Voltage transformer explosion

[hhhansen]

Hi all,

Recently a local power plant suffered from an explosion initiated by a voltage transformer inside a closed room. This caused a double busbar fault on 132 kV. During the fault a steel door was blown out of the room. At the adjacent room a huge 500 kg steel door was blow 32 out in the yard. Photo´s attached in document.

Regarding the incident I have a number of questions which puzzles me:

1) The voltage transformer contained 50 kg of Nynas Nytro oil. During the explosion I assume the oil content is vaporised and ignited. But does the vaporised oil really contain enough energy to cause the damages?. In case how is the potential energy calculated for the oil? (I assume that the short circuit current does not have enough energy to cause the damages, since short circuit current is interrupted within hundreds of milliseconds)

2) What is the likely reason to cause the break down of the voltage transformer? It is installed during 2005.

3) What must be paid attention to avoid similar future incidents?.

Best Regards
Hans-Henrik.

 

[waross]

The oil must not only be vapourized but also mixed with air in order to explode.
I was in the room when covers were blown off an MCC due to a phase to ground fault in air. The 800 Amp breaker cleared the fault almost instantly (8000 Amps instantaneous trips).
8000 Amps at 600 Volts (347 line to neutral Volts) is vanishingly tiny compared to the possible energy of bus bars at 137 kV.
There is an incredible amount of energy in an electrical fault.
The temperature of the arc may instantly vapourize copper and steel. The heated air in the vicinity of the arc expands violently and explosively.
Remember that the asymmetric current in a fault may reach almost 3 times the symmetrical current on which Available Short Circuit currents are based.
If this was a through fault, the magnetic forces in the transformer may have initiated the damage.
I suspect that any oil fire would be subsequent to the initial electrical explosion.
Was this transformer feeding the fault?
Alternately;
Consider only a small amount of oil vapourized and forming an explosive mixture with the volume of air in the room. Compare this volume to the volume in one cylinder of a diesel engine.
Either way, you are dealing with a potent bomb.
I'll step aside now and let ScottyUK continue.

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

 

[oldfieldguy]

I'd like to see pictures.

I concur with Bill's assessment. The forces generated by the fault's arc, the subsequent vaporization of metal, the creation of plasma, air expansion, all these can create some frighteningly impressive effects without the addition of a fuel-air explosion.

old field guy

 

[ScottyUK]

Er, thanks Bill...

When the copper of the busbar is vapourised its volume expands massively. This rapid expansion in its own right can generate severe over-pressure, especially if the fault is not cleared quickly. When the VT exploded it is quite possible that the casing ruptured and formed a jet which acted as an atomising nozzle much like a diesel injector in an engine. Clearly you have a very power ignition source in the arc, so you have all the ingredients for a fuel-air explosion to add to the shockwave and over-pressure from the arcing fault.

Calculate the area of the door and you might be surprised just how big a force even a moderate overpressure can generate. This link shows the aftermath of an OCB failure on an 11kV board; I would guess that the double doors and frame, which landed 20m away from the switch house having passed through a fence en-route, probably weighs quarter of a tonne or more.

At 132kV the available fault energy will be an order of magnitude larger than at 11kV, and the 132kV VT will contain roughly the same amount of oil as the 11kV OCB.


As for the cause - first guesses would be one of: physical damage; partial discharge; or bad oil. Impossible to tell without a lot more information and maybe not possible even then. You need to engage the OEM's help with the investigation.

 

[itsmoked]

They use just a few kJ of capacitor bank discharge thru a tiny wire to explosively form 1/4" steel plate into deep intricate bell-housing shields.

The oil is entirely inconsequential with regards to any explosive aspect. A measly 3psi overpressure will remove a steal walled building from its concrete foundation. So if we limit ourselves to 2psi those doors had 80" x 36" x 2psi => more than 2tons of force applied to them.

Those building should really have simple blast roofs.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[prc]

Let me answer the questions ad seriatum:
1)The research on this subject was reported in USA ( 1960's &70's by EPRI) and recently in Japan( Tokyo Electric) & France(EdF & SERGI).As per current understanding, the arc inside oil generate a gas volume which is a logarithmic function of the arc energy.ie first one MJ of energy will generate 2.3 cubic metre of explosive gas from oil while a 100 MJ arc will generate only an additional 2 cubic metre of gas from 99 MJ of energy. Arc energy is proportional to the arc or fault current.As soon as as arc occurs, gas is generated which cocvers the arc, preventing it from coming in contact with surrounding oil.This arc quenching propery was used in oil circuit breakers. The intense heat from arc, breaks down the gas molecules converting it in to plasma.This moves out as a shock wave with accelerations of hundreds of "g", hitting the nearest part of the equipment (in this case porcelain) shattering it and throwing out hot gas and oil. Oxygen from atmosphere cause burning of oil and a fire breaks out.In large power transformers, with massive quantity of oil, such tank rupture rarely cause a fire as the temperature of gas cools dwon by the time it comes in contact with atmosphere.But a failure of a bushing or OLTC can result in a transformer fire.

2) It is hoped the VT is of electromagnetic type.Major cause of failure of such VTs is ferro-resonannce. But the VT was in service since 2005 and since no change in service conditions were made( like, new capacitors in to nearby grid, cable connections etc) we have to look in to other reasons. Apart from design or manufacturing lacunae, a possible cause is water ingress inside through any of the gasketed joints.Recently two tutorials came out on ferro-resonance which may be useful to power engineers - IEC Technical report IEC TR 61869-102: 2014 Instrument Transformers-Part 102-Ferro-resonance in substations with inductive voltage transformers and CIGRE Technical Brochure No.569 Resonance and ferroresonance in Power networks,published in Feb,2014.

3) How to prevent violent VT failure? First the cause is to be idetified and then counter measures may be taken.Primary fuse will not help. Insulation power factor checking can be a preventive diagnostic tool.

 

[scottf]

One thing to look at in regards to a VT's ability to withstand an insulation failure is the robustness of the connection of the ground layer of the insulation to the tank (which is then connected to the station ground. Some designs use a relatively small conductor for this connection, as under steady state conditions, there is very little current flow.

However, in the event of an insulation failure, that connection has to carry the available fault current. In too many designs this connection is inadequate, which is a primary cause of catastrophic failure.

When evaluating a potential CT or VT design, always look at this connection and how much fault current it can take. Technically, it's more than just this connection, but the entire insulation ground circuit, commonly called the Fault Current Carrying Path of the insulation. For VTs, it's the outside layer of the insulation and for most head-type CT designs it's the core housing, plus the stem, plus the connection to the CT's base.

 

[7anoter4]

why transformer explode:

http://www.sergi-france.com/why-transformers-explode.php

There are "explosion proof"potential transformers:

http://www.siemens.co.in/pool/about.../instrument_transformer____ct_transformer.pdf

 

[prc]

Scottf, my experience does not go with your solution. Even before the fault current reach the earth connection, the shock wave from failure point used to blast the porcelain /lower tank.

 

[scottf]

prc...in my past life, I worked for HV instrument transformer manufacturer for about 15 yrs or so. The fault current carrying path was essential to our "explosion-proof" designs. We did many type tests to prove the design and it was proved in the field over many decades. As long as the tanks and insulators are strong enough and as long as the fault current carrying path is robust enough, the unit will not explode. "explode" was always defined as nothing can land outside of a circle with the radius set to the unit's height. Essentially what would happen is that the expansion bellows and cover would fly off, but the unit itself would remain intact. Those tests and ratings (originally done in conjunction with WAPA here in the US) were the early origins of what is now included in the HV test addition to IEEE C57.13 and CSA-C60044-X.