## Nuclear reactor strange power production

Don't know how to use a machine? Looking for efficient setups? Stuck in a mission?
mmmPI
Smart Inserter
Posts: 2552
Joined: Mon Jun 20, 2016 6:10 pm
Contact:

### Re: Nuclear reactor strange power production: bug?

Good story

50,000 / 485 = 103.092783505 , it's worse than just 103.09 !

But one doesnt need to use this number, it's easier to math the heat exchanger (10MW) for the reactor(40MW+max adjacency). And the water for the turbine and offshore pump considering 1=1 conversion water steam, 60/s and 1200/s. Then make sure you provide excess water

It leaves a discrepency between turbine(5.82 MW) and reactor (40MW+adjacency) which could be solved for large number but it's not too bad to have a little too many turbines (the two engineers need to agree on that).

In game when you consume less electricity than what reactor can sustain, steam will accumulate in those excess turbines and it provide the ability for the nuclear power plant to handle little spike of a small %. At least that's my routine when i build nuclear plant with modded things and i don't want to redo too much math

If the plan is tileable reactor (to infinity) one is bound to always have excess turbine and heat exchanger anyway before it is complete as reactor efficiency keep increasing as the share of the side reactor on the total diminshes. So it wouldn't be the most compact possible.

I had not checked all the math and the plan of expansion when i removed the turbines. So it's very well possible i did something wrong regarding the number of reactor.

Tertius
Filter Inserter
Posts: 616
Joined: Fri Mar 19, 2021 5:58 pm
Contact:

### Re: Nuclear reactor strange power production: bug?

The calculating for my own reactor stuff goes like this.

Prerequisites by the game:
- offshore pump: 1200/s
- heat exchanger: 103/s; 10 MW
- steam turbine: 60/s
- reactor: 160 MW (assume infinitely tiled; ignore the existence of borders)

- width of one heat exchanger row in tiles: 1/2 per heat pipe, 2 per heat exchanger, 1/2 per steam pipe = 3 [highest entity density]
- this minimum width demands this row structure: heat pipe - heat exchanger - steam pipe - heat exchanger - heat pipe - and so on.
- so the minimum tileable width of multiple heat exchanger rows is 6 (one regular row and one mirrored).
- minimum tileable width with reactors plus highest density heat exchanger rows: reactor width is 5. Least common multiple: 5 * 6 = 30.
- so the minimum size of a tile is 6 reactors wide, which can have up to 10 heat exchanger rows.
- The maximum power output of such a tile is 6 reactors * 160 MW per reactor = 960. This is per side, because we can mirror the setup to get a 2xN row setup for 300% neighbor bonus, so the complete output of one tile is 2*960 MW = 1920 MW.

I want 1 offshore pump per heat exchanger row and every heat exchanger supplied with full water.
So maximum number of heat exchangers per row = floor(1200/s / 103/s) = 11

How far do we get with 11 heat exchangers in 10 rows (110 heat exchangers)?
Maximum power output of one heat exchanger field is 10 MW * 11 * 10 = 1100 MW.
This is slightly above 960 MW we produce for this field. So it fits! We need only 96 heat exchangers to consume 960 MW.

So how to arrange these?
The highest density is 10 rows. We can do 8 rows with 10 heat exchangers plus 2 rows with 8 heat exchangers.
We can also do 9 rows. 6 rows with 11 heat exchangers plus 3 rows with 10 heat exchangers. However, this will give us only 2 tiles space, not really useful, and adds vast asymmetry because the border rows get 1 and not 1/2 of the heat resp. steam capacity with their heat/steam pipes.
So I chose 8 rows with 10 heat exchangers plus 2 rows with 8 heat exchangers. The hole of the 4 missing heat exchanges can be used to place a roboport.

This is what you see in the picture I linked.

Now you need to find a place/space for the offshore pumps. They can be placed near the reactors or on the far side (both valid options).

Then there are the steam turbines.
One side needs at least ceil(960 MW / 5.82 MW) = 165 steam turbines.
One steam turbine is 3 tiles wide, so the highest steam turbine density is 30 / 3 = 10 rows of steam turbines.
Our 6-reactor-tile has 5 steam pipe outputs. This conveniently matches the maximum of 10 rows. We can either split each steam pipe and connect 2 half length steam turbine lines, or fully consume the steam of our 5 steam pipes with 5 steam turbine lines (you see the latter version in the picture I linked).
The steam pipes with the 10 heat exchangers produce 2 * 10 * 103 = 2060 steam/s, so you need ceil(2060/60) = 35 steam turbines.
The steam pipes with the 8 heat exchangers produce 2 * 8 * 103 = 1648 steam/s, so you need ceil(1648/60) = 28 steam turbines.
This is what you see in the picture I linked.

A different approach is to split the steam pipes, so the steam turbine lines can be half length and higher density is achieved.
Unfortunately, 35 is an odd number, so we need 18 turbines per row for the rows behind the 10 heat exchangers.
So we get 8 steam turbine rows with 18 turbines each (these are behind the 10 heat exchangers)
and 2 steam turbine rows with 14 turbines each (these are behind the 8 heat exchangers).
Added up these are 8 * 18 + 2 * 14 = 172 steam turbines. This is a valid number, however a bit of slack as well (it's 172 - 165 = 7 turbines too much).

Now personal preference comes into play. Up to this point, we are in the corset the game numbers puts upon us. Designing an actual solution is now where designs will start to differ vastly from player to player.
It's also necessary to squeeze power poles into the steam turbine field. And roboports.

---

If you don't mind, a short presentation of my solution. Personally, I wanted to get rid of these 7 slack turbines.
According to my personal preference, I also wanted to have steam buffer and implement wasteless nuclear fuel cell supply. So I needed to add storage tanks as steam buffers. After much experimenting, I finally added them to the end of the steam turbine rows, so reactors → heat exchangers → steam split → steam pumps → steam turbines → storage tanks. (if you care to look, it's here).
The buffer at the end connect all steam turbine rows, so it's possible steam flows from one row to other rows, so I can actually get rid of some of these 7 slack turbines. If you look at my solution, I actually used 166 steam turbines per side. That's only possible because all rows are connected, so all steam is distributed to everywhere in the end)

astroshak
Filter Inserter
Posts: 596
Joined: Thu May 10, 2018 9:59 am
Contact:

### Re: Nuclear reactor strange power production: bug?

Tertius wrote:
Wed Nov 15, 2023 5:45 pm
- minimum tileable width with reactors plus highest density heat exchanger rows: reactor width is 5. Least common multiple: 5 * 6 = 30.
- so the minimum size of a tile is 6 reactors wide, which can have up to 10 heat exchanger rows.
If your design therefore is 2x6, does that mean that every 2x6 you have a doubled up Heat Pipe or Pipe (depending on which are outside of the BP)?

Tertius
Filter Inserter
Posts: 616
Joined: Fri Mar 19, 2021 5:58 pm
Contact:

### Re: Nuclear reactor strange power production: bug?

astroshak wrote:
Fri Nov 17, 2023 2:29 am
If your design therefore is 2x6, does that mean that every 2x6 you have a doubled up Heat Pipe or Pipe (depending on which are outside of the BP)?
There are no double heat pipes or pipes. On the contrary, each pipe is accessed from both sides.
According to how I see it, you either need 9 or 10 rows of heat exchangers to not require more than 11 exchangers per row to satisfy the maximum throughput of 1200/s for one offshore pump for the row.

If you do use 10 rows, you will completely fill all space with heat exchangers, heat pipes and steam pipes, and all the pipes are accessed from both sides, i. e. by up to 22 heat exchangers.
The general spacing for 1 blueprint tile (6 reactor width) looks like this:
Screenshot 2023-11-17 160445.png (522.78 KiB) Viewed 1428 times
The top and bottom heat pipe in the blueprint will overlap and merge, if you place multiple blueprints next to each other to tile the blueprint.

If you do use 9 rows, one line of heat pipes and one line of steam pipes is accessed by one side only, and you have a gap of 2 tiles. The other lines are accessed from both sides. As far a I see it, this gap cannot be used in a meaningful way, so I propose to go for the full 10 row field that's a little shorter, so the whole plant can be 2x3=6 tiles narrower. A shorter heat exchanger line also means you have less maximum throughput at the end of the heat exchanger line, so you can have longer steam pipes behind it without pumps.

Belter
Fast Inserter
Posts: 203
Joined: Sun Mar 14, 2021 12:38 pm
Contact:

### Re: Nuclear reactor strange power production: bug?

Tertius wrote:
Tue Nov 14, 2023 11:09 pm
There were two different engineers. One designed the heat exchanger and the other one designed the steam turbine. Both agreed upon working with 500°C steam.
Yeah, they should have talked a bit more..

Also TY for the long post! I had the same conlusions but the merged steam pipes did not want to work so tried to force separated one. Will come back to that... In that case a column of 11 heat exchangers work.

So after some proper calculations I had to realize that a column of 11 heat exchangers are just not enough, need 12 - but that needs 1236 / sec water. So recycling water! At the end of a row of 12, the missing 36 / sec water is coming from the row of 11 where there is a surplus of 67. I also re-use the remaining steam, from 2 rows I put it to 5. Producing 9.4 GW long term but the reactors are going up to 99X degrees. Wish Factorio would show electricy w/3 digits. 9.44 GW should be the production.

This is the best I can do w/Robort coverage and w/o pumps, power mgmt, solar panels etc. In the savegame there are BPs for building (2 steps) and the gadgets for testing. The building block is 2x10 reactors. I've managed to improve the look also. By removing the Roborts it could be really nice!
Attachments
202310 MapEdit-Nuclear4.zip
Layout-recyclewater.png (1.81 MiB) Viewed 1400 times

Tertius
Filter Inserter
Posts: 616
Joined: Fri Mar 19, 2021 5:58 pm
Contact:

### Re: Nuclear reactor strange power production: bug?

Belter wrote:
Fri Nov 17, 2023 11:12 pm

Also TY for the long post! I had the same conlusions but the merged steam pipes did not want to work so tried to force separated one.!
There is one thing left that needs to be mentioned: fluid throughput and pipe length. The higher the desired throughput, the shorter the distance this throughput can be kept up.

Look at these:
Screenshot 2023-11-18 154317.png (1.38 MiB) Viewed 1360 times
The top line will not work, because 20 heat exchanger produce 20 * 103 = 2060 steam/s. The distance to transfer this throughput is 4 resp. 5 pipe segments. According to the wiki, a throughput of 2060 cannot be transferred over a distance of 4 segments without intermediate pump. 3 segments is max. So the steam turbine line to the right cannot receive the full 2060/s.

The middle line will work, because it splits the steam output to 2*7 and 2*3 heat exchangers, producing one output with with 2*7*103 = 1442 steam/s and another with 2*3*103 = 618 steam/s. These can be transferred over a distance of 7 resp. 300 segments. Both merge after 4 resp. 5 pipes and split at the same time, so the maximum throughput of one pipe never exceeds its capacity.

The bottom line will also work, because it increases throughput with pumps. All three pumps are required. Build this, inspect the pump throughput and compare with the maximum pipe length from the wiki to understand why high fluid throughput becomes tricky above approx. 1400/s.

Belter
Fast Inserter
Posts: 203
Joined: Sun Mar 14, 2021 12:38 pm
Contact:

### Re: Nuclear reactor strange power production

I've added tons of pumps, just another wip version for the merged steam pipes. 10 x 16 heat exchangers are the perfet ratio for 10 reactors, could not stand...

2x2 tiles can be saved for sure (1 at the well, 1 at the heat exchangers I added for pumps but removed), even more maybe.

Upd: savegame was wrong, fixed. BP added as TXT (cannot fit to the 60.000 char limit here :O )
Attachments
bp-nuclear-v8.txt
202310 MapEdit-Nuclear8.zip
v8-zoom.png (1.14 MiB) Viewed 1285 times
v8.png (18.64 KiB) Viewed 1285 times

Belter
Fast Inserter
Posts: 203
Joined: Sun Mar 14, 2021 12:38 pm
Contact:

### Re: Nuclear reactor strange power production

Anybody has any suggestions how to measure power more accurately?

3 blocks of 2x10 reactors can sustain 9.4GW so it looks 100% but as the temp of the reactors are 99X degrees, there might be some minor bottlenecks...

The polished version:
- saved 16 tiles from height
- optimized cables, removed lot of poles
- less RoboPorts but still fully covered
- two radars per block to fully cover also
- removed a full row of pumps, probably more can be cut from the turbines (won't save space though)
- perfect # of heat exchangers
- maybe 1 extra turbine - test showed it is required

This is the best I can do so far. Perfect would be:
- no pumps at all but still 100% power
- first BP can build the full electrid grid (can be done only if I add 1+1 tiles) - here North and South separated until 2nd BP is applied due to water wells
- at the top/bottom I'm merging the remaining steam. Ran into some messy situations, sure this is not ideal for UPS

Out of scope: fuel saving, starter solar cells etc.
Attachments
202310 MapEdit-NuclearA.zip
bp-nuclear-v10.txt
v10-zoom.png (2.55 MiB) Viewed 1167 times
v10-map.png (29.81 KiB) Viewed 1167 times

mmmPI
Smart Inserter
Posts: 2552
Joined: Mon Jun 20, 2016 6:10 pm
Contact:

### Re: Nuclear reactor strange power production

Belter wrote:
Tue Nov 28, 2023 9:14 am
Anybody has any suggestions how to measure power more accurately?

3 blocks of 2x10 reactors can sustain 9.4GW so it looks 100% but as the temp of the reactors are 99X degrees, there might be some minor bottlenecks...
I'm going to look at the savegame, but one way to measure bottlenecks could be to measure fuel consumption. On its own its not enough, but it allows to measure if you consume 9.412432 GW worth of fuel, that one could be very precise.

Now maybe the power plant is producing the exact same, or maybe 9.400321 GW which from the fuel consumption measurement alone one couldn't tell.

But this can be tested against around 20K beacons. If you try to "draw" as much power as you consume fuel, are the entities consuming power behaving well or not ? As your earlier test seem to suggest beacons lacking power even just a little bit would empty their buffer and have very visible impact on energy graph. It is also possible to use the entity from the /editor and set it to consume 9.4 GW or 9.350 GW, and then just add a few beacons or other entity with less consumption for more precise measurement.

Maybe just using the energy consumer from the editor is enough now that i think of it.

Edit : there are mods too maybe this one or those listed at the bottom of its description https://mods.factorio.com/mod/Circuit_P ... Combinator

Belter
Fast Inserter
Posts: 203
Joined: Sun Mar 14, 2021 12:38 pm
Contact:

### Re: Nuclear reactor strange power production

mmmPI wrote:
Tue Nov 28, 2023 10:12 am
Edit : there are mods too maybe this one or those listed at the bottom of its description https://mods.factorio.com/mod/Circuit_P ... Combinator
Yeah, thank you!
Nixie.png (196.23 KiB) Viewed 1131 times

Belter
Fast Inserter
Posts: 203
Joined: Sun Mar 14, 2021 12:38 pm
Contact:

### Re: Nuclear reactor strange power production

OK added some pumps. The output is... well. 9.444 GW. No. 9.438 GW.... Noooooooo
Attachments
v11-Nixies.gif (3.69 MiB) Viewed 1120 times
202310 MapEdit-NuclearB.zip

mmmPI
Smart Inserter
Posts: 2552
Joined: Mon Jun 20, 2016 6:10 pm
Contact:

### Re: Nuclear reactor strange power production

Belter wrote:
Tue Nov 28, 2023 9:32 pm
9.340307 GW
I had not thought of Nixie tube and i'm really confused because adding them to the save game where i was measuring with beacons trying to reach thermal equilibrium, just adding the mod, seemed to have changed the power production when i thought measurements were finished and only publication was needed so here are my findings without nice display for now i will need to investigate another day for that:

To me the savegame provided can power closer to 19657 additionnal beacons, and then the reactors temperature are somewhat stable, losing 0.1°C when the inserters for fuel are working and almost regaining it over time, the drift is very very low less than 0.1°C cooling off per hour. With 19658 the temperature visibly decrease, with 19656, it visibly increases.

That woud mean between 9.434 880 GW and 9.435 840 GW as "additionnal" available power, but the power plant itself is consuming some variable amount which makes it difficult to say precisely how much total power is produced when at max capacity but it is a little more as it include the electricity of its radar and robotports...

With 19657 beacons and the radars and inserters and everything, it shows it can sustain almost 9.439 986 GW total electricity production as measured with the mods. As at this level of draw, only the additionnal fluctuations due to inserter for fuel are responsible or the variation of temperature very slightly trending down. ( possibly roboports too ).

(also i added 4 combinators which represent 0.000004 GW or 4kW and are part of the imprecisions)

Given that the theoric output of reactors is ( 56*4*40)+(4*3*40) = 9.440000 GW and that the mod rate calculator approximate it as 157.3*60 = 9.438000 GW the efficency is between 99.99985 % and 100.02104% which i think shows no problem in that regard, all the reactor burning capacity can be turned into elecricity (and even more as it possible to "cool" the reactors by drawing too much power).
202310 MapEdit-NuclearA1.zip

Belter
Fast Inserter
Posts: 203
Joined: Sun Mar 14, 2021 12:38 pm
Contact:

### Re: Nuclear reactor strange power production

mmmPI wrote:
Tue Nov 28, 2023 11:01 pm
Given that the theoric output of reactors is ( 56*4*40)+(4*3*40) = 9.440000 GW and that the mod rate calculator approximate it as 157.3*60 = 9.438000 GW the efficency is between 99.99985 % and 100.02104% which i think shows no problem in that regard, all the reactor burning capacity can be turned into elecricity (and even more as it possible to "cool" the reactors by drawing too much power).
TY again, this is very useful. I had some issues also about different power scenarios, when altering the the configuration. What I've found is stability comes if
- the reactors are not increaing their temp (so not tending to 1000)
- the last hear pipe is 50X hot
- 10 hours is a safe bet to stabilize

The last savegame above has some extra pumps mainly at the end of the steam turbines. If you check the water consumption or the first steam pumps, still see some chaos...

Belter
Fast Inserter
Posts: 203
Joined: Sun Mar 14, 2021 12:38 pm
Contact:

### Re: Nuclear reactor strange power production

v12
- see the attched savegame
- cut another 4 tiles from the height by removing a bunch of pumps
- I cannot compress this further - and still keep robort coverage.
- there is some chaos still going on, power is fluctiating between 9.43 and 9.44 GW but temperatures are stable
v12-nixie.png (1.46 MiB) Viewed 951 times
Attachments
v12-map.png (5.83 KiB) Viewed 951 times
v12-zoom.png (730.99 KiB) Viewed 951 times
202310 MapEdit-NuclearC.zip

Tertius
Filter Inserter
Posts: 616
Joined: Fri Mar 19, 2021 5:58 pm
Contact:

### Re: Nuclear reactor strange power production

It's definitely possible to squeeze every single Watt out of bigger power plants.
I did it with the setup I referenced earlier (not the single image - the thread I referenced). Connecting the power measurement combinator to a bunch of nixie tubes and waiting for temperatures to stabilize yields exactly the expected output and it is 100% stable.
For a 120 reactor setup (tile size 2x6)
full roboport coverage (it's too dense, but it's the most sparse the tile size allows)
power pole, roboport and turbine placement <-> 1 week headache

It's my most polished setup. I'm currently working on pump placement near the reactors, but it doesn't have priority. Having them outside decreases reactor temperature and increases reaction speed, which is somewhat important for me since I always add circuit control.
In this state, reactor temperatures range from 745.56°C (the border reactors) to 754.25°C (the center reactors).

Screenshot 2023-12-03 232918.png (659.38 KiB) Viewed 881 times
Screenshot 2023-12-03 232955.jpg (2.69 MiB) Viewed 881 times
Screenshot 2023-12-03 233245.jpg (810.69 KiB) Viewed 881 times
Screenshot 2023-12-03 233423.jpg (215.2 KiB) Viewed 881 times

Belter
Fast Inserter
Posts: 203
Joined: Sun Mar 14, 2021 12:38 pm
Contact:

### Re: Nuclear reactor strange power production

Tertius wrote:
Sun Dec 03, 2023 10:47 pm
It's my most polished setup.
Looks great, if you have a savegame or BP that would be nice also!

Tertius
Filter Inserter
Posts: 616
Joined: Fri Mar 19, 2021 5:58 pm
Contact:

### Re: Nuclear reactor strange power production

I posted blueprints and instructions here: viewtopic.php?f=208&t=103591

Savegame of my sandbox: (it's the huge reactor left to the mall at the left side. Everything else is something else and irrelevant. Mods are irrelevant as well; for measurement and visualization keep Nixie tubes and the power meter combinator.)
Attachments
Sandbox Entwicklung.zip