Harmonics Help

[cjhut]

I have a question concerning harmonics. We have a 240v single phase lighting panel that is feeding some 400w metal halide high bay lighting fixtures. The main CB is 100A and the branch circuits are 20A. When all of the lights are turned on the main breaker is seeing about 50 amps. Each of the branch circuits are only seeing about 10amps. The branch circuit conductors are all on the same conduit for about 10feet before they branch out from a junction box. However, they have been derated appropriatley and are 10guage. From day one when this panel was installed, the circuit breakers seemed to be unusually warm. The conductors are fine, but even the main breaker is warm. An infrared temp gun shows temps of 150 defF on the breakers, in 70 deg ambient temperature. Being that heat seems to be transferring from the brakers themselves, and the currents are by no means excceding any of the ratings, it must be harmonics generated by the all inductive load of the ballasts. This wasn't a cheap panel. I used a Seimens bolt-in type panel. I am wondering what I did wrong, and how do I prevent this in the future. Also, if someone could give me a simplified harmonics explanation, it would be nice.

 

[Skogsgurra]

First question: How are you measuring the current?

If you use a clamp and it is a TRMS clamp, then the thermal effects from the harmonics should be evident from the measurement.

If you calculate current instead of measuring it, the harmonics will add to thermal load and the effect will be exactly what you see. Your lamps do probably have a spec saying what the power factor (not displacement, but total power factor, sometimes denoted with greek letter 'lambda') is. Use that power factor when you calculate breaker and panel load.

There is so much written about harmonics on the web that you can find all information you need. But avoid the 'popular and simplified' explanations that involve ropes, boats, trains, beer froth and whatever people come to think of. Just think of harmonics the way they occur IRL, as deformed waveforms with peaks that are higher than needed and thus causing more heat than needed because heat is proportional to current squared. Twice the current is half the time does the same work, but heats conductor and breakers twice as much. Simple as that.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[cjhut]

I was using a fluke true RMS clamp on ampmeter. Would I then use the formula I=P/VxPF? I don't have the power facor or total watts available right now, but assuming that I did and calculate it, would That show me the load that the panel is seeing with the harmonics?

 

[Skogsgurra]

No, you just use the current the Fluke shows. No calculation needed.

Does your Fluke also show crest factor? If it is very high, then it is possible that the fluke saturates and shows too little current. Have you tried to measure with the next higher range?

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[jghrist]

The breaker temperature of 65°C at 50% load does not necessarily mean that the temperature would be too high at full load.

 

[jghrist]

There will be more heating with the same current at higher frequencies (harmonics) because skin effect will increase the resistance.

 

[cjhut]

This is all good stuff guys. I appreciate the input. I have done many lighting panels in our plant, and have only seemed to see problems with 2 of our panels. I don't have a complete understanding of harmonics, but I do try to oversize conductors and mains, circuit breakers to allow for this. I don't use any special formulas, but I take into account the loads and just overkill everything a little bit. For example, this panels total load is a little under 50%, and I used 10gage copper Thhn conductors with 20amp branch circuit breakers, where the loads are only 10amps at the most. In one of our shops we just have 480v available, so we put in 480v highbay lighting, with a molded case circuit breaker panel. These breakers are by nature larger frame sizes than a 240v breaker. The loads are almost identical as far as 20amp branch cicuits and about 50% load on each. These breakers are room temperature. I am wondering, because of the increased buss size and frame size(even though its 100a main as well), that the skin effect is reduced. I guess that I am wondering, how in the future can I prevent this problem. Should I go with a 480v panelboard for 240V lighting for better heat dissipation just in case? Could I calculate what the tatol harmonic distortion would be, and then size it from there? And how exactly would I do that?

 

[marks1080]

I still question that your problem is actually due to harmonics. Be sure to diagnose the problem properly before trying to find a solution.

 

[jghrist]

I think you would need to get a PQ meter to measure harmonic distortion rather than try to calculate it.

Like marks1080, I question whether harmonics are really the problem.

What is the normal temperature of the thermal element of a thermal-magnetic breaker with 50% load? I'm not convinced that there is even a problem.

 

[cjhut]

Maybe you are right jghirst and marks1080. The obvious problem would be heating due to poor mechanical connections-- but that's simply not the case. I never thought of the thermal element itself causing heat, but that just seems awful hot under a normal 50% load. Not to mention that my IR temp gun is probably underreporting a little bit from the actual temerature right at the element. If it's not harmonics and it is the breaker itself, I am not used to seeing that kind of heat in normal situations. I have many of these exact Siemens bolt in panelbaords, and i don't see these kinds of temps.

 

[marks1080]

The lights you mentioned wouldn't cause harmonics by themseleves. Is there anything else off the board besides lighting? Is there a computer lab or something being fed? Large motor controllers? Inductive loads by themselves will not cause harmonics.