Page 1 of 1
Scalable Production Supply and Product Transport
Posted: Tue Apr 24, 2018 9:37 pm
by Bauer
Problem:
Scalable production setups are favouralbe in late phases of the game. However, the transport of supplies and products does not scale well. For example, additional belts are required or new train lines have to be set up.
Solution:
I suggest to use a train based solution with progressiv loading of the wagons, i.e. the train is loaded stepwise from production unit to production unit. At the loadinig/unloading station next to a production facility, I use only 2 or 3 swings of every inserter to load the train. At the next station, the train is loaded/unloaded again with another 2 or 3 swings. Thus, the train collects/delivers progressively the goods station by station.
Details:
My 4 wagon trains are nuclear powered and have a cycle time of 253 ticks for a 2 swing stop and 274 ticks for 3 swing stop. For 4x12 stack inserts (stack size 12) this corresponds to a throughput of 273 items/s and 378 items/s, respectively, for an individual station. This is the equivalent of 6.8 or 9.5 blue belts. For a 100 stack size item, the train can load 4x40x100=16.000 items. Hence the total troughput of the train line is 3794 items/s, which is the equivalent of 95 blue belts. If this is not big enough, you can easily scale by adding more wagons (and engines).
At the end you loop the train line. If you need more production units, you move the loop of the train tracks and add another unit.
All stations have the same name, including the central station. 2 stops are programmed with a leave condition G=1 from the circuit network. In case of a production facility, the G signal is generated by a timer.
The video and the following image show a green circuit production.
Per unit 273 Cu and Fe plates are being processed into 356.5 green boards per second. This corresponds nicely with the loading/unloading capacity of a 3 and 2 swing unloading station, respectively. Hence, the setup scales up upto 14 production units with a consumption of 3794.5 copper plates/s.
Re: Scalable Production Supply and Product Transport
Posted: Fri May 04, 2018 11:09 am
by Bauer
No feedback? No interest?
Lacking novelity? Incomprehensible? Boring?
Re: Scalable Production Supply and Product Transport
Posted: Fri May 04, 2018 8:45 pm
by zggz
try to make big factory with out so big stocks and you will receive feedback.
Re: Scalable Production Supply and Product Transport
Posted: Tue May 08, 2018 2:12 am
by Dry Hairy Tree
I like the way you've worked this out. There's a devil in the details of scaling up and you kicked its ass.
Feedback round these parts is rare. Just check your thread views people come and look. Then some go off and do what you thought up on youtube and call it their own. These are Factorio's 'experts'.
Play cos you love it. Get another activity for positive feedback, unless you're a Dev, cos the Devs round these parts are legends.
And I do like your work I'm just over bots and beacons for now.
Stand up comedy is my favorite way of getting lots of positive feedback. But you do need to be freakin funny. Not something I try cultivate online.
Re: Scalable Production Supply and Product Transport
Posted: Wed May 16, 2018 7:56 pm
by herkalurk
Why 4 locomotives for each train? Seems a bit overkill.
Re: Scalable Production Supply and Product Transport
Posted: Thu May 17, 2018 11:11 am
by Bauer
herkalurk wrote:Why 4 locomotives for each train? Seems a bit overkill.
Well, is there such a thing as "overkill" in factorio?
Of course you're right. My demo project had a 1:1 ration and I didn't want to recalculate.
However, I'm pretty happy with this setup, though.
Re: Scalable Production Supply and Product Transport
Posted: Thu May 17, 2018 2:29 pm
by mrvn
Doing 4 parallel stations instead of 2x 2 serial seems smarter. You are wasting the time the train needs to accelerate and decelerate between stations. With more stations that adds up and with less locomotives it becomes more pronounced. Try it with LCCCCCCCC trains and stations in series for example.
As for scaling I don't quite see what your solution brings. More production means more trains. Normally that means more trains sitting at more stations. But in your setup that means more trains visiting each station. You can only push so many trains per minute through each station and your demo shows the limit already.
Re: Scalable Production Supply and Product Transport
Posted: Thu May 17, 2018 2:35 pm
by herkalurk
mrvn wrote:Doing 4 parallel stations instead of 2x 2 serial seems smarter. You are wasting the time the train needs to accelerate and decelerate between stations. With more stations that adds up and with less locomotives it becomes more pronounced. Try it with LCCCCCCCC trains and stations in series for example.
WIth bots based stations it's easier to stack 4 stations next to each other. If you have to account for belts and inserters space quickly becomes a factor.
Re: Scalable Production Supply and Product Transport
Posted: Thu May 17, 2018 8:18 pm
by Bauer
@herkalurk and mrvn:
You don't get the point, I fear. This about expandability / ease of use. If I need more green circuits I just add another block and that's it. The train line has a huge capacity, i.e. it's enough for 14 of such blocks.
I is actually designed to work with recursive blueprints to allow for an auto expand of the factory but I didn't want to make it too complex for this post.
Re: Scalable Production Supply and Product Transport
Posted: Fri May 18, 2018 2:17 pm
by herkalurk
How many bots did you have in each block? Looks like a couple thousand flying around.
Re: Scalable Production Supply and Product Transport
Posted: Tue May 22, 2018 8:49 am
by mrvn
Bauer wrote:@herkalurk and mrvn:
You don't get the point, I fear. This about expandability / ease of use. If I need more green circuits I just add another block and that's it. The train line has a huge capacity, i.e. it's enough for 14 of such blocks.
I is actually designed to work with recursive blueprints to allow for an auto expand of the factory but I didn't want to make it too complex for this post.
I do get your intention but I'm saying your solution works against that.
Taking your design to extremes on every stop every inserter only makes one swing. That means it takes out 12 items from the train. That gives you 12 items in the time for a train to go from one station to the next + 14 ticks for the inserter to rotate. And the time for a train to go from one station to the next greatly increases with the number of cargo wagons.
So while your supply might go up going from 1 to 2 to 3 stations it will then go down and building more stations would make it slower overall. Secondly it gets wrse the longer the trains are, it doesn't scale with train size.
Re: Scalable Production Supply and Product Transport
Posted: Tue Jun 05, 2018 7:25 am
by Bauer
herkalurk wrote:How many bots did you have in each block? Looks like a couple thousand flying around.
There are about 1000 per station.
Re: Scalable Production Supply and Product Transport
Posted: Tue Jun 05, 2018 9:04 am
by Bauer
mrvn wrote:Bauer wrote:@herkalurk and mrvn:
You don't get the point, I fear. This about expandability / ease of use. If I need more green circuits I just add another block and that's it. The train line has a huge capacity, i.e. it's enough for 14 of such blocks.
I is actually designed to work with recursive blueprints to allow for an auto expand of the factory but I didn't want to make it too complex for this post.
I do get your intention but I'm saying your solution works against that.
Taking your design to extremes on every stop every inserter only makes one swing. That means it takes out 12 items from the train. That gives you 12 items in the time for a train to go from one station to the next + 14 ticks for the inserter to rotate. And the time for a train to go from one station to the next greatly increases with the number of cargo wagons.
So while your supply might go up going from 1 to 2 to 3 stations it will then go down and building more stations would make it slower overall. Secondly it gets wrse the longer the trains are, it doesn't scale with train size.
The point is that the train transport capacity is close to infinity
as it is. If your infinity is larger than mine (because you have a faster CPU, RAM, etc.), you can use 6/6 or 10/10 trains. Or 2 4/4 lines, depending on what you intent to do at the individual modules.
Having that said, what can be easily scaled up is the number of stations and thus the production capacity of -- in this case -- green circuits.
This is not about optimal production or transport. This is about ease of scaling the production up with the needs of a growing factory. You just need to blue print a new station with the production facility and that's it. No need to add more transport which I always find difficult to integrate into an existing factory, i.e. you cannot easily add another 10 belts to you main bus.
Re: Scalable Production Supply and Product Transport
Posted: Tue Jun 05, 2018 11:04 am
by mrvn
Bauer wrote:mrvn wrote:Bauer wrote:@herkalurk and mrvn:
You don't get the point, I fear. This about expandability / ease of use. If I need more green circuits I just add another block and that's it. The train line has a huge capacity, i.e. it's enough for 14 of such blocks.
I is actually designed to work with recursive blueprints to allow for an auto expand of the factory but I didn't want to make it too complex for this post.
I do get your intention but I'm saying your solution works against that.
Taking your design to extremes on every stop every inserter only makes one swing. That means it takes out 12 items from the train. That gives you 12 items in the time for a train to go from one station to the next + 14 ticks for the inserter to rotate. And the time for a train to go from one station to the next greatly increases with the number of cargo wagons.
So while your supply might go up going from 1 to 2 to 3 stations it will then go down and building more stations would make it slower overall. Secondly it gets wrse the longer the trains are, it doesn't scale with train size.
The point is that the train transport capacity is close to infinity
as it is. If your infinity is larger than mine (because you have a faster CPU, RAM, etc.), you can use 6/6 or 10/10 trains. Or 2 4/4 lines, depending on what you intent to do at the individual modules.
Having that said, what can be easily scaled up is the number of stations and thus the production capacity of -- in this case -- green circuits.
This is not about optimal production or transport. This is about ease of scaling the production up with the needs of a growing factory. You just need to blue print a new station with the production facility and that's it. No need to add more transport which I always find difficult to integrate into an existing factory, i.e. you cannot easily add another 10 belts to you main bus.
Except with your design adding a new stations (eventually) reduces the NOT INFINITE transport capacity below the production capacity of even just the first production complex.
Re: Scalable Production Supply and Product Transport
Posted: Tue Jun 05, 2018 1:04 pm
by Bauer
mrvn wrote:Bauer wrote:mrvn wrote:Bauer wrote:@herkalurk and mrvn:
You don't get the point, I fear. This about expandability / ease of use. If I need more green circuits I just add another block and that's it. The train line has a huge capacity, i.e. it's enough for 14 of such blocks.
I is actually designed to work with recursive blueprints to allow for an auto expand of the factory but I didn't want to make it too complex for this post.
I do get your intention but I'm saying your solution works against that.
Taking your design to extremes on every stop every inserter only makes one swing. That means it takes out 12 items from the train. That gives you 12 items in the time for a train to go from one station to the next + 14 ticks for the inserter to rotate. And the time for a train to go from one station to the next greatly increases with the number of cargo wagons.
So while your supply might go up going from 1 to 2 to 3 stations it will then go down and building more stations would make it slower overall. Secondly it gets wrse the longer the trains are, it doesn't scale with train size.
The point is that the train transport capacity is close to infinity
as it is. If your infinity is larger than mine (because you have a faster CPU, RAM, etc.), you can use 6/6 or 10/10 trains. Or 2 4/4 lines, depending on what you intent to do at the individual modules.
Having that said, what can be easily scaled up is the number of stations and thus the production capacity of -- in this case -- green circuits.
This is not about optimal production or transport. This is about ease of scaling the production up with the needs of a growing factory. You just need to blue print a new station with the production facility and that's it. No need to add more transport which I always find difficult to integrate into an existing factory, i.e. you cannot easily add another 10 belts to you main bus.
Except with your design adding a new stations (eventually) reduces the NOT INFINITE transport capacity below the production capacity of even just the first production complex.
(Eventually) correct! Given that a wagon is filled by ~28 swings, the "infinite" capacity is reached with 14 stations. As stated in my initial post, this corresponds to a capacity limit of 3794 items/s for the show-cased 4/4 setup, corresponding to 95 blue belts. If you plan to go bigger, you need to start the system with say 8/8 trains to get 7588 items/s, which is pretty close to infinity -- at least when it comes to my computer. (I initially started with 6/12 trains until I realized that this corresponding capacity is way larger than "infinity".)
Let me point out again: This is about scalability. I'm not playing in creative mode, so I actually have to evolve my base into a megabase. Hence, I cannot setup a production in full size. It needs to grow with the rest of the factory. In order to reduce my work, I came up with this construction. It allows me to double my production almost 4 times without the need to increase the transport of supplies or products. The latter is not only a lot of work but also tends to need more space than initially allocated for.
In fact, I am using a version employing Recursive Blueprints. Thus, the production increases by itself without my doing if the demand increases. It is some kind of self-organized organic growth, limited by my speed to add mining areas, which is very difficult to automate.
Re: Scalable Production Supply and Product Transport
Posted: Tue Jun 05, 2018 1:16 pm
by mrvn
Bauer wrote:mrvn wrote:
Except with your design adding a new stations (eventually) reduces the NOT INFINITE transport capacity below the production capacity of even just the first production complex.
(Eventually) correct! Given that a wagon is filled by ~28 swings, the "infinite" capacity is reached with 14 stations. As stated in my initial post, this corresponds to a capacity limit of 3794 items/s for the show-cased 4/4 setup, corresponding to 95 blue belts. If you plan to go bigger, you need to start the system with say 8/8 trains to get 7588 items/s, which is pretty close to infinity -- at least when it comes to my computer. (I initially started with 6/12 trains until I realized that this corresponding capacity is way larger than "infinity".)
Let me point out again: This is about scalability. I'm not playing in creative mode, so I actually have to evolve my base into a megabase. Hence, I cannot setup a production in full size. It needs to grow with the rest of the factory. In order to reduce my work, I came up with this construction. It allows me to double my production almost 4 times without the need to increase the transport of supplies or products. The latter is not only a lot of work but also tends to need more space than initially allocated for.
In fact, I am using a version employing Recursive Blueprints. Thus, the production increases by itself without my doing if the demand increases. It is some kind of self-organized organic growth, limited by my speed to add mining areas, which is very difficult to automate.
And that is where you are wrong. If a 4/4 setup gives 3794 items/s then an 8/8 setup will give more like 5000-3000 items/s, not 7588 items/s. The longer trains means it takes more time to get from station to station and longer for the next train to stop. So your throughput actually sinks. And that's why it isn't scalable.
Re: Scalable Production Supply and Product Transport
Posted: Tue Jun 05, 2018 1:29 pm
by Bauer
mrvn wrote:Bauer wrote:mrvn wrote:
Except with your design adding a new stations (eventually) reduces the NOT INFINITE transport capacity below the production capacity of even just the first production complex.
(Eventually) correct! Given that a wagon is filled by ~28 swings, the "infinite" capacity is reached with 14 stations. As stated in my initial post, this corresponds to a capacity limit of 3794 items/s for the show-cased 4/4 setup, corresponding to 95 blue belts. If you plan to go bigger, you need to start the system with say 8/8 trains to get 7588 items/s, which is pretty close to infinity -- at least when it comes to my computer. (I initially started with 6/12 trains until I realized that this corresponding capacity is way larger than "infinity".)
Let me point out again: This is about scalability. I'm not playing in creative mode, so I actually have to evolve my base into a megabase. Hence, I cannot setup a production in full size. It needs to grow with the rest of the factory. In order to reduce my work, I came up with this construction. It allows me to double my production almost 4 times without the need to increase the transport of supplies or products. The latter is not only a lot of work but also tends to need more space than initially allocated for.
In fact, I am using a version employing Recursive Blueprints. Thus, the production increases by itself without my doing if the demand increases. It is some kind of self-organized organic growth, limited by my speed to add mining areas, which is very difficult to automate.
And that is where you are wrong. If a 4/4 setup gives 3794 items/s then an 8/8 setup will give more like 5000-3000 items/s, not 7588 items/s. The longer trains means it takes more time to get from station to station and longer for the next train to stop. So your throughput actually sinks. And that's why it isn't scalable.
I see. I compensated (but not fully) by putting additional signals between the chests in the station. I measured the loss factor for longer trains but don't have the data with me. It wasn't very big, like 10%. Even going to fewer engines per wagon has only a small impact on throughput.
Since the train size and configuration is a constant in my setup anyway, this effect doesn't effect scalability anyway.