petrol boiler shenanigans

[bjp]

TL;DR:

In my dlc game I was thinking about tossing together a petrol boiler now that I have the oil wells capped.  Instead of just re-doing my normal cookie-cutter boiler+ex-changer
 

Spoiler

I wanted to try something different.  This post is a very messy summary of what I found out and the two new (to me at least) designs I came up with for a petrol boiler.

# Searching for a more interesting petrol boiler

Looking around, I was inspired by [this thread] about a waterfall heat ex-changer.  It takes about a really neat heat ex-changer that can be used in place of the typical zig-zag/stair/whatever heat ex-changer.  At the end of the thread is also a survival example of a build by @blakemw .

## Testing

I don't really have any experience with waterfalls, so after [reading up] I did some testing in debug and found that they are pretty easy to create.

Some of my test setups:

Spoiler

 

I've read everything I can on heat ex-changers and (oni) heat transfer in the past, but I still don't know how to estimate heat ex-changer performance beyond trial and error.  Using this "method",  I found that a waterfall heat exchange was *much* better than the typical zig-zag ex-changer (this is mentioned in the thread).  

From my testing, my old boiler's heat ex-changer would raise the temperature of oil from ~98.8dC to ~370dC at 10kg/s.  This required 5 rows of aluminum radiant pipes (61 total segments).  The waterfall ex-changer was able to match this performance with only 10 segments, which made it the same height as the old one (accounting for separation tiles) but *much* thinner.  I don't think I'll ever build the old ex-changer again....

I didn't test it very thoroughly, but you could probably shorten it by splitting the waterfall:

Spoiler

## Design considerations

### Stopping/Starting

I'm a huge fan of designs that can be stopped and restarted without too much hassle ("batched" mode).  I usually turn off my boiler when my petrol tank is full.  This is one of the reasons I really like @Gamers Handbook 's [boiler design] (and my modified version).  This doesn't change with the new heat ex-changer, but as @mathmanican mentioned in the waterfall thread, there's not much thermal mass to maintain temperature so restarting can stutter a little bit but eventually smooths out.

The main issue is restarting the waterfall.  @blakemw mentioned in the thread that he used a bead of lead to create the waterfall and that it occasionally gets displaced and he needed to build a lead wire that melted in place.  I did see this while testing.  What's more, I kept seeing the lead phase between liquid and solid a lot.  Either way, I wanted a no-intervention way of restarting the waterfall.

As mentioned in @mathmanican's hydrodynamics thread, you can create a waterfall by using a mechanical door.  With a hydro sensor you can make it re-create the waterfall automatically as soon as the petrol starts flowing.  This worked extremely well in testing.

Spoiler

The only part that was necessary was the hydro sensor on top (>1kg) which opened the door when there was enough liquid to create a waterfall.  The rest of the automation was a backup.  If the petrol was flowing but there still wasn't a waterfall, it would toggle the doors every 5s.

*edit:* After writing that previous paragraph, I realized the hydro sensor could be moved to just above the door (i set it for >0), then it will auto-correct without the need for the extra automation:

Spoiler

### Maintenance

Maintenance is simple.  It's a very open design so you can just build ladders next to the waterfall when you need to.

### Failure modes

The 2 main failure modes of a petrol boiler are running out of oil and running out of heat.

* oil source failure: Since the boiler can handle running in batches, this is the same as simply stopping and restarting when the oil source is remedied.
* heat source failure: The temp sensor simply shuts off the oil valve if the heat source can't keep up.  It restarts itself after the heat source is fixed.  If the heat source is just weak, this also automatically boils the oil in batches.

### Steam buffer

GH's original design (and my oldest ones) has the boiling chamber pull heat directly from the heat source (magma).  For a very hot source, this can inject heat very quickly, causing large spikes and oscillations in temperature.  This wastes heat energy and can risk popping your pipes if you have a good heat ex-changer.

Because of this, I started using an intermediate "steam buffer" between the heat source and the boiling chamber.  It let's you control the temperature gradient, and therefor the rate at which heat is injected.  You can tune it so that the petrol temperature *actually* never exceeds what you set on the temp sensor.  I find a 1T block of steam at ~500 works well for this setup.  It is also too low a temperature to create sour gas, which is a nice benefit.

*note* A hotter steam buffer will reduce the stuttering on start/restart, but wastes energy.  Also, if your heat source is weak you don't need a steam buffer.

## The design

As mentioned before, the design is @Gamers Handbook's simple boiler design, but with the following modifications:
* 2 temp sensors in the boil chamber to control the throughput.
  * The one on top controls the oil valve (>401 for this setup).
  * The one on the bottom controls the heat injection door (>403 for this setup).
* A 1T steam buffer at 500.
* A auto-correcting waterfall heat ex-changer.

Advantages:
The waterfall ex-changer itself has a lot of advantages (listed by @mathmanican in the waterfall thread), but the whole design itself has a few:
* Robust:
  * No popped pipes (even when i shut the valve and let the oil sit in the pipes).
  * No sour gas.
  * No oil dumped in output.
* Can be started/stopped pretty easily (batch-able).
  * It is completely self restarting, you just need to enable the flow of oil into it.

### Initial Startup

The only thing you need to do for an initial startup is to to fill the boil chamber with enough liquid to make 2 tiles of petrol (>750kg).  This way both temp sensors will remain in liquid.  After that it should be entirely automatic.

I'm going to start using this design in my games and see how it goes.

 

# beyond a good design, into the realm of crazy (or how i wasted a whole day playing in debug)

While reading the waterfall ex-changer thread, I noticed a [post] by @mathmanican  where he was boiling petrol without a boiling chamber.  

Now I've heard of flaking before, and experienced it in the oil biome, but I've never seen anyone show a practical use of it outside of a debug map. To be fair, his was also a debug setup, but it made me wonder if I could make a survival petrol boiler work off of flaking.

I didn't really understand flaking very well (still don't completely), but from what @mathmanican said and what I could [find], as long as a 5.01kg *bead* of oil passes by the block of hot igneous rock in the wall, and it is above some minimum temperature, it will instantly flake to a 5kg bead of petrol.  

This is pretty easy to do in debug, but in survival you have to get an initial batch of oil above the minimum flaking temperature to kick start the ex-changer.  Once started, the ex-changer will keep the oil well above the required temp.

Also, if I wanted to maintain my boiler's ability to batch/restart, I would need to make this pre-heating step easy, if not automatic.  

I spent way too many hours trying different designs trying to get this to work, and almost gave it up as completely impractical.

For all designs I used a simple steam buffer at 450dC that injects heat into an Igneous tile with tempshift plates and a conveyor bridge.

My first (failed) attempt, was to put the flaking block above the old boiler as-is, and let normal boiler operation get it started.  Once flaking starts, the boiler chamber shouldn't be needed again unless I need to restart it:

Spoiler

This failed miserably.  The thing about the bead of oil/petrol falling down is that... it's a bead.  As soon as it hits my 1-tile wide boiler chamber it acts like a bead pump and ejects the contents, throwing liquid everywhere, messing up the temp sensors, and popping pipes.

My first thought was to break the bead with an airflow tile.  I tried a few different configurations of this, and while I was able to break the bead, it was still very unstable and threw a lot of liquid everywhere.

I decided to abandon the boil chamber and moved to a ledge of diamond that I could use to try and boil some petrol to get it started.  Once running this worked great, but the startup was very finicky, messy (dropped oil), and popped some pipes.  It required sending small amounts that would boil on the ledge before rolling off and then increasing it.

Spoiler

After my best design (bucket design below), I was able to come back and 'fix' this design.  You can make the startup painless by building a solid block to 'catch' the the initial oil (like in the bucket design).  Once the petrol starts flowing you can mop the oil and destroy the block.

Spoiler

*note* Sometimes when i tried to start the 1 tile version, it would throw oil everyone on start, but sometimes it wouldn't... not sure why (bead pump effect?).

The temp sensor is to control pre-heating to 200dC, and the element/hydro sensor shuts off the oil valve if the heat source fails (we'll start seeing oil instead of petrol).  If you don't want the safeguard, you can make it a tile thinner, but it requires more building/deconstructing on setup.

Spoiler

The design that I finally came up with and really liked was this little bucket design:

Spoiler

Like the ledge during setup, there is a pocket to catch the initial oil, and once the petrol starts flowing it just flows over the oil.  The 2 tiles lets the liquids move sideways so there's no jumping.

The temp sensor is to control pre-heating (>200), and the hydro sensor shuts off the oil valve if the heat source fails.  

The reason I like this one better is that once it's built, it's completely automatic, you just turn on the flow at 5.01kg/s and it will start itself.  The only thing that needs to be changed is the hydro sensor (set it to whatever its reading is + 10 after it's running), but even this is optional.  

If you don't like the look of the oil, it's quantity is small enough that you can mop it, though then you have to wait for the bucket to fill with petrol before it starts flowing again.  This also changes how you detect heat source failure.

The hydro/element sensor can be removed from both designs if you don't care about heat source failure.  Also, the temp sensor and heat injector (steel door) can be removed after start.  Their only purpose is to warm up the thermal mass (and protect against warming up too much).

The practical differences between the two designs are that the bucket is easier to start, and it keeps oil out of the output.

### Advantages

Both designs (ledge and bucket) can restart instantly as there is more than enough thermal mass to bring new incoming oil to the minimum flaking temp.

Using flaking means you don't have to hit a narrow temperature window to get good thermal efficiency, your "donor" just has to maintain a temp hotter than (i think) 406.  You should also never see popped pipes or sour gas.

### Disadvantages

The only downside I wasn't able to address was the max throughput of 5kg/s, but sometimes that's enough.

You also need to make sure you don't get packets smaller than 5.01kg.  You can use a liquid reservoir to buffer the input.

 

Hopefully I didn't miss anything or get anything too wrong.  If anyone has any ideas for improvement let me know.

 

 

 

 

[nakomaru]

I like the idea of splitting it. I also definitely want to try flaking, as that looks like a great way to isolate the boiler from the exchanger. Since flaking doesn't care about TC, but does care about HC, you could use a very hot igneous rock insulated tile to safely transfer heat right next to your pool.

I recently put this together to nicely improve upon the efficiency of a straight shot waterfall. (~96.5% heat recovery at 10kg/s) (requiring about 40% the heat to boil compared to a waterfall - lots of math regarding heat exchangers there if you are interested) Essentially just a bunch more stuff contacts each liquid tile, and we also triple the piping contact area. A door can be used instead of naphtha as you know. 

[bjp]

Oh nice, I don't know how I missed that thread.  Your setup looks really cool, although it uses a lot of material.  No conveyor bridges? :b

[nakomaru]

Later on I add them to get the 390C number. This was just a nicer picture.

The insulated tiles and airflow tiles are not needed, just two metal tiles per row and three times the radiant pipes.

[ghkbrew]

I don't know if you ran in to this, but one issue I had with testing a flaking boiler is that you can't have the oil too hot coming in.  It has to be at least 3C (IIRC) below the nominal "evaporation" temperature.  This limits the theoretical efficiency of the design.  With a long aluminum pipe waterfall exchanger it's easy to preheat the oil too well and break the boiler. 

My other complaint was that you have to run it at just over 5kg/s.  That's a really awkward flow rate.  It's about half the max flow rate of a pipe, but you can't just use two of them, because then the second boiler only be getting 4.999 kg/s and wouldn't flake.

Lastly, I don't see much reason to optimize a design for intermittent running.  The whole point of a petroleum boiler is that it make  the reservoir -> boiler -> petrol gen -> reservoir loop resource positive.  You should just run your generators flat out once your reserve tank is full. Otherwise you're missing out on potential resource production.

[bjp]

1 hour ago, ghkbrew said:

I don't know if you ran in to this, but one issue I had with testing a flaking boiler is that you can't have the oil too hot coming in.  It has to be at least 3C (IIRC) below the nominal "evaporation" temperature.  This limits the theoretical efficiency of the design.  With a long aluminum pipe waterfall exchanger it's easy to preheat the oil too well and break the boiler. 

My other complaint was that you have to run it at just over 5kg/s.  That's a really awkward flow rate.  It's about half the max flow rate of a pipe, but you can't just use two of them, because then the second boiler only be getting 4.999 kg/s and wouldn't flake.

Lastly, I don't see much reason to optimize a design for intermittent running.  The whole point of a petroleum boiler is that it make  the reservoir -> boiler -> petrol gen -> reservoir loop resource positive.  You should just run your generators flat out once your reserve tank is full. Otherwise you're missing out on potential resource production.

I didn't hear about the "too hot" issue, but my ex-changer didn't get that close.  The 5.01kg/s is weird, I was trying to think of a way to stack the remainder for a second one but you would have an empty packet every now and then.

 

As for running flat out, I know a lot of people like to do this but not everyone does.  I like to throttle when I'm overproducing (not just power, but water too) instead of just dumping excess into space or something similar.

[ghkbrew]

2 hours ago, bjp said:

The 5.01kg/s is weird, I was trying to think of a way to stack the remainder for a second one but you would have an empty packet every now and then.

Yeah, you can do it with a reservoir pretty easily.  And you still end up with 99.8% up time, not too bad, just a hassle.  I don't remember what the issue was but I also wasn't able to get 2 stacked bead droppers in the same boiler even with fluid packet stacking.  Have you tried that?

 

2 hours ago, bjp said:

As for running flat out, I know a lot of people like to do this but not everyone does.  I like to throttle when I'm overproducing (not just power, but water too) instead of just dumping excess into space or something similar.

Fair enough.  I do find it tends to be simpler to slightly over-produce and void the excess (or put it in infinite storage).  It gets complicated when you're using 3 outputs from the same production (water, power, CO2).  You have to exactly match your usages or you end up needing to void something anyway (or running short of something).

Not that it's too big of a deal.  The waterfall heat exchanger is actually really good for periodic/batch usage.  It's warmup time is basically just how long it takes the petroleum to fall which is just (height / 5) seconds.

[bjp]

36 minutes ago, ghkbrew said:

Yeah, you can do it with a reservoir pretty easily.  And you still end up with 99.8% up time, not too bad, just a hassle.  I don't remember what the issue was but I also wasn't able to get 2 stacked bead droppers in the same boiler even with fluid packet stacking.  Have you tried that?

I did try that in one of my tests.  For some reason the second beat simply wouldn't flake so I got a mix of petrol and oil at the bottom.  Not sure why.

 

[nakomaru]

This version gets 10kg/s, except for once every 500 seconds where it skips 2x5kg in a row. No reservoir - using an in-line packet stacker. It might miss sometimes, not sure yet.

[ghkbrew]

14 minutes ago, nakomaru said:

This version gets 10kg/s, except for once every 500 seconds where it skips 2x5kg in a row. No reservoir - using an in-line packet stacker.

Nice, but its not actually flaking all the beads. There's a bead of crude oil that reaches the bottom in the gif. What temperature is it at?

[nakomaru]

Those are phantoms that didn't get flashed, I think. They disappear as they have less than 10g.

I tried with a variety of temperatures. For example 276C oil will flash against the 442 to 449 C ceramic tile. Also 393C.

Sorry, stuff does get by. Pretty tough to be perfect on this.

[bjp]

The little packets that get thrown off and disappear were mentioned in one of the threads I referenced.  When I tried stacking one vent above the other I ended up with full 5kg beads flowing past and not flaking.  I couldn't figure out why, they were hot enough, and it was only occationally, so the pool of oil at the bottom would slowly build up.

[ghkbrew]

44 minutes ago, bjp said:

The little packets that get thrown off and disappear were mentioned in one of the threads I referenced.  When I tried stacking one vent above the other I ended up with full 5kg beads flowing past and not flaking.  I couldn't figure out why, they were hot enough, and it was only occationally, so the pool of oil at the bottom would slowly build up.

I think a hybrid design with flaking above a regular hot plate could work. Most of the oil would flake and then the little bit that gets by would be boiled normally. The main advantage would be that you could run cold oil on startup with before the exchange is up to temperature. But fast warm up is one of the strength of the waterfall exchanger. I dont know that it'd be worth it.