Factorio Forums

www.factorio.com
https://forums.factorio.com/

Is nuclear power really bad for megabases?

https://forums.factorio.com/viewtopic.php?f=208&t=101392

Page 2 of 2

Re: Is nuclear power really bad for megabases?

Posted: Wed Feb 02, 2022 3:12 pm
by gGeorg
SoShootMe wrote: ↑
Wed Feb 02, 2022 7:12 am
 1k SPM "load",

Electric energy interface: 1916
Solar panels/Accumulators (19x 8000 panels and 6720 accumulators): 1938
Nuclear (3x blueprint below, stacked): 979

nuclear power resulted in a 50% reduction in UPS.
Translated your result to other point of view :

1k SPM powered by Nuclear or 2K SPM powered by solar ~ the same UPS

Re: Is nuclear power really bad for megabases?

Posted: Wed Feb 02, 2022 5:10 pm
by SoShootMe
gGeorg wrote: ↑
Wed Feb 02, 2022 3:12 pm
 Translated your result to other point of view :
1k SPM powered by Nuclear or 2K SPM powered by solar with the same UPS.
No need for translation - I wrote a direct equivalent already. But being "worst case" leads to an important caveat you have glossed over (emphasis added):
SoShootMe wrote: ↑
Wed Feb 02, 2022 7:12 am
 That is, solar could let you build up to twice as big before UPS starts to drop, but most likely significantly less than that.
For example:
coppercoil wrote: ↑
Wed Feb 02, 2022 9:58 am
 I took my real 430 SPM factory and replaced 3.8 GW nuclear power to electric energy interface.
Nuclear: 320 UPS.
Electric interface: 380 UPS.
This is ~19% higher SPM for the same UPS, rather than ~98% higher.

Re: Is nuclear power really bad for megabases?

Posted: Wed Feb 02, 2022 8:14 pm
by quyxkh
Try this one?

Re: Is nuclear power really bad for megabases?

Posted: Thu Feb 03, 2022 2:54 am
by gGeorg
SoShootMe wrote: ↑
Wed Feb 02, 2022 5:10 pm
gGeorg wrote: ↑
Wed Feb 02, 2022 3:12 pm
 Translated your result to other point of view :
1k SPM powered by Nuclear or 2K SPM powered by solar with the same UPS.
No need for translation - I wrote a direct equivalent already. But being "worst case" leads to an important caveat you have glossed over (emphasis added):
SoShootMe wrote: ↑
Wed Feb 02, 2022 7:12 am
 That is, solar could let you build up to twice as big before UPS starts to drop, but most likely significantly less than that.
For example:
coppercoil wrote: ↑
Wed Feb 02, 2022 9:58 am
 I took my real 430 SPM factory and replaced 3.8 GW nuclear power to electric energy interface.
Nuclear: 320 UPS.
Electric interface: 380 UPS.
This is ~19% higher SPM for the same UPS, rather than ~98% higher.
Your post is perfect answear of the OP.
The Nuclear cost has range, from ~19% to ~98% SPM compared to solar on the same UPS.

Would you add to compare my Cloverleaf plant powering 1k SPM optimised base ?
viewtopic.php?f=208&t=96233
I have worked on optimised design for a year, so I would like to know the rate.
Thank you

Re: Is nuclear power really bad for megabases?

Posted: Thu Feb 03, 2022 7:41 am
by SoShootMe
gGeorg wrote: ↑
Thu Feb 03, 2022 2:54 am
 Your post is perfect answear of the OP.
The Nuclear cost has range, from ~19% to ~98% SPM compared to solar on the same UPS.
What I've written is far from perfect :). There's a range, but it depends on how UPS-optimised both production and power are. For the former, I think it's pretty unlikely any "real" game (...today) will be better than Stevetrov's 1k SPM cell supplied by infinity chests/pipes (I expect belting/piping in from surrounding sources would be similar, but not terribly practical) - that's why I chose it to get a "worst case" figure. But I write "worst case" in quotes because it's obvious nuclear power can be less UPS-optimised than the relatively simple 2N design I used. I chose that design because I had it to hand, it is appropriate for large scale nuclear power, and I figured that scale requirement points towards similar designs that won't have a huge range of UPS at a given output (that may be a bad assumption).
Would you add to compare my Cloverleaf plant powering 1k SPM optimised base ?
viewtopic.php?f=208&t=96233
I tried, but output oscillated until I added an EEI to consume excess power - yielding a result with 14x 480MW of 745 UPS. However, UPS comparison between nuclear designs is probably best done in isolation, like the save from the original post.

Re: Is nuclear power really bad for megabases?

Posted: Thu Feb 03, 2022 9:26 pm
by ptx0
SoShootMe wrote: ↑
Thu Feb 03, 2022 7:41 am
 I think it's pretty unlikely any "real" game (...today) will be better than Stevetrov's 1k SPM cell supplied by infinity chests/pipes
you'd be surprised how much UPS those infinity items consume.

Re: Is nuclear power really bad for megabases?

Posted: Fri Feb 04, 2022 1:15 pm
by SoShootMe
ptx0 wrote: ↑
Thu Feb 03, 2022 9:26 pm
SoShootMe wrote: ↑
Thu Feb 03, 2022 7:41 am
 I think it's pretty unlikely any "real" game (...today) will be better than Stevetrov's 1k SPM cell supplied by infinity chests/pipes
you'd be surprised how much UPS those infinity items consume.
That is well known, has always surprised me, and is a valid point, but doesn't change what I wrote. Do you think there are "real" games with better SPM per UPS?

Re: Is nuclear power really bad for megabases?

Posted: Fri Mar 04, 2022 3:57 pm
by causa-sui
Alt-F4 blog has an interesting post on this topic that rekindled my interest.

I'm interested in this reddit post linked off this alt-f4 blog as well.

Re: Is nuclear power really bad for megabases?

Posted: Thu Feb 09, 2023 4:55 pm
by leathermodiste
A usual nuclear reactor produces 1 gigawatt (GW) of electricity which you can't simply replace it with a 1 gigawatt coal or renewable plant.
Based on the capacity factors above, you would actually need almost two coal or three to four renewable plants to generate the same amount of electricity onto the grid.
If we see in this manner then we can say that Nuclear Power is not bad for megabases.

Re: Is nuclear power really bad for megabases?

Posted: Mon Aug 14, 2023 10:40 pm
by Premu
My first megabase was a 1k SPM base without any regards to UPS, and used nuclear power. I still managed to achieve 60 UPS.

My next megabase was a 2k SPM base, but there I applied at least some optimizations. I still used a nuclear power plant, just bigger. Still I managed to achieve 60 UPS.

For a 2.7k SPM base with an even larger nuclear power plant (it produced around 17 GW), I dropped to something like 45 UPS. Obviously the nuclear power plant isn't the sole reason for that - a lot of the rest of the base needs a lot of computing power as well, but at least here it might make sense to replace nuclear power with solar panels. For anything up to 2k SPM I wouldn't bother with the massive effort to build something like 300 000 solar panels and accumulators.

Right now I build a new 2.7k SPM base, but with time with solar power and further UPS optimizations.

Re: Is nuclear power really bad for megabases?

Posted: Sat Sep 09, 2023 4:22 pm
by brakes59
I find it interesting that most commenters here completely missed the point of the original post, yet were so passionate about their preconcieved ideas.

I personally find the question posed by the original poster very interesting and I would really like to see a more thorough analysis of ups cost resulting from nuclear power usage. I would also ask a question of a similiar nature: for each evolution factor, which defence design is sufficient, but associated with minimal effort? I have been wondering, if, for a maximal evolution facotr, three rows of laser turrets with walls and dragon teeth are enough.

Re: Is nuclear power really bad for megabases?

Posted: Sat Sep 09, 2023 7:37 pm
by Tertius
brakes59 wrote: ↑
Sat Sep 09, 2023 4:22 pm
 I find it interesting that most commenters here completely missed the point of the original post, yet were so passionate about their preconcieved ideas.
If you ask me, and this is only my personal opinion: Questions about UPS come up, because the computer Factorio is running on isn't powerful enough to provide 60 ups. This is a shortcoming of the player's equipment, not the shortcoming of some factory design. Because of this, I (personally) don't deem UPS something worth discussing. I'm interested in solutions that solve challenges from within Factorio, not challenges because I have a too slow computer. This is the completely opposite opinion compared to the opinion expressed in the OP. His name of the game is UPS, my name of the game is "solve factory challenges".

If you want to achieve something as a technician, you need the proper tool. If you don't have the proper tool, it gets tedious and you achieve less. In terms of Factorio UPS and the person playing Factorio, the proper tool for a bigger factory is a better computer Factorio is running on, not tedious work to design strange factory designs just because of computer+engine performance. This is not a challenge worth doing, because the next computer generation or Factorio release is obsoleting that research.

Re: Is nuclear power really bad for megabases?

Posted: Sat Sep 09, 2023 10:25 pm
by mmmPI
Tertius wrote: ↑
Sat Sep 09, 2023 7:37 pm
 If you want to achieve something as a technician, you need the proper tool. If you don't have the proper tool, it gets tedious and you achieve less. In terms of Factorio UPS and the person playing Factorio, the proper tool for a bigger factory is a better computer Factorio is running on, not tedious work to design strange factory designs just because of computer+engine performance. This is not a challenge worth doing, because the next computer generation or Factorio release is obsoleting that research.
As a technician i could rate my own designs by their output on my tool, the better rated giving the more SPM, ( or the most SPM/ real time taken to make the factory ) , if i improve my design, with the same given tool, i will produce more science, or get faster to the desired SPM, it is even more important if i have a poor tool to choose amongst my design the one that uses the least amount of computing power for said task, be it energy production, or material transformation :)

I don't think it's fair to say it gets obsolete from new hardware, only from when the game changes its internal, and that one player now suddenly has to reconsider if doing one way is still the "most optimal" in terms of UPS.

Such "research" is used even if you have a super good tool, if you want to get the most SPM out of it and push it to its limit, although the designs are different due to scale, they may use the same tricks to achieve the "best perfomance".

Plus everyone consumes less real life electricity when using UPS-optimized designs :)
brakes59 wrote: ↑
Sat Sep 09, 2023 4:22 pm
 I personally find the question posed by the original poster very interesting and I would really like to see a more thorough analysis of ups cost resulting from nuclear power usage.


It is possible to test individual nuclear reactor, since their designs can be very different, their perfomances too. I recommend having a test map, with the /editor mode, where you only build 1 nuclear power plant, and you press F4 to show " time-usage"

you can then compare the update time in ms that it takes to your machine to run factorio on an empty map, and on a map with a nuclear power plant, and on a map with solar pannel. The worst the computer and/or the bigger the nuclear power plant, obvisouly the better the solar. On a very very old computer, you will use all your UPS trying to run a 40MW single reactor, on a super good computer i don't know how many terawatt you can produce ,but eventually you will tank your UPS too.

Re: Is nuclear power really bad for megabases?

Posted: Mon Sep 11, 2023 8:50 pm
by brakes59
Tertius wrote: ↑
Sat Sep 09, 2023 7:37 pm
brakes59 wrote: ↑
Sat Sep 09, 2023 4:22 pm
 I find it interesting that most commenters here completely missed the point of the original post, yet were so passionate about their preconcieved ideas.
If you ask me, and this is only my personal opinion: Questions about UPS come up, because the computer Factorio is running on isn't powerful enough to provide 60 ups. This is a shortcoming of the player's equipment, not the shortcoming of some factory design.
I believe in your case the shortcoming is the scale of your factory. If you ever get to 5k SPM, UPS will become an issue no matter how powerful your computer.

Re: Is nuclear power really bad for megabases?

Posted: Mon Sep 11, 2023 11:51 pm
by Tertius
brakes59 wrote: ↑
Mon Sep 11, 2023 8:50 pm
Tertius wrote: ↑
Sat Sep 09, 2023 7:37 pm
brakes59 wrote: ↑
Sat Sep 09, 2023 4:22 pm
 I find it interesting that most commenters here completely missed the point of the original post, yet were so passionate about their preconcieved ideas.
If you ask me, and this is only my personal opinion: Questions about UPS come up, because the computer Factorio is running on isn't powerful enough to provide 60 ups. This is a shortcoming of the player's equipment, not the shortcoming of some factory design.
I believe in your case the shortcoming is the scale of your factory. If you ever get to 5k SPM, UPS will become an issue no matter how powerful your computer.
It just happens that I'm building such a megabase, and my old PC with a Core i7-6700K went just under 60 ups when I got near completion. I bought a new PC with a Core i5-13600k half a year ago, and ups isn't a thing with it yet. The new CPU has about 4 times the performance according to benchmarks, and Factorio ups roughly doubled. Fps suffers if I zoom out too much (zoom mod installed), but regular zoom is 60 fps / 60 ups.
The base is almost completed, the stuff that's remaining is connecting some open ends, so the last infinity chests can go away (I'm developing it in map editor mode). Larger production lines aren't missing, so what's remaining will not have much performance impact.

I built a little bit with ups in mind, so I tried to minimize inserter swings and maximize the sleep time of the inserters, because inserters had the biggest impact in terms of time used. But I didn't explicitly design for ups. I just tried to be somewhat efficient. Stack inserters instead of blue inserters do help a bit, if they're able to grab more than 3 items for a swing, and blue inserters instead of yellow inserters help as well, because they're able to go to sleep faster due to their faster swing. Less splitters, if possible. And I tried to reduce sideloading, if possible.

But that's about all. I tried to evaluate the impact of a multi GW nuclear power plant by stamping some 160 GW tiled setup on the megabase map, but impact was almost negligible compared to the time used by factory inserters and assembling machines. As far as I see it, there are other pieces of a factory that have much bigger impact on UPS than the nuclear power plant, so dealing with these seems more promising.

Re: Is nuclear power really bad for megabases?

Posted: Tue Sep 12, 2023 1:54 am
by SoShootMe
Tertius wrote: ↑
Mon Sep 11, 2023 11:51 pm
 As far as I see it, there are other pieces of a factory that have much bigger impact on UPS than the nuclear power plant, so dealing with these seems more promising.
This is an (important IMO) implication of my earlier conclusion. UPS will definitely start to drop at a lower SPM if you scale up power with nuclear rather than with solar, but unless you have highly UPS-optimised science production, there is a good chance there are bigger potential gains with changing science production than power generation.

Re: Is nuclear power really bad for megabases?

Posted: Tue Sep 12, 2023 2:30 am
by mmmPI
Tertius wrote: ↑
Mon Sep 11, 2023 11:51 pm
 I tried to evaluate the impact of a multi GW nuclear power plant by stamping some 160 GW tiled setup on the megabase map, but impact was almost negligible compared to the time used by factory inserters and assembling machines. As far as I see it, there are other pieces of a factory that have much bigger impact on UPS than the nuclear power plant, so dealing with these seems more promising.
Maybe you are correct on where the bigger gains lies , however the original post is also 160GW in nuclear power, and it takes a lot of time for the update ( 12 ms ), so you must have a more UPS-friendly nuclear plant, or a faster computer that the one whose spec are shown. ( in the op 2 ms are taken by "heat manager" which you don't need without nuclear, and you only have 16 ms before it start dropping under 60 UPS, so that's a significant amount ( 12.5% ), and that's not counting the entities themselves, just the heat manager otherwise 12ms/16ms is 75% just for the power source :/

Re: Is nuclear power really bad for megabases?

Posted: Tue Sep 12, 2023 2:00 pm
by Tertius
Well, I admit I have a personal flaw. I judge too fast. I see one image, one statistics, and say to myself: "this is it", than it is this forever. Even if I missed some aspect or looked at the wrong spot. I apologize.

I tested now my nuclear reactor again, on an empty sandbox map, similar to the OP. 1008 reactors, 161 GW. As far as I see, I didn't get vastly different results (surprise, surprise). It's way faster on my PC, but that's probably I have a faster PC - OP is a Intel i7-9750H (12) @ 4.5GHz, mine is a i5-13600K. Passmark benchmark score 10984 versus 38359 (1:3.5) . OP got 12.6 ms update time, I got 6.5 ms update time (1:1.93).
Screenshot 2023-09-12 153349.png
Screenshot 2023-09-12 153349.png (52.38 KiB) Viewed 1607 times
I guess I dismissed the impact of a nuclear power plant, because I didn't test it separately but in combination with the whole megabase also on the map. The relevant entity time usage times are these:
Screenshot 2023-09-12 153422.png
Screenshot 2023-09-12 153422.png (114.87 KiB) Viewed 1607 times
I saw boilers and generators only about 2 ms, and this was for me probably the sign that it is just a little bit more in addition to the whole base - after all, a 161 GW nuclear power plant is a quite huge installation. Every item takes its performance toll, and I found it justifiable to have this in addition.

And my 5k SPM base has a peak energy demand of about 34.4 GW, so 161 GW seem far, far away. I'm actually using a combination of solar and nuclear energy. I have 348k panels + 292k accumulators for a peak of 21 GW of energy daytime, and the demand that goes beyond that is satisfied by a (steam buffered) plant of 84 reactors, peak production 13.5 GW.

Ah, and revisiting the debugging values, I now remember why I said: "negligible". This is the entity timing info of the whole 5k SPM base:
Screenshot 2023-09-12 155505.png
Screenshot 2023-09-12 155505.png (202.88 KiB) Viewed 1607 times
Boiler and Generator 0.148 ms and 0.182 ms - that's really negligible in comparison to the other components. In my opinion.

This is the whole timing of the base:
Screenshot 2023-09-12 155651.png
Screenshot 2023-09-12 155651.png (165.46 KiB) Viewed 1607 times

Re: Is nuclear power really bad for megabases?

Posted: Thu Sep 14, 2023 10:01 pm
by mmmPI
Tertius wrote: ↑
Tue Sep 12, 2023 2:00 pm
 Ah, and revisiting the debugging values, I now remember why I said: "negligible". This is the entity timing info of the whole 5k SPM base:

Boiler and Generator 0.148 ms and 0.182 ms - that's really negligible in comparison to the other components. In my opinion.

This is the whole timing of the base:
I somewhat agree with your conclusion upon seeing the results. Your computer only need 6ms to produce 160 GW in game with the blueprint you used. For the mentionned factory that produce 5K SPM, you only use 13 GW of nuclear power, during peak.

So a rough estimation would be (13/160)*6.4 = 0.52 ms maybe too rough , but around 3.25% of the 16 ms available ; or 3.5% of the 14.6 ms used ) So very few ms taken in the real base, disregarding the upscaled testbench.

With the Boiler and Generator accounting for more than half of it (0.148+0.182)=0.33 ms /0.52 = 63% .

0.33 ms though is also roughly 25% of the time taken for mining drill and it can get hard to gain 25% UPS efficiency on mining drill blueprint depending on the efficiency of your current mining drills blueprint, that's also the same "price" as the assembling machine. (sure without inserter they are useless and those cost a lot)

It's negligible for your current setup ( computer + blueprints + share of nuclear in base ), but it's still relevant to compare it with other components i think, even though inserters and mining drill are the biggest source of optimization, switching from nuclear to solar would still be at some level of optimization a choice that become easier to put in practice rather than perfecting the previous 2. I guess there is a correlation between the "badness" and how serious one is when using the word "megabase" where the deeper you go on UPS-optimization, the worse nuclear appear eventually. ( equivalent to rolling back several iterations of improvements ).
All times are UTC
Page 2 of 2

Powered by phpBB® Forum Software © phpBB Limited