|
Post by shurugal on Dec 20, 2016 12:42:50 GMT
I don't understand why we don't have steam turbines. Our ships are already huge and massive, why can't we have a 10-ton steam turbine that will be half the size and twice the efficiency of a 100-ton thermocouple? A Carnot engine with a hot side of 3000K and cold side of 2500K has an efficiency of 16.7%. Reactors we run ingame at those conditions come in at about 14.6%. They are so close to Carnot engines as is that another engine's efficiency gain wouldn't really matter (aside from maybe increasing the delta T). IIRC, turbines tend to get ~ 8 kW/kg power density on the high end IRL, whereas ingame total reactors (minus radiators) get about 40 kW/kg. Would still be interesting to try to make a turbine that beats them given we could probably stretch delta T to be much wider and ignore creep limitations of materials, but I would rather have a few other things first (eg. coilguns not exceeding 100% efficiency). part of the point of using steam is dropping the cold side further, putting us into much more favorable carnot ranges. If we get the waste heat down, then it won't matter that the radiators are less efficient, there will be drastically less heat for them to reject.
|
|
|
Post by coaxjack on Dec 20, 2016 21:25:00 GMT
Are the quoted power to weight ratios for real-world existing reactors? If so, I think we can easily squeeze a substantial amount of power out of a notional turbomachine design if we run them suicidally close to redline typical of designs in this game.
|
|
|
Post by newageofpower on Dec 20, 2016 23:10:14 GMT
Are the quoted power to weight ratios for real-world existing reactors? If so, I think we can easily squeeze a substantial amount of power out of a notional turbomachine design if we run them suicidally close to redline typical of designs in this game. Military fighter engines (indeed, the entire fighter) have sub 5% safety margins. I design my civilian grade reactors with an 80 kelvin margin and my military reactors with a 5 kelvin margin.
|
|
|
Post by shurugal on Dec 20, 2016 23:33:23 GMT
Are the quoted power to weight ratios for real-world existing reactors? If so, I think we can easily squeeze a substantial amount of power out of a notional turbomachine design if we run them suicidally close to redline typical of designs in this game. Military fighter engines (indeed, the entire fighter) have sub 5% safety margins. Sauce on that? I'm hardcore into flight simming and military aviation, and I can tell you that there is usually a safety factor of at least 30-50% anywhere possible.
|
|
|
Post by newageofpower on Dec 20, 2016 23:38:34 GMT
Hmm. I recall reading an article (on SciAm, I think, so I really should check another source) explaining how a 5% deformation in the airframe could lead to the total destruction of the aircraft during high supersonic flight.
|
|
|
Post by cuddlefish on Dec 21, 2016 0:01:29 GMT
Hmm. I recall reading an article (on SciAm, I think, so I really should check another source) explaining how a 5% deformation in the airframe could lead to the total destruction of the aircraft during high supersonic flight. That doesn't really sound equivalent? 5% deformation of most precision machinery would cause disaster, but that's not what's usually is being talked about in terms of safety margins, is it?
|
|
|
Post by shurugal on Dec 21, 2016 0:08:27 GMT
Hmm. I recall reading an article (on SciAm, I think, so I really should check another source) explaining how a 5% deformation in the airframe could lead to the total destruction of the aircraft during high supersonic flight. while that is true, the stresses required to generate a 5% permanent deformation are quite significant. Most fighters, for example, are rated for ~9G along the lift vector. Permanent deformation typically sets in around 12G, depending on loadout and fuel-state. There is a significant difference between tight tolerances and safety factor. It might only take a small deformation to exceed tolerance, but it takes a significant force beyond what is rated as the maximum safe limit.
|
|
|
Post by n2maniac on Dec 21, 2016 5:22:53 GMT
A Carnot engine with a hot side of 3000K and cold side of 2500K has an efficiency of 16.7%. Reactors we run ingame at those conditions come in at about 14.6%. They are so close to Carnot engines as is that another engine's efficiency gain wouldn't really matter (aside from maybe increasing the delta T). IIRC, turbines tend to get ~ 8 kW/kg power density on the high end IRL, whereas ingame total reactors (minus radiators) get about 40 kW/kg. Would still be interesting to try to make a turbine that beats them given we could probably stretch delta T to be much wider and ignore creep limitations of materials, but I would rather have a few other things first (eg. coilguns not exceeding 100% efficiency). part of the point of using steam is dropping the cold side further, putting us into much more favorable carnot ranges. If we get the waste heat down, then it won't matter that the radiators are less efficient, there will be drastically less heat for them to reject. Please see: qswitched wrote on his blog that: I respectfully disagree, if you make sufficient assumptions. These are that the thermocouple exit temperature can be adjusted more or less freely for a given core temperature and electric output power, without significant mass penalties. Solving the Carnot efficiency equation and the Stefan-Boltzmann law yields these graphs. A few conclusions are readily apparent: 1) You should always run your core as hot as possible 2) There is a single optimum radiator temperature, and it is always three quarters of the thermocouple inlet temperature. Thermocouple delta T maximum is 500K, and the reactor hot side hits limitations at about 3100K. Efficiency gains may be about 1.5, but the radiator area would shrink by only about 10%.
|
|
|
Post by zuthal on Dec 21, 2016 10:26:07 GMT
Specifically, assuming that both heat engine efficiency (as a percentage of carnot) and radiator efficiency are independent of temperature, the optimum cold side temperature is 75% of the hot side temp, to optimise for radiator area per electrical power output. Earth-bound powerplants can push the cold side down to only a few tens of degrees above room temperature because on Earth, you can do open-cycle cooling, as you generally have large reservoirs of cold, dense fluid. You cannot do that in space.
|
|
|
Post by shurugal on Dec 21, 2016 13:05:03 GMT
Thermocouple delta T maximum is 500K, and the reactor hot side hits limitations at about 3100K. Efficiency gains may be about 1.5, but the radiator area would shrink by only about 10%. Thermocouple dT max is only 500k, but steam expansion has no such limit. If we expand the steam through a turbine first, then we convert a considerable portion of what would otherwise be waste heat to mechanical and electrical energy. Yes, a carnot engine operating between 3100 and 1600 has shit efficiency, but one operating between 3100 and, say, 600 has an efficiency of 80%. If we use a thermocouple to drop what is left from 600 to 100... you see what i'm getting at?
|
|
|
Post by newageofpower on Dec 21, 2016 16:51:18 GMT
Thermocouple delta T maximum is 500K, and the reactor hot side hits limitations at about 3100K. Efficiency gains may be about 1.5, but the radiator area would shrink by only about 10%. Thermocouple dT max is only 500k, but steam expansion has no such limit. If we expand the steam through a turbine first, then we convert a considerable portion of what would otherwise be waste heat to mechanical and electrical energy. Yes, a carnot engine operating between 3100 and 1600 has shit efficiency, but one operating between 3100 and, say, 600 has an efficiency of 80%. If we use a thermocouple to drop what is left from 600 to 100... you see what i'm getting at? Steam turbines rarely go over 45% efficient (modern natural gas plants tend to use a two stage approach; with the exhaust from a higher temperature gas turbine heating a lower temperature steam turbine). A quick google search shows the highest gross efficiency around 49%, with a net exceeding 48%. linkNote: article written in 2002. 2016 designs likely to prove superior. The majority of power generating steam turbines in existence (older designs due to long service lifespan) are 33-37% in efficiency. That being said, the potential efficiency gains are enormous, and that's without ludicrously cheap diamond, noble metals, and other 'exotic' materials available. Plus, powerplant turbomachinery is designed (and operated) with a 50 year operating lifespan in mind, cutting this margin may improve performance.
|
|
|
Post by n2maniac on Dec 21, 2016 18:28:11 GMT
Thermocouple delta T maximum is 500K, and the reactor hot side hits limitations at about 3100K. Efficiency gains may be about 1.5, but the radiator area would shrink by only about 10%. Thermocouple dT max is only 500k, but steam expansion has no such limit. If we expand the steam through a turbine first, then we convert a considerable portion of what would otherwise be waste heat to mechanical and electrical energy. Yes, a carnot engine operating between 3100 and 1600 has shit efficiency, but one operating between 3100 and, say, 600 has an efficiency of 80%. If we use a thermocouple to drop what is left from 600 to 100... you see what i'm getting at? I do. I also know it is a losing battle below about 2300K. Radiators scale by T^4. Lets say you have a system generating 1GW with 80% efficiency with waste heat at 600K versus another system generating 1GW at 15% efficiency with waste heat at 2400K. Each GW of waste heat on the 600K radiator requires 4^4 = 256x the area. But it generates less heat: 2400K: 1/0.15 = 6.67GW 600K: 1/0.8 = 1.25GW 6.67/1.25 = 5.34x less waste heat. So yea, I understand you are seeing the 5.3x number and saying it is better. I am claiming the 256x number will hurt a lot more. As long as the heat source is limitless (which is a reasonable approximation with nuclear fuel sources ingame) and radiators are non-negligible, this will be the case. Solving a little calculus will yield Tc = 0.75 * Th as the optimum for radiator area (ignoring pumping), and corrections for the rest of the effects will not drag the overall power density down much. If you don't believe me, mod the ultimate and yield strengths of tungsten and tantalum to be ~10x higher and try making reactors on ships with low exhaust temperatures. The radiators will be unreasonably large.
|
|
|
Post by shurugal on Dec 21, 2016 22:56:10 GMT
Thermocouple dT max is only 500k, but steam expansion has no such limit. If we expand the steam through a turbine first, then we convert a considerable portion of what would otherwise be waste heat to mechanical and electrical energy. Yes, a carnot engine operating between 3100 and 1600 has shit efficiency, but one operating between 3100 and, say, 600 has an efficiency of 80%. If we use a thermocouple to drop what is left from 600 to 100... you see what i'm getting at? I do. I also know it is a losing battle below about 2300K. Radiators scale by T^4. Lets say you have a system generating 1GW with 80% efficiency with waste heat at 600K versus another system generating 1GW at 15% efficiency with waste heat at 2400K. Each GW of waste heat on the 600K radiator requires 4^4 = 256x the area. But it generates less heat: 2400K: 1/0.15 = 6.67GW 600K: 1/0.8 = 1.25GW 6.67/1.25 = 5.34x less waste heat. So yea, I understand you are seeing the 5.3x number and saying it is better. I am claiming the 256x number will hurt a lot more. As long as the heat source is limitless (which is a reasonable approximation with nuclear fuel sources ingame) and radiators are non-negligible, this will be the case. Solving a little calculus will yield Tc = 0.75 * Th as the optimum for radiator area (ignoring pumping), and corrections for the rest of the effects will not drag the overall power density down much. If you don't believe me, mod the ultimate and yield strengths of tungsten and tantalum to be ~10x higher and try making reactors on ships with low exhaust temperatures. The radiators will be unreasonably large. To start, your math on waste heat is off: at 15% efficiency, we have a total power input requirement of 6.67GW, minus 1 GW of output electricity, for a waste energy of 5.67 GW at 80% efficiency, we have a total power input requirement of 1.25GW, minus 1 GW of output electricity, for a waste energy of 250 MW ingame requirements to dissipate 5.6 GW @ 2400k gives us a 3000 m 2 radiator. ingame requirements to dissipate 250 MW at 600k gives us a ~300,000 m 2 radiator. okay, bad trade. but we can improve our situation: concentrate that heat energy with a refrigeration pump. if we take our 600k final temperature and pass it through a heat pump to crank it back up to 2400k, we get a total energy factor of 1.33 (ideal pumping efficiency of 75%), for a total heat output of 333.33 MW @ 2400k, which mean we only need ~180 m 2 of radiator area. Sure, we use 83.33 MW of power to run that heat pump, but it lets us run a ridiculously tiny radiator. Now, if we want to just flat-out go nuts with it: That 600k is the temperature output of our steam turbine. We can run a 500k drop off that across a thermocouple, to 100k, giving us ~80% recovery of that waste 250 MW. now we bring our 1 GW output up to 1.2 GW, with 50MW of wasted energy. We can use that extra 200 MW to pump our outlet back up to 2500k, and have a radiator area of only <50 m 2
|
|
|
Post by newageofpower on Dec 21, 2016 23:04:20 GMT
Just took a look @ your math.
Heat pump will generate more waste heat. But yes, it checks out... With the given inputs.
Assuming 80% gross efficiency seems a little questionable. Something closer to 60% combined steam and thermocouple stage seems more reasonable.
|
|
|
Post by shurugal on Dec 21, 2016 23:05:14 GMT
Just took a look @ your math. Heat pump will generate more waste heat. But yes, it checks out... With the given inputs. Assuming 80% gross efficiency seems a little questionable. Something closer to 60% combined steam and thermocouple stage seems more reasonable. if my math is off, please let me know where i goofed it. I'm learning this stuff as i go here. and yes, 80% is the idealized efficiency. This assumes, of course, that we have frictionless bearings and that our turbine components perfectly reject the heat of the working fluid back into the fluid. Does the game calculate machine losses, or does it only work on idealized numbers at this point?
|
|