"Big power" options, low UPS hit

[zOldBulldog]

@zb, there doesn't seem to be anything out there, there's nucs that'll serve perfectly well in 10GW maps, there's the electric energy interface that costs whatever you think's fair and produces whatever you think's right, but that got met with silence, bob's mods add a bunch of tiers to the vanilla solar panels, but those are also somehow not good enough. How would anyone trying to help know what would be good enough?

Where do you get that solar/accumulator tier mods aren't good from?

- Bob's probably isn't good enough because it would totally change the game with its other stuff, but I don't know if any of the other solar/accumulator tier mods would do the trick of duplicating the "general feel" of using nuclear.
- Bio Industries seems to have something as well and it looks very interesting but also changes too much of the core game. Still, I might use it someday in a separate game.
- I also saw 3 or 4 Advanced Solar mods that *might* do the trick with the various solar tiers (even one that has a 10x factor for the tiers), but I don't know enough about them to know if they are viable and I have not been able to find much useful information to get a feel for it even in the documentation they have.

Also, there must be some other power mods. I saw some geothermal, wind, etc... but they all seem to generate too little electricity to match nuclear. But again... I don't know enough.

Thus... the thread. ***Somebody*** must know.

[mrudat]

How about Recursive Blueprints + Bot Landfill?

It's not quite 'just build this thing', but once setup it should require 0 attention for infinite power.


There are also a number of mods that offer more dense nuclear power (in terms of W/entity). Notably Bob's power and Advanced Electric, for example.

One thing that works for small nuclear power is a mod that adds an Enormous tank that holds 1M fluid; it's large enough (in terms of surrounding the tank with turbines and heat exchangers) that you don't need pipes for a 40MW power plant, or at least you need only a very few pipes.

There's another that merges 1-4 turbines into a single larger entity.


There's a mod called Dyson Swarm that launches solar collecting satellites and beams power back to your factory.


There's Fission and Fusion, but I'm uncertain if it actually saves UPS or not.

[zOldBulldog]

Those are really good hints mrudat.

And I think I might have found the Holy Grail... if I understood it right.

Luzivras Factorio Power: https://mods.factorio.com/mod/Luzivras_Factorio_Power

It looks like 8 tiers of advanced solar panels and accumulators where each one is equivalent to 8 of the previous one. If the cost of each tier is the same as all the panels/accumulators that go in 8 of the previous (plus maybe a small additional cost to pay for the convenience) and I did my math right... it would allow to setup a production chain to build the stuff and then concentrate all solar into one chunk... up to 704GW average power through the day once you reach his tier 8. And if it also has zero hit on UPS like vanilla solar... then it might just be the Holy Grail.

Fingers crossed. I hope it is all that.

[Zavian]

But it is well known that nuclear doesn"t scale [...]

The answer *should* be already well known and easy to name for experienced players

That's what I'm trying to get at: I see you and lots of people repeating the _claim_ that nucs aren't good enough, and I certainly see lots of designs that'll eat any rig like it's going out of style, but when I try for a scalable-enough design that'll handle a modest megabase, it seems to work. It's where I'm going after I get my new launch base gizmos all worked out, that's for sure. And maybe goose the train-routing circuitry, see if I can't find the LTN-in-vanilla grail. And and and.

Anyway, i'm pretty far from the sharpest nuc plant designer in the shed, I look at what rad and that guy on reddit posted and think "why couldn't I see that?", but even with my setup I get ~200µs/GW, tested just now, 0.3ms with nothing but an eei and 20K beacons, 2.2ms with the nuc plants and the beacons, more than I thought but not much.

So I think nuclear does scale, if you're conservative with the design. Could you point at what you're looking at that says different, give something concrete to work with?

Just a example (Using the timing figures you posted above, plus timing figures I made up for a hypothetical megabase design).

Let's assume that player Alice wants to build a megabase with as much science/minute whilst still getting 60 ups. She builds a nice modular design that makes 1250 science does minute. During the test phase, she uses the electrical energy interface for "free" power, for easier testing. She tests and refines it, optimising for lowest possible ups cost. She ends up with a design that can do 1250 science/minute using 2ms for entity update. She says "Ah. I can build 8 of these, flipped and rotated for a nice symmetrical megabase that will do 10k science/minute at 60 ups". (This is a simplified example, please ignore the fact that such designs don't always scale linearly, and that flipping can be problematic. I know that).

So she plonks down a nuclear power plant, and suddenly the update cycle now takes 4.2ms. So she looks at her power consumption, and does some math, and realises that she will need 4 of those nuclear power plants to power her 10k science/min factory, and that if things scale linearly, she will be down to 40 ups. Or she can spam solar and get 10k science/minute at 60 ups.

When optimising for ups, solar has essentially zero ups impact. Nuclear is good enough, until you decide that you want those extra ms that power is using for more science. At that point solar wins.

No matter how much the devs improve performance, solar will still be better than nuclear, for a player who is optimising their design for ups. (Assuming the existing nuclear model. the devs could rework nuclear into build plant, feed U-235 in, get power out, without all the heat pipes turbines etc).

Also every time the devs improve game performance, players build larger factories. When I started 1 rocket per minute was considered good. In 0.15 player were building 4k sci/minute. Now I've seen youtube videos of 10k sci/minute bases. eg https://www.youtube.com/watch?v=Av8DZo368Jw . 10k science/minute at nearly 60 ups. that base uses over 100 GW. I'm not sure how many GW your 2ms power plant can sustain, but even with this 20GM design, you would still need over 5 of them. https://www.reddit.com/r/factorio/comme ... gn_tested/

[eradicator]

She ends up with a design that can do 1250 science/minute using 2ms for entity update. She says "Ah. I can build 8 of these, flipped and rotated for a nice symmetrical megabase that will do 10k science/minute at 60 ups". (This is a simplified example, please ignore the fact that such designs don't always scale linearly, and that flipping can be problematic. I know that).

I don't think i will ever understand why people post "arguments" of which they know they're based on false premises, it's just a null-statement. Also i don't think anything in factorio ever scales linearly. No offense meant. I just don't get it :p (Edit: Hm, i guess the method used to derive the maximum repetition count is irrelevant. But then why even mention a method... i should get some food.)

When optimising for ups, solar has essentially zero ups impact. Nuclear is good enough, until you decide that you want those extra ms that power is using for more science. At that point solar wins.

THIS is the problem. A zero cost solution exists. Nothing can be smaller than zero (no, you can't make an entity that gives you more UPS for each one you build :p). Nuclear consumes fuel, computing fuel consumption costs time. You can't optimize fuel consumption down to 0 cost. Thus the only thing even better than solars would be...solars that work at night. The dyson-swarm mod mentioned above is probably the closest you get to that. You have only one entity that produces all the power (without consuming anything), optimizing away the last bit of cost that solar has - the memory cost of storing all the passive entities. And because it works at night it also optimizes away the need for accumulators (which also have 0 computing but non-zero memory cost).

[zOldBulldog]

Thank you zavian, I knew that kminus the math) but was obviously having trouble explaining clearly.

Do you know if the solar mods that concentrate the panels into less space by making higher tier panels preserve the "zero impact on UPS" characteristic?

[orzelek]

Thank you zavian, I knew that kminus the math) but was obviously having trouble explaining clearly.

Do you know if the solar mods that concentrate the panels into less space by making higher tier panels preserve the "zero impact on UPS" characteristic?

They should. From what I recall game groups solar and accus per type for the calcualtion purposes.

[eradicator]

They should. From what I recall game groups solar and accus per type for the calcualtion purposes.

I'd be suprised if the types played any role at all in the calculation. They all have the same daylight regression, so you simply calculate current_power = (max_power * brightness) each tick. Where max_power is a number updated only when solar types are built or removed. But - regardless of engine internals - mods have no way to change how it works, even if they wanted. Thus all solar prototypes must behave the same.

[orzelek]

They should. From what I recall game groups solar and accus per type for the calcualtion purposes.

I'd be suprised if the types played any role at all in the calculation. They all have the same daylight regression, so you simply calculate current_power = (max_power * brightness) each tick. Where max_power is a number updated only when solar types are built or removed. But - regardless of engine internals - mods have no way to change how it works, even if they wanted. Thus all solar prototypes must behave the same.

I think it was mostly meaningfull for accumulators - to work as one they need to have same capacity and be at same load level. And game groups them based on that so in theory once you load your whole field to full they should be neatly grouped depending on their capacity.

[zOldBulldog]

They should. From what I recall game groups solar and accus per type for the calcualtion purposes.

I'd be suprised if the types played any role at all in the calculation. They all have the same daylight regression, so you simply calculate current_power = (max_power * brightness) each tick. Where max_power is a number updated only when solar types are built or removed. But - regardless of engine internals - mods have no way to change how it works, even if they wanted. Thus all solar prototypes must behave the same.

I think it was mostly meaningfull for accumulators - to work as one they need to have same capacity and be at same load level. And game groups them based on that so in theory once you load your whole field to full they should be neatly grouped depending on their capacity.

Wouldn't it be common sense to use 1-n of the same capacity, where n is the number needed for the next capacity? The upgrade to one or two of the next capacity? Basically upgrading by a bigger MW amount each time you upgrade.

(Note: I am just going in theory. I have not started using the mod yet)

[mrudat]

Given UPS concerns, it would be marginally more efficient to manufacture the maximum tier, and just place that, rather than building increasing tiers of solar panels.

Getting to megabase scale it may or may not be better to keep upgrading tiers, or just using vanilla solar until you've unlocked the maximum tier.

It's a tradeoff of robot+player time to upgrade existing solar vs robot+player time to place more landfill for solar to sit on.

[mrvn]

@quyxkh:

I fiddled around with your design and meh. Every tiny change i tried on it made it worse. The main problem is that the maximum distance you can carry heat away from a reactor seems to be pretty much exactly the size of that design. If i add a tiny bit of spacing between the turbines power output drops immediately. Water transfer is rather trivial, much more so than i expected (feared) from your original design. With a handful of landfill and a medium sized lake i get this:

1120mw_nuclear.png

You should add at least one solar panel so the pumps can work in a black out situation. Otherwise it's impossible to start up the reactor.

[Aeternus]

Can confirm that anything with fluid mechanics tends to be a nightmare on UPS. I'm glad the fluid mechanics are getting a rework. Keeping the fluid system for your nuclear plant and refineries simple (minimizing pipe sections and pumps) helps a little.

I've had to scrap a 10k SPM megaplant idea with 8 sections of individual science production. When I finished the third segment (~30 GW of power in use) UPS tanked below 20. It just wasn't fun anymore to play that. My machine's an I7, playing from a SSD with 8 GB memory, so processing power should not be an issue. The base was mostly nuclear powered, with a single huge dull column of reactors flanked on both sides by a row of heat exchangers and beyond that some storage tanks and turbines. The whole thing was a single pump -> heat exchanger -> tanks -> turbines row, not crosslinked anywhere.

I'll try again, this time with pure solar and probably a centralized refinery to keep duplicate pipeworks to a minimum. We'll see how it goes. Plastic deliveries will be a pain in the neck.