Your Power Plant Might Have a Drinking Problem

While at an energy conference (ARPA-E, 2014) I found out that power plants account for 40% of water draw in the US. Simply put, they use a lot of water. The good news is that it doesn't have to be fresh water. Brayton Point, for instance, uses grey water. In other words, it uses water that came from your toilets and sinks that has been reprocessed. Others use saline water from oceans (all water in the oceans is saline, cause it is salt water).

No math this time, just review the math from my thermal power plants post.

Why do power plans needs water? Cooling purposes. The way a turbine works is that high-pressure air has to drive through it. The way this happens is water is flashed to steam. Steam takes 1600x the space at 1 atmosphere compared to water. So it creates a massive pressure differential on one side of the turbine, turning the fans, turning the turbine, and generating electricity. The steam needs to be cooled on the other side to either create the vacuum that drives the pressure differential to turn the turbine, or, if it's a close cycle and the same water is used, to cool the steam back to water. It needs to be water again, otherwise it cannot expand and drive the turbine.

Schematic of a thermal power plant. It needs water to cool the water used to drive the turbine.

Okay. That was complicated. Let's break it down further. This section if very detailed, and most of you will want to skip this paragraph. Here goes: There are two major ways to run a thermal power plant. Combined cycle, and single cycle. Combined cycle is more efficient. How? It uses several turbines to extract energy rather than a single one. Think about it this way: when you have 300 degree celcius steam coming from the coal-burning reactor, it is all steam. There is no water-phase droplets in it. This is called dry steam. It can be directed to special high-efficiency turbines that can extract a lot of energy. The steam then loses pressure and temperature, and some water droplets begin to form. If this mix of steam and water were directed at the same turbine, it would pit and tear at the turbine blades, destroying it. Two things could be done with this steam. Either it could be directed to another turbine, or it may not be reused. The second turbine will be designed differently for steam that is lower pressure and lower temperature. Having multiple turbines like this increases efficiency. Inefficient plants use only one turbine

(everyone else should join back in now) Eventually you end up with a mix of water a steam. As I said before, this has to become water again, so it can expand to steam and drive turbines. Or, if a plant is doing a once-through cycle and expanding water from a stream, it needs to dump the water back into the environment. Dumping near-boiling water into the environment is a terrible idea. That would be a bit of a disaster. So, in either case, you need a lot of water from the environment to cool the water used for the steam cycle in the plant. An alternative scenario is using evaporative cooling towers. They evaporate water, which requires heat to go into the water, which then cools other water. No matter what, cooling a plant requires a lot of water.

So here we come back to the end point. Power plants use an insane amount of water. "Ahh, but Jason," you ask, "these are just thermal power plants. I use solar power. So my plant is water-efficient!"

Not so, I say! Solar plants also use water for cooling and cleaning. And this is from NYT, an ostensibly liberal paper that likes solar. This is because major solar plants use solar thermal, rather than solar PV. Solar PV is pretty much water-free, other than for cleaning mirrors. But that electricity is too expensive to be useful at the grid scale (recall from a prior post that it costs about 3-7 cents to produce a kwh of electricity, but we buy it for 20 cents, so it makes sense for us to put solar panels on our houses at the cost of 20 cents per kwh, while it doesn't make sense for power companies to use solar panels since they mark up prices 3x to get power to us).

How does this compare to coal? In the link above, we have solar power using 1.2 billion gallons a year to produce 500mw of power. Your average coal plant produces 600 watts of power. A once-through plant draws "between 70 and 80 billion" gallons a year. But a closed-cycle plant, the one that uses the same water in the plant and only uses other water for cooling at the end of the power cycle, uses 1.7 to 4 billion gallons a year. So the efficient ones are comparable to solar in water use.

So here we have a chart showing all this:

Water use by power plant type, source 

Note that you can find different graphs using different information sources, but the general point always remains: power plants use a lot of water.

Wind power, however, doesn't use water. Unfortunately, wind power is only available in a few places. How about hydro power? It passively uses water, so it doesn't really count. Great, right?

Not so fast. How much of the US power generation comes from thermal and solar sources? According to the EIA, 87%. I reproduce the info here:

In 2012, the United States generated about 4,054 billion kilowatthours of electricity.  About 68% of the electricity generated was from fossil fuel (coal, natural gas, and petroleum), with 37% attributed from coal.

Energy sources and percent share of  total electricity generation in 2012 were:

  • Coal 37%

  • Natural Gas 30%

  • Nuclear 19%

  • Hydropower 7%

  • Other Renewable 5%Petroleum 1%

    • Biomass 1.42%
    • Geothermal 0.41%
    • Solar 0.11%
    • Wind 3.46%
  • Other Gases < 1%

So yeah. Your power plant has a drinking problem.

thanks for reading!

- Jason Munster

Other Alternatives

Here we will cover a few more electricity producing alternatives, specifically geothermal in its different flavors, and the waste of money that is tidal power. Before that, let's make a quick roadmap of what we have covered, where that is going, and what we haven't yet covered.

Pretty much, we have talked about electricity producing resources. We have only briefly touched on energy as a whole. In the US, for example, 35% of all energy use is petroleum for transportation. None of the stuff we have discussed is useful to replace that without better battery technology. Nonetheless, it is likely that at some point in the next century, much of our domestic energy needs, including transport, will be covered by electricity. And we will require a lot more of it. An upcoming post will assess all the different tech for producing energy we have discussed, and which ones can be potential solutions.


Geothermal energy exists because it gets hot underground. In general, the temperature of the Earth rises by 30 degrees C for every km you go underground. This temperature increase in depth is called the geothermal gradient. If you go 7km underground, you are pretty much guaranteed temperatures of higher than 200 degrees C. Which, as we all know, is hot enough to boil water. 7km underground is pretty deep, however. In some places, the underground is much hotter. The temperature rises much more rapidly. It could be due to volcanic activity in the area, or radioactive decay underground. Either way, when hot temperatures are closer to the surface, that heat can be harnessed to drive a turbine.

Map of the geothermal resources available in the US. In general, this represents areas of higher temperature gradients.

Geothermal comes in two main flavors. One directly harnesses the steam from the ground to produce electricity (called flash steam, cause the pressurized hot water comes out, flashes to steam at atmospheric pressure, and drives a turbine), the other uses a heat-transfer mechanism where pressurized hot water from the subsurface (it needs to be pressurized, because it is above the boiling point of water at atmospheric pressure) is run through a set of heat-exchanging pipes before being put back underground. There is a third type discussed later in the article called hot dry rock.

Surprisingly, the first mechanism of directly using steam, in practice, is unsustainable and produces pollution. This is because the used steam is often vented to the atmosphere. The steam produced underground has pollutants. Like CO2 and sulfur, amongst other things. If the steam is used directly in a turbine and then expelled to the atmosphere, these pollutants come with it. If the used steam is instead re-injected into the formation, this problem is avoided.

Reinjecting the steam is easier and more common in the heat-exchange mechanism. The super heated stream of steam from underground is already isolated, and re-injection is pretty simple.

And here comes the fun part. If the steam is used and then vented rather than reinjected, the formation will run out of water. Instead of being a renewable resource, the geothermal will be a depletable resource that will only last for 10 or 20 years. This is because the pressure of the formation will drop, and the steam will no longer be able to rise. Does this sound familiar? Oil and gas production need to do this all the time to get maximum recovery rates. Reinjection of fluids is rather easy, and has been pretty well developed by the oil and gas industry.

Hot Dry Rock

The next major innovation takes a cue from the oil and gas industry. Hot dry rock is exactly what it sounds like. The rock starts off hot and dry. It has plenty of heat in it, but there is no steam or water to produce and then make energy from. How is this dealt with? Hydrofracking and injection. A well is drilled, the drill hole goes horizontal, it is fracked to drastically increase the surface area that the well hole can be exposed to, and water is injected into the rock. The water heats up a lot, then it is produced via a separate well to make steam. It is fairly complicated, and costs a lot more.


EGS stands for enhanced geothermal systems. You will run across this term a lot these days. It more or less means that the heat in the field is managed by either fracking beforehand, injecting water afterwards to maintain pressure in the field and extend the life of the geothermal power plant, or a combination of both. It drastically increases the lifetime and viability of a geothermal site.


The capital costs of geothermal pretty much will dictate the average cost of electricity produced. It looks like flashed steam will cost the least. In reality, unless the steam is reinjected afterwards, the field won't last as long, and the capital investment costs will have to be paid out over a shorter period of time, resulting in higher costs. Hot dry rock will undoubtedly always remain more expensive because of the costs associated with fracking and reinjection.

Footprint and other

Most of our power plants produce heat above ground, and need storage for either spent nuclear fuel or a coal pile (except for gas plants. They just need pipelines). So geothermal power plants don't take up a ton of space

fun uses of geothermal: geothermal heat is produced and used in Iceland to melt snow on the roads and such.


I tend to think that tidal power sucks. In part because it is very expensive, and in part because at best it could provide all of 1% of world electricity.

Tidal power has two main problems: it uses salt water and it has only a few areas that it will work. There needs to be tides of sufficient strength that it can produce electricity, and even then, salt water is corrosive, limiting the lifetime of these power plants and making the levelized capital cost very high.

Tidal power also comes in two main flavors. One is tidal impoundment. Think of it as creating a hydroelectric dam every time the tide goes out. The tide comes in, fills up the area behind the impoundment dam, and then as the tide goes out, the area behind the impoundment dam is filled, and as it flows out, it generates electricity. As you recall from the hydropower article, the energy produced from a hydro dam directly relates to its height. The height of a tidal impoundment dam is limited by the height of the tide. In most parts of the world, this is not very high, so it is not very efficient. Moreover, it kinda messes with natural habitats.

The other type of tidal power is more or less an underwater wind turbine. The problem is that all the moving parts are underwater in the ocean. Where decay and breakdown happens quickly. Moreover, looking at the equation of energy produced via such a turbine:

 E = \frac{1}{2} A \rho v^3

where A is area, and v is velocity,

we quickly realize that the area of the rotor for a tidal turbine is small (wind turbines have 40m blades, and we aren't gonna have 80m of water depth in most places to replicate that scale in tidal areas), and the speed is slower (water doesn't flow at 6-8m/s very often). Tidal power can't scale and produce as much energy as wind. And the environment is unfavorable. In short, this is not a viable resource for large-scale energy production. And it costs a lot.

Hokay, that is all for today! Thanks for reading!

-Jason Munster


Oil exports and imports

I am going to be lazy this week and post a very short one.

We discussed the Bakken before. It is producing nearly a million barrels a day of oil. This is $100 million in oil sales a day. It costs up to $45/barrel to lift the oil out of the bakken. The Bakken makes profits of about $50 million a day, or about $18 billion a year.

Let's compare this to Saudi Arabia. They pull oil out of the ground for $1 per barrel (it is very efficient there). They make profits of roughly $100/barrel. They produce 12 million barrels a day. In other words, they make $1.2 billion per day.

The US currently imports 11 million barrels of oil per day, and we use a total of 20 million. We transfer about $1.1 billion out of the US per day to drive cars.

What is the end product of this money? Many countries with oil have what is called the "resource curse." It is also known as Dutch disease. It turns out if your country has a ton of natural resources, it becomes inefficient and actually has less growth than it would have otherwise. This has several reasons. First, if a government makes a ton of oil money (or copper or gems), they tend to provide stuff for free to the populace. They take the money they make and just use it to pay for social things, like schools, hospitals, etc. Citizens pay less or no taxes. Corruption and handouts tend to be rife, but since the citizenry gets stuff for free, they don't even care. It is highly inefficient, and the country does not develop a real economy. They often will not use the money to diversify the economy, and once the resource runs out, the country is literally worse off than it was before they found the resource. Finally, having resources tends to allow despotic regimes to thrive. Iraq, Iran, Venezuela,

Interesting, eh? Maybe is a good reason to drive a fuel efficient car. Except that China and India and other developing countries with increasing numbers of cars will ensure every barrel of oil finds a buyer.

One last thing. The US gets its oil imports almost entirely from North America, from Mexico and Canada. The Middle East oil goes to China and the rest of the east. Why does the US spend so much money to maintain peace in the Middle East? Cause if the oil produced there came off the market, the former buyers will have to go elsewhere to buy their oil. Despite the fact that we don't get oil from the middle east, the global price (and our price) is much lower than it would be if war caused the oil to stop flowing from there. So why doesn't China start also enforcing peace in the Middle East? How come the US is the major country to fund this peace keeping? Frankly put, no other country has the experience of the US with having troops on foreign soil. China doesn't yet have the experience to effectively do this sort of thing. But they are already practicing. Eventually, as China continues to rise, they will likely begin to shoulder responsibility of ensuring the flow of oil.