even distribution of ore between stations of same name

Don't know how to use a machine? Looking for efficient setups? Stuck in a mission?
varundevan
Fast Inserter
Fast Inserter
Posts: 167
Joined: Mon Jan 20, 2020 4:41 am
Contact:

Re: enable stations of same name in sequence

Post by varundevan »

5thHorseman wrote:
Mon Jun 15, 2020 10:26 am
If your stations are unloading so fast that the trains only go to the first one, then is balancing going to help in any way?

balancing is needed because,

as the first station is unloaded more times , only the first station has resources accumulated .. and the trains come in for loading , wait for the first station to clear (because , i have added a condition to disable station if there are not enough resources)..
i observed that first station has more resources while other is empty
if that has been balanced , the waiting trains can access that ...
i hope that makes sense .. what are your thoughts ..

User avatar
5thHorseman
Smart Inserter
Smart Inserter
Posts: 1193
Joined: Fri Jun 10, 2016 11:21 pm
Contact:

Re: enable stations of same name in sequence

Post by 5thHorseman »

varundevan wrote:
Tue Jun 16, 2020 3:50 am
5thHorseman wrote:
Mon Jun 15, 2020 10:26 am
If your stations are unloading so fast that the trains only go to the first one, then is balancing going to help in any way?

balancing is needed because,

as the first station is unloaded more times , only the first station has resources accumulated .. and the trains come in for loading , wait for the first station to clear (because , i have added a condition to disable station if there are not enough resources)..
i observed that first station has more resources while other is empty
if that has been balanced , the waiting trains can access that ...
i hope that makes sense .. what are your thoughts
I missed that this was LOADING trains, not unloading them. But the same basic idea applies.

If a train arrives at a station and can not be fully loaded - and you've set it up this way - it will stay at that station taking up that slot. The next train in should go to the next station in line automatically. The trains will self-regulate in that case. It "costs" a train but it saves you having to fiddle with logic. And if that train cost is too much you probably have too many stations for what you need to do.

You could also just set the stations to be disabled if they have less ore in them than a train can hold. That way, trains will only go to stations that they can then immediately leave from.

User avatar
disentius
Filter Inserter
Filter Inserter
Posts: 694
Joined: Fri May 12, 2017 3:17 pm
Contact:

Re: enable stations of same name in sequence

Post by disentius »

So, I went a bit overboard with this.

Surprisingly refreshing to think about other people's challenges and trying to find solutions.
I made several setups, but your train grid made everything cramped, and all those chest and inserters with no good transport mechanism gave me a headache. Then I remembered the diagonal trick... :D
(feel free to hate it)

Features:
- round robin station filling
- made train grid symmetric, with less poles/roboports
- Ore output stations enable only if there is a full train to be loaded
- Waypoint stations to enable a sequential stacker
(runs with 16 trains, for testing)

I am still working on the outpost question.

Enjoy:)
exchange stationV2.gif
exchange stationV2.gif (6.12 MiB) Viewed 1579 times

User avatar
Impatient
Filter Inserter
Filter Inserter
Posts: 880
Joined: Sun Mar 20, 2016 2:51 am
Contact:

Re: enable stations of same name in sequence

Post by Impatient »

You asked about a logic to sequentially enable stations. So my posting is off-topic.

From your map screenshot of the facility I deduct that it is a reloading complex or transfer site. Trains bring ores from mines there and other trains pick up ores there to bring them to the 60 spm complexes. Am I correct with this deduction?

I use reloading complexes a lot for big factories. They make sense as logical single point of entry of a resource into the base. But, depending on desired throughput, they can be big bastards. The usual approach is to implement M unloading stations and N loading stations, where M>=N and a big balancer and maybe a buffer inbetween. Therefore I find your approach of pairs of unloading and loading stations quite interesting and actually very smart. It completely ommits the big balancer, can ommit all buffers if desired, it has almost the same throughput as in a M=N setup and (BIG PLUS!) it can be scaled (a reloading complex following the usual approach can only be rebuilt for a higher throughput), by adding more pairs and enlarging the stackers.
The only thing that 1:1 reloading can not handle at all is to balance wagon load in case the trains to be unloaded come unblanced.

AND, I think for a 1:1 reloading complex with X pairs, there does not have to be any intra-pair dependencies, like the sequential activation you asked about. How? You only need to ommit the inter-pair buffers. And set up an inter-pair circuit controller for de-/activation of the loading station. (And with a bit more of inter-pair circuit logic you also can have inter-pair buffers).

How in detail? All unloading stations have the same name and all loading stations have the same name. A train to be unloaded can chose among any of the free unloading stations. When it entered and is ready for unloading, the associated loading station is activated. Any train to be loaded will enter that loading station. If none is available the train to be unloaded has to wait. The throughput is slightly lower than in a M=N setup in the usual approach, because the transfer has to wait for the train to be loaded to move from the stacker to the loading station. Also the stacker for the trains to be loaded need circuit controlled rail signalling for the case when all of the loading stations are deactivated.

In case inter-pair buffering is required, the condition for a loading station to be active just shifts from "train to be unloaded present" to "buffer contents > 0". And the condition for a unloading station to be active shifts from no condition to "buffer contents = 0".

After writign all this, do you still need sequential activation of stations?

If yes, maybe you are planing for one pair of stations to be associated with a specific 60 SPM complex? And by sequential activation you want to distribute resources evenly to these complexes? Then you probably want that, because you don't want to have a half full train sitting at a 60 SPM complex, while others are without resources. And you want the buffers for times when there are more resources than demand. That is the only explanation I can come up with, for your requirement.

In case that is your intention, how about this general idea for balancing:
viewtopic.php?t=67235
BUT think of the chests being the buffers of the reloading station pairs, and the unloading stations being the inserters, which insert into the chests.

The resulting behavior would be this: In case all the buffers between station pairs are empty, it does not matter to which 60 SPM complex the resources are shipped, they are starved on inputs anyways and always the first one is as good as all of them in a sequence. But if resources start stacking up in the buffers, only those unloading stations are activated, where the inter-pair buffers contents is lower than the average in all the inter-pair buffers. Thus ensuring an even distribution among the pairs and hence the 60 spm complexes. It only requires one combinator per pair of reloading stations and one math combinator for all reloading pairs. It easily upscales by adding more pairs of stations and setting the divisor in the math combinator to the total number of pairs.

Would that solve your problem? I am pretty sure the solution to what you want underneath is one of these two suggestions. And each one is simpler and more robust than sequential activation.

User avatar
Impatient
Filter Inserter
Filter Inserter
Posts: 880
Joined: Sun Mar 20, 2016 2:51 am
Contact:

Re: enable stations of same name in sequence

Post by Impatient »

OK and damn. Only after I wrote my previous post, I read the other thread of yours ( viewtopic.php?f=18&t=85687 ) and got the xontext.

What I recommend now is the 2nd suggestion I made in my previous post. That ensures that no train travelling between the mines and the reloading complex has to wait for unloading (unless all buffers are full, but then it does not matter). It ensures that each pair of reloading stations has the same amount in the buffers, because, as explained, those with amounts below average (of amounts in all buffers) get refilled first. And thus it does not matter from which buffer is taken (=at what loading station is loaded).

simple+robust

varundevan
Fast Inserter
Fast Inserter
Posts: 167
Joined: Mon Jan 20, 2020 4:41 am
Contact:

Re: enable stations of same name in sequence

Post by varundevan »

disentius wrote:
Wed Jun 17, 2020 10:35 pm
So, I went a bit overboard with this.

Surprisingly refreshing to think about other people's challenges and trying to find solutions.
I made several setups, but your train grid made everything cramped, and all those chest and inserters with no good transport mechanism gave me a headache. Then I remembered the diagonal trick... :D
(feel free to hate it)

Features:
- round robin station filling
- made train grid symmetric, with less poles/roboports
- Ore output stations enable only if there is a full train to be loaded
- Waypoint stations to enable a sequential stacker
(runs with 16 trains, for testing)

I am still working on the outpost question.

Enjoy:)

exchange stationV2.gif
The train grid is the first thing i freezed , for 60 spm and i like to cramp it ..
in the end .. a 1000 spm would be achievable in a small space ..

all this time i never thought of putting trains inclined, design is different and I LIKE IT , but the number of loaders is limited to 4 per side.
i shall try to use in my design and share it ..
i really appreciate it ..THANKS FOR THE SHARING THE INTEREST
requesting to share bp with trains and stations names .. please (edit : sorry you did , i saw after bp after this post)

also , disentius , after observing for some time ,
i would like to also mention that , having only one input to the waiting stacker, clogs up the train route , have to have multiple entries (from different directions) to the waiting area ..
so went back to the old design , (the first image i shared, looks ugly , but works a bit better)

varundevan
Fast Inserter
Fast Inserter
Posts: 167
Joined: Mon Jan 20, 2020 4:41 am
Contact:

Re: enable stations of same name in sequence

Post by varundevan »

Impatient wrote:
Thu Jun 18, 2020 1:53 am
OK and damn. Only after I wrote my previous post, I read the other thread of yours ( viewtopic.php?f=18&t=85687 ) and got the xontext.

What I recommend now is the 2nd suggestion I made in my previous post. That ensures that no train travelling between the mines and the reloading complex has to wait for unloading (unless all buffers are full, but then it does not matter). It ensures that each pair of reloading stations has the same amount in the buffers, because, as explained, those with amounts below average (of amounts in all buffers) get refilled first. And thus it does not matter from which buffer is taken (=at what loading station is loaded).

simple+robust

hi IMPATIENT, i appreciate for being PATIENT and read through all the posts and giving an elaborate reply.
thanks for appreciating the design , yet it needs lot of refinement

i made this loading stations for 6 complexes (6x 60 spm tiles ) ... its an assumption ... i tried to cramp as many as can .. and after 2 days of thinking.. ended up with this design ..
actually i used and observed that sequential activation is limiting the throughput.. and discarded that idea..
from your post what i understood is that , to enable the stations that has ore less than the average of all stations ( in this case 5), sounds good ..
it works good, here is the bp that i used ..



one other problem i am facing is , since 12 trains of loading and 12 trains of unloading , total 24 trains is trying to access the stations .. and trains are getting clogged .. so the placement of this loading stations matters ... which is an another problem with this design ...
feel free to share your thoughts and ideas ..

I highly appreciate your interest and effort ...
thanks for being here

varundevan
Fast Inserter
Fast Inserter
Posts: 167
Joined: Mon Jan 20, 2020 4:41 am
Contact:

Re: even distribution of ore between stations of same name

Post by varundevan »

hi,

Got a new layout idea on the thought of increase the throughput,

in this design , trains can access the stations in more than one routes at the same time.
the drawbacks are, because of lack of stacker ... sequential enabling and even filling tricks cannot be used ...disabling the stations severely limits the throughput.
posting this to trigger a brainstorming.
share your ideas and thoughts if interested.
6.png
6.png (4.81 MiB) Viewed 1474 times
here is the bp


thank you
disentius
for making improvements in train grid, i appreciate it

Post Reply

Return to “Gameplay Help”