Balancing the power grid is a tough problem, made even tougher by the fact that power plants can't always be relied on. Many plants only run in certain conditions- solar during the day, wind when it is windy- and all plants have downtime for maintenance or refueling.
For a electric grid operator, this
presents a problem. Power grids are closed electrical systems, and as
such, need to remain balanced between generation and load. In the US, all electrical equipment is built to
run at a grid frequency of 60 Hz- have too many or too few plants generating power, and you'll deviate from that special frequency and start damaging equipment all over the grid. Once the equipment fails or trips offline due to frequency variation, you could see a cascade into widespread blackouts like in 2003[1].
Not only that, but grids
currently have very little ability to store energy in any way-
meaning that the power coming from all of these power plants must be
carefully balanced at all times to keep the grid stable. Grid operators call power plants into action as load increases, and turn them off when load decreases, maintaining the balance of power throughout the day. In order to do this, operators need a good understanding of the capacity available to them.
Capacity is
the maximum normal power output of a generator. If you recall from
our discussion of power and energy units, power is the
instantaneous output of energy- it's exactly what grid operators are
balancing.
Grid operators
have to plan their capacity deployments carefully. Energy
can't be stored, and operators have to hit a moving
target as electricity demand fluctuates through the day. This means
that they need to ensure they have enough capacity available for peak
demand, when grids have their highest power requirements. Let's take as an example the total load in New England on an August weekday:
A
power plant's capacity factor
is a measure of how often they are generating power at or near their
rated capacity. It can be thought of as their average output compared
to their maximum:
In other words, the capacity factor is how often a power plant is running. For plants that run on a fuel, this is often influenced by fuel costs. Cheaper plants will typically be called to run more often, resulting in higher capacity factors.
When
trying to meet demand, grid operators must use different power plants for
different purposes. Baseload plants
stay on almost around the clock to meet demand that never sleeps, and are typically the cheapest to run per unit of energy. They often won't- or can't- shut down easily, meaning that 24 hour operation is a necessity rather than a business choice. Load-following power plants generate
power to cover the daily ramp up of usage during the day, and have
the ability to come online quickly to meet daily increase in demand. A
few others are peaking power plants,
designed to operate only a few times a year at times of extreme peak
demand. These are typically the most expensive to run, and often they
are some of the last resources for a grid to call before running out
of power.
How do you know which is which? Well, we can tell easily by looking at the capacity factors- baseload plants running most often with the highest capacity factors, and peakers running least.
How do you know which is which? Well, we can tell easily by looking at the capacity factors- baseload plants running most often with the highest capacity factors, and peakers running least.
Above: Average capacity factors of fuel-based power plants in the US, 2013. Data from the EIA[3].
But the above data doesn't take into account intermittent renewables. We should make the distinction between dispatchable and non-dispatchable generation- renewable sources often run without being dispatched, only when conditions are right. Due to advanced weather forecasting, operators are more than ever able to predict and plan for wind and solar output, so they typically generate in the same role as a baseload power plant- contributing as much as they can whenever they can, at low cost. Solar power also often generates at times very close to the actual peak demand, which helps to offset the need for more expensive load-following and peaker power plants.
Above: Average capacity factors of non fuel-based power plants in the US, 2013. Data from the EIA[3].
Above: Average capacity factors of non fuel-based power plants in the US, 2013. Data from the EIA[3].
We can see that primary among the baseload plants is nuclear power, with a whopping 90% capacity factor. High energy fuel and rare shutdowns make it easy for nuclear plants to run around the clock, and shutdowns can be difficult and expensive. Biomass plants are able to run often as well due to low fuel costs- typically municipal trash or other waste streams which are easy to procure. Geothermal and hydroelectric generators provide reasonably stable support, and wind and solar round out the rest when they can, though with lower capacity factors. Often large, more efficient natural gas and coal plants make up any remaining baseload needs.
Load-following power plants are most often coal and natural gas, although gas has the ability to respond more quickly. These plants jockey for lower fuel costs, and the recent shale gas boom has given the economic edge to natural gas. Even so, in most southern and midwestern states, coal still serves a significant baseload and load-following role, despite the gas dominated northeast closing many coal plants in favor of further cheap gas development.
Peaker power plants typically involve smaller petroleum and gas plants that are able to fire up quickly and run only a few days a year. Because of their limited operating hours, these plants are typically old, expensive, and inefficient, but get paid handsomely for their performance. Petroleum based plants rarely serve a role outside of peak hours, which is why they earn the lowest average fuel-based capacity factor at 22%.
So, in order to meet The total load of a grid, we end up with a power profile that looks like this:
Above: Approximate Generator roles for New England, Aug 20th 2014
This isn't exact, but it's a good idea of how grid operators need to think about their assets. A baseload power plant can't usually perform the same role as a peaker or load-following plant, as it can't switch on and off quickly. Conversely, a peaker power plant can't handle running 24-7 and would likely fail. This stack is one of the reasons that maintenance on larger power plants must be carefully accounted for- operators need to have enough of a diversity of power plants to fill each needed role and meet total demand.
At least the grid operators know what they have to work with. If you're curious for more insight, many system operators offer access to their real-time system data- check out ISO-NE's real time energy dashboard[4]. and phone app[5]. Both of these applications give you good insight into what's really happening when you flip your light switch on in the afternoon. That energy doesn't come out of thin air- but where exactly it comes from certainly depends on many factors.
Sources:
[1] NY Daily News: Northeast Blackout of 2003. 08/14/2013
[2] ISO-New England: Energy, Load and Demand Reports. 08/20/2014
[3] Energy Information Agency: Electric Power Monthly. 09/25/2014[2] ISO-New England: Energy, Load and Demand Reports. 08/20/2014
[4] ISO-New England: Real Time Maps & Charts. 12/04/2014
[5] ISO-New England: ISO to Go. 12/04/2014
