What's power factor with a nonlinear load?

[Bruce Varley]

This topic has been aired in some recent posts, so there may have been an
answer, but I stopped following the threads when the invectives started.

Could someone please explain clearly what the definition of power factor is
in the case of a nonlinear load? Preferably something official, such as
maybe from the IEC standards.

I keep seeing references to the power factor of things like CFLs being
'low', I'm not sure on what basis that's being stated. For a true voltage
source, if the current spikes are right on the voltage peaks, then there is
a good argument that the PF should be 1.

 

[Phil Allison]

"Bruce Varley"

This topic has been aired in some recent posts, so there may have been an
answer, but I stopped following the threads when the invectives started.

Could someone please explain clearly what the definition of power factor
is in the case of a nonlinear load? Preferably something official, such as
maybe from the IEC standards.

I keep seeing references to the power factor of things like CFLs being
'low', I'm not sure on what basis that's being stated. For a true voltage
source, if the current spikes are right on the voltage peaks, then there
is a good argument that the PF should be 1.

** The general definition of Power Factor is the ratio of Watts to VA.

ie PF = Watts / VA

where V = rms voltage

and A = rms amps.


( Notice there is no mention of " cos phi" )


So, since the AC supply is a 240 volt rms sine wave

- VA is determined by the rms amps draw alone.

Consideration of phase angle or " cos phi" ONLY applies where the current
draw is also a sine wave.

See:

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


In a nutshell,

PF compares how hot cables get to watts consumed.

The ratio of 1 then equates to " good as possible ".




....... Phil

 

[Bruce Varley]

"Bruce Varley"


** The general definition of Power Factor is the ratio of Watts to VA.

ie PF = Watts / VA

where V = rms voltage

and A = rms amps.


( Notice there is no mention of " cos phi" )


So, since the AC supply is a 240 volt rms sine wave

- VA is determined by the rms amps draw alone.

Consideration of phase angle or " cos phi" ONLY applies where the
current draw is also a sine wave.

See:

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


In a nutshell,

PF compares how hot cables get to watts consumed.

The ratio of 1 then equates to " good as possible ".
...... Phil

Thanks for the clear definition, Phil. I hope people will bear with me while
I take this a bit further.

I thought this formula would calculate a PF of 1.0 for a nonlinear load
typical of an SMR waveform, with sharp current spikes around the voltage
peaks, but fortunately decided to shut up and try a few simulations before
making any more comment. To my surprise, when I applied Phils formula to
such a current waveform, I came up with a 'result' (whether it's 'power
factor' or not is a very moot point) that is well under 1.0.

Why is this so? The reason a PF in the sinusoidal (linear, reactive load)
situation is low is that the voltage and current are out of phase, so that
there are portions of the cycle where the voltage and current are opposite,
and the reactive load behaves like a generator, returning energy back to the
source. So on average, the energy dissipated in the load is less than the
total volts * amps. But with the SMR waveform, the voltage and current are
always of the same sign at any one instant, so there is no time when energy
is being passed back from the load ot the source. Shouldn't the PF be 1.0?

I can't come up with a clear explanation for the result yet, it might be
just a mathematical artefact, with no clear physical significance.

Is this important? IMO, yes. Because we're talking of two completely
different loss mechanisms, that may require quite different approaches for
mitigation. In the case of true, low PF, the losses are due to that excess
current, that heats the transmission system up but doesn't show on your
meter and doesn't do anything useful (ignoring reactive power system
stability issues). In the case of the nonlinear load, the problem is the
nonlinear effects, of which 'harmonics' might be only part of it (because
superposition doesn't apply, if you want the technical reason for that).
Conventional PF correction is unlikely to help here, in fact providing nice
big fat caps to help the high frequencies to circulate could well make
things worse. I don't know what you do for a power network driving millions
of switchmode devices, all the way from tiny phone chargers up to big VVVF
drives.

All this might be related to whether it's smart policy to chuck out trannies
and light bulbs too...

 

[Poxy]

Thanks for the clear definition, Phil. I hope people will bear with
me while I take this a bit further.

I thought this formula would calculate a PF of 1.0 for a nonlinear
load typical of an SMR waveform, with sharp current spikes around the
voltage peaks, but fortunately decided to shut up and try a few
simulations before making any more comment. To my surprise, when I
applied Phils formula to such a current waveform, I came up with a
'result' (whether it's 'power factor' or not is a very moot point)
that is well under 1.0.
Why is this so? The reason a PF in the sinusoidal (linear, reactive
load) situation is low is that the voltage and current are out of
phase, so that there are portions of the cycle where the voltage and
current are opposite, and the reactive load behaves like a generator,
returning energy back to the source. So on average, the energy
dissipated in the load is less than the total volts * amps. But with
the SMR waveform, the voltage and current are always of the same sign
at any one instant, so there is no time when energy is being passed
back from the load ot the source. Shouldn't the PF be 1.0?
I can't come up with a clear explanation for the result yet, it might
be just a mathematical artefact, with no clear physical significance.

Is this important? IMO, yes. Because we're talking of two completely
different loss mechanisms, that may require quite different
approaches for mitigation. In the case of true, low PF, the losses
are due to that excess current, that heats the transmission system up
but doesn't show on your meter and doesn't do anything useful
(ignoring reactive power system stability issues). In the case of the
nonlinear load, the problem is the nonlinear effects, of which
'harmonics' might be only part of it (because superposition doesn't
apply, if you want the technical reason for that). Conventional PF
correction is unlikely to help here, in fact providing nice big fat
caps to help the high frequencies to circulate could well make things
worse. I don't know what you do for a power network driving millions
of switchmode devices, all the way from tiny phone chargers up to big
VVVF drives.
All this might be related to whether it's smart policy to chuck out
trannies and light bulbs too...

Phil quoted this link in another post:
http://sound.westhost.com:80/articles/external-psu.htm

Which covers some of the points you made, and references this document:
http://www.elec.uow.edu.au/iepqrc/files/technote3.pdf

Which specifically addresses the impact of harmonic distortion on the power
system, and mentions the negative side-effects of power factor correction
caps when confronted with harmonic distortion.

Incidentally, today I was talking to a pump supplier about a new unit for a
site, and they suggested a variable-speed drive unit to drive a 3-phase
pump - these units commonly quote a power factor of 1, but on a 6 KW pump I
imagine the harmonics would be a significant issue that seems to be ignored.
Then again, while I think a 6KW pump is kind of a big motor, it seems every
Dick and Harry is installing 6KW split-system heat pumps in their homes,
many of which are "inverter" style, variable speed units...

 

[Phil Allison]

"Bruce Varley"
"Phil Allison"

Thanks for the clear definition, Phil. I hope people will bear with me
while I take this a bit further.

** Frankly folks - I wos expecting this ............

Bruce ..........


Remember this bit " A is determined by the rms amps draw alone" ?

Got any idea what " rms " is all about ??

The fact that you do NOT is the problem.

YOU need to do some basic arithmetic.

Compare the cases of pure DC current and pulsed DC current.

Figure out why pulsed current heats the supply cables a WHOLE bunch more
than DC current does - when the average power in the load is just the
same.

Shame if you are maths challenged.

I am NOT gonna do it for ya.





........ Phil

 

[Franc Zabkar]

This topic has been aired in some recent posts, so there may have been an
answer, but I stopped following the threads when the invectives started.

Could someone please explain clearly what the definition of power factor is
in the case of a nonlinear load? Preferably something official, such as
maybe from the IEC standards.

I keep seeing references to the power factor of things like CFLs being
'low', I'm not sure on what basis that's being stated. For a true voltage
source, if the current spikes are right on the voltage peaks, then there is
a good argument that the PF should be 1.

I agree, it is counterintuitive (because of the way that PF is
taught), but here is a worked example that cleared it up for me:

http://groups.google.com/group/aus.electronics/msg/fe5b24eacbfeb277?dmode=source&hl=en

- Franc Zabkar