Power factor importance

[Joe90]

I am looking to design hub and I was wondering How important the power factor may be.

First motor

Power factor 0.60
Efficiency: 0.80

Second motor

Power factor 0.80
Efficiency : .75

Say with a 10kw controller (current limited on phases like Kelly)
Is the second one will transmit 6000W to the ground before hitting the controller limit
and the first one only 4800W to the ground before hitting the controller limit.

If I understand correctly, the second one will feel and put out more power even with less efficiency !

Thanks

 

[incememed]

Very interesting! It would be helpful if you elaborate some on your relation between power factor (PF), the actual motor design and the way you intend to control the motor.

Talking about hub motors, and BLDC motors in general, I'm not so sure they can be readily subjected to PF analysis (or if such analysis would be of any interest). Especially not considering they are normally controlled with block commutatation (rotor position updates with 60 degree resolution), producing square wave voltages and currents (ideally). Such waveforms have high harmonic content, and would require some kind of fourier analysis to even be available for a PF calculation.

I'm on loose grounds, wiser people will chip in if necessary. I have not found any model explicitly for the BLDC motor. The permanent magnet synchronous machine comes close in physical ressemblance, but the model always assumes sinusoidal symmetrical three phase voltages/currents. The DC-motor model seems to be the best fit.

I'm not very familiar with Kelly products, but from what has been related on this forum, such controllers normally use battery current sensors only. Hence, controller would look at some kind of DC-valued U_batt * I_batt calculation to know its max power limit.

 

[liveforphysics]

A Kelly drives it with a trap wave... a worse power factor motor with a square BEMF shape would he the optimal choice for it.

This is DC getting chopped into square blocks. PF does not apply.

 

[incememed]

This is DC getting chopped into square blocks. PF does not apply.

Yes, I figured as much. (Still a bit curious though in what respect it is mentioned in your first paragrapgh, and by Joe90, if not in the sense of PF).

Actually, what I'm fishing for is a mathematical description/model/reasoning concerning advanced timing in a regular 60 degree block commutation scheme. (I admit the connection to the subject of PF is far fetched, but at least both topics share a current lagging an applied voltage). I'm aware the topic of advanced timing has been widely discussed, and I was hoping someone would know of any analytical source of information dealing directly with hard degree, inductance and current values, perhaps with some illustrative equations and diagrams.

I see two different takes on the subject. 1. Advance timing as in field weakening to enable higher top speeds. 2. Advance timing as in making sure the area formed under a current vs. time curve during one commutation period has its total area divided equally around the instance in time when the permanent magnet rotor vector lags the stator magnetic field by exactly 90 degrees, resulting in maximum torque per current.

I'm interested in #2. Block commutation as we know it is set up to commutate 30 degrees before the two magnetic vectors form the sought after 90 degree relationship. It will again commutate 30 degrees past that point. In an ideal situation, the product of current and time from each of the 30 degree periods would be exactly equal. This would be so because the torque contribution of a current will be weighted by how far away it is in time from the 90 degree instance. Moving away from the 90 degree instance in any direction yields diminishing returns.

As has been explained many times by this fantastic forum, for reasons of inductance, duty cycle and BEMF, the current will not rise instantly when applied in the beginning of the 60 degree commutation period. Instead it will "slowly" rise and reach its peak perhaps by the end of the 60 degree period, when time has come to commutate and begin all over again. So, before the 90 degree instance there is very little current*time area, and after the 90 degree instance there is a whole lot more, with diminishing contribution to torque as the motor rotates away from the 90 degree instance. The center of the current*time area is very far away from the 90 degree instance! For large currents, compensating for this by advancing the timing might be of good use.

Still on loose grounds here, but if the above reasoning is correct and relevant, perhaps it has been analysed more in detail somewhere?

 

[rhitee05]

I have to disagree with LFP here, I think PF does apply. It would not be easy to calculate in the traditional way given non-sinusoidal waveforms and the like. I'm not enough of a power expert to know exactly what to do there. However, you can take a step back and look at PF in the big-picture sense. All PF does is describe how much useful power is generated and how much power is wasted. We can do that and we know that it depends on timing. Optimal timing should yield the best power factor; advanced/retarded timing will reduce it.

Based on that, I definitely think it's inaccurate to give a single PF that describes a motor. Unless the controller compensates, timing will vary with speed and thus PF will vary. I don't know exactly how to calculate it, but I think the controller will have some impact too based on the waveforms generated. There definitely should be a difference between block and sinusoidal waveforms, for example.

PF doesn't apply to the DC bus, though, it's a strictly AC concept.

 

[liveforphysics]

Its not the ratio of used to wasted Power at all.

Its the ratio of real to appearent power. AKA, it only applies when you were measuring power atanst a standard, in the case of a power company, they don't have true RMS meters, they measure against the assumption of a perfect sine wave. So a crappy pf means you get billed for using more power than you actually drew.

It has zero bearing on reality of your power usage or power wasting etc.

 

[Hugues]

Hi guys,

I more or less follow your discussion, seems interesting.

Anybody could share a link on some basics behind this concept of Power Factor ?

thanks

Hugues
Switzerland

 

[theRealFury]

http://en.wikipedia.org/wiki/Power_factor

Wikipedia is your friend (well sometimes).

 

[rhitee05]

Its not the ratio of used to wasted Power at all.

Its the ratio of real to appearent power.

Fair enough, I was sloppy in my usage of terminology. The definition of power factor is the ratio of real power (watts) to apparent power (volt-amperes, VA) OR the cosine of the phase angle between voltage and current waveforms. It is also related to reactive/imaginary power (volte-amperes reactive, VAR) because P^2+Q^2=S^2. It is very easy to see how this can apply to BLDC in that the concept of timing is related to the phase alignment between the current and BEMF waveforms. The non-sinusoidal waveforms would make it difficult to actually calculate, but conceptually it is the same as with sinusoidal.

My comment about wasted power was not referring to the reactive power as wasted. Reactive power doesn't isn't absorbed, it just sort of sloshes back and forth. But if you have a bad power factor and lots of reactive power, that makes the apparent power larger. The currents flowing in the system are related to the apparent power, so having lots of reactive power means higher currents and thus more losses.