The Rock Gasifier - A regolith melter for polluted dirt

[ghkbrew]

The Rock Gasifier - My answer to the un-asked question: "How an I turn 10kg/s of pdirt (or sand) into gigawatts of power?"

The theory is simple.  It works by exploiting the fact that molten glass has an SHC of 0.2, but when gasified becomes rock gas which has an SHC of 1.  A 5x increase in heat.

 

The core of the build is essentially a single heat exchanger between pdirt/molten glass going in and rock gas/magma coming out.  It's complicated by the phase change on both sides which occur at different temperatures.

First the pdirt is pre-heated slightly a (1).  This drops the temperature of the magma below it's freezing point to ~1250C.  The now ~900C pdirt is then melted at (2) into molten glass.  The glass is pumped in tungsten pipes at 1kg/s to avoid phase changes.  This let's us increase the energy creation by about 2 fold by increasing the temperature at which it evaporates.  The molten glass is counter-flowed (shut-up that's definitely a word) against the rock gas in (3) bringing it up to ~2924C (just under the melting point of wolframite).

We're limited to a theoretical high temperature of 2926.85 as the highest melting point of any material you can make a liquid vent out of.  If we could somehow prevent the gas from touching the vent, we could go up another 495 degrees to the melting point of tungsten and increase the power output by another 37%.

Area (4) is heater to keep area (2) above the melting point of pdirt while rock gas counter-flow warms up, after which it is no longer needed.

Area (2) is under a moderately delicate balance to keep it in the range where rock gas will condense, but magma won't freeze.  This is the reason for the initial pre-heating in (1).  If the pre heater is much longer or shorter than the current one the build will break.

The shape of the rock gas chamber is the result of some trial and error.  The idea is to avoid heat being conducted down the diamond tile walls (which are the only thing that won't melt at these temperatures).   This shape with diamond tiles actually makes the exchange more efficient than vertical walls of insulated insulation. Presumably because of the decreased gas flow rate.

 

The reservoirs in the heating element are from left to right. 1) 5 tons of molten steel kept at almost exactly 2926C.  2) a reservoir of "cold" steel which has been used to heat the bottom tiles.  3) a reservoir of heated steel fresh from the metal refinery.  and 4) a reservoir of steel waiting to be reheated.  The temperature sensor on the output of reservoir 1 activates the liquid shut off to add more hot steel whenever it drops below 2926C.  An equal amount of "cold" steel will then flow past the bridge input towards the metal refinery to be heated.

With 5T of steel and refining iron to reheat the steel ~350C each time, the output temperature fluctuates between 2626 and 2629C.

Each pass through the diamond tiles drops packet temperature by ~3C which works out to a heat drain of ~12kDTU.  Meaning to keep this up to temperature 1 load of iron ore needs to be refined every 7.5 cycles. Catalina is very bored...

Plumbing

Spoiler

Shipping

Spoiler

 

Automation

Spoiler

 

I'm honestly shocked at how simple this build ended up.  The least feasible part of it is actually finding 10kg/s of pdirt in survival.  As an alternative sand can be used as a drop in replacement since it has the same thermal properties.  Plus with sand you can use a rock crusher to close the resource loop, at the cost of 2-5 dupes punching a button all day.

 

[catrue]

this is compactness. Debag was not surprised yet

[wachunga]

Great stuff as usual.

 

I've been tinkering with the same idea off and on, please excuse the brain dump that follows.

1 hour ago, ghkbrew said:

heat drain of ~12kDTU

At that rate you could power this with a kiln to be extra silly (keep the refinery for preheating though). Partially covering it with molten material is necessary to boost heat output from 4 kDTU to 12 KDTU.

You're probably losing heat to the 1.5C boiling bug. The great thing about counterflowing 0.2 SHC against 1 SHC is that your 0.2 material comes out at virtually the same temperature as the 1 material going in. Which means in theory you only need to account for the 1.5C "heat of vaporization". With 10 kg/s of 0.2 SHC glass, that should be 3kDTU. In my fiddling, I tried this geometry to avoid the bug:

The molten glass exchanges heat with the tile of rock gas before it boils and loses the 1.5C. No heat exchanges with the insulated tiles so extra 1.5C deletions (from the bug) don't happen. In practice, the steady state temperature loss you need to make up for is ~1.8C. The kiln had an uptime of ~1% for 0.12kDTU. That's less then expected. I suspect the simulation or uptime reporting starts to fall apart at this scale. You would of course have to rejigger things a bit to make room for a similar setup.

[ghkbrew]

6 hours ago, wachunga said:

8 hours ago, ghkbrew said:

heat drain of ~12kDTU

At that rate you could power this with a kiln to be extra silly (keep the refinery for preheating though). Partially covering it with molten material is necessary to boost heat output from 4 kDTU to 12 KDTU.

I keep looking for a reason to use kilns as heaters, but every time I realize that a single batch of steel in the metal refinery provides as much heat as 13 cycles worth of kiln time. And it seems too silly even for me

I also have vague plans to use smart storage bins as electric heaters since they don't have an overheat temperature, but you need 100 of them to even match the heat output of a kiln.

There's also the oxylite refinery.  If you feed it super heated oxygen it'll make oxylite that immediate melts into magma, plus 12kDTU/s of heat.  Artificial volcano anyone?

 

6 hours ago, wachunga said:

You're probably losing heat to the 1.5C boiling bug.

Definitely, I tested it

6 hours ago, wachunga said:

I tried this geometry to avoid the bug:

Yup, these two also work:

The second one is interesting.  Because the vents emit the glass simultaneously the liquid tiles don't exchange heat with each other (since they're the same temperature) and avoid the bug.

 

This one is basically what I have in the build:

It deletes about ~8.7C vs ~1.9C for the first two.

I decided to live with the bug, since the heat input is still almost negligible and I really wanted a build that didn't use any insulation (and thus could be built in the current DLC).

I also toyed with the idea of the heating just the pipe (like you're doing) to avoid any excess heat going into the rock gas, but then decided against it because it would double the width of the exchanger since you have route the pipes up, around and back down.  But then I doubled the width of the gas chamber anyway so... maybe I should rethink that decision.

 

[wachunga]

In the back of my mind I knew you went through all the same stuff and I was being a gas bag. 

[MooChiChi]

I like your build and the idea behind it.

On 1/15/2021 at 11:15 PM, ghkbrew said:

Plus with sand you can use a rock crusher to close the resource loop, at the cost of 2-5 dupes punching a button all day.

This is really inspiring to me. Thanks!