Harmonic Distortion

[harlista]

Good morning;

I’m new to this forum and have a question for the collective. I work in a data center in Texas. We use two utility feeds entering our building, each side feeds a medium voltage switchgear unit which, in turn, feeds 4 delta-wye transformers (27KV in 480V out) per side. Two sets of switchgear per side have a load of about 100KW and 300KW and run about 10% to 12% total harmonic distortion. The remaining two sets of switchgear on each side are loaded at 400KW and 750KW and run at about 5% to 7% THD per phase. The 400KW switchgear side feeds 3 UPS systems, 480V delta in and 480V wye out. On the output side of the UPS I'm reading a THD of 25%, 19% and 10% phase A, B and C respectively. Each UPS feeds 4 or 5 PDUs with a 480 delta input and 208/120 wye output.

I know the delta-wye configuration inherently mitigates some of the harmonics but was wondering if the delta-wye to delta-wye to delta-wye configuration we are using was causing a grounding issue and possibly the high harmonic content on my UPS output. Or does the non-linear load the PDUs supply reflect harmonics back to the UPS? Or is a THD value of 25% considered high and worthy of farther investigation?

Thanks in advance for any insight on this situation.

 

[Senselessticker]

THD is a strange bird. Harmonic distortion is not a problem *unless* its a problem (i.e. putting harmonics back on the utility at the point of common couplng). I'm not sure what PDU stands for, however if its a non-linear load, then it is putting "noise" on the system as seen by the UPS.

 

[dpc]

When talking about harmonic distortion, you need to specify if you mean voltage distortion or current distortion. I'm going to assume you mean voltage distortion, because that's the simplest to measure.

The UPS inverter is probably one source of the distortion. Non-linear loads dowstream will contribute as well. The UPS draws current in a non-linear fashion and this creates harmonic voltages on the supply side.

Measuring distortion on the output of the UPS may or may not be meaningful, depending on the type of UPS inverter and the equipment you use to do the measuring. Measurement on the input side of the UPS should be more meaningful.

Transformer connection will have little impact on the distortion from one UPS. For multiple UPS, you can improve the voltage distortion created on the line side by using a combination of wye-wye and delta-wye transformers to take advantage of the 30 deg phase shift that will provide some cancellation effect.

But as already mentioned, if you're not having problems, it may not be worth spending much more time worrying about, especially if it has been in service for a while without problems.

If you're laying 10% voltage distortion onto the utility's incoming feeders, they may have a beef with you, but until another customer complains, I doubt that they will get involved, especially if your kwh meter is spinning nice and fast.

 

[ScottyUK]

As dpc said, are you talking about current distortion or voltage distortion? 25% voltage THD is poor performance for any modern UPS to be producing and could give you problems with your load; 25% current THD certainly isn't inconceivable, especially with computer SMPS loads. If the voltage waveform is good and the UPS isn't showing any problems I wouldn't worry about current THD on the load side.

The PDU is an integrated isolation transformer and distribution package. They're quite common for large computer installations.

Your PDU should be shielding the UPS from the triplen harmonics by preventing them propagating through the delta, but it shouldn't add harmonics of its own.

----------------------------------

I don't suffer from insanity. I enjoy it...