ReactorCraft Fission Reactor Designs

[Demosthenex]

Feel free to post to the wiki if you like.

 

[madnewmy]

I think that's the RF generator, not the HP turbine.

yup, i meant i would love one for this xd

and there is the turbine http://www.minecraftforum.net/topic...worldgen-civilization-and-more/page__st__8700

 

[Demosthenex]

Here's my next batch of test results. Each reactor was connected to a single HP turbine whose output was stored into an industrial coil. A computer controlled clutch allowed me to capture exactly one hour of output. Each reading was taken after allowing the reactor to power up and stabilize.

The top three are just some I was tinkering with. I was disappointed in to 5 GJ model, while very impressed with the 85 GJ one. Food for thought, see further discussion on boilers.

The middle one at 32 GJ was a test of heat conductivity where the boilers adjacent to the reactors were NOT water cooled. These insulated boilers were surrounded by blocks (hollow steel slabs so I could check temperature through the hole). The bottom boilers were connected and generating steam. I intend to test an interlaced pattern to see if I can increase the heat transfer as I believe the border region was the limiter here.

In the first lower column I've been experimenting with whether consecutive boilers draw away more heat and are more efficient. The answer is "it depends". Too few boilers and you can't extract all of the heat (26 GJ/31 GJ/41 GJ), but after that additional boilers seem to reduce the efficiency (29 GJ).

The second lower column is where I tried to limit the dimension I was measuring in. This series produced mediocre gains with additional boilers. The fluctuation in values concerns me because this leads me to believe the topology of the steam piping may have something to do with the total output. All of these were combined in a grid, linked across the front of the reactor array, and that single connection to the HP turbine.

The third lower column was measurements in two dimensions. This was really unusual in that the hottest reactor at the top had the most output.

As a result, I have a few new questions.

Does the topology of the steam lines matter? I know you can get steam "stuck" sometimes if you alter then while running, but these started new and empty. Do steam lines have to balance? Are there issues combining outputs?

Do boilers also react with neutrons? I had the impression only the cores did. If so that could explain the reduced efficiency in some configurations where the neutron had to travel through too many potential obstacles between the core and the reflector.

How much heat is generated when a core does interact with a neutron? What is the chance of a neutron passing through a core interacting with it? Are reflectors always 100% effective?

My next round of tests will be to determine if I can be more efficient by utilizing 3 or 4 sides of the cores for heat transfer. This could be less efficient if the boilers absorb neutrons.

Are there blocks that transfer heat, but never absorb neutrons? My understanding was all other solid blocks are insulators. I know some insulators absorb neutrons, are there any that do not?'

I received a few questions regarding vertical stacking and heat transfer. I ran one reactor that isn't listed here that had miserable performance. I'll have to test this more, particularly with the insulated boilers or other heat transferring blocks. I can say that many of these designs tile vertically as both the cores and boilers combine into multiblock structures. Perhaps at some point I'll study how much they improve.

 

[Reika]

Does the topology of the steam lines matter? I know you can get steam "stuck" sometimes if you alter then while running, but these started new and empty. Do steam lines have to balance? Are there issues combining outputs?

Yes. Since they only transfer steam between each other if there is a difference greater than four units, more complex and/or longer piping means more "pressure" is required to drive steam through the lines.

Do boilers also react with neutrons? I had the impression only the cores did. If so that could explain the reduced efficiency in some configurations where the neutron had to travel through too many potential obstacles between the core and the reflector.

Not unless the boiler is of the wrong type for the reactor.

The following differ by reactor type. Here is the data for mainline fission reactors:

How much heat is generated when a core does interact with a neutron?

20C for uranium fuel and 30C for plutonium fuel.

What is the chance of a neutron passing through a core interacting with it?

25% if loaded with fuel, 0% if empty. Note that if the core is full of nuclear waste, the neutron may be absorbed yet generate no fission event.

Are reflectors always 100% effective?

No, 75% of neutrons are simply absorbed.


If you know Java, I strongly recommend reading the "onNeutron" function in each of the relevant reactor TileEntity classes, which is fired each time a neutron impacts the block.

 

[Demosthenex]

If you know Java, I strongly recommend reading the "onNeutron" function in each of the relevant reactor TileEntity classes, which is fired each time a neutron impacts the block.

I know many languages and have read your Github repository extensively. While I had followed onNeutron, *heat, setTemperature, I never could find it. There's alot of inheritance going on that I hadn't fully understood. You have remarkably clean code!

Thanks for the additional information![DOUBLEPOST=1404353976][/DOUBLEPOST]

Yes. Since they only transfer steam between each other if there is a difference greater than four units, more complex and/or longer piping means more "pressure" is required to drive steam through the lines.

What would you recommend then in regard to the steam lines given you must aggregate the output from the boilers to power something as large as the HP turbine? Would straight lines be preferred to grids? Minimize junctions?

 

[Demosthenex]

No, 75% of neutrons are simply absorbed.

That's significant! So inter-core reactions are more important than reflected reactions.

To be clear a reflector will never allow a neutron to pass through, it will be reflected with a 25% chance or absorbed?

 

[Reika]

What would you recommend then in regard to the steam lines given you must aggregate the output from the boilers to power something as large as the HP turbine? Would straight lines be preferred to grids? Minimize junctions?

Yes. The pipe pump also helps.

To be clear a reflector will never allow a neutron to pass through, it will be reflected with a 25% chance or absorbed?

Correct.

 

[Demosthenex]

Please bear with me as I have discovered that some of my power output measurements are inaccurate due to size of the HP turbine I used. They are not completely off, but I have to go back and inventory where I used a maximum size HP turbine vs a 3 or 4 block version. I discovered this while comparing the output for the same reactor design in anoter build, and got significantly less power. In debugging the turbine was larger on the first.

For the columns of like design, those numbers are good because they all used the same 3 block HP turbine. Single reactors are not reliable.

I'm going back to try and setup a standardized way to measure output. I'm already using a bedrock industrial coil to store output for one hour to base my numbers on, and I'm trying to nail down the conversion from the coil to kilowatt hours as I think sustained output is a better measure than spot checking the MW.

 

[Demosthenex]

Here's another batch testing vertical tiling and power scaling. It looks like a simple reactor made tall has great output!

 

[Reika]

Please bear with me as I have discovered that some of my power output measurements are inaccurate due to size of the HP turbine I used. They are not completely off, but I have to go back and inventory where I used a maximum size HP turbine vs a 3 or 4 block version. I discovered this while comparing the output for the same reactor design in anoter build, and got significantly less power. In debugging the turbine was larger on the first.

For the columns of like design, those numbers are good because they all used the same 3 block HP turbine. Single reactors are not reliable.

I'm going back to try and setup a standardized way to measure output. I'm already using a bedrock industrial coil to store output for one hour to base my numbers on, and I'm trying to nail down the conversion from the coil to kilowatt hours as I think sustained output is a better measure than spot checking the MW.

You are in luck: Cutting the turbine length reduces the torque (and thus power) output by a simple ratio, and has no other effects, so you can simply divide by the same ratio to get the "theoretical" total energy.

  @Override
   protected double getEfficiency() {
     switch(this.getNumberStagesTotal()) {
     case 1:
       return 0.0125;
     case 2:
       return 0.025;
     case 3:
       return 0.075;
     case 4:
       return 0.125;
     case 5:
       return 0.25;
     case 6:
       return 0.5;
     case 7:
       return 1;
     default:
       return 0;
     }
   }

 

[Demosthenex]

So folks given Reika's list of HP turbine efficiencies and my frequent testing with 3 deep turbines, that means all of my numbers are taken with a 7.5% efficiency. This represents a huge underreporting of the potential power done during testing for convenience.

That means the little 6 core reactor design that reached 28 GWh at 7.5% efficiency could reach 373 GWh (avg output of 106 MW/s) with a full size HP turbine at 100% efficiency.

I find that number to be enormous! I'll setup a test with a full HP turbine to confirm.

 

[Scottly318]

I read this thread and all I can hear is a booming voice saying "BEHOLD THE POWER OF SCIENCE!"

 

[madnewmy]

So folks given Reika's list of HP turbine efficiencies and my frequent testing with 3 deep turbines, that means all of my numbers are taken with a 7.5% efficiency. This represents a huge underreporting of the potential power done during testing for convenience.

That means the little 6 core reactor design that reached 28 GWh at 7.5% efficiency could reach 373 GWh (avg output of 106 MW/s) with a full size HP turbine at 100% efficiency.

I find that number to be enormous! I'll setup a test with a full HP turbine to confirm.

It would be fun if you could also leave us an approximation of the torque and speed it produces

Also, did you try them with one/multiple LP turbine? It might be beneficial for smaller reactors (I think, not totaly sure about that)

 

[Demosthenex]

It would be fun if you could also leave us an approximation of the torque and speed it produces

Also, did you try them with one/multiple LP turbine? It might be beneficial for smaller reactors (I think, not totaly sure about that)

So I think the HP turbine as a fixed max speed, and dynamic torque. However much of my testing has been with low efficiency, and so only the total wattage matters. I'd recommend if you want to split up the output you use Electricraft or charge up bedrock industrial coils.

 

[Demosthenex]

Wow! The single layer 6 core reactor produced 1.2 TJ in an hour with a full size HP turbine. That's.... mindblowing.

 

[Reika]

So I think the HP turbine as a fixed max speed, and dynamic torque.

This is true for both turbine types.

 

[Demosthenex]

I was just made aware on IRC that the vertically tiled multiblock reactors are NOT pushing the waste down to the bottom for extraction even though the fuel pellets do move to the bottom and populate all of the cores. I'm on 23c still.

Be aware this can completely ruin the output from the taller reactors and invalidates my testing results of the vertically tiled cores.

Nice catch DevonX!

 

[Reika]

I was just made aware on IRC that the vertically tiled multiblock reactors are NOT pushing the waste down to the bottom for extraction even though the fuel pellets do move to the bottom and populate all of the cores. I'm on 23c still.

Be aware this can completely ruin the output from the taller reactors and invalidates my testing results of the vertically tiled cores.

Nice catch DevonX!

I will fix this.

 

[Demosthenex]

I've started to experiment with waste disposal.

Here's what I've found for easiest automation. While the waste is dangerous and so both the short term waste container and long term disposal drum should be fully submerged with water on as many sides as possible, item ducts work fine for transporting the waste to them.

The short term container will accept any kind of waste, while the long term drum will only accept waste with a half life over 5 years. With item ducts this means you can set the vacuum priority mode on the drums, and the waste they reject will goto the short term containers instead.

I do not yet know if the short term containers need to have waste moved to long term storage later. The experiment continues!

 

[Reika]

item ducts work fine for transporting the waste to them.

Bad idea. They like to forcibly stack the waste inside the containers, leading to rapid overheating.