Optimized Steam Engine Setup

[Patric20878]

8MJ per lump of coal
/2 for boiler
/1800 for 2 engines at 900 kW each
* 26.67 (or really 40 * 2/3)

equals 59.26

Thanks for the explanation. Then I'm convinced by evidence it's really coincidence - the 26.67 is 2/3 of 40, which is a round even number instead of some seemingly randomly picked decimal that wouldn't be likely at all. Matches with my estimate of 119.5-119.8 also, now that I realize I forgot to multiply the downtime by 2 since 2 engines were fluctuating with coal shortages, making the correct estimate 119.0-119.6. For it to supply 60 boilers then, the throughput would have to be 27 items/s, but since that would break the x1/x2/x3 multiplier on belt speed, I'll cancel my idea of making a suggestion to buff it.

Aside the boiler/engine ratio bug then, all that's left is deciding which of the following I should do ~ 59/118 for 1:2 ratio and no resource efficiency decrease from only using 1 engine for the last boiler, 60/119 for the highest integer of engines a red belt can support, or 60/120 for gaining the final 240 KW, maximizing output per row, but either causing constant minor fluctuations in steam due to coal shortages or adding overhead to extend the ~20 minutes or so buffer time for the ~0.74 engines worth of coal the red belt can't supply. Can't say any of the 3 are quite clean solutions.

(Edit: 59/119 corrected to 60/119)

[mrvn]

8MJ per lump of coal
/2 for boiler
/1800 for 2 engines at 900 kW each
* 26.67 (or really 40 * 2/3)

equals 59.26

Thanks for the explanation. Then I'm convinced by evidence it's really coincidence - the 26.67 is 2/3 of 40, which is a round even number instead of some seemingly randomly picked decimal that wouldn't be likely at all. Matches with my estimate of 119.5-119.8 also, now that I realize I forgot to multiply the downtime by 2 since 2 engines were fluctuating with coal shortages, making the correct estimate 119.0-119.6. For it to supply 60 boilers then, the throughput would have to be 27 items/s, but since that would break the x1/x2/x3 multiplier on belt speed, I'll cancel my idea of making a suggestion to buff it.

Aside the boiler/engine ratio bug then, all that's left is deciding which of the following I should do ~ 59/118 for 1:2 ratio and no resource efficiency decrease from only using 1 engine for the last boiler, 59/119 for the highest integer of engines a red belt can support, or 60/120 for gaining the final 240 KW, maximizing output per row, but either causing constant minor fluctuations in steam due to coal shortages or adding overhead to extend the ~20 minutes or so buffer time for the ~0.74 engines worth of coal the red belt can't supply. Can't say any of the 3 are quite clean solutions.

In all but brownout situations there should be more production capacity than consumption. So I would go with 60/120. That gives a little extra room for spikes. And when not in a spike steam will collect, coal will back up so when the next spike hits it will take a really big and long alien attack to wear that last boiler down to having no steam and fuel again.

I know wood is half the fuel value compared to coal so 30/60 for wood fueled furnaces. Just out of interest: Have you done the math for solid fuel, rocket fuel, vehicle fuel (AAI), wood bricks, coke (Angels), ...?

I'm currently running my steam engines on vehicle fuel because converting wood from greenhouses to vehicle fuel is an energy gain. Problem for me is that the boilers run out of water before they run out of fuel. I have two offshore pumps and two rows of boilers with the fuel on a red belt in the middle. So I'm not worried about the belt running dry.

[Patric20878]

In all but brownout situations there should be more production capacity than consumption. So I would go with 60/120. That gives a little extra room for spikes. And when not in a spike steam will collect, coal will back up so when the next spike hits it will take a really big and long alien attack to wear that last boiler down to having no steam and fuel again.

I know wood is half the fuel value compared to coal so 30/60 for wood fueled furnaces. Just out of interest: Have you done the math for solid fuel, rocket fuel, vehicle fuel (AAI), wood bricks, coke (Angels), ...?

I'm currently running my steam engines on vehicle fuel because converting wood from greenhouses to vehicle fuel is an energy gain. Problem for me is that the boilers run out of water before they run out of fuel. I have two offshore pumps and two rows of boilers with the fuel on a red belt in the middle. So I'm not worried about the belt running dry.

In normal operation, a base wouldn't make use of the extra steam engine. In a brownout situation, I don't think 1 extra steam engine is going to make any difference, and storage tanks store a lot more steam anyways. Not to mention I could probably push it to 61/122 in that case. And adding overhead just for 1 engine sounds really inefficient. If I make it 60/120 I'd need a stronger reason. I'm probably most leaning towards the 60/119 option (59/119 was a typo, oops), followed by 59/118. Will need more discussion and playing devil's advocate to confirm.

No math done for stuff besides coal. And the bug with boilers not actually producing enough steam for 2 engines probably has something to do with your setup, if not insufficient water pressure.

[BlakeMW]

I'd definitely go with 60/120. Why? Because a 60/120 setup will always produce as much or more energy than a 60/119 - if it has fuel buffered it produces more, if the fuel buffer runs out it produces as much, as such unless there's something really useful you can do with that 120th steam engine slot it's better to just stuff a steam engine in it to get that little bit of spike accommodation and make the whole thing look elegant.

The only argument against 60/120 would be if you like to *strictly* separate baseline power generation and load balancing, i.e. using steam for baseline and accumulators for load balancing. If you want to be able to know exactly what your baseline capacity is by looking at the power screen then using a stingy number of boilers is better.

[Patric20878]

I'd definitely go with 60/120. Why? Because a 60/120 setup will always produce as much or more energy than a 60/119 - if it has fuel buffered it produces more, if the fuel buffer runs out it produces as much, as such unless there's something really useful you can do with that 120th steam engine slot it's better to just stuff a steam engine in it to get that little bit of spike accommodation and make the whole thing look elegant.

The only argument against 60/120 would be if you like to *strictly* separate baseline power generation and load balancing, i.e. using steam for baseline and accumulators for load balancing. If you want to be able to know exactly what your baseline capacity is by looking at the power screen then using a stingy number of boilers is better.

The main argument I'd have against it is the ugly fluctuations at 60/120. As much I really want 60/120, the fluctations are way uglier than 60/119. But your post did give me an idea on overhead - I could probably just have a belt loop around and feed back to main belt, or just a chest with some inserters. In the meantime though, I'd like the absolutely ugly fluctuations issue to be addressed before I can be satisfied with 60/120. 60/119 can just as easily be argued to be elegant since it never fluctuates, and so can 59/118, for being optimal on boiler usage.

[mrvn]

I'd definitely go with 60/120. Why? Because a 60/120 setup will always produce as much or more energy than a 60/119 - if it has fuel buffered it produces more, if the fuel buffer runs out it produces as much, as such unless there's something really useful you can do with that 120th steam engine slot it's better to just stuff a steam engine in it to get that little bit of spike accommodation and make the whole thing look elegant.

The only argument against 60/120 would be if you like to *strictly* separate baseline power generation and load balancing, i.e. using steam for baseline and accumulators for load balancing. If you want to be able to know exactly what your baseline capacity is by looking at the power screen then using a stingy number of boilers is better.

The main argument I'd have against it is the ugly fluctuations at 60/120. As much I really want 60/120, the fluctations are way uglier than 60/119. But your post did give me an idea on overhead - I could probably just have a belt loop around and feed back to main belt, or just a chest with some inserters. In the meantime though, I'd like the absolutely ugly fluctuations issue to be addressed before I can be satisfied with 60/120. 60/119 can just as easily be argued to be elegant since it never fluctuates, and so can 59/118, for being optimal on boiler usage.

But it only ever fluctuates when you use 100% of the power for a long long time.

Why would I use 60/120?

1) easier to build up to it by adding 4 boilers + 8 steam engines in a blueprint every time power runs low, no need to delete the 120th engine or 60th boiler.
2 ) 2 boilers and 4 engines each share a power pole. With 59 boilers that leaves 2 poles half used. Unsymetrical.
3) Makes all boilers run equal, not one running half speed.
4) If it fluctuates then you are using 100% of the power all the time and you have bigger problems than a fluctuating boiler.

PS: 60/120 fluctuates because the belt can't bring enough coal to run 60 boilers full time. With 60/59 the last boiler only runs half the time and the belt is fast enough. Nothing to do with water pressure (although that has to be watched too).

[Patric20878]

But it only ever fluctuates when you use 100% of the power for a long long time.

Why would I use 60/120?

1) easier to build up to it by adding 4 boilers + 8 steam engines in a blueprint every time power runs low, no need to delete the 120th engine or 60th boiler.
2 ) 2 boilers and 4 engines each share a power pole. With 59 boilers that leaves 2 poles half used. Unsymetrical.
3) Makes all boilers run equal, not one running half speed.
4) If it fluctuates then you are using 100% of the power all the time and you have bigger problems than a fluctuating boiler.

PS: 60/120 fluctuates because the belt can't bring enough coal to run 60 boilers full time. With 60/59 the last boiler only runs half the time and the belt is fast enough. Nothing to do with water pressure (although that has to be watched too).

Only 20 or so minutes, not that long.
1) Maybe good reason, but by that logic I rather not even want to count how many steam engines I pasted already every time I paste one. More effort than deleting an engine.
2) Maybe good reason, but a flat line at 108 MW for the first 20 minutes, followed by haywire fluctuations is more unsymmetrical.
3) If anything, 60/120 makes 2 boilers run unequal - the last 2 boilers have something around 61.33% uptime or so. Not good reason.
4) It's the fact that 60/120 doesn't actually sustain 60/120 that bothers me - though what you say makes me consider that maybe the setup fluctuating after 20 minutes of 100% consumption could be seen as a good thing, to indicate that the setup is overloaded.

The discussion is helping though - right now, maybe 40% leaning towards 60/120, 40% towards 59/119, and 20% towards 59/118. Help me make it 80%+ for 60/120! Like how do I even explain how much a 60/120 produces, it'd be wordy af going like, it has 120 engines, and produces 108 MW, except when your base consumes full power, then it sustains for 20 minutes until it fluctuates like crazy, so actually it doesn't quite produce 108 MW...yeah, just pretend that last steam engine doesn't exist if you ever run any risk of running at full power for more than 20 minutes at a time.

And dunno about your setup, but running out of water typically means insufficient water pressure or water output. I use precise offshore pump positioning and underground pipes to distribute water pressure correctly in my setup.

[mrvn]

But it only ever fluctuates when you use 100% of the power for a long long time.
Only 20 or so minutes, not that long.

That's 2 full days. Add a solar cell and that will probably make up for the shortfall of the belt.

But seriously. My power consumption fluctuates heavily. And not just when aliens attack. Inserters draw power when they work, same for assembelrs. A train arrive and power spikes because loading/unloading activates the inserters. So 20 minutes of using 100% power means a lot of the time something is falling short. That's no way to live.

[BlakeMW]

Another argument for 60/120 is that there are also other things that can result in the less than expected output, for example some of the close-to-the-pipe-limits steam setups I've built will run at 100% at one orientation, but not quite 100% at another orientation, or it might depend on the part placement order whether it can sustain 100% or not. But if it manages 100% most the time and at least 97% all the time then that's good enough for me. Because the thing is that much as mrvn reasons, a powerplant running at full load isn't optimal: for short periods it can be explained by recharging accumulators, but if the powerplant is at full load for 20 minutes or longer that probably means the accumulators are going flat sometimes and machines are actually being starved of power. Consider the cost of an Assembler 3 with 4x Prod 3 modules, to be optimal/efficient you don't want lost ticks of production and it quickly becomes cheaper to build a surplus of generation capacity than to lose assembler time. Hence it should be assumed that a powerplant should never need to run at higher than 90% load for prolonged periods and that it will have full steam reserves for burst power.

[mrvn]

And I rather waste 1% of the steam engine power generation capability by building that 120th steam engine than waste 49% of the boilers steam production capability by not building it.

[Patric20878]

This is supposed to be low-tech compatible, like literally before medium pole tech - solar cells and drones don't count Someone at drone tech not needing to remove the 120th engine gets offset by someone who needs to place the 120th engine manually, all you'd be doing is making it trivially longer to do it early-game so you can do it trivially shorter later-game - there is no significant net gain whatsoever here.

And if we're talking about burst, 120 is the near-start (119.266... being start) of the slippery slope of burst power vs burst time. Like really, I could make it burst at 122 for 7 minutes too, why stop at 120? I could probably make it burst at 124 for a couple minutes too. Or burst at 500 engines, for really short but intense burst, lasting until all them run out of steam. Where do you draw the line? 120 isn't even the start of it to make it a significant number beyond number aesthetics. For me, there's a clear line between indefinitely sustainable at full load, and temporarily sustainable, and 59/119 is it. And I don't see why this setup should force the 120th steam engine on it just for burst power when I could just start a new row, adding the 120th engine to the new row to get sustainable power. If shore length isn't limited, this is no issue, and if it is limited, the 120th engine doesn't increase sustainable power, and one should add a lot more than 120 engines to provide for higher bursts of power, with basically no limit. For the last boiler only having 1 steam engine, this can be solved by having alternating rows of 59/118 and 60/120, using some splitters to prioritize 59/118 lanes (an idea possibly worth considering? ) Previous designs probably could've had burst power too, but I think I value sustainability over the gains of 1 steam engine, which is nothing, and by the same logic, if full consumption was required for an extended amount of time, new rows ought to be built in advance. I don't like it being 60/119 instead of 60/120, but the difference between sustainable and unsustainable is a hard binary difference for me that overrules it not being a clean multiple of 30.

Here's a random ASCII graph on how long x number of engines can sustain max output for visual purposes:

Code: Select all

                                                         <=119                                                         >=120
----------------------------------------------------------------------------------------------------------------------x
                                                                                                                      |
                                                                                                                      |
                                                                                                                      |
                                                                                                                      |
                                                                                                                      |
                                                                                                                      \
                                                                                                                       y
                                                                                                                        \
                                                                                                                         \
                                                                                                                          \

Where the flat line is infinitely sustainable at max load (excluding the bug with boilers/engines not being exactly 1:2 ratio atm), the vertical drop indicates shift from infinitely to finitely sustainable (starting at 20 minutes of time, decreasing), and the diagonal slope is decreasing amounts of burst time as you add more engines. x and y signify the 119th and 120th engines, respectively. Obviously not to-scale but I hope it illustrates the point I'm making. And as you can see in this graph, 120 is a way uglier place to be at than 119 is.

[BlakeMW]

You should probably come up with a convincing case where someone would run a power plant at 100% load for 20 minutes (meaning their grid is over-capacity and machines are slowing down due to low power) and then be upset by the loss of a few percent power output.

(The only real case which I can think of would be for emergency steam - that might realistically run at full load for 20 minutes if a real clusterfuck hits the nuclear fuel supply or something - but even here it'd be doubtful that the slight drop in power & power quality after 20 minutes would be upsetting - not to mention that emergency steam might be limited by a fuel buffer anyway, or use solid fuel instead of coal)

And if we're talking about burst, 120 is the near-start (119.266... being start) of the slippery slope of burst power vs burst time. Like really, I could make it burst at 122 for 7 minutes too, why stop at 120? I could probably make it burst at 124 for a couple minutes too.

That's simple enough. Because the 1:2 ratio and multiples of 10 are simple, elegant and convenient, like when crafting stuff you queue it up in 5's. If two designs are very nearly as good as each other and one uses convenient numbers and the other doesn't, I'd go with the convenient numbers.

[mrvn]

This is supposed to be low-tech compatible, like literally before medium pole tech - solar cells and drones don't count Someone at drone tech not needing to remove the 120th engine gets offset by someone who needs to place the 120th engine manually, all you'd be doing is making it trivially longer to do it early-game so you can do it trivially shorter later-game - there is no significant net gain whatsoever here.

At the point where you need 120 steam engines you probably have the science to have solar pannels and much more. But that is beside the point.

Even at the start I regularly blueprint things and then run along and place stuff. The ghosts are just a guide where to place things. Well, and now I use the bluebuild rewritten mod, so as I stand next to ghosts they get placed.

And if we're talking about burst, 120 is the near-start (119.266... being start) of the slippery slope of burst power vs burst time. Like really, I could make it burst at 122 for 7 minutes too, why stop at 120? I could probably make it burst at 124 for a couple minutes too. Or burst at 500 engines, for really short but intense burst, lasting until all them run out of steam. Where do you draw the line? 120 isn't even the start of it to make it a significant number beyond number aesthetics. For me, there's a clear line between indefinitely sustainable at full load, and temporarily sustainable, and 59/119 is it. And I don't see why this setup should force the 120th steam engine on it just for burst power when I could just start a new row, adding the 120th engine to the new row to get sustainable power.

There is a big difference between 120 and 121 engines. With 120 engines you can burst power till the steam is gone, the coal in the boiler is goine and the coal on the belt is gone. A belt going past 30 boilers holds a lot of coal. That's why it taks 20 minutes to run out. With 121 engines you run out of steam for the last engine pretty much emidiately, shorter than the usual alien attack, which makes it useless.

To sustain bursts past the 120 engines you need steam tanks. And then you would go with a 60x boiler-steam-steam-steam-tank-tank-steam or similar. It's a good setup to work in parallel with solar power. During the day you store steam in the tanks while the steam engines are (hopefully) not using any and during the night all 4 engines per boiler can run.

Not sure the 2 tanks are the exact size to run the steam engines in a 60/40 split but it works well for me.

If shore length isn't limited, this is no issue, and if it is limited, the 120th engine doesn't increase sustainable power, and one should add a lot more than 120 engines to provide for higher bursts of power, with basically no limit. For the last boiler only having 1 steam engine, this can be solved by having alternating rows of 59/118 and 60/120, using some splitters to prioritize 59/118 lanes (an idea possibly worth considering? ) Previous designs probably could've had burst power too, but I think I value sustainability over the gains of 1 steam engine, which is nothing, and by the same logic, if full consumption was required for an extended amount of time, new rows ought to be built in advance. I don't like it being 60/119 instead of 60/120, but the difference between sustainable and unsustainable is a hard binary difference for me that overrules it not being a clean multiple of 30.

Here's a random ASCII graph on how long x number of engines can sustain max output for visual purposes:

Code: Select all

                                                         <=119                                                         >=120
----------------------------------------------------------------------------------------------------------------------x
                                                                                                                      |
                                                                                                                      |
                                                                                                                      |
                                                                                                                      |
                                                                                                                      |
                                                                                                                      \
                                                                                                                       y
                                                                                                                        \
                                                                                                                         \
                                                                                                                          \

Where the flat line is infinitely sustainable at max load (excluding the bug with boilers/engines not being exactly 1:2 ratio atm), the vertical drop indicates shift from infinitely to finitely sustainable (starting at 20 minutes of time, decreasing), and the diagonal slope is decreasing amounts of burst time as you add more engines. x and y signify the 119th and 120th engines, respectively. Obviously not to-scale but I hope it illustrates the point I'm making. And as you can see in this graph, 120 is a way uglier place to be at than 119 is.

That graph is totaly backward and completly wrong: It's more like

Code: Select all

engines
   ^
   ||
   ||
   ||
121|\
121| \________________ 20 minutes
120|                |_____
   |
   |                                  infinite
   +------------------------------------>
     time

[Patric20878]

@Blake: No need, either way the one extra is trivial. The burst is no more important now than it was before, and no one said anything about burst power before. Number aesthetics don't overrule what I drew in the graph.

@mrvn: Incorrect. 121 engines lasts about 11-12 minutes of coal and can sustain steam just fine. And so can 122, at about 7 minutes. It should be clear from my graph that the difference between 120 and 121 is small compared to the difference between 119 and 120. Burst power holds no ground whatsoever to me unless someone can clearly define just how much burst time is enough, since right now, I don't see 10 minutes of burst time to be too short either by that logic. If anything, all previous versions had no burst time. And my graph is fine, all you even did was flip the axis.

[BlakeMW]

Burst power holds no ground whatsoever to me unless someone can clearly define just how much burst time is enough, since right now, I don't see 10 minutes of burst time to be too short either by that logic.

Well first of all let's clearly define "burst" as when the steam engines are drawing more power than the boilers can sustain.

I consider in general around 30s of burst power to be enough, this is related to how long it tends to take an alien attack to die without causing undue damage to defenses, so the time it takes for the bugs to run up and die "instantly" to a barrage of lasers. The 30s is actually in case two attacks happen in close succession and is derived mostly from experience: like my experience is powerplants which can burst for 30s (such as those using a tankless 1:3 ratio which can burst at full power for 27s) usually don't run out of steam under heavy (i.e. Deathworld) biter attacks. There are also other things that can cause a short duration burst draw such as fulfilling a large logistics request (bots recharging afterwards).

For a solar/steam powered factory the relevant time is the length of the night: most pessimistically this can be considered to be 208s, the duration when solar power produces less than 100% output altough the steam engines should only be at full power during 100% darkness which is a mere 42s.

These seem to be two reasonable durations which a steam power plant might be built to burst for, and both are much lower than 1200s. I can't imagine nominal conditions in Factorio where a steam power plant would be running at 100% load for that long - the most imaginable being a mod with a greatly increased day/night duration.

[Patric20878]

Yes, so 120 isn't a significant number far as burst power goes. For just 30s burst time, you could probably use 128 engines. As the previous designs have all set the precedent for the standard setup to be fully sustainable, if one wants burst power, they can add their own variant, like how I used to have storage tanks for the extended variants.

And I believe you're slightly missing my point on the whole concept about engines being sustainable or not. Let's consider 2 cases, where power consumption is:
less than or equal to 119 engines: 119 and 120 are practically equal.
120 engines: You get 1 engine to add some burst power. 0.9 MW extra every 107.1 MW is so beyond trivial it'd make no difference at all. Either 119 engines are enough or you'll need *way* more than just 1 engine to provide burst power. Whether or not the base consumes at max load is totally irrelevant. Hence 120 is a totally pointless number to set it to - either set it to a significant number like 119, which is max sustainable engines, or actually give it some useful amount of burst. Which then just ties into my first point - depending on how much burst you want, add more or less steam engines to the setup, which really only applies if you're literally out of shore length to just add more to a new row. It doesn't need to be part of the standard variant of the design. What I previously said about fluctuation after 20 minutes possibly being a good thing for being an overload indicator can possibly be a bad thing too for masking other fluctuations that should be looked into, like if coal isn't actually max compressed on the belt, resulting in coal shortages, so that line of thought can be discarded.

At the moment, I'm fairly convinced that alternating 118/120 rows is the best way to go. No wasted boilers, good for pasting 1:2 blueprints for, is indefinitely sustainable, and should look decent too. And since red belt throughput is 26.67 items/s, it also makes more sense that it's matched with 119 engines, so if it ever gets buffed to 27 items/s, it can be matched with 120 engines.

[mrvn]

How do you get 121 engines to last 11-12 minutes? How do you even connect that to 60 boilers? If you connect 3 steam engines to one boiler then that third engine will not last 11-12 minutes for sure, right? Are you adding pipes across all boilers so that the 121th engine draws steam from all of them? But then that is a whole different setup and the cost of pipes outweighs the benefit of the 121th engine I thing.

And I believe you are missing the point about power draw not being a straight line. It's not like your base needs exactly 107.1MW of power. It's more like you need 98-108MW of power. With 119 engines you have moments where you are lacking 0.9MW of power, with 120 engines the spikes can still be satisfied while the coal belt can recover in the moments where you need less than 107.1MW.

0.9MW might seem irrelevant but when you are missing them they suddenly become very big.

Lets do the math:

A belt can supply 26.67 coal/s. If 60 boilers and 120 steam engines need 27 coal/s then each steam engine needs 27/120 = 0.225 coal/s. So 119 steam engines need 119*0.225 = 26.775 coal/s. Ups And 118 engines need 118*0.225 = 26.55 coal/s.

I was trying to show that with 119 engines you waste nearly a full engines worth of coal. But the math says 119 engines isn't even sustainable at 100%. So the argument is more about 59/118 vs. 60/120. With 59 boilers wasting very little belt capability and 60 boilers being purely to smooth spikes and make the rows even.

Well, I still think 60/120 is a good setup. Since laser turrets and alien waves have huge spikes I even go so far as to suggest 64/128 for that extra buffer because rows of 16 boilers with one water pump each work well and you can blueprint them in e.g. chunks of 4. Or go 1,2,4,8,16,2 rows, 4 rows at the beginning. After that you need a new full belt of coal and probably have roboports and assembler to produce boiler and steam engines and just blueprint whole setups as needed.

[Patric20878]

By doing 61/121, and I make it last 11-12 minutes cuz MAGIC. :> But no, same reason I can make 120 steam engines work when the previously thought to be best was 60.

And no, I don't need exactly 108 MW of power. Almost never will there be a situation where my base needs 108MW instead of 0-107MW, and also won't have any more shore length to add more engines to place more at the same time. And I addressed all this already in the post right before this one. Read the post I made with the graph and understand what the actual axis are - neither of them are power draw. The straight line in my graph is how long the setup *given max consumption* can sustain for, not how much it's consuming.

By the math leitk posted, a red belt sustains 59.266.. boilers at max load, or 118.533 engines..hmm. I mistakenly thought it'd be 119. Then I'll test 119 long term, and if it runs out, I'll take out 59/119 as an option. Good, since it'd simplify things to not have 119 as one.

And as for your suggestion about about 64/128, already covered it in the previous post, please review. I also covered why 60/120 isn't that useful a setup in the same post when going the burst power route, since 2 engines extra provide next to nothing compared to 118, and number aesthetics has no functional impact, no matter how compelling it is to make it look neat. And neither belong in the standard setup. That can go in a variant and the burst power can be adjusted as arbitrarily needed, but I've always made the standard setup to be indefinitely sustainable, and any variants involving temporary power, like with the old storage tank water batteries, were variants of the standard design. I can do further testing to see if there's any significant number past 118 for a good candidate as a variant, but otherwise, I think I'm pretty much decided on 59/118 as the standard design, if all the math and testing checks out. The discussion was most insightful, thanks~

I think the best thing we can do now if we want 60/120 (ideally, I still do) is to make a suggestion to the devs to buff belt throughput from 13.33/26.67/40 to 13.5/27/40.5 or something. That would cleanly allow for 30/60 engine setups on yellow/red belts. Or if they want to keep numbers more even, 14/28/42 or 15/30/45 instead. Doubt it'd happen anytime soon but I'd support the change.

[mrvn]

Doubtful you will get them to change belt speeds as that would break ratios in everything else. As far as they are exact there anyway. I bet a lot of things are currently at 99% or 101% belt usage instead of having a perfect ratio. So would you fix the boilers or the furnaces for example.

[Patric20878]

A 1.24% decrease in boiler coal consumption would work too, calculated from 27 / 26.67.

And also I just tested 119 again - it sustains fine. 118.5 is from wrong math.
Recall the formula is 8000 / 2 /1800 * 26.67. The 1800 comes from 2 steam engines x900 KW each boiler. In a 60/119 setup, not every boiler has 2 engines! Thus, to get the actual average, you take 119*900/60, and get 1785 KW.

So if you apply the formula again, 8000 / 2 / 1785 * 26.67 = ~59.7647, or ~119.5294 engines, averaging 1785 KW each boiler. Seems like my original estimate of 119.5-119.8 was actually right then, and later second thoughts had units of measure confused.