Dist. Transformer - Short-circuit on Stabilizing winding 1

[TY18]

Hi All,

I am facing one critical matter where my energised (Vector group:ynYNd01) oil distribution transformer is down and basically the consultant and contractor disagreed with factory's view.
a) Factory made this transformer according to consultant spec, it was tested twice before energised.

b) Contractor used bear copper tape to earth it with our provided point at the bottom of the transformer and they also used another copper bear copper tape to earth the stabilizing winding point from the top of transformer (with both copper tape jointly touched on bottom and grounded. But the top stabilizing earth bar was touched with the side body of the transformer. The burn mark was showing at the point from stabilizing bar touched the body of txf.

c) We open up and lift the winding and saw some burn point from the bottom of the stabilizing winding which also burn the bottom structure steel of the transformer. Is already out of shape and touched the structure steel.

d) Factory engineer explained that it was the wrong earthing mathod that stabilizing shall use insulated copper to earth and shall not touch the body earth bar. It's already touched the body earth bar and stabilizing non-insulated touched the transformer top body created big flow of current force to cause the internal stabilizing winding to curve and touch the internal winding structual steel to spark....

e) Contractor forgot to install our provided Oil Thermometer (with contact), which is totally weird... I am totally shocked! and we dont know is this overheated.

Now, the consultant disagreed our view. Please help to look into this and advice us are we wrong in the factory explaination?

P/S: the transformer is specially design for solor system to sell back electrical power back to the power authority.
It faced a major trip in the northern city during one same day that all the other solar system trip at the sametime. Does it matter? What to check?

 

[waross]

Do you have a four wire connection on the energized wye winding? If so, the stabilizing winding may be trying to stabilize the grid. I have seen quite a few issues with four wire Wye/Delta transformers and transformer banks. When a four wire wye/Delta tries to stabilize the grid a heavy current may flow in the delta and in one or more of the wye windings, regardless of the grounding of the delta. If the heat developed causes the delta winding to fault to ground, a ground loop will be formed with even heavier currents.
One ground on the delta winding should not allow ground currents to flow.

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

 

[FPelec]

I agree with waross; in order to better understand what happened, can you give us some additional information: which are the voltage levels and which is the neutral grounding strategy on LV and HV sides of the transformer?
let me add that IMHO, if not strictly necessary, stabilizing winding should be avoided. For distribution transformers, in most cases, the 3-limb construction can allow for a sufficiently low zero sequence impedance also without a delta winding.

Si duri puer ingeni videtur,
preconem facias vel architectum.

 

[prc]

It seems quite strange and normally unnecessary to provide a stabilizing winding in a distribution transformer.

What happened to your transformer: when the line tripping happened (must be a LG fault as for 3L fault there will be no flow of current in stabilizing winding) heavy circulating current flowed in the closed delta stabilizing winding. The winding could not withstand the forces from the fault current and must have buckled inward or axially bent or collapsed, resulting in inter turn insulation damage and shorting. When it touched the grounded steel it became an additional 2L fault on tertiary winding.In a distribution transformer it is quite difficult,of course not impossible,to design/make a strong stabilizing winding due to small size. So poor quality transformer and you deserve a free replacement/repair.

Factory engineer's explanation: please ask him to show you by a sketch how current can flow in the delta winding when one corner of it touches the tank, which in turn is grounded. It is true that when tertiary is earthed usually touching the tank surface is avoided. This is to avoid arcing or slight current flow in tank due to un equal voltage between top and bottom of grounding bar. It is a normal practice with some utilities to solidly earth one corner of the stabilizing winding to tank top externally or internally with out bringing out on a bushing.

Oil Temperature Indicator: Don't worry about your contractor missing to fit the oil temperature indicator. It would not have showed any thing. Winding failed before any overheating due to mechanical forces. Remember oil time constant of transformers is couple of hours and no fault current will be sustained so long as to show a jump in oil temperature indicator to trip the unit out of grid.

Contractors' omissions : are not new - let me give a recorded classic case. GE supplied a small power transformer in US -time 1920's. Those days due to transport limitations even small transformers were sent with out oil filling, in disassembled condition, and oil was supplied separately in drums. Contractor forgot to fill oil in the transformer and the unit was energized. Thanks to the large design margins used those days and due to the light load on transformer, nothing happened for a couple of years. But when connected load increased, one day fumes came out of transformer and when they opened tank,station engineer was shocked to find no oil inside the tank and transformer instruction manual was hanging from a bus bar. Finally oil drums were located in the store for final filling in the tank.

 

[TY18]

Hi PRC,
Thanks for your suggestion. May I clarify what is "LG" Fault? 3L & 2L that you mentioned is ...?

 

[prc]

LG- one phase line to ground 3L- three phases shorting 2L -two phases getting shorted

 

[Neellal]

Hi Waross, your point is interesting..but could you please elaborate why for four wire system it happens.is it possible to avoid it with just a simple Dy configuration eleminating the stabilzing winding completely and feeding grid by delta 3 wire...

 

[waross]

Hello Neellal;
My internet has been out for about 5 days. I tried to answer you on my phone, but no joy. I will try to provide an explanation later this evening.

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

 

[krisys]

Please note that the component of current circulating through the delta loop (also called as “buried delta”, because it is not connected to the load/supply) is only the triplen harmonic components, predominantly third harmonic components.

One and only one of the terminals of the buried delta winding of the transformer (also called as “corner delta grounding”) shall be connected to the ground. As the buried delta winding is floating, in the absence of corner delta grounding, a continuous fault current cannot flow for the accidental touching of delta winding to the tank. However, in the floating system, an arcing ground phenomena could happen and would result in transient surge voltages. Eventually it would cause a multiple insulation breakdowns.

As a result of multiple insulation breakdown, two points of the buried delta winding would come into contact with ground. This would result in a real short circuit current, thus aggravating the damages and spreading the damages.

It is a complex phenomena and the problem might have developed slowly over the period of time. Hope my explanation/argument makes sense and the fraternity understands.

 

[waross]

Sorry! Hands on experience does not support this statement:

Please note that the component of current circulating through the delta loop (also called as “buried delta”, because it is not connected to the load/supply) is only the triplen harmonic components, predominantly third harmonic components.

Note: This explanation applies to four wire connected wye/delta transformers and banks.
In a perfect world the voltages impressed on the primary windings of a transformer or transformer bank with a four wire wye connection are equal in magnitude and the phase displacement angles are equal.

However life is not perfect when long distribution circuits are subject to uneven loading, particularly single phase loading.
When the vectors of a delta connected secondary are plotted, the result should be a closed, equilateral triangle.
However when uneven loading causes a voltage drop on one phase, the resulting secondary vector will be slightly shorter and the triangle will not be closed.
Uneven loading implies a neutral current and a voltage drop on the neutral conductor also. The voltage drop on the neutral causes a displacement of the neutral and a resulting phase angle error. The phase angle error causes a further error in the secondary vector plot.
The four wire wye/delta transformer (or bank) will act to correct these errors.
The resulting gap in the unclosed triangle represents a voltage which causes the circulating current in the delta winding.
This circulating current is limited by an impedance that is three times the rated impedance of the transformer.
A quick and easy estimate of the circulating current may be derived by comparing the percentage voltage error with three times the rated transformer (bank) impedance.
Eg: rated impedance = 2%, 3 x 2% = 6%. In this example, if one phase voltage is 6% low, 100% of rated current will circulate in the delta.
This statement assumes that the delta is rated for 100% current. If the delta winding is a tertiary winding then the impedance value of the delta winding shall be used in the calculation. In the real world, the phase angle error will cause the actual current to be slightly higher so this calculation should be taken as a rule of thumb rather than a rigorous solution.
But the phase to neutral voltages are equal and we still have high currents!
Voltage drops, both phase to phase drops and single phase drops are common on long distribution circuits. This condition is often corrected by a bank of three wye connected single phase voltage regulators. While this corrects the phase to neutral voltages, it tends to make the phase to phase voltage errors and the phase angle errors worse.
In the real world, the source impedance far out on a distribution circuit tends to be relatively high. This may act with the transformer impedance to reduce the circulating current. The rule of thumb should be used with the terminal voltages of the energized transformer, not the open circuit voltages.
What happens when a phase is lost? (This refers to a phase open feeding the distribution circuit, not an open phase at the transformer) The source impedance of that phase becomes the impedance of the load connected to that phase.
At this point the relationship between the transformer rated impedance and the source or line impedance becomes important.
At one extreme, with a small transformer on a relatively high capacity, heavily loaded line the source may be considered as infinite and the voltage drop of a missing phase will be 100%. The current circulating in the delta may then be estimated as 100% / (percent impedance x 3)
In our example of a transformer rated at 2% impedance the current when a phase is lost may be estimated as 100%/(2%x3) = 1667% of full load current.
At the other extreme is a transformer bank that is the major load on the distribution circuit. The transformer may back feed the missing phase with no ill effects to the transformer.
Perhaps the worst case is when the loading on a missing primary phase causes a current that is slightly less than the rating of the primary fuses. Primary fuse protection is often in the range of 250% to 300% of rated current. If the circuit characteristics are such that the transformer current with a missing phase is in the order of 280%, then 300% primary fuses may never clear the circuit.
All this explains why the three wire wye/delta connection is not recommended.
By the way; I spent a few years as system engineer of a small system. I inherited a number of problematic four wire wye/delta installations. I eventually had them changed to wye/wye connections and the problems went away. A large neighboring utility never did change out their wye/delta connections. It was common to see one fused cutout hanging open on their transformer banks. The banks operated as open delta bank with one transformer idling. The circulating current is just enough to magnetize the idling transformer from the delta winding.

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

 

[prc]

Failure of any complex equipment in service is like a murder.Only super detectives will be able to locate the real killer through evidence from site and by keen investigations. Capable lawyers by dramatic arguments may derail the whole case acquitting the real killer.
The failure mode arrived at by krisys is interesting one.This can be verified through the evidence at site. 1) Any arc mark on tank can be due to also- the winding failed and if it is to touch the structural there must have been severe fault current to distort winding.When the winding touched structural, it became a 2 L fault as one corner of delta was already touching the tank through earthing bar. That current flow through tank can result in burnt marks.2) How arc can occur? The current for that can be fed only through the winding capacitance to earth of other two phases.Will there be sufficient energy in such an arc? 3)What type of over voltage can arise from such arcing? Did anyone noticed such arcing earlier to failure? If it was touching properly with tank, why arcing should occur?4) The failure marks in winding shows an insulation failure or mechanical failure from dynamic forces ? 5) The case diary tells that there was a line tripping just before transformer failure.The delta circulating current in such a situation is extremely high,some times more than a 3L fault at delta terminals. May be TV 18 can throw more light and other experts may guide us.

 

[sibeen]

When the vectors of a delta connected secondary are plotted, the result should be a closed, equilateral triangle.However when uneven loading causes a voltage drop on one phase, the resulting secondary vector will be slightly shorter and the triangle will not be closed.

I know I'm fairly thick, but I can't get my head around this.

If the delta secondary is closed, than how does the resultant vectors, when the primary is unbalanced, end up with an open section? I can understand that the vectors won't be the same length etc, but no matter how I draw them they still end up with a closed loop.

 

[waross]

Draw the conventional delta vector diagram. This will be an equilateral triangle. Now assume a 10% voltage drop on one of the phases. Take your eraser and shorten one vector by 10%. Now you will have a gap of 10%. Now one of the vectors will be 10% short and the triangle will not close.
or, a little more accurate:
Construct an equilateral triangle, A,B,C with sides of 10 units to represent the actual transformer connection.
Now construct on top of this triangle another triangle to represent the voltages.
A'=10 units
B'=10 units
C'=8 units
Draw A' over A.
Construct vectors B' and C', so as to meet at point B'C'
You will see the error between point BC and point B'C' This illustrates both the magnitude error and the much smaller phase angle error
That 10% gap will drive the circulating current around the delta winding. In the event of a primary line to ground fault the circulating current may be as much as 1/3 of the available short circuit current. That is enough to do similar damage to the transformer as an actual short circuit.
I understand that many of the original distribution schemes in North America were delta/delta. When industrial expansion took off after WWII a quick and cheap way to gain a 73% increase in distribution circuit capacity was to run a neutral conductor, raise the voltage from 2400 Volts to 4160 Volts and change the distribution transformer primaries from line-line to line-neutral connections.
Browsing very old text books and handbooks from this era I have several times come across the advice:
"When wye/delta transformer connections are used, a primary line to ground fault may result in widespread primary fuse blowing throughout the system."
In support of both krisys and prc,

Failure of any complex equipment in service is like a murder.Only super detectives will be able to locate the real killer through evidence from site and by keen investigations. Capable lawyers by dramatic arguments may derail the whole case acquitting the real killer.
and
5) The case diary tells that there was a line tripping just before transformer failure.The delta circulating current in such a situation is extremely high,some times more than a 3L fault at delta terminals.

This may have been the first damage resulting in:

As a result of multiple insulation breakdown, two points of the buried delta winding would come into contact with ground. This would result in a real short circuit current, thus aggravating the damages and spreading the damages.

This may be an instance where there was initial damage which led to subsequent damage.
The underlying cause may have been the the decision to use a a four wire wye/delta scheme.
The delta/wye connection is a good connection. However when there is the possibility of a back feed, the back fed power will see the connection as a wye/delta and that is where the problems start.
After my time in wye delta land, I have to agree with both krisys and prc as to the development of this damage.

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

 

[prc]

Bill, you had mentioned many times problems of Yd transformers at distribution end. In many countries (like India), only Dy distribution transformers are used. Now the case we are discussing is a YNyn unit with a delta stabilising winding. Now the tripping ( assuming LG) reported -I took it as on LV side which can draw zero sequence current from grid.TV 18 may please explain the extensive tripping that he reported. Which side and nature of it.What can be the reason for it? Another point why Yd unit is not used for this application as normally with conventional step up transformers?

 

[waross]

Hi prc
Dy transformers are common. Yd transformers are problematic. If the wye primary has the neutral connected, (four wire connection) the delta winding will not tolerate any unbalances on the wye side. The delta stabilizing winding on the YNdn will try to stabilize any voltage unbalance or phase angle error on the primary.
The circulating currents in the delta winding may be avoided if the primary neutral is left floating. Then there is the possibility of 173% over voltage transients when the transformer or bank is energized.
Dy transformers are often a good choice for GSU transformers. A generator will tolerate unbalanced loading better feeding into a delta/wye transformer.
Problems arise when a delta/wye transformer is back fed and becomes a wye/delta.
A wye/wye transformer with a three legged core may not like unbalances because of the phantom delta. A better choice is a delta/wye if there is no possibility of a back feed.
I have read about and actually saw one legacy installation of a wye/delta bank with four fused cutouts. When the transformer bank was to be energized, the neutral cutout was fused first. Then the three phase fuses were installed. with the neutral connected, there was no over voltage transient. Then the neutral fuse was removed. Now there were no circulating currents in the delta.
The four wire wye/delta connection was common in the area of Central America where I spent some years. The problems were manifold.

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

 

[prc]

Thank you Bill for your lucid explanation,as always. But I believe YNyn ( ie star /star with solidly earthed neutrals on both sides) transformers can handle unbalance easily as it is the sole connection that I have seen in sub transmission sector in this part of the world. Of course due to the virtual delta, the tank surface will get heated up slightly during unbalance operation.

 

[sibeen]

Draw the conventional delta vector diagram. This will be an equilateral triangle. Now assume a 10% voltage drop on one of the phases. Take your eraser and shorten one vector by 10%. Now you will have a gap of 10%. Now one of the vectors will be 10% short and the triangle will not close.

Yeah, I get that, but how does that relate to the delta secondary of a transformer which is physically closed?

 

[waross]

The equilateral triangle represents the physical closed delta secondary of the transformer. The "triangle" with a gap represents the the transformer voltages when the primary voltages are unequal. The gap represents the voltage which drives the circulating current.

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