Strange Transformer Turns Ratio Test Results 5

[rcw retired EE]

It is a new 250/333/416/450 MVA, 23kV delta - 240 kV Wye Generator Step Up Transformer getting ready to energize.

All looks good except for (TTR) test(s) on one phase. H1: X1-X2.

Tap 1 is 1% low, Tap 2 OK, Tap 3 1% high, 4 1.5-2.5% high and 5 is 2.5% to 3.5% high. Other phases look OK. We have seven different tests counting the factory tests. Five of the results match. The other two indicate the unit is OK. (See chart)

Factory test shows good results: <0.5% on all phases. The manufacturer assembled and tested the unit at site. Their first on site TTR test got the bad readings listed above.

The vendor blamed the temporary power supply, came back with a different test set and got good TTR readings. (We don’t know which test sets were used.)

Not trusting the supplier, our startup crew ran their own test and got the bad readings. Two more tests with different test sets gave similar readings. (Multi-Amp TR800 & Meggar Three Phase 550503).

The vendor still said it's OK, and said we must be having interference issues with site power. (Power is from a 1500 kVA bank. Testing is done when the rest of the site is not working.)

I asked for an applied voltage test. They put 600V 3-phase on H1, H2, H3 and measured the X terminal voltages. Results were a little better but still off by 2.5%. A second test was done using 600/347 volt power connected to H0, H1, H2, H3. Same results: phase B & C good, phase A off by 2.5% on tap 5. Note that the 600V is only 0.25% of rated H1 voltage.

Winding resistance measurements and excitation current measurements all look reasonable. The results compare across the phases.

The vendor is still saying, no problem, go ahead and energize because their factory test and second set of tests show it is OK.

Are we just not doing the tests correctly? (The testing company has many years of HV testing experience. I have a few years of cranking smaller TTR test sets). I figured testing problems would show up on randomn phases with multiple tests.

What might cause the different test sets to give consistent "bad" readings on one phase and "good" on another, but still give one set of good readings with a different test set?

Is this a typical problem with large transformers? Is there something special we should be doing during the test to isolate outside influences? If so, why doesn’t it show up on Phase B & C? Nothing is energized near by.

Does anyone believe the vendor's claim that the voltages will balance out when we put the 240 kV on it?

It is my decision on whether to energize and see what happens. There is enough variation in the test data for me to suspect some testing issues. But the applied voltage test has me worried. It is too simple to goof up much.

Comments?

 

[waross]

How about energizing a phase with all the taps in the circuit and taking a voltage measurement from tap 1 to each tap in turn. Measure from #1 to #N, #1 to #2, #1 to #N, #1 to #3 #1 to #N, #1 to #4, #1 to #N, ...... #1 to #N,
If this test does not give equal indications, phase to phase, you may have a problem.
As a second check, I would try to determine the volts per turn. I would then check to see if a one or two turn error in tap connection would explain the bad readings.
Consider checking resistance from tap to tap.
I would also investigate the possibility of putting DC through the taps only and measuring the voltage at each tap.


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

 

[jghrist]

The error increases by 1-1/4% each tap position. I assume that the taps are 2-1/2%. I suspect that the error has something to do with the off-load taps.

 

[rcw retired EE]

Good points, gentlemen! I think the bad tap changer is the most likely problem. Right now the off load tap changer is under a lot of oil. Maybe I can get the vendor's reps to go inside the tank and do some checks.

 

[Elmir]

Be sure to rule out error by your test equipment by verifiying measurements with calibrates voltmeter, make sure to record results and compare to your automated ratio test set.

 

[prc]

I dont think it is due to instrumentation eror.If so how in two phases ratio is ok ?Did the connections to tap changer tampered during erection?If tap changer connection is doubted, check the single phase resistance of HV in each phase and compare.

 

[rcw retired EE]

We tested the transformer again, same results. But this technician noted that the nameplate shows two selector switches for each tap and theorized that one of the tap selectors is not moving and is stuck in tap 2. Tap 2 always tests good. (See nameplate)

That explains the constant increase in error on taps 1, 3,4,5 that Jghrist noted. But when I look at the tap changer layout, I can't come up with a connection that gives these results.

We start draining oil to look at it on Monday.

Another theory is that the core is magnetized and putting DC current through the winding will cure it. I'm skeptical.

 

[Elmir]

I agree with PRC and would perform winding resistance tests.

 

[Zogzog]

I think you found your problem, please update us after the tap changer is inspected.

Refrering to the vendors site power excuse, most TTR's are battery powered (Older are hand crank, I still keep one of those around) so site power shouldnt be an issue. Kudos to you for being suspect and getting a testing company to recheck (Were they NETA certified?).

The only other thing to consider would be the calibration of the test equipment, I dont think that is the issue this time but always good to keep in min. Is the test equipment calibrated yearly? Can the testing company provide calibration record upon request?

 

[rcw retired EE]

A transformer expert looked at the test data for us and did the folowing analysis:

"The technician states that the tapchanger has "two decks" therefore I assume the HV winding in effect has been split into two halves.
As we do not know how many turns on each winding I have assumed the LV has 100 Turns. Therefore, allowing half turns for calculation purposes:

LV turns Tap Calc HV Turns Total Calc ratio
1 633(316.5 + 316.5) 6.33
2 618(309 + 309) 6.18
100 3 602(301 + 301) 6.02
4 587(293.5 + 293.5) 5.87
5 572(286 + 286) 5.72

However if one half is stuck on tap 2 the expected ratio would then be

LV turns Tap Calc HV Turns Total Calc ratio
1 316.5 + 309 = 625.5 6.25
2 309 + 309 = 618 6.18
100 3 301 + 309 = 610 6.10
4 293.5 + 309 = 602.5 6.02
5 286 + 309 = 595 5.95

As has been measured."

These ratios match the "bad" test results.

His simple analysis provides a theory that matches the data.

When in doubt, go back to basics.

We are going into the unit this week. Thanks for the tips.

 

[rcw retired EE]

We finally got into this transformer. The A phase tap changer was stuck on tap 2. A loose bolt allowed the interconnecting shaft to slip between phase B & C.

I'm still concerned that the nameplate winding configuration does not match the actual windings. The nameplate shows two half windings, each with three 2.5% taps labeled a,b,c on one half and d,e,f on the other. Nameplate says the tap switch connects a-f, b-f, b-e, c-e and c-d for the five taps. Each step picks up another 2.5%.

But the two half windings actually have five taps each, with 1.25% between each tap. The switches connect tap 1 on the upper winding to tap 1 on the lower winding, tap 2-tap2 for tap 2, tap3-tap3 etc. 1.25% + 1.25% = the 2.5% tap.

The two representations are the same functionally, except when problems occur.

My question for the more experienced transformer EE's: Is it normal for the nameplate to show the functional winding configuration instead of the actual connections?

 

[rcw retired EE]

I should have said that the loose bolt allowed the interconnecting shaft to slip between the two tap switches on phase A. One switch was stuck, the other moved. That's why the error increased 1.25% in each step and was 0% on tap 2.

 

[edison123]

Excellent post and discussions.

 

[prc]

Rcwilson.Unfortunately it is the industry practice to show only functional winding configuration in name plate. I always thought and propagated to show the actual physical winding arrangement in the name plate,but not with much success.

 

[eraengineering]

Could you please advice me if this TR800 device can automatically detect vector group Dy1, Dy5, Dy11?

In case, we perform the turn-ratio test of other vector groups, please let me know whether we will have to key vector group in TR800 or not. And please also let me know whether the other vector groups will be shown on TR800’s screen once we finish the test?

What if we key the wrong vector group? Is there any alert message on TR800‘s screen?

Is there any guide how to use the TR800?

Thank you very much!