Tutorial: Nuclear Engineering, How to blow up your Nuker with style...

Be advised, this thread is now OUTDATED. Please refer to the NEW GUIDE!

Nuclear Engineering
Note: This IS goddamn complicated. Do not ask for "Pls make youtube tut", because i doubt you will want to watch a 1-hour video tutorial either. If you are not capable of comprehending the complexity of this system, don't worry, but pay a visit to the Nuclear Engineering section of this forum. You will probably find a design matching your personal requirements in there and don't need to bother understanding the mechanics, at all.

1. Your best friend: the Nuclear Reactor
The first thing you should keep in mind while thinking about Reactors: They're expensive. Reeeeeeeaaaaally expensive. In redo, they can convert Uranium into ridicolous amounts of energy.
A reactor consists of a main unit, the "Nuclear Reactor" and up to 6 "Reactor Chambers". A Reactor by itself merely consists of an 6by3 grid for installations. Each Reactor chamber placed adjacent to the original Reactor will provide some cooling, more hull integrity and a new column to place stuff in. This means, with full "upgrades", you will get a 6by9 Reactor.
To get the Reactor into producing energy, you will need 2 essential parts, no reasonable Reactor gan run without: Uranium Cells and Coolant Cells.
Uranium Cells are, trivially, the part producing the energy. Each uranium cell lasts quite long, 200000 ticks to be exact. This is approximately a length of 160 minutes, in effect 8 full Minecraft Days. However, as every knows, Nuclear Fission generates vast amounts of HEAT, as well. To absorb this heat, a basic and cheap component is the Coolant Cell. As well, there are one-use coolants, Heat derivators and additional Reactor Plating. We'll come back at this destinctive components later on.
To regulate your reactor (remotely, or even automatically), it has an additional function: If you apply redstone to the Reactor (or any adjacent block, f.e. chambers), all Uranium cells will be locked, instantly stopping to pulse. The remaining reactor components will procede as normal and usually cause a slow cooldown.
To set up a new reactor, you will not only need the hardware, but as well a good idea of the mechanisms to create something profitable. What is profitable? That depends on the requirements you want for the reactor.
1. Efficiency - A single uranium cell can result in 1-5kk EU, depending on the reactor's efficiency
2. Safety - 100% safe reactors are hard to design. Sometimes you will want to sacrifice safety for another attribute instead
3. Cooldowns - One can design a reactor which is capable of running non-stop. Other reactor designs tend to heat up and will need some cooldown periods in between.
4. Breeder - It's possible to enrich uranium to gain more uranium cells from one chunk of uranium ore. You need a special reactor for this, though...
5. Hungry - "Hungry" Reactors need additional ressources to keep running. For example they tend to burn through Coolant Cells (which need to be replaced) or use one-use coolants.
Generally, you won't be able to construct "Teh ULTIMATRON 5000 super-efficient, no-risk, never-cooldown-ned, breeding-integrated, never-hungry Reactor". It is, with a lot of planning, possible to create a reactor fullfilling one or two attributes to their max, but that's it.

Now, after covering the basics, let's move on to the actual mechanisms behind it:

2. Nuclear Fission, How to (don't)
For simplicitys sake, we will work with a "heat"-based system. Heat is a basic value, representing the (duh) heat of the reactor. Generally, heat is bad. Heat will affect the surrounding of the reactor, damp water, set fires and even (if too hot) smelt solid materials into lava. You don't want to have a fully-overheating reactor either, it will tend up to suddenly vanish... with a loud bang.
On the other hand, Breed-Reactors work more efficiently when heated up to certain degrees. We'll cover this in the seperate Breeder-Section.
Generally, you can assume a reactor to explode when reaching 10000 heat. This is not related to degree or any known measurement of Earth, but just the well-known MCHI (MineCraft Heat Index), commonly referred to as "heat" (Hydro-Eating, Annoying Temper).
Quite every (cooling) component of the reactor can store 10000 heat. How it does obtain and how it does work with this heat depends on the component.
Let's first answer the question: How does a reactor even heat up?
By Nuclear Fission done in and between uranium cells.
Due to the laws of block-ish physics, uranium reacts extremely linear and can be accustomed for in advance:
A single uranium cell will "pulse" once. Pulses mean energy and heat production. Generally, all uranium cells will pulse at once every few moments (to be exact, once every second), which is called the "Reactor Pulse/Tick".
Each pulse will produce 100 EU, spread over 1 second (20 ticks), resulting in 5 EU/t.
The heat generated by each pulse is, for unknown, quantum-physic-based (balance) reasons, dependant on the amount of coolant elements surrounding the Uranium Cell:
4 (fully surrounded): 1 heat @ each element
3 : 2 heat @ each Element
2 : 4 heat @ each Element
1 : 10 heat @ Element
0 : 10 heat @ Reactor Hull
As you can see, fully surrounding Cells with Coolant Systems (which are consisting of Coolant Cells, Reactor Plating and Heat Derivators) is generally something advised. However there is an important aspect AGAINST this:
Uranium cells "interact". For each adjacent uranium cell, the original cell will pulse another time. This does NOT reduce the cells longlivity. Effectively, this means a single cell will produce 1kk EU. If you place another cell into the same reactor, but distant of the first cell, the Reactor will produce 2kk EU.
If you place both cells next to each other, however, the Reactor will produce 4kk EU.
The problem of this behavior:
1. More pulses mean more energy generation, but as well more heat, too.
2. If you surround a cell with further Uranium, you trivially don't surround it with Coolants, thus the derivation of heat becomes more difficult.
This effect is called "Reactor Efficiency".
In best case, you surround an uranium cell by 4 other cells. In this case, the middle cell will pulse 5tuble, causing the maximum 5kk output. In worst case the cell is isolated and will only produce 1kk total. The higher the efficiency, the better... as long as the heat doesn't melt your reactor, that is.

But how do we create a good reactor if they will just overheat? By using:

3. Coolant Systems, A must-ave for Nuclear Engineers (and inhabitants of deserts)
There are 4 Coolant Systems the reactor can make use of.
1. Coolant Cells
A coolant cell can, as most components store 10000 heat. Next to this important function, it does cool itself down by 1 heat per reactor pulse.
2. Reactor Plating
Reactor Plating serves multiple functions:
It can store 10000 heat and will (though 10 times slower then a coolant cell), cool itself down as well.
More important, it can transport heat to adjacent components. If a reactor plating is heated up by something, it will instantly divert the heat among all surrounding coolant systems. Even to other platings (which will again divert the heat, however, NOT to other platings anymore).
Additionally, each piece of plating will increase the Reactor Hull's integrity, causing it to to take more heat before melting.
3. Heat Derivatos
These are important, highly-advanced and expensive gadgets for advanced Nuclear Engineers and can as any components store up to 10000 heat without melting.
HD's posess a unmeltable, highly-intelligent Circuit to measure, recognize and rate the heat of itself, surrounding components and the Reactor itself.
Thus, it will, each pulse, check all nearby components and attempt to balance their heat levels, by either absorbing or emitting up to 6 heat per pulse. As well, it will absorb (or emit) up to 25 heat from/into the Reactor Hull.
4. Lastly, the most underestimated Coolant System: The reactor itself
Folloring the basic laws of heat distribution, a rector will emit heat to it's surrounding. Each cubic meter (block) of air surrounding the reactor in a 3by3 sphere, will provide 1/4th of cooling. As well, water will provide a full point of cooling each. As well, the Reactor itself can vent 1 heat per pulse into the world by itself. Additionally Reactor Chambers possess intern heat vents, each chamber emitting additional 2 heat per pulse.
The sum of all these fators is called "Outward cooling".

Next to these Coolant Systems, there are "One-Use" Coolants, which can be used to ensure a reactor stays cooled even with moderate heat generation.
The most basic way of cooling a rector, trivially, is pouring fresh cold water in. Thus, if the reactor is heated up enough (>=4000), putting a bucket of water into the system will evaporate the water, resulting in a loss of 500 heat.
As well, one could fill the reactor with ice, which will instantly melt once the reactor is hotter then 200 heat (quite always) and will reduce it's heat by 200.
Whereas Water clearly is more effective, Ice can be stacked, whereas buckets would need to be refilled constantly. Your choice.

You now have the background knowledge to understand a reactors needs. Let's build something!

4. Constructing the Reactor; count the screws mate, count the scews...
As a review of chapters 1&2; the main part you will want to worry about is a rectors efficiency.
Do you want a low-efficiency reactor, that's running savely all-day? Or are you low on Uranium and rather would like a slow, interrupted by cooldown periods, reactor, which however can produce amyazing 4-5kk from each uranium cell?
To design a reactor, i usually would recommend a few sheets of paper and a pencil. First, you will need to think about a design for the uranium cells. You can place them in a whole lot of different manners and patterns. The trivially "best" (with max surrounding) pattern is a square. However, squares thend to isolate uranium cells without any adjacent coolant, drastically boosting the heat generation. A "line" of Uranium is semi-efficient, but will produce much less heat. A single uranium cell surrounded by 4 coolant cells is the most basic reactor, producing minimum efficient energy while not producing any heat at all.
Based on this factors, decide how the uranium cells are meant to be placed. Draw a draft of this placement. In next step, calculate the pulses sent by each piece of uranium. Keep in mind it's 1+# of adjacent cells. Thus a 2x2 block of uranium would pulse 3 times with each cell.
Now you can already see the output of your future reactor, it's 5EUt*# of all pulses.
In the next step, calculate the amount of heat generated, for each cell individually. Keep in mind, the heat generated per pulse is dependant on the amount of coolant surrounding the cell (4 Surrounding Elements = 1 heat each, 3 = 2 each, 2 = 4 each, 1 = 10 each, 0 = 10 @ Reactor Hull) and of course the number of pulses. Multiply these two factors and (only an advice) draw arrows from cells to nearby tiles, labelled with the amount of heat produced. Now sum up all arrows leading to a single tile and write the number into the tile. That's the total amount of heat produced towards this tile per reactor pulse. You will quickly be astonished how much heat this will be...
Now you need a way to deal with the heat. the most basic way would be placing coolant cells. However, as coolant cells can only diminish 1 heat per pulse, you will usually quickly overheat them. F.e. a production of 12 heat/pulse will cause a cell to melt in 10000/(12-1) ticks. That's less then 1000 (Reactor!) ticks, not even 10% of the total runtime!
However, now you can use Platings and HD's to redistribute the heat.
Plating will transmit the whole heat it gains towards the adjacent tiles. Thus, if you place a plating next to a 12-heat generating uranium, it will (trivially uranium isnt a coolant system itself) emit 4 heat each to it's 3 surrounging tiles. You see, 4 heat is much less then 12 for a single tile to handle. Depending on your reactor layout, you may even be capable of placing one plating adjacent to the primary plating again, further distributing the 4 heat to 3 more tiles, effictively causing the secondary plating (the plating "behind" the first plating) and it's 3 adjacent tiles to take mere 1 heat / pulse. Which is low enough to be handled by single Coolant cells each.
Keep in mind though: If a plating cannot find any viable adjacent tile to emit heat to (f.e. because it's surrounded by other platings, but itself a secondary plating (secondarys cannot emit heat further to more platings), it will store the heat itself. And as we read before, plating is quite sucky when it comes to cooling down again.
While planning, you should already draw these sort of behavior into your drafts. The system is quite reliable and can be pre-determined by this way.
The stuff written above was the best case: You get some heat, relink it via a few platings and redirect it into multiple 1 heat/pulse streams, which get easyly absorbed by coolant cells.
However, in many setups you will run out of space easyly. For example because the platings would intersect with other plating, causing unwanted heat flow, resulting in platings overheating (and if in such a system the platings melt, you're really screwed up). For this sake, you can create the HD's, also known as "Heat Teleporters".
HD's will emit or absorb up to 6 heat / pulse to/from nearby coolant systems and then emit/absorb up to 25 heat to /from the reactor hull. As every HD can access the reactor hull, you can consider each HD a "trainstation", linked to each other HE via the "Rail"/Reactor Hull. Effectively, this means you can place a HD in spots where a lot of heat is produced, but cannot get deriviated further (keep in mind not to overpower the HD though. 25 / pulse is it's max). Then you place another HD anywhere in the reactor and link it up with another row of cooling elements.
TA-DA, you successfully send the heat from the cells across the reactor hull into another set of cooling elements.
This is the spot where you can just drop your draft and say "hey, why even bothering, i just place a lot of HD's and link my reactor to sets of cooling systems... then i don't need to actually count the single heat production values. And this is quite correct, as long as you can ensure each HD doesn't get more then 25 heat/pulse (because then it won't be able to effectively emit the heat anymore, heating up and melting quickly).
If you design a HD-based reactor, you will usually have to accustum 3 values, only.
1. The total amount of heat produced by all uranium cells.
2. The amount of outward cooling (recheck chapter 3).
3. 1.-2. = The amount of heat not dealt with.
In the next step you need to deliver the amount of heat specified in 3. to an equal amount of coolant cells, spread across your reactor, connected via HDs. If you succeed in this, you created a reactor which will never overheat, always capable of dropping it's whole heat into a set of coolant cells.
With bigger and more efficient setups you probably won't succeed, though. In this case you should check whether your reactor is still "safe":
4. Calculate 3.-# of your coolant cells. This is the amount of remaining heat per pulse.
Now count together all components linked via HD's + 1. If the sum of all components exceeds 4., you got a safe reactor. Why?
Because too overheat during a full cycle of 10000 ticks, the reactor would need to produce 1 heat / pulse to each storage, including each coolant component (linked via HD) and the reactor hull itself. If it produces any less, the whole system (system = all components including reactor hull) WILL heat up... but probably not to any dangerouds degree. Exception:
If the sum of components barely exceeds your reactors heat production, it could cause the system to heat up to (f.e.) 9500. With that level of heat, the reactor could probably melt it's sourrundings, reducing his outward cooling or in worst case even melting a nearby reactor chamber. This can either be an annoying loss or in worst case exactly the fatality ned to make your reactor overheat.
It's up to you how to set up your reactor. But keep all these things in mind.

And, don't forget:

5. Breeding radioactive eggs: How to 8tuble your yield
Sometimes, upon running out, Uranium Cells won't be consumed, but turn into Near-Depleted Uranium Cells.
You don't need to be worried about these cells clocking your perfectly-designed reactor, they will not count as Uranium Cells for chain reactions anymore and will merely emit 1 heat per tick, regardless of their surrounding. Which is, compared to the heat production of a normal cell, quite neglectable.
Filling this depleted cell with some Coal Dust can create a less-radioactive material, generally referred to as "Depleted Isotope Cell". For itself, this cell is worthless.
However, you can now place it back into a reactor, slowly refining it into a fully functional Uranium Cell again. This process isn't automatized though: It requires nuclear fission.
Frankly spoken: A Isotope cell will slowly refine by nearby Uranium Cells pulsing.
Listen closely now. Uranium Cells will pulse one time for themselves and one time for each surrounding Uranium Cell... OR Depleted Isotope Cell. In latter case, this will generate the normal heat, but NO energy. Instead, the energy will be used to enrich the material inside of the Isotope cell. The Isotope itself does merely produce 1 heat per tick, regardless of the amount of pulses it receives.
Additionally it should be mentioned: Breeding does, for random physic reasons, run at much higher speed if the reactor is HOT. Literally.
Experiments resulted in a double of breeding speed for each 3000 heat. This makes breeding at 9000+ heat extremely efficient, yet ridicolously dangerous. Instead of running on high temperatures, you can of course just surround a isotope cell with more uranium cells, though. If you can deal with the heat produced by the pulsing uranium, that is.
If a Isotope cell is completely refined, it will turn into a "Re-Enriched Uranium Cell". Don't worry, it won't start producing energy and overheating your reactor right away. You first need to adulterate it with some coal dust again, and voila, you got a fully working Uranium Cell again.

As well, one should mention you can intentionally create Depleted instead of Full uranium cells, by splitting a single uranium ingot among multiple cells.
If you got a well-designed breeder reactor, you can potentially 8tuble your Uranium Yield this way...

6. For true engineers: radiated l33tspeak
As this is an own science for itself, you will probably will be one among MANY people designing Nuclear Reactor Setups.
To bath in glory and protect your intellectual property, you are free to public your designs accompanied by information like drafts/blueprints/screenshots.
Of course you're free to name your reactors however you want to , but there's a small naming convention to ensure everyone is understanding the same:

Reactors are assigned to "classes" representing their layout:

Mark I
Mark I reactors are the true art of Nuclear Engineering: Safe reactors, which never heat up.
A reactor labelled "Mark I" must fullfill the restriction of not producing any excess heat per frame, which effectively means the whole system does not accumulate heat. Not a singlepart of it!
There is, however, a sub-class convention: There are Mark I-I and Mark I-O reactors.
Mark I-I is defined as "A Mark I reactor capable of running anywhere, by colling itself with it's components." whereas a Mark I-O is "A Mark I-O reactor requiring a defined amount of outward cooling to maintan it's Mark-I state."

Mark II
Mark II are the next level of generators, following right after the "utopicly safe" Mark I ones.
A mark II is a reactor capable of running a full operation without going into "critical heat" (>8500). The reactor-system is allowed to heat up to any degree below that number, given not a single component does melt.
This definition ensures a Mark II is still quite safe, even if it can only run once before needing a cooldown.
As well, there's an additional clause used by some Engineers to point out their Mark II is much better then the other one. Thus, the naming "Mark II-x" was invented. x is a number representing the amount of full cycles this reactor can run non-stop without infringing any of the criteria for being a Mark II. Usually, this number is around 1-4, but in some cases you will design a reactor actually being nearly a Mark I, but just producing a tiny amount of heat to much. If your reactor can at least carry on 16 full cycles, you're permitted to label it "Mark II-E" for "endless", representing a reactor which will, under normal circumstances, run just like a Mark I.

Mark III
Mark III are usually the more-efficient reactors, producing too much heat to be designed as Mark I or Mark II.
This class is defined by not being able to run a full cycle without melting system components. For this reason, a Mark III needs an operator (whether human or automatized), who is in charge of shutting the reactor down in correct periods to prevent it from overheating.
As theoretically any reactor can run, given enough cooldown time, the classification Mark III is additionally bound to the restriction "Must be capable of running for at least 1/10th of a full cycle, non-stop, without melting system components." This is approximately the duration of 18 minutes, roughly equivalent of one MC-Day.

Mark IV
Actually something a good Nuclear Engineer would never design, but could actually be interesting for people with a ridicolous sense of effective uranium usage:
Mark IV reactors are a variation of the Mark III definition, stating "A reactor capable of running at least 1/10th of a full cycle, while not exploding." However, this vast definition does not include any restrictions regarding produced heat, molten system components or any risks, at all.
A extreemly efficient reactor may be designed as Mark IV, if enough other ressources (to replace molten system components) are avaible, but uranium is sacre.

Mark V
Rather a joke then an actual classification, the Mark V is defined as "A reactor incapable of running smoothly and solely designed to prove the fact reactors can explode."
Design one if you like, but don't exspect any real Nuclear Engineer to pay much attention to your complaints after your house (and the surrounding landscape) is gone.

None of the definitions above include any detail regarding the usage of one-use coolants. Thus there's a rule regarding the use of these: Any design is free to include as many one-use coolants as it likes, under the restriction that none of the collants needs to be refilled while the reactor is running. If it's a Mark III (or higher), the reactor trivially has cooldown periods, which are allowed to be used to exchange/refill one-use coolants.
However, if your reactor uses any one-use coolants at all, you must mark this by adding the "-SUC" Sufix after the classification, representing the term "uses Single-Use-Coolants".

Additonally, many people will be interested in "How efficient is your reactor?"
We remember, a single uranium cell can produce 1kk-5kk, representing the effectiveness of a reactor, dependant on the amount of uranium cells surrounding the given cell. For this sake, another convention regarding Efficiency Ratings was introduced to Nuclear Engineering:
To determine the "Efficiency Ratio" of your reactor, count the amount of all uranium cell pulses (excluding those used for breeding, as they don't produce energy!) across the whole reactor and divide it by the number of uranium cells used.
The result will range from 1 to 5 (with a lot of decimals in between). It should be noted, however, that 5 can't ideally be reached, unless you got a infinite-sized reactor where each uranium cell is fully surrounded by 4 other cells each.
Depending on this number, your reactor can be ranked among following classes:
Efficiency E: 1
Efficiency D: >1 and <2
Efficiency C: >=2 and <3
Efficiency B: >=3 and <4
Efficiency A: >=4
Reaching a high efficiency class usually, but not necessaryly, means a "high" Mark-Classification as well, as more efficiency = more heat = less stable system. However, it IS (f.e.) possible to design a "Mark-I EC" reactor. (A system not producing any heat while still having an efficiency ratio of 2.17 .)

Of course, if one designs a Breeder Reactor, there are different laws to apply. Breeder Reactors are more effective when heated up, thus designing a Mark I Breeder is actually quite stupid. For this sake, Engineers can either add a "Breeder" tag to their existing reactor, if breeding is included (example: Mark I ED Breeder) OR use a whole different naming convention:

Negative-Breeder
The most common design for normal reactors is a setup which produces less heat then cooling. Achieving this, exspecially for smaller breeders and low-efficiency reactors, is considered easy.
A Negative-Breeder is a reactor which is meant for breeding and runs entirely safe by cooling down just by itself.

Equal-Breeder
The best, but hardest to accomplish, design for a Breeding reactor is a stable "Equal-Breeder". This reactor is fine-tuned to keep his heat. Some Engineers even use Isotope-Cells stored in the reactor to adjust the amount of heat it generates upwards, as well as enclosing their reactors in solid rock to prevent outward cooling.
An equal-Breeder usually quickly be heated up by replacing the Isotope Cell with normal Uranium (as it generates, in any situation, far more heat then a Isotope Cell). As soon as the reactor is heated up enough (usually beyond the 6000-heat mark, unless you are Duke Nukem and want to go with 9000+), the isotope cell is used again, resulting in the reactor to balance out, maintaning a high level of heat to accelerate breeding.
As it's hard to exactly hit the spot, the definition is forgiving, merely requesting "A Equal-Breeder must not alter his heat more then 1000 points in any direction while running a full cycle.", which effectively allows a production of +-0.1heat/pulse. Many Nuclear Engineers thus prefer the labelling "Perfect-Breeder" if it's truly +-0heat.

Positive-Breeder
Risky to use, but easy to set up and more effective (usually) then Negative-Breeders, this class of Breeder DOES produce heat. The definition does not limit the amount produced, thus any badly designed reactor including an Isotope Cell can be classifed as this. On the advantageous side, a carefully watched Positive-Breeder is quite useful to maintain high temeperatures for breeding, merely needing a cooldown then and when.
Some Engineers prefer "Aquatiq-Breeder" reactors, which are filled with multiple layers of water buckets to prevent the reactor from going past 4000 (the point where water damps to create cooling), while never going below 3000 either. Given the heat production is low enough for the water buckets to last for a whole cycle, this is an actual safe and convenient setup.

This is it, start designing Reactors now.