Differential

[a430]

We have a transformer 15MVA, 132/11KV, Yy0. Primary CT ratio = 100/1A, Secondary CT ratio = 1200/0.577A.
We connected a differential scheme using Siemens ICTs (type 4AM22 20-7AA, T40/A) connected in star/delta on both sides. The primary was connected in 1:1 ratio using the windings 10:10 and the secondary in the ratio 1:1.732 using the windings 26:15.
The transformer trips at about 15% loading with this scheme. We connected so many similar schemes, either transforming LV current up or transforming HV current down. The result is always the same.
A scheme with which the transformer was able to carry a little above 50% of its rated capacity is connected as follows: HV ICT Ratio = 6/12, LV ICT Ratio = 7/8
With another scheme, the current into the relay from the HV side was almost equal in magnitude to the one from the LV side. But the two adds up instead of canceling.
Note: The transformer is solidly grounded on the HV side, and grounded through a resistor on the LV side. And there is REF connected to the neutral of the HV CTs. We are using three (Duo-bias) Electromechanical relays Reyrolle type C21).
We intend to create a floating Delta with the remaining ICT windings when we have a stable scheme.
The interposing CTs we are using are Siemens (type 4AM22 20-7AA, T40/A). The transformer was initially stable when the secondary CT was of ratio 800/1. After a fire incidence we installed a new secondary panel with a CT that has ratio of 1200/0.577 (Alstom type VMXspec BS/1S). After installing the new panel we were able to achieve a stable scheme after some manipulations, but right now not even that scheme is stable.
Can you please help me by suggesting what may be wrong with our design?

 

[DanDel]

Did you adjust the relay taps for the change in CT? A 1200:0.577 ratio is not typical.
Typically, with electromechanical relays and a wye-wye transformer, the CTs should be connected in delta on both sides. I am not familiar with 'Reyrolle' relays; do they require the ICTs? What is their input?

 

[DiscoP]

There is probably nothing wrong with the design.

If I have interpretted your question correctly, the HV side ICT is connected in delta, and the LV ICT does the delta connection and CT ratio correction.

My guess would be that the scheme wasn't installed and commissioned correctly - there are plenty of oppurtunities to make a mistake with a protection scheme like this.

 

[Bung]

YOu need to connect the ICTs for the primary (132kV) side in D-D, with no ICTs on the 11kV side, ie star connected 100/1 + 1/1 D-D ICT on 132kV, and delta connected 1200/0.577 on 11kv side (effectively 1200/1 for phase currents out of the CT) [0.577 = 1/sqrt(3)]

Make sure the 30deg phase shifts are in the same direction for the CT Y-D-D connection on the HV side and the D on the LV side.

The Areva protection applications guide (see the Areva T&D website) explains it all very nicely.


Bung
Life is non-linear...

 

[mc5w]

For this type of transformer your secondary CT ratio needs to be 12 times the primary CT ratio. Since you do not somehow have that you need a ratio correction transformer. The odd ratio plus the correction transformer violates the Keep It Simple Stupid principle.

Just be sure that you have the right phase shift in your delta connected CT secondaries. Sounds a lot like you have the right ratio correction transformer but if hooked up with primary and secondary swapped would produce a 3:1 error. Same goes for having one CT secondary phase shift +30 degrees and the other -30 degrees.

What you might need to do is to set up a simulator to check your design, adjustments, and connections. For instance, if you have American 120Y208 volt power you could make a scale model transformer that steps down to 10Y17.3 volts and connect a set of CTs to that. You would then connect your protective relays to the CTs. To simulate an internal transformer fault you would connect some 120 volt heaters between the primary CTs and your scale model power transformer. This would help you verify correct design, connections, and correct ratio correction. Doing this with a 15 KVA benchtop transformer will be a lot less frustrating.

There should also be some shunts in your CT circuits both on the 15 MVA transformer and the scale model transformer. You could then connect a dual trace ocsilloscope to the CT citcuit to see if you have the right phase shifts.

I am also getting the impression that you are using a summing transformer to connect your CTs to your relay. Am I correct? On US power systems the transformer manufacturer usually is consulted as to what differential relay is the best to use.

There is also such a thing as a differential relay with restraint coil. This has a forward torque coil ( or equivalent software ) that responds to internal transformer faults and a reverse torque coil that responds to through currents. This gets around the shift in current phase due to reactive voltage drop when there is external load or an external fault. This way the relay can be made more sensitve to transformer internal faults. The relay should also have adjustments to help it ignore transformer magnetizing current, core losses, and inrush current.

 

[DiscoP]

Bung

I thought the only option was to use a Y-D ICT on the HV Side to allow for REF protection, leaving two options for the remaining connections.

1. 1:1 Y-D ICT on the HV side with a 1:1.732 Y-D connected ICT on the LV Side (this is what I think a430 is proposing)

2. 1:0.577 (eg 15:26) Y-D ICT on the HV side, with the LV side CT connected in delta.

This might be a dumb question, but I've tried drawing it up with a D-D ICT but can't seem to get it (eg I don't know where to connect the Star point of the 100/1 CTs - and also where to connect the REF CT).

Can you please give me some clues.

 

[Bung]

Is REF protection needed on a delta HV winding? There is no star point, so e/f protection applied to the delta is effectively restricted anyway. You just need a zero sequence path on the secondary side of the CTs (some more comment below on this).

The CTs only have to do a few things:
1] correct for the phase shift in the Dy transformer (usually + or - 30deg).
2] correct for the ratio of the main transformer (in this case 132/11 or 12/1).

If the constraints are that you have some 1200/0.577 CTs, 100/1 CTs and some ICTs that can be connected in any ratio you like (within reason!), then you just have to arrange those connections to suit what you want to measure. The grounding of the CT secondaries (provided you've taken care of all the zero sequence traps) is just for safety purposes.

I suggested a D-D ICT connection just to keep the phase and ratio corrections at unity. I didn't consider a REF requirement. I have never applied a REF in that situation, so it is probably just as easy for you to work out as for me! But it does suggest that a YNYN ICT connection might be better if the REF is to work (but I haven't tried working it out). I will go away and try to draw it up, but in the meantime maybe somebody else will have the answer?


Bung
Life is non-linear...

 

[Bung]

AAH-HA the light dawns: the power transformer is Yy0, not Dyn! Maybe I should read things and not jump to conclusions! I've never come across anything other than Dyn or YNd transformer at 132/11 voltages, so I just mentally jumped to a conclusion. Sorry about that.


Bung
Life is non-linear...

 

[a430]

I will like to thank everybody for the solutions and clues,what i discovered was that, there was a polarity reversal on secondary cts.When corrected, the Transformer was able to carry full load and is in service since.