Prelude to Oil Cooling a hub motor

[Sancho's Horse]

“By three methods we may learn wisdom: first, by reflection, which is noblest; second, by imitation, which is easiest; and third, by experience, which is the most bitter.” Confucius

I am preparing to head down the bitter road again. So, this is my prelude to oil cooling a hub motor.

My thoughts:

My need for low-end torque in stop and go situations means there will be lots of heat being developed. The motor I have is a nine continents 3000W hub motor.

I am not an engineer, nor do I have access to tooling (other than a drill press, angle grinder, metal cutoff saw, a jig saw *not currently working, a dremel, bench vice, and files). I have been working through lots of calculations, because I am using a nuvinci N171 (developers kit), and am not at the edge of the envelope, but close enough it is worth keeping tabs.

Torque, Power, and Speed Limits* of nuvinci N171

Maximum Sustained (Pulse-free) Input Torque 65 Nm (48 lb-ft)
Maximum Instantaneous (Spike) Input Torque 130 Nm (96 lb-ft)
Maximum Sustained Input Power 5 kW (7 hp)**
Maximum Recommended CVP Speed Input: 1000 RPM / Output: 1000 RPM

The motor is set-up for 700RPM no load. It has 3 speed (80%, 100%, 120%) switching. It also has ebrake.

Now....Here are my thoughts on cooling.

I am a big fan of drilled, centrifugal, internally finned air cooling methods. They are relatively easy, and they work. And...I think someone else has pointed out, no matter what cooling method you use (air, water, oil) you are ultimately using air, because in all three the final energy sink is the atmosphere. Other oil cooling attempts have met with various levels of success, ranging from failure to pretty good.

I have reported elsewhere, self-fan cooling vs. liquid cooling showed liquid cooling arrives at 2-3 times the level of Amps per mm^2. This requires a pump and radiator to get anywhere near these levels of difference. Most approaches to oil cooling currently modded use the oil to move the heat from internals (windings, etc.) to the cover more efficiently than with air, due to the huge difference in thermal conductivity between air (0.0243 W/mK @ 25C...0.0299 @80C) and oil (>0.138 @25C). Water has better thermal conductivity (0.06W/mK @25C) but will still have an air gap from the electronics, decreasing its effectiveness, and will require a heavier apparatus/tubing/channelling to make exchange. Oil can go right on the electronics (with some modifications).

Now, I brought up, and I am sure others have as well, the covers just do not provide the level of surface area required to really achieve the levels of cooling. I saw b0b did a pretty cool mod using aluminum radiator fin on crossbreak's Converting a hubmotor to middrive thread: http://endless-sphere.com/forums/viewtopic.php?f=28&t=45245&start=135

I like it, but I think there is far greater possibility for passive oil cooling hubmotor design.
So...without further word interruptions, here is how I intend to oil cool my motor.

First, I am going to do testing to establish unaltered motor temps.

Next, I am going to use common, readily available pipe elbows with flare nuts, and clear tubing to establish oil movements out of and back into the hub motor during spin up/ebraking using a mounting device I have created and a video camera I call the Nam cam.


Finally, I will put a series of elbows and flare nuts at intervals around the hub, with aluminum tubing traversing the space in between spoke flanges, with a prepared patterned laminate of aluminum flashing housing and intimately in contact with the tubing.


My feeling is that centrifugal cooling goes both ways. I hope to show that using it with oil cooling and a radiator of much greater surface areas housed in the space between the spokes is a big win for cooling, and a proposal which will give oil cooling its due.

Feel free to chime in.

 

[Sancho's Horse]

I saw veloman's thread Cooling a Mac hub motor thread:http://endless-sphere.com/forums/viewtopic.php?f=2&t=40243 and he uses precisely the real estate where I am seeing possibilities.

He uses a wrap of c channel aluminum without any oil cooling. He reported it as relatively ineffective.

I think this is some really good thinking on using this space. It definitely increases surface area.

He unfortunately can't do air cooling easily because of the motor type. However, to even approach the level of air exchange being found with centrifugal air cooling methods, it needs way more surface area. By my thinking, that old description of the lungs containing a tennis court of surface area is more along the lines of what is needed.

The thermal conductivity of aluminum @ 25 c is 205 W/mK so this is clearly not the problem. It is between the aluminum and the air, and the oil and the aluminum that the real limiting factors are found. Now the thermal conductivity of oil @25 c vs. air means oil is ~5.7 times as conductive. However, both will need some serious attention.

I am starting with the air/aluminum interface. I have only recently started down the path to creating calculations giving estimations of how much energy can be dissipated, and these still involve some serious fudge factor intuition if they are going to be remotely useful. So, in order to sharpen my fudge factor, I plan on making the whole apparatus modular, and removable. If I build fins of various dimensions, I will get a better idea of what is going on with very little added expense.

Now...I have seen some of the venting mods and the unique approaches. For those of you who have seen some "things" on ES, I looked for the ballshaft venting pics, but no. But still...One thing I like about this is that there is the real possibility of making something which fits your personality and "improves" the look of your rig. Now, at 700 rpm's getting something to look the way I want at those speeds is a pretty tall order, but I definitely think getting an effect is doable. And at slower speeds or stopped, you can definitely make something which exactly fits the look you want for your bike.

Cutting flashing is super easy. You can easily use scissors. Or...use a blade to score a pattern, cut a small start and rip. It rolls out flat with a rolling pin.

And torn to reveal the pattern

You can get pretty intricate, but whatever you create, remember: a cromotor may shred the snowflakes.

 

[Sancho's Horse]

Ok, been sorting through lots of information. I am still just as hopeful as earlier, but a bit better informed. Here is a pretty simple illustration from a review of a 2010-2013 Electric Motor Thermal Management program by the Dept. of Energy(http://www1.eere.energy.gov/vehicle...v_power_electronics/ape030_bennion_2011_o.pdf which illustrates why we work so much on thermal management:

Basic stuff. Here is another figure from that review comparing a water jacket case cooling versus oil cooling at the windings. Apparently, their winding cooling did not involve added mechanical losses like the method I am considering.

Somewhat interesting.

My hope is that I will be able to use a larger volume of oil in my approach,
that the oil will reliably move into the tubing without problem,
the cooling at the perimeter with increased surface area will be better optimized for shedding heat faster than through the case,
that a small amount of convective heat flow through the tubing will continue after spinup,
and that during stopping the cooled oil will return to the case housing to provide cooled oil during the high heat high torque generation of motor start-up.

The single biggest area for concern is the oil going into the tubing reliably. I am comfortable that I can optimize the opening placement and geometry to get oil into the tubing, but doing this without creating cavitation forces is my chief concern. Cavitation is the formation and bursting of bubbles when a fluid goes from an area of high pressure to an area of low pressure, or more technically, where the static pressure drops below the vapor pressure. Dexron III has a very low vapor pressure...<0.01 mm Hg or <1.33 Pa, and it has some anti-foaming agents, so that is very helpful. I have read some papers on cavitation, and possibly have some others on the way.
View attachment cavitation oil finland.pdf
A couple of things I am going to need to change: I need to replace the elbows with straight flare nut couplers, and change the angle of coupler insertion to achieve a better orifice geometry.

I am also likely going to spin an oil filled drum with coupled tubing on my drill press to test the oil flow at a speed similar to my motors rpm.

Oh, the joy of technical papers! May the technical papers be unto you! And happy holidays!

 

[Sancho's Horse]

The effect I am hoping to make use of is called the magnus effecthttp://en.wikipedia.org/wiki/Magnus_effect

The hope is that the oil will spin to the outside of the hub, and the differences in forces due to the Magnus effect between one opening and the other will be enough force to push oil into the tubing, and push large less energetically problematic bubbles out the other side.

Bursting bubbles generates a large amount of force, but the larger the bubble...the less force is generated.

This is all beyond my pay rate...so, the only real way for me to know if this is something which will work is to find the right person with appropriate specific knowledge, or test it.

I vote test it.

I have used my pie plates to create a mock-up, and found my drill press can be adjusted to run at approximately the rpms of my motor (660 for the drill, 700 for the motor). I will later add a lexan window to the side of one of the pie plates, and I have painted the interior white to aid in resolution.

If my ideas are bunk, the lexan mock-up will definitely still help visually for understanding some of the forces involved. I cannot think of a way to mock the counterveiling forces created by the rotor, without using an actual motor. So, it will be a very imperfect representation, unless I can think of a way of mocking the rotor. Let me think about it some more.

 

[ian.mich]

drill a needle hole, seal her with gasket 'cone, fill with dex iii

 

[Sancho's Horse]

OK, I have done a lot of work on this the past several few months. My initial setup was simply some reinforced pie plates, with various tubing sizes and configurations, and a plexiglass side cover. Some of them I may post in the future. The primary things I looked at were hole/tubing sizes, hole geometries, hole placement configurations, and overall movement of the fluids at various rpms. I also looked at the levels and types of bubble formation. There were a fair number of tiny bubbles, but my feeling is that there are likely far more bubbles formed through the agitation of the stator as it moves than by moving the synthetic ATF out through the tubing. I think a bit of epoxy coating for the windings would not hurt to prevent mechanical damage from the fluids. That being said...here is a pic of an early setup...nothing fancy.

Simple Newtonian physics says the oil will fling out. This was observed even at relatively low (not quantified) rpms, and with all hole configurations. However, the configuration which I have moved forward with involves only 4 holes, and seemed to enjoy the best movement/least complexity of motor modification approaches. I am also still considering a spoke conformation due to the ease of establishing geometric symmetries which is critical to maintaining good bearing life.

I have experience with casting, so I am currently finishing my Master Mold. For the initial prototype I am using aluminum cast a la plaster. The most critical feature I am considering now is overall balancing of forces sufficient for high rpm service (hub motor high...not rc). I have two inlets and two outlets in my initial design nearly diametrically opposed. They do have differing geometries with the intake more recessed in the sideplate, and the exhaust protruding slightly into the motor housing. The exhaust will also have a pressure release mechanism designed into the casting, which I keep changing my mind on... design wise...as I read more. The benefit I see from the two inlets and outlets is relative symmetry. However, beyond that it allows good movement of ATF during spin-up and as the motor slows the oil dumps back into the motor housing nicely. I have attachment points for my radiator fin design. Another benefit I am seeing from the fin design is it is going to allow me to do some very effective balancing to achieve a good, balanced motor spin. This pic is prior to a final model which... eventually...and still...is changing. Right now, my primary focus is on tweaking outlet geometry and the pressure release mechanism, and improving my modelling of forces so that I can improve fin configurations to make a perfectly balanced motor.

 

[m52 power!]

So what ever happened to this?!