Many posts have discussed about transformers in parallel, however in one of them many people commented: "There is not circulating current if both primary transformers are connected at the same bus".

That answer is wrong and is a myth to think that.

Let me share you the following:

Two three-phase 10 MVA 66,000Δ-12,470Y volt transformers were in parallel operation in a substation. The primaries of the two transformers are connected to a 66 kV transmission line through a single air break switch.

This switch is designed to interrupt magnetizing current only, which is less than 1 A. The transformers were being removed from service and the secondary loads had been removed. A switchman then started to open the air break switch, expecting to see a small arc as the magnetizing current was interrupted.

Instead, there was a loud "bang" and there was a ball of flame where the air break switch contacts had vaporized. Something was obviously wrong.

What happened?

The answer is: Circulating current.

Closer inspection, revealed that the two transformers had been set on widely different taps: The first transformer was on the 62,700 V primary tap and the second transformer was on the 69,300 V primary tap. Both transformers had a 7% impedance. Because the turns ratios were unequal, a 6 ampere per phase circulating current was set up even without any secondary load.

I need to say that this story is very instructive, but, unfortunately, with terrifying consequences. This story needs to be narrated because it is a big warning to the rest of us.

If the primaries of the 2 transformers are both connected on the same bus and their secondaries are independent of each other (not connected to a common bus) then I can say that the 2 transformers are NOT in parallel.

If the primaries of the 2 transformers are both connected on the same primary bus and their secondaries are connected on the same common secondary bus, I can say that the 2 transformers are in parallel.

If 2 transformers are not in parallel. I cannot see a reason to have a circulating between the 2 transformers. There can only be a circulating current in any one of the 2 transformers if the winding taps on any phase of that transformer is not the same with the the other 2 phases of that same transformer. Thn there will be circulating current within that delta winding.

On the air break switch, if the switch is not an "interrupter" type , at 69 KV it will not be safe to operate it to de-energize the transformer. Safe practice is to de-energize the transformer by tripping the primary circuit breaker and then opening the air break switch to isolate the transformer from the primary bus. Even if the magnetizing current is only 1 amp, at 69 KV if corresponds to 1 X 1.732 X 69 KV = 120 KVA.

Both primaries (Δ) were connected to the same 66KV bus and both secondaries (Y) were connected to the same 12.47 KV bus, that is obvious, otherwise they are not in parallel.

The problem in the other posts was just that. No one analyze in detail but all give opinions without basis, that is the reason why i call this post the Myth.

Is wrong to say "There can only be a circulating current in any one of the 2 transformers if the winding taps on any phase of that transformer is not the same with the the other 2 phases of that same transformer..."

The circulating current flows on both transformers not only in one of them, the circulating current flowing in the secondary obviously also flow in the primary modified just by the tap ratio of each transformer.

The transformers have not primary circuit brakers.

why? I don not know, is a very old installation.

If you take your time to analyze in detail you will find that secondary circulating current is 331 amp. (without any secondary load connected) and this reflects to primary as 65.8 amp in one transformer and 59.6 amp in the other one.

So, the net current through the air break switch is the difference in the primary currents:
65.8-59.6 = 6.2 amp. ==> 6 amp for practical purposes.

6 amp at 66 Kv is enough to produce a big damage if equipment is not designed to handle.

In conclusion, in parallel transformers if tap ratio have not the same setting there will be circulating current.

Additional causes for circulating currents between paralleled transformers are from automatic load tap changers operating (slowly) to adjust the secondary voltages. The LTCs also affect the impedance of transformers. However the secondary voltage differences can be more worse and longer duration without the LTCs if the primary feeders are supplied from different upstream substation buses or entirely different substations. In these cases the settings for directional relay elements to protect paralleled transformers from through faults caused by faults upstream of just one of them also need to be set to accommodate the lower magnitude circulating currents when they are unavoidable by system design.

You right Walter;

Also when by any reason is it necessary to put in parallel two transformers with different impedance.

What happen in this case?

Ir order to avoid overload one transformer is it necessary to select the appropiate tap ratio in each transformer, doing this we can have a good load sharing, however the appropiate tap ratios never are the same because transformer impedances are different.

So we back to the inicial condition: Different tap ratios will produce circulating currents.

I think that connecting in parallel any electrical equipment is a dangerous matter and it needs very detailed analysis.

The impedance can be different but as long as the impedance are close at each transformers rating then they should take their own appropriate (fiar) share of load. For example if a 50MVA unit is 6% on 50 MVA (12% on 100 MVA) and if you were to install a 75 MVA in parallel and specified it as 6% on 75 MVA (8% on 100 MVA) then each unit should take its "fair" share of the load. This can be helpful if you want to partially replace your fleet with larger sizes or if adding a 2nd,3rd or more transformers at a site and don't want to get a new unit of of the smaller size. Of course you need to specify the new unit with the same DETC/LTC taps or study what you need and can operate with.

Differences in impedances causes differences in VA flows (but this should not be considered circulating flows). Difference in tap ratio creates circulating VAR flows.

First, let me say that like many advanced electric power engineering topics, understanding the finer points about circulating currents is a very important yet very poorly understood subject, except for the really easy cases.

To your point, it depends upon whether these are NLTCs or OLTCs. If NLTCs then only operating procedures that are rigorously followed will work; i.e., the operators must be instructed that the taps are on the same settings, if one is moved they're both to be moved. If the transformers are not identical then someone who understands these things has to create a procedure that specifies what combinations of taps are allowed and which can't be used.

If they're OLTCs then the controls have to be set for Master/Slave so that the Master follows the voltage and the slave is always set to the same tap as the master. This can be accomplished in the control circuitry, and many modern controllers already have this built-in.

Enrique, your warnings are well advised, but your story seems a bit out of line with reality. If the transformers were set with their taps at opposite ends (10% apart), the circulating currents would load the transformers to nearly 70% of their rating without any load being drawn.

That means that each 10MVA transformer could only carry 3MVA additional load! It seems unlikely that the substation would have endured this operating condition for very long unless it was very lightly loaded. But your point about paralleling without investigating all the system impacts should be heeded by all.

What constitutes paralleling is the real question. With two transformers there are four different connection scenarios, but only the case as described by Enrique constitutes intentional paralleling and is relatively easy to analyze. It is possible to set up unintentional paralleling upstream or down stream from the obvious cases, especially during system switching or emergency operations, and these can only be detected by rigorous system analysis.