Odd Harmonics in Power System - reduction

[OliskaP]

Hi,

could anyone explain to me how odd harmonics in the power system are reduced or removed? I've read that in transformers the harmonics produced by the load circulate if the secondary side is wired as delta configuration, yet star is usually used on secondary side i believe?

I've looked around a little and I saw that zigzag transformer could help reduce the harmonics, how? I could not find any good explanation about this, also I do not fully understand how a zig zag transformer work.

Are there any other common techniques used to reduce the influence of harmonics in the power system?Best regards.

 

[anorlunda]

Shunt capacitors are sometimes used for that purpose. The higher the frequency, the more a shunt capacitor looks like a short circuit.

 

[OliskaP]

Shunt capacitors are sometimes used for that purpose. The higher the frequency, the more a shunt capacitor looks like a short circuit.

So shunt capacitors actually gives two benefits in a power system. Harmonic reduction and reactive power compensation, i.e. helping such that voltage stays within specified limits.

Are shunt capacitor designed to do both tasks at the same time, or do you have to design them individually?

 

[Svein]

So shunt capacitors actually gives two benefits in a power system. Harmonic reduction and reactive power compensation, i.e. helping such that voltage stays within specified limits.

Yes. but a capacitive load also introduces a phase shift which we do not want ( I am sure "anorlunda" can explain that better than I can).

 

[jim hardy]

If you draw on a sheet of graph paper three sinewaves 120 degrees out of phase, draw first one brown, draw second one red and third one orange,
(or purple or whatever colored pencil you have)
and then on same paper draw three more sinewaves of 3X frequency , aligning the first zero crossing of each with its corresponding 1X frequency sinewave zero crossing,

you will see that the third harmonics for all three phases are exactly in phase. You'll have to draw all three 3rd harmonic colors overlaid.
http://zone.ni.com/reference/en-XX/help/373375C-01/lveptconcepts/ep_harm_interharm/

That fact has some interesting ramifications
one of which is that in a delta connected winding third harmonic current just runs around the triangle .

Look at picture 1A below, KVL says all 3 windings are in series for 3rd harmonic and the only thing limiting current is the combined impedance of the transformer and its power source.
The third harmonic is short circuited at the Delta winding and the transformer will accept as much of the third harmonic current as its source can deliver. That can burn up a small transformer that's connected to a large source. But a big transformer capable of handling all the 3rd harmonic its source can produce will simply circulate them and clean up the sinewave.
In a wye connected winding the third harmonics combine and flow through the neutral. That can burn out the neutral wire in a distribution system that feeds big nonlinear loads like computers. For that reason NEC was revised to require neutral wire for computer installations be at least as large as the phase conductors.

It's common to use a delta winding to short out 3rd harmonics which cleans up the sinewave.
Big power transformers are often equipped with a "tertiary" winding for that purpose. It gives a handy place to get power for the cooling fans.
"o"in this picture means "zero sequence" , for this discussion just call it third harmonic

see also
http://www.hammondpowersolutions.com/files/HPS_article_Zero_Sequence_Harmonics.pdf
http://www.allaboutcircuits.com/textbook/alternating-current/chpt-10/harmonic-phase-sequences/Zigzag transformer uses two windings in series whose voltage is shifted by 60 degrees, 1/6 cycle.. At third harmonic that's half a cycle, 180 degrees, so the third harmonic voltage in one winding cancels that in the other winding.
http://static.schneider-electric.us/docs/Electrical Distribution/Low Voltage Transformers/Harmonic Mitigating/7400DB0301.pdf

Hope that helps you get started.

old jim

 

[anorlunda]

So shunt capacitors actually gives two benefits in a power system. Harmonic reduction and reactive power compensation, i.e. helping such that voltage stays within specified limits.

Are shunt capacitor designed to do both tasks at the same time, or do you have to design them individually?

Yes they are used for both purposes. There are many designs of capacitors. A design has ranges of voltage, current, temperature, vibration, duty cycle, off gas removal, and so on. I'm not familiar with all the considerations of capacitor design.

But I'm reminded of an incident where my firm recommended $2 milllion in capacitors to filter harmonics from a scrap steel arc furnace in Mexico. But after only two weeks of service, the entire $2M capacitor bank exploded. Explosions are not something you can predict using the capacitance value and Ohm's law.

Yes. but a capacitive load also introduces a phase shift which we do not want ( I am sure "anorlunda" can explain that better than I can).

Transformers also cause phase shift. In fact anything that will be effective will probably cause phase changes. Detailed calculations including harmonics for the whole nearby power grid can be complex, but that might be what you need to predict all the side effects of a change.

To filter harmonics with shunt capacitors without changing power factor can mean adding shunt reactance to compensate the shunt capacitance. The L and C values might cancel at the fundamental frequency, but not at higher frequencies. But with both installed, beware resonances. There are also nonlinear power electronic solutions for filtering and reactive compensation, that can do a better job, but cost more money.

Anything like this costing thousands or millions of dollars for investment (or for consequences) needs a professional engineering study, not a quick answer on an Internet forum.

Good luck

 

[OliskaP]

Thank you all. I hope that some time in the future I will be as knowledgeable within power system as you guys.

 

[LagCompensator]

Zigzag transformer uses two windings in series whose voltage is shifted by 60 degrees, 1/6 cycle.. At third harmonic that's half a cycle, 180 degrees, so the third harmonic voltage in one winding cancels that in the other winding.
http://static.schneider-electric.us/docs/Electrical Distribution/Low Voltage Transformers/Harmonic Mitigating/7400DB0301.pdf

old jim

Hi, sorry for interrupting your thread @OliskaP.

@jim hardy, How exactly do a zig zag transformer manage to shift 60 degrees? I neither fully understand the zig zag transformer, and can not really seem to understand how this 60 degree phase shift is created.

best regards.

 

[jim hardy]

.

@jim hardy, How exactly do a zig zag transformer manage to shift 60 degrees? I neither fully understand the zig zag transformer, and can not really seem to understand how this 60 degree phase shift is created.

Actually I'm not expert on them either.
There's a presentation on transformers her, zigzag tarts around slide 39.
http://www.slideshare.net/akashsolanki/2-equivalent-circuits-of-power-transformers

Let's try a graphical approach.

Here's a representation of a three phase transformer in 3-d
only one winding is shown.

Now connect the windings to make a wye

Let's draw the phasor diagram

That's straightforward enough.
Now in order to not have a messy drawing let's run a thought experiment.

What I've drawn so far could be showing a primary winding exciting a transformer core .
It could also be a secondary winding on a core that's excited by a primary that isn't shown, couldn't it ?
Indulge me and assume the latter. I'm not good with Paint and it'd take me a week to draw multiple windings, so just remember the core is excited by an invisible primary... Be kind to the old guy, please...

Now i will attempt to split my secondary windings in two and redraw the phasor diagram
.

Now let's look at how we'd connect one arm of the secondary to make a zigzag, i'll start at N and go up to A1, then over to C3 and come out C2

system is acting up, my post in progress disappeared for a while so i'll double post this then try to edit it
got about three hours in it and don't want to lose that.

Now let's look at how we'd connect one arm of the secondary to make a zigzag, i'll start at N and go up to A1, then over to C3 and come out C2
and i'll call that new exit point A'

Now let's try for a phasor diagram but so as to get not overwhelmed i'll only draw the one leg we've hooked up.
I'll draw segment N-A1 and add segment C2-C3 to it. Since we went backward through C2-C3 i'll swap his head and tail then slide him up and right to connect C3 & A1 like we wired him.

That's how it does 60 degrees, remember from middle school geometry adjacent angles are supplementary.

Now what are the ramifications ?

Consider the flux in the core. Go back to that 3D diagram.
What is mmf from currents flowing in leg A.?
Since it has current from two different phases flowing in opposite directions,
third harmonics will be canceled and so will any DC content.
That is a big deal - core losses are lessened, impedance of the transformer winding to third harmonics is lessened, and it won't try to walk up the BH curve and saturate from a DC content in the load current.

Aha!
At last i understand what they were trying to tell me about the zigzag connected generator on my control rod drive system. It supplied a half wave rectifier thyristor control that made a few hundred amps of DC to move the control rods.
I've never before succeeded at figuring out the zigzag step by step, just remembered some words... So i thank you for the question.

And i hope anyone who sees a mistake will correct me. First drafts are prone to error.
?

old jim

 

[jim hardy]

.This post is redundant. It didn't carry the images from previous post and i don't know what became of the attachmed files so I'm leaving it alone for now. If the previous one still looks complete in the morning i'll delete this one, but am not leaving all the eggs in one basket for now.

jh

EDIT , the morning after
Images are still there so i'll get rid of this redundant one.
As always i see little things i missed in my drawing, like one too many turns in phaseB of split secondary 3D sketch, but that's cosmetic.

jh

 

[LagCompensator]

@jim hardy - Amazing, fantastic explanation.

 

[OliskaP]

@jim hardy , excellent graphical approach - as always.

 

[jim hardy]

Thanks guys for the kind words...

old jim

 

[P Gray]

Hello Folks, Interesting forum topic and varied comments. I design harmonic filters that provide both PF Compensation with VARs, and harmonic attenuation via LC Shunt filter design. If I am reading my colleagues responses correctly above, there appears to be a misunderstanding concerning applying capacitor KVAR into an electrical system and its benefits / challenges.

Jim is absolutely correct in that in todays modern commerical and industrial power systems, due to the propogation of harmonic current sources / loads throughout, harmonic currents of different frequencies, I.E. Positive sequence (1,4,7 harmonics, etc), Negative sequence (2. 5. 8. 11 th order harmonics, etc), and zero sequence components (3, 6, 9, 12, 15th order harmonics) pose different effects in any given power system.

Positive sequence harmonics are largely harmless. Negative sequence harmonics create current flow into motors in opposite direction of fundamental current and create overheating in windings and insulation failures, a consistent problem in industrial power systems with many many motor loads. Also, the 5th harmonic, a negative sequence component, is by and large the most prevalent harmonic current injected into power systems from six pulse drives. Many harmonic filters I design are tuned or detuned specifically around the 5th as the predominant harmonic in amplitude encountered. Finally, ,the Zero sequence harmonics, I.E. the 3rd being the most prevalent in most modern power systems, is frequently the frequency driven into resonance from applying capacitors or poorly designed / tuned harmonic filters into clients power systems. the fact that one of our colleagues above mentioned a million dollar capacitor failure for a steel client illustrates that NOTHING can create power system failures or instablity quicker than a misapplied capacitor bank.

In modern, harmonic rich power systems, it has been my experience that an engineering study to examine harmonic complexion of the power system, the current / desired power factor improvement goals of the power system, and incremental steps of capacitance introduced to the power system in KVAR Capacity must be modeled prior to final filter recommendation / to qualify design. I model the system expected FREQUENCY RESPONSE to check for resonance prior to finalizing any capacitor / or harmonic filter design, to ensure that the design I propose will be compatible with the clients impedance profile. In this way, we use mathematics and software to ensure the any stepped capacitance, or any combinations of stepped capacitance we propose will NOT drive the power system into resonance. As a primer, Resonance is a condition where a current (Frequency) at or near a harmonic ( a harmonic multiple of 60 hertz) at or near a dominant harmonic oscillates UNIMPEDED (no damping) between harmonic source ( A drive, computer network, UPS, or any harmonic source) and dominant XL or network (typically the utility service entrance transformer), where Xc equals Xl at or near a power system harmonic frequency.

In summary, one should typically be wary of applying straight capacitor bank only into todays modern power systems; for Power Factor improvement, for harmonic attenuation, and voltage stability, a harmonic filter of some sort should be applied.

If you need harmonic attenuation or PF Correction, the company I work for is one of several resources. See us on the web at [Advertising link redacted by the Mentors] Blessings to all; and Jim youve done a great job helping this discussion.

Paul