 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Off topic archive. : How much saturation is good for you.
| Author | Message | ||||
oztules Guru  Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Hmmm Well Dinges, I decided to do some homework and found the article on fl between fin and flux Oztules dies here Not much more to say other than I think I have been sunk 
.... oztules wanders off n a daze, and mumbles something about dutchmen and leaky dikes.....  Village idiot...or... just another hack out of his depth |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
On tracking down more stuff, I may be alive still. It appears that the iron drag will not increase much more after saturation, so it may be that I didn't extrapolate Flux's statements enough. More flux= less copper loss. If after a decent saturation, the iron losses don't go up further, we can decrease cogging, and increase power density for almost free after that point Dinges... I may live again! I'm thinking Flux was dealing with what was physically possible/feasable rather than what was going to happen after unreal T values. Upon thinking, if industrial gennies are running in full saturation.... 1.8t it may be the cost of going into further saturation that stops them and they then elect to go copper loss instead. I can feel the paddles being removed from my chest, and the voltage being turned off. ......oz (still breathing)tules Village idiot...or... just another hack out of his depth |
||||
| Dinges Senior Member  Joined: 04/01/2008 Location: AlbaniaPosts: 510 |
Depends on whether you're talking about alcohol saturation in your blood or magnetic saturation in your stator, I suppose... Interesting link. Not the one I have in mind (can't find it again, I'm 100% sure I even saved it about a month ago for offline reading). Somehow I managed to entirely miss that discussion. I think I know why I missed it too. Very interesting observation, completely new to me. He's basically saying that even with saturated stator teeth the iron losses would be small. If that's the case it may well be good to have a stator that is so far saturated till iron losses equal copper losses (for a particular point on the generator curve; as RPM and power varies this optimum may actually shift ?). Flux's remark indicated that we practically are still below the optimum flux density where this equilibrium exists, so you may have been very well correct (practically speaking) with your suggestion to strive for as high flux density as practically possible, even if it means saturated teeth. Next interesting thought that flares up in this brain... Would it warrant the expenses of going from N40 to N50 ?... (or would it be cheaper to simply use a slightly larger conversion using more N40 magnets,  )
This makes sense, the question is of course, 'how much is too far?'. Is anything below saturated stator teeth too little flux ? With my limited knowledge I think any flux where copper losses > iron losses is 'too low'. And any flux where iron losses > copper losses is 'too high'. Where this point lies exactly for a particular genny ? Impossible to say without a decent mathematical model. The way I understand saturation is that it's impossible to increase flux density in the stator iron beyond saturation, as the stator simply can't hold anymore flux lines. The air (mostly filled with copper) in the stator slots could take up some more flux lines but, in the end, this flux has to be guided to the next stator slot via the stator yoke (backing iron, behind the teeth). When that is saturated it can't handle this extra flux anymore. It's impossible to have flux densities above saturation level. (Edit: on 2nd thought it may actually be possible; not sure anymore. After saturation the stator iron starts behaving as vacuum, with magnetic permeability mu-zero (greek 'mu'), essentially behaving as an aircore generator. I.e. the stator iron magically 'disappears'? As I said, not sure anymore. I used to think saturation proved to be a hard upper limit on flux densities) TANSTAAFL... (Edit: or maybe there is... ?) BTW, I'm still curious as to your reasons for saying that the only relevant efficiency was copper efficiency (resistive losses, copper losses). You must have had reasons for claiming this. |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Yes, after further reading, I think I might be gaining the upper hand . If we cram in more magnets of higher flux .. say 100t, our iron loss only counts for the first 2t at best, after that,the flux doesn't see the steel anymore and treats it as an air core. This means that at low power (say less than 50w)for this alternator, iron loss>copper loss, but as the power increases, the iron loss remains almost constant (but for frequency shift), but the copper loss can be 1/100th of < 1T operation. This means a clear win for lots of flux, as it turns out I need only compensate for the initial iron drag of say 2t, and the rest is win win win in a huge way. I live again.
What this means in practical terms is if you want light wind operation for <100w, then don't make a iron cored machine unless you use ferrites. Then your losses will all occur after cut in (made easier by almost no drag). After cut in your copper losses will mount up as a square, but input power is a cube, so you will still have increasing output, but stator heat becomes an issue.... Unless you design for high cut-in rpm... then you can have the best of both worlds, (and auto mppt because of the curve matching) ...... Strange,.... this is how Flux built his generators for the first 30 years of his career from what I gather. They never had the stator heat problems we experience now days... because they had high cut in, and high cut in rpm and gearing was the way to decrease copper losses and keep them acceptable. So... (circling for the kill) it seems if we can live with the initial losses from saturation, which we appear to be almost doing now with 1.6 and 1.8 being already present in current designs, any increase in flux we can manage to drive it further into saturation is all benefit.... including auto decogging if we can get the flux in the air gap and inter tooth gap up over 2t.... see it just gets better (oztules feeling pretty good again) The only reason commercial designs don't go much over 1.8t, seems to be the cost of getting more flux, so they settle for copper loss at this point.. it's cheaper and the heat is manageable. ie if you use neo's it gets prohibitive from a cost point, and if you do it via electromagnets, the energy loss will overcome the gains...why.. because the flux coils will be suffering from excessive copper loss as well. I think I can safely rest my case now... OK... "BTW, I'm still curious as to your reasons for saying that the only relevant efficiency was copper efficiency (resistive losses, copper losses). You must have had reasons for claiming this. " The reasons are simple. The only thing that makes a big generator physically big for large output is wire size. Nothing else. If we had superconductors, we could use little stepper motors with 1" shafts, to develop 10kw. We could use hair thin wire in tiny coils of hundreds of turns to get any turn/volt combination at any power level in a matchbox sized unit... suddenly alternator design is a doddle. You could use a weak magnetic field for the stator, and spin thousands and thousands of turns of 1 micron wire to generate whatever you wanted in a few cubic inches... shaft size would be the limiting factor for size. The flux strength would not matter, just add more turns of thinner wire to achieve the result. It all comes down to resistance. The second we get some, the above scenario collapses in a pile of dung. And once again, more power means bigger wire, and so bigger devices to carry it. What can we do to mitigate this some.... only increase the flux any way we can to allow for less turns. Nothing else matters. Every other problem that crops up is immaterial compared to resistance. We can machine for higher tolerence and less gaps, wind tighter for same turns but thicker wire, make the coils go inside the magnet diameter in air gap machines to fit in a few more grams of copper for the same turns, mount dual rotors with magnets on each, use focusing devices (iron cored stators) and any manner of gimmicks to achieve less turns so we can use larger wire and so get the resistance down. It has massive implications to the dynamics of the machine. If we want power from the wind at sensible wind speeds, and we want direct drive, this is what it all comes down to. 1. more power means bigger blades 2. bigger blades means less rpm for bigger power 3. the alternator design increases in size at a squared rate because of the resistance, but is multiplied by the problems of lower speed to work with. So because of 1 and 2, 3 goes horribly pear shaped very quickly. So the only thing you need to keep your eyes on is resistance and how to minimise it... every thing else will fall into place when you get this done. It is at the base of all the calculations for an alternator. How big it needs to be physically is controlled by resistance and nothing else. So the efficiency in an air core is all about resistance of the copper. (I suppose you could include bearing loss too and eddy current in the copper... but I have found it too small to consider and I use 1.8mm wire), but thats it... there are just no more efficiencies to consider. It is the damn RPM requirements that cause the problems. It it ran at 10000 rpm, 1 turn of huge litz would be enough. In an iron core, the same problem, but with some iron loss as well, but it is linear in nature, and so does not figure as a problem. It is the squared losses the cause the concern..... and they are resistance. resistance resistance. Get them under some sort of control, and you win. When you wind the stator on your beast... how will you determine the winding. It will be rpm (governed by power input requirements... not by wish). Once that is established, it is all geared towards getting as fat a wire into the stator as possible... nothing else matters after the rpm is established, and whatever flux you can afford is installed. .........oz(back from the dead)tules Village idiot...or... just another hack out of his depth |
||||
GWatPE Senior Member Joined: 01/09/2006 Location: AustraliaPosts: 2127 |
Hi oztules, maybe we should experiment with super-conductors instead of windmills. My axial flux testing pointed to low turns, high rpm and thick wire with the most powerful magnets and no iron. Gearing is not as much a problem with no cogging. The gains from higher rpm with lower resistance outweigh the gearing losses. This allows as large a blade as needed to capture the required energy and gearing. The large commercial windmills with induction drives would not work without gearing either. Gordon. become more energy aware |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Yes , the ones over here have large gearboxes as well. Im sure I read that there are some late models with pmg arrangements but radial I think, and inverters to match. sigh...250000 watt inverters must be something else to see. Village idiot...or... just another hack out of his depth |
||||
| Dinges Senior Member Joined: 04/01/2008 Location: AlbaniaPosts: 510 |
Are you sure of this ? I'd expect the iron losses to go up (at a fixed maximum flux density of, say, 2T) by frequency; twice the RPM, twice the power losses. I think Flux also hinted at this in the link you gave. Obviously. An air-cored generator (e.g. the axial flux) doesn't have iron so doesn't have iron losses. The downside is it needs to be much larger and needs more/stronger magnets to achieve the same output power as a motorconversion. I wouldn't call it auto DE-cogging; there is still a lot of cog, just that the amount has a fixed upper limit. As you increase flux density the genny essentially behaves more and more like an air-core generator. The absolute amount of cogging remains the same, but the relative part of the cogging power in the total power becomes smaller, so the effect of cogging gets 'diluted'. Err...For transformers, it's not just the amount of copper wire (kg) that determines power handling ability, but also the amount of iron. Even using super conductive wires you couldn't push 100 kW through a tiny wall-wart transformer. (unless you were still talking of using insanely high flux densities, in which case it would essentially become an air-cored transformer) Your stator would still need to be able to guide all that flux so would have to be of a decent size too, I'd expect. It's not just the copper that determines the physical size of a transformer/motor ? (even if you used super-conducting wires). Unless you're still talking of the situation where we use insanely large flux densities... but obviously you aren't, judging by your next remark: In an inductor or transformer, one can lower flux densities to below saturation by using more turns of wire; this is what you were basically stating before. And as the superconducting wire is lossless you could do it impunibly w.r.t. copper losses. But I expect the stator would still have to be sized to handly the actual kW you're pushing through it ? I suppose now everything hinges on this question: how large does the transformer iron need to be to handle 100 kW of power using superconducting wires. Could the core be something you held in your hand, or would it need a kitchen-sized room to house it. I don't have the knowledge to be able to judge this other than believing you on your 'scout's honour'. (as you can see my limits are beginning to show, not being an electrical engineer). On another topic: the high pole count that you mentioned Zubbly advising earlier in this thread makes sense; by using a large number of poles you can build a low resistance generator, as you need relatively few turns per coil to end up at your desired voltage. Few turns/coil means space for thicker wire, i.e. low resistance. This would also explain why for low RPM gennies you want a short, large diameter stator and for fast generators/motors they use small diameter/long stators. Peter (slightly confused but not defeated...yet) |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Dinges, I will have to give up. I can't sustain an obviously flawed argument. My original statement cannot be supported beyond about 2T and it sure can't stand up if you start to bring real physics into it. (Thats patently unfair.) The inductive reaction, reactance limiting and core area just ruin a perfectly good flat earth argument. Beyond about 2t, It doesn't appear that I can find any real advantage of more saturation. (and I've explored all possibilities... only to get shot down). Core area and permeability are the problem with the superconductor idea. With silly T, it is ok, as the core becomes irrelevant fairly quickly, and receeds into the background. we are then aircore. so in order point 1 yes you are right losses go up linear. point 2 just padding point 3 I tend to agree with you, but I'm sure flux has mentioned otherwise, perhaps he was being cavilier as well. points 4,5 and 6 are the same point, the core is critical until saturation, and then air core only rules apply so the little 10kw stepper would be fine.(particularly if you throw out the iron core all together...In the superconductor world, we don't need the core. You can use infinite turns to achieve the inductance you need in air core, and ignore the iron, so space is not an issue...so I could wiggle a bit here, but in practical terms... you win these ones too. Just shows how hard it is to argue an extremist point of view and win. Still within the limits of practicality, the original statement seems reasonably sound. I don't know if many will get past 2t. There is emf reactance as well and this muddies it around saturation, and possibly extends the unsaturated zone a little. ULR, Flux and ghurd had a discussion about this in 2006, can't find it, but recall it is not as bad as it would seem, and the tooth jumping of the flux, changes as we went into saturation, and effectively changed the footprint of the magnet, but I can't recall what the outcome was. From memory, you were part of that discussion (that wasn't why it was forgetful though). The design criteria for the genny remains the same get the resistance down... however that is done. By doing it, the other parameters are taken care of in order to get the resistance down anyway. ............ oztules
note to self... must not write stuff at 4 in the morning. Village idiot...or... just another hack out of his depth |
||||
| Dinges Senior Member Joined: 04/01/2008 Location: AlbaniaPosts: 510 |
Oztules, I was preparing to accept defeat till I read your post and was pleasantly surprized. Victory at last! But I think you are right in your superconductor example too. (What ?! You claimed to be wrong and now I claim you were right ? Will this discussion never end then ?!) I'll be the first to admit that it was a highly abstract/non-realistic situation, nevertheless was an interesting thought experiment. Had to work quite hard to be able to keep up with you, studying books and websites. This is the part that I'll be remembering best of this discussion. The insane flux densities and superconducting wires extremes and the conclusions that follow from it, even for the less-extreme cases such as the gennies we build normally. I maintain my original point: in theory there could be an 'optimal' flux but I'm by now convinced that in practice this flux is much higher than we are currently attaining (see e.g. Flux's remark), so in practice you were right: get as much flux density in there as you possibly can, even if it saturates the teeth. And get the resistance down as much as possible. I'll go even further and say that, in practice, to compensate for losses, I'd do as you do: just make the blades a tad larger. Still important to strive for the best efficiency, reasonably achievable, but not at the price of everything. There comes a point where it's simply smarter and more economical to accept some losses and compensate by going for a larger genny. Anyway, to really be able to have meaningful discussion on this topic you need a good mathematical model of the genny and blades, explicit assumptions and a solid background in electro-magnetics and aerodynamics. Neither of which I have. A good textbook on motordesign would be nice to have too. Wouldn't mind spending a few hours plowing through this matter. Rosenberg just doesn't cut it in that respect. Thanks for your well-presented reasoning. I hope we didn't scare the other board users off too much though.  I think the real outcome of this discussion is a draw. And I don't know about you but I ended up a little wiser. (do I hear Ron's sarcastic laughter in the background ? Nah, must be my imagination)
Dinges says in a Goldfinger voice: 'But we shall meet again, mr. Bond'. |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
It was an interesting discussion, and it has crystalised some aspects in my head... I hope others may take something useful away too. Where and when do you hear Rons laughter... I haven't tracked him down for a while.... Draws self up to full height and with best Arnie expression says...: I'll be back... 
Tested out the chainsaw blades today.... I'm impressed. very steady 500-800 watts, with small gusts pushing it up to 1300 watts Weather bureau feels we had 29.6 kph gusts for the day. I felt is was pretty mild day, so even down here on the flatter ground, with a 4M prop, there is some power to be had. The prop is on the test pole still, so I can play with the furling etc, allowed me to feel the stator after a sustained (5min)1kw.... was almost stone cold. Will do a story perhaps a bit later ........oztules Village idiot...or... just another hack out of his depth |
||||
| Dinges Senior Member Joined: 04/01/2008 Location: AlbaniaPosts: 510 |
Mostly he lurks in the background, waiting to strike when you least expect it. Just because you don't see him doesn't mean he doesn't see you... Definitely looking forward to the story. Was going to tell you to make sure it furls on time (then again, 'don't teach your grandmother to suck eggs'... ). 500-800W average without the stator even feeling lukewarm is definitely impressive. You could then safely rate it at, say, 1kW at least ? Probably a bit more.
All boils down to the copper losses, in this case. |
||||
 Print this page |
