Oil for 1 rocket per minute

[Xarovin]

I'm trying to plan out how much stuff I need for 1 rocket per minute, and I'm just wondering if anyone has already worked it out for oil and can compare their results to what I have?

Here's the totals I'll apparently need:
All heavy oil -> light oil
10500 light oil -> solid fuel
remaining light oil -> petroleum gas
29850 petroleum gas -> plastic bars
2325 petroleum gas -> sulfur

These are the results I have for breaking even:
Crude oil: 43680/m or 728/s
Oil Refineries: 364
Heavy Oil Cracking: 91
Light Oil Cracking: 344

For the record, here's the code I used, it isn't perfect but I think it works alright in this case:

Code: Select all

from __future__ import division
import math
def calculate_oil_needed(petrol, light, heavy):
    num_refineries = 0
    remaining_petrol = -1
    while remaining_petrol < 0:
        num_refineries += 1
        heavy_processing = int(math.ceil((12 * num_refineries - heavy) / 48))
        light_processing = int(math.ceil((54 * num_refineries - light + 36 * heavy_processing) / 36))
        remaining_petrol = 66 * num_refineries - petrol + 24 * light_processing
    return {'Refineries': num_refineries, 'Crude Oil': num_refineries * 120, 'Heavy Oil Cracking': heavy_processing, 'Light Oil Cracking': light_processing}
print calculate_oil_needed(29850 + 2325, 10500, 0)

[DaveMcW]

Looks about right.

But the smart kids use productivity modules and speed beacons, which makes the numbers smaller and the math harder.

[siggboy]

Here are some really good optimized blueprints for turning oil into PG and Solid Fuel, respectively: https://imgur.com/a/k55lw

Just use as many as you need for the 1 rocket/minute.

(These are not my designs, but I think they're really good.)

Design #1 for Oil->PG (the most efficient one in my opinion) has a bug, he forgot to connect one of the chemical plants (the input for the light oil). It's easily fixed by adding the missing pipe.

You also wont have problems with balancing the intermediate products (light, heavy, PG), just blueprint the respective amounts of PG and SF mini-factories and put all the oil into them that you can get. Remember to use pumps (Dave has posted some nice designs for parallel pumps) so you get the necessary pressure on the input pipes for the oil. These high-speed refineries suck up a lot of oil per second.

[Xarovin]

Ah thanks, I was going to say I'd try wing it for the PG production, but after seeing the designs, I don't think I could do better haha. Is it easy to make a logic network that'll pump something from one tank to another should the first tank get full? I've never looked into it so I don't know what they're capable of

Edit: Is there a way to test out the design through cheating? It appears a pumpjack can't extract more than 10 oil per second. I guess I should assume each pipe can hold just over 100 fluid per second to be on the safe side? Also I now know what parallel pumps are, in long distance pipes, how often should I ideally be placing them to keep the speed up?

[siggboy]

You should not even use tanks (they're in the blueprint but you should remove them). Maybe add one tank so you can see if the PG is flowing and maybe measure throughput with a combinator contraption.

Tanks also limit throughput (they're basically pipes with a very large size, but in Factorio smaller pipes have higher throughput, even though that makes no sense at all).

I think it's a bad idea to create a buffer for no reason like in the blueprint. It's not going to produce rockets faster if you have, like 100k PG in the buffers at all times. It's simply wasted product waiting to be consumed later.

You cannot move tanks without losing the contents. If you have a large amount of fluids in tanks, and then you realize that you need the space for expansion, or if you simply want to move the setup somewhere else, you will lose the fluids, or you need to go to great lengths to pump it all somewhere else.

(The same is true for ore buffers, plate buffers, etc. I try to avoid them as much as possible, I even limit the input to my furnaces to avoid their internal buffers filling up.)

If your buffers are low at all times, then it's also easier for you to test your factory under full load. You will get realistic numbers very soon, without having to wait for the buffers to empty. Everybody can produce a "fake rocket per minute" for a while when the buffers are full -- it doesn't mean anything.

Simply consume the PG that is produced right away. Most of it will go into plastic bars, and a smaller amount will go into sulfur. I don't think you should use any of it for solid fuel, your solid fuel factories already produce that more efficiently.

Regarding the pumps, here's the relevant thread: viewtopic.php?t=6066

If you want to test your design, you should use Creative Mode and TestMode (and maybe ToyBox). Flatland is good for making sandbox maps (but you can also make them yourself in the map editor and then use TestMode to remove the doodads and the trees). Don't make the sandbox too big (400x400 is more than enough for most purposes) because it will create large save files and saves will be slow.

[Xarovin]

Thanks, so you'd recommend building a parallel pump every 4 spaces then, or is that speed a bit overkill?
The petrol to solid fuel was just an idea to avoid overflow, but I guess if I have separate refineries for each thing then it's not a problem. As to plate buffers, I made a design that seems to work quite well, I found you can eat a fully loaded belt with 6 stack inserters, and create a(n almost) fully loaded one again with 3 http://i.imgur.com/INqbhIH.jpg
I've only tested that with the steel production, but each section (contained within the walls) makes about 4.5k plates/900 steel per minute, so hopefully it should also work for all the other iron and copper parts.

I have a sandbox world I'm doing the testing on, though it's infinite and the power requirement is now insane haha, 23000 solar panels isn't enough, and construction bots are a pain to use over long distances if I try build any more. I wasn't aware of testmode (to get a fully loaded belt I'd fill up each chest using cheats, which took ages) so I'll get that downloaded asap. Just a quick question so I can jump right in, if I shouldn't use tanks at all, how do I create the required oil flow using testmode (since it's not like I can just fill a tank and just pump it out)?

[siggboy]

Thanks, so you'd recommend building a parallel pump every 4 spaces then, or is that speed a bit overkill?

Maybe you should spread the oil over more pipelines so you don't have to use pumps quite as often. Also keep in mind that 1 underground pipe counts as 1 pipe, and the connection can run over 11 tiles. So with 2 pipes (1 enter 1 exit) you can cover 11 tiles of distance; that really helps with throughput.

The petrol to solid fuel was just an idea to avoid overflow, but I guess if I have separate refineries for each thing then it's not a problem. As to plate buffers, I made a design that seems to work quite well, I found you can eat a fully loaded belt with 6 stack inserters, and create a(n almost) fully loaded one again with 3 http://i.imgur.com/INqbhIH.jpg
I've only tested that with the steel production, but each section (contained within the walls) makes about 4.5k plates/900 steel per minute, so hopefully it should also work for all the other iron and copper parts.

That base looks really neat, but I think you're using too many beacons (each of the double row of beacons basically wastes half of the beacon capacity. Also maybe use productivity modules in the furnaces instead of speed modules. That's more efficient (the speed beacons amplify the effect of the productivity modules, so you get the speed module for free, effectively. The same principle applies to assembly machines and oil production (those blueprints for the oil use prod 3 everywhere + speed beacons).

I have a sandbox world I'm doing the testing on, though it's infinite and the power requirement is now insane haha, 23000 solar panels isn't enough, and construction bots are a pain to use over long distances if I try build any more. I wasn't aware of testmode (to get a fully loaded belt I'd fill up each chest using cheats, which took ages) so I'll get that downloaded asap. Just a quick question so I can jump right in, if I shouldn't use tanks at all, how do I create the required oil flow using testmode (since it's not like I can just fill a tank and just pump it out)?

Yeah that's crazy, these tests are much easier with cheating mods.

If you need infinite fluids and fluids sinks you have to use Creative Mode. TestMode has no support for fluids. You can use both at the same time.

For a real game you should look into Nucular for the power production. It's a lot more interesting to build than solar farms, but it's not easy mode. So there's a good balance of power output and fun. Nucular can easily support multi-gigawatt factories, if you use the correct ratio of breeder reactors to fission reactors you don't need a lot of Uranium either.

[Xarovin]

Oh right thanks, though I can't use both mods at the same time as it complains there's some GUI problem (and creative mode crashes on my sandbox world), passed on the bug though so maybe it'll get fixed
Spreading the oil over multiple pipes sounds like a really good idea, but more maths, gonna have to figure out what I'd need for a fully loaded pipe at the end, thanks for the tip though

Anyway, I know I could cut down on the beacons a little bit, but I like everything being as dense as possible haha. I'd also totally forgot about using productivity, but I'll get them added in
Nuclear sounds really cool, but I also want my achievements in the main game so I'll survive with solar power

[siggboy]

Oh right thanks, though I can't use both mods at the same time as it complains there's some GUI problem (and creative mode crashes on my sandbox world), passed on the bug though so maybe it'll get fixed

I leave Creative Mode disabled but TestMode enabled, and then you can give yourself the items from Creative Mode with the TestMode item generator.

So for example you can use the TestMode "tool box" to give you a void chest (which is an item from Creative Mode).

That way you can have both at the same time.

Spreading the oil over multiple pipes sounds like a really good idea, but more maths, gonna have to figure out what I'd need for a fully loaded pipe at the end, thanks for the tip though

The image gallery with the blueprints mentions how much oil you need per second per blueprint. You also don't have to spread it exactly, you just need to make sure that you don't need to have pumps every 4 tiles.
You could just make enough parallel lines so they won't carry more than 90 oil/second. If your oil comes out of assemblers that are emptying barrels, then you can easily create pipelines that carry just enough oil but not too much.

Anyway, I know I could cut down on the beacons a little bit, but I like everything being as dense as possible haha. I'd also totally forgot about using productivity, but I'll get them added in

Beacons draw a lot of power, and you need 2 tier 3 modules in each of them, that need to be crafted. I think it's worth to min-max a little here and don't use beacons that are not efficiently placed.

Prod3 is pretty much mandatory with speed beacons in ALL cases where it's possible, because the combination is extremely potent. The energy usage is very high, though, but since you have solar you have free energy anyway, so it should not be a concern.

[Xarovin]

So it turns out when using test mode to make creative mode items, the fluid spawner doesn't get any options. However the bug about crashing on my sandbox world is fixed now, and it all seems to work fine

Sometime soon I'll get round to attempting to build the full oil production then, I'll follow the designs on the images, but possibly reply back here if something is going very wrong haha. I was thinking about avoiding assemblers with barrels since I've never done them before, but looking at the stats they might actually work really well. Looks like I'd need around 30 non module'd assemblers and 30 barrels per second for the entire oil needs, which doesn't sound too bad.

With your productivity module suggestion btw, it's resulted in some of the ore piling up which I was trying really hard to avoid lol, I may have to redesign it with more furnaces now

[siggboy]

I've used the designs on the images, they work fine (remember to get the blueprint strings, don't build it all by hand, the links to the blueprint strings are in that imgur album). If you use design #1 (for PG), then you need to remember to place that missing pipe as mentioned, or it won't perform 100%.

Oil in barrels is a good way to store and ship oil, even if you have a Rail Tanker, because 400 barrels of Oil (capacity of 1 wagon in 0.13) is way more than 1 Rail Tanker can load. Loading and unloading is also faster.

However, barrel handling is challenging in the trains. I've got a complex train scheduler, so that base is covered, but it's certainly more complex than using a tanker. When you design the oil shipment, you need to think about getting EMPTY barrels TO the outposts, and not worry about how to retrieve full barrels. Basically, you make sure that your train has empty barrels when it goes to the outpost, and over there you unload enough empty barrels to keep it stocked. You should use a combinator circuit so you wont unload too many barrels. Keep the barrels total (empty + full) a set number at each outpost.

Simply pick up any full barrels you can find, at the same time you deliver the empty ones. Then handle the reverse at the oil unload.

With the new possibilities of 0.13 trains it should be running smoothly with some combinator work.

About prod modules: if you use them everywhere, keep in mind that they are expensive to craft and they need time to amortize their cost. Which is why speed beacons are good in combination with prod modules, because they make the machines work faster, and that means the prod modules amortize more quickly (they work per machine cycle, so more cycles = more effective bonus).

Electric furnaces take 15 hours to amortize the cost of the prod 3 modules, if you use full beacon speed boost on the furnaces... so maybe not worth it if you're strapped for resources.

Always use prod3 + speed on green and blue circuits, and on plastic bars, also on copper wire for the green circuits, and on rocket parts, and in the rocket launcher (well that last one is too obvious, doh). And possibly on oil too. I also use them on red circuits, but that's another area where it takes a while to amortize (not 15 hours, though...)

[Xarovin]

So, I've done the first part, and it's turned out a lot easier than expected for the full amount haha, it seems 1 express belt with 8 assemblers is more than enough, with 2.2k oil barrels per minute (55k oil out of the required 44k, no productivity allowed though). It seems to output oil at a rate where it needs 4 pumps to free up the pipes (otherwise the assemblers fill up), though it stretches a bit thin after that, so I'm not 100% sure I've done it correctly, here's the design anyway - http://i.imgur.com/A0yHIDY.jpg

The barrel handling doesn't sound so bad, I'd just have 2 stations at each point, one to unload then the next to load (until empty/full), plus it shouldn't be too hard to design a station for the unloading since it only needs 1 express belt.
However, I just realised stopping barrels pooling up at an empty outpost might be a bit tricky, I'm thinking maybe have an extra stop, so that once the train is full of oil barrels, it'll fill the remaining spaces with empty barrels from the outpost, so it only takes the same amount of barrels back with it again. I've avoided smart inserters like the plague up until now, but I can't really see another way to get around that now lol (and I'm not sure what you can/can't do with combinators)

Anyway, with the increasing richness as you go out, that's why I'm happy with using a ton of resources haha. 15 hours is insane, but I want to go full out (also without the help of any mods), so it's not an option taking the easy route and just adding more un-moduled furnaces

Edit: Just noticed those oil processing designs need a water mod, looks like I'll have to try wing it after all

[siggboy]

It seems to output oil at a rate where it needs 4 pumps to free up the pipes (otherwise the assemblers fill up)

In your layout you use 400% speed assemblers, each will put out about 125 oil/second, which is slightly more than 4 pumps can handle.

The pumps are limited to their maximum output. You can never get more than 30/s out of a pump, no matter how fast you put it in. Four pumps in parallel can therefore not output more than 120/s.

I think you should split the output from each assembler into two, with 60/s each, and those could feed about 2 of the blueprinted designs each. 60/s each can easily be maintained over very long distances with 2 pumps in parallel (you probably won't even need it).

The barrel handling doesn't sound so bad, I'd just have 2 stations at each point, one to unload then the next to load (until empty/full), plus it shouldn't be too hard to design a station for the unloading since it only needs 1 express belt.

You don't need 2 stations at each point.

Here's in more detail how you'd do it (without mods or complicated schedulers):

For this example we assume that your trains have 2 wagons. The front wagon will take only full (black) barrels, and the rear wagon will take only empty (white) barrels.

At the refineries (where the oil is consumed), you have an unloading station that retrieves full barrels, and a loading station where the empty barrels are collected.

The trains will arrive at the depot with up to 400 black barrels in the front wagon, and a rear wagon that might be empty (or still contain some white barrels). At the unloading station you take all barrels from the train. The black ones go to the assemblers, and the white ones go to the loading station. When the train is empty you send it to the waiting bay for the loading station (you send it to the loading station, but it might have to wait in the bay).

At the loading station the train will wait until the rear wagon is full (contains 400 white barrels). Then you send the train to the outpost. If you have many outposts you probably want to dedicate one train to each of them and have multiple loading stations, so you can send trains in parallel for better throughput. I guess you can figure that out quite easily (if you give stations identical names they are treated as one by the game).

At an outpost, you do the following: white barrels are taken out of the rear wagon, but only until the outpost has a set amount of total barrels in stock. So you take "white + black barrels" in a combinator, and stop unloading white barrels when that sum reaches a certain value. In our example that value might be 400, because we can not load more than that in a single trip anyway. The exact value does not matter, as long as it is not too small -- since the outpost can not buffer more oil than it has barrels for.

While you're unloading the white barrels, you pick up as many black barrels as you can (up to 400 in our train). You allow the train to return to the depot when its cargo does not change for a few seconds (the inactivity rule).

When you set up a new outpost you need to craft the appropriate number of white barrels, because every outpost will have that many at all times (but not more).

However, I just realised stopping barrels pooling up at an empty outpost might be a bit tricky, I'm thinking maybe have an extra stop, so that once the train is full of oil barrels, it'll fill the remaining spaces with empty barrels from the outpost, so it only takes the same amount of barrels back with it again.

Too complicated, just use combinators to never unload more than (white + black) barrels at an outpost.

I've avoided smart inserters like the plague up until now, but I can't really see another way to get around that now lol (and I'm not sure what you can/can't do with combinators)

All the inserters are "smart" now, maybe you're talking about the filtering inserter. However, if you use decicated wagons for white and black barrels you don't even need a filter. The white barrels are always unloaded from the rear wagon, and the black barrels are put into the front.

You need 1 arithmetic combinator to calculate (empty barrels + full barrels) and use the result to disable the inserters so they'll stop unloading empty barrels when the limit is reached. This is the only combinator required at all in the entire setup.

Edit: Just noticed those oil processing designs need a water mod, looks like I'll have to try wing it after all

It's not a "water" mod, the mod simply gives you additional recipes for oil processing that have swapped inputs and outputs, so it makes it easier to create compact layouts.

I HIGHLY recommend you use it, it's updated for 0.13.

[Xarovin]

Found a huge flaw with my idea anyway, unless I found a way to only load full stacks, 'wait until full' also includes non complete stacks of oil barrels, so it gets stuck. As to your idea (thanks a bunch for taking the time to write it out), since you'll always have barrels on belts traveling to and from assemblers, can they still be counted in the total amount? I would have thought it'd only be the stuff in containers. Until I learn the basics of combinators however, I found another way that works, similar to your idea, but it just involves having 2 trains instead of 2 wagons (since I can't say 'wait until wagon 1 is full and wagon 2 is empty'). All that is needed is for the train to be fully loaded before it gets on the tracks, so the only barrels in circulation are actually on the trains

I'll end up splitting the outputs depending on how thirsty the designs are (I was wondering how much per second it was making oil, so just added more pumps until it started fluctuating). For example I just got the solid fuel production to 12k/m, and it needed 2 full outputs to be comfortably supplied. I remade one of the imgur blueprints with an extra space in height, so not having the mod isn't causing much of a problem (plus like I mentioned, I'm not missing out on achievements, so I'm going 100% vanilla in my main world ) - http://i.imgur.com/6OrAcj0.jpg

I had no idea about naming stations the same btw, that sounds ridiculously useful, my original plan was assign 2 trains to each station and enough waiting space for the 2nd aha.

[siggboy]

You seem to be terribly afraid of using combinators. It's really a lot easier than you think.

So here's exactly what you'd have to do at the outpost:

The front wagon is loaded from buffer chests containing full barrels.
The rear wagon is unloaded into buffer chests that will contain empty barrels.

The amount of buffer chests does not matter, you could use up to 12 but that would be ridiculous overkill, but never mind.

1. Connect all buffer chests with red wire (to eachother, like daisy chaining). The signal on the red wire will now be the grand total of ALL barrels in all chests. E.g. if there are 12 full barrels and 8 empty barrels buffered, there will be 2 signals on the red wire: "12 Filled Barrel" and "8 Empty Barrel"

2. Put an arithmetic combinator (the dark one) near the chests. Connect any of the chests to its input side with red wire. When you mouse over the combinator you'll see that it now gets the information what's in the chests.

3. Set the operation on the combinator to "Empty Barrel + Full Barrel", and set the output to "Green signal". You can use any signal for the output, it does not matter at all which one.

4. Connect all the inserters that are unloading the empty barrels from the rear wagon to eachother with red wire.

5. Set the condition on the inserters to "Green signal < 400". "400" is the maximum amount of barrels you want to exist in the buffer chests. You need to set this condition on all inserters between the rear wagon and the buffer chests.

6. Connect the output of the arithmetic combinator to any of the inserters from step 5 with red wire.

Now, each time the content of any buffer chest changes, the "green signal" is updated. This actually happens with 1 game tick of delay, because that's how long it takes to compute the result. The inserters will keep working, and unload empty barrels, until this value reaches or exceeds the limit (400), then they'll immediately stop (with 1 game tick of delay, which does not matter in this case).

Of course, since we're at the same time loading full barrels into the front wagon, the green signal might also go down again (it's the sum of full and empty). At some point, there will be no more full barrels and it will stop. Then the content of the cargo wagons will not change any longer and the train will go home due to the inactivity rule.

This system has some limitations due to delays and stack bonuses, so it's possible that more than 400 total barrels end up in the outpost, but it's not a big deal. You don't need the counts to be exact, you're just interested in it not drifting too far in either direction.

I don't recommend to use 2 trains, just to avoid having to use a combinator. The combinators are there for a good reason, and this is a really good way to use them, even very simple. Actually the perfect way to get your feet wet in that area.

Using 2 trains for different purposes is very awkward and it will lower your throughput. I know that because I've done exactly that (different trains for full and empty barrels), and it didn't work well and it's just a bad idea.

[brunzenstein]

You seem to be terribly afraid of using combinators. It's really a lot easier than you think.

So here's exactly what you'd have to do at the outpost:

screenshot of wiring?

[Xarovin]

Ah sweet thanks, just did my first basic combinator I'm usually alright with that sort of stuff (I found redstone great in minecraft and made some cool stuff), but just never thought of a use on here before haha.

Anyway, I'm just thinking, instead of doing a sum of full + empty, it seems to me if I was to fully saturate the empty barrel area when building the outpost (and let it run through to saturate the non empty barrel area), then it'd load and unload everything at the same rate. I can't test right now as I've gotta go in 10 mins, but should that idea work? I'll get your method up and running if that fails

As to brunzenstein who just replied while I was writing this, the wiring is literally like 2 wires lol, I don't think a screenshot would help very much :p

[siggboy]

Yes, you can keep all the outposts saturated, then the limit (full + empty) is basically the space in the buffers. I guess that should work as well, but then you're a bit constrained with the number of buffer chests that you can use and also setting up the outposts is a bit more work, because saturating might require a lot more barrels than a single train can carry. So you need to bring more, maybe in a larger train that you make for that purpose.

Also, while you're scaling up you might not want to start with the full amount of 400 barrels, maybe just 100 or something like that, for the mid-game.

So the combinator solution is a bit more flexible in my opinion. You'll make a blueprint for the whole setup anyway, including the assembly machines that fill the barrels and pipe connections and everything. That way this slightly more advanced solution is not more work than the simple one.

With regards to the train station condition: if the inactivity rule is not working for you, there's an alternative approach. You can make the train leave when "Full Barrel = 0 AND Green Signal = <limit>" (you set that condition in the train schedule), then you don't even need to wire the inserters, because they obviously stop unloading empty barrels anyway when the train is leaving...

In that case you put the Green Signal (output from the combinator) to the train station directly, and also the chests.

This will also make the train leave straight away when the buffer chests contain no more full barrels. The inactivity rule does have a delay, and it might even hold back the train at the station if new barrels are coming in too quickly (barrels that are just getting filled from the pump).

@Brunzenstein: I can make a screenshot, but I actually don't have a setup yet for 0.13, only for 0.12 which is very similar but more complex since it's using my train scheduler logic. The explanation above makes it look more complex that it is, like he said, it's really just 1 combinator and a few wires, nothing much to phone home about.

[Xarovin]

After a bit of testing, I think your idea may be the only one that works lol. First attempt was it'd set off once the train became inactive since it'd be full of crude oil and finished unloading empty barrels, but I didn't account for the delay to turn empty barrels to full, so it'd set off before fully unloading barrels. I then used a combinator to check all chests were full before setting off, and it worked, but then the train is sitting there until the chests are fully replenished, which by that point it could have looped around again

I did this based on what I remembered of your design since I don't want to copy totally step by step. At the main station, the rules are wait until oil = 0 and barrel = 400, and at the outpost, the rules are wait until oil + barrel > 1920 (4 chests, there are 8 in total), and oil = 400, and I set 1920 to a constant combinator since I had a spare space - http://i.imgur.com/g5n39Uv.jpg
One thing I'm wondering about, can I set the train logic to something like 'a AND (b OR c)', instead of '(a AND b) OR c' which it seems to be? I want the outpost to be 'oil barrels = 400 and (total > 1920 or empty barrels = 0)', since it got stuck once due to the empty barrels running out. If not, I guess inactivity will do the trick, but there's potential it'll mess up a little bit.

One other question, when you have to stretch a red wire somewhere and it's too far, what's the usual go to thing to use as a midway point? It looks messy if I just connect to a random belt or inserter.

[siggboy]

Why do you keep so many barrels at the outposts? If the train brings as many empty barrels as it will pick up full barrels, then the additional barrels are not necessary, they're just a buffer of oil that's not really useful. If you have enough trains then these buffers should never build up in the first place, it's oil waiting to be consumed, it should be at your factory . OK it's not entirely true, some additional barrels are useful as an output buffer but probably not 2000...

I don't think you can do "a AND (b OR c)" in the train station since it does not allow for parenthesizing the expressions. If you really need this you need to take "b OR c => d" in a combinator and then you take "a AND d" in the train station.

So one combinator does "total > 1920 => red 1", another one does "empty barrels = 0 > red 1", you put both of these outputs to the station and then in the station you do "oil barrels = 400 AND red > 0"

I've used this technique before to make an "OR" condition with combinators, it's pretty basic.

If you want to extend a wire you use a power pole, or maybe there's a combinator that you can abuse for that purpose (but that's often confusing, so a power pole is better).