My longest train filled with steam from heat exchanger (262 wagons) + math explanations

Clever and beautiful constructions, bigger than two chunks
- Defense: killing biters as an art
- Castles, Throne Rooms, Decorations (comfortable living in the Factorio World)
- Main Bus Concepts
- Modular Systems, Factory Streets, show how all works together
- Megabases
Please provide us with blueprints or saves, if that makes sense of course.
Forum rules
Clever and beautiful constructions, bigger than two chunks
Smart Inserter
Smart Inserter
Posts: 1601
Joined: Mon Jun 20, 2016 6:10 pm

My longest train filled with steam from heat exchanger (262 wagons) + math explanations

Post by mmmPI »

Hello !

Why would you do such things ?
that wasn't my idea in the first place to make the longest, i don't know why people have those idea, i just like solving math problem in practice and also in theory.

Constraint necessary :

1)train must be filled from heat exchanger directly into the fluid wagon and nothing else.

2)train must be the longest.

Other constraint subsidiary :

1) Each used fuel cell used must yield at least 24 GJ.

2) All reactor used at least once must be adjacent

3) Each wagon must be filled at the same time and at the same rate


1) no space and time constraint

2) infinity pipe for water

3) infinity chest for train fuel and reactor fuel

Due to the size i'm not going to try and post a full picture or it would just look like a long and thin horizontal line, the map is available at the end. This is the centerpart :

slice.png (87.94 KiB) Viewed 1043 times
It's composed of a 2x18 core in the middle.
center.png (229.72 KiB) Viewed 1043 times
And this pattern on the 2 sides :
side.png (336 KiB) Viewed 1043 times
The use of 3 rows or reactors ( 555 on each side) is made to abuse their heat-propagating-property, they are 5 tiles wide compared to 1 for a heat pipe. So you only need a 1°C differential per 5 tiles, and not per tile. It is also used because those reactor can buffer heat, at least 1/2 fuel cycle worth of heat at the scale of the whole design. The train triggering the refuel when arriving at a station, the other 1/2 is carried away by the train under the form of heat while the fuel is still burning.

This is the longest train filled with steam i have ever seen. It can contain up to 6 550 000 steam in its 262 wagons !

But instead of filling it entirely it is only filled with 5 773 100 +roundings. This is because the heat comes from a 2x18 reactor in the middle. Meaning a refuel represent ([160(2x18)-160]/0.097)X200 = 11 566 391 steam. This makes it easy to trigger the refuel 1/2 of time when a train arrive at the station.

This is how it looks :
refuel logic.png
refuel logic.png (274.94 KiB) Viewed 1043 times
I couldn't get fueling inserters to work on a 1 pulse signal from this inserter loop. I don't know why. Instead i made the signal last longer by reading a fish on a belt every 2 trains. And it worked.

If you are precise, you will realise that the refuel is a little too frequent, there is roughly 191 extra steam that are produced compared to what is removed. This means the reactor is potentially not cooling back down to the same temperature after a refuel and 2 train load but instead it overheat by 0.001% which is some fuel efficiency lost which i utterly neglected for the reasons below:

With a 2x18 reactor, you produce {([160(2x18)-160]/0.097)X200}/36= 320 733 steam/fuel cell, if you take a smaller reactor like a 2x16 You produce only {([160(2x16)-160]/0.097)X200}/32}= 319 587 steam/fuel cell which is only 99.64% of the 2x18. So yes there is some calculated loss on the 2x18 i'm showing you, yet it's still more efficient in steam/fuel cell than a shorter reactor like a 2X16 even if the 2x16 is "wasteless".

Of course you'd be more efficient with a 2x20 or a 2x40 reactor, there is no limit to this, so you can't really max out efficiency in steam/fuel this is not a valid metric to try and maximize, 24GJ is a lower bound to me, if the reactor overheat, then some energy is lost; it must not represent a % significant enough to alter the fuel efficiency and make it drop under 24 GJ per cell, which is what you get when you only have 2 neighbour bonus in a 2x2 power plant for example.

If you count the size of train as the metric to compare design, then this one is the longest i've ever seen, but as i'll explain, size of train is not a good metric either.

Now if you count something like steam/s as a deciding factor in case another design would also be 262 wagon long and you don't know which one is the best. Then this design is not very good because the heat exchanger are not working 100% of time due to the train being very slow. But it would require another 262 wagon train to compare.

Why is the train so slow you may ask ? i used a fluid wagon in front of the train which is bad for acceleration, and i fuel the train with solid fuel, while i could have used nuclear fuel.

Yes for this i will use the analogy of a honey dispenser for cake topping.

What ??

i know it's not a straightforward comparaison, but bear with me i'm adding pictures.

honey disp1.png
honey disp1.png (8.02 KiB) Viewed 1043 times
This represent the honey dispenser, under it are some cakes. The black box is the reservoir of honey, and the yellow rectangle is a refuel of a big block of viscous honey. It's still a fluid, but not like water, very fluid, instead it's a slow moving fluid exactly like heat in the game.

From the little vignette on the right side which is the thing seen from above, you should understand that if you put too much honey at one time, above the red line, it will spill over the side of the dispenser, and not the side you want it to spread. This honey is lost for the cakes and will make the plant messy. That is the same as if you put fuel in the reactor, and the reactor overheat before spreading the heat. The heat/honey is lost.

What you want instead is open the many littles holes under the dispenser by removing the green lock so that the honey is dropping to top every cakes.
honey disp2.png
honey disp2.png (4.49 KiB) Viewed 1043 times
Over time the honey, like the heat, will spread and start to fill the whole bottom of the honey dispenser( the whole array of reactor).As long as you didn't put too much in the first place, then you just have to wait for the honey level to stabilize and there is no loss.
honey disp3.png
honey disp3.png (5.08 KiB) Viewed 1043 times
When the honey is stabilized, then you can open the the valve ( green line removed )and the flow will be the same on every little cake. When the machine is initialized, then you can start filling wagon, they will all receive the same amount of steam/s, and if you let this situation for a certain amount of time and then close the valve again, you can make sure all cake received the same amout of honey/ the wagons received the same amount of steam.
honey disp4.png
honey disp4.png (5.07 KiB) Viewed 1043 times
Ofc in the game heat is not honey, it will more look like this where the middle is hotter than the outer part.

Now time to think of it a little.

1) there is reason to be willing to have all cake filled with the same flow, otherwise some have more and some other less. It's what i did to show it was possible to reach even 262 wagons without using the most rationnal method to make the train the longest. But the wagons have a definite capacity, unlike the cakes, this means if some wagons start getting filled before/faster than the others, the only thing that will change is the amount of time required for the train to be filled. For a train, it would mean the inner wagon would fiil in faster than the outer wagon. Thus you would need to wait more time for the outer wagon to be filled 100%. In other word steam/second will decrease as a cost to increase train lengh.

2) The more you wait before 2 refuel/ the more you wait between 2 honey refill = the more time you give the dispenser to stabilize/ the more time you give heat to spread everywhere, and especially to the furthest part :).

This means, if you lower your expectation of steam/s you can make way longer train. Just make it slow enough so that heat has time to propagate this means it's easy to make a train longer than another one at the cost of steam/s. Which again doesn't matter in the slightliest if the aim is to make the longer train what could a reduction in steam/s do? it's not the main metric.

This is why the train is fueled with solid fuel, has only 13 locomotives for 262 wagons, and has a fluid wagon in front

It is calculated slowness to give time to the heat to spread further than it would if the train was faster, this gives time to center part for cooling. Like honey's level is flattening with time.

Also using nuclear fuel for a train would reduce efficiency on uranium, compared to using oil products as oil is infinite.

But i want more steam/s !!!

That's not what this build is about, this is a different metric than making the longest train, if you want more steam/s you can always double this design.

Think about it for the cake machines. The honey dispenser is this size because the boxes in which cakes are put to be sold are this size. Making a 10% smaller dispenser that load 50% faster is not the requirement, what good would partly filled boxes do to you ? you can only sell them when the number written on the box is the same as the number inside the box. Customers won't care how fast they were put in boxes. They only care about buying the biggest box of cake they can find.

If you build 2 of them train size doesn't change mind you :) , you just get 2x as much steam/second. The only thing that changes is the amount of rolling stock used. If you wanted to have a sensible metric to measure what is happening when you build another one of those longest reactor then you would need to look at Steam/s/wagon. But not the number of wagon of 1 train, the total number of wagon used. What happen then, you have 2x steam/second and also 2x wagon, so the steam/second/wagon doesn't change. This means Steam/s/wagon is a good metric to evaluate train-based design. It mesure the efficiency of a wagon in the task of carrying steam which is more useful than having the longest train for most people.

Side note : the reactor could sometimes reach 1000° or close, not all of the reactor, only the hottest 2 in the middle reaches 999.9° for a few tick only, this is due to the 0.001% overheating, this means every refuel the reactor doesn't actually overheat, instead it waste the excess 0.001%, negating the drift towards upper ° on a design that also has no risk of overcooling this means it's pretty stable :) and still more efficient than what you could get with a smaller reactor. The same way using a 10x2 power plant only 10% of time is more efficient than producing the same amount of electricity with a smaller plant running higher % of time.

The honey dispenser would require someone to clean around it due to some droplets of honey falling around the dispenser :).

Conclusion :

production.png (19.98 KiB) Viewed 1043 times

Maximizing fuel efficiency means using larger than needed reactor, less often than needed if they were smaller. Meaning the slow train not only 1) allow for a longer train 2) allow for better fuel efficency. You can add as many locomotives as you want in front of a train to increase its acceleration, and it makes it trending to infinity usually but in this kind of design you don't need to. It would help maximizing the amount of steam carried per rail. Or the % of time the reactor is allowed to burn fuel. Or the steam/s carried away. But those are not the train lengh :).
(3.29 MiB) Downloaded 213 times

Global Moderator
Global Moderator
Posts: 6756
Joined: Fri May 30, 2014 8:54 am

Re: My longest train filled with steam from heat exchanger (262 wagons) + math explanations

Post by Koub »

[Koub] OK enough is enough. Every single thread where mmmPI and mrvn will restart the endless fight that pollutes the forum will be cleaned past the first post and locked.
I have split the argument that was starting here and sent the posts to the trash area. And I'm locking the thread.
Koub - Please consider English is not my native language.


Return to “Medium/Big/Gigantic Sized Structures”