Transformer fire modelling 1

[Power0020]

I wonder if there is any software tool to model a power transformer (180 MVA) fire to see temperature distribution around it up to 15 m or more away from the transformer?

this will be useful in assessing the hazard of fire propagation for nearby buildings

 

[krisys]

There is a French company which is dealing with nearly similar kind of things.
I do not remember the name of this vendor, but they have thrown some light on this subject.
They doe the early detection of transformer fire with some condition monitoring technique, which is unique!

 

[aussiejohn2]

Hi Power0020,
I have somewhere a paper with some measured results (i.e. full scale testing) from one of the Australian organisations - possibly ENA or ESAA, or maybe TransGrid. You might try them. The resulting temperature profiles were strongly skewed in the direction of the wind at the time.
John.

 

[prc]

Why you want to model a transformer fire? Better prevent it ! I believe it will be extremely difficult. No one can predict the pattern and intensity of a transformer fire. You will get everything on the subject from CIGRE technical Brochure N0. 537-2013 Guide for transformer fire safety practices. This 139 page document prepared by CIGRE WG A2.33 can be down loaded free from

http://www.cigre.org

for members. French Company SERGI is making nitrogen injection system to prevent and spread of fire from transformers. You can get lot of details from their web site too.

 

[ScottyUK]

GL Noble Denton acquired Advantica, a company specialising in this type of work, at some point since I last dealt with Advantica. They seem to have retained the Spadeadam fire & explosion test laboratory and have software tools which can probably do what you need. I'm not sure exactly which bit of the new organisation you need to speak to, but this link might help.

https://www.dnvgl.com/oilgas/labora...d-blast-research-testing-dnvgl-spadeadam.html

 

[dpc]

Lots of variables. I'd be very suspicious of any predictive model. Even if testing is done, there are so many factors involved. Generally there are building and fire codes that address this concern, so that will be a minimum requirement. If this is a high downside situation, firewalls can be built.

 

[ScottyUK]

You might want to post a link to this thread in on of the Process Engineering fora, perhaps forum798, and get their take on it. Modelling software such as Phast is quite established as a means of modelling scenarios which can't reasonably be simulated in a laboratory, although it is usually applied to hydrocarbon processing facilities rather than electrical plant.

 

[krisys]

French Company SERGI is the one who provides a unique solution. I have attended one of their presentations and I liked their concept.

It is mainly for the industry and utility application, where the transformer application is critical for the sustenance.

It is rather focusing to minimizing the damage to the transformer than preventing the fire.

 

[krisys]

Here is the link for SERGI TRANSFORMER PROTECTOR:

Link

 

[aussiejohn2]

Hi Power0020,

I agree with others that modelling is not justified. Scotty is right that Phast can (easily) do it, but the uncertainty in the inputs vastly dominates the results. Was there an oil mist that ignited (the famous video on the ?Florida? golf course)? Did the bushing explode? Did the tank fail at the lid, or on a corner? Was the pool contained in a bund (consider the ?Westwind? fire in the US)? etc. etc.

For building protection, building codes or wiring standards are usually explicit about appropriate distances. There is also some guidance in Factory Mutual's Loss Prevention Sheet for transformers, as well as IEEE 979. I'm fairly sure the ENA reference I referred to above is their Guide for Fire Protection of Transformers.

If you're concerned about fire risk, why not just pay the small premium for natural ester dielectric?

Regards,

John.

 

[aussiejohn2]

Hi all,

Hijacking the thread a little. Most utilities/mining/industrial I have worked with over the years have had a rock filled, closed base concrete bund around their transformers. The bund is sized to hold (selected and maintained) rocks, plus all oil, plus storm (and sometimes fire) water. It is considered that any fire will be suppressed by the rocks and oil released will not cause a pool fire. I know of some examples that support that - windings and oil left in the tank burned for many hours despite attempts at extinguishing (no Sergi!), but oil in the bund did not.

My new utility instead uses a shallow bund, piped into a below-ground tank sized to hold full oil volume. An underflow from this tank goes to the top of a second, identical tank, and an underflow from that tank via a valve to waste at grade. Oil floats, so any oil stays in the first tank, with any rain water underflowing to the next tank. Normally the water in the next tank is tested before release and any small amounts of carryover oil pumped out via hatch for disposal.

The pipe from the bund is taken from a small sump. Standards say that this should have a "suitable" flame trap. We are electrical folk, unsure what is suitable, and have been unable to find or buy any advice. Does anyone have any pointers?

Many thanks,
John.

 

[Power0020]

Thanks John, I am familiar with AS/NZS oil containment practices.

The point in that particular project is that there is a building pretty much near to the substation, the 15.2 m fire zone is encroaching the building and firewalls will cause additional hindrance for transformer installation, maintenance in addition to cost.

The client is of the opinion that with active fire suppression, the need for firewalls aren't critical. However, failure of fire suppression system (water spray) will literally, leave the building endangered.

How probable is that? well, I don't have a figure on top of my head to say that the unfortunate instance of transformer fire will-or would- happen once every x years, but, the consequences may well be catastrophic given the building is classified as a residential one (it is designed to accommodate substation and nearby power plant workers and families...).

The transformer is about 7 meters away from the building (in plan). An EPC contractor will simply answer "in the case of fire, the building needs to be evacuated"....I believe that also won't be the answer here, as the fire may occur at 2 am night!.

Probabilistic approach will balance the likelihood of event versus the cost of consequence and associated fatality risk on annual basis, I doubt that may fall within a value less than 1 in 1000000 fatality per year but again, is a transformer fire subject to a probabilistic risk approach? something similar to earthing?.

 

[ScottyUK]

My involvement in modelling a transformer fire with Advantica happened after a devastating explosion at our plant which killed three colleagues. I can say with confidence that anything which was 7m away from that transformer would have been in a whole lot of trouble from both fire and blast damage, and would also be in the line of fire for all manner of projectiles.

The deluge fire suppression may have helped contain the fire afterward, but the initial release of oil occurred with so much violence that it virtually atomised and detonated, with remnant oil continuing to burn afterward. A heavy (40kg) blockwork fire wall suffered structural failure due to the blast, and although it was more-or-less still standing it had to be demolished immediately afterward. By contrast the reinforced concrete blast walls fared reasonably well with only surface spalling due to the fire. It was around that point the unpleasant truth hit me that the blast walls were there to protect the plant, not the personnel: the blast was directed away from the generators, toward us.

The transformer which blew up at our site was a 48MVA unit auxiliary transformer. Yours is considerably bigger. Build the wall: if possible specify a blast wall, not a fire wall.

 

[krisys]

Power0020 (Electrical),

By the way, what is the size of the transformer are you talking here? The requirement goes with the quantity of oil in the transformer. What is the rating of the transformer?

There are few mitigation available, such as deluge, providing the oil soak pit and spreading the transformer area with metals (i.e. the crushed rock pieces), using dry type (cast resin) transformer (if it is a residential area), specify vegetable oil instead of mineral oil (this has higher fire point) etc.

Also the local regulations shall be complied.

I am sure ScottyUK is mentioning about large power transformer, installed in the utilities or a large industrial complex. Your case may be less critical.

 

[Power0020]

It is 220/66 kV, 180 MVA, with about 80,000 L of oil!

 

[krisys]

Certainly it is falling under the category of large transformer by any standard. So, as a minimum you have to comply to the local/national regulatory requirement. If the industry standard is exceeding the local requirements, it is recommended to follow the industry standard.

 

[prc]

If you have not yet ordered the unit, the following steps will reduce or eliminate the fire risk:
1) Specify Resin impregnated dry type bushings instead of conventional oil impregnated condenser bushings. Most of the transformer fires are from blasting of bushings.
2) Specify 2 nos pressure relief devices mounted at two opposite corners of cover,.
3) Ask for rupture resistant or proof tank
4) Specify the nitrogen injection system of SERGI or provide water deluge or water mist fire protection systems.
4) You can also consider ester filling in place of mineral oil. But I will not recommend from cost benefit angle.

 

[aussiejohn2]

Hi Power0020,
Of course it is possible to apply risk analysis - as for anything. It is difficult to get high accuracy estimates of either likelihood or consequence but it is certainly possible to get adequate for purpose ones. In your case the transformer is so large and the building so close that i think the outcome will be very clear despite the uncertainty. A reasonable estimate is 1 fault per 100 transformer years. For your size of transformer, 1 fire per 10 faults is an ok starting guess. There is data for failure on fire suppression systems available - consult a fire engineer if you can't get it elsewhere. I would say that even with the mitigations others have suggested, the residual risk is still significant. The possible exception is the k class fluids. There are, apparently, no reported pool fires with natural ester (and presumably synthetic also). It is definitely worth tracking down the US substation fire i mentioned earlier - West coast, maybe around LA. One tx, from memory of the order of 750 MVA, ignited another four or five.
John.