In recent months, I’ve been digging into pumped hydro storage of electricity, specifically closed loop pumped hydro storage. This was triggered in part by the Australia National University study on global potential resource for the technology, which found that there was about 100 times the resource as required by global electrical demand in locations that were close to transmission at that, and 250 times as much in the US.
One piece CleanTechnica published made a fun pitch to Elon Musk to tie the Boring Company’s underground tunneling expertise to Tesla Energy’s storage business with the complementary pumped hydro storage technology. Zach Shahan and I discussed that on an episode of CleanTech Talk recently. Another made a more serious pitch to Democratic candidates for President to include pumped hydro storage in their climate action plans. And another piece viewed the latest Mark Z. Jacobson study out of Stanford on 100% renewables by 2050 partly through the lens of pumped hydro storage.
Through this process, I’ve had fascinating and deep conversations with pumped storage hydro developers in the United States and Scotland, social license communication experts in the US interested in this space such as Mike Casey of Tigercomm, and Australian academics working in the space. The last is the focus of this article, specifically the work of Phil Connor and his team out of New South Wales there and in India. Connor is an electrical engineer who worked with Australia’s excellent CSIRO federal research agency for close to three decades.
Connor’s focus for almost two decades has been on floating solar on hydro reservoirs. He was able to connect to one of India’s leading conglomerates, Tata (which is a fascinating megacorporation that was founded to give back as much as it could and is India’s largest car manufacturer), and have access to one of its hydroelectric reservoirs to test his theories and designs in the 2000s.
There are several pros and cons to putting floating solar on pumped hydro reservoirs: evaporation control, panel efficiency, reuse of transmission connections, volatile water levels, shadowed reservoirs, water movement, and relative cost.
The first is reducing evaporation. This is a pretty serious drain on hydro reservoirs.
Lake Mead, on the Colorado River in the southwestern United States, is a case in point. It loses six feet of water level to evaporation annually, in a region that regularly experiences deep droughts. That’s on a reservoir that’s 247 square miles | 640 square kilometers, and 120 miles | 191 km long. That’s a lot of water lost to evaporation, about 600,000 acre feet or 740 billion liters annually. That’s almost 300,000 Olympic Swimming pools worth.
That gives a sense of the scale of the problem, but also a specific challenge. One of the biggest solar parks in the world, the Longyangxia Dam Solar Park in China, is a fraction of the size of Lake Mead. Its 4 million panels only cover 25 square kilometers, about 4% of the area of Lake Mead, and it has an 850 MW capacity.
Evaporation is a function of three things, more or less. Temperature, surface area (especially as a ratio of volume), and wind. The higher the temperature, the bigger and shallower the reservoir, and the higher the wind, the greater the evaporation.
Lake Mead’s location in a hot part of the US, and one that’s getting hotter due to global warming, as well as its massive size, lead to the massive evaporation. It’s also so big that covering it in solar panels is an absurd idea. Presuming the same ratio as the Chinese site, it would suggest a capacity around 21 GW and a price tag in the trillions. And it would eliminate recreation on the Lake, which is a very big use case. That doesn’t mean that there’s no value in covering a portion of it.
It’s difficult to say how much evaporation would be saved by the floating solar array. It’s not an area that is well described in the literature. Connor suggested that evaporation savings might be 40% to 50% in a semi-desert environment with full reservoir coverage, but didn’t have solid numbers or references for it.
For pumped hydro storage, however, there are two aspects of the evaporation equation which are diminished. The first is that the reservoirs are small. Let’s take the example from the ANU study:
“PHES system with twin 100 hectares (ha), 1 gigalitre (GL) reservoirs separated by a height difference of 500 m is able to contribute 1 gigawatt-hour (GWh) of storage capacity (assuming an usable fraction of 85% and an efficiency of 90%), or 200 MW of power with 5 hours of storage to the electricity system – equivalent to a large gas-fired power plant.”
One of the big points of pumped hydro storage is that the greater the vertical distance between the upper and lower reservoirs (the head), the smaller the reservoirs can be, and the lower the volume of water required for substantial storage. 100 hectares is a square kilometer, about 40% of a square mile or roughly 250 acres. It’s about 0.15% of the size of Lake Mead. If there’s water in the upper reservoir, it isn’t in the lower reservoir, and vice versa in a closed loop system. Evaporative losses are about 600 times lower by that measure.
And temperature diminishes with height. The upper reservoir is about a third of a mile higher than the lower in this example. All else being equal, that means that the upper reservoir will be in air that’s about 5.7 degrees Fahrenheit or 3.2 degrees Celsius cooler. Since the point of pumped hydro is storage, and that’s in the upper reservoir, a lot of the time the water will be in cooler air. Pulling on that thread, in general water is pushed uphill at night when demand is low, and flows downhill during peak periods during the day, which tend to be late afternoon more of the time. The combination suggests that the water will spend more of the hottest parts of the days in the upper reservoir.
But the scale of the reservoirs are much more amenable to floating solar projects. While covering Lake Mead with floating solar is an absurd idea, covering a couple of 100 hectare ponds with floating solar is much more in scale. Per NREL, the US sees about 6-8 acres per MW of solar capacity, or 2.4 to 3.2 hectares per MW. Assuming both reservoirs were 80% covered in floating solar, that would suggest a 50-60 MW capacity wind farm, which isn’t bad.
So far so good, but there are still six additional factors to consider, and they aren’t all in favor of the concept. The second article deals with them.