Thread by @ElephantEating

A conversation off of Cody's tweet brought up Train Energy Storage, which I think is a neat idea but have never run the numbers (and hadn't seen elsewhere either). The question was about whether they could be good seasonal energy storage, so consider this the reality check.

https://twitter.com/cody_a_hill/status/1035270390098972672

Train Energy Storage (TES) is a neat idea - put a bunch of sand into traincars, connect power lines to the (already-electric) locomotive, and pull the train up a mountain to store energy & recover the potential energy (locomotive motor) on the way down. interestingengineering.com/concrete-gravi…

The upside is that this is mostly off-the-shelf technology. But could it be used to store enough energy to be useful for seasonal storage? What is the potential scale? Let's do some math!

The stored potential energy is easy to calculate, just mass X height X 9.8 = Joules

Let's look at some ridiculous scale to start: There are about 500,000 covered hopper traincars in the US. Let's imagine that I filled them all with sand (~140 tonnes/car) and brought them all to the highest point in California (top of Mt. Whitney). How much stored energy is that?

I calculate that it works out to 800 GWh of stored energy. That is a lot - enough to run the entire state for about 35 hours. It is starting to get into the realm of the amount of energy we might want for seasonal storage (for CA). But we probably couldn't build that system.

The highest trainline in CA is at an elevation about half as high, so bringing all the loaded trains to that height would store half as much energy.

And we probably couldn't get every covered traincar in the US. That sounds expensive. Plus it is 9,000 km long, so it reaches not just all the way up our mountain, but across the entire country. Where would we store them all at the top of the mountain?

Let's say we could get 10% of that - 50,000 traincars (still 900 km long!) loaded with sand and took them from sea level to 2,000 m, which is still pretty high. We would have 40 GWh of stored energy, enough to run all of California for a bit more than an hour. Not much.

Trains are amazingly efficient, which should actually be a clue that stored energy is small. Consider: a single tank of diesel can tow a mile long train over several mountains, showing that the fuel has more stored energy that the train's potential energy at the top of the pass.

That illustrates how much more potential energy there is in chemical fuels of any type. Another: Aliso Canyon Field, a natural gas storage facility in California stores 86 billion CF of gas, which is 25,000 GWh of energy - 30X more than the ridiculous TES scenario (800 GWh).

Maybe you don't like natural gas. Ammonia (convertible to hydrogen & zero carbon) has about 4 kWh of energy per liter. Large oil tanks (like at giant tank farms) can hold 150M L. If it was ammonia, that is 600 GWh of energy for one giant tank.

Back to the trains: if I were using them for seasonal storage, what do I do at the top of the mountain? Do I build the world's largest train yard and leave trains there for months at a time? I'm tempted to say no - those train cars are expensive.

Consider that the traincars are the costly part of the system (relative to just sand). So I might propose that I use the train system to move sand up the hill then drop the sand at the top. I would need some sand-handling equipment, but that already exists.

Or, if I'm smart, I might realize that it is easier to handle liquids than sand and the cheapest liquid is water. So I could fill my traincars with water, drive it to the top and pump it into big tanks. That seems more efficient and cheaper than sand.

But at this point, I'm just designing an inefficient pumped hydro system. Which brings me around to the original question: Is train energy storage the right solution for super-large-scale storage? I don't think so. For a large scale, it is hard to see it cheaper than pumped hydro

However, I do hope to see more investigation - it is certainly could be cheaper at scale than many battery systems and an option in places that can't have pumped hydro (though, of course, both require mountains, so...)

One more general point: a lot of times we see a new technology and are tempted to say, "The known technologies that we've done the math on don't provide easy answers, but this new technology that we haven't looked into will surely solve our problems." Usually that is incorrect.

And if you want to learn more about the technology, check out @drvox's writeup here: vox.com/2016/4/28/1152…