Atomic science power generation. Tips and tricks.

[TonyVS]

In the new version of tekkit I feed liquid deuterium direct

 

question does anyone know what the difference is between the Electromagnets and the Atomic Electromagnets in the new atomic science? all I know is it takes 3 Electromagnets to make 2 Atomic Electromagnets

 

Edited February 19, 2014 by TonyVS

[GaMEChld]

Alrighty, here's what I've come up with so far. It's not perfect, because it still requires a little bit of manual manipulation, but will run at long stretches without intervention.

 

Main Fission Reactor:

Fuzzy Export - Fissile Fuel Rods split at 50% (will not feed the reactor any rod that is more than half damaged).

Fuzzy Import - Any Fissile Fuel Rods (take rods of any condition from reactor) [active on pulse, tied to a timer]

 

 

ON SEPARATE SUBNET

Breeder Reactor Setup

Breeder: Basic Export - Breeder Fuel Rod

Center: Fuzzy Export - Fissile Fuel Rod ANY (feeds damaged fissile rods for healing)

Center: Precise Export - Fissile Fuel Rod (perfect) [connected to main network] (fully healed rods sent back to main network for use)

 

The feed for the reactors in this subnet is a chest containing Breeder Rods and the damaged Fissile Fuel rods that I wish to heal. The chest has a Basic Export Bus attached connected to the main network that will stock Breeder Rods for this breeder project. The only manual manipulation is the periodic deposit of damaged fissile fuel rods into this subnet for healing. I toss a batch of damaged fissile fuel rods into the chest, and it will feed them to the breeder, heal them to 100%, then send them back to the main network when fully healed, where the main reactor is already configured to use any rod that is >50% health.

 

Work in progress, but functional for now.

[Silmenume]

I apologize if this is dragging the conversation in a direction that it shouldn't go - but here is my Fusion Reactor Power Generation and Automatic Resupply Set up. (With Pictures!) (If this belongs in another thread please move)

This is all predicated on the current recommended build - NOT BETA! Which I believe is the 1.1.10 build.

 

When I first started playing with fusion power I knew very little about how to really make it work right.  FREX - I found that a single fuel cell was consumed in a handful of seconds and it would go through a stack of cells in about 10 minutes (give or take) I did not understand that the reactor would then continue to produce steam for a good long time afterwards.  This lead to some pretty daunting tin consumption numbers - numbers so high that one would need a quarry constantly running to provide enough tin ore.  (Turns out I was wrong about that - but that was a first principle I was working with) So when I started experimenting I was very concerned about squeeeeeezing every MJ I could from the process.  My design here is the fruit from those original assumptions.  I've succeeded in that when I'm basically in an "idle" mode I'm only consuming maybe 1 cell an hour.  Conversely the system can throttle up to a current max output of some 450 MJ/t without any interference on my part.  I was able to accomplish this through judicious use of liquid managers and monitoring my storage system levels and turbine "activity".

This is a self-sustaining, self-regulating operation. The Reactor is what I call a "Fat-Diamond" flood reactor. The internal open space from electromagnets surrounding the fusion reactor to the outside electromagnets is 5 empty squares. The points, however are 3 wide instead of one which allows for more spawning space on the flat sides.

Here is the overall set up. I'm a bit embarrassed because I have not yet spent much time making it look pretty, but it works well. It is set in a re-purposed underground quarry hole. You can see the reactor, the steam buffer storage, power turbines and some power tessaracts. The structure in the upper right is immaterial so please ignore.

 

 

The image below focuses in a bit more on the fusion reactor itself.  Again notice the diamond points which are 3 wide instead of 1 wide.  I saw the desgin somewhere on the net and to my discredit I do not recall where so I cannot give attribution.  I had also read that while one can make a flood reactor as wide as 6 at the points the plasma rarely ever reached that far - IOW it was a waste of resources to build out to capture a very rare event.  However it was postulated that the diagonal sides could be on block wider (further away from the reactor core) as the plasma spawning space "might" be based on a circle.  If 6 (7 less 1 for the inner most electromagnet) was the max radius of plasma spawning why not try and capture some of that lost potential spawning on the diagonals.  I don't know if its true or not but it was worth pursuing on the idea of extra power per deuterium cell.  You can also see the Deuterium Cell feed on the top of the reactor basin.

 

 

This image is an "eye level" of the reactor at one of the "points".  You can see some water which can heated by the electromagnets which can be seen behind.  On top of the water are steam collectors which collect the generated stem then transfer said steam to liquiducts.  One can see the arrows in the liquiducts pointing up indicating which way the fluid, steam in this case, is to move.  On top redstone block are positioned to "power" the liquiducts.

 

Below is an image of the bottom the flood reactor.  One can see the electromagnetic glass on the bottom which contains the plasma but does not in and of itself contribute anything to the steam creation process.  Visible through the electromagnetic glass are the electromagnets which are directly involved in the production of steam.  When the plasma spawns, near as anyone can tell, only those electromagnets which come into direct contact with the plasma actually heat the water.  The more times and the longer the electromagnet is in contact with the plasma the more steam that is generated from that electromagnet - or so it seems.

 

Also note the redstone energy conduit, redstone energy cell and the lever.  I had read from jakalth in a thread that one does not want to give max power to the fusion reactor core for there is a fairly narrow window where the process is most efficient.  Thus the redstone energy cell; it serves both as a battery and a choke limiting the amount of energy, 55MJ/t in my set up.  The lever is there to shut of the power to reactor should it become necessary to do so.

 

(The foreman on the right makes sure all safety regulations are followed)

 

 

Below is my steam buffer system.  Given the behavior of liquiducts I found that I could feed the reactors entire steam output through one duct without "apparent" loss.  Each "tower" of liquid managers is constructed from 8 levels of 8 liquid managers giving a total of 64 per tower.  In my experimenting I found that a pair of towers, 128 liquiducts (which equals 16,384 buckets), was sufficient for my storage needs when I was running 20 large turbines.  On average about 70% of capacity was used per deuterium cell.  As has been noted several times before the amount of steam generated by a given cell varies quite widely in a flood reactor design and so with this quantity of storage I don't think I ever ran out of space.  The reason for the second pair of towers deals with a disappointing glitch with the autarchic gates.  This is a mode in the autarchic gates whereby they are supposed to send a single pulse thus drawing one and only one item from inventory when presented with a redstone signal.  Alas this feature does not work or fails so consistently as to require a very messy work around.  TonyVS was the poster who helped me with this particular problem.  Further down I show the solution to this issue.  (Thanks TonyVS!)  Why the second set of towers?  Because the kludgey autarchic gate work around does not draw a reliable fixed number of deuterium cells from storage when requested.  Frequently 2 cells manage to get delivered into the reactor and thus the need for the extra steam storage.  I did want to squeeeeeeeeeeze every bit of energy I could from each cell so I had to provide enough storage to cover the steam output when 2 cells were burned in rapid succession.  The output of these buffers all feed into a single liquiduct pipe when then goes off to feed my large turbines.

 

 

Below is where the steam is converted to MJ's via large turbines.  Right now I have 20 large turbines giving me a theoretical output of 450 MJ's/t.  I have read that 25 is probably a better number but I haven't tested that yet myself.  I know my design isn't particularly space efficient but my layout gives my plenty of space for more large turbines.

 

 

The entire output of my large turbines is fed into a single REC which can be seen in the lower left of the below image feeding a two tier storage redstone energy cell storage system.  This two tiered energy storage system is a major part of the self-regulation of my energy production system.  Rather than feed deuterium cells into the fusion reactor at regular intervals whether I need the energy or not I can now monitor my energy reserves and deuteriums cells only when my reserves are empty.  Yet I have no interruption of power and this is accomplished via the two tier design.  I will explain fully as we go along but right now I'll just go over the REC's themselves.  The entire energy output of the large turbines directly feeds 18 REC's which you see on the bottom two rows.  The REC's inputs and outputs are both set to max at 100MJ's/t.  These 18 REC's are ganged together and their combined output is then fed into the topmost tier of 6 REC's.  Their settings are different at 100MJ's/t in and 80MJ's/t out.

 

(You can also see the monitoring gates on the bottom 18 REC's - this will be explained shortly)

 

 

Below we see the 6 REC's.  One can see the redstone energy conduits that are drawing power from the REC's and which then leaves to feed everything else I run.  The input side is hidden from view behind the 6 REC's.  Why do I have the 6 REC's here and why do I have their settings set to the values of 100 in and 80 out?  The answer to the first question is a function of the setting of 80MJ's/t out.  6 REC's times 80MJ's/t gives me a maximum potential output of 480MJ's/t which is just over my theoretical max generating capacity of 450MJ's/t.  In this particular arrangement of values I have full unfettered access to the maximum generating capacity of my large turbines should I have need.  Remember that my bottom tier of 18 REC's is set to 100 in and 100 out so they do not hinder the flow of MJ's from the large turbines.

 

So why do I have my top 6 REC's set to max in and 80 out?  So they can develop a storage charge reasonably quickly when the system is not under max load.  When these top 6 REC's are full then the bottom 18 REC's begin to fill with power - and they do.  My experience is that over time I get an average of 7,000,000 MJ's per deuterium cell so I do get end up with extra energy frequently despite the 3.6M MJ storage of the top 6 REC's.  This also means that the bottom 18 REC's with their settings at 100/100 will run out of stored power before the top 6.  Because the bottom REC's are set to 100 MJ's/t output while the top REC's are set to output 80 MJ's/t in almost all instances the top REC's are still charging while the bottom rows are discharging. 

 

This is key.  When the bottom 18 REC's ALL show empty a redstone signal is sent to the computercraft computer.  The computer watches for the signal and when it arrives sends a timed redstone signal to the autarchic gate to pull a deuterium cell out of storage and feed it to the reactor.  The 6 REC's act as a buffer while this process happens so there is no interruption of power.  This process of pulling the deuterium cell out of storage is imperfect due to a glitch in autarchic gates and I'll show later all that is involved trying to make it work.

 

 

Below is the fundamental unit of the monitoring process.  I used Diamond gates as sensors for their access to all 4 colored wires.  I really only needed gates that sensed one event but they only output to red wire.  This was annoying because how long it takes to make a diamond gate, but that's how it goes.  There was an occasion or two where I could use gold gates because access to only the blue and red wires was sufficient and you can see them scattered here and there.  As you can see I have 3 gates feeding their outputs to a 4th gate.  Each of the three gates monitoring a REC is watching to see when the REC is empty then set to output a signal to a unique wire color when that condition is met.  The 4th gate which is position behind the 3 the REC's can be seen to monitoring the three wire colors used by the gates on the REC's.  Being an AND gate this 4th gate will output a signal also to a colored wire (the remaining color which has not been used by the 3 monitoring gates) only when all three colored wires are high (carrying a signal).  So this 4th gate is actually functioning AS an and gate.  Because gates don't differentiate between incoming and outgoing wires this limits the number of wires it can monitor to three because one color of the 4 available must be reserved for this gate's output.  Hence the limit of 3 monitors feeding into one AND gate.

 

 

Below is the reverse image of the above...You can see the yellow output wire from the 4th gate leading to the next tier of logic gates.

 

 

In the image below you can see groups 2 and 3 of the REC's and their attendant monitoring gates.  From this view you can see 3 levels of monitoring and logic system.  The first tier being the gates monitoring the REC's leading to the AND gate right behind them as discussed above.  Also in this image is the 3rd tier that collates the outputs from the 3 AND gates meaning this foreground gate is actually managing the data from all 9 of the bottom REC's.  It too is an AND gate under the same restrictions outlined above where it can only monitor a maximum of 3 wires if one plans on using a wire to carry the output result.  In this case we can see that this AND gate is monitoring 3 wires (Red, Blue, Yellow) and will output a signal to the Green wire when the conditions of the AND are met - in this case when all 9 REC's are empty.

 

 

In the below image we can see the AND gates that are monitoring both the lower 9 and upper nine REC's.  In the right bottom quadrant of the image we can see the output wire of this AND gate, which happens to be yellow, leading out to our last and final level of the system.

 

 

In the below image not only do we see the final AND gate level, but we can see all for levels of the monitoring and logic system!   This AND gate monitors the inputs from the upper and lower systems and when both are true outputs a redstone signal.  For this condition to be true all 18 REC's must be empty.  You can see the rednet cable which accepts the redstone signal and will carry it to the computercraft computer.  So now we have a system that tells us when the lower tier of REC storage is completely empty and at this point one would think this would be a sufficient condition to add fuel to fusion reactor.  But its not!!  I found there was another situation that needed to be monitored to be truly efficient and we shall soon see...

 

Note - the blue wire on this last AND gate serves no purpose.  I'll have to go back sometime and remove it.  Sorry if it causes any confusion.

 

 

I found that under high demand conditions the bottom REC's can be completely empty while the turbines are still generating power.  Either the fusion reactor was still generating steam or the steam buffer system still had steam and was still delivering it.  In either case nothing would be gained by feeding the reactor a deuterium cell as the system still was generating power.  Thus I added this gate to monitor the redstone energy conduits. The gates seem to sense the conduits as batteries and not power transmission devices so I could only set them to send a redstone signal when the conduit was "empty".  You can see this in the picture below...

 

 

This view just shows the rednet cable running from the gate monitoring the generation system to the computer craft computer.

 

 

Below we see the 2 rednet cables leading to the computercraft computer - one from the REC's and one from large turbines.  Note - computercraft computers do not recognize rednet cables so you have to convert from the cable to redstone dust.

 

 

Below is the kludged deuterium cell delivery control system.  It looks like a Rube Goldberg contraption because of the above mentioned glitch in the autarchic gates.  In this setup the comptercraft computer serves two roles.  The first is to monitor the two rednet cables.  When both are true - there is no power stored at all in the bottom 18 REC's AND the turbines are not generating power then send out a redstone signal to the autarchic gate telling it turn on its "energy pulser".   The second role is to "control" the pulser in the autarchic gate.  The autarchic gate has two pulser modes - continuous and single.  Ideally the "single pulse" mode is the I needed here, but it is the one that is glitchy.  In my case it just doesn't work at all - I have to use the continuous mode of the "energy pulser."  Hence the computer.  The real purpose of the computercraft is to control the duration of the redstone signal to the autarchic gate so as to get it to run just long enough to pull a single deuturium cell from the ender chest.  2/10's of a second is as close as I can get to making this happen.  I reality typically get anywhere from zero to four cells.  This is why I have the extra two towers of liquid managers, the distribution pipe, the chest and the import bus (which is attached to the chest but from behind and can't be seen in the pics I have).

 

The lever in the foreground is just a "manual override" to force the system to feed deuterium cells to the fusion reactor.

 

 

This is the program I use to run the system.  Unfortunately the first conditional line got cut off.  Basically it is an and conditional testing the "left" redstone input and the "back" redstone input. If both are true then the computer turns on the "right" redstone output and then "sleeps" for .2 seconds leaving the output on.  After .2 seconds the computer "wakes up/comes out of pause" and turns off the "right" redstone output.  This is my effort at forcing something to do something it wasn't designed to do - to make the "continuous pulser" act like a "single pulse".  It does not work very reliably which accounts for the distribution pipe, chest, import bus and me cable.  By the way it was TonyVS who clued me into this arrangement and gave me a starting program to work with.  Thanks TonyVS!

 

Note -  I have the computer sleep for 5 seconds whether or not it sends a signal because I found that running the computer continuously lags the servers as well as causing the computercraft computer to throw some sort of "thread lock" (or the like) error.

 

 

Below is the part of the system that deals with the Deuterium cells directly.  We start with the ender chest.  This is paired with another that remains topped off with D cells by a chemical extractor in another part of my base.  A tip from TonyVS suggest I fill all but one stack of the ender chest with filler so it isn't filled with deuterium cells.  Using this arrangement leaves only one stack of cells is sitting idle - less the 1 - 4 that get pulled by the autarchic gate at need.

 

To make sure ONLY deuterium cells get pulled you need to use an advanced wooden transport pipe to pull the D cells out of the chest.  The pipe needs to be powered to pull items out of storage; hence the autarchic gate.  You use the GUI on the advanced wooden transport pipe to tell it what items are of interest.  You then click the button to set it to "These Items Are Required" so it know what to do with the items indicated in the spaces above.  In this case this adds up to only pull deuterium cells.

 

Next we see the setting on the autarchic gate which is to turn on the continuous pulser in the presence of a redstone signal.  This is the signal that the computercraft computer outputs for .2 second when the REC's are empty and no power is being generated.  Unfortunately the result is unpredictable and I get anywhere from 0-4 cells which leads to the need for the distribution pipe, chest, import bus and ME cable.

 

The way I set up the distribution pipe it basically acts as a divider.  No more than half (rounding up!) of the cells go on through to the reactor.  So if 4 get sent only 2 continue on through.  The others are sent the chest which then recycles them back into the system via the ME import bus.  If 0 are sent of if 1 is sent and gets shunted to the return in 5 seconds the computer will send another request and more will be sent.  Like I said kludgey, but its the best I could figure out.  So as I mentioned above the regular chest accepts those deuterium cells that got shunted by the distribution pipe.  The import bus attached to the chest then vacuums up the cells and puts them into storage where they will be called back into service by a system that I will describe later.

 

 

The following three images follow the path the deuterium cells travel from the ender chest to the fusion reactor core.

 

 

 

 

So the above system automatically adjusts itself to need without outside input.

 

Below is the system that keeps the deuterium cells in automatic supply.  I start with an aqueous accumulator to provide an "infinite" supply of water to the chemical extractor

 

Below is the chemical extractor which is powered by the fusion reactor power system, fed water by the aqueous accumulator, fed empty cells by the ME Interface and feeds its output back into the ME network.  The ME Interface contains a pattern that starts with an empty cell and outputs a deuterium filled fuel cell.  This allows for a device on the ME network to call for a deuterium cell and have a device (the chemical extractor connected to the ME interface) to present itself to the self as being able to fulfill such requests.  This is important!!  I have a Molecular Assembler (not shown) that has a pattern that constructs the empty cells on need (all the way up from sand and pulverized tin if necessary) if the request for a deuterium cell meets an empty inventory.  Nice and automatic!!  Look ma no hands!

 

Note - the ME interface crafting mode is set to "Export stored items one at a time, or craft"  This is critical!  We want this ME interface to make the chemical extracter to fill the empty cells with deuterium when the request is made when none are in storage.  On the right is the import bus which takes the completed deuterium cells from the chemical extractor.

 

Finally we come the ender chest that is paired with the one I showed earlier next to the fusion reactor.  This is the chest that gets filled with newly created or "recycled" deuteriums cells (those that had been sent to the chest by the distribution pipe down by the reactor) that the other ender chest draws from.  The export bus attached to the ender chest must have its "Output Controls" set to "Export stored items one at a time, or craft..." or the auto-filling of the ender chest will not work "flawlessly".   With this setting the export bus will attempt to fill any inventory it is attached to that has space left with the items indicated in spaces at top (in this only deuterium cells) if any are available in storage or craft them if they are not in storage.  Thus if any cells had been shunted away from the fill pipe of the fusion reactor to the chest and pulled into the ME network the cells can now be placed back into the ender chest when other cells are removed.  No cells are wasted.  If there are no cells stored in the ME network when a space opens in the ender chest then the export bus will put out a request for a deuterium cell to be manufactured.  In either case the ender chest stays fully supplied as long as ME network has the necessary raw materials - sand, water and pulverized tin.

 

 

When I am away from the server for a few days or a week the power system is incredibly frugal with its consumption of resources.  When I need lots more power its right there.

 

I don't know if I've shed any new light on the fusion reactor power generation scheme, but this is how I've designed what I believe to be a very resource friendly and hands off yet flexible power generation system.

 

Thanks to jakalth and TonyVS who have helped me directly and all those whose work has allowed for the accumulation of the data necessary to make this work.  I hope this is helpful to someone out there!

 

Best

 

edit for clarity and spelling errors...

Edited February 22, 2014 by Silmenume