0

How does a BLDC motor performance change at high altitude where the air density is really low? For example, say I have a given BLDC motor with given torque-RPM performance curves. Now, I assume these curves have been generated based on thermal limits at STP conditions. If I want to operate this motor at 50,000 feet, how would I account for the different temperature and low density on the performance curves at these conditions?

EDIT: My (limited) understanding of how torque-RPM curves are generated comes from this datasheet: https://www.logic-control.com/datasheets/8/Catalog/Intro/011-Understanding%20Torque%20Curves.pdf

This told me that the upper bound of the continuous torque curves comes from testing to thermal limits, which led to my understanding that if the thermal environment changed, these curves themselves would change. From some of the answers, it seems like this is not the case.

With regards to the purpose of the motor, it is to drive a propeller load with a known torque and RPM demand. Thus, I was mostly curious about whether I can trust a BLDC motor's torque-RPM curve from a manufacturer, given that they probably test at STP. And if so, what corrections do I make.

dumbpropnerd
  • 53
  • 1
  • 6
  • 1
    Well, the torque-speed curves won't change, but its thermal characteristics will. Probably. This is well outside my area of expertise. – Hearth Oct 15 '18 at 18:02
  • The ability of the motor to dissipate heat will be greatly reduced at 50,000 feet due to lower air density. I guess reduced air drag will be an effect, but it seems like that would be a small effect in most cases. Will it be inside of a heated cabin? Ambient air temperature at 50,000 feet will be VERY LOW. And this will effect everything (not just your motor) if you fail to guard against it. – user57037 Oct 15 '18 at 18:11
  • What is the motor being used for? Providing that information will probably lead to more helpful answers or advice. – user57037 Oct 15 '18 at 18:12
  • @mkeith Thank you, I have edited the question above. The thermal design (e.g. heated cabin or cooling loop) is not started yet. It is the next step. Although the temperature is very low, based on previous data, motor overheating was a problem (due to no air to dissipate) rather than getting too cold – dumbpropnerd Oct 15 '18 at 19:23
  • The torque required to spin the propeller at a given RPM will be much less at 50,000 feet than at sea level. The thrust produced by the propeller at a given RPM will be much less at 50,000 feet than at sea level. My apologies if you already know this basic information. But it seems that it is worth pointing out. – user57037 Oct 15 '18 at 19:42
  • 1
    I did a quick check, and it appears that the air density at sea level is around 7x the density at 50,000 feet. So thrust will be reduced very dramatically, as will convection cooling. – user57037 Oct 15 '18 at 20:11
  • @mkeith Thanks. Yes, I am aware of the different prop requirements. The torque and thrust propeller curves I'm using are calculated for the aerodynamics at the required altitude. I'm still not satisfied regarding how to deal with the motor information though. Seems like all the answers point to "take care of it thermally" which we will do. I'm academically curious what *does* happen to the performance if nothing were done to actively take care of the motor at high altitudes. – dumbpropnerd Oct 15 '18 at 20:26
  • To a first approximation, I would guess that the cooling will be about 15 percent of what it would be at 1 atmosphere. So the max stator current needs to be de-rated by around 85%. OR you need to devise a more effective way to remove heat from the stator. Ideal thing would be if you could test in a temperature controlled vacuum chamber maintained at equivalent altitude. – user57037 Oct 15 '18 at 20:50

4 Answers4

2

I've sometimes discussed motors for space applications with high-end motor vendors. While I can't say this is conclusive, different torque curves for ground testing vs space were never brought up by the vendor.

However, one thing that did come up is lubrication. As mentioned by several comments and answers, your motors are at risk of running far hotter than on the ground due to reduced convection cooling. However, you also have a risk at the cold end. Will your motor ever be powered off for a significant period of time? (I'm guessing this is a balloon payload?) If powered off, the motor could cool until the lubricant thickens too much, causing the motor to bind when you turn it on. Your motor should list its minimum operating temperature.

Selvek
  • 1,947
  • 7
  • 12
  • Thank you. I will most definitely consider heating and cooling requirements (partly why I'm asking this question) of the motor assembly, and may end up having some active thermal design. However, I am academically curious how to correct for motor performance in the absence of thermal design, at high altitudes. I do not see much data out there. I ask because in the link I posted, they hint that torque-RPM curves of motors are generated based off of thermal limits. – dumbpropnerd Oct 15 '18 at 20:29
  • 1
    "In the absence of thermal design", your torque-RPM curves won't change even if they are thermally dependent, because you've specifically decided to ignore thermal effects :) – Selvek Oct 15 '18 at 20:53
  • Ok, in that case, how is the "continuous torque" curve determined for a given applied voltage? If you check out the link I posted, they say that they apply a load until the motor stabilizes at 85 C when it draws the current required for that torque load, and that is how they determine the torque curve. The natural extension of that (I mistakenly assumed) was that the torque curve would thus change in a different thermal env. So I guess my question is how is the torque curve determined, if not by having a thermal limit? If temp was no issue, couldn't it draw as much current reqd per load? – dumbpropnerd Oct 15 '18 at 21:20
  • I think you're quite correct that the torque curve will change in a different thermal environment. My point is, I don't believe that altitude has any significant effects that are NOT thermally related. When you said "in the absence of thermal design", I interpreted that as you'd like to know what motor performance changes there may be, specifically *not* including thermally-caused changes. – Selvek Oct 15 '18 at 22:40
  • 1
    @dumbpropnerd: If you look at their "continuous torque" curve it's got a big fat arrow from the phrase "thermally limited". I'd take that as a clue. In a permanent magnet motor (BLDC, brushless AC, or brushed), the torque is proportional to current, and the heating is almost entirely due to I^2 R effects. So if you know the thermal resistance from armature to ambient, and the ambient temperature, then you can figure out what your continuous current (and hence torque) is. – TimWescott Oct 15 '18 at 23:07
  • @TimWescott I did say in my OP that "thermally limited" was my original understanding, but there were comments/answers saying that the torque-RPM curves stay the same (excluding the comment on this thread since it was a misunderstanding). Which is why I asked how else the curves are made? I appreciate the clarification, and I think using the steps you outlined will be useful for what I'm trying to do. Thanks – dumbpropnerd Oct 16 '18 at 02:28
  • 1
    @dumbpropnerd: Torque vs. current and zero-load speed vs. voltage are characteristics of the physical construction, how the motor is wound, and how strong the magnets are. In fact, if you use mks units for torque, current and voltage, and radians/second for speed, the ideal torque/current constant is exactly the same number as the ideal voltage/speed constant -- and the dimensions work out. For a motor with rare-earth magnets, the limit on short-term torque at any given RPM is the resistance of the wire (which limits current based on the difference between back-EMF and supply voltage). – TimWescott Oct 16 '18 at 15:48
  • "For a motor with rare-earth magnets..." In case you get caught in a time vortex, or are using really cheap motors, ceramic magnets and AlNiCo magnets would get demagnetized if the motor current was too high, ruining the torque/current constant -- in other than extreme cases the no-load speed would go up, but the motor would be wimpy. – TimWescott Oct 16 '18 at 15:49
  • @TimWescott "the limit on short-term torque at any given RPM is the resistance of the wire (which limits current based on the difference between back-EMF and supply voltage" Thank you, this is a very useful clarification for me. – dumbpropnerd Oct 16 '18 at 16:51
  • @TimWescott "the ideal torque/current constant is exactly the same number as the ideal voltage/speed constant" This is exactly what I was told. Which is why I was confused when a manufacturer had different values, as I asked here: https://electronics.stackexchange.com/questions/399593/bldc-motor-constants-why-is-my-kt-and-ke-different (spec sheet posted as image somewhere in the comments). Apparently, it scaled with root(3) and had something to do with AC configurations, all of which was a bit over my head, but something I eventually want to get to the bottom of, understanding wise. – dumbpropnerd Oct 16 '18 at 16:54
  • 1
    @dumbpropnerd: Ah, yes, that. I've always been graced with data sheets that compare apples to apples. If one number is specified for RMS and the other peak, or one for a speed control's input voltage and the other for the AC current in the windings, then yes, things don't match up. If you do the basic physics, though, they *have* to match up -- just start from the definition of an ideal motor as being lossless and not storing energy, and you quickly find that shaft power out **must** equal electrical power in (or visa versa, for a generator). That leads to the one and only motor constant. – TimWescott Oct 16 '18 at 17:22
1

Altitude brings with it two characteristics

  1. reduction in ambient temperature (assuming some part is exposed).
  2. reduction in pressure.

With decreased temperature comes increase magnetic strength. Take Samarium-cobalt, there is usually a +-5% change centred around 25C (-55 -> 105), however there is usually a +-5% manufacturing tolerance on the magnet strength. At first glance this increase in magnet strength (increase in Kt) is beneficial? However, is it of practical use? Just because the ambient is at -55C or -30C does not mean the rotor is going to be at that temperature.

Case 1. unit is continually operated from takeoff until altitude. The decrease in ambient will help with the cooling of the stator (copper and iron losses) and in turn keep the rotor cooler, but it had already been operating for an extended period of time and depending on your operating duty and loading, might not be significantly cooler.

Case 2. unit was un-powered during accent and then was power at altitude, with suitable soaking time. Yes the magents would be at their coldest and your system would benefit from the higher torque per current. However... usage will still still warm the stator and rotor up until it again stabilises like Case 1. It will take some time, a time associated with the thermal mass and thermal resistance.

A product I have worked on was driving against a mechanical drive chain. This presented a higher drag at cold temperature. At -55C Qualification testing the overall performance was lower than nominal - motor-drive was performing better but the mechanics were not. At +90C ambient the overall performance was lower than nominal - motor-drive was performing worse than nominal while the mechanics were slightly better. All within spec but a noted overall characteristic.

Pressure.

Pressure brings with it two characteristics.

A reduction in pressure means a reduction in air density and thus less air to push against to maintain lift.

A turboprop has operated upto 20km ( https://en.wikipedia.org/wiki/Grob_Strato_2C ) and when you consider you are looking into characteristics up to 50,000ft (15km) and a "propeller load", this is something to consider.

At altitude the voltage breakdown characteristics also reduces. Pachen's curve. This means you could suffer from partial discharge or corona effects, if you operate at "high voltage" where high in this case is 1000V. This will drive improvements in winding insulation to ensure they do not break down. You have not stated the operating voltage thus this might not be a concern for you.

enter image description here

0

The biggest difference will be due to lower heat dissipation. As copper gets hot, its resistivity goes up, causing your motor to be less efficient.

If you know the "sea level" parameters for the motor, you could probably get close by derating the duty and/or max current with info from here: Altitude Corrections

TuxRobotics
  • 84
  • 3
  • 1
    A BLDC motor has the windings on the stator, which means that it can be profitably cooled by thermal conduction. That means that the altitude corrections will only apply to the degree of cooling happening due to air flow. – TimWescott Oct 15 '18 at 19:07
0

There are two effects that change the motor performance-

  1. Winding temperature - a low ambient will noticeably reduce the resistance of the windings, which brings the torque-speed curve up - so long as the reduction in air density doesn't reduce the cooling enough to offset that reduction in temperature.
    1. Magnet strength. Most magnets have a negative temperature coefficient on Br, so they will get stronger at lower temperatures. Some magnets, especially samarium alloys that contain some gadolinium, can have very low TCs however. Stronger magnets won't change the torque speed curve much, the Kt value (torque generated per amp) improves so the current at a particular load reduces, but the Ke value also goes up, so the applied voltage rises, so apart from the resistive loss in the windings, the power consumed changes little. Maximum speed achievable on a given supply will fall though.
Phil G
  • 5,609
  • 1
  • 10
  • 18