Page 1 of 1

New Reactor design - PLS check my construction

Posted: Wed Nov 06, 2019 8:47 pm
by Cyrilis
First post here!

OK, my reactor is desigend to produce 2.05 GW, but he only brings 1.7 to 1.8 GW. Where is the problem?

This is my baby:

Image

Blueprint can be found here:

https://pastebin.com/K9bY6UUa

Idea behind the design:
I wanted something close to the Neighbourhood efficience of a 2N row design, so I came up with this "double sixpack + 2" layout. While a 14 reactor 2N setup has a single core equivalent of 52, mine has 50, so close enough for me. Also th 2N rows are incredibly wide and I wanted something more "compactish".

Features:
- 14 reactors
- 200 heat exchangers
- 350 turbines 18 pipes of water input
- 154 Tanks for Steam Storage
- simple circuit controlled reactor
- Steam-meter
- Kovarex on site
- recycling on site
- fuel celll assembling on site
- non liquid input: iron plates and 238-U

given my math is correct, the reactor should provide 2.0something GW, but for a reason I can't find, he stalls at 1.7 to 1.8 GW. I have the feeling there is a throughput issue somewhere with either stem or water, but I was not able to locate it, nor did I manage to change the design in a way that results in 2 GW power output.

Could someone pls have a look at the block and help me to identify the problem and fix it?

thanks a lot

Re: New Reactor design - PLS check my construction

Posted: Thu Nov 07, 2019 4:29 am
by jcranmer
You have water and steam throughput issues. You might have heat throughput as well, but I don't know the heat transfer equations in Factorio well enough to be sure.

Each set of 12 heat exchangers consumes ~1237 units of fluid/sec, an offshore pump provides merely 1200 fluid/sec. Water is therefore throttling you to a touch over 1.94GW.

Steam is throttling you worse than that. Pipes have a max throughput that's dependent on length, and when you're pushing well above 1200 fluid/sec, that maximum length is very short. If you look from the heat exchangers by the iron input line, you've got a total of 1800 steam/sec dumping into that junction, which is 7 tiles from the next pump. But 7 tiles only lets you push 1500 steam/sec. It's even worse because the tank in the pipe reduces max flow rate. To top it all off, the fluid is going to have more slowly slosh itself around the tanks to actually reach the steam turbines. I'm actually surprised you're hitting 1.7-1.8GW: I suspect that's not a true steady state production rate but instead slowly draining the accumulated steam, since the tank-to-steam-turbine paths are much shorter.

Re: New Reactor design - PLS check my construction

Posted: Thu Nov 07, 2019 11:37 am
by astroshak
Unless your Kovarex process is horribly inefficient, I’d not worry about steam storage. Uranium mines can power nuclear power plants for a *very* long time. I have a hard time imagining depleting even a 2M uranium patch using both nuclear power and green ammo based defenses.

Elimination of steam tanks from your design would allow the turbines to be closer to the heat exchangers, reducing the pipe throttling that jcranmer speaks of, as well. On those times when I did add steam storage (and the requisite wiring to control the inserters) I tend to put it after the turbines, rather than before. Although, figuring out how to pump the steam in and out has eluded me.

Re: New Reactor design - PLS check my construction

Posted: Thu Nov 07, 2019 12:47 pm
by mrvn
You can put a steam tank after the turbines or between the last and second last. Steam turbines have a high flow rate and will fill the tank when they don't use the steam. And when the heat exchangers don't produce enough steam it will flow back from the tanks.

Re: New Reactor design - PLS check my construction

Posted: Thu Nov 07, 2019 3:06 pm
by Serenity
I found that having junctions of any kind is bad and result in certain parts not running for no apparent reason. It's better to have contained sub-systems like 1 pump -> 12 heat exchangers -> 20 turbines. That will result in everything running as much as possible.
You can still have tanks in the system like that if you really want, but just give each such unit a few tanks and don't connect all the tanks to each other

Re: New Reactor design - PLS check my construction

Posted: Thu Nov 07, 2019 3:49 pm
by mrvn
Serenity wrote:
Thu Nov 07, 2019 3:06 pm
I found that having junctions of any kind is bad and result in certain parts not running for no apparent reason. It's better to have contained sub-systems like 1 pump -> 12 heat exchangers -> 20 turbines. That will result in everything running as much as possible.
You can still have tanks in the system like that if you really want, but just give each such unit a few tanks and don't connect all the tanks to each other
In a junction the connections are processed in a fixed order. That means you either have at least 2 inputs or two outputs to the junction that compete against each other. One input or output will be favoured. It's no 50:50 split.

Also consider a X junction with two inputs and two outputs. The first input supplies 70 fluid. The second input tops that off to 80 fluids leaving 80 fluids behind. Then the first output takes 50 fluid and the second is left with 30 fluids. You don't get the full 160 fluids the two inputs could supply. The throughput is limited to that of a single pipe.

Re: New Reactor design - PLS check my construction

Posted: Thu Nov 07, 2019 3:52 pm
by Cyrilis
Hey guys, thanks a lot for the feedback... It may take a while until I come up with a design that includes your suggestions. I'll update this as soon as I can find the time to tinker around with the steam issue.

Re: New Reactor design - PLS check my construction

Posted: Thu Nov 07, 2019 5:16 pm
by astroshak
With the understanding that 2N reactor setups are the most efficient and most scalable I took a 2x4 setup and put enough stuff on each side to provide 640 MW of power. I then made that into a blueprint taking only two of the reactors and all the stuff on one side. I also made sure to include room for roboports so I could simply issue a build order.

With all of this on landfill (by design, this allows for a good placement of the offshore pumps) it actually takes two build orders to build, but the only time I run into any issues is when there happens to be pre-existing land where I need the offshore pumps. I tend to look for long lakes. To help make finding good lakes easier, I have the offshore pumps right after the heat exchangers.

My design is slightly underpowered, due to the throttling effect of pipes, and that I only have 60 heat exchangers per side rather than the 64 needed. 65 is doable but only at the expense of lengthening the already long blueprint.

I chose 640 MW per set of 4 reactors because with the exception of the first and last pairs of reactors they are all producing 160 MW due to neighbor bonus and I wanted to get it all I o e. p not two or three.

Just some food for thought.

Re: New Reactor design - PLS check my construction

Posted: Fri Nov 08, 2019 8:27 am
by gGeorg
Cute creation. That layout for 3 Neigbor bonus for 2 cores is nice touch.

In the new tooltip you might easy see, Heat exchanger consumes 103 water and creates 103 steam.
As long as 12 x 103 = 1236 and shore pump gives 1200 at max here is the first shortage.
You need more offshore pump. Also keep in mind you better need surplus for fluid slowing down in pipes.

Piping is quite a challenge. I would recomend you get an info about pump attached to tank storage bonuse, this will make your creation much sleak design.

Also reconsider number of steam storages. Raw info from Factorio cheat sheat is misleading. Get noticed important fact : Each heatpipe, each Heat Exchager and each Core has a capacity 0,5MJ. Its quite significant diference for design. Cheat cheet recommends 40 tanks for four core setup but My Latest four leafe design uses ONLY 8 steam storages and is able to store full 96GJ energy of the cycle.
Based on this, my guess is you are using about 100 steam storages but 20 would be enough. You would better make precise counting by yourself.

Looking froward your second version.

Re: New Reactor design - PLS check my construction

Posted: Mon Nov 11, 2019 11:31 am
by mrvn
Yesterday my steam engines power run low leading to some brownouts. So I build a nuclear reactor from scratch again. A very simple 2x4 design going west-east with 14 heat exchangers (a water pump at each end), pump, underground pipe (to make space for the offshore pump), pump and 24 steam turbines per reactor going up and down. The ends are capped with 2 steam tanks each, so 50k steam buffer per reactor. Reactors are getting fueled when steam < 50%. So far all boring. But it gave me some thoughts.

A row of heat exchangers with a steam pipe on one side and heat pipe on the other is 4m wide. A reactor is 5m wide. In my design it made it a double row of heat pipes for extra energy storage. But what if you put the heat exchangers closer together? That won't leave space for an offshore pump at the reactors end of the heat exchangers but you wouldn't need it I think. Instead of 4 rows of 14 heat exchanges you have 5 rows of 11 (one less), or 12 (more that you have heat and water so one runs part time). And then 5 rows of 19 or 20 steam turbines.

Anyone tried something like that? Would reduce the height of the reactor a fair bit.

Re: New Reactor design - PLS check my construction

Posted: Mon Nov 11, 2019 12:56 pm
by astroshak
If you plan for 2xN reactors, don’t figure each Reactor is going to be producing 120 MW apiece. Only the Reactors at the ends of the 2xN line will, every other Reactor will produce 160 MW apiece. That tells me that the BP should accommodate 640 MW production per 2x2 to be truly infinitely scalable.

Reactors are 5x5. Pipe, Heat Exchanger, Heat Conduit are, in total, 4 wide. That means for a 2x4 Reactor BP, you have 20 spaces wide to put 640 MW on each long side of the Reactor block, or five rows of steam generation. 640MW/5rows=128MW per row. According to the Cheat Sheet and the Wiki, Heat Exchangers consume 10 MW apiece, meaning you need 13 per row. That idea of putting Offshore Pumps at both ends of each row is an intriguing one that did not occur to me (my BP uses 12 HE per row, part of the reason its slightly underpowered). As the Steam Turbines put out 5.82 MW of power, you need 22 Steam Turbines per row to provide the 128 MW (think mine is 20 per row, the other reason its underpowered. Yes, I need to fix my BP).

Make it as long as you need to for the Offhore Pumps and the Roboports.

Due to said length, I made mine by half : 2x2 Reactors, but still 20 squares wide. Its kind of in the following shape, where X is everything else, R is 2x2 Reactors :
XXXXXXX
XXXXXXXR
Note that it is longer than the number of X’s I’ve put here for representational purposes.

Re: New Reactor design - PLS check my construction

Posted: Thu Nov 14, 2019 10:54 am
by mrvn
astroshak wrote:
Mon Nov 11, 2019 12:56 pm
If you plan for 2xN reactors, don’t figure each Reactor is going to be producing 120 MW apiece. Only the Reactors at the ends of the 2xN line will, every other Reactor will produce 160 MW apiece. That tells me that the BP should accommodate 640 MW production per 2x2 to be truly infinitely scalable.

Reactors are 5x5. Pipe, Heat Exchanger, Heat Conduit are, in total, 4 wide. That means for a 2x4 Reactor BP, you have 20 spaces wide to put 640 MW on each long side of the Reactor block, or five rows of steam generation. 640MW/5rows=128MW per row. According to the Cheat Sheet and the Wiki, Heat Exchangers consume 10 MW apiece, meaning you need 13 per row. That idea of putting Offshore Pumps at both ends of each row is an intriguing one that did not occur to me (my BP uses 12 HE per row, part of the reason its slightly underpowered). As the Steam Turbines put out 5.82 MW of power, you need 22 Steam Turbines per row to provide the 128 MW (think mine is 20 per row, the other reason its underpowered. Yes, I need to fix my BP).

Make it as long as you need to for the Offhore Pumps and the Roboports.

Due to said length, I made mine by half : 2x2 Reactors, but still 20 squares wide. Its kind of in the following shape, where X is everything else, R is 2x2 Reactors :
XXXXXXX
XXXXXXXR
Note that it is longer than the number of X’s I’ve put here for representational purposes.
I thought 2x4 reactor would be plenty. But now I'm already at 50% usage at night. Now I have the choice to simply copy&paste a second 2x4 reactor or go for 2xN and expand it as needed. As you say, when you want to make it expandable then you have to figure for 640MW instead of 560MW.

Unless I forget I will post my blueprint and a screenshot tonight.