Comprehensive power management

[mrvn]

2) Excellent. But now I am a bit confused, why is it that most people try to mimimize pipe distances from reactor to heat exchangers?

3) Again good. And again confused, if the setup is equally efficient from 500-1000, why are people so focused on heating things up to 999 degrees?

Actually people would love to run the reactor at 500°C or as close to it as possible. It has to run a lot hotter though for maximum power generation. You never want it to hit 1000°C but anything below that is fine. The closer your heat exchangers are to your reactor the cooler it can run leaving a bigger margin for when not all energy is consumed. Why?

A heat exchanger will start working at 500°C using up a set amount of heat. For it to keep working that heat has to be replaced from the heat pipe, meaning the heat pipe has to be x°C hotter to allow enough flow into the heat exchanger. Same between each pair of heat pipes. So the maximum distance a single heat exchanger can be from the reactor is 500/x m.

But usually you have more heat exchangers. So where the second heat exchanger pulls it's energy the temperature difference has to be 2x. With 3 heat exchangers it has to be 3x. For 16 heat exchangers the first must be within 500/x/16 m from the reactor, which isn't very far. You can calculate it exactly from the game values or experiment with placing heat exchangers at different distances. What you will see is that the heat pipes will be at a lower temperature the further away from the reactor they are even with the reactor at 1000°. At some point the heat exchangers will stop working 100% because the heat difference simply won't allow enough heat to flow. The temperature will be below 500+x°C. Think of it as a river. The steeper the gradient the faster the water flows and you need a certain throughput.

Also: On fuel insertion a heat exchanger closer to the reactor starts working faster. That means the steam reserve doesn't have to be so big.

[JimBarracus]

Actually people would love to run the reactor at 500°C or as close to it as possible. It has to run a lot hotter though for maximum power generation. You never want it to hit 1000°C but anything below that is fine. The closer your heat exchangers are to your reactor the cooler it can run leaving a bigger margin for when not all energy is consumed. Why?

annotation: you can also build more turbines and heat exchangers than required for an optimal ratio, because you can use the excess heat (500°-999°) to power more than your constant output.
Sure the power will drop after a certain time because the reactor cools of too much but its a good way to deal with power peaks.
Especially when your current setup is at its limit.

[mrvn]

Actually people would love to run the reactor at 500°C or as close to it as possible. It has to run a lot hotter though for maximum power generation. You never want it to hit 1000°C but anything below that is fine. The closer your heat exchangers are to your reactor the cooler it can run leaving a bigger margin for when not all energy is consumed. Why?

annotation: you can also build more turbines and heat exchangers than required for an optimal ratio, because you can use the excess heat (500°-999°) to power more than your constant output.
Sure the power will drop after a certain time because the reactor cools of too much but its a good way to deal with power peaks.
Especially when your current setup is at its limit.

Sure. That's using heat a energy storage. You can also add tanks to store excess steam and have more steam turbines than possible to feed with the heat exchangers.

If you add solar cells you can use ~3 times the turbines and steam tanks or also 3 times the heat exchangers and only use them at night. But I think you can't fit 3 times the heat exchangers on a reactor and have them all work 100%. So steam tanks or a mix would be my guess of how to make it work. Personally I think fuel management to run the reactor only as much as needed is simpler. But you can do all three. Fuel when needed, buffer heat for the night, overproduce steam and run e.g. twice as many steam turbines per reactor.

But hey, if you add a daylight clock you could throw in fuel some time before night hits and spin up the reactor for night power instead of having it demand driven.

[Greybeard_LXI]

Lots of interesting information in this thread about setting up nuclear power in this thread. But...

I tend to get very focused on what I am working on, so I end up not monitoring the power graphs and get brownouts/blackouts if I don't have alerts. And it "feels" like a waste to waste resources significantly overproducing power just to avoid them.

Maybe one way to keep up with power is to plan it when you start expanding.

Say you want to build a green circuit module with 50 assembler 3s that have an average of 4 fast inserters. (Plus other stuff like maybe beacons and modules I'm not going to count.)

50 Assembler 3s @ 325KW = 16,250 KW
200 Fast Inserters @ 46KW = 9,200 KW
Total power needed = 25,450 KW

At 60 KW per solar panel you need 424.1666... (425) solar panels and an appropriate number of accumulators. (I'm not sure I'm taking night into account correctly.)

Just include this in your plan for the green circuit module. Don't go on to red circuits until you build out the power for the green.

Making a spreadsheet to calculate how many panels you need for any module shouldn't be too hard.

Edit to add notes:
Note 1: Maybe you already do this. Sorry if I'm not giving you credit for your experience.
Note 2: Do as i say, not as I do.

[Koub]

(I'm not sure I'm taking night into account correctly.)

You're not. Solar panels average 42 kW when night is accounted. they only produce 60kW at peak, when they get enough light

[mrvn]

First problem with that idea: When you first need green circuits you don't have an solar panels.

Second problem: When you've got solar panels there is no space next to the green circuits to retrofit them because stupid you didn't leave space.

Third problem: If you play without solar panels adding in steam engines or *gasp* nuclear to each blueprint is wasteful because you can only increase power in 1.8MW lots (or 900kW wasting half a boiler) and getting fuel adds more overhead.

I, and probably most other people, tend to build power in larger chunks. Like add 100MW at a time and then add a ton of stuff before noticing power is down again. Counting every assembler and inserter and producing solar cells to match is just too much math.

[Bauer]

The brown-out is your friend!

It's not the end of the world if you don't have enough power to keep everything going.
The only really bad thing that could happen is a black-out, i.e. your power production including the mining is out of power.
Hence, if you put that (the power production) on a separate grid which is 100% reliable (solar) nothing really bad will happen (only lower production speed and lower UPS).

I also recommend to stock the base material for the next step of your power plant extension.

[Ranger_Aurelien]

I usually notice power is low when I am working with inserters and see the power bar for it on the right is not green, or at worse, when the inserter arms seem to have trouble grabbing. You could set up an alarm tower (aka "Programmable speaker") with a green wire to an accumulator that sounds map-wide when the power drops below a threshold (say 10%). Since accumulators drain last after solar, boiler, nuclear, that would mean power is indeed low.

https://wiki.factorio.com/Programmable_speaker


I agree mid- and later- game I tend to add power in blocks.

1 offshorepump powers 20 boilers powering 40 steamengines I have my own blueprint for that, and it seems to be a good unit to add at a time for mid-game...

https://wiki.factorio.com/Offshore_pump
https://wiki.factorio.com/Power_production#Steam_power


Later, with robots to deliver solar and accumulators, I have a blueprint for a chunk that includes four roboports I picked up from the multiplayer stress map.... My personal preference is to assume 1:1 solar to accumulators, which allows some leeway for lasers and avoiding brownouts.


For nuclear, there is a "neighbor" bonus for the nuclear cores, so having only one or two cores is inefficient, so I have a blueprint for a long facility with four nuclear cores (with four offshore pumps for 83 turbines) that is my basic module if I need nuclear, but I prefer solar when possible.

https://wiki.factorio.com/Tutorial:Nuclear_power

KatherineOfSky's Nuclear (I based my four-core off her design, at about ~15:00 mark)
https://www.youtube.com/watch?v=yA21zrqMMq0


In the end it is technically wasteful to build too much more than you can need in a reasonable amount of time (Maybe better use of time in the short term to do /spend resources on something else, but bases just grow so power is normally a good investment) -- but you don't want to run out of power either. Like many answers in this forum, "build as many as you need plus a few more in case biters eat them".



Regarding backup systems, I agree having auxiliary "starters" can be critical to avoid a snowball brown-out (Inserters are low on power so they can't feed more coal into boilers so boilers go out and there is even less power, so all power stops.)
You can "clip" all the wires from a tower using SHIFT+LEFTCLICK (I believe), then manually run copper coil between them. That way you can create independent overlapping networks.
I set up small coal plants that power only the inserters of the main coal plants, plus the electric coal miners.

https://wiki.factorio.com/Keyboard_bindings


For nuclear I just add two dozen silos to store steam, so they can be self-starting (the nuclear cores always make steam when running hot enough and the nuclear turbines only burn steam when they are needed to provide electricity beyond solar). It's also part of an "efficient" nuclear fuel usage system that uses Circuits to reconcile the two.


Circuit wiring by this guy (Tuplex) at about ~29:00 mark
https://www.youtube.com/watch?v=_ReduT-3Wx4

Take care!

[Koub]

You could set up an alarm tower (aka "Programmable speaker") with a green wire to an accumulator that sounds map-wide when the power drops below a threshold (say 10%). Since accumulators drain last after solar, boiler, nuclear, that would mean power is indeed low.

I'd want to be warned as soon as the accumulator gets below 100% for any non solar power source. An accumulator holds a ridiculously small amount of energy (on a factory scale), and if it ever gets below 100%, the brownout and blackout are coming.

[SuicideJunkie]

I set up some circuit logic to store a buffer of 6 steam level readings, and use that to get a MW draw value directly off an individual nuclear plant.

There is also a steam lock, where I pump a tank to 100% full, then toggle the pumps over to pumping it into the main storage buffer.
At the moment it toggles, I add that 1 tank worth of steam to the old memory cell values and connect the lock tank to the reading for steam levels.
The steam lock ensures that the measurements are accurate, and I don't get negative readings when the reactors are repressurizing the storage.

That hooks into a digital lamp display and a speaker for alarms, so I can be alerted if the plant overloads on MW or dips dangerously on the GJ stored.

When the alarm goes off, it is time to upgrade power, or fix a broken belt of fuel cells.


On the power usage side, I also like the accumulator slosh pile. Handy for low duty cycle, high demand, low priority things.
1) when main storage is rising or full, an accumulator bank is hooked into the grid to trickle-charge from any excess solar generation.
2) when the bank gets full, it is connected to the consumer. Usually it disconnects from the main grid due to the accumulators dropping to 99% at this point.
3) the consumer burns through the power to run beacons and a high demand production area until empty, at which point the consumer is disconnected again.
4) goto 1

[Bauer]

accumulator slosh pile

I understand the concept. Out of curiosity, what application do you have for a high demand, low priority setup that's worthwile to setup a huge accumulator field? I just cannot think of any (my fault).

[SuicideJunkie]

I used it for oil processing under Nauvis day to keep the pollution down; bursty power to get an efficiency + speed combo for minimum pollution per product, with solar input.

Also for my over-unity rocket fuel production.
That needs 7x speed + 4x productivity, but turns 100GJ of SF into ~115GJ of RF and only costs about 4GJ of electricity.
It operates so fast that the duty cycle is quite low, and the slosh pile lets it keep up all those beacons without tripping an emergency on the main grid every couple days when it kicks on.


I suppose you could also use a similar scheme (cap gap, but no logic) to halve the cost of a beacon setup:
If you can't pay the drain, but can otherwise pay the operating costs, all the devices will still operate at 100%, and beacons have stupendous drain to save on.

Once you have nuclear power running it is probably irrelevant, but it is quite nice when you're on solar with chemical backups and large spikes will burn resources or set set off alarms.

[Bauer]

Thanks. Understood. Idle beacons use a lot of energy for nothing.

I usually come from the other side: I switch off parts of the factory when they are idle (RS-latch on a few buffer chests).
In order to make this work, my solar-accululator-ratio is closer to 1:1 in this phase of the game. And the steam setup is also RS-latched to the grid because the demand doesn't synchronise well with the day-night-cycle. This is the advantage of your setup.