Change in Transformer Inrush

[mgtrp]

We are having issues with one of our step-up transformers tripping on inrush almost every time we energise it.

The transformer is 3/5 MVA and is a star-connected autotransformer with delta tertiary. It is a step-up between a four generators and the grid, and has been in-service and trouble-free for almost 10 years.

We are able to energise it from the grid without issue, but when we try to energise from the generators, it almost always trips immediately due to "switch-on-to-fault" instantaneous overcurrent. I am confident that the relay setting is not a problem because it has not changed, and we are also seeing the high currents on the generator protection relays (pick up of overcurrent elements on all four units, and one has seen sufficiently high current to trip).

I presume that an increase in transformer inrush would be caused by the core/windings shifting, or alternately by the transformer becoming magnetised. We will look to have magnetising current measured at some stage, although I'm not sure that we'd have past test results to compare against on this unit.

Any other suggestions of possible causes, and methods to diagnose/repair?

Thanks!

 

[PHovnanian]

Just thinking out loud. What would the effect on inrush be if the delta tertiary were opened up (and nobody noticed)?

 

[dpc]

Do you have fault current data and waveform capture from the relay that is tripping? That is the first thing to look at.



David Castor

http://www.cvoes.com

 

[DTR2011]

how about energizing the transformer from a single generator. slowly (and manually) increase the voltage on the generator, up to nominal.

 

[mgtrp]

The event report is attached, showing transformer saturation during inrush followed by a trip. To avoid confusion, the large "VN" is actually the poorly-labelled synchronising voltage.

I'm sure that if we were to slowly energise the transformer, we would be fine. But that is not a long-term solution, especially as there are other loads off the generator bus that normally should be in service. We are able to back-energise from the grid, and then synchronise across the transformer circuit breaker, but that doesn't help us when we have a grid outage, which is generally when we would be using the local generation. The town in question is supplied by a single transmission line which is affected relatively often by lightning.

 

[DTR2011]

if it is a case of residual magnetism, sometimes the slow ramping up/down of the voltage can cure this.

 

[odlanor]

we have seen this problem with four generators connected to a transformer, while the basic design.
The energization of the first generator in the transformer may cause an inrush current so high that could damage the generator.
The solution adopted was the same proposed by Smallgreek. In our case the slope of the automatic voltage regulator, has changed during starting.

 

[mgtrp]

Thanks.

I guess that we'd measure "as-found" excitation current, try ramping up slowly the voltage from ~0 to 4.16 kV +10%, then slowly ramp back down to zero. After demagnetising, measure the excitation current again and hope that it's decreased.

Does this sound reasonable?

Providing a soft-start on a permanent basis is not feasible, given the other loads connected to the same bus.

 

[MAGTiger]

Can you determine how stable the generator output frequency is druing the switching event? If it slows down too much, the xfmr could saturate until the freq recovers.

Neil

 

[VTer]

How big are your generators? With a weak system, the magnitude of inrush alone might not be the issue, but the time it takes to magnetize the transformer might be much longer compared to your utility, therefor tripping your protection. Just a thought.

"Throughout space there is energy. Is this energy static or kinetic! If static our hopes are in vain; if kinetic — and this we know it is, for certain — then it is a mere question of time when men will succeed in attaching their machinery to the very wheelwork of nature". – Nikola Tesla

 

[electricpete]

I agree with David, it seems like more info would be helpful.

We are able to energise it from the grid without issue, but when we try to energise from the generators, it almost always trips immediately due to "switch-on-to-fault" instantaneous overcurrent.

Is this associated with all phases or one phase? Same phase each time? Does more than one relay trip? If single relay, could be relay malfunction or problem with low-side CT’s (including residual magnetism, shorted turns).

methods to diagnose/repair?

Looking at the transformer itself, transformer single phase excitation test should shed some light if there is a residually magnetized core. Results are often compared to previous results, but you can also do a quick check that you have two high similar readings and one lower reading.

There are a wide variety of tests geared toward detecting transformer movement (leakage reactance, FRA), but to my understanding they rely heavily on detecting changes, so not very useful without a baseline.

=====================================
(2B)+(2B)' ?

 

[waross]

Is my understanding in error here?
I understood that residual magnetism was more or less limited by the hysteresis curve and was much much less than the normal working flux. Further to this, the flux density during normal operation of the transformer is more than enough to overcome and remove any residual magnetism. I understood that any residual magnetism was a result of the de-energization and did not persist from one operation to the next.
I understood that energization surges were dependent on not only the amount of residual magnetism but also on the point on the wave that the switch closes. The result being that even with a high residual magnetism, 50% of the time this would tend to reduce the inrush rather than increase it.
Have I been wrong? It has happened before.

My understanding notwithstanding, I seem to remember some discussions regarding a circuit that would de-magnetize a transformer so as to avoid residual flux from pushing a transformer into saturation on energization. Does anyone remember those threads?

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

 

[dpc]

Are you sure nothing has changed in the CT wiring or relay settings? I see what appears to be one of the phases saturating, possibly due to dc offset.

A loose CT connection can create a high resistance causing saturation. I've seen some poor crimp connections come back to cause problems after a few years of oxidation and thermal cycling.

I would be more suspicious of the CTs, CT secondary wiring, and relay.

Is it always the same phases that are high?

 

[mgtrp]

Ok, I'll try to address most of the replies:
1. Generator frequency is "probably" stable during inrush. Certainly, we are not picking up our underfrequency protections, set at 58 Hz. Furthermore, I doubt whether the frequency could drop much in 4-5 cycles, and pre-energisation frequency is generally fairly close to 60 Hz.
2. While we do have a relatively weak system, the generator total size is approx 4.3 MW @ 0.8 pf, in comparison to 3/5 MVA transformer size. Furthermore, this site has been in service for almost 10 years under the same configuration without issue until recently.
3. I don't have waveform captures from many energisations (none others available to share at this present time, although I have looked at them and they are similar), but from the relay event log file and other wave forms that I have seen, the phase with highest current is not consistent.
4. We are taking an outage on the unit later this week or early next week, and will be measuring excitation current @ 20% rated kV, and hopefully short-circuit current @ 600V. I also have the factory test report and some field test data from the middle of last year (excitation current only).
5. My understanding of residual magnetism in power transformers is not as good as I'd like! Certainly in instrument transformers (e.g., CT's), magnetising a core is a possibility. I had previously located a thread about demagnetising a transformer:

http://www.eng-tips.com/viewthread.cfm?qid=91058&page=403.

We do not trip every time we energise, although I'd certainly say more than 50%. When I was onsite, we managed 1 successful close and 2 trips. During previous outages we managed 1 successful close and >10 unsuccessful closes, although this statistic is blurred because the operator tried both protection settings groups, one of which (intentionally) results in trip-free operation of the circuit breaker.
6. I'm fairly confident that it's not a CT or relaying issue because A/ the settings in question have remained consistent for at least 3 years (I have past test reports etc), and B/ the upstream generator protection is also picking up the high currents and, in one case, operated. The generator protection oscillography also shows high currents and saturation (I can upload this oscillograph if it's useful, although it is in Beckwith *.osc format rather than COMTRADE). It is not always the same phase.

A further development:
Site staff have reported today that the transformer is now "growling" - making a lot more noise than previously.

As mentioned above, we will be measuring excitation current and short circuit impedance (both at reduced voltage), are there any other checks that might be useful?

 

[veritas]

mgtrp

You mentioned that things have been going okay for years. Has anything changed that you know of that could give rise to these trips?

Has the transformer recently been tested? Meggar, gas analysis, etc.? I'd be wary of a "growling" transformer!

 

[davidbeach]

Growling is a sure sign of a saturated transformer. Solve the growling problem and most likely your other problem goes away as well.

 

[dpc]

Growling is a bad sign for the transformer. I may have to change my diagnosis - like my doctor always does.

dpc

 

[mgtrp]

Saturated transformer causing both the growling and heavy inrush is my working theory. As to what's causing it, that's a tougher question to answer! I'm not aware of any issues with either low frequency or high voltage at this site, and there are numerous independent sources from which to check both.

 

[electricpete]

if transformer is moving to the top of your suspect list, definitely go for transformer oil sample as mentioned by veritas. It can be taken even with transformer energized, but only by qualified person.

Also when transformer is denergized,there is a standard battery of electrical tests performed with a Doble tester or similar device. This includes transformer turns ratio, CH, CL, CHL, power factors, single phase excitation currents, etc.

Less standard denergized tests to look at (as mentioned above) include leakage reactance and FRA, depending on what test capabilities are available.

=====================================
(2B)+(2B)' ?

 

[PHovnanian]

mgtrp

Transformer saturation under normal voltage/frequency conditions could be the result of the core's magnetic path opening up. Perhaps this unit needs to be untanked and hardware tightened up.