All the different ways to increase the efficiency of a nuclear power plant

[mrvn]

Can someone show proof what the optimal number and placement for heat exchangers is? Assume a tileable setup, i.e. an infinite reactor.

I don't have math, but what I did with my 2xN layout was to have every single tile not strictly being used for something else (fuel in/out & power supply) be a heat pipe. Because of the tiling requirements I needed a horizontal bus *anyway*, and I simply made that one tile wider than needed in order to gain a lot of buffer capacity without a major lengthening of the heat exchanger rows.

Images here: viewtopic.php?f=208&t=87432&p=504270#p504270

Obviously going wider instead of longer is a benefit and using any tile left over can't hurt.

Your heat exchangers are further away from the reactors compared to my setup though. I just have one line for inserters and one for underground belts and buffer chests and then a water hole for the offshore pump. The belt is half fuel cells and half spend fuel cells, doing double duty. Getting them nearer to the reactor increases the temperature range one can use. So my thinking is that the extra heatpipes would be better on the other end, between the heat exchangers and the turbines.

But your design touches on a question I have about the best layout of the heat exchangers. I have a single line of heat exchangers per reactors. Don't even have them cross connected with heat pipes at all. You have double line of heat exchangers, potentially a more dense design. But then you waste one tile between heat exchangers where the steam comes out. (I do too because it aligns to each reactor for me and I don't have to cheat by landfilling the offshore pumps then). But a denser design should give more heat capacity if one goes for the optimal setup.

[foamy]

Can someone show proof what the optimal number and placement for heat exchangers is? Assume a tileable setup, i.e. an infinite reactor.

I don't have math, but what I did with my 2xN layout was to have every single tile not strictly being used for something else (fuel in/out & power supply) be a heat pipe. Because of the tiling requirements I needed a horizontal bus *anyway*, and I simply made that one tile wider than needed in order to gain a lot of buffer capacity without a major lengthening of the heat exchanger rows.

Images here: viewtopic.php?f=208&t=87432&p=504270#p504270

Obviously going wider instead of longer is a benefit and using any tile left over can't hurt.

Your heat exchangers are further away from the reactors compared to my setup though. I just have one line for inserters and one for underground belts and buffer chests and then a water hole for the offshore pump. The belt is half fuel cells and half spend fuel cells, doing double duty. Getting them nearer to the reactor increases the temperature range one can use. So my thinking is that the extra heatpipes would be better on the other end, between the heat exchangers and the turbines.

But your design touches on a question I have about the best layout of the heat exchangers. I have a single line of heat exchangers per reactors. Don't even have them cross connected with heat pipes at all. You have double line of heat exchangers, potentially a more dense design. But then you waste one tile between heat exchangers where the steam comes out. (I do too because it aligns to each reactor for me and I don't have to cheat by landfilling the offshore pumps then). But a denser design should give more heat capacity if one goes for the optimal setup.

Reason for the extra space between the exchangers is fluid flow concerns. To avoid the use of pumps, I need two pipes of output; each pipe pulls from 12 heat exchangers, for just a smidge over 1200 steam/sec. The interval between the end of each row and the start of the turbines is six pipe segments in the worst case, which can handle ~1500 steam/sec. If you use pumps you could go double-wide on the heat pipes, but then you're significantly more vulnerable to a brownout spiral and need to allocate a fairly beefy UPS to avoid it.

Thinking about it now, though, you might be able to tweak things so that the initial heat piping is double-width and the fluid piping is single-width, and then trade part-way down. That'd get you a bit more buffer space in each spike. However you'd also need two lengths of fluid pipe and an extra vertical tile in order to realign the exchanger rows, and it's already pretty tight tolerances on getting water and heat flow to the terminal exchangers. Might be worth experimenting with.

The thing about putting the buffering behind the exchangers is that they'd have reduced thermal range, which cuts into their efficiency relative to a steam tank. I think if I were to try and add more buffer capacity in, by far the simplest way would be to stack a tank at the end of every turbine row. It's another three tiles per side, but it doesn't cause any kind of flow concerns and would give you a fairly significant boost of ~4GJ of buffer per reactor.

[Impatient]

Can someone show proof what the optimal number and placement for heat exchangers is? Assume a tileable setup, i.e. an infinite reactor.

Hmmm ... if we are talking infinite then ...

1. A reactor in a double row (2xN) layout with x3 neighbor bonus produces 160MW
2. A HX consumes 10MW

=> One reactor needs 16 HX

3. 16 HX produce 16x103steam/s = 1648 steam/s

=> 1648 water/s in, 1648 steam/s out, 1648 fluid/s => safely within the limits of fluid throughput

4. A reactor is 5 tiles wide and a row of HP-HX-pipe is 4 tiles wide

=> you could tile the row of HX comfortably with each reactor. One connection to the reactor would be sufficient. Using the 5th tile to insert and extract fuel cells (with 2x longhanded inserters and interwoven ubelts, no roboports needed)

The connection to each reactor would be very close.

2021-09-29 Factorio HX setup.jpg (144.43 KiB) Viewed 5890 times

But IMO it is not important how close the HX are to the reactor in a setup that is tilable with a reactor. Having space for the fuel cell logistics close to the reactor is as good.

This also produces 160MW of steam and is much more comfortable.

2021-09-29 Factorio HX setup comfy.jpg (131.52 KiB) Viewed 5883 times

Hm ... by your comments, I think the only thing to be more optimal is the water supply. I use two offshore pumps per water pipeline.

[Impatient]

This one accounts for the limit of offshore pumps.

2N, 6 reactors infinitely tillable HX layout
- For 2N reactor layouts
- Every 6 reactors infinitely tillable
- 6R x 160MW/R = 960MW = 10MW/HX x 96HX
- Max. water intake/pipeline: 103.1w/s x 10 = 1031w/s (safe for offshore pump)
- Max steam output/pipeline: 2062 st/s (allows for 3 pipes between pumps)
- Substations are necessary for steam pumps.
- Robot logistics for infinite tillability.
- Dimensions: 30x41 (including u-pipes), 30x40 (with offshore pumps)

0eNq1nUtv3MgVhf8K0asZQO1hvUiWgOySwFnMLJxFBkgMoyVRFpFWd6cftgVD/z1syR61bR7XOdejlSU3+ekWeevBe3iqP84ulod+sx1W+9n5x9lVv7vcDpv9sF7Nzmfz6u/rbeV/q7b94nI//rhc3K0P+91/VvPqb+/67V3VfP5oVw2r62E17PvlXbUflsvFxbI/Hte8qj5Urql//dcvr6q/VPn40/ivO/7Hy9/Hz3Lz8vfjgb8uPryo3i/2/XZE7Rf/7X/ZDJt+Oaz68/Ho8MK9/2V3RNUPZ4eHX3/aLa776nqMbH19vbtZb/tqc7jd/PyJV+32/eK2GkPeHPYnOF83fvzsCFgsl+v3uwdEqI5H7KqLfv++71cPpN0D6p+Hi91+cbwou2ox/o1Vf9nvdoux/cfzHv/Iw9Evjke/Wl+s99Vy/XbY7YfLR/bni/Pp0gzLYX/3cPBfh9t+tTuSz6tQf4iu+mlYXS4PV8PqbXWYP0T089nDR3X10/thf/NlU48BXtxV/7jdjPH1q/3YNu/mdZ77PDubDZcjeHb+74+z3fB2tVgeb/H+btOP93YM5nY8YrW4Pf62Olwu+8V2/uluzu7Hc1dX/YfZubs/K5590y/2D6GenOfZ8/oPlzeL1dv+9I+GyZOvl4fh6unsh+t+clK8f302Gy/BMF6Ix0Y//HL3ZnW4vRjx5w4292y2We+Gx6z/OBtZ8/Qinc3uxh9807xI98d4vqL5ieZ/y6mfON00J6DL8Q3MnbDGZl8N2/7y8eM4wY00N0rcRHNbidvw16GWwC0PDhK448GNBM48OEtgV9Nk7zWy48lJI3ue3Gnkp653HEq/5QV32ne/4DVTvKcudxwN5vv1/O12fVhdTZD9l+NCmZ0KscZa4zV8rN13YvVT7KceN6x2/XY/da/m/ssR9kvq2Wycufbb9fLNRX+zeDeMY/R42uWwvTwM+ze366v+zfr6zXrTbxePZ4SpodU99dDP0/H88qbf7ceB/3+H8d/JwNzXgX06+M31sBzPeJxaPs+Of/yBw3axGg638+tDv5xf9svl7NiGw3FZler711PRZXUGacFMVKsgNKW5L5LiuzMQQvgCwpcRoYAIZUQsIIjpPRUQsYxoCoimjGgLiLaM6AqIXEbkAqIrIkJdSq1yboViepZTPJTy05UTNJQS1JUzNJQy1JVTNJRS1JVzNJRy1JUzLJSS1JUTPZSy1BE5VkpTV071WMpTX86xWMpTX8716Avz/XGIQ7PnFK+Us76c97GUs76cs7G0jjmOeEq7+HWMz99ZcUytkWIpr305J2NnWmdxbc/0mpDipVpctyXmKiZHrU5OsHH6UiaPVnPjQ/zwrp9vtut3w1VpTZcAPVBhunKYUV2NIVBSQahpjaXEMI1qLVWFaVRnKSRMo7KpdjDJampTuWCa5UwVgmmWNxUFplnBVAaYZkXTg/80K5ke9adZDf1wH5lhrGlND/ccu6MHco6XTZNOZCaJtrZVUd302NQ6dZDzACTXYwMAUROBK3OoeYCIh5oGfJnTWIZuT9RXWybAUA6wYzixzKHqKqnI6WrLBEVcsI7K+bYcIJXyTZkTTNMn01KqD+RyhFQf6Mqchurb5c7dUTnvyr2760zLDebaU93AlQeOTJUXXbmDZ9tyiGhrpnqCK/f5zA3+5UEoU5nvyp08U6nvyr08N6YFJHP1uc5Q7uaZmgFcuZ/nbFrgekZBo3qDLw8hrqamAV8TJCr9fSBIVP57T5BszwXULeBWRIkIkpoPfCRIrenRhWou1St8SwRJTQm+PJY4x/WCjiA5+jnNU8Km86YHNRIe6Cc1EhhNj2qe0mFPRGOuoOeeS4g9kZu5SGru6rXqQ59DadippBqR8p9SpkT4E7HXpmm7Z9W0ixJyJJroefUWMgKvvUJG5MVXyEi8dAoZDa/gQkYraKcQ0gkiLoRkQYFFkLKcHAiIE/RTCPGCkAshQVBQISQKUi6EJEGHhZBGEGIhpBVUVAjpBDkXQrKgnSJIWVgmMrasLBMZG70g40JIoHXcmlqrxGgScrl1QkyCkgub/JTVuz9eCP/eetWNka364e3NxfrwMMWG2ExOmrE1LQLJC8uLCyQwmxSABBRydyI7f++6zv1nUPf1dY1u+rrK0nPKKEZVXcCkoJI6RIqiYo1jSgZpIOVpVmOomiNWaylMI5ip0opg2VLKBDBFcs5FmLOUvhDMW4o4CBYsJRIEi2zZIJHvQydL2YCFN+wYzAJNM0biXt8+kaOph/X0bO9vN/nHHnbT877A3aovXsM5UJbBU4tI3vCyT2qIOiQnjDMhRk1hxyDTvEW1tdFEdxyiKI5jkKiOY5Aoj0OQSR+nrj4nkDdEiF5T2jHIJJFzjRU1chyjKJJjUCOK25jUino7JtlWb9QNyKK+DYMklXJi2LBJ5VRzZa0cB6mK5ZgURZEbk5Kou2OSSS/nbkEr6tw4yE6U3jHJJJkzzfWyZJ6g0U2VzDHJi0I3JgVRfMckk2TO3YIkCt04yEYU3zHJJJlzzVUlcxykKplDkiyZYxItmSfK5+JtkjkLpyVzFmiSzBNlgfGqZJ4SF3OjViIjuvntnyEvp4TwnVpUhIGq9mZIkv3NsHWCwRlH42l7MmYE2iWNGbzHGTN4kzNm8C5nzOBtzpjR0Uo7ZmRaaYeMsjRNpHxZmiYyVfE6Y4hgdsaQyCvtGJJ4pR1DGl5px5CWV9oxRHA8Y4hgeYaQsjRNJFtZmibSnnc9J8qs5BXbM45K8D1jCG18Zptmcj4nykDmFeszbrLJ+8w2nzY/k8BkM7Ylh/YPceJKMjyTiuJPxGsuEurVVX+iZNvEmfCc4ow/kcd/ZKHs0e1V3dkpIFKjkmBMrUX7AUnTWbQVwMoWpWCapYjldRHmTFVlAPOmmi2ABVP1EcCiqbYHYMlUpQKwxlQDArCWrotwI1zTmeoiJJyfwDhgW5umW8pK4E8E6u36Yr1Zb/dT4+YnaMzTm1l5tQqCJtdW3a0Dk9TtOjApibJ2jUCNCIIRtaJiD0GdqGJDULbMKjVR8u1qUbFHIXLiM7EA5MTnSICCZfqkLloUhXYYYhKFdggybQPLNbYVhXYYYycK7RCUVaEdkUjlmRgzsm1JxNwAUnkmhhFSeSZ6e46mRRvV3KQK7TDIRhXaIalVhXZI6lShHZJM2x0ztyBwyrMrd9XAKc8uEyRvWoxTzQ2q0A6DjKrQDklJFdohqVGFdkhqTY8w1C3oVKEdBplVoR2RlF2/O625zqlCOwzSq0I7JAVVaIck/iVzyocUnO0lcxLOv2ROAm0vmVMuonAiNCvl1gj3Bs7aQ1PMaIPgWgSh73Uois2JYHhaoMUM3gqNGbwVGjN4KzRm8FZozBCs0BgiWKExRLBCQ4hihcYQwQqNIYIVGkMEKzSGCFZoDBGs0BgiWKExRLBCY4hghcYQwQoNIYoVGkMEKzSGCFZoDKGt0JHyXwWbFTpSzrGgWKFxk0UrdMxfW3bdtBU62KzQ7IWlrdAsMGvKanwup1xIP7gtTHxWp1zgnNgn4bToKw+8CILfnRAMlejYTrOioUCLWMlSAEWwxlLMQ7DWUipDsM5S9EEwk3cBwASB+XR2ADBnefRGMM8+jkbqTerQBMvjKAuP9GBLApNpaqDe+w6NuEFZTFzMrah4xoi+p6X7M15niQnhsyiowkBlUzQmqaZo2DpOeGZAlm9rjJGopKmmaByiuF84Blk2DOfaKpqicYiiKRqDRFM0BJlM0dRFU03ROETRFI1BJlM011jRFI1jFE3RGKSaojFJNUVjkskUzd0A1RQNg5RN0ZhkMkVTzZVN0ThI1RSNSaopGpNUUzQmmUzR3C1QTdE4SNUUjUm2BwuiuVE2RaMgo2yKxiTVFI1Jqikak0ymaO4WqKZoHKRqisYkkymaa65qisZBqqZoSJJN0ZhEm6IjZYGJNlM0C6dN0SzQZIqOlDUmnsjWklbr0L1qtB0fY/imfJzq15NkdUPwGFCM6obg0SNSpk3HkOFr2nSMGY42HWOGoEdDRuB1cciIvKYNGYnXtCGj4TVtyGh5TRsyOkHThpAsaNoIUpajiTQry9FEvpflaCJZy3I0ka1lOZpI17IcTeRrWY4mErZsfyYytixHM8mWBU0bQU7l6O+bjiPllYllaZroApI0DSG8NE02zSZNUwamKEnTsMmNaSFENr+ll20kUNxqNT7XN7REWcqm3iGMP6xLP+vXlcQTXfpHVJUafdm5um04XDLL24bjmKJFwnDTLMsGrIhl2TgcsUwbhyOYrcoKYKaNwwHMtnE4gpk2Dkcw08bhCGbaOBzB6He6yRHOtnE4C6ff6WaBtverqHe6Y9OJancAL2FHWY6GJFmOxiTxHSYMEtVoDAqajopBUdOMMShpOioGNZqyi0GtJptiUKfJphiUNW0TgjjLc0eAnKhtYpIXtU1MCqIAiUlRFCAxSXV0YpLq6MQk1dGJSaqjE5OyKOVBEikeE50lO1FvwyQv6m2YpH7bLiZFUW/DJFUUwyRVFMOkVlSuMKkTlStMyqJy9Uh6fTYb9v3teNbF8tBvtsPquI/KcnHRL8f/87+dVU31SVvZVcPqeliNhy/vqv2wHA9a9tXL36vl4m59OJ72bnwUflyNdW6cVnwbQzOmZ3d//38HQB2Q

Copy blueprint

2021-09-29 Factorio HX setup comfy 6.jpg (491.97 KiB) Viewed 5870 times

[xaetral]

Nice, here's mine:



Notice that the pumps would be quite close from a side to another, which helps a lot fitting it into a natural lake (you may also use 3 pumps per two reactors instead of 4 pumps).

[mrvn]

Can someone show proof what the optimal number and placement for heat exchangers is? Assume a tileable setup, i.e. an infinite reactor.

Hmmm ... if we are talking infinite then ...

1. A reactor in a double row (2xN) layout with x3 neighbor bonus produces 160MW
2. A HX consumes 10MW

=> One reactor needs 16 HX

3. 16 HX produce 16x103steam/s = 1648 steam/s

=> 1648 water/s in, 1648 steam/s out, 1648 fluid/s => safely within the limits of fluid throughput

4. A reactor is 5 tiles wide and a row of HP-HX-pipe is 4 tiles wide

=> you could tile the row of HX comfortably with each reactor. One connection to the reactor would be sufficient. Using the 5th tile to insert and extract fuel cells (with 2x longhanded inserters and interwoven ubelts, no roboports needed)

The connection to each reactor would be very close.

2021-09-29 Factorio HX setup.jpg


But IMO it is not important how close the HX are to the reactor in a setup that is tilable with a reactor. Having space for the fuel cell logistics close to the reactor is as good.

This also produces 160MW of steam and is much more comfortable.
2021-09-29 Factorio HX setup comfy.jpg


Hm ... by your comments, I think the only thing to be more optimal is the water supply. I use two offshore pumps per water pipeline.

An offshore pump only does 1200 fluid/s. So you need 2 offshore pumps for 16 heat exchangers.

Nice, here's mine:



Notice that the pumps would be quite close from a side to another, which helps a lot fitting it into a natural lake (you may also use 3 pumps per two reactors instead of 4 pumps).

Similar problem, the pumps only manage 1200 steam/s so that's a bottleneck.

You also have a lot of free space between the heat exchangers. The idea was to move them closer together so more heatpipes fit.

[foamy]

I hate having to power pumps. If it weren't for the fact it's flat impossible to use a burner inserter to extract the fuel cells I wouldn't even run power poles through the reactor input/outputs.

[Impatient]

Assuming you write that, because you don't like to depend on power to be available to produce power: Have an inner and an outer power grid. The inner power grid would just power the infrastructure to insert fuel cells and it would have its own steam buffer+turbines.
Edit: The inner grid ofc must also include the pumps.

But let's face it: If you build lossless and circuit controlled on-demand reators, you will never experience an unplanned shutdown unless your supply of cells runs dry or your reactor or the infrastructure gets destroyed. But in those cases an unbuffered not circuit controlled reactor will also shut down unplanned.

[Impatient]

Hm ... by your comments, I think the only thing to be more optimal is the water supply. I use two offshore pumps per water pipeline.

An offshore pump only does 1200 fluid/s. So you need 2 offshore pumps for 16 heat exchangers.

Which I literally recognized in the sentence above your answer

[Bauer]

Assuming you write that, because you don't like to depend on power to be available to produce power: Have an inner and an outer power grid. The inner power grid would just power the infrastructure to insert fuel cells and it would have its own steam buffer+turbines.

But let's face it: If you build lossless and circuit controlled on-demand reators, you will never experience an unplanned shutdown unless your supply of cells runs dry or your reactor or the infrastructure gets destroyed. But in those cases an unbuffered not circuit controlled reactor will also shut down unplanned.

We all know how quickly a brown-out can turn into a black-out. The inner grid of a nuclear plant only needs to power a few inserters, pumps and maybe the fuel cell recovery and/or a few bot charging stations. Plunking down a few solar panels for this is no doubt a best practise.

[foamy]

Assuming you write that, because you don't like to depend on power to be available to produce power: Have an inner and an outer power grid. The inner power grid would just power the infrastructure to insert fuel cells and it would have its own steam buffer+turbines.
Edit: The inner grid ofc must also include the pumps.

But let's face it: If you build lossless and circuit controlled on-demand reators, you will never experience an unplanned shutdown unless your supply of cells runs dry or your reactor or the infrastructure gets destroyed. But in those cases an unbuffered not circuit controlled reactor will also shut down unplanned.

The issue would be with overbuilding without uprating the reactor, which I've done a time or five :v As you say, separating the grids works -- I usually pull from the mainline turbines and do some circuit magic to control whether or not that's connected to the outside world or not -- but...

Well, you can make a reactor work without pumps. You can even feed it with burner inserters and belts. That leaves the only thing requiring power in the entire array being the ejection inserters... which means you're still stuck running power through the whole thing. If I could only eliminate that one dependency I wouldn't need power poles at all.

[Bauer]

Well, you can make a reactor work without pumps. You can even feed it with burner inserters and belts. That leaves the only thing requiring power in the entire array being the ejection inserters... which means you're still stuck running power through the whole thing. If I could only eliminate that one dependency I wouldn't need power poles at all.

You can fuel burner inserters with burner inserters. The later fuel themselves.

[foamy]

Well, you can make a reactor work without pumps. You can even feed it with burner inserters and belts. That leaves the only thing requiring power in the entire array being the ejection inserters... which means you're still stuck running power through the whole thing. If I could only eliminate that one dependency I wouldn't need power poles at all.

You can fuel burner inserters with burner inserters. The later fuel themselves.

I suppose, but you're no longer saving space over a powered one. You would still be brownout proof, though.

[gGeorg]

• Larger reactors should be fueled all simultaneous. So you should check that each reactor has a fuel cell ready to go in. If your fuel production is too slow and you only fill a few reactors then you loose the neighbour bonus and you go into a death spiral with each cycle producing less energy so less fuel.
  
  I don't control the fueling of the reactor using steam or accumulators but rather I have one reactor that is designated as master. For the master the inserter removing the spend fuel cell is controlled by whatever method you prefer.

Check the presence of fuell for each reactor is not needed. In fact for bigger systems or for higher reliablility is wise to do not check all. Also, it is better to use as "master" any reactor. As result, the logic is "when any number of reactors is ready to start cycle, then do it in case steam is needed." I used this logic in my Perfect clover leaves desing, and after 2 years of usage, I can say it is bullet proof.
viewtopic.php?f=208&t=96233
Try it yourself, then let me know if you find a failure. All the control logic cost just 5 combinators.

[foamy]

Try it yourself, then let me know if you find a failure. All the control logic cost just 5 combinators.

That's four combinators more than you need for synchronous insertion, and three more than you need for synchronous insertion with guaranteed available cells or steam monitoring. Two and a half (counting a CC as a half) if you want both.

[farcast]

• Larger reactors should be fueled all simultaneous. So you should check that each reactor has a fuel cell ready to go in. If your fuel production is too slow and you only fill a few reactors then you loose the neighbour bonus and you go into a death spiral with each cycle producing less energy so less fuel.
  
  I don't control the fueling of the reactor using steam or accumulators but rather I have one reactor that is designated as master. For the master the inserter removing the spend fuel cell is controlled by whatever method you prefer.

If fuel cells aren't being replenished fast enough for every reactor to have a fuel cell when needed, then that's a supply problem, not a circuit problem. Forcibly waiting for every reactor to have a fuel cell would just trade a gradually worsening brownout for a periodic blackout that lasts longer and longer. Sure, the fuel cells would be used more efficiently, but that will only buy you time and not solve the actual problem.

[Khagan]

All the control logic cost just 5 combinators.

That's four combinators more than you need for synchronous insertion, and three more than you need for synchronous insertion with guaranteed available cells or steam monitoring. Two and a half (counting a CC as a half) if you want both.

Make that five more than you need for synchronous insertion with steam monitoring.

If fuel cells aren't being replenished fast enough for every reactor to have a fuel cell when needed, then that's a supply problem, not a circuit problem.

Indeed.

[foamy]

Make that five more than you need for synchronous insertion with steam monitoring.

I'm not sure how you'd do a combinator-free regulated insertion? I'd be interested in seeing the logic. Combinator-free steam control for a limited-capacity reactor is simple, but without some form of memory I don't see how you don't wind up inserting multiple cells.

EDIT: I suppose synchronicity is assured by simply having a sufficient fuel supply and buffer, which technically requires no circuits at all on the reactor end of things, so I was sloppy in my wording earlier. I meant in terms of ensuring all reactors get precisely one cell in the context of a regulated reactor that's designed to be low-loss or lossless via storing the energy in a buffer of some from.

[mrvn]

• Larger reactors should be fueled all simultaneous. So you should check that each reactor has a fuel cell ready to go in. If your fuel production is too slow and you only fill a few reactors then you loose the neighbour bonus and you go into a death spiral with each cycle producing less energy so less fuel.
  
  I don't control the fueling of the reactor using steam or accumulators but rather I have one reactor that is designated as master. For the master the inserter removing the spend fuel cell is controlled by whatever method you prefer.

Check the presence of fuell for each reactor is not needed. In fact for bigger systems or for higher reliablility is wise to do not check all. Also, it is better to use as "master" any reactor. As result, the logic is "when any number of reactors is ready to start cycle, then do it in case steam is needed." I used this logic in my Perfect clover leaves desing, and after 2 years of usage, I can say it is bullet proof.
viewtopic.php?f=208&t=96233
Try it yourself, then let me know if you find a failure. All the control logic cost just 5 combinators.

It is very much needed

A) at start up when fuel cells might just not have reached every reactor
B) when short of fuel

Running only part of the reactors is hugely wasteful as you loose neighbor bonuses and this thread is about efficiency. Doing that at startup just delays the time till the reactor is ready to produce power. And when being short of fuel it increases the slope of the death spiral. It might prevent a larger power shortage by running with partial reactors but at the cost of more fuel making the shortage worse.

While you might say that situation is broken anyway and you should never let it get the fact is shit happens. The biggest reason to check for fuel in the reactor is to connect an alarm to it so you notice. And if you check anyway you might as well use it in the control logic too. Only costs an extra wire then.

[mrvn]

Make that five more than you need for synchronous insertion with steam monitoring.

I'm not sure how you'd do a combinator-free regulated insertion? I'd be interested in seeing the logic. Combinator-free steam control for a limited-capacity reactor is simple, but without some form of memory I don't see how you don't wind up inserting multiple cells.

EDIT: I suppose synchronicity is assured by simply having a sufficient fuel supply and buffer, which technically requires no circuits at all on the reactor end of things, so I was sloppy in my wording earlier. I meant in terms of ensuring all reactors get precisely one cell in the context of a regulated reactor that's designed to be low-loss or lossless via storing the energy in a buffer of some from.

You need a few steps in the control logic:

1. every reactor needs a buffer chest with an inserter wired to insert exactly one fuel cell
2. designate one reactor the master, use a chain of 2 inserters (Master A and Master B) to remove the spend fuel cell there
3. wire Master A to the steam tanks and remove a spend fuel cell when steam is low
4. wire Master B to all buffer chests, set to enable when fuel cells >= reactor count, read hand content
5. wire all inserters that insert fuel to Master B and enable when spend fuel > 0

The spend fuel cell acts as a timer preventing the insertion of multiple fuel cells. The Master A/B inserters replace a combinator, setting the condition for fueling to "steam is low" and "fuel is available".

For me the problem is this control circuit is that it requires to program in the number of reactors in the setup. For a tileable reactor design that is annoying. I use extra combinators in my design to negate the signal from the fuel chests so I have a "fuel is missing" signal that is independent of the number of reactors. But you don't need any combinators if you really don't want to.