Transformer Inrush - Limited by short circuit availability? 1

[vandal06]

Hi all,

I'm aware that transformer inrush is a phenomenon caused by energizing the transformer at a non-zero crossing of the voltage waveform that causes the remanent flux to interact with the energizing flux such that you get large increases in magnetizing current, often leading to terminal currents around 12xFLA. Is this limited at all by the short circuit availability of the source energizing the transformer? I thought perhaps the saturation of the transformer core reduced the impedance of the transformer so much that it was primarily winding resistance that limited the current flow. I also know that there are all sorts of off-nominal harmonic currents such as 2nd and 5th involved that I have not been able to reconcile with this possible interpretation.

My specific scenario is I'm doing a protection study on a system that involves a 115/34.5 kV two-winding, delta/grounded-wye transformer protected by a numerical differential relay. A instantaneous overcurrent element has been set in the high-voltage winding protection to back up the differential for winding faults. Typical setting philosophy is to set it above 12xFLA (restrain in inrush) or 1.25% of the secondary bus fault current, whichever is higher. This ensure that 1) the relay does not operate on inrush and 2) only operates for transformer faults. This relay has been set according to that principle, but when running my short circuit studies it is apparent that the system cannot supply enough short circuit current for the element to ever operate when trying to set above 12xFLA.

My first thought was to reduce the pickup to operate for transformer faults (while avoiding secondary faults) and time delay it 0.1 to 0.2 seconds to ride through the inrush. However, if the inrush current is dependent on the short circuit availability of the system this is a moot point. I have been perusing the forum topics for the last hour and have not seen this explicitly addressed amongst the other transformer inrush topics, so here we go.

 

[healyx]

Hi Vandal,

My feeling is that inrush would be limited by upstream short circuit capacity, but should you ever have a system where the upstream infrastructure can't provide 12xFLA to downstream equipment?

 

[waross]

The Available Short Circuit Current is the steady state current into a short circuit. It is calculated by dividing the rated current by the P.U. impedance. The available short circuit value is used to select switch gear which is rated in available current. The first cycle fault current depends on the point on the wave form that the fault occurred, the X:R ratio of the supply transformer, and for your application, the impedance and the X:R ratio of the transformer being energized.
If you are using an "Available Short Circuit Current" value be aware that the first cycle peak current may be over twice the calculated (or measured) available short circuit value. We have some gurus here that can do the math, but I would not be surprised to see a peak inrush in excess of the calculated available short circuit current.
Note: when switch gear is tested and rated, the discrepancy between actual peak fault current and available short circuit current is taken into account. The initial peak fault current decays to the available short circuit current.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[mgtrp]

A few comments - firstly, modern numerical relays tend to be very good at filtering out the DC offset, harmonics, etc, and are thus not as susceptible to inrush currents as older relays and fuses. Secondly, in large power transformers (which a 115-34.5 kV unit probably is), inrush currents tend to be of lower magnitude but longer duration than smaller transformers. Thus you might find that you are being overly conservative in assuming 12xFLA.

For what it's worth, I generally set transformer instantaneous elements at 130% of the fault current on the secondary terminals, and have never had an issue. Admittedly, though, the short circuit strength of our system is relatively low, and as you noted, weaker systems will - generally - not to provide as high inrush.

 

[dpc]

thread238-238946

This has been discussed a few times before. Weaker systems will result in lower transformer inrush. There was an old S&C paper that discussed this, but the link in the old thread seems to be broken.

 

[stevenal]

Inrush is not a single point, it is a curve. I use the Cooper suggested points of 25X at .01s, 12X at .1, 6X at 1, 3X at 10, 2X at 100. If system impedance cuts off operation at the higher points, the others remain valid. A definite time element at 12 or 6X might might work, but why not use a time over-current curve fit between the inrush and damage curves? Through faults would be be captured too, but with transformer impedance in the mix, coordination with downstream devices is usually possible.

 

[vandal06]

Healy, the system I'm studying has less fault current available than the 12xFLA at 0.1s for inrush. The transformer is rated at 25MVA, which gives a 12xFLA approximation of 1500A. The available fault current from the system is only 600A.

waross, so with that in mind I'm looking at a 1st cycle peak current of 1200A, which is still less than the 1500A for 0.1 seconds that can be assumed for a strong system.

mgtrp, agreed on the points about harmonic filtering in relays. This particular digital relay's overcurrent elements only operate on the fundamental component. In the SEL-387 manual (different relay from what I'm dealing with, but similar operation) they even state that you can ignore the inrush current directly and just set based on load and fault conditions. That didn't sit right with me as all the older guys here have never ever done that with a modern relay.

Dpc, somehow I missed that post. If I had found that topic this thread would not needed to be posted.

stevenal, I was aware that inrush is a curve but I only typically look at the 12x at 0.1 point and ensure my overcurrent elements (be they instantaneous, definite time, or inverse time) do not operate at that point. As to your second point my intention is to also use a 51P to back up for system through faults that would fit between the inrush and damage curves and coordinate with downstream devices.

I appreciate everyone's contributions, it looks like my suspicion was correct!

 

[pie314]

Vanda106
The short circuit impedance has a definite affect on the peak inrush current. High inrush currents occur when the core saturates during the energization. Once the core saturates the current is the voltage divided by the impedance ohms of the transformer.The point at which the core saturates is a function of the residual magnetizim of the core,The crossing of the wave form voltage, and the rated flux density of the core

 

[pie314]

I of course meant the AIR CORE impedance of the primary winding.