fusion reactor help

[eric167]

ok, I need serious help automating this thing.
the threads that are out there are few and far between, and the information can be confusing.

I want to use AE.
I don't mind using overclockers and transformer upgrades to speed things up.

I want to have separate lines for tritium and deuterium, with as little excess as possible.
minimum rate is 1 cell of each every 6.4 seconds.

I would like pictures of well-laid out setups.

 

[netmc]

With AE, things are pretty simple. There is instant movement from one area of the network to the other, so you don't have to situate machines next to each other to deal with latency.

Automation for tritium and deuterium is pretty straight forward. I'm not going to go into overclockers except to say that in the version of GT I am on, it reduces production time to 70% and increases energy usage to 160%. This has changed a bit in various GT versions, so I'm going with the standard recipes. I'm also not going to bother with the EU costs as if you are powering a fusion reactor, you should have plenty of energy generation.

The base time for a tritium cell is 150 seconds. If you need one every 6.4 seconds, take 150 seconds and divide by 6.4 which gives 23.43, or 24 rounded up. So, to create one tritium cell every 6.4 seconds, you need 24 centrifuges producing tritium.

The base time for a deuterium cell is 150 seconds as well, so again 24 centrifuges. This is the final production stage for the fuel to feed the fusion reactor.

To generate the deuterium needed to power the tritium reaction (you need 4 deuterium for each tritium produced), you will need 4 times the number of tritium. If you use the simple number of 24, you will need 96 additional centrifuges creating deuterium to power the tritium reaction. If you use the actual production value for 6.4 tritium/second, you actually end up with only 93.75 (or 94 rounded up) saving two centrifuges.

So, we have 24 centrifuges producing tritium, and 118 (24+94) producing deuterium.

Now, each deuterium cell take 4 hydrogen cell to produce. 118 centrifuges producing deuterium time 4 hydrogen cells needed for each cycle giving 472 hydrogen cells needed every 150 seconds (run time of each centrifuge cycle). The time it takes to create 4 hydrogen cells from water in an industrial electrolyzer is 38 seconds. So, we get just shy of 4 full cycles (3.94) in the electrolyzer for each centrifuge cycle. If you divide the total number of hydrogen needed by the 4 hydrogen cells produced then by 3.94 cycles, it ends up at 29.89 (30 rounded up). Thus, 30 industrial electrolyzers are needed to produce the 472 hydrogen needed to power the other reactions.

Grand totals:

24 Industrial Centrifuges producing tritium
118 Industrial Centrifuges producing deuterium
30 Industrial Electrolyzers producing hydrogen

This will give you one set of tritium/deuterium fuel every 6.4 seconds.

(I don't think I missed up any of the math in this, but you may want to double check.)

There are all kinds of ways you can do the actual automation. You can just use AE import and export buses on each machine, or factorization routers feeding the different machine banks, or any number of alternatives. I'll leave the specifics of how you automate the machines with AE up to you.

Edit: These machine counts should give you a good starting place on setting up space and power for them. With AE, the layout doesn't have to be the most efficient, as long as all the machines are accounted for and connected to the AE network.

 

[eric167]

does block facing matter for AE?
do I have to stick the I/O busses on the right side of the electrolyzer/centrifuge, or is that taken care of for me?

 

[netmc]

For the electolyzer and centrifuge, it doesn't matter now. The cells and inputs can go in on any side. Output can be extracted on any side. There are several of the GT machines that have a square on one of the faces. That is the output slot. For those machines, all other sides can be used as input except the front.

Edit: Clarified the last sentence.

 

[eric167]

guess I have to wait for all the centrifuges to fill up, as right now half of em arnt on, and im not getting the rates that I need.

 

[netmc]

That's the one downside to AE automation. It's hard to balance the inputs if the machines aren't full of resources. A few routers from factorization can be a big help since they automatically distribute in a round Robin fashion.

Make sure to turn off the power to your final machines as you don't actually want to have them run yet if you are pre filling their inventories.

 

[eric167]

im just not charging the reactor, which lets everything back up from there.

 

[Harvest88]

Better yep get a crapton of OCed Heatvents if your sinking in Copper and a decent amount of gold. If your not rich in copper but don't mind forking a ton of diamonds you'll want Advanced Heat vents then goes crazy with them. Use 3 sides of the centrifuges to automate it (top for inputting cells and side for pulling out, also a power line on the bottom of the machines) then the remaining 3 faces just spread out them to make sure every vent is exposed to air. Done for the entire production you'll use 9% less energy (more profit per reaction!) and reduce machines mats down by 9% so you'll only need 91% of the machines required as well as power. How you ask? Every OCed or Adv. heatvent exposed to air on a machine will decrease operation time by 3% and therefore at the same time decreasing energy by that same 3% so 3x3 will reduce both machines and power supply demand to 91% if the entire facility is optimized like that. Only bad thing you'll need quite some room for all air circulation for the vents to work. Edit oh I forgot you need four sides for the electrolyers so for those you'll only have two faces so you'll miss out on a solid 9% reduction per reaction However you'll still save quite some juice.

 

[tonic316]

Is there a machine that will output energy at 8192 eu/tick besides an IDSU? Because once you have a fusion reactor you're limited to the 2048eu a tick from a plasma generator which kind of sucks if you want to have a industrial blast furnace with 3 overclockers. I know you can upgrade their internal storage to 100million eu which means it can go at 3 overclockers while taking an energy loss being powered by a plasma generator. But is there an AESU type option that will output at 8192 so we can have a double buffer for things like the industrial blast furnace running at 3 overclockers?

 

[eric167]

supercondensator.

you want two.

the first accepts the output of your generators and bumps it to 1 million per packet.
the second dumps into something that can take 8192.
you WILL need two HV transformer upgrades for any machine to take the 8192 though.

 

[Siro]

Look at upgrading the various charging boxes or the machine box. If I recall correctly, they can be upgraded to output 8192. Machine boxes are also a way to transfer items, power and liquid through one face (via conveyor/pump modules), which will allow you to optimize energy usage on the electrolyzers better. And they can also be a connection point for a router for extracting. This means you could heat vent 5 faces on both centrifuges and electrolyzers. Quantum tanks should be use for storing tritium and the majority of the deuterium (level emitters should keep a minimum stock for tritium processing) so that the cells can be reused.

Personally I prefer to quarry ridiculous amounts of endstone from an endstone mystcraft age and do the he-3 h2 reaction instead as it keeps me involved in the process (moving the quarry every now and then) and separates my fuel production lines. It needs a lot of macerators, but far fewer centrifuges. The caveat to doing this is that 6 energy injectors are needed instead of 4 to start the reaction, but need less continuous energy to keep it running (meaning a minimum of 1 AESU instead of 2).

 

[tonic316]

supercondensator.

you want two.

the first accepts the output of your generators and bumps it to 1 million per packet.
the second dumps into something that can take 8192.
you WILL need two HV transformer upgrades for any machine to take the 8192 though.

supercondensator in the new GT outputs 10,000,000 eu a tick :/ not 8192. And making a industrial blast furnace take 8192 eu/tick gets blown up when being powered by a super condensator.