Two 3-Phase Unit Transformer in Parallel Operation

View "Parallel 2 Transformer" post 23/6/2011.

Essential requirement - Both have same Low Voltage phase shift relative to High Voltage side. The HV inputs must be essentially the same supply with very little phase difference.

Normally, the transformers have the same number of phases, turns ratio and regulation. Otherwise, load sharing is not ideal and/or currents circulate between the transformers, even with no load. To be exact, phase angle of the impedance causing regulation must be same for both. With transformers of more than one phase, there are a lot of possible winding configurations and the two must be compatible.

If the two transformers do not have the same rating, tolerances might cause overload of the weaker one. Even with nominally equal ratings, attention must be given to tolerances, to avoid overload of one unit at the combined full-load.

67model

Thanks for the advice, I research on the other references, and this is what I found.

for paralleling 2-3phase unit xformers the requirements are:

1. Same tap change

2. Same wdg. arrangement

3. Same vector group

4. same kva rating

The supply voltage should be identical. If this 2 xformers are designed on a secondary selective system or double ended substation and supplying a low voltage switchgear is it possible to parallel the 2 xformers by phased out by using phased out meter?

Another important consideration is the percent impedance of the transformers to be paralleled (%Z). they must be as close to equal as possible otherwise you will just be nuisance tripping on the overload protection of the lower %Z XFMR non-stop!

Correction! Two trransformers being paralleled need not have the same kVA rating. They can be in the ratio 1:3 also.

Further correction!! E65, Please provide references for your assertion that the ratings "...can be in the ratio of 1:3 also." I have not been able to verify that in any of my references.

Ref. your points :-

1. True if you have a tap change. But you do not have to have a tap change and I wrote [with care!] "the same ratio". A transformer could meet your four requirements and have a different ratio! Like 11/0.44 versus 11/0.4 kV - which would be trouble.
2. Probably true, in practical terms to get the maximum joint rating. If you put a core type in parallel with a shell type, this might be more difficult, but theoretically only the percent regulation [and its R/X ratio] has to be the same. M.G. Say's "Performance and Design of Alternating Current Machines" only lists same percentage impedance (without qualification by winding arrangement), voltage ratio, polarity, phase sequence and zero relative phase displacement. With transformers of 100s of kVA rating for power distribution, efficiency and X/R ratio are so high that matching X/R is insignificant as a requirement. Since you gave no statement of rating and X/R is much lower for small, less efficient transfos, I was careful to mention the angle of the impedance.
3. True.
4. Not true. In normal planning terms, having identical transfos is most economic. But I am sure that few utilities, with an overloaded transfo (awaiting repair of its twin) would flinch at putting a smaller portable emergency transfo in parallel - if it was available and could stop the overload safely. I am sure there are circumstances where permanently paralleling an available transfo, of smaller rating, is an economic and timescale solution to an increased load.
   

As I wrote, the supply voltage should be (ideally) identical in magnitude and phase. In practice, voltage and phase difference will cause circulating currents - but how much is acceptable is a matter of how much circulating current you are prepared to accept and that depends greatly on how much the load is compared to the transfo ratings and the cost of the extra losses.

Answering your final paragraph, I am not sure I understand the meaning of :-

• secondary selective system.
• double-ended substation.
• by phased out.
• phased out meter.

A system diagram would help a lot! And I would think it essential that the connections gave the same voltage, phase angle and sequence at the LV paralleling switch (commissioning tests of voltage, phase and phase sequence ensure this is really true) . It is always possible to parallel two systems, you just close a breaker between them, but the effect in damage and trouble may be bad.

There are systems where plant is normally tied together, but a fault could change this. Reconnecting the parts together may not be possible at any load because of limits such as protection tripping or system stability and a kind of check-sync relay [which recognizes that voltage and phase difference are acceptable and that the parts to be joined are still synchronized] is needed as an interlock.

According to Westinghouse T&D Reference Book Ch. 5, Section XVII, Topic 43, " (Single phase) Transformers having different kva ratings may operate in parallel, with load division such that each transformer carries its proportionate share of the total load. To achieve accurate load division, it is necessary that the transformers be wound with the same turns ratio, and that the percent impedance of all transformers be equal, when each percentage is expressed on the kva base of its respective transformer. It is also necessary that the ratio of resistance to reactance in all transformers be equal...For satisfactory operation the circulating current for any combinations of ratios and impedances probably should not exceed ten percent of the full-load rated current of the smaller unit."

Topic 44, "The same considerations apply for the parallel operation of symmetrical three-phase transformer banks as have been outlined for single-phase transformers. In addition it is necessary to make sure that polarity and phase-shift between high-voltage and low-voltage terminals are similar for the parallel units...When three-phase transformer banks having any considerable degree of dissymmetry among the three phases are to be analyzed it is necessary either to set up a complete three-phase equivalent circuit, or to interconnect equivalent sequence networks in manner to represent the unbalanced portion of the circuit according to the rules of symmetrical components."

Looks like some power engineering expertise is required to do this properly.

I have left out the drawings to respect the copyrights.