Big Reactors

[pizzawolf14]

OK, I've seen a ton of threads an big reactors lately, and people have made some brilliant spreadsheets for the mod. I want to know the best reactor/turbine combo one can make as far as efficiency to output ratio. Assume infinite resources.

Runners take your marks
Get set
*Boom*

 

[GreenZombie]

*Sigh*

Similar to passive reactors, actively cooled big reactors can be made arbitrarily efficient at any size. I havn't worked out the math yet, but it looks like each fuel rod block in an actively cooled Big Reactor will contribute ~200mB/t of steam.
Don't use any material other than cryotheum as it is the best moderator, best heat conductor, (and, for passive reactors, 2nd best absorber).

Basically, assuming infinite resources, it should always be possible to beat the effeciency of any particular design, by making a slightly larger one.

Well, unless RF/volume is the efficiency metric. What DO you mean by efficiency? RF/mB of fuel? RF/m³? something else per something else entirely?

 

[McJty]

I actually have a similar question but this time not assuming infinite resources
I currently have a passive cooled 7x7x7 reactor which produces about 16000 RF/tick. It uses liquid ender and it has 9 control rods (arranged in a cross). For now it is the most efficient design that I was able to come up with.
I'm playing survival so I don't have infinite resources. Given the 7x7x7 reactor I have now, what would be the best turbine setup be to get the most out of that reactor when I switch it to active cooling?

Thanks!

 

[GreenZombie]

The first thing you can do is rearrange your control rods so that they are cardinally aligned. That many fuel rods (9*5) should be yielding 22500RF/t. Arranged as an X you are getting no cross rod fertilization.

 

[McJty]

Sorry can you explain that? What exactly do you mean by 'cardinally aligned'?

 

[GreenZombie]

Each fuel rod in a big reactor sends 4 radation packets each tick - one directly north, one east, one west, one south. Each radiation packet travels 4 blocks, or until it hits the edge of the reactor. Each packet is given a radiation intensity (which starts at a base number, but is boosted by the emitting rods fertilization), and a "hardness" which specifies what percentage of the radiation is hard. This starts at 20%, and scales up as the fuel rod temp goes up. At 1000degrees about 40% of the radiation emitted is hard.

If the radiation packet intersects another rod, its slow radiation boosts the rods fertilization (meaning that rod will fire off a boosted intensity radiation packet). If the radiation packet intersects a coolant block, a % of the slow radiation is absorbed, (and a % of that is converted to RF), and a % of the hard radiation is moderated to slow radiation (using the absorbtion, heat efficienty and moderation properties of the coolant respectively).

This means that efficient Big Reactor designs (a) ensure that fuel rods are lined up n/s/e/w of each other, (b) if they are separated from each other, they are separated by a material that has good moderation properties, and (c) there is at least one, but hopefully more, blocks of a coolant with good absorption properties between each rod the edge of the reactor.

Conversely, blocks that are not directly n,s,e or w of a fuel rod, do not play any part in the above calculations, and as such can be filled with glass, air, or the material of your choice.

It also means that, if you can keep the rod temperature near 1000 degrees where most of the radiation is not hard, there is no point to separating fuel rods: you can densely pack them as attempting to moderate them by spacing the core with moderating material simply means that so much less radiation arrives at the next fuel rod to fertilize it.

 

[b0bst3r]

Yet in my tests an X formation gives out more power and better loss than a + formation.

In a 7x7x7 reactor using 5 rods:

X formation
1333 C
12.0 KiRF/t
0.162 mB/t
428%

+ formation (no spacing)
1346 C
11.8 KiRF/t
0.166 mB/t
419%

Cooled with Gelid

Oh and this is using Ag Skies pack.

 

[GreenZombie]

A 7x7x7 reactor has interior dimensions of only 5x5x5. With both X and + designs the radiation from rods only travels 1 to 2 blocks before escaping the reactor interior.
The odd number of rods means that, in the X formation, the center rod is not partaking in fertilization, but, each of the other rods can "see" two rods to fertilize.
In the + formation each rod can still only see two other rods to fertilize, and all 5 rods are partaking in fertilization. This should be good, but the relatively small interior volume and dense packing (of both designs) is pushing the core rod temp up past 1000C, raising the proportion of hard radiation, and hence the necessity for moderation blocks between fuel rod blocks.

If you have the space and materials you can get a far more efficient reactor by increasing the casing to 8x7x8 - with a 2x2 fuel rod core (i.e. drop a fuel rod entirely). This should run cooler, use less fuel and provide more power than your current 5 rod design (for the cost of a bit more space)

The X shape is "ok" up to 3x3 as 4 fuel rods can partake in fertilization, but scaling X shapes up to a 5x5 core area means that none of the additional corner fuel rods are within 4 blocks of any other rods.

 

[McJty]

I think I understand what you mean now but I'm still a bit uncertain about how to actually place the rods for best results. Is it something like this (ascii art):

r.r.r
.r.r.
r.r.r
.r.r.
r.r.r

 

[GreenZombie]

"chessboard" packing, as you presented in ascii art is a space efficient way to interleave rods and moderator. Don't forget to border them with coolant.
If the fuel rod / core temperature can be kept below 1200 degrees, then omitting the moderator is usually better. i.e. make the core as dense as possible.

Of course, one interesting way to understand what you have presented is that there are actually two separate reactor cores sharing the same space.

A set of 9 rods, and a set of 4 rods, that are cross fertilizing.

4.4.4
.2.2.
4.4.4
.2.2.
4.4.4

Here, the rods marked "4" are all in range of 4 other rods, and cross fertilize. And the rods marked 2 can each see two rods.

I have a 22kRF/t reactor that has an interior layered like this (as a 11x7x11):

gggcccggg
gggcccggg
gggcccggg
ccc444ccc
ccc444ccc
ccc444ccc
gggcccggg
gggcccggg
gggcccggg

g is glass. c is cryotheum. 4 is a fuel rod.
Each fuel rod can see 4 other fuel rods. Radiation has to travel through at least 3 blocks of coolant to leave the reactor. The "glass" blocks are filled with glass, because the material there doesn't matter.
~22500RF/t peak production, with a fuel efficiency of ~110kRF/mB when throttled to 60% where my current usage has it.

 

[Skyqula]

If you guys havent seen my spreadsheet yet, take a look: https://docs.google.com/spreadsheet...sV5eFqBWUeGGvtyQ6e1agJN2s/edit#gid=1439247654

The best design pattern I got sofar is a repeated square pattern with cryotheum in its center. It produces 46,6 Buckets of steam per tick at an efficiency of 75.161 buckets of steam per yellorite ingot. (Actively cooled reactor)

Spoiler

This reactor could probably be made abit bigger (my gues is 2 less width, 4 more length) to produce 50 Buckets/t (maximum input/output). Youd then hook up 25, 1800 RPM turbines shown in this spreadsheet: https://docs.google.com/spreadsheet...UxOUGxMRlpERWtPMmtGT213bmc&toomany=true#gid=0
Unless another, better pattern shows up. Thats probably going to be your best bet. It would produce around 601.725 RF/t or 904.525.055 RF per yellorite ingot (assuming same efficiency but at 50B/t instead of 46,6B/t).

Don't use any material other than cryotheum as it is the best moderator, best heat conductor, (and, for passive reactors, 2nd best absorber).

This is not true. Cryotheum is the best coolant to use inbetween rods (The best thing between rods is... another rod). Between rods and the reactor casing you want resonant ender. Resonant ender has a realy high change of preventing radiation to travel trough it. Therefor, little to no radiation will ever reach the reactor casing and will therefor not produce heat. (Assuming actively cooled)

This means that efficient Big Reactor designs (a) ensure that fuel rods are lined up n/s/e/w of each other, (b) if they are separated from each other, they are separated by a material that has good moderation properties, and (c) there is at least one, but hopefully more, blocks of a coolant with good absorption properties between each rod the edge of the reactor.

And we can add D: are flat reactors (3 high) to get the most rods sending there radiation to eachother.

 

[GreenZombie]

I am looking at the code trying to understand how coolant heat generation differs from RF generation. I am clearly reading it wrong because it looks like actively cooled reactors should still be generating RF... There also doesn't seem to be a NetBeans or Eclipse project for BigReactors so Im having to use notepad++ to browse the source...

I know that in the passive case, Cryotheum has 66% absorbtion but 95% heat effiency, whereas Ender has 90% absorbtion but only 75% heat efficiency. Over a single block they both convert about 67% of the radiation into RF.

In Cryotheums case however, 34% of the radiation is available in the next block, whereas Ender has only 10% of the original intensity remaining. So (in passive reactors) more than one block of coolant favors Cryotheum as the coolant of choice.

The reason I mention this is that the coolantContainer.onAbsorbHeat method get passed "rfAbsorbed" implying that actively cooled reactors are still "generating RF" but are looping it into the cooling system instead of manifesting as actual available RF. Which implies that Cyrotheum might be better at transferring heat into an actively cooled reactors "coolant" store.

further reading shows rfTransferred has nothing to do with radiation generated RF.

 

[Skyqula]

I did another coolant test, this time with ender on the outside and cryotheum everywhere else. Its a small increase in efficiency, from 75.161 Buckets of steam per yellorite to 76.461 buckets of steam per yellorite. I also removed the corners, they are indeed useless. This changes the previous reactor this this:

Spoiler

 

[colonelhomer815]

If I were to use resonant ender as a moderator, am I able to use flowing blocks or is it only source blocks that can contribute for TE fluids?

 

[Padfoote]

If I were to use resonant ender as a moderator, am I able to use flowing blocks or is it only source blocks that can contribute for TE fluids.

Flowing will work.

 

[colonelhomer815]

cheers.

 

[GreenZombie]

I did another coolant test, this time with ender on the outside and cryotheum everywhere else. Its a small increase in efficiency, from 75.161 Buckets of steam per yellorite to 76.461 buckets of steam per yellorite. I also removed the corners, they are indeed useless. This changes the previous reactor this this:

Spoiler

What if you tried staggering the coolant blocks in the reactor. You have a lot of fuel rods with 2 faces adjacent to coolant and a lot with no faces adjacent to coolant - which means they can only move heat to adjacent rods which is bad for passive reactors, and probably active ones too.

Offsetting the coolant blocks should result in most fuel rods having 1 of their faces adjacent to a coolant without diminishing the overall density.

 

[Skyqula]

What if you tried staggering the coolant blocks in the reactor. You have a lot of fuel rods with 2 faces adjacent to coolant and a lot with no faces adjacent to coolant - which means they can only move heat to adjacent rods which is bad for passive reactors, and probably active ones too.

Offsetting the coolant blocks should result in most fuel rods having 1 of their faces adjacent to a coolant without diminishing the overall density.

I did a test looking at what gives the best results in terms of coolant inbetween rods. I also tried a Controll rod at 100% and 0%. The result was that an active rod is the best thing inbetween rods. The only reason coolant is used inbetween rods is to keep the temperature down. I made 2 more sheets where I can layout the reactor and it tells me how manny rods are radiating each rod. In the best design the avarage is 10.8!

Annyway, this is the puzzle behind big reactors. Finding the best pattern of rods and coolant that gives the lowest temperature and the highest radiation with the desired output. As that will lead to high fuel efficiency and power generation. You then apply the simple rules of 1) Dont ever build up and 2) 1 row of resonant ender touching the casing 3) Corners can be annything and 4) Cryotheum everywhere else.

 

[GreenZombie]

Annyway, this is the puzzle behind big reactors. Finding the best pattern of rods and coolant that gives the lowest temperature and the highest radiation with the desired output. As that will lead to high fuel efficiency and power generation. You then apply the simple rules of 1) Dont ever build up and 2) 1 row of resonant ender touching the casing 3) Corners can be annything and 4) Cryotheum everywhere else.

There are two (known) factors governing a reactor cores effiency: Fertilization, and cooling. Above some critical fuel rod temperature, the hardness of the radiation will be high enough that the radiation will be unable to fertilize adjacent rods, and moderator blocks will be required. That temperature seems very high, as, as noted, we have both found that the best thing to have between to fuel rods, is another fuel rod

That said, the way heat is modeled in the reactor is interesting, particularly to passive reactor design as heat emission directly generates RF. Heat can be conducted out of a fuel rod into adjacent blocks - either fuel rods (in which case it stays in the core and maintains core heat) or to an adjacent block of coolant (Given its heat conductivity). If each fuel rod gets 4 chances to emit heat, one from each face, and each face is facing another fuel rod it receives as much heat per tick as it emits, and consequently runs extremely hot.

As such, is appears possible in your design (I am at work and can't test it), to offset the rows of coolant blocks so they are not lined cardinally. This should expose alternating faces of fuel rods to coolant.

Something like this.

.xxxxxxxxx.
.x.x.x.x.x.
.xxxxxxxxx.
.xx.x.x.xx.
.xxxxxxxxx.
.x.x.x.x.x.
.xxxxxxxxx.
.xx.x.x.xx.
.xxxxxxxxx.

With that layout, it seems that every fuel rod has at least one face adjacent to coolant, but the overall fuel rod density has not changed, so, hopefully, the average path still fertilizes enough fuel rods.

I don't understand actively cooled reactors enough as it seems counter intuitive that lowering temperatures will produce *more* steam, but as I think that steam is principally produced by heat leaving fuel rods into the reactor environment, this tweak should have potential.

 

[Skyqula]

There are two (known) factors governing a reactor cores effiency: Fertilization, and cooling. Above some critical fuel rod temperature, the hardness of the radiation will be high enough that the radiation will be unable to fertilize adjacent rods, and moderator blocks will be required. That temperature seems very high, as, as noted, we have both found that the best thing to have between to fuel rods, is another fuel rod

That said, the way heat is modeled in the reactor is interesting, particularly to passive reactor design as heat emission directly generates RF. But, heat can be conducted out of a fuel rod by an adjacent block of coolant (Given its heat conductivity). As such, if a fuel rod has no faces adjacent to coolant, it receives as much heat per tick as it emits, and consequently runs extremely hot.

As such, is appears possible in your design (I am at work and can't test it), to offset the rows of coolant blocks so they are not lined cardinally. This should expose alternating faces of fuel rods to coolant.

Something like this.

.xxxxxxxxx.
.x.x.x.x.x.
.xxxxxxxxx.
.xx.x.x.xx.
.xxxxxxxxx.
.x.x.x.x.x.
.xxxxxxxxx.
.xx.x.x.xx.
.xxxxxxxxx.

With that layout, it seems that every fuel rod has at least one face adjacent to coolant, but the overall fuel rod density has not changed, so, hopefully, the average path still fertilizes enough fuel rods.

I don't understand actively cooled reactors enough as it seems counter intuitive that lowering temperatures will produce *more* steam, but as I think that steam is principally produced by heat leaving fuel rods into the reactor environment, this tweak should have potential.

Interesting! Ill take a look at this. Just made a 27x29x3 reactor. Controll rods at 50%. Almost at 80,000 buckets of steam per yellorite ingot, almost!