Friday Facts #430 - Drowning in Fluids

[Panzerknacker]

Thank you so much for a update regarding this, for me, highly important subject. Nice and fitting FFF title also hahaha, I can already imagine.

It is good to read that the fears I had for the new system actually came out to be true and were also discovered by the dev team. I am reading a lot of good things in this post that were thought about and corrected, also I like the graphics while building. Really something when people are able to admit flaws in their design.

Letting this go through my mind tho, it still feels like you are trying to fix something that was never really broken, and the 'fix' brings a lot of complications with it apparently that need to be addressed individually while they worked naturally by itself under the old system. Many new problems and quirks will surely be found after release also.

For me the old system was really simple: if machines were starved for fluid, you simply need more production or pipeline capacity. There was no problem at all with the build order that some were talking about; if a machine was not getting fluid you were simply not producing enough. Nothing unintuitive or wrong with it at all, wether it internally depended on the build order or not, so what?? And yes, you had to TEST setups without being able to pre-calculate everything, but hey, that's engineering.

I think you learned a lot overthinking all of this, invented some nice QoL features and graphics, and are doing a great effort, but I really am more convinced than ever that you need to let this go and simply rollback to the 1.1 simulated fluid system, WITH the new QoL stuff and maybe the steam ratio and some of the other nice stuff that is clearly a improvement.

As an engineer, I bet you know that simple solutions always win from complex ones. The tough one tho is coming up with such simple elegant things. That's exactly what this is turning into. The old fluid system was a, programmatically, beautifully simple and clever system. This new system is turning into something rather complicated now, and it will probably get even worse from here on.

Anyway, you guys have this in your hands now. You decide. For Factorio: Spage Age being the best it can possibly be.

[mmmPI]

Let's get you all a more practical example of why limiting even pumps to 1200/s breaks reactor designs to the core!

Not it doesn't , you can design nuclear reactor without any pumps, since you can read the steam level and the temperature of the reactor, it's easy to only refuel when you have enough storage capacity for the steam to be stored.

The fluid flow at infinte throughput and speed inside a 250x250 area, as if it was only a single tank holding all the fluid.

Unless you use pumps, then it will break the network into 2 different network that can share fluid at the rate of "1200" per pumps.

But if you look at the previous FFF about fluid, you could see builds that look like a long pipe 5 pumps in parralels feeding into another long pipe. This provide 1200*5 6000 fluid/s of potential transfer between the 2 networks, and if you want more you can just add more pumps.

20 heat exhanger need 2 pumps in parallel if you want to use pumps, so that 2400 fluid , or more than 1977.6 per second could be transfered to the network behind the pump.

But since fluid do not require pumps to flow distances and high throughput, i don't see many reasons to still want to use pumps between the heat exchanger and the storage tanks array

[Taipion]

I mean... even ASSUMING everything works fine, there is no limit and no bottleneck, no weird quirks in the mechanic and all that...

So, both to utilize steam storage tanks, and to bridge segments (when building really big reactors),
you are actually supposed to to pipe-to-pump-to-pipe-lines, an I mean LINES OF PUMPS to match the requirements of your reactor?!?
Who thought that would be a good idea?
So on my 10gw reactor I have to build a line of 87 pumps to bridge segments, and to load it into storage tanks with a full 10.080mw?!?
...and then I need another line of 87 pumps to load it back from the storage tanks into the turbines?!

Again, who thought that limiting pumps to 1200/s would be a good idea?!?

I can clearly see the flawed flow of thought, and the limited consideration, clouded by not knowing what happens outside ones own little box...
"yea there won't be any more limit on pipe throughput, so we can just nerf pumps 10x like we nerfed personal lasers into oblivion (no one will use them anymore), for a little more realism in a game that is in no way meant to be realistic, I mean what can go wrong, right? ...right?!?!?

[Loewchen]

What does the capacity of tanks have to do with needing to use pumps? You have infinite throughput without pumps. And especially for a reactor setup you can keep the steam network size easily below the 250 limit.

#3: to make full use of the storage tanks capacity, and ignore current fill lvl in the pipes and ignore fluid mechanics, you do need to use pumps, actually 2 per storage tank

You don't need to pump steam to build a regulated reactor setup without any energy losses with 1.1 fluid simulation and you obviously do not pumping in 2.0.

[Theikkru]

[...]
#3: to make full use of the storage tanks capacity, and ignore current fill lvl in the pipes and ignore fluid mechanics, you do need to use pumps, actually 2 per storage tank

Well there's your problem. Ignoring fill levels looks like it's entirely unnecessary due to this bit:

[...]there is a hardcoded limit of 100 fluid per flow operation (6000/s). This limit is multiplied with the fullness ratios of the source and sink to produce the actual flow value[...]

So, even without any pumps, to flow limit the exchangers you'd need greater than 98% segment fill, and to flow-limit the turbines you'd need less than 1% fill. 97% effective tank capacity is plenty to use as-is.

[mergele]

That does osund generally good, but I got 2 issues:
1) 250x250 and one tile more makes it instantly zero sounds extremely binary and artifical. I would suggest making it a scaling throughput reduction to all fluid operations on that pipe instead.
2) This is just for clarification. Does that 6k/tick limit mean no matter how many pumps I add on, a system can never deliver more and there will need to be a second parallel one set up?

[Taipion]

Let's get you all a more practical example of why limiting even pumps to 1200/s breaks reactor designs to the core!

Not it doesn't , you can design nuclear reactor without any pumps, since you can read the steam level and the temperature of the reactor, it's easy to only refuel when you have enough storage capacity for the steam to be stored.

Say no more, you already explained in detail that you have no clue of how efficient nuclear reactors work in factorio.
Being able to read out heat lvls is "nice to have", but does not help a lot with building one, as you need to (damn... how many times I explained that already?!?!) ...as you need to be able to buffer 1 fuel cycle, the temperature at which you start the next will not always be the same as the load differs,
and steam is a much more informative source to tell when to do that, as you, PREVIOUSLY, NOT ANYMORE AFTER THE EXPANSION, were able to, with steam, to tell exactly when the heat exchangers output falls below the demand of your base, and you could design a very small and efficient steam storage to just bridge that gap.
Now, you can totally rely on reading out the heat, and build bulky reactors with way too many heat-pipes to buffer the heat, go ahead, do that...
Most people will have to do that, as pumps are crippled now and steam management simply won't work anymore. gg wube, gg

[Taipion]

What does the capacity of tanks have to do with needing to use pumps? You have infinite throughput without pumps. And especially for a reactor setup you can keep the steam network size easily below the 250 limit.

#3: to make full use of the storage tanks capacity, and ignore current fill lvl in the pipes and ignore fluid mechanics, you do need to use pumps, actually 2 per storage tank

You don't need to pump steam to build a regulated reactor setup without any energy losses with 1.1 fluid simulation and you obviously do not pumping in 2.0.

Yes, you explained already that you have no clue of reactors, now please read into it and come back when you do, thank you.

[GregoriusT]

I mean... even ASSUMING everything works fine, there is no limit and no bottleneck, no weird quirks in the mechanic and all that...

So, both to utilize steam storage tanks, and to bridge segments (when building really big reactors),
you are actually supposed to to pipe-to-pump-to-pipe-lines, an I mean LINES OF PUMPS to match the requirements of your reactor?!?
Who thought that would be a good idea?
So on my 10gw reactor I have to build a line of 87 pumps to bridge segments, and to load it into storage tanks with a full 10.080mw?!?
...and then I need another line of 87 pumps to load it back from the storage tanks into the turbines?!

Again, who thought that limiting pumps to 1200/s would be a good idea?!?

I can clearly see the flawed flow of thought, and the limited consideration, clouded by not knowing what happens outside ones own little box...
"yea there won't be any more limit on pipe throughput, so we can just nerf pumps 10x like we nerfed personal lasers into oblivion (no one will use them anymore), for a little more realism in a game that is in no way meant to be realistic, I mean what can go wrong, right? ...right?!?!?

You do not need inline pumps at all ever anymore (unless your setup is bigger than 250 tiles), that is how it works. You do not need to pump into a Tank, the Tank will just work without the Pumps. Also what you really fail to consider, quality Pumps have higher throughput so the 2000 per second that your imaginary setup (wont) need are achievable easily.

[linktothepast83]

I don't know how to feel just reading all these big changes until i actually get a feel of them ingame. But i will say this i feel a bit sad when i think that way too many of the designs i made over the years catering to the old fluid system issues are meaningless now by upgrading to 2.0.

If it weren't for computational issues i would rather see a fluid tranportation system that works like belts do and not an instantaneous one. The only instantaneous action would be input and output from the pipes. Everything would move from the direction of initial input, every side loading would follow the initial direction like items on belts do and pipes fill up gradually and get congested like belts do as well. Similarly in every junction throughput would split up to a third if it is a cross shaped one.

[Taipion]

[...]
#3: to make full use of the storage tanks capacity, and ignore current fill lvl in the pipes and ignore fluid mechanics, you do need to use pumps, actually 2 per storage tank

Well there's your problem. Ignoring fill levels looks like it's entirely unnecessary due to this bit:

[...]there is a hardcoded limit of 100 fluid per flow operation (6000/s). This limit is multiplied with the fullness ratios of the source and sink to produce the actual flow value[...]

So, even without any pumps, to flow limit the exchangers you'd need greater than 98% segment fill, and to flow-limit the turbines you'd need less than 1% fill. 97% effective tank capacity is plenty to use as-is.

97% if that whole system works exactly as you think it does,
why do you even play factorio, if 97% is good enough for you?

[Taipion]

I mean... even ASSUMING everything works fine, there is no limit and no bottleneck, no weird quirks in the mechanic and all that...

So, both to utilize steam storage tanks, and to bridge segments (when building really big reactors),
you are actually supposed to to pipe-to-pump-to-pipe-lines, an I mean LINES OF PUMPS to match the requirements of your reactor?!?
Who thought that would be a good idea?
So on my 10gw reactor I have to build a line of 87 pumps to bridge segments, and to load it into storage tanks with a full 10.080mw?!?
...and then I need another line of 87 pumps to load it back from the storage tanks into the turbines?!

Again, who thought that limiting pumps to 1200/s would be a good idea?!?

I can clearly see the flawed flow of thought, and the limited consideration, clouded by not knowing what happens outside ones own little box...
"yea there won't be any more limit on pipe throughput, so we can just nerf pumps 10x like we nerfed personal lasers into oblivion (no one will use them anymore), for a little more realism in a game that is in no way meant to be realistic, I mean what can go wrong, right? ...right?!?!?

You do not need inline pumps at all ever anymore (unless your setup is bigger than 250 tiles), that is how it works. You do not need to pump into a Tank, the Tank will just work without the Pumps. Also what you really fail to consider, quality Pumps have higher throughput so the 2000 per second that your imaginary setup (wont) need are achievable easily.

It is fascinating, how everyone here believes they know it all (but ofc don't) and ignore literally everything one says...

A little bit oversimplified: A storage tank won't fill to 100% if the pipes to and from it are not also 100% AND assuming there are no weird flow quirks like in the current system.

So you will definitely need to use pumps to fill tanks to 100%, and also to empty them, let alone to control the flow.



...btw did I mention yet, that these nerfs to pumps are totally uncalled for

[GregoriusT]

97% if that whole system works exactly as you think it does,
why do you even play factorio, if 97% is good enough for you?

um why are you bottlenecking yourself to 20 percent with your arbitrary reactor setup?

[Theikkru]

[...]
97% if that whole system works exactly as you think it does,
why do you even play factorio, if 97% is good enough for you?

Are you seriously trying to tell me that you can't buffer 1 cycle of your reactor if the effective capacity of your tanks drops from 100% to 97%!? You don't have space to shove in even 1 more tank to compensate, even if you can take out the pumps!? This is really starting to sound like a skill issue on your part.

[GregoriusT]

It is fascinating, how everyone here believes they know it all (but ofc don't) and ignore literally everything one says...

A little bit oversimplified: A storage tank won't fill to 100% if the pipes to and from it are not also 100% AND assuming there are no weird flow quirks like in the current system.

So you will definitely need to use pumps to fill tanks to 100%, and also to empty them, let alone to control the flow.

It is amusing how you think Tanks are separate from Pipes somehow, which they are not. Tanks are just a bigger Pipe segment by now, and if the pipes leading to and from tanks are 100 percent, then so is the Tank, because they are all merged into ONE entity.

[Camusensei]

[...]

A little bit oversimplified: A storage tank won't fill to 100% if the pipes to and from it are not also 100% AND assuming there are no weird flow quirks like in the current system.

So you will definitely need to use pumps to fill tanks to 100%, and also to empty them, let alone to control the flow.

So your point is "Because of the pump nerf, my tanks won't be 100% full because I'll have to stop using pumps. I don't want that so I'll have to use thousands of pumps which is unreasonable." ?

[mmmPI]

Let's get you all a more practical example of why limiting even pumps to 1200/s breaks reactor designs to the core!

Not it doesn't , you can design nuclear reactor without any pumps, since you can read the steam level and the temperature of the reactor, it's easy to only refuel when you have enough storage capacity for the steam to be stored.

Say no more, you already explained in detail that you have no clue of how efficient nuclear reactors work in factorio.
Being able to read out heat lvls is "nice to have", but does not help a lot with building one, as you need to (damn... how many times I explained that already?!?!) ...as you need to be able to buffer 1 fuel cycle, the temperature at which you start the next will not always be the same as the load differs,
and steam is a much more informative source to tell when to do that, as you, PREVIOUSLY, NOT ANYMORE AFTER THE EXPANSION, were able to, with steam, to tell exactly when the heat exchangers output falls below the demand of your base, and you could design a very small and efficient steam storage to just bridge that gap.
Now, you can totally rely on reading out the heat, and build bulky reactors with way too many heat-pipes to buffer the heat, go ahead, do that...
Most people will have to do that, as pumps are crippled now and steam management simply won't work anymore. gg wube, gg

You probably didn't understood the part i underlined. If you have 1 million steam storage capacity, you know when you can refuel because you can calculate how much steam a refuel would produce even if power consumption was to stop immediatly ( 500K ? 750K? ). Therefore it is easy to know when to trigger the refuel and to be able to buffer one fuel cycle. You can also cut on the storage array capacity if you are sure your nuclear reactor will always have a drain. In no way this is requiring pumps. Just a wire connected to a single tank, and to the inserters.

Why would you want the tank filled to 100 % ?
What makes you think you need pump to achieve 100% fill with 2.0 ?

[Taipion]

It is fascinating, how everyone here believes they know it all (but ofc don't) and ignore literally everything one says...

A little bit oversimplified: A storage tank won't fill to 100% if the pipes to and from it are not also 100% AND assuming there are no weird flow quirks like in the current system.

So you will definitely need to use pumps to fill tanks to 100%, and also to empty them, let alone to control the flow.

It is amusing how you think Tanks are separate from Pipes somehow, which they are not. Tanks are just a bigger Pipe segment by now, and if the pipes leading to and from tanks are 100 percent, then so is the Tank, because they are all merged into ONE entity.

Yes, you already showcased your fabulous ability to neither read nor understand anything, it's fascinating,
but unfortunately I can't waste time with people like you, be it genuinely not understanding or really sluggish trolling attempts, I do not care.

In case you're not trolling, try read into this and actually understand it, mechanics in factorio are fun, once you get a hang of them!

[Taipion]

Let's get you all a more practical example of why limiting even pumps to 1200/s breaks reactor designs to the core!

Not it doesn't , you can design nuclear reactor without any pumps, since you can read the steam level and the temperature of the reactor, it's easy to only refuel when you have enough storage capacity for the steam to be stored.

Say no more, you already explained in detail that you have no clue of how efficient nuclear reactors work in factorio.
Being able to read out heat lvls is "nice to have", but does not help a lot with building one, as you need to (damn... how many times I explained that already?!?!) ...as you need to be able to buffer 1 fuel cycle, the temperature at which you start the next will not always be the same as the load differs,
and steam is a much more informative source to tell when to do that, as you, PREVIOUSLY, NOT ANYMORE AFTER THE EXPANSION, were able to, with steam, to tell exactly when the heat exchangers output falls below the demand of your base, and you could design a very small and efficient steam storage to just bridge that gap.
Now, you can totally rely on reading out the heat, and build bulky reactors with way too many heat-pipes to buffer the heat, go ahead, do that...
Most people will have to do that, as pumps are crippled now and steam management simply won't work anymore. gg wube, gg

You probably didn't understood the part i underlined. If you have 1 million steam storage capacity, you know when you can refuel because you can calculate how much steam a refuel would produce even if power consumption was to stop immediatly ( 500K ? 750K? ). Therefore it is easy to know when to trigger the refuel and to be able to buffer one fuel cycle. You can also cut on the storage array capacity if you are sure your nuclear reactor will always have a drain. In no way this is requiring pumps. Just a wire connected to a single tank, and to the inserters.

Why would you want the tank filled to 100 % ?
What makes you think you need pump to achieve 100% fill with 2.0 ?

See? You believe it's right to make the call on a certain number of steam "missing" to start the new fuel cycle...
...well, if you're ok with 2nd best solutions, go ahead, I fail to be able to care anymore...
Iit's just ridiculous how dense people are around here,
and given previous examples on laying out problems, I doubt that (yet again) any dev will even read any of this, and the whole thing will go unnoticed.
That's the actually sad part about it.

[GregoriusT]

Yes, you already showcased your fabulous ability to neither read nor understand anything, it's fascinating,
but unfortunately I can't waste time with people like you, be it genuinely not understanding or really sluggish trolling attempts, I do not care.

In case you're not trolling, try read into this and actually understand it, mechanics in factorio are fun, once you get a hang of them!

I think you are trolling, because you neither seem to be able to write a coherent explanation of your Issue,
nor do you seem to understand the new Fluid Mechanics and how everything including your Turbines is a merged Tank without throughput limits now,
nor do you seem to understand the fact that a Steam Tank at or near 100 percent fill level is an absolute waste of power no matter what your Reactor Design is,
nor do you seem to understand that the new way to read heat from Circuits is better than reading Steam itself,
nor do you seem to be able to read that higher Qualify Pumps will solve your specific throughput Issue.
nor do you seem to understand that you should redesign everything in 2.0 anyway, because things are different now.

So yeah, please explain in an actually coherent way, why or for what you need the Pumps at all.

Edit: And I am pretty certain Turbines do not need to be filled 100% to actually do their electric Job