Hot Sand

Guest Post by Willis Eschenbach

I live up at the top left of the map in Figure 1, in Northern California between Santa Rosa and the Pacific Ocean. Down the coast on the far side of San Francisco from me is Monterey Bay, and the town of Moss Landing.

Monterey Bay is famous for fish and fishing because there is a submarine canyon that runs all the way in to the shore at Moss Landing. This brings in the deepwater currents with loads of nutrients, which feed a rich marine ecosystem.

Half a century ago, I fished commercially for three years in Monterey Bay, two of them fishing out of Moss Landing. There was a huge old power plant in Moss Landing that was the friend of everyone who fished those waters, because it had two giant chimneys. We fished nights, not days, and at any time of the night, it was infinitely comforting to see the rings of red lights on the chimneys, visible from all over the Bay. They marked home, and land, and safety. Here are the stacks during a full moon.

Now, fifty years later, the power plant is shut down but the chimneys still remain, mute obelisks of an earlier time. You can see their shadows in the upper right of this aerial view of Moss Landing.

And what are the white boxes up at the tip of the shadows of the chimneys? They’re one of the subjects of this post. Those make up one of the largest battery installations on the planet. It’s comprised of hundreds of Tesla Megapack batteries. It stores on the order of 7.3 gigawatt-hours of electric energy (GWh, or 10⁹ watt-hours). Here’s a photo from the ground.

So … what’s not to like about lithium megabatteries?

Well, the first thing not to like is cost. The Tesla Megapacks cost about $327 per kilowatt-hour of storage, a huge amount. And with lithium prices skyrocketing, that will only go up. So building them at grid-scale is stupendously expensive.

Next issue is environmental damage. Lithium mines are not very pretty and are destructive to the environment without special procedures … procedures that are unlikely to happen in the countries where lithium is mined.

Next issue is safety. Here’s a recent story

Second battery malfunction in less than 6 months reported at Moss Landing power plant

7:11 PM PST Feb 14, 2022: MOSS LANDING, Calif. — In Moss Landing, firefighters responded to another battery meltdown at the Vistra Energy Storage Facility Sunday night, when they arrived roughly 10 battery racks were melted.

It’s the second incident at the plant in the last five months alone.

Firefighters say the two incidents should provide a learning opportunity to make any needed adjustments or improvements.

One concern is this plant is going to get bigger.

A Tesla Megapack costs about one million dollars … and ten of them went up in smoke. That’s an expensive “learning opportunity”.

And a final issue is lifetime. Lithium batteries can only be cycled a certain number of times before they wear out and need to be replaced.

---

With that list of the issues with lithium batteries as prologue, folks that know me know that I’m very skeptical about new technologies. I’ve seen lots and lots of “stunning breakthroughs” announced with great fanfare that never made it off of the drawing board.

But today, I came across an energy storage technology that might actually work. Here’s a drawing of the idea. It’s being developed both privately and by the National Renewable Energy Laboratory (NREL). NREL calls its incarnation of the technology the “Enduring” system.

ORIGINAL CAPTION: In a new NREL-developed particle thermal energy storage system, silica particles are gravity-fed through electric resistive heating elements. The heated particles are stored in insulated concrete silos. When energy is needed, the heated particles are fed through a heat exchanger to create electricity for the grid. The system discharges during periods of high electricity demand and recharges when electricity is cheaper. Image by Patrick Davenport and Al Hicks, NREL.

TL;DR Version: Electricity is used to heat sand. When you need electricity, the hot sand is used to boil water to drive steam turbines for electricity.

So why do I think this one is possible? Several reasons:

First, it is very cheap. Instead of using expensive lithium for storage, it uses cheap silica sand. This brings the cost down from the $327 per kilowatt-hour (kWh) of lithium batteries to an NREL estimated cost of $2 – $4 per kWh. And even if the final cost is three times that, it’s still only a few percent of lithium battery cost.

Next, it’s safe. Sand can’t catch on fire. Lithium can, and does, and is very hard to put out once it starts burning.

Next, it’s scalable, and it’s cheap to scale. Add more insulated tanks of sand and you add more storage capacity.

Next, it can be built on the sites of closed coal-fired power plants. All the infrastructure is there—train tracks to bring in the sand, turbines, generators, substations, transmission lines, and the like.

Next, it doesn’t require any new or unproven technology. We know how to heat sand, and how to build boilers and steam turbines, and how to do all the things shown in the drawing above.

So will this be the secret technology that sets solar and wind loose to make an actual difference in the real world? Because up to now, solar and wind ain’t doing diddly squat.

Seems doubtful that it will change things that much. Storage is only one small problem with sun/wind. A much larger problem is that most of the electricity from sun/wind is used immediately, and so there’s not a lot left over to put into storage. Next, both technologies require dangerous/rare/poisonous materials, are short-lived, and are hard to recycle. Plus, wind turbines massacre raptors, for a curious reason discussed here.

And there’s another big problem … there’s not a lot of solar/wind energy there to harvest because it’s so spread out, and many of the good sites are already in use. So this storage technology could help at the margins, but won’t be a revolution.

However, sand storage would still be useful for load balancing on the grid, and should be quick to ramp up and down to meet variations in demand.

There’s already a Finnish company that is commercially testing the technology. It’s called Polar Night Energy, and they’re using the heat directly, not for electricity, for district-wide heating of towns in the far north. Here’s their test installation:

Store heat in the summer when it’s not needed, and release it in the winter when it is needed … works for me.

Anyhow, that’s the good news for today … yeah, I know that compared to the ongoing global lunacy it ain’t much, but it’s what I’ve got.

My best wishes to all,

w.

PS: As always, I politely ask that when you comment you quote the exact words you’re discussing. This lets us all know exactly what and who you are responding to, and it avoids endless misunderstandings.

Technical Note: I ran some numbers to see if this all pencils out … seems like it does. R computer language code and results below. Lines starting with “[1]” are the computer output. Anything on a line after a hashmark (#) is a comment.

(us_electric_consumption = 3.9e15)# watt-hours Wh
[1] 3.9e+15
(moss_landing_battery = 7.3e9)
[1] 7.3e+09
(enduring = 26e9) # enduring storage, watt-hours Wh
[1] 2.6e+10
(ca_electric_consumption = 280e12) # Wh
[1] 2.8e+14
(sf_electric_consumption = 5e12) # Wh
[1] 5e+12
(ny_electric_consumption = 51e12) # Wh
[1] 5.1e+13
(enduring/ny_electric_consumption*secsperyear/3600/24) # days of NY City supply [1] 0.19 (moss_landing_battery/ny_electric_consumption*secsperyear/3600/24) # days of NY city supply, Moss Landing Battery
[1] 0.05225152
(degrees_temperature_swing = 900) # °C
[1] 900

(sand_specific_heat = 800e3) # joules/tonne/°C
[1] 8e+05
(storage = degrees_temperature_swing*sand_specific_heat) #storage joules/tonne
[1] 7.2e+08

(storage_whr = j2wh(storage)) # storage wh per tonne
[1] 2e+05
(tonnes_needed = enduring/storage_whr) # tonne
[1] 130000
(sand_density = 1.6) #tonnes/m^3
[1] 1.6
(volume_needed = tonnes_needed/sand_density) # cubic metres
[1] 81250
(tank_num = 5) # number of tanks
[1] 5
(cube_side = volume_needed^(1/3)) #metres per side
[1] 43.31196
(cube_side_per_tank = (volume_needed/tank_num)^(1/3)) #metres per side
[1] 25.32899
(cube_side_ft = m2ft(cube_side)) #metres per side
[1] 142.0993
(sand_per_ton = 40) # sand cost, $/tonne
[1] 40
sand_cost=tonnes_needed*sand_per_ton
paste0("Sand cost = $",format(sand_cost,big.mark=","))
[1] "Sand cost = $5,200,000"