Published on August 5th, 2011 | by John Farrell8
When Picking Solar Power Options, It’s the Water, Stupid
August 5th, 2011 by John Farrell
Concentrating solar has promised big additions to renewable energy production with the additional benefit of energy storage — saving sun power for nighttime — but there’s a catch. Most of the new power plants are big water users despite being planned for desert locations.
With solar PV prices dropping so rapidly, does concentrating solar still make sense?
Concentrating solar thermal power uses big mirrors to focus sunlight and make electricity. Think: kids with magnifying glasses, but making power instead of frying ants. The focused sunlight makes heat, the heat makes steam, and the steam powers a turbine to make electricity. In “wet-cooled” concentrating solar power plants, more water is used to make power than in any other kind of power plant. The following chart illustrates the amount of water used to produce power from various technologies.
Water consumption can be cut dramatically by using “dry-cooling,” but this change increases the cost per kilowatt-hour (kWh) of power generated from concentrating solar power (CSP). In the 2009 report Juice from Concentrate, the World Resources Institute reports that the reduction in water consumption adds 2-10 percent to levelized costs and reduces the power plant’s efficiency by up to 5 percent.
Let’s see how that changes ILSR’s original levelized cost comparison between CSP and solar PV. Here’s the original chart comparing PV projects to CSP projects, with no discussion of water use or energy storage.
To make the comparison tighter, we’ll hypothetically transform the CSP plants from wet-cooled to dry-cooled, adjusting the levelized cost of power.
Using the midpoint of each estimate from Juice from Concentrate (6-percent increase to levelized costs and 2.5-percent efficiency reduction), the change in the cost per kWh for dry-cooling instead of wet-cooling is small but significant. For example, all three concentrating solar power projects listed in the chart are wet-cooled power plants. With a 6% increase in costs from dry cooling and a 2.5% reduction in efficiency, the delivered cost of electricity would rise by approximately 1.7 cents per kWh.
The following chart, modified from an earlier post, illustrates the comparison.
With the increased costs to reduce water consumption, CSP’s price is much less competitive with PV. A distributed solar PV program by Southern California Edison has projected levelized costs of 17 cents per kWh for 1-2 MW solar arrays, and that a group purchase program for residential solar in Los Angeles has a levelized cost of just 20 cents per kWh.
In other words, while wet-cooled CSP already struggles to compete with low-cost, distributed PV, using dry cooling technology makes residential-scale PV competitive with CSP.
But there’s one more piece: storage.
To make the comparison comparable, we’ll add the cost of 1 hour of storage to our two PV projects, a cost of approximately $0.50 per Watt, or 2.4 cents per kWh. The following chart illustrates a comparison of PV to CSP, with all projects having 1 hour of storage (Nevada Solar One has been removed as it does not have storage).
When comparing CSP with storage (and lower water use) to PV with battery storage, we have a comparison that is remarkably similar to our first chart. Distributed PV at a commercial scale (1-2 MW) is still cheaper than CSP, but residential PV is more expensive.
Even though dry-cooled CSP competes favorably on price, it still uses much more water than PV. That issue is probably why many solar project developers are switching from CSP to PV technology for their large-scale desert projects.
Concentrating solar thermal power had its moment of cost advantage a few years ago, but the rapid pace (and zero water use) of solar PV installations has quickly eroded even the energy storage advantage of CSP.
Check out our new 93-page EV report, based on over 2,000 surveys collected from EV drivers in 49 of 50 US states, 26 European countries, and 9 Canadian provinces.