Contest: Longest train based nuclear reactor

Power Plants, Energy Storage and Reliable Energy Supply. All about efficient energy production. Turning parts of your factory off. Reliable and self-repairing energy.
Locked
mrvn
Smart Inserter
Smart Inserter
Posts: 5703
Joined: Mon Sep 05, 2016 9:10 am
Contact:
Contact mrvn

Contest: Longest train based nuclear reactor

Post by mrvn »

I'v been playing around with putting steam from nuclear reactors into fluid wagons. I started with L>LFFFF trains but you need a lot of trains and congestion sets in.

So I've been thinking: What is the longest train you can fill from a nuclear reactor directly (heat exchanger -> pump -> fluid wagon)?

I was going for maximum speed so 3 pumps per fluid wagon. But if you can produce more trains per minute with less ... No idea what is ideal.

The steam per fuel should be maximized, which I believe only allows a 2xN reactor design.

Results fall into 2 categories:

1) reactor overheats so you have to delay refueling a bit every cycle

For this the delay should be minimal. Add another fluid wagon if you can to take away more steam to keep the reactor cooler. But overall the steam per second it the deciding factor.

2) reactor cools down so the train waits a bit on filling

For this the waiting should be minimal. The question should be: how short a train can you fuel without the reactor overheating when burning fuel without pause?

But again the steam per second is more important. If a longer wait produces more steam then go for it.

Example for category 1:

I figured out experimentally that you can fill a 15 fluid wagon train with an end stop with heat exhangers on both sides of the train when using heat pipes. With 16 fluid wagons the last heat exchanger drop to 500°C with the reactor at 1000°C. It might be impossible to not waste fuel with 15 fluid wagons.

Can you beat that with heat pipes? Maybe have a through station with the reactor in the middle of the train and using only 2 pumps per fluid wagon? Fill 2 trains in parallel.

Example for category 2:

Here is my test setup with a 50 fluid wagon train. 50 you say? I thought at 16 the heat exchangers had low temp? Well, nuclear reactors conduct heat a lot better:
50-wagon-nuclear.png
50-wagon-nuclear.png (808.22 KiB) Viewed 6490 times


It uses a 2x4 reactor that is actually fueled and all others are just heat conductors. A 2x4 reactor can sustain 112 heat exhcnagers. The 50 fluid wagons use 150. And even with the pause between trains the 2x4 reactor cools down. The train should probably be shorter to meet the rules above. So this doesn't qualify, it's just an example of what you can do. It's also cheating with infinity pipes for the water. YOu should have offshore pumps or bring water by train.

Can you suggest improvements to keep it at 50 fluid wagons or make it even longer? Fueling a larger reactor becomes a problem.

Maybe a small section of heat pipes and then switching to nuclear reactors is the way to go. That would leave room for inserters to fuel a larger reactor. Other ideas?
Top
coppercoil
Filter Inserter
Filter Inserter
Posts: 472
Joined: Tue Jun 26, 2018 10:14 am
Contact:
Contact coppercoil

Re: Contest: Longest train based nuclear reactor

Post by coppercoil »

Should the design be tileable?
Top
mrvn
Smart Inserter
Smart Inserter
Posts: 5703
Joined: Mon Sep 05, 2016 9:10 am
Contact:
Contact mrvn

Re: Contest: Longest train based nuclear reactor

Post by mrvn »

Wouldn't hurt but up to you.
Top
mmmPI
Smart Inserter
Smart Inserter
Posts: 2728
Joined: Mon Jun 20, 2016 6:10 pm
Contact:
Contact mmmPI

Re: Contest: Longest train based nuclear reactor

Post by mmmPI »

mrvn wrote: ↑
Mon Oct 25, 2021 11:02 pm
 So this doesn't qualify, it's just an example of what you can do. It's also cheating with infinity pipes for the water. YOu should have offshore pumps or bring water by train.
But then that makes it an example of what you CAN'T do no ? i' m lost here.

What If instead of infinity pipe i use the editor to put water where i need and use offshope pump it's different than infinity pipe ?

What if just take a map with lots of water and landfill everything except the area where i need water so i can put an offshore pump ?
Top
User avatar
DaveMcW
Smart Inserter
Smart Inserter
Posts: 3700
Joined: Tue May 13, 2014 11:06 am
Contact:
Contact DaveMcW

Re: Contest: Longest train based nuclear reactor

Post by DaveMcW »

The highest possible train speed is 1.38 tiles per tick. This works out to 11.83 fluid wagons per second.
11.83 fluid wagons per second can carry 295714 steam per second.
To produce 295714 steam per second we need 2869 heat exchangers.
2869 heat exchangers require 28685 MW of power.
28685 MW of power requires a 2x120 reactor.

Therefore, we can power any size train with a 2x120 reactor and enough steam pipes!

We should build a bigger reactor, because we will need to power a lot of pumps to distribute steam to our infinitely long train.
Top
mrvn
Smart Inserter
Smart Inserter
Posts: 5703
Joined: Mon Sep 05, 2016 9:10 am
Contact:
Contact mrvn

Re: Contest: Longest train based nuclear reactor

Post by mrvn »

mmmPI wrote: ↑
Tue Oct 26, 2021 10:24 am
mrvn wrote: ↑
Mon Oct 25, 2021 11:02 pm
 So this doesn't qualify, it's just an example of what you can do. It's also cheating with infinity pipes for the water. YOu should have offshore pumps or bring water by train.
But then that makes it an example of what you CAN'T do no ? i' m lost here.

What If instead of infinity pipe i use the editor to put water where i need and use offshope pump it's different than infinity pipe ?

What if just take a map with lots of water and landfill everything except the area where i need water so i can put an offshore pump ?
It's a starting point that would need to be optimized to qualify. Cut it down to 38-40 fluid wagons so the reactor doesn't cool down as much and the delay for trains is minimized. Maybe qualify is the wrong word. Cutting down the design to the proper size would beat this one. It's not going to be the winning reactor. Replace the infinity pipes with offshore pumps somewhere because that really disqualifies it. But with a single train stop like this that's easy to do, I was just lazy last night.

Using waterfill to place offshore pump or the editor is perfectly fine. Using a big lake would be the way in a real game where you don't have waterfill. The result is all the same. The difference is that you have to leave space for the offshore pumps and the 2x3 water holes. You can landfill them after placing the offshore pumps but I always feel bad about that hack.
Top
mrvn
Smart Inserter
Smart Inserter
Posts: 5703
Joined: Mon Sep 05, 2016 9:10 am
Contact:
Contact mrvn

Re: Contest: Longest train based nuclear reactor

Post by mrvn »

DaveMcW wrote: ↑
Tue Oct 26, 2021 10:50 am
 The highest possible train speed is 1.38 tiles per tick. This works out to 11.83 fluid wagons per second.
11.83 fluid wagons per second can carry 295714 steam per second.
To produce 295714 steam per second we need 2869 heat exchangers.
2869 heat exchangers require 28685 MW of power.
28685 MW of power requires a 2x120 reactor.

Therefore, we can power any size train with a 2x120 reactor and enough steam pipes!

We should build a bigger reactor, because we will need to power a lot of pumps to distribute steam to our infinitely long train.
But can you:

1) Break the train into the loading station and accelerate it out of it fast enough?
Any locomotives on the train use up space that could be used for steam. What % of the train would be locomotives? What's the average speed then? You can't fill a train while it is moving. So the carrying capacity of rails is basically useless information.

2) Keep 2869 heat exchangers supplied with heat from a 2x120 reactor?

The reactor can only be 1000°C and it has to have some temperature range left for when the train changes. 2869 heat exchangers would need 6694 tiles to be spread along a train. Heat pipes fail after 15 wagons. I'm not sure how far reactors can spread the heat but 1116 reactors seems a bit much.

3) Who says I only have one rail track?

2 rail tracks can carry 23.66 fluid wagons per second. I don't think the rail track is the limiting factor. If you manage to make a tilable design the steam/s would be infinite. Those would have to be judged on steam/s/tile or the amount of dead time (either in the reactor or filling trains). I kind of gets into the I-know-it-when-I-see-it country.
Top
User avatar
Impatient
Filter Inserter
Filter Inserter
Posts: 883
Joined: Sun Mar 20, 2016 2:51 am
Contact:
Contact Impatient

Re: Contest: Longest train based nuclear reactor

Post by Impatient »

@DaveMcW: That seems to be the highly theoretical case, where one train can be loaded continuously while moving at max speed.

Love that approach, because it leaves space for my engineering.
Here is my unrequested attempt to transform it to ingame-reality:
1.) Needs S sibling stations + infrastructure to handle S trains from the start of their decelleration stage to the end of their acceleration stage (+an adequately sized stacker to account for fluctuations in demand and rail network throughput - if buffering is decided to be done that way) + the circuitry to load trains sequentially and not in parallel.

2.) A throughput of 300k steam/s needs 100 pipelines (with a pump-upipe-pump-upipe-... composition, which has a mxc throughput of 3k/s) per station (100 pipelines x S stations).

3.) Tillability says there are 7 tiles available per wagon for loading infrastructure. But I approach the use of the 7th tile with caution. Depending on the design, it could pipe-connect the infrastructure of neighboring wagons and thus, because of the un/predictability of the fluid system, lead to unequal and maybe therefor longer loading times. But that I would need to think through case by case or test.

4.) 7 tiles can acommodate 7 pipelines with a joint throughput of 7x3k/s = 21k/s

5.) Only 2 3k/s pipelines can be merged for loading as the min pipe distance between pumps in such a merge is 1, which has a max throughput of 6k/s. Therefor one loading pump can load with max 6k/s, there for 4 loading pumps are required for 21k/s. Only 3 loading pumps can connect and thus there is another limit of 18k/s and only 6 pipelines per wagon are required -> 6 tiles/wagon suffice - which all can be merged on one side of a wagon -> one side loading suffices. Edit: I tested this and because the fluid system has its on will, this is not possible. 2x2 pipes would need to be merged on one side and the remaining 2 on the other side.

6.) The total throughput of 300k steam/s divided by 18k steam/s/wagon computes to 16.67 wagon(s). So the number of wagons per train needs to be 17 unless parallel loading is an option (which comes with its own pletora of complications and buffering requirements, depending on if circuit controlled or not, when trying to maintain a continuous i/o of trains).

7.) When 300k/s throughput is a strict requirement, I would also implement some sort of train buffering capability and/or priority track signals on the output side, to account for congestion in the rail network. A side note.

.) bakc a the loading procedure, the pumping itself only needs 25k/18k/s= 1.39s . Pump at- and detachement time can be measured in a test. The max time of a standing train in the stacker to a standstill in a station can be measured. Max leaving time into the output buffer or the rail network can be measured. The sum of all those times dictates the number of required sibling stations S.

.) The by far most funny challenge is to accomodate 100 pipelines for S stations. But unless one wants a belt of several hundret piplines all round the array of heat exchangers and the whole complex to stretch accross a huge area, the same 100 pipelines have to be used for all S stations. Which requires a fork system at each station, which is only 7 tiles wide, while maintining a loading and piping throughput of 18k/s.

... which i think is the end of this design/implementation attempt. :? :lol:
Top
mrvn
Smart Inserter
Smart Inserter
Posts: 5703
Joined: Mon Sep 05, 2016 9:10 am
Contact:
Contact mrvn

Re: Contest: Longest train based nuclear reactor

Post by mrvn »

That's kind of missing the challenge. It's not a heat-exchanger - pump - wagon setup.

What you describe is a standard nuclear reactor to one side and replacing the steam turbines with loading stations with long pipes transporting the steam from reactor to trains. Multiple heat-exchangers go into each pump so they can get close to 1200 fluid/s each.

If you want that kind of design then build a 2xN tileable reactor of your choice, remove the steam turbines and put parallel through stations there. Every 7 reactors would feed 6 fluid wagons so you have to do some balancing. As you mention connecting fluid pipes leads to oddities. So I would have 7 parallel loading stations and load 6 wagons each from one of the 7 reactors. In such a design there is also no reason not to have steam tanks. So 7 parallel loading stations the length of the 2xN reactor (N a multiple of 7) is really all you need and 7 on the other side, so 14 total.

The challenge for the contest is to transmit the heat along the train, not the steam.
Top
User avatar
Impatient
Filter Inserter
Filter Inserter
Posts: 883
Joined: Sun Mar 20, 2016 2:51 am
Contact:
Contact Impatient

Re: Contest: Longest train based nuclear reactor

Post by Impatient »

mrvn wrote: ↑
Tue Oct 26, 2021 5:01 pm
 That's kind of missing the challenge. ...

The challenge for the contest is to transmit the heat along the train, not the steam.
Sounds far too easy. ;)

Edit: I am just joking. Also I meant in comparrison to my own pipe attempt, which addmitedly missed your specs and also just in response to daves theoretical numbers.
Top
User avatar
Impatient
Filter Inserter
Filter Inserter
Posts: 883
Joined: Sun Mar 20, 2016 2:51 am
Contact:
Contact Impatient

Re: Contest: Longest train based nuclear reactor

Post by Impatient »

In a bottom up step, this is the one tillable pattern for a direct, 3 connection design I found:

tillable hx2wagon setup 01.jpg
tillable hx2wagon setup 01.jpg (119.98 KiB) Viewed 6373 times

Measuring the water throughput, it gave me a reliable 3k/s.
('writing "reliable", because in the upper row it is ~500/s higher ... which makes absolutely no sense to me, when comparing the pattern repetition in both rows. But whatever ... fluid system).

Just wanted to share this finding.
3k/s is good for 50 hx (for everyone who doesn't know).
3k/s is good for (3k/s)/(103.1/s) = 29.9 hx (how could I forget that)

tillable hx2wagon setup 01 water tp.jpg
tillable hx2wagon setup 01 water tp.jpg (141.66 KiB) Viewed 6373 times

But it does not align with reactor connection points, to keep them an option for heat conduction. ....
Last edited by Impatient on Tue Oct 26, 2021 10:54 pm, edited 1 time in total.
Top
User avatar
Impatient
Filter Inserter
Filter Inserter
Posts: 883
Joined: Sun Mar 20, 2016 2:51 am
Contact:
Contact Impatient

Re: Contest: Longest train based nuclear reactor

Post by Impatient »

I found a tillable pattern that alligns with reactor connection points at 70 tiles.
(No surprise, as 70 is the least common multiple of 7 (6 wagon length + 1 gap) and 10 (2x5 reactor width) and where in a single wagon setup all connections could be made. Doesn't mean there aren't shorter ones. But there had to be one at that length for sure.)

tillable hx2wagon setup 02 reactor aligned.jpg
tillable hx2wagon setup 02 reactor aligned.jpg (408.73 KiB) Viewed 6347 times




The longest "unpumped" distance in this pattern is <hx-pipe-hx> and testing water throughput, it gave me 3600/s in both rows (good for 60 hx/row) (good for 3600/103.1 = 34.92 hx/row).
Last edited by Impatient on Tue Oct 26, 2021 10:56 pm, edited 1 time in total.
Top
mmmPI
Smart Inserter
Smart Inserter
Posts: 2728
Joined: Mon Jun 20, 2016 6:10 pm
Contact:
Contact mmmPI

Re: Contest: Longest train based nuclear reactor

Post by mmmPI »

mrvn wrote: ↑
Mon Oct 25, 2021 11:02 pm
 Can you suggest improvements to keep it at 50 fluid wagons or make it even longer? Fueling a larger reactor becomes a problem.
if you want the longest train : do not use 3 pump per wagon but just 1 and put the reactor on the side parallel to the track. load steam from only 1 side of the train.

with a setup like this you can get minimum 216 heat exchanger working at full load in a row. X2 because i tested only the propagation of heat on 1 side !!!! ( the temperature for 217th is still rising in the lab.)
181-2.png
181-2.png (253.84 KiB) Viewed 6331 times

Green is some non-fluid-wagon wasted space, so may as well put loco there, but you should also put loco on each side of the fluids wagons too !!

LLLLLLLWWWWWWWWLLLLWWWWWWWWWWWLLLLLLLL

you can put infinite amount of locomotive on each side, that's easy win since there can't be longer train than infinite number of locomotive !
With 2 pump per wagon and the setup shown above you could probably go with 400 wagons. But probably more with only 1 pump.

The more time you wait and the more fuel you are willing to burn the further the temperature will spread. probably til the 999-501=498th reactor, which is 2490 tiles, or 355 wagons, or 711 exchanger per side.

If we account for what's on left and on the right of the hot core, we total 710 fluid wagon being loaded simulateously each by 2 exchanger for an equivalent of 1420 MW.

For the test however i used a 2x28 reactor in the middle, which is wasteful.
181-1.png
181-1.png (417.54 KiB) Viewed 6331 times

a row of unfueled that doesn't qualify because it has 2 pumps per wagon. This means the heat is going away on the side and can only reach 181 reactor at the moment of the test, but it was still ramping up in the others , just very slowly.
( there is room for the waterhole and pump ).

nodowntime.png
nodowntime.png (519.35 KiB) Viewed 6331 times
To reduce train down time : build X number of parralel tracks 4 tile apart and park X number of train there. Then use pumps to make steam go through all of the train, wagon per wagon so that only the train in X position receive steam, when this X train is full it leaves the stations, and then the X -1 train will get filled immediatly with no down time apart from the internal buffer of the 2 pump per wagon transfering from X-1 to X.

When X-1 is getting loaded with steam, it is possible that X has enough time to come back, the loading will then transfer what was in X-1 to X, and there would have been 0 downtime since the internal buffer of steam is now in X and available again.


LLLLLLLWWWWWWWWLLLLWWWWWWWWWWWLLLLLLLL (X-2)
LLLLLLLWWWWWWWWLLLLWWWWWWWWWWWLLLLLLLL (X-1)
LLLLLLLWWWWWWWWLLLLWWWWWWWWWWWLLLLLLLL (X)


then it's just a game of cloning area in the editor to the infinite and waiting infinite amount of time and i ain't no more time for that :(
Top
User avatar
Impatient
Filter Inserter
Filter Inserter
Posts: 883
Joined: Sun Mar 20, 2016 2:51 am
Contact:
Contact Impatient

Re: Contest: Longest train based nuclear reactor

Post by Impatient »

mmmPI wrote: ↑
Tue Oct 26, 2021 8:27 pm
To reduce train down time : build X number of parralel tracks 4 tile apart and park X number of train there. Then use pumps to make steam go through all of the train, wagon per wagon so that only the train in X position receive steam, when this X train is full it leaves the stations, and then the X -1 train will get filled immediatly with no down time apart from the internal buffer of the 2 pump per wagon transfering from X-1 to X.
I just love that.
Top
mrvn
Smart Inserter
Smart Inserter
Posts: 5703
Joined: Mon Sep 05, 2016 9:10 am
Contact:
Contact mrvn

Re: Contest: Longest train based nuclear reactor

Post by mrvn »

mmmPI wrote: ↑
Tue Oct 26, 2021 8:27 pm
 you can put infinite amount of locomotive on each side, that's easy win since there can't be longer train than infinite number of locomotive !
A train with an infinite number of locomotives takes infinite time to fully leave and return. So your actual steam transported is 0 steam/s.

Anyway, I wouldn't count the locomotives aside as a deciding factor when the number of fluid wagons is the same.
mmmPI wrote: ↑
Tue Oct 26, 2021 8:27 pm
 nodowntime.png

To reduce train down time : build X number of parralel tracks 4 tile apart and park X number of train there. Then use pumps to make steam go through all of the train, wagon per wagon so that only the train in X position receive steam, when this X train is full it leaves the stations, and then the X -1 train will get filled immediatly with no down time apart from the internal buffer of the 2 pump per wagon transfering from X-1 to X.
That's just using the train as steam tanks. The first train will never leave so I would say that doesn't count. You could add some logic so a different train is used as buffer over time so every train leaves. But still not liking it. Pumping steam out of a fluid wagon feels like cheating in this case. It's not heat-exchanger - pump - wagon.

But good idea about putting locomotives in the middle and having heat exchangers spread out both side. You can mirror the setup top and bottom to have 2 trains. You can fill each train only when the other has enough of a head start so it can leave and return before this train is full.
Top
mrvn
Smart Inserter
Smart Inserter
Posts: 5703
Joined: Mon Sep 05, 2016 9:10 am
Contact:
Contact mrvn

Re: Contest: Longest train based nuclear reactor

Post by mrvn »

I think we can wave the need to offshore pumps and watering holes.

Every time you have a infinity pipe between two heat exchangers that leaves a 2x3 hole between the pumps connected to a fluid wagon. You can have water there. And there is one land tile left to place the offshore pump:
watering-hole.png
watering-hole.png (190.57 KiB) Viewed 6312 times
So feel free to use infinity pipes to showcase your ideas.
Top
mmmPI
Smart Inserter
Smart Inserter
Posts: 2728
Joined: Mon Jun 20, 2016 6:10 pm
Contact:
Contact mmmPI

Re: Contest: Longest train based nuclear reactor

Post by mmmPI »

mrvn wrote: ↑
Tue Oct 26, 2021 9:50 pm
 A train with an infinite number of locomotives takes infinite time to fully leave and return. So your actual steam transported is 0 steam/s.
Not if you make a double headed station and you put loco only on 1 of the 2 side alternating for each train left or right the one that doesn't have to be in the way.
Then the train come from one side and goes back the same side. Next train swoops in directly without any gaps from the other direction and alternate 1 2 1 2.
Maybe little 2 or 3 tick delay before the incoming train read the station opens on the other side so it doesn't collide. You don't need signals in this case. Or you can just spam signal every left tile available if the trains are supposed to move away with that steam.

How close do you think the last wagon from train 1 and 2 can stay when train 1 goes back and train 2 advance if you use no signal only train limit ? Distance can be 3 4 ticks maybe. And acceleration depend on the number of locomotives you can place behind. Especially if you have 400 fluid wagon, the amount of locomotive to balance out the deadweight and get close a good ratio of acceleration will be what 18x400 ? a train with 18 loco and 1 cargo has less acceleration than a train with 20 loco and 1 cargo, so may as well add more loco, and more loco, and more loco.

L>L>L>L>L>L>L>L><L<L<L<<L<L<L<L>LWWWWLWWWW ( where WWW is fluidWagon in hundred of unit, and L> and L< are locomotives facing each way, in millions of unit).

Ok they do not count at train size, but the contest is not clearly setting a goal, wether it's train lengh, steam/s, number of train, fuel efficiency , ( how's counted the penalty if we burn uranium 10x too many to increase the number of wagon by 2 ?) that's more you needing help figuring out stuff x)

switch.png
switch.png (219.5 KiB) Viewed 6284 times
mrvn wrote: ↑
Tue Oct 26, 2021 9:50 pm
 That's just using the train as steam tanks. The first train will never leave so I would say that doesn't count.
we are already using fluid wagons, isn't that the same as tank ? you said it was the same as a tank to me and really insisted upon that so i thought it would be ok to use the fluid wagon as a storage tank since the fluid wagon is storage tank in the sense that it can have fluid that flows in and out and can be used for buffer already exactly like the tank.
You even showed me some logic to wire the pump to activate either one or the other so that fluid can flow in and out of the fluid wagon to make it act like a storage tank, i'm a tad disappointed not to be allowed to use that cool trick and find it arbitrary.

mrvn wrote: ↑
Tue Oct 26, 2021 9:50 pm
 But good idea about putting locomotives in the middle and having heat exchangers spread out both side. You can mirror the setup top and bottom to have 2 trains. You can fill each train only when the other has enough of a head start so it can leave and return before this train is full.
if you mirror the setup top and bottom you will have 2x as many pump/heat-void per distance, the train will be smaller because the heat flow will be diminishing faster as distance from hot core increase i predict same as using 3=>2=>1 exchanger per wagon

But then if you mirror it again, and again and again and again, then the ratio wagon/reactor is getting closer to 1/1 again, the more you do it. instead of 2 train 1 array of reactor, that will cut the lengh of the train.

But then that makes a design infinitly long, also infinitly large, or the other way around and there's nowhere left to carry the steam.
Top
mrvn
Smart Inserter
Smart Inserter
Posts: 5703
Joined: Mon Sep 05, 2016 9:10 am
Contact:
Contact mrvn

Re: Contest: Longest train based nuclear reactor

Post by mrvn »

I had this long text all written up when factorio made my Xorg freeze. After killing it and recovering firefox the forum needed me to log in again and that lost the whole text I wrote. GRRRRRR. So a condensed version with just the highlights.

That train switch idea is golden. I might even use that for ore smelting.

Without signal you can make the train leave in the same tick. They will maintain a constant separation. How close together can you have the trains before they start? Having a stop right behind the other train I think gives 1 tile separation. Otherwise its something between 0 and 2. Any more than 2 and you could move the train stop closer.

Making the new train start moving before the old one leaves is also possible. They can start of with more distance and you would have to time it so they never collide. Minimum distance would be approaching 0 if you can time it right.

I was just about to write: You can force the old train out when the new train changes a signal. But no signals. So it has to be train stops doing all the timing. With the train coming from some distant outpost where the steam was used timing would be hard to get right. The train would be coming in full speed and braking for the stop while the other train accelerates away. You have to start the new train way before the old one is filled and for every outpost the timing would be different. Not practical.

But if you have a stacker before the reactor where all incoming trains stop. Then you can time it once and make the trains leave the stacker at the right tick so they never collide.

Making a train network without signals is a big challenge. You can trivially solve it by not having trains cross ever. The train limit can trivially prevent all collisions if trains only move in a circle station to station (never skip a station). Making trains leave in circuit conditions so they don't collide makes it a challenge. There is that one forum post of a smelter where ore comes in horizontal and plates leave vertical and the trains drive through the gaps between wagons. So much fun to watch.


Fluid wagons used to move steam from the reactor to some outpost that uses the steam are not tanks. Whatever gave you that idea? Fluid wagons that just stay in place and could be replaces by fluid wagon shaped tanks is what I call tanks. And having the fluid wagon move 6 rails between filling and emptying so it moves is not making it different. If the intention is to buffer steam then it is a tank. If the intention is to move steam to an outpost then it's a fluid wagon.
Top
mmmPI
Smart Inserter
Smart Inserter
Posts: 2728
Joined: Mon Jun 20, 2016 6:10 pm
Contact:
Contact mmmPI

Re: Contest: Longest train based nuclear reactor

Post by mmmPI »

mrvn wrote: ↑
Wed Oct 27, 2021 12:44 am
 Without signal you can make the train leave in the same tick. They will maintain a constant separation. How close together can you have the trains before they start? Having a stop right behind the other train I think gives 1 tile separation. Otherwise its something between 0 and 2. Any more than 2 and you could move the train stop closer.
you could have them almost collide and insta-break if you disable all their next station except one that says "no path" because it's on unconnected rail. that require timing. then disable that no path station and the train will skip it. and use limit for departure condition.
mrvn wrote: ↑
Wed Oct 27, 2021 12:44 am
 But if you have a stacker before the reactor where all incoming trains stop. Then you can time it once and make the trains leave the stacker at the right tick so they never collide.
you can't really make stacker without signals, trains will path through the same spot at least entrance and/or exit even if you name each lane different .

If the amount of time they spend "outside" is not the same or a multiple of the same value collision will happens. ( 1 train need the path at tick 100 200 300 400, other train need the path at tick 75 150 225 300, bim collision.)
mrvn wrote: ↑
Wed Oct 27, 2021 12:44 am
 I was just about to write: You can force the old train out when the new train changes a signal. But no signals. So it has to be train stops doing all the timing. With the train coming from some distant outpost where the steam was used timing would be hard to get right. The train would be coming in full speed and braking for the stop while the other train accelerates away. You have to start the new train way before the old one is filled and for every outpost the timing would be different. Not practical.
Those doesn't have to be the same train, you can have train with unknown time travel parks at definite spot which triggers other train on a different non-signal network to park parralel for transfer of material. one system fully controlled isolated from the other one :)
mrvn wrote: ↑
Wed Oct 27, 2021 12:44 am
 Fluid wagons used to move steam from the reactor to some outpost that uses the steam are not tanks. Whatever gave you that idea? Fluid wagons that just stay in place and could be replaces by fluid wagon shaped tanks is what I call tanks. And having the fluid wagon move 6 rails between filling and emptying so it moves is not making it different. If the intention is to buffer steam then it is a tank. If the intention is to move steam to an outpost then it's a fluid wagon.
But fluid wagons have only 3 connections contrary to tanks, so if you use 2 pumps per wagon to add steam then that wagon can only use 1 pump to transfer to the next wagon,so eventually it will fill up and have to move.

Plus it's also possible that the last train leaves, and before it comes back the X-1 leave, and before X-1 come back, X-2 leave, and now X comes back, but there's a gap and nothing is transfered from X-3 to X and the wave continue until reaching train X that then starts moving !

moving 6 rail is not enough to qualify as fluid wagon what's the minimum distance the outpost need to be? :D

And what if my intention is to move steam to an outpost to buffer it in the outpost ? Does the fluid wagon becomes a tank the moment it arrive at the oupost ? or does the fluid wagon is already considered a tank the moment he leaves the heat exchanger because his intention is clear from this moment ?

i can't tell the intention of my storage tanks, what if they want to move but can't because they have no wheel ? are they still storage tank ? or should i treat them like fluid wagon ? maybe it's a line of 2000 tank that i use to transport steam to an oupost, where i use a fluid wagon as a moving buffer to distribute that steam equally between machine using the train timer condition ? who's the tank ? who's the buffer ?
Top
mrvn
Smart Inserter
Smart Inserter
Posts: 5703
Joined: Mon Sep 05, 2016 9:10 am
Contact:
Contact mrvn

Re: Contest: Longest train based nuclear reactor

Post by mrvn »

mmmPI wrote: ↑
Wed Oct 27, 2021 1:44 am
 you can't really make stacker without signals, trains will path through the same spot at least entrance and/or exit even if you name each lane different .
Have one train stop at the entrance with train limit 1 called entry. So only one train can arrive at a time. Then have one train stop per lane with train limit 1 all the same name exit. Each train will pick one lane.

You have to build your intersections with train stops if you want more than one train moving. Not terrible efficient or at least not easy to make efficient but that's why there are signals.
mmmPI wrote: ↑
Wed Oct 27, 2021 1:44 am
mrvn wrote: ↑
Wed Oct 27, 2021 12:44 am
 Fluid wagons used to move steam from the reactor to some outpost that uses the steam are not tanks. Whatever gave you that idea? Fluid wagons that just stay in place and could be replaces by fluid wagon shaped tanks is what I call tanks. And having the fluid wagon move 6 rails between filling and emptying so it moves is not making it different. If the intention is to buffer steam then it is a tank. If the intention is to move steam to an outpost then it's a fluid wagon.
But fluid wagons have only 3 connections contrary to tanks, so if you use 2 pumps per wagon to add steam then that wagon can only use 1 pump to transfer to the next wagon,so eventually it will fill up and have to move.
Then you are dual using it. Part is buffer, part is transport.

But when you fill the train from 2 heat exchangers (206.x steam/s) and empty it with a pump (1200 steam/s) it's not going to fill up.
mmmPI wrote: ↑
Wed Oct 27, 2021 1:44 am
 Plus it's also possible that the last train leaves, and before it comes back the X-1 leave, and before X-1 come back, X-2 leave, and now X comes back, but there's a gap and nothing is transfered from X-3 to X and the wave continue until reaching train X that then starts moving !

moving 6 rail is not enough to qualify as fluid wagon what's the minimum distance the outpost need to be? :D

And what if my intention is to move steam to an outpost to buffer it in the outpost ? Does the fluid wagon becomes a tank the moment it arrive at the oupost ? or does the fluid wagon is already considered a tank the moment he leaves the heat exchanger because his intention is clear from this moment ?

i can't tell the intention of my storage tanks, what if they want to move but can't because they have no wheel ? are they still storage tank ? or should i treat them like fluid wagon ? maybe it's a line of 2000 tank that i use to transport steam to an oupost, where i use a fluid wagon as a moving buffer to distribute that steam equally between machine using the train timer condition ? who's the tank ? who's the buffer ?
Just more ways to nit-pick.

The situation was a fluid wagon you fill while a train unloads so you can empty it while the train switches for the next one. Intention: buffer. can't be any clearer.
Top
Locked

Return to “Energy Production”