Input power factor for electronic equipment?

[nukem2k5]

This question is for an industrial installation. I'm looking for somewhat theoretical insight.

I'm doing some rough preliminary UPS sizing. We will have a significant amount of 'system cabinets' for the process control system, fire & gas system, and SIS system. The cabinets comprise various components like

• AC/DC power supplies for the logic solvers and analog/digital IO modules
• IS barriers,
• AC/DC power supplies for field devices (e.g. Phoenix PSU)

The UPS output will be 230VAC single phase delivered to these 'system cabinets'. There's a debate among my team regarding the appropriate input Power Factor to assume for these cabinets. Obviously, assuming a 1.0 PF will result in the most conservatively sized UPS (i.e. assuming a 0.8 PF standard output rating, the unity PF will increase the Watt output rating of the UPS and hence its overall rating).

My thinking is that these electronic devices, particularly the various power supplies, are not purely resistive due to the nonlinear elements in their circuitry (coils, capacitors, etc.).

To elaborate, have a look at this spec sheet, page 152 of 166. It's tough to glean any consistent insight from the input voltages, power consumption/current values, and heating values. I would think that the heating value is basically the approximate 'real work' being done by the equipment (plus inefficiency), so the (vectoral) difference in input power would basically be the Reactive power consumed by the load (i.e. VARs). By that logic, the first item in the list would have a very poor power factor when supplied at 230VAC (120W / 230VA = 0.52 PF). However, if you were to supply DC voltage, the power consumption is (24VDC * 5.5A = 132 W), with only 120 W heating value. In that case, I assume the difference in DC power consumption (132W) and the heating value (120W) is just an accepted margin of error in the approximation.

Further, the high in-rush currents listed on pg 154 suggest to me the presence of significant amount of nonlinear components, so clearly these devices consume reactive power (although I can't find a definition of the column headings 'Primary' vs 'Secondary). Similar to an induction motor, their steady-state reactive power consumption may be low, although the only 'real' work being done by electronics is effectively thermal/heat creation.


Anyway, based on my logic that the devices have nonlinear components, would it be more appropriate to assume 1.0 load PF (purely resistive), or something like 0.85? Anyone have any previous experience with this?

 

[Jraef]

Most things electronic will have a Switch Mode Power Supply (SMPS) and as such are now usually required to have Power Factor Control (correction) to 0.75 or better based on EU regulations that went into effect in 2001. This is not your typical displacement Power Factor, it is distortion Power Factor, but the net effect on sizing your UPS is the same. What happens on the load side of the SMPS is irrelevant. So although not required here in North America yet, the fact that it is a "world market" now means that we benefit from those regulations. But there are different levels of requirements for different types of systems, here's how it breaks down.

SMPSs with passive PFC can achieve power factor of about 0.7–0.75, SMPSs with active PFC, up to 0.99 power factor, while an older (pre-2001) SMPS without any power factor correction have a power factor of only about 0.55–0.65. If you assume a .75 on average, you are probably good to go in terms of sizing your UPS.

Anything that looks like a "drive", as in a VFD, DC drive or Servo Amplifier will be .95PF or better.

 

[nukem2k5]

although not required here in North America yet, the fact that it is a "world market" now means that we benefit from those regulations.

Thanks. I should've mentioned that this installation will be the North Sea region and hence subject to EU and local UK regulations (including IEC design standards, obviously).

Suppose I ought to do some reading on SMPS


Sent from my SAMSUNG-SM-G930A using Tapatalk

 

[Besoeker]

Anything that looks like a "drive", as in a VFD, DC drive or Servo Amplifier will be .95PF or better.

Have to disagree with you on that one, old chap.
I grew up with DC drives of the phase-controlled variety. SCRs. Often in paper mill applications. Typically machine speed would be from "crawl" at 10mpm to production at 1000mpm. Metres per minute.
Power factor varied with firing angle from dreadful on crawl to about 0.8 at full speed. It was the same for competitors of course.

 

[Jraef]

Have to disagree with you on that one, old chap.
I grew up with DC drives of the phase-controlled variety. SCRs. Often in paper mill applications. Typically machine speed would be from "crawl" at 10mpm to production at 1000mpm. Metres per minute.
Power factor varied with firing angle from dreadful on crawl to about 0.8 at full speed. It was the same for competitors of course.

No argument, but most new DC drives being sold now are PWM output using a 6 pulse diode bridge rectifier; basically the same as an AC VFD, just with a DC output instead of an AC output. They are smaller and cheaper and more efficient.

 

[Besoeker]

No argument, but most new DC drives being sold now are PWM output using a 6 pulse diode bridge rectifier; basically the same as an AC VFD, just with a DC output instead of an AC output. They are smaller and cheaper and more efficient.

Yes, possibly so and maybe it depends on ratings. We did develop a range of DC drives called UPF as in unity power factor. These were used in process industries where there were a number of DC drives on the same machine. These had a common DC input and individual choppers for each drive.The last one I recall had an an uncontrolled 24-pulse rectifier fed by two 2MVA transfomers and producing 710Vdc. Some of the drives had to be regenerative and fed back into the common DC supply.

 

[nukem2k5]

This is not your typical displacement Power Factor, it is distortion Power Factor, but the net effect on sizing your UPS is the same.

Do you mind elaborating on that a little bit, to help me better understand the physics? I try to think of power factor simply as a phase shift between the current and voltage. The way I interpret what you said, I think more in terms of harmonics, which isn't the same thing. Are you saying that the switching done by the PSU is also shifting the current and voltage to be out of phase? I would think that the cause of any phase shift is due to the capacitors and inductors in the PSU circuit.

 

[Jraef]

Do you mind elaborating on that a little bit, to help me better understand the physics? I try to think of power factor simply as a phase shift between the current and voltage. The way I interpret what you said, I think more in terms of harmonics, which isn't the same thing. Are you saying that the switching done by the PSU is also shifting the current and voltage to be out of phase? I would think that the cause of any phase shift is due to the capacitors and inductors in the PSU circuit.

Hmmm... well here is my super condensed version. Traditionally, what we have referred to generically as "Power Factor" is what you described, and is technically "displacement" power factor, because it describes the displacement of the voltage and current. But non-linear loads draw current from the line in "pulses" because the diodes in a bridge rectifier only conduct at the peaks of each sine wave. So for every voltage half sine wave, the associated diode conducts all of the necessary current during a narrow window of time that is defined by the diode's "Forward Conduction Voltage" (FCV). Below the FCV the diode blocks current flow, above it, it allows current flow, and when the sine wave declines, it stops conducting again until the next wave. So that is what defines the term "non-linear" loads; the current draw is not taking place in the same smooth "linear" sine wave fashion, it pulses within each sine wave. That pulsing then causes distortion in the sine waves, what we refer to as Harmonics because that distortion is the result of additional higher frequency sine waves being superimposed on top of the fundamental. So instead of the Power Factor now being a simple triangular formula relationship to current and voltage of one sine wave, there is now a complex mixture of higher frequencies taking place simultaneously and the total Power Factor takes on a 3D relationship with frequency in the mix. So if you take the total PF as simply kW/kVA and calculate the displacement PF as that angular difference between the fundamental voltage and current, they will no longer be the same, and the difference is called the distortion PF because it is taking place at higher than fundamental frequency as a result of the harmonics caused by the device in question.

Capacitors and such in the power supplies are on the DOWN STREAM side of the rectifier, so they correct the PF issues created by the LOAD, but not those created by the rectifier itself.

Found a graphic for you:

 

[Besoeker]

Hmmm... well here is my super condensed version. Traditionally, what we have referred to generically as "Power Factor" is what you described, and is technically "displacement" power factor, because it describes the displacement of the voltage and current. But non-linear loads draw current from the line in "pulses" because the diodes in a bridge rectifier only conduct at the peaks of each sine wave. So for every voltage half sine wave, the associated diode conducts all of the necessary current during a narrow window of time that is defined by the diode's "Forward Conduction Voltage" (FCV).

For continuous Idc, each diode conducts for 120deg of the cycle. On load it is continuous except for el cheapo power electronics.
Consider the phases to be A, B, and C. The A+ diode conducts through the load to the B- for 60deg, then to the C- for a further 60deg.
From a VSD for a cement works.

The distortion power factor is 0.958. It would be much lower for discontinuous Idc.
I know that doesn't directly answer the OP's question, but there is no general one size fits all answer.

 

[nukem2k5]

Capacitors and such in the power supplies are on the DOWN STREAM side of the rectifier, so they correct the PF issues created by the LOAD, but not those created by the rectifier itself.

Very interesting. Thanks for the explanation. Is it accurate to say that the 'total' PF is a (vectoral?) combination of [1] the voltage and current displacement (phase shift) at the fundamental frequency, as well as [2] the displacement (phase shift) at the higher-order frequencies (where [2] only exists due to the non-linear elements in the circuit)?

In other words: the harmonics also affect the 'total' Power Factor (in the form of 'Distortion Power Factor') because the harmonic (higher-order) currents are phase-shifted (displaced) with respect to the harmonic voltage waveforms? Or is Distortion PF completely independent of phase shift, i.e. the higher-order voltage and current waveforms are 100% in phase with each other but still somehow increase the VAR demand?

Or is it simply that the distortion (harmonic) current is purely reactive, thus lowering the PF beyond the simple "displacement PF"?