Can Permanent Magnets Efficiently Heat Water via Induction?

[Fish4Fun]

I am idly curious about induction heating; specifically for heating 80L-100L of water from 20C to 100C as efficiently as possible with an emphasis on as little waste heat going to the surrounding environment as possible. I have done some precursory reading ( http://en.wikipedia.org/wiki/Induction_heating among other related searches), but have more questions than answers, lol.

Assuming a thin ferrous skin mechanically attached to a multi-layered aluminum/stainless pot designed specifically for induction or resistive heating, would it be practical to use a rotating disk with permanent magnets embedded in it to heat the water inside the pot? Obviously the amount of heat in watts delivered to the water will be less than the amount of energy used to rotate the disk, but this does not preclude the efficiency of heating the water via induction being higher than using mechanical energy to create electricity that is in turn used to drive conventional resistive heating elements. That is to say, if the available energy source is an IC, wind, hydro or other mechanical energy source, can it be more efficient to use this energy to drive a rotating disk populated with permanent magnets than an alternator/generator than a resistive heating element?

Given a roughly 18 inch diameter pot and a similarly sized rotating disk and a practical rotational speed of 2000 to 4000 RPM, is it generally better to have more magnetic pole pairs, or stronger magnetic flux densities? I assume, like everything in engineering, there is a min/max relationship where increasing the magnet pairs (increasing the "frequency") is overshadowed by the physically related decrease in magnetic field density per pole.

I am not asking for anyone to work out the maths for me, but rather a bit of guidance on the practicality of an engineering solution and some tips on a good starting point based on either engineering knowledge, or real world experience. If the idea of using permanent magnets to heat water in a pot is simply impractical for any reason I would just as soon abandon the idea to folly w/o building any prototypes or agonizing over the math; on the other hand, if the idea has some merit, a bit of guidance on a starting point would be much appreciated. For instance, if 1/4" Neodymium disk magnets were placed in concentric circles of 184 magnets, 178 magnets, 170 magnets etc, etc would this be more or less effective than say concentric circles of 1" disk magnets beginning with 50 magnets, then 46 magnets, 42 magnets etc, etc. For the outer circle, at 3000 RPM, we would have an effective "frequency" of 270kHz (1/4" magnets) vs 75kHz (1" magnets). Obviously as we move inward on the disk the effective frequency will decrease. Might it be more efficient to use 1/4" x 1/4" x 1" rectangular magnets than circular magnets?

Again, I am not looking for someone to work out the actual math, I would just like some general "off-the cuff" thoughts on where to begin if the idea is viable enough to move to prototyping. I would begin prototyping by starting with a much smaller model and a much larger rotational speed range (~4" disk, 1000 -20,000 rpm) perhaps three disks: one with 1/8" diameter disk magnets, one with 1/2" diameter magnets and the final one with 1/8" x 1/8" x 1/2" rectangular magnets.

Things I am assuming that may or may not be true:
1) The heat gain to the magnets/disk will be minimal and meliorated by air cooling.
2) It is practical and safe to construct an 18" disk capable of maintaining integrity @ 2k-4k RPM ( @ 4k rpm, the outer edge of an 18" disk would be rotating @ ~314 sfps ==> 1.5 * pi * 4000 / 60 )
3) The heat delivered to the pot will be = to the mechanical energy input to rotate the disk less frictional losses.
4) Mechanical input to the disk of ~10hp = ~7.5kW ==> Actual heat gain of the pot should be 6.7kW (90% efficiency) to 3.7kW (50% efficiency) for pot temperatures close to ambient.

Again, thanks for any insights or thoughts, not asking anyone to solve the math, just guide me to determining if it is worth the time and money to investigate. It would take me weeks to muddle through the math, days to build the prototypes, I just don't want to start down either path if the general idea is flawed or impractical. Obviously I have a specific purpose in mind, I am not limited to inductive heating with permanent magnets, there are other, simpler ways to heat water to 100C, but if the idea is practical, offers efficiency gains and/or is more cost effective than resistive heating elements then I would be interested in pursuing at least a prototype.

Thanks in advance!

Fish

 

[Windadct]

Due to the circuitry & aux systems involved - I can not see an induction heating solution ever being better than good ol' resistive heaters - this best resolves " specifically for heating 80L-100L of water from 20C to 100C as efficiently as possible with an emphasis on as little waste heat going to the surrounding environment as possible". 2c

 

[mfb]

Due to the circuitry & aux systems involved

If the energy source is a rotating shaft with the required angular velocity?
Looks like a possible design.

The total power output could be a bit low, however. To heat 80l by 80°C, you need ~26MJ. If you can dump ~7kW into this device (which is completely unclear), you can heat the water in about one hour. Power loss to the environment can be significant.

 

[pumila]

Conversion of mechanical energy to electrical energy is usually pretty efficient, resistive heating driven from that is 100% efficient, and it is relatively easy to use building insulating foam to reduce thermal losses over the heating cycle.

However induction heating using mechanical drive has the mechanics of coupling the power source to the rotating induction magnets to consider, and they have to be coupled through the insulation layer. External parts if the device must not be ferrous in order to prevent the power heating the outside of the assembly too.

Go for the generator with resistive heating.

 

[Mike_In_Plano]

Just as an aside, I once had a centrifuge availbale that had a flat, circular, aluminum plate that held capillary tubes in place. Also having some Neodynium magnets in place, I became curious as to whether this 10KRPM centrifuge could levitate the magnets as they sat over the speeding aluminum plate.
With the magnets fastened to a steel bar, they did indeed float about 1/8 of an inch over the aluminum, but the also had a considerable dragging force, and I could hear the load on the centrifuge's motor.
I pressed them closer to the plate with the end result that the lifting force was considerable, as was the drag and the loading of the motor.
After finishing the experiment, I wound down the machine and found the plate was quite warm, which I attribute to the induced currents.
So, yes, you can make a kind of magnetic clutch that transforms mechanical energy into heat using magets and aluminum.
However, I would have been hesitant to attempt this with Iron as the lift due to the eddy currents would have been competing with the attraction of the iron and I would possibly have wrecked my machine.

 

[Fish4Fun]

For the outer circle, at 3000 RPM, we would have an effective "frequency" of 270kHz (1/4" magnets) vs 75kHz (1" magnets).

Ooops, I blew this calculation...

184 Magnets/2 = 92 Magnet Pairs  ==> 92 * 3000rpm = 276,000 Flux Changes /  Min

(not Hz)

276,000 / 60 = 4600 Flux Changes / Second = 4.6kHz Equivalent

For the second case, 75,000/ 60 = 1.2kHz

As the literature I have read suggests induction heating frequencies in the 20kHz to 200kHz range, it would appear that an electronic approach would be better suited. To get 20kHz directly from 1/4" diameter magnet pairs with 184 magnets (92 magnet pairs) ==> (20,000/92) * 60 = 13,043 RPM. An 18" diameter disk spinning @ 13,000 RPM would have a rim speed of 735,132 inches per minute, (1021 fps, 696 mph). A 3000rpm motor would need to be geared up 4.33:1.

I consider an 18in disk rotating @ 2000-3000rpm already on the marginal side of safe; @ 13,000rpm the same disk would be just plain dangerous without turbine-like precision of all of the mechanical components. (A Car with an 18in diameter tire is turning < 100 rpm @ 60mph).

The best solution to inductive heating (as stated both above, and implied in the literature about induction heating) would be to use coils switched @ rectified 240Vac by either IGBTs or Mosfets. While I may at some point in the future play with a small prototype circuit out of intellectual curiosity, I cannot envision a full-scale 10kW version being practical from a budget/time perspective. Resistive heating is awfully cheap and easy, lol.

Fish