Why do we sometimes choose delta over wye for transformers and generators?

[psparky]

Speaking of delta and wye transformers and generators, why do we sometimes use one over the other?

The WYE's seem easier to ground...but why pick a delta over a wye...or a wye over a delta in certain situations?

 

[psparky]

Clearly, the WYE has the advantage of the line to neutral as well...so why on Earth use a Delta?

 

[jim hardy]

One reason is sort of sneaky: it attenuates third harmonic distortion.

graph your third harmonic voltage. On all three terminals it's in phase.
So a delta connected winding is a near dead short circuit for third harmonics.
So it'll clean up a sinewave.

https://www.physicsforums.com/showthread.php?t=601222

http://www.epc-website.com/PDF/Myth1.pdf

 

[Aussielec]

I'd say the biggest advantage of a star connection over a delta is that you have two voltages available, as opposed to just one. That said you can centre tap one of the delta windings creating a high leg system where you actually have three voltages available.

 

[psparky]

One reason is sort of sneaky: it attenuates third harmonic distortion.

graph your third harmonic voltage. On all three terminals it's in phase.
So a delta connected winding is a near dead short circuit for third harmonics.
So it'll clean up a sinewave.

https://www.physicsforums.com/showthread.php?t=601222

http://www.epc-website.com/PDF/Myth1.pdf

Oh...that's right, cleans up harmonics...now I remember.

Thanks, Jim.

Oh, and aren't most transformers Delta primary, WYE secondary? I guess that would give the best of both worlds?

 

[Windadct]

IN general the Delta is only used for dedicated loads, which only need a Line to Line supply, and where grounding is not as important or other technical reasons ( for large rectifiers for example often multiple winding configurations are used because they each have a different phase shift - good for another discussion)
Early in electrification delta was often used because it would NOT shutdown on a single leg ground fault - common in older large plants - they would have a ground detection system, and service the ground fault when they were off shift, but the improvement in insulation and the need for improved safety caused this scheme to fall out of favor.
Also note - most transformers still use Delta for the primary - or better described "source" side depending on power flow. e.g. a Step-up transformer for a generator will usually have the low voltage ( generator ) side be Delta and the HV - distribution or transmission side as wye.

 

[jim hardy]

And since a "separately derived" power system is to be grounded at its originating power transformer, it is convenient (but certainly not mandatory) to use a wye transformer winding ground its neutral. Windadct described it well.

The generator ground resistor I described [STRIKE]earlier[/STRIKE]
makes that generator-transformer pair a "High impedance grounded system" , where fault current is limited. Some such systems will operate with an accidental ground on a phase. EDIT: Oops- that was in different thread- 'how does grounded hv...'

Solidly grounded systems, in contrast, pass enough current into a fault to trip something.

"IEEE Green Book" is the best reference I know of on this subject - your facility needs to have one in the office. I bought my own, company library got it for me (with substantial library discount) some years ago.

 

[cabraham]

Actually both Y and delta have salient advantages. As Jim Hardy correctly pointed out, a delta does a great job at suppressing triplen harmonic magnetizing current, as well as maintaining balanced voltages when load currents are severely unbalanced while using only 3 wires. A delta-delta xfmr using 3 single phase units has the advantage that 1 unit can be removed for repair w/ the other 2 still providing balanced 3 phase service. But the total VA rating gets reduced to 57.7% of all 3 units, not 66.7% as intuition suggests.

A Y is ideal for connection to Earth ground because of the neutral point symmetry. A Y also provides lower insulation stress, important for high voltages. A delta is well suited for lower voltage high current venues. Ultimately a good power grid uses both. Each copper section should have at least 1 Y for grounding. Each iron section should have at least 1 delta for triplen harmonic and unbalanced load current suppression.

A grid w/ just 1 or the other is not as effective. If a grid is carefully laid out using both, system performance is optimum. Done right, only 3 wires are needed regardless of how unbalanced load current may become.

Claude

 

[jim hardy]

Actually both Y and delta have salient advantages. ... Each iron section should have at least 1 delta for triplen harmonic and unbalanced load current suppression.

Claude

Now THERE is a succinct summation.
Thank you Claude - that word picture paints a perfect physical picture in my alleged mind.
I'll remember that sentence easily because it's already formed a mental image..

old jim

 

[gerbi]

[..] delta does a great job at suppressing triplen harmonic magnetizing current [..]

Well.. it does not supress 3rd harmonic, delta connection creates a closed circuit so 3rd harmonic mag current can circulate.

Regarding transformers - delta requires some extra copper, so it's more expensive than Y. But all in all, it's used after all in different configurations (reasons as stated by guys above).

Rotating machines - regarding higher harmonic voltages in typical 3phase machines - it's better to use Y connection (3rd, 6, 9... n*3 harmonic are then canceled in line-line voltages as they are in phase). In delta connection the 3rd harmonic mag current circulates in closed winding loop (this creates extra looses in winding). Y connection usually requires less copper (less turns in series with bigger cross section, lower voltage -> better slot fill with copper).

Big generators are almost exclusively Y connected (working with insulated star point to reduce short circuit currents and it's effects on the machine).

 

[cabraham]

Well.. it does not supress 3rd harmonic, delta connection creates a closed circuit so 3rd harmonic mag current can circulate.

Maybe I should have worded my statement as follows. Delta connection suppresses the triplen harmonic flux distortion by providing a closed path for triplen current. However, "suppression" is actually a valid description. The mere fact that the delta provides a closed path does not suppress flux distortion. It involves Ampere's law.

Without a delta winding, if a 3-wire Y-Y xfmr was observed, the secondary line-to-neutral voltages would contain triplen harmonic distortion due to triplen flux. This triplen flux is due to absence of triplen harmonic magnetizing current. Under 4-wire connection (low Z neutral path provides easy path for triplen current), the flux is fundamental w/o harmonics, triplen or otherwise. Without the neutral path (open path), the equivalent circuit is normal operation with a negative triplen current set injected onto the waveform. This model is perfectly valid and gives the right answer.

The objective is to suppress this negative triplen current or triplen flux, either view works. If the secondary Y connection is modified to delta, the triplen currents circulate inside the closed path of the delta. How does this suppress triplen core flux?

When secondary triplen current exists, the secondary winding has a triplen mmf due to the triplen amp-turns. By the Law of Lenz the polarity of the triplen mmf is counter to that of the primary magnetizing mmf. So this delta triplen mmf tends to reduce the triplen flux component in the core.

With non-triplen secondary mmf, as soon as core flux gets reduced due to Lenz law, the primary mmf (amp-turns) increases to counter the counter mmf. The increased primary current due to secondary loading is "counter-counter-mmf". This happens because non-triplen secondary load current can be amp-turn balanced by primary mmf since non-triplen current can exist in a 3-wire Y.

But when the delta secondary carries triplen mmf, these amp-turns cannot be balanced in a 3 wire Y because the path is very high impedance for triplen frequencies. So when the secondary triplen mmf reduces the triplen core flux, there can be no primary counter-counter-mmf to cancel the triplen flux reduction. The secondary triplen current will generate a triplen counter flux which cancels the core triplen flux due to absence of triplen current in the primary.

So the delta secondary does indeed suppress "triplen core flux, triplen line-neutral voltage, and triplen negative magnetizing current". By providing a closed path for triplen current, the magnetic flux associated w/ triplen secondary mmf cancels the triplen core flux distortion. Also "suppress" is valid in another sense. If the primary Y has a 4th wire, a delta secondary will provide a low Z shunt path for triplen current such that nearly all triplen magnetizing current is in the delta secondary, instead of the primary Y neutral and hot wires. So a delta does indeed suppress triplen currents here.

Anyway, I just wanted to clarify, your point is well taken, the delta does indeed carry the triplen current, in doing so it suppresses triplen core flux, phase voltage, and triplen current is the Y primary if a neutral wire were present.

Claude

 

[psparky]

Man, you guys are smart. Reading your posts can be truly humbling.

I'll reply with some questions when I have some time to digest your thoughts.

 

[Windadct]

Hey PSP ... IMO we are, if anything, knowledgeable, the "smart" people figured all of this out to start with... but thanks, this is the strength of a forum.

 

[gerbi]

Cabraham, I agree.. and very well written, that's +1 from me :)

[..] delta does indeed carry the triplen current [..]

Indeed. For me, using word 'suppress' regarding 3rd harmonic mag current without clarification, was an abuse. It could be taken as: delta = no 3rd mag current in transformer, which is not true. That's why I replied. Good that's clear now.

Cheers

Edit
and here is a little something for all interested in Y connections of transformers
http://www.google.pl/url?sa=t&rct=j...=CUDlBlvk8fJNr2dS8YNiPw&bvm=bv.44770516,d.Yms