We left off with 1 poor power plant and 5 little houses. Obviously the modern electrical system is WAY bigger and WAY more complicated than this. One thing that might have occurred to you in our little example is the fact that when there are only 5 houses there can be wild swings in demand. Instead if there were 100 or 1000 or 100,000 houses then when just 1 home's demand doubles the effect on the grid as a whole would be much less. Put another way, 1 home in 10 is 10%, but 1 home in 100,000 is only 0.001%. And in fact this really is one way that the modern grid is able to stay so consistent and stable - by combining thousands of homes and businesses into one unit. By the same token instead of having just one power plant there are many.
The grid is physically controlled by people and computers at large grid operation centers. From what I am told, they look like a combination of an air traffic control center and a stock trading floor. Large numbers of grid operators with lots of computer screens are monitoring demand, generation, the weather, and much more. Because these centers are crucial to national security you (nor I) know where they are, and security is intense. These centers are staffed and operating on a moment to moment basis 24 hours a day, 365 days a year.
In a very simplified example, they work something like this. You have a region such as southwestern Ohio (where I live in Cincinnati). My electricity is provided by Duke Energy. Duke has grid operators making sure that their area has just the right amount of electricity. The grid operators have a "menu" of power plants at their disposal from which to draw power. As an example (that approximates my area), let's say they have 6 coal plants, 2 natural gas plants, 1 nuclear plant and 1 solar electric plant. How do they choose which plant to use? In their lingo they choose using "efficiency," which you might think refers to which power plant makes the most electricity with the least emissions. But no, when they use "efficiency" they are referring to the most electricity for the least amount of money. In other words, they use the power plant that is the cheapest at that moment (these are businesses after all). So how does that work?
First come renewables such as solar or wind. A solar or wind power plant costs a lot to build, but their "fuel" (the sun or wind) are free. Can't get cheaper than that! So whenever the sun shines or wind blows the electricity made is typically allowed to feed straight into the grid. Note that a grid operator can't control when the sun shines, so they just have to take it when they get it. As an example, grid operators have no control over a homeowner's solar panels. So really the grid operator doesn't "choose" renewables, it's more like they just "happen" and the grid operator adjusts accordingly.
Second comes nuclear. Yes the nuclear plant costs a ton to build but once it's built and running... it's running... and running... whether you use the power or not. The uranium lasts years, so over the long haul the cost of the actual fuel is almost negligible.
Third, traditionally, come coal plants. Until the recent past, coal had been cheap compared to natural gas (something that has changed a lot in recent years with the fracking boom). Coal plants are not very "agile," and can only increase or decrease their output within certain limits. If a coal plant is just being run at a constant rate, there is no problem, but as more solar and wind are added to the grid, a coal plant may be required to start and stop daily. A coal plant needs 2-3 hours to get going, so when a grid operator looks ahead toward the peak demand hours of the day (usually early evening), he or she knows to go ahead and start ramping up that coal plant way in advance. And given that 1) coal is cheap and 2) you have to ramp it up now anyway -- why not choose it?
Fourth comes natural gas - although this is changing in recent years. A recent survey of the United States as a whole found that the utilization of coal vs natural gas power plants in 2015 was essentially neck and neck (source: http://www.eia.gov/todayinenergy/detail.cfm?id=25652). The reason for this is that natural gas has become dramatically cheaper because of the increased supply from fracking. For this reason, grid operators are quicker to utilize natural gas power than they used to be. Natural gas power plants come in 2 varieties: the traditional simple cycle "peaker" plant, and the newer combined cycle gas turbine (CCGT in the industry lingo). The first kind of plant called the "peaker" plant has been around the longest and is typically reserved for "peak" demand times just as the name would suggest. This type of plant is very agile and can be ramped up within seconds, so it is a great way to meet spikes in demand. The problem is that this type of plant is not very efficient in the traditional sense - i.e. it does not convert a very high percentage of the energy in the fuel into electricity (on the order of around 20% to 40%). For this reason the cost is also higher, and therefore the "peaker" plants are used sparingly. The new type of natural gas power plant, the combined cycle type, is the one you may have heard of where they use a jet engine (pretty much the same kind as on an airplane!) that is set up to run on natural gas. You may be wondering what the "combined cycle" refers to. This is how it works. The jet engine is mounted on some huge stand, and uses natural gas for its fuel. As it runs, super hot exhaust gas comes out the back. This exhaust spins a first set of turbines to make electricity. But then instead of just letting the exhaust go, they use it to heat water and make steam! Then the steam turns a second set of turbines. This way they milk as much power as they can from from the generator as a whole. And this is why they are called "combined cycle." By using the two "combined" cycles (first the exhaust and then the steam) this type of generator is much more efficient in terms its fuel use. In other words a much higher percentage of the energy in the fuel ends up as electricity (more on the order of 50% to 60%). This type of plant is also fairly agile (though not as agile and the peaker plant) and can ramp itself up within minutes. For all of these reasons more and more of these new combined cycle natural gas power plants are being built.
So in summary here is the typical order in which a grid operator will "recruit" or start to use each type of power plant:
- Renewables such as Solar or Wind (again these just "happen when they happen")
- The newer Combined Cycle Natural Gas
- "Peaker" Natural Gas
Recall however that 4 & 5 are pretty much neck and neck these days with combined cycle natural gas expected to move up a spot in the future. Nuclear and Coal, given that they are not very "agile" are termed "base load" plants. There is never a time when no electricity is being used at all - even at 4am there is still a lot of electricity being used. This baseline amount of electricity is termed the "base load." And when electricity demand peaks, that is called "peak load." Nuclear is the ultimate "base load" plant - it just runs and runs with really no change. Traditional Coal plants are also mostly base load. The newer combined cycle natural gas plants are kind of in between - they can be used for both base and peak loads - this often termed "intermediate load." Finally as you already know, the "peaker" natural gas plants are used exclusively for peak loads.
The next part in the overview series is an "Interlude" with further background we need to review before we really get to see the grid in action.