Many power plants need lots of water. They burn fuel or split atoms to boil water to produce steam to turn a turbine that drives a generator to produce electricity. This process produces lots of waste heat, which must be dissipated to prevent overheating, and so even more water is needed to cool things down. In rethinking our electricity landscape, water is a key factor.
Which is more important: Meeting energy demand, lowering carbon emissions, or conserving water? How about all three?
The Three Big Challenges Facing the Electric Power Industry
The U.S. electric power industry has huge challenges to meet in the coming decades.
First and foremost it has to meet growing demand for electricity. By 2050 it is estimated that U.S. demand will grow by about 37 percent*.
And then there's climate. In 2012 the U.S. electric industry emitted 2,039 million metric tons of carbon dioxide (CO2) -- that's almost 40 percent of all U.S. energy-related emissions. To meet the climate commitments made in the Copenhagen Accord, U.S. emissions will almost certainly have to come down substantially even while total power generation goes up.
And finally there's water. The electric industry uses huge amounts of water, primarily to cool the excess (steam) heat produced in thermoelectric plants. "In 2005," the U.S. Geological Survey reports, "about 201,000 million gallons of water each day ... were used to produce electricity (excluding hydroelectric power)" with 71 percent** of that supply coming from freshwater sources.
Even though a good portion of the water withdrawn for cooling is returned to surface waters, albeit at higher temperatures, electric power represents a large strain on water resources in many regions of the country. Moreover, the industry"s dependence on water makes it highly vulnerable to water shortages. And with growing consumer demand for water and the prospect of more intense droughts as a result of climate change, that vulnerability will likely become more acute over time. Already, recent droughts and high temperatures have caused power companies to curtail electricity generation. (And this isn"t just a U.S. problem.)
So, in addition to increasing production and lowering emissions, the electric industry must find a way to reduce its water withdrawals. Figuring out the best way to do all three is critical, and decisions need to be made soon. In all likelihood much of the nation's fleet of power plants will have to be replaced between now and 2050. If wrong infrastructure decisions are made in the next decade, the investors of the most capital-intensive [pdf] industry in the country -- and we -- will have to live with those bad decisions for a long, long time.
A Look at Texas for Answers
So how should the industry design its future infrastructure? Can we let economics dictate things and have confidence that carbon emissions and water usage will be taken care of? Should reducing carbon emissions take precedence over conserving water, or the other way around? To get a handle on these questions, Mort Webster of MIT and co-authors carried out a series of model simulations to investigate how the industry should best address the "three-way tension among efforts to meet growing energy demands while reducing greenhouse gas emissions and water withdrawals." They reported their results last week in the journal Nature Climate Change.
Using a case-study approach, the authors focused on electricity generation in Texas. Why Texas? With both population and electricity demand on the rise [pdf], its own independent grid system (see also here), and the state’s vulnerability to water constraints from extended droughts, as epitomized by 2011"s lowest rainfall on record, Texas is a veritable poster child for the three-way conundrum Webster et al. set out to address. (See here too.)
Webster et al. considered three hypothetical scenarios for how Texas's electric sector might evolve to meet demand in 2050:
- The industry would not at all be constrained by carbon emissions or water usage.
- The industry would cut carbon emissions by 75 percent by 2050 but would be unconstrained by water.
- By 2050 the industry would cut carbon emissions by 75 percent and water usage by 50 percent.
In each case the authors assumed that cost optimization would be used to design an electric-generating system capable of meeting a peak demand of 136 gigawatts.
What the Future Could Look Like Under Different Scenarios
Not surprisingly, the makeup of electricity sources for scenario 1 differed quite a bit from the mix for scenario 2. But what surprised me was the difference in the mix of power sources between scenarios 2 and 3. Here's a summary of the findings:
- Without constraints on carbon or water, the most cost-effective mix would consist of roughly equal amounts of coal and natural gas, with natural gas split between combined-cycle units, which require water cooling, and the (currently much rarer) combustion turbines, which do not. Average CO2 emissions would be about 270 million metric tons and water withdrawals about 380 billion gallons annually.
- Limiting carbon emissions by 75 percent sharply curtails (but does not eradicate) the use of coal, and introduces a significant amount of nuclear power into the mix. Interestingly, this scenario, while obviously reducing emissions, actually leads to about 65 percent more water usage than scenario 1. The reason: nuclear power's large thermal inefficiencies require more water cooling than coal-fired power generation. So Webster et al.'s calculation suggests that a policy focused on carbon emissions alone could have the unintended consequence of greater water withdrawals. Not exactly a win for the environment.
- Limiting both carbon emissions and water withdrawals takes coal completely out of the mix, but leaves a significant share for what the authors call a "nuclear hybrid," meaning nuclear power plants that use a hybrid cooling system involving both water and air. It also pushes virtually all the natural gas generation to thermal plants or "dry" combined-cycle plants and it introduces a small but non-negligible role for wind.
A Good Start for Power Plant Planning but …
Fascinating stuff, but a couple of caveats. I wish the authors had worked in the cost of environmental externalities and included more economic information in all scenarios, specifically how much more scenarios 2 and 3 would cost the industry and consumers relative to 1.
And then there's the potential role of end-user and demand-side strategies -- strategies involving efficiency and load shedding (power companies’ interruptible rate programs). As David Brewster, co-founder of the Boston-based energy company EnerNoc (and, full disclosure, a Nicholas School alum [video]) told TheGreenGrok: "If you look at the overall electricity grid, it is a highly inefficient system, and, in fact, about 10 percent of the power plants that we build are only utilized one percent of the time." Aggressively implementing strategies to make the system more efficient could change the electric power supply-and-demand landscape and ameliorate much of the need for building new power plants. However, the models Webster et al. used were unable to consider such strategies***.
And don't forget that prognosticating about the technologies we'll be using in three to four decades is a dicey proposition. Lots of really clever folk out there are working 24/7 to make all those predictions about how we generate electricity today horse-and-buggy news.
Still Webster et al.'s overall message is well taken: in designing the electric grid of the future we should be very thoughtful before we leap -- it's not all about economics. And when you think about it: Not a bad recipe for living sustainably in the 21st century.
* The Energy Information Administration projects an annual growth rate of about 0.9 percent out to 2040, which we have extrapolated to 2050.
** As per this chart, 143,000 million gallons of freshwater are withdrawn for thermoelectric generation. This number divided by the total amount of water withdrawn for electricity (201,000 million gallons) is 71 percent.
*** While their model did not address demand-side savings directly, Webster et al. attempted to get at this in the supplemental information via two reduced demand scenarios, the results of which were qualitatively the same as the main findings in terms of fuel mix.