First attempt on a megabase - train design

[Premu]

After finishing several games with different patches I finally wanted to go build a mega base with a science production of 1000 per minute.

I've started with a modest factory to get all the non-repeatable research and acts as mall. Now the research is done, and I want to prepare everything for going big.

This is my current base:



For now my powersupply is done by a large solar array good for 110 MW, and the old factory is completely belt-based with a central bus. I've cleared out and secured a large area north and west of my base for grabbing those resources and getting building space. I might enlarge that area again if necessary. I've plenty of U-235 available, so I'll also be able to switch to nuclear power on a larger scale.

My overall strategy is to build several belt-based subfactories like Iron works, electronics, oilproducts, specific science packs, etc. and connect those with trains. Each subfactory might need several inputs and provide several outputs. So I must be able to handle this.

I've created an excel table which calaculates how many machines, miners, etc. I need taking into account speed bonuses and productivity. While working out an easily scalable design for production chains isn't that hard, the logistics cause headaches for me.

For example, I plan to use miners with one productivity and two speed modules. With that (and using productivty modules everywhere possible in the chain) I'll need almost 600 miners for both copper and iron. There are lots of small and big ore patches around, but I have somehow to collect all of them, and bring them to central smelting areas. Now these smelting areas would have 24 rows each which can fill up a blue belt fully with plates. This means I'll need train stations with a complexity I've never used before.

The first question I was pondering about was how large should I make the trains. First I did some math. Apparantely, three staple inserters with their maximum capacity are capable to fill up a blue belt completely. Of course, if I start to load and unload from trains I'll have times in which no trains are present, so filling or emptying one belt with just three such inserters will not be sufficient. I think using six for a belt is a good compromise - which happens to fit perfectly to a side if a train wagon. This means that such a train station needs to be actually unloading 50% of the time at least. I'll assume that this is managable.

Now, I use six inserters to fill up a single belt, and I use six wagons per train, two unloading/loading stations should suffice to serve a smeltery. So my first plan is to use two locomotives and six cargo wagons for those smelteries with two unloading stations for each. For those with more experience - is that a feasiable plan or are my assumptions too optimistic?

The next issue are those trains coming and going to factories with different goods. For example I want to process the iron plates directly into steel in my iron works. I might also produce gears and engine units directly as those don't need any thing else than iron and steel as input. Now my iron works would have four different products to offer. Many other factories will require some of those, but most likely not all of them. And if the requesters need steel, iron plates and gears, they'll need it in different ratios each.

So, should I use different trains for each product? This would lead to lots of train stations and lots of additional trains. Or should I use specially designed trains for each requester, bringing in all needed goods in an appropriate ratio? The latter will make loading and unloading far more complicated, as these ratios will most likely not fit to one wagon for each product.

[coppercoil]

There’s no “right” answer for it. There are many different ways to make the same things. You need to think which approach looks more interesting for you. Do you like to optimize things and make complex designs? Do you like uniform setups regardless of some redundancy? Do you like 1 belt per wagon side? Do you like fast trains or long slow trains? Or maybe, long fast trains? Every way is possible. Every way has its own challenges.

[Amarula]

The first question I was pondering about was how large should I make the trains.
...
So my first plan is to use two locomotives and six cargo wagons for those smelteries with two unloading stations for each. For those with more experience - is that a feasiable plan or are my assumptions too optimistic?

I am using L4CL (one locomotive, four cargo wagons, one locomotive) with two unloading stations for ore delivery and it is working okay for me. L6CLL should be feasible but you need to consider how many intersections/crossings, and how much traffic. The longer the train, the heavier it is, and the more time it needs to speed up and slow down. There are several posts here that go into train length and intersection design. If I recall, there was a recommendation to go to L4CLL4CL, for 8 wagon capacity...

So, should I use different trains for each product?

I tried many variations of mixing products on one train, and having one train serve more than one customer, and it was a nightmare: constantly my customers were running out of product because the train was waiting somewhere else. So now I use one train for one product for one customer. I do have a very few exceptions: the acid train serves all my uranium mines; the train fuel train delivers fuel for all my trains; and my personal acme train delivers a variety of building materials to me on demand. Other than that, one train, one product, one customer. As coppercoil said, there is no right way, so I am not saying this is what you have to do, only that this is what ended up working for me.

Also to consider is how you are going to provide fuel for your trains, and what type of fuel. I am using nuclear fuel, the extra speed really makes a difference in timely delivery, and the extra fuel life means a single fuel train can keep up with all my trains.

I've created an excel table which calaculates how many ...

I <3 excel! I have several workbooks in use for planning my base and expansions. I think you are off to a great start.

[mmmPI]

The first question I was pondering about was how large should I make the trains. First I did some math. Apparantely, three staple inserters with their maximum capacity are capable to fill up a blue belt completely. Of course, if I start to load and unload from trains I'll have times in which no trains are present, so filling or emptying one belt with just three such inserters will not be sufficient. I think using six for a belt is a good compromise - which happens to fit perfectly to a side if a train wagon. This means that such a train station needs to be actually unloading 50% of the time at least. I'll assume that this is managable.

I share the reasonning for maintaing throuput of ressources, with some little tricks you could even increase that 50% margin of time. Though my personnal opinion would be that train lengh require other thing to take in account before deciding such as the other points you mentionned later and some others I will try to mention.

doing : cargo->inserter1->chest->inserter2->belt is one trick to increase the 50% margin. Inserter1 in this case has higher capacity than Inserter2. But it forces you to balance out not only the wagon quantity, but also the chest. ( using circuits or splitters if ratios are not perfect ).

Now, I use six inserters to fill up a single belt, and I use six wagons per train, two unloading/loading stations should suffice to serve a smeltery. So my first plan is to use two locomotives and six cargo wagons for those smelteries with two unloading stations for each. For those with more experience - is that a feasiable plan or are my assumptions too optimistic?

On theory, assuming no mistake in your math for inserter capacity, you still may face uneven consumption of goods in all the wagon. Depending on how good your ratios of consumptions are and your balancing between lanes. This could then cause some trains to unload slower than theory and cascade into bigger trouble if one of the wagon isn't unloaded at all because of back up goods, while the other are empty.

Here i want to bring your attention on train schedule conditions. Those are quite significant parameter that shape how your network function ( IMO ).

In the example it could force train to move after inactivity time, while the regular schedule would use empty/full condition from load to unload. This is not a precise/elegant solution to the problem but something to keep in mind.

The next issue are those trains coming and going to factories with different goods. For example I want to process the iron plates directly into steel in my iron works. I might also produce gears and engine units directly as those don't need any thing else than iron and steel as input. Now my iron works would have four different products to offer. Many other factories will require some of those, but most likely not all of them. And if the requesters need steel, iron plates and gears, they'll need it in different ratios each.

So, should I use different trains for each product? This would lead to lots of train stations and lots of additional trains. Or should I use specially designed trains for each requester, bringing in all needed goods in an appropriate ratio? The latter will make loading and unloading far more complicated, as these ratios will most likely not fit to one wagon for each product.

I would start by this, train lengh and train schedule being for me consequences of those choices.

( ofc you can force yourself using only plenty of tiny trains , and design the rest of the factory around it , decide 1 trains is 1 good , all train same size, there are many way of thinking about the problem and having fun solving it)

Those are somewhat arbitrary choices, or as you expressed it trade-offs between complexity here or there. At this point you can consider your "naming convention" for the station. This is a thing with "train lengh" and "train schedule" that i would consider altogether for the train network. ( those thing i look at when i join multiplayer game to "understand" the network people made ).

By this i mean some networks function with station with identical names, some with each station having unique name, some a mix of both. And you can also have them open/closing depending on the quantity of material in chest.

On large base i often have several station called [Acid Unload] i connect them to a tank, and i open them when the tank is under 5K, trains leaves when above 20K, train leaves [Acid Load] when full. This make sure i have some sulfuric acid in all uranium mining area using only 1 train that don't even move most of the time and it keeps schedule conditions fairly simple.

Some consideration of item density and richness of map ressources are relevant too;

For convenience , i prefer smelting iron to steel and only move the steel with trains, for this i would use small blueprint of furnaces and place them near each miners. I don't precisely math the time the outpost will run, but roughly i don't bother when it's not relevant compare to game lengh/objectives. The main smelting area is another option that doesn't force you to relocate as time goes, but ressoources will be further and further away from this main area as closer patches deplete, which means you will need more and more trains to carry them in time, which means the main smelting area would eventually need scaling and so on ...

( one could also make green circuits in between an iron and a copper patch using the furnaces array to save on trains like near your southern wall )

Deciding what is being carried from A to B ( which material you want to move around ) ; then what is point A and what is point B ( where you place your assembly/furnace on the map ) ; enventually we come to HOW ? we move those.

The rail layout themselves, with junctions, and signals, and stations, train size and schedule, naming convention ... all those from my point of view comes as technical answer once you have a plan for the network.

Signaling rails is a topic on its own, making a loading station , making an unloading station , those you can find some great guides and creations on the internet to compare with your experiment. From very simple to very complex, i guess given how much time you are enjoying testing those you develop preferences

Choosing the train lengh and the number of trains is a parameter that is more related to your own map and your own objectives as i see it. It is hard to say if it's optimistic or not 1000 science per minute is quite an objective and we can't tell how much iron those 3 patches inside the left wall represent. If that is enough for your goal or if that is 1/10th of the requirement, or if you want 1000 science per minute for 10 minutes , 10 hours , or 100 hours would also play big role in the design.

For practical testing purposes, making 5 X 200 science, could be a way to test design on 1 module adjusting before it takes too long to fix the problems you discover along the way and would also allow to spread train traffic accross the map to avoid traffic jam.

Different train for each product is the 'easier' way to make it work, that require the same logic duplicated everywhere, sharing goods in same train but different wagon or even in the same wagon, requires in-depth planning to avoid error. But you can mix both depending on what thing you move around !

[Premu]

Thanks for the hints!

I now also cleaned up the other side of my base, and now I think the next best step to set up a basic rail network first with a few central lines. After that I'll experiment with some prototypes of stations. If those prototypes work, I can apply them and start to build the actual factory. For start I might just build up for 250 science per minute, but keep the space reserved for the further extension. That way I can see faster if there are problems in my designs which I need to address.

Because one thing is clear - no matter how well you plan something out, somewhere something goes wrong.