If you read enough news articles about renewable energy development, eventually you'll see a statement along the lines of "this plant will produce X megawatts of electricity, enough to power X thousand homes."
Reporters are generally encouraged to translate figures, especially mindbendingly wonky ones in fields like electrical engineering, into terms the lay reader can more easily grasp. That's why this particular phrase is so widely used. The figures are usually simple. The reality is a bit more complex.
Take this sentence from a BLM press release:
Located 40 miles south of Las Vegas, Nevada, Silver State North is a 50-megawatt plant that will use photovoltaic technology to generate enough power for about 9,000 Nevada homes.
We'll pick on this example, though there are plenty out there. There are a couple of issues here.
According to the US Department of Energy, the typical Nevada household consumes 986 kilowatt-hours of electricity per month. The "average month," accounting for the "30 days hath September" and Leap Day factors, has 728.48 hours in it. Do the math (986 divided by 728.48) and this means that in an average moment, an average Nevada home is running the equivalent of 1.35 kilowatts of appliances.
BLM says the 50 megawatt plant will power 9,000 homes, but if you multiply 1.35 kilowatts by 9,000 you don't get 50 megawatts. You get a little under 12.2 megawatts. Why the discrepancy?
The discrepancy in this particular example very likely results from the BLM's attempt to account for the plant's capacity factor, which is a concept we define here. The short version: 50 megawatts is the plant's maximum output, but it isn't always going to be producing power at maximum. Sometimes sections of the plant will be down for maintenance. Sometimes a cloud will drift over the solar cells, reducing their output by a varying percentage. And for about half of each day, the sun is going to wander over to the other side of the planet and the plant won't put out any energy at all. Solar power installations' average output runs somewhere between 20-25 percent of their nameplate capacity. It looks like someone responsible for this press release took 50 megawatts, divided by a capacity factor somewhere between 20 and 25 percent, got 12.2 megawatts, divided THAT by 1.35 kilowatts and rounded off the result to a nice round 9,000 homes.
At first glance this seems to make sense mathematically, but when the sun goes down and the plant stops producing electricity, every one of those 9,000 homes is going to need another power plant to run their lights. In other words, baldly stating that the plant will power 9,000 homes is greatly oversimplfying things. It's much better to phrase things as Silver State North's neighbor BrightSource does, saying that their Ivanpah plant produces enough power
"to serve more than 140,000 homes in California during the peak hours of the day."
The other issue is in determining how accurate the stats for how much power an "average" house consumes. The above average for Nevada households, for instance, includes homes in broiling Laughlin, where residents face average temperatures above 85 from April through October, and South Lake Tahoe, where the average never breaks 81. One household may be cranking the AC for more than half the year while the other turns on a ceiling fan a couple times in a summer. One home may be energy efficient while another's weatherstripping has gaps big enough to let small animals pass. Averaging households across a wide region made up of varying climates, cultures and lifestyles can create numbers that aren't particularly useful in any real-world situation.
While it may be tempting to resort to numbers of homes as a way to convey how much power a plant will generate, in the end it's often not very accurate, nor really all that clear to most readers.