Optimized Steam Engine Setup

[Patric20878]

I never considered water tanks to be useful. Boilers are cheap, there's no major logistical challenge in connecting them to engines, they're small, reliable, so it seems like one of the more inappropriate places in factorio for some kind of buffer (that is, between boilers and steam engines). A chest with coal would be better.

And that's the top reason I'm probably just going to do away with the water tanks completely. All the problems about temperature, pressure, output and everything that comes with trying to attach water tanks to 15-engine rows would simply go *poof* by just storing coal instead. And it'd store way more in coal than tanks can store in hot water too, enough that you'd have more along the lines of HOURS to react to it rather than just 5 minutes lol.

[BlakeMW]

I generally find little merit in storing hot water except to deal with the predictable daily cycle. For backup steam or for "load topping" because you don't know how long you're going to need the extra power for, it's best to have the steam engines able to operate at full power indefinitely, there's no level of hot water storage which is logical or optimal if you don't know how much storage is needed.

A slight semi-exception is "laser capacitors", especially on heavier biter settings if you beeline laser turrets instead of using gun turrets (which IMO is usually a poor decision, but if you have little iron and lots of oil it's a solid choice) then during an attack the lasers can use like 10x the power as the rest of your factory. For this you can build like a 1/14/60 - although in these designs the storage tanks are pretty much optional because a steam engine stores enough water for 2s worth of generation which is usually enough burst electricity to wipe out a biter attack. These laser capacitors are very cheap and significantly reduce repair costs and perhaps save laser turrets from destruction (however steam engines are significantly more iron-expensive per MW than accumulators - so if you have ample oil and copper then logically you'd just use accumulators)

Another niche use which is almost useless in vanilla is for sensitive load-monitoring, you build something like a 2/28/21 - it just needs to be slightly over-ratio on steam engines. Then you add a storage tank and link it to the circuit network. If the storage tank starts being drained it means the power plant can't keep up with the load. In vanilla it's only really useful to activate backup oil power if the coal power isn't enough, or to monitor the "health" of the power plant and output it as a signal for display at another location. But it works great with mods which add alternative base-load generation, such as nuclear reactors, thorium reactor or perhaps hot spring. It's one of the best ways to catch that power consumption is slightly higher than what the boilers can keep up with, without requiring an actual brownout, but in 0.13 we'll simply be able to use an accumulator to detect this condition (< 99% charged).

Btw I tend to think that steam engines are optimally built over-ratio. If your factory consumes say 30MW on average, it actually consumes more than 30MW some of the time and less other times, as factories always have an at least slightly jagged power graphs. So the truly optimal power plant will account for this and have a few percent more steam engines than the optimal ratio. A power plant with 7 pumps and 72 steam engines (6x12) can be considered at least as optimal as one with 7 pumps and 70 steam engines. Storage tanks are not required for this micro-load-balancing but they don't hurt. Of course, as usual accumulators do the same thing.

[Patric20878]

Btw I tend to think that steam engines are optimally built over-ratio. If your factory consumes say 30MW on average, it actually consumes more than 30MW some of the time and less other times, as factories always have an at least slightly jagged power graphs. So the truly optimal power plant will account for this and have a few percent more steam engines than the optimal ratio. A power plant with 7 pumps and 72 steam engines (6x12) can be considered at least as optimal as one with 7 pumps and 70 steam engines. Storage tanks are not required for this micro-load-balancing but they don't hurt. Of course, as usual accumulators do the same thing.

But then you're dropping output by 2 steam engines. It has its own buffer like you said to account for slight variations in consumption.

And yeah, in the end, water tanks still can't act as capacitors for laser defense like accumulators can, so I'm removing the water battery variant from my setup.

[BlakeMW]

You're not dropping output. The point is that for an exactly on-ratio power plant then under normal operation water will inevitably accumulate in the steam engines and without an extra steam engine that water cannot be consumed. It is water you've paid to heat, but cannot make any use of.

10 steam engines store 8.5MJ of hot water (85kJ * 10 * 10), equal to 1.7 accumulators. An extra steam engine to consume that hot water is thus equivalent to 1.7* accumulators (also coincidentally in wattage, as 510kW / 300kW = 1.7).

* 1.87 accumulators when you take into account the up to 10 water stored in the 11th steam engine.

So throwing on an extra steam engine is equal to 1.7 accumulators.

[Patric20878]

14 boilers let you heat water fast enough, without the excess, it doesn't heat 10 engines well. What do accumulators have to do with this? 72 steam engines produce 1020 more KW than 70.

[Arpus]

I just wanted to ask why 21 boilers per 15 engines. Wouldn't 20 be enough.
According to wiki 1 engine needs 1,31 even if it was rounded down and i counted with 1,32 then 20 is still enough(20/1,32=15,15).

[Qon]

I just wanted to ask why 21 boilers per 15 engines. Wouldn't 20 be enough.
According to wiki 1 engine needs 1,31 even if it was rounded down and i counted with 1,32 then 20 is still enough(20/1,32=15,15).

I tested a .5GW powerplant with 21 boilers to 15 engines and all boilers are active all the time, the last one seems to burn slower though.

[Patric20878]

I covered why not to use 20 boilers in the last few pages of this thread. It can't heat up the water to max power, hovering at some 45.6/45.9 MW instead. Mathwise, 20 should be able to support 15, but water pressure probably has some factor here that causes it to not be so simple of a division.

[Qon]

0.13 version is here!
New ratio is 3/29/19

Still tileable, works with electric inserters if you prefer. But with rocket fuel and stack ability pickup for all inserters from belts burner inserters are going to idle so much that their 0 energy drain will make their energy draw less than that of electric ones! Electric ones now have no benefits over burners.

So you are probably going to need landfill to get a shore like this. Good thing it's a vanilla object now. And it's super easy to make that shore, a single tile wide line is all that is needed.

You can probably compress it slightly by shifting some steam engines. But the overall design is almost finished. If I don't post a version with slightly better steam engine placement soon I expect Patric to do it q:

Because landfill is now canon.

LandfillSteam_0.13.jpg (485.94 KiB) Viewed 8059 times

Should be easy to use an express belt of ROCKET FUEL instead
Just shift it some tiles right and split off with an express splitter to basic belts for a maximum of 4.5 GW power/blue belt!

Edit:

Here it is

LandFillSteam_RocketFuel.png (1.15 MiB) Viewed 8057 times

More improvements possible

Probably a bit overkill on offshore pumps but whatever it works and pumps are cheap. Remove some if you want. Now hitting the pipe limit so not all offshore pumps are used at max... Maybe the extra water can be piped past the engines like my earlier post for even more power/shore length! Then you could even add 1 more row of offshore pumps further out shifted one tile, or combining 2 pumps facing in opposite direction and then undergorund pipe it to land for even more water.

[BlakeMW]

Hmm, are you sure that works? I haven't managed to get more than about 24 boilers in a row to contribute at full effectiveness, and in your screenie the steam engines are generating 37.2MW which is consistent with the amount of water which can fit down a boiler string, also 3 offshore pumps are no better than 2 in my testing.
Something new enabled by landfill pump-packing though, is running a cold water pipe between the two center boiler strings and joining into them not further than 14 boilers from the end, this slightly improves efficiency, up to about 96% of maximum.

What I've found in general though is there are two general problems with long boiler strings:

• The first is inadequate pipe capacity where the water throughput is bottlenecked and ends up less than that required to operate 20 steam engines.
• The second is excessively fast flow caused by too much "suction" where the liquid is moving too fast for the boilers to contribute their full 390kW of heating and ends up with water at less than 100 degrees even if by the maths there are enough boilers (water only gets heated while it is dwelling in a boiler, it doesn't get energy magically transferred just by passing through).

I you accept reduced boiler efficiency and add even more boilers to help heat the water (say 31 boilers instead of 28) you exacerbate the flow bottleneck problem, so you can only choose between an adequate volume of inadequately heated water, or an inadequate volume of adequately heated water, attempting to improve either variable worsens the other.
Basically, designs with more than 24 boilers in a row just don't work and they are running very close to the limit, 21 boilers is a more sensible limit which gives a little wiggle room in the fluid simulation.

[Patric20878]

What is this landfill madness? What's the tech level needed for landfilling?
And Qon's design produces 37.4 of the 38.8MW it should produce, wonky water simulation/pressure strikes again.

(And why did I sign up to keep updating a design for a game I don't play anymore )

Oh and, unless the landfilling thingy is as low tech as small poles and yellow split belts, it'll be made into a variant rather than an upgrade.

[BlakeMW]

Something I dreamed up is kind of "braided" strings of 14 boilers relying on offshore pump stacking:



This layout gets full efficiency and the individual strands stack well both with direct feed and another belt. One problem though is the maximum steam engine string length is about 25, it gets pretty hard to cram in the steam engines as tightly as the boiler strings without the steam engine strings exceeding this maximum length. A possible solution is to use a bypass underground pipe running between rows of steam engines delivering water to the end of the steam engine strings and extending the maximum length of the strings such that 80 or 100 steam engines can be packed into 3 rows of 26.7 or 33.3 steam engines (occupying the same space as 8 or 10 stacked offshore pumps).

I get the feeling this design is too crazy to use in a real game, when straightforward 1/14 designs are well, straightforward. Nevertheless it possibly offers the maximum possible vertical space utilization (while in principle you could probably braid 2 strings of 21 steam engines, you'd be very hard pressed to fit the steam engines into the same vertical space and would also need to make sure each string of 21 boilers feeds into exactly 15 steam engines, which would require some impressive underground pipe wrangling).

Edit: To answer the landfill question, it's a 50 green science tech which no prerequisites and with the new tech costs I believe it is the cheapest early game green science tech (tied with turret damage 1). Landfilling costs 20 stone per tile.

[Qon]

Hmm, are you sure that works? I haven't managed to get more than about 24 boilers in a row to contribute at full effectiveness, and in your screenie the steam engines are generating 37.2MW which is consistent with the amount of water which can fit down a boiler string, also 3 offshore pumps are no better than 2 in my testing.

Basically, designs with more than 24 boilers in a row just don't work and they are running very close to the limit, 21 boilers is a more sensible limit which gives a little wiggle room in the fluid simulation.

It works. It provides power. It gives more than 15 steam engines in a row. Yes I'll have to optimise it, but as a proof of concept it is good enough.
Yes the pumps are overkill. I said that in my post. The point is that you can fit a lot more pumps now. You can fit even more pumps if you want for even more power if you use tricks like my first proof of concept of using more than 10 steam engines in a row.

Something new enabled by landfill pump-packing though, is running a cold water pipe between the two center boiler strings and joining into them not further than 14 boilers from the end, this slightly improves efficiency, up to about 96% of maximum.

Yes, someting like this is what I'll do when I have completed my design. I might get even more steam engines lined up than 19.

What is this landfill madness? What's the tech level needed for landfilling?
And Qon's design produces 37.4 of the 38.8MW it should produce, wonky water simulation/pressure strikes again.

(And why did I sign up to keep updating a design for a game I don't play anymore )

Oh and, unless the landfilling thingy is as low tech as small poles and yellow split belts, it'll be made into a variant rather than an upgrade.

Landfill is green science, 50 red and green packs.

37.4 instead of 38.8 is good enough for a proof of concept. It's not like that 0.8 extra steam engine at the end is too expensive if not utilised at 100%. But it's a proof of concept of using even more pumps. And you can fet even more that that as I described.

And I enjoy that you stay even if you don't play :>

Something I dreamed up is kind of "braided" strings of 14 boilers relying on offshore pump stacking:

pic

This layout gets full efficiency and the individual strands stack well both with direct feed and another belt. One problem though is the maximum steam engine string length is about 25, it gets pretty hard to cram in the steam engines as tightly as the boiler strings without the steam engine strings exceeding this maximum length. A possible solution is to use a bypass underground pipe running between the steam engines delivering water to the end of the steam engine strings and extending the maximum length of the strings such that 80 or 100 steam engines can be packed into 3 rows (occupying the same space as 8 or 10 stacked offshore pumps).

Yes, that looks kinda similar to what I described. Good idea to interleave them vertically. And yes bypassing the steam engines is what I would do once you get above 15~20 steam engines in a row. I did it before, I can do it again q;
Looks really nice .

[Qon]

New ratio is 3/42/30! Three times as much as the regular setup. And all steam engines are almost at 100% now with 61.1 MW compared to theoretical 61.2.

A bit long though...

Time to compress it further. Even more water can be drawn from the shore still.
You can probably just merge the pipes towards the end for the last half when pressure is lower.

3/42/30

SuperSteam2.png (1.41 MiB) Viewed 8868 times

Closeup

SuperSteam2PumpCloseup.png (3.05 MiB) Viewed 8867 times

SuperSteam2BoilerCloseup.png (3.36 MiB) Viewed 8867 times

SuperSteam2BoilerEngineCloseup.png (3.36 MiB) Viewed 8867 times

[Patric20878]

Heh. As much as the 30 engines per row thing works, you're fitting pipes between practically every engine. That's gotta be removed in a real design.

For now I'll let yar experiment with the changes that come with 0.13. Maybe once you're more familiar we'll work further on seeing how much advantage this can give over my current design.

Oh and, 50 red and green packs? Afraid that won't fit in with the emphasis on minimal tech that came with my design after v1.2, so a setup using landfills will probably be a variant rather than an upgrade.

[BlakeMW]

I've been thinking about rocket fuel in boilers. You lose 10% of the fuel value by converting solid fuel to rocket fuel but otoh you can use productivity modules and with 4x Prod3 module in assembler 3 you get 26% more fuel value - a gross profit of 45MJ of electricity. The net profit is trickier, I think the minimum cost to perform the conversion (fully utilized alternating rows assembler/beacon) is about 15MJ leaving a net profit of 30MJ. Of course, if supplementing with solar power the electricity expense can be further discounted.
Rocket Fuel also saves energy on inserts (for all inserter types, not only burner inserter) but actually increases total factory inserts because the 10 solid fuel had to the inserted into the chemical plant to make the rocket fuel - you're not saving energy on inserts, only changing where those inserts happening and causing more inserts.

[Patric20878]

From a practical and teching up standpoint, I have little interest in rocket fuel optimization with a steam engine setup, as if you have modules to upgrade machines, you should be on solar power already.

[BlakeMW]

From a practical and teching up standpoint, I have little interest in rocket fuel optimization with a steam engine setup, as if you have modules to upgrade machines, you should be on solar power already.

Not necessarily. If you're speed running or something you may not have solar power.

In fact it came up in my game I just played to get the "There is no Spoon" achievement (win game in < 8 hours). The rocket was just being built, and my power quite suddenly dropped to under 50%. I ran over to my larger power plant and discovered the biters had obliterated all the mining drills and the power had gone off once the boilers ran dry. So I had the problem the grid was under full load and I needed the new mining drills to operate at 100% to fuel the powerplant - happily I had a heap of rocket fuel in my inventory (I was saving it for the satellite) which I dumped into the boilers to restore full power immediately.

0.13 does significantly simplify large power plants due to landfill allowing the creation of straight coasts on demand and also specific coast configurations for pump stacking. Additionally the power switch and accumulator logistic connection makes it much simpler to make backup steam - in fact we can now make a fully blueprintable backup steam powerplant - just prepare the coast with landfill and add a fuel supply. This makes solar/accu with steam backup more appealing and the steam side more scalable. Rocket fuel should at least be considered as the backup fuel - although the only real advantage of rocket fuel (besides stretching oil with productivity modules) is with deep boiler stacking, you could stack boilers really really deep, about 5x deeper than was possible with solid fuel in 0.12.

[Qon]

From a practical and teching up standpoint, I have little interest in rocket fuel optimization with a steam engine setup, as if you have modules to upgrade machines, you should be on solar power already.

All the things BlakeMW said and...

Steam is much more compact than solar.The size requirements get quite ridiculous when you want tens of GW of power. Steam is now blueprintable, and resources are now getting richer the further away from your base you go so mining replacement ceases to be an issue if you go place your power plant far away from the origin. With big low frequency patches you can get a GW/coal patch that lasts for tens or hundreds of hours. To build really high throughput factories space requirements can be so high that it can be one of the biggest time sinks. With steam you just build a GW powerplant in minutes. The most time consuming part is setting up coal mining + trains.

The reason people think solar is so good is because the energy doesn't suddenly stop (disregarding accumulators running out at night lol). With 0.13 resource generation the richness can potentially be so high that it won't run out for steam either.

Solar is CPU efficient though since it uses a constant and negligable amount of your cpu time.

Steam is definitly a valid alternative for mega factories and shouldn't be dismissed without analysing the situation properly. I would probably do a mix on both solar and steam depending on how many really big and rich coal/oil patches I find.

[Patric20878]

The tech rocket fuel requires though, lol.
Not gonna bother. And changing all inserters to burners would require no change in physical layout anyways.

And 0.13 increasing resource patch density by distance, welp, that's new. Guess it would make steam more viable late game then.