Non Linear Load Calculation

Status
Not open for further replies.
T

timm333

Senior Member
Location
Minneapolis, MN
Occupation
Electrical Design Engineer
I am trying to figure out the load calculations of non-linear office loads (computers). When we have motors, we use capacitors to improve the power factor. And when we have VFD's then we don't have the issue of low power factor, but we use line-reactors to filter the harmonics. I have following two questions:

1) What is difference between transformer sizing for linear-loads and transformer sizing for non-linear-loads; if we have 20 kW non-linear office load (computers) at 120/208 V, then will a 30 kVA 3-phase transformer be enough?
2) Non linear office loads have a low power factor (70%). How do we improve the power factor of non-linear office loads, should we use capacitors or reactors or something?

Thanks.
 
Personally, I don't think I have sized more than (1) non-linear general purpose transformer in the past 40 years. The only time I see them is when an consulting firm is doing a public/government project.
Since the advent of switched mode power supplies, desktop and laptop computers do not present the amount of harmonics they did 20 years ago.

That said: If your loading is really non-linear 20kW @.70PF, I would consider a 30kVA transformer as being too small. Up size it or specify a non-linear design.
 
The document at the link above states that computers have two power factors: the displacement power factor and the distortion power factor. The displacement power factor of the computers is 100%.

In addition to the computers, we also have some motors which have power factor of 85%. As the displacement power factor of computers is 100%, will the overall power factor of the system improve (become more than 85%) when motors and computers are running together?
 
Also if we want to reduce the harmonic current drawn by the computers, how to reduce this harmonic current? Thanks.
 
The document you posted is almost 20 years old. Computer power supplies are much better now. Typical office computers won't put enough harmonic content on your system to matter. If you are really concerned, get a good recording meter and hook it up to your computer for a while and see for yourself.

If the loads are fairly consistent and the harmonic content is known, you can cancel out the important portions with matched, phase-shifting transformers. I would not expect this to be the case with the scenario you described. If you are having issues meeting the utility company's requirements at the PCC, about the only way to handle distributed, inconsistent harmonic loading is with an active filter system at the building's service entrance.
 
Personally I wouldnt worry about non linear loads for all but a few very specific applications. Consider the following:
1. modern power supplies are much better now (as has been mentioned
2. NEC load calcs are very conservative
3. There will typically be other loads besides the computers, and the computers themselves are not that big a load (unless we are talking a data center, that being one of the few specific applications, but even then presumably there is lot of HVAC, potentially making the non linear component not that significant).
 
Status
Not open for further replies.
Top