We have been informed by a super computer manufacturer to assume a near unity power factor for one of it's products. Can anyone confirm is this is a typical value?
It is reasonable that the displacement power factor (angle between fundamental voltage and current) can be close to 1.0.... that is always achievable by simple reactive compensation. However the total power factor includes another factor sometimes called distortion power factor which is approx equal to the rms fundamental current over the rms total current (exactly equal if the voltage is sinusoidal). This factor cannot be 1 if there is any distortion (harmonics) present in the current waveform. Historically it has been difficult to eliminate harmonics from the current waveform. But I have heard of newer power supplies which sense the input current and use this to control the switching in a manner that keeps input current close to sinusoidal.
So in summary, it is possible. Displacement power factor can be controlled by simple reactive compensation. Distortion power factor can be controlled by sophisticated input-current-controlled switching. Total power factor (defined as real power divided by [Irms*Vrms])is the product of these two.
Suggestion: The power factor can be adjusted to be unity by using active harmonic filters or by hybrid harmonic filters. Additionally, certain types of power converters produce power factor unity on their inputs, e.g. unity power factor frequency changers, unity displacement factor frequency changers, etc.
See Standard Handbook for Electrical Engineers, 13th Ed by D.G. Fink and H.W. Beaty.
The supercomputer manufacturer can achieve to have the power factor near unity for its products.
Incidentally, the distortion factor for voltage Total Harmonic Distortion less than 5% is very close to unity.
