utility sector and established competitive markets, frequency regulation is typically procured by system operators through an auction process in advance, similar to the day ahead and real-time energy market. The market operator (like PJM) announces how much frequency regulation capacity is needed, and generators submit offers to be able to provide that frequency regulation. This establishes a separate price for frequency regulation.

In most electricity markets, offering frequency regulation to the grid operator means that the generator is willing to increase or decrease output (known as "regulation up" and "regulation down") by some amount. (The ERCOT market in Texas works a little differently, where there are separate markets for regulation up and regulation down.) This means that the generator is, at the same time, removing capacity from the day-ahead/real-time energy market and is committing to being able to produce some amount of power. For example, let's say that a generator with a capacity of 100 MW offered 5 MW of capacity to the regulation market. This means that the generator stands ready to reduce output by 5 MW if needed and to increase output by 5 MW if needed. Thus, the generator could not offer more than 95 MW of its capacity to the day-ahead/real-time energy market and would need to make sure that at least 5 MW was cleared in the day-ahead/real-time energy market.

The payment for regulation has two components. First, generators are paid for the capacity that they dedicate to providing regulation. This is sometimes called the "capability" price and takes units of $ per MW of capacity. Second, when a generator is called on to increase or decrease output in response to a frequency deviation event, it is paid for the energy that is produced or not produced. This is sometimes called the "performance" payment and is often set equal to the real-time price energy (so takes units of $ per MWh).

For example, let's take our same generator providing 5 MW of regulation. The regulation capability price is $5 per MW. The generator is dispatched to produce 50 MWh of energy through the real-time market at a price of $10 per MWh. Because of a frequency deviation event, the generator is asked to produce an extra 2 MW of power for 10 minutes. The generator's total revenues for this hour would be:

  • Energy market revenue: 50 MWh * $10/MWh = $500
  • Regulation capability: 5 MW * $5/MW = $25
  • Regulation performance: 2 MW * (1/6 Hour) * $10/MWh = $3.33
  • Total revenue: $528.33