Anyone use barrels?

[Kahnugo]

I'm going to be a little biased in the way that I don't necessarily play optimally, all options given, rather I tend to play concept builds where I decide that I'm going to do a specific thing regardless of how solid of a strategy it actually is.

That said I think generally barrels do have a place, but if the situation allows for a fluid build that probably will be the best way to do it in most cases. Both methods of handling liquid comes with it's own infrastructure challenges (pipes or barrel handling).
Barrels are more flexible and small amounts of liquid might be a lot nicer to deal with if using barrels (and bots). Ironically the same can be the case for very high amounts of liquid in my experience (to the point where it become a pain in the butt trying to fit in pipes everywhere.

Specifically transporting by train, fluid wagons allow for faster shipment, they can unload slightly faster and hold slightly more liquid than barrel trains can. You might have other priorities though.

[Deadlock989]

Barrels were invented before fluid wagons existed in vanilla, so in that sense they are a relic. There was an overlap period where they were a better choice for moving fluid in most situations, and then they were nerfed to where they are now.

The only time I ever use them is in games where I'm allowing logistics bots in crafting "malls" and I want sulphuric acid or lubricant on demand without having to run pipes all over the shop.

[SKKiro]

They are obviously the way to go if you plan to use bots to transport fluids, since you have no choice at all in that regard.

Important to also think that barrels are very precise; as each one is 50 units of fluid. So if you are into being really really precise on how you use certain fluids then they are your best bet too.

[blazespinnaker]

I've used heavy oil barrels to prime coal liquefaction. Kind of a killer app for them, tbh.

They can be useful for creeping on biters with flamers, though probably by that time you'll use torpedos. Flamers are fun tho

[ptx0]

I'm probably going to create a oil factory to process and create all the oils needed and ship them off by train to the base, but why even bother with barrels if you can simply transport the oil via trains and load into tanks at the destination, no barrels required?

Has anyone found cases where barrels are a better choice than fluid train cars?

yes when pushing really high SPM with modded belts and high item rate, you might be better off with barreling for plastic production since pipe logistics are a pain, comparatively.

[NotRexButCaesar]

torpedos

I guess if nauvis is a white dwarf, the air must be a liquid.

[SoShootMe]

Has anyone found cases where barrels are a better choice than fluid train cars?

Barrels have some uses in general (as others have written), but specifically for transport of fluid by train, the biggest advantage I know of is that it can occasionally be useful to have a single wagon holding fluids, including different types of fluid, and items at the same time.

For moving large volumes of fluid by train, I think the advantages fluid wagons have of reduced load/unload time and increased wagon capacity outweigh other considerations. Even where you want to move large volumes of fluid around a production area with logistic robots carrying barrels rather than pipes, I think using fluid wagons and filling/emptying barrels locally can make sense, even if you are doing it at both ends of the train transport.

[jodokus31]

Barrels are awesome in modded games with many fluids. Either have an easy distribution via bots or load a single train with different fluids and items, where throughput is not the most important factor.
I cannot imagine seablock without barrels.

[ptx0]

oh right you can unload with more inserters than you can pumps, though i'm not sure that the throughput is better with vanilla stack inserters compared to vanilla fluid pump, probably not.

but you can unload on a curve

and alignment is not as important as pumps.

[blazespinnaker]

torpedos

I guess if nauvis is a white dwarf, the air must be a liquid.

i get lazy sometimes, especially when the ideas are boring.

[blazespinnaker]

yes when pushing really high SPM with modded belts and high item rate, you might be better off with barreling for plastic production since pipe logistics are a pain, comparatively.

Anyone have a good handle on the UPS efficiency here? Very interested.

[Yoyobuae]

oh right you can unload with more inserters than you can pumps, though i'm not sure that the throughput is better with vanilla stack inserters compared to vanilla fluid pump, probably not.

12x stack inserters with at least 10x stack size = Unload a cargo wagon full of barrels in 1.3 seconds = 20000 fluid per 1.3 seconds (once unbarrelled) = 15385 fluid/s
3x pumps = Unload a full fluid wagon in 1.8 seconds = 25000 fluid per 1.8 seconds = 13889 fluid/s

So it's a tiny bit faster using barrels but I think that benefit will be offset by having to process 25% more trains to get same amount of fluid.

EDIT: Nevermind, I've just found a more optimized fluid wagon unloading setup which empties it in 0.7 seconds.

[Yoyobuae]

yes when pushing really high SPM with modded belts and high item rate, you might be better off with barreling for plastic production since pipe logistics are a pain, comparatively.

Anyone have a good handle on the UPS efficiency here? Very interested.

Pipes should be more UPS efficient then having several assemblers+inserters+belts or bots for barreling and unbarreling. The inserters in particular are bad for UPS.

Pipes get all this unwarranted bad reputation for being UPS inefficient when in fact they were the first to be multi-threaded, way ahead of belts which just got the multi-thread treatment in version 1.1.

[blazespinnaker]

oh right you can unload with more inserters than you can pumps, though i'm not sure that the throughput is better with vanilla stack inserters compared to vanilla fluid pump, probably not.

but you can unload on a curve

and alignment is not as important as pumps.

Interesting. 13.85*50=692ps/stackinserter*12=8310 fluid unitsp/s[1].

capcity of fluid wagon is 25K[2]. unloads in 108 ticks(1.8 s)[2] cargo wagon filled with barrels can carry 20,000 units of fluid[2].

20K/8310=2.4s unload time with stack inserters.

Fluid/pumps are much faster than barrels.

1. https://wiki.factorio.com/Stack_inserter
2. https://wiki.factorio.com/Fluid_wagon

[blazespinnaker]

Pipes get all this unwarranted bad reputation for being UPS inefficient when in fact they were the first to be multi-threaded, way ahead of belts which just got the multi-thread treatment in version 1.1.

Yep, sounds like barrels suck pretty bad. 1 storage tank require 5 steel + 20 iron, the equivalent in barrels requires 25K/50=500 steel!

Other than priming coal liquefication, I don't see what their point is.

Maybe on some speedruns they could be used to move sulfuric acid without having to spend time on pipe infra. Maybe. Its a stretch.

If biters attacked pipes more, barrels might be handy.

I think it'd be a good deathworld difficulty option for factorio - biters attacking belts/pipes/power/rail.

[Yoyobuae]

Maybe on some speedruns they could be used to move sulfuric acid without having to spend time on pipe infra. Maybe. Its a stretch.

I was just looking into that. An Any% speedrun requires something like 2200 processing units to be made, which translates into 11000 sulfuric acid, which can be carried with 220 barrels (22 stacks). So less than one iron chest of barrels could supply the sulfuric acid needs for processing units for the entire run.

One problem with that is that steel is kinda a precious resource during speedruns, and using up 220 (212 if prod 1s are used) steel to make those barrels might not be worth.

[Amarula]

The only compelling case I have seen for barrels is storage density (now why you would ever want to store liquids is a totally different question but if you did):
One tank = 3x3 = 9 spaces; volume = 25K.
9 spaces = 9 steel chests x 48 stacks x 10 barrels per stack x 50 units per barrel = 216K.
Obviously barrels also have overhead to fill and empty the barrels, and the robots to deliver the barrels to the chests, but in a large storage facility those costs are very small compared to the volume savings overall.
And yes, I have done that, storing over 200M crude oil, don't ask me why

[Kahnugo]

oh right you can unload with more inserters than you can pumps, though i'm not sure that the throughput is better with vanilla stack inserters compared to vanilla fluid pump, probably not.

12x stack inserters with at least 10x stack size = Unload a cargo wagon full of barrels in 1.3 seconds = 20000 fluid per 1.3 seconds (once unbarrelled) = 15385 fluid/s
3x pumps = Unload a full fluid wagon in 1.8 seconds = 25000 fluid per 1.8 seconds = 13889 fluid/s

So it's a tiny bit faster using barrels but I think that benefit will be offset by having to process 25% more trains to get same amount of fluid.

EDIT: Nevermind, I've just found a more optimized fluid wagon unloading setup which empties it in 0.7 seconds.

I am a bit curious about the numbers, 1.8 seconds seems right for fluid wagons (tested ingame with a wagon and 3 pumps, emptied in 106 ticks) about the edit, how do you bypass the attachment delay (which I forgot about in my earlier post)?

About the stack inserter value, what is the math behind that? I was assuming that 27.6 per second for stack size 12 translated linearly into 23 per second for stack size 10. edit: Which gives 1.45 seconds to empty a wagon.

[blazespinnaker]

"Which gives 1.45 seconds to empty a wagon."

Hmm, did you test that in game? I'm not seeing that.

[blazespinnaker]

Maybe on some speedruns they could be used to move sulfuric acid without having to spend time on pipe infra. Maybe. Its a stretch.

I was just looking into that. An Any% speedrun requires something like 2200 processing units to be made, which translates into 11000 sulfuric acid, which can be carried with 220 barrels (22 stacks). So less than one iron chest of barrels could supply the sulfuric acid needs for processing units for the entire run.

One problem with that is that steel is kinda a precious resource during speedruns, and using up 220 (212 if prod 1s are used) steel to make those barrels might not be worth.

I think the search space for optimization is very complicated. A lot of people try to reason it out, but I think it has to be done algorithmically because it's so vastly large. Some stuff at scale can be obviously written off, but when you look at bootstrapping, it starts to get tricky. For example, maybe using barrels for lubricant is good as it can help squeeze out a few more bots, which in turn can have a compound effect on times. (x bots could put down x miners/pipes/etc, which leads to more production, more bots, etc.)

Not saying this is true about bots/lubricant, just an example of the possible complications in reasoning it out.