Friday Facts #274 - New fluid system 2

[vanatteveldt]

update sounds very good, and thanks/congrats to TheYeast for the contribution!

what does the green and the blue column in the demo gif exactly means? And will there be an option to show these columns to get an easy way to identify bottlenecks and something like that?

It is the volume and speed - numbers you can find in the tooltip, but visible all at once this way. It is already available - go to debug settings ingame (f4) and check show-fluid-box-fluid-info.

actually it would be great if you could see something about this through the windows so you can see the fluid flowing through then... maybe show volume as fluid level through the window?

[Byproduct]

That scorpion thing is so cool, congrats!

[5thHorseman]

The current limit is 0.05 fluid, not volume or %. So 100x less than what you write.

Just to be clear, then, if you put 100 pipes in a long line, and slowly empty them so that each pipe ends up with 0.05 units of a liquid, that liquid will stay in all pipes. Then if you keep emptying it, all 100 pipes will have 0.04 units, then 0.03, etc until they get all the way down to 0.0005 units each (for a total of 0.05 in the entire 100-pipe line), at which point the entire 100-pipe chain will have 0.05 units and that 0.05 units will be deleted?

Or, will all the liquid get deleted when each of the 100 pipe segments is lowered to 0.05 units, which is the impression that Caddie (and myself) had prior to your reply.

[Nightinggale]

* The actual speed is the speed of light, which is 299.792.458 m/s. Depending on the metal, it can lose a little bit of speed, but it's still very close to that number. 300.000.000 is a common approximation as it's often close enough and it makes the calculations much easier, like allowing all calculations in this post to be done without a calculator. It doesn't really matter if the answer is 4k or 5k km/tick. Both are way too fast to justify Factorio not assuming the speed to be instant.

Actually, the travel speed depends on the dielectric constant of the volume around the wire and switching speeds depend on capacitance. This is why microchip manufacturers are trying to reduce the dielectric constant in the interconnect insulators, but increase it in MOSFET (generic gate-dielectric-channel type transistor however it's formed) gate insulator. It is usually a significant fraction of the speed of light, but is still significantly slower.

You are right, but that doesn't mean I'm not right at the same time. The thing is you are talking about tightly packed wires next to each other with insulation in between with a not insignificant dielectric constant. I on the other hand only wrote about the power grid. This is uninsulated copper wires in mid air and without insulation at all, we can ignore the capacitance in the wires.

Since the capacitance in wires can be ignored for the Factorio power grid, I decided to not mention it at all. I could also have mentioned that the spread of current in a wire depends on frequency, meaning AC doesn't scale linearly. Twice the cross area of a wire will allow twice the current for DC, but less than twice the AC current. That's another detail I decided not to include in order to keep it simple.

Capacitance in wires leads to increased power loss for AC wires, which is actually the main reason why high voltage wires tend to be uninsulated and placed far into the air. It's not because it's the cheapest way to build wires. That's just a bonus. The concern is about how much power a powerplant has to produce to power your 100 W computer and uninsulated wires are the best for that job.

Underwater cables suffers from huge capacitance in the insulation and unless they are fairly short, they tend to be DC because DC doesn't change voltage, hence doesn't suffer from cables not wanting to change voltage. Underwater cables have a certain distance where AC just won't pass as in 100% power loss. The distance is rather short, like a few hundred km. DC doesn't suffer from this and the UK is building a power cable to Iceland because Icelandic powerplants run on steam coming out of the ground, hence no fuel or emissions.

99% of the reason we have AC instead of DC on large scale power grids is because it allows us to use non-switching transformers to step the voltages up and down (and I can't remember what the other 1% of the reason was.)

The primary reason is historical. Edison built the first powerplant in Pearl street on Manhattan. It delivered 110 V DC. This meant low voltage, hence high current and the distance from powerplant to consumer was 100-200 meters (or yards. The uncertainty is much greater than the difference between a meter and a yard). Westinghouse went with AC because transformers allows high voltage, hence low current and low power loss. This allows consumer and powerplant to be placed in different states. The problem was that DC motors worked ok while AC motors was unusable. A Westinghouse employee named Nikola Tesla figured out how to produce a 3 phased motor, which could compete against the DC motor. Westinghouse then pushed forward with poly phased power as it was called back then. It won the war of the currents, resulted in country wide AC power grids in most countries and ages after everything was up and running in large grids, somebody invented a way to make a "DC transformer". The need was however not great as it doesn't make sense to replace existing and working AC for DC, particularly because the price of doing so would be really high.

DC is used for special cases, like underwater/underground cables, connecting two powergrids not using the same frequency or using the same frequency, but are out of phase etc. This requires another invention (or 2), which allows moving between DC and AC.
If you wonder where such conversions takes, place, then here are some examples. Back in the day when power grids were just for a city and cities weren't connected, one Japanese power company bought generators in America and another bought in Europe. The result is that today Japan has both 50 and 60 Hz, which is causing problems, but not significantly enough to change. Instead transformer stations goes AC-DC-AC to bridge the frequency areas. Another example is the fact that Scandinavia runs 50 Hz like the rest of Europe, but being isolated prior to introduction of underwater cables, the voltage doesn't peak at the same time, meaning connections have to use the Japanese approach even though both sides are 50 Hz. There is also the fact that railroads doesn't always use the same frequency as the power grid. Germany use 16.7 Hz for trains (again for historical reasons. Early AC engines got hotter with high frequency and 50 Hz would melt the cobber in such high power motors and Austria/Germany invented/pioneered AC trains, hence using a much older standard than other countries, like pre WW1 vs other countries deciding on frequency during the 60s to 80s), meaning power to overhead wires have to use AC-DC-AC conversions here as well.

Wire resistance cares about current and not voltage, while the total power is the product of the two.

Power delivered: P = U*I, as in power(W) = Voltage(V) * Current(A)
Power loss per km/mile/whatever: P = k*I*I (Ohm's law combined with the previous equation)
k is a constant based on the thickness of the wire. This means lowering k makes the wire thicker, hence more expensive and less bendy, which in turn makes it more difficult to put it up.

Combine this and you see that you want a low I. Since you want a constant U*I, you naturally want a high U, hence high voltage.

This is getting way off topic. The fact remain that for power grids (and in general), Factorio doesn't have the size to make it make sense to consider the speed of electricity in the power grid to be anything other than instant.

[Amarula]

Heartfelt thanks to everyone for contributing to these discussions. Even if it is off-topic, I have learned a lot about electricity and fluid dynamics! I especially want to thank those who take the time to snip out exactly the words they want when responding to another post. Do you have any tips for making it easier than going through the quoted material and deleting everything you don't want? For those who want to reply, but not to any specific part, please consider using "@previousPoster'sName" instead of quoting - especially instead of quoting more than 3 or 4 lines. When I see a long post quoted entirely, I hear the voice of the teacher from Charlie Brown "BLAH BLAH BLAH"


Thank you to Dominik, and all the other devs and staff, for being here on the forums, and for the magic that is Factorio.
And to all of you, a very Merry Christmas and best wishes for a safe, happy and Factorio-filled New Year!

[vanatteveldt]

Heartfelt thanks to everyone for contributing to these discussions. Even if it is off-topic, I have learned a lot about electricity and fluid dynamics! I especially want to thank those who take the time to snip out exactly the words they want when responding to another post. Do you have any tips for making it easier than going through the quoted material and deleting everything you don't want? For those who want to reply, but not to any specific part, please consider using "@previousPoster'sName" instead of quoting - especially instead of quoting more than 3 or 4 lines. When I see a long post quoted entirely, I hear the voice of the teacher from Charlie Brown "BLAH BLAH BLAH"


Thank you to Dominik, and all the other devs and staff, for being here on the forums, and for the magic that is Factorio.
And to all of you, a very Merry Christmas and best wishes for a safe, happy and Factorio-filled New Year!

yeah I agree

[Dominik]

The current limit is 0.05 fluid, not volume or %. So 100x less than what you write.

Just to be clear, then, if you put 100 pipes in a long line, and slowly empty them so that each pipe ends up with 0.05 units of a liquid, that liquid will stay in all pipes. Then if you keep emptying it, all 100 pipes will have 0.04 units, then 0.03, etc until they get all the way down to 0.0005 units each (for a total of 0.05 in the entire 100-pipe line), at which point the entire 100-pipe chain will have 0.05 units and that 0.05 units will be deleted?

Or, will all the liquid get deleted when each of the 100 pipe segments is lowered to 0.05 units, which is the impression that Caddie (and myself) had prior to your reply.

Sorry, my bad, I misunderstood what you wrote.
So yes, if you put 5 fluid into 101 pipes, it should eventually disappear. But that is not a very frequent case. Most common case is the pipes are full all the time and that loss does not seem very large to me. Not great, sure, but lower limit means more time when you do want to empty it. If you have another idea how to solve the thing it would be great.
In my original simulator what I did was that instead of destroying the fluid I sent it to some output so it never got lost. But that only worked because the output was not using the fluid, just counting it. In the game machines can't make use of some small amount of fluid so it would still be stuck there.

[FasterJump]

I don't want fluids to disappear either. [...]

Actually you made me changed my mind about the 0.05 threshold, I'm less enthusiastic.

What if pumps had an option, a "Power Purge" checkbox that would activate the 0.05 threshold in the drained pipe block?

Most common case is the pipes are full all the time

I'm not sure. If you consume more that you produce, the pipes will be 99% empty, even if the flow-rate is important. This will happen especially if the pipe segment is short and/or if the fluid has low frictions like steam.

[Nightinggale]

I don't want fluids to disappear either. [...]

Actually you made me changed my mind about the 0.05 threshold, I'm less enthusiastic.

What if pumps had an option, a "Power Purge" checkbox that would activate the 0.05 threshold in the drained pipe block?

I have been thinking something along those lines. Real life oil pipelines have something called a pig. It has multiple purposes and in this specific case, what it would do would be to enter the pipe and push the remaining liquid to the end. It would clean the pipe, hence removing the hardcoded liquid type and making it ready for a new type. The old liquid would end up in a single pipe at the end where we could make it vanish.

Now if entities adding liquid to pipes could have a "send pig" option, then we can clear pipes without rebuilding while at the same time preventing deleting liquid against the will of the player.

Alternatively we can make a special pipe with a send pig option. That might actually be better as it will keep things simple. It will also allow adding circuit control. This in turn would allow some rather interesting setups like having pumps with multiple types of liquid pumping into the same pipe, but only one active at a time. Combinators would then grant time for each liquid and effectively we could end up with a pipe, which can handle multiple liquids, but only one at a time. Real life pipelines do in fact switch liquid, like gasoline and diesel.

This would require some way of filtering the output to make it really useful, though just being able to switch one liquid to another in order to match the input for say an assember when switching recipe would be really useful.

Most common case is the pipes are full all the time

I'm not sure. If you consume more that you produce, the pipes will be 99% empty, even if the flow-rate is important. This will happen especially if the pipe segment is short and/or if the fluid has low frictions like steam.

I have had cases of lack of liquid due to various, often uneven arrival of solid input, making liquid(gas) production stall or refineries stalling because the tank for say heavy oil is full, allowing the light oil to run out. In other words it is a very realistic scenario to have liquid you want to keep, which just happens to run low in the pipes and emptying the pipes will just make the problem worse.

[featherwinglove]

* The actual speed is the speed of light, which is 299.792.458 m/s. Depending on the metal, it can lose a little bit of speed, but it's still very close to that number. 300.000.000 is a common approximation as it's often close enough and it makes the calculations much easier, like allowing all calculations in this post to be done without a calculator. It doesn't really matter if the answer is 4k or 5k km/tick. Both are way too fast to justify Factorio not assuming the speed to be instant.

Actually, the travel speed depends on the dielectric constant of the volume around the wire and switching speeds depend on capacitance. This is why microchip manufacturers are trying to reduce the dielectric constant in the interconnect insulators, but increase it in MOSFET (generic gate-dielectric-channel type transistor however it's formed) gate insulator. It is usually a significant fraction of the speed of light, but is still significantly slower.

This is uninsulated copper wires in mid air and without insulation at all, we can ignore the capacitance in the wires.

Aluminum. Other than that, it's refreshing to find someone who knows what they're talking about. Double conversion does require inverters, which means it is switched and somewhat more complex than a simple transformer. The most common type are the ones in consumer electronics which change DC voltages from, for example, 1.5V-ish of a single AA battery in a cordless mouse to the 5V that its electronics needs. Where I lived which had LRT (Calgary, AB), the traction electricity was 600VDC via pantograph; I believe Siemens manufactured the cars.

Back to the fluid topic, I do believe there are some mods which add pig-like mentality ...er, functionality. I haven't used any though.

[nosports]

So yes, if you put 5 fluid into 101 pipes, it should eventually disappear. But that is not a very frequent case. Most common case is the pipes are full all the time and that loss does not seem very large to me. Not great, sure, but lower limit means more time when you do want to empty it. If you have another idea how to solve the thing it would be great.

You could insert the Fluid into a Barrel when carrying one or the resources for Said Barrel.
Place the füllend Barrel into User inventory Ort to the ground

[Zavian]

So yes, if you put 5 fluid into 101 pipes, it should eventually disappear. But that is not a very frequent case. Most common case is the pipes are full all the time and that loss does not seem very large to me. Not great, sure, but lower limit means more time when you do want to empty it. If you have another idea how to solve the thing it would be great.

"Most common case is the pipes are full all the time". That is only true if supply exceeds demand. Experienced players might achieve that most of the time, but newer/less experienced players often underestimate how many pumpjacks/refineries/chem plants they need, and are often chronically short of at least something for substantial period of time. I would even suggest that amongst most players, demand exceeding supply, of at least something is normal.

Consider a new player with one refinery (using advanced oil processing) and a chem plant making 10 lubricant every 300 ticks (the chem plant is limited by heavy oil output of that one refinery). They then pipe that lube to their bus/production area where the pipe runs past their utility manufacturing area where it splits off 4 time to supply 4 different assemblers for blue belts/splitters/undergrounds/plus 1 assembler for bot engines, and then up the bus to 5 assemblers for electric engines for science. Total length of pipe could easily be over 200 pipes if they aren't using undergrounds, and there are a lot of split offs, each of which will halve the flow of fluid. (Remember this design is fluid starved, so both sides of every split will want fluid). By the time the pipes have 8 junctions the 2 outflow pipes from the last junction get 0.039 lube from each pulse of 10 lube the chem plant produces. (This is simply splitting the fluid evenly between each outgoing pipe in each junction. Viscosity and fluid flow mean that that 0.039 lube will probably trickle into the pipe segments over multiple ticks). What are the odds that the player will lose lube regularly before it gets to the last electric engine assembler? Even if he adds another 4 refineries (so he is producing 10 lube once every 60 ticks), will those last electric engine assemblers get any lube? (Obviously if the blue belt/splitter/underground assemblers fill their output/run out of materials, those junctions will no longer be hungry for lube, and things will improve, but this demonstrates that in cases where demand outstrips supply, multiple junctions where both sides want fluid can quickly decimate fluid pulses below 0.05 units, which leaves me worried that the assemblers on the end of 10 junctions might not get any fluid, and fluid might regularly 'vanish', when demand exceeds supply).

As for suggestions to solve the issue. How feasible would it be to keep a count of the number of fluid sources that a set of pipes is connected to? Then only activate the clearing mechanism if there are no fluid sources connected to the pipe. (That is potentially more performant as well, since instead of checking the fluid in multiple pipes per tick, you can simply maintain a list of pipe networks that don't have a connected fluid source, and simply check one of those every X ticks).

EDIT: You could even track the 'free end' (the pipe that was connected to the last fluid source/the pipe that was adjacent to the pipe that was removed to remove the connection to the last fluid source). Then once that free end reaches 0.05 fluid, just start pushing fluid into any adjacent pipe(s) until you reach the fluid box of an assembler/chem plant, and only destroy any fluid that couldn't be pushed that way.

[Brathahn]

Sorry, my bad, I misunderstood what you wrote.
So yes, if you put 5 fluid into 101 pipes, it should eventually disappear. But that is not a very frequent case. Most common case is the pipes are full all the time and that loss does not seem very large to me. Not great, sure, but lower limit means more time when you do want to empty it. If you have another idea how to solve the thing it would be great.
In my original simulator what I did was that instead of destroying the fluid I sent it to some output so it never got lost. But that only worked because the output was not using the fluid, just counting it. In the game machines can't make use of some small amount of fluid so it would still be stuck there.

what if you have a single pumpjack on the end of a long pipe? will all the oil vanish even before it gets to the other end of the pipe?
this kind of setup is quite common for spread apart oil patches.

[pleegwat]

In my original simulator what I did was that instead of destroying the fluid I sent it to some output so it never got lost. But that only worked because the output was not using the fluid, just counting it. In the game machines can't make use of some small amount of fluid so it would still be stuck there.

In game assemblers can't use arbitrary small amounts. But if the player is actively emptying the system, they're probably using a pump which can pump arbitrary small amounts.

[Zavian]

In my original simulator what I did was that instead of destroying the fluid I sent it to some output so it never got lost. But that only worked because the output was not using the fluid, just counting it. In the game machines can't make use of some small amount of fluid so it would still be stuck there.

In game assemblers can't use arbitrary small amounts. But if the player is actively emptying the system, they're probably using a pump which can pump arbitrary small amounts.

Also whilst the assemblers can't necessarily use small amount, it is quite common for me to have an assembler with 9.9 out of 10 fluid, so pushing a few small amounts might be enough to let the assembler cycle again. Better to push it somewhere where it might be useful, than just destroy it. (If I'm changing recipes, then typically I just accept that some assemblers will waste some fluid).

[Dominik]

Guys, keep in mind that the 0.05 is a limit for EVERY fluidbox in the system. If any one has more, it does not purge. So in order to trigger that you really have to empty the pipes and wait. In realistic cases, there would be more fluid by some producer.
If someones oil production and it stalls and some pipes empty (sure happens to me too) then so what if once in a while 5 fluid out of 10k gets lost?
I will test it out on some very bad setup to make sure.
Adding an entirely new mechanism is not likely though, too much work to do on guis too

[Lubricus]

Guys, keep in mind that the 0.05 is a limit for EVERY fluidbox in the system.

It don't need to apply to fluid boxes inside assembly machines and boilers... 0.05 is a very small amount so i think it's OK. Modders out there don't do as super high value low volume fluid. Picker extended have a delete fluid command "Picker Pipe Cleaner" https://mods.factorio.com/mod/PickerExtended

[Lizzy]

When I put a pump in the middle of a fluid system, does it split the system in two? If that's the case, I would think working with a depleted pumpjack (no modules of course) could cause up to 50% disappearing crude oil.

[Blacky007]

Guys, keep in mind that the 0.05 is a limit for EVERY fluidbox in the system. If any one has more, it does not purge. So in order to trigger that you really have to empty the pipes and wait. In realistic cases, there would be more fluid by some producer.
If someones oil production and it stalls and some pipes empty (sure happens to me too) then so what if once in a while 5 fluid out of 10k gets lost?
I will test it out on some very bad setup to make sure.
Adding an entirely new mechanism is not likely though, too much work to do on guis too

try to use a variable that can be modified eg. testing 0,05 till 0,01

[Dominik]

When I put a pump in the middle of a fluid system, does it split the system in two? If that's the case, I would think working with a depleted pumpjack (no modules of course) could cause up to 50% disappearing crude oil.

pumps don't split systems, for efficiency reasons.