Power Transformer Micarta Diaphagms vs Relief Valves

[Bambie]

thread238-9659

The manufacturer's manual for their 30 year old power transformer infers that the micarta diaphragm and goose neck vent protects the gas detector relay from seeing pressure transients above 5 psig, which will cause loss of calibration. It also credits the diaphragm with mitigating the severity of damage resulting from explosive pressurization from a fault.
Can a relief valve set to 5 psig with a discharge pipe routed down to the berm (to satisfy environmental concerns) also claim to provide these attributes?
Would the insurance company find a loop hole if a fault occurred?

 

[prc]

Pressure diaphragm at the end of a goose neck protruding over the conservator (some times an additional diaphragm provided at the pipe joint where it starts from cover) is an old design that no one adopts these days. I don't know micarta- it used to be phenolic resin impregnated paper board (mica sheet) of 0.5-1.0 mm thick to break at an overpressure of 0.5 kG/cm2.PRDs are much superior and reliable with the advantage of resealing property ie after releasing excess pressure it will close automatically after giving indication at control panel. These are provided to protect the tank from rupture in case of dynamic overpressure from a severe internal arc flash from fault . So insurance company should be happy with PRD over diaphragm plate.

When you say gas detector relay is it the gas operated relay mounted on pipe connecting tank to conservator?

 

[Bambie]

prc,
A ruptured diaphragm and gradually tapered goose neck imposes very little dynamic backpressure in the transformer.
An open relief valve imposes set or blowdown pressure in the transformer plus any dynamic backpressure from the deflection shield and long discharge pipe.

The gas detector relay is mounted on the top of the transformer with a maximum of 5 feet of oil above it in the conservator. It is set to trip the transformer at 2.8 psig, which may not prevent a pressure rise past 5 psig due to thermal inertia. Loss of calibration means the pressure relay may not detect the next transient.

The ruling given B.G.E. may have been different had they substituted a relief valve for a rupture disc (attached).

 

[prc]

Knowledge changed a lot during the past 60 years. The honorable judge is referring to the small valve in kettle. What we are now referring is 150mm (6") opening PRD that is far reliable and superior to diaphragm.
It is also to be understood that PRD or diaphragm may not work in case the fault is in a faraway place inside transformer than PRD position. Then tank will rupture. Nothing can be done about it. A survey done by CEGB,UK (CIGRE 1960) proved this beyond doubt. That is why in large transformers two or more PRDs are provided on tank cover.
The sequence of operation of relays in case of a severe fault inside the tank is - first electrical relays(over current /differential), PRD / gas operated relays.[pre][/pre]

 

[Bambie]

prc,

What pressure should the PRD be set to open at (ie., WHAT is it protecting)?
If it is supposed to protect the gas operated relay bellows, then shouldn't it be set at or below 5 psig minus the static head to the top of head tank?
Why wait for the head tank to go solid (5 feet static head above the PRD = 1.8 psig) before opening?

 

[bacon4life]

PRD's usually are to protect the structural integrity of the tank. Why wouldn't you just replace the fragile relay after a PRD opening? Even if you had the PRD set statically low enough to protect the PRD, the dynamics of a fault will cause highly unequal pressures throughout the tank.

 

[Bambie]

bacon4life,

The gas detector relay can't detect pressurization rates less than 1 psig per minute and the PRD set pressure tolerance increases to +/-2 psig on 5 psig at pressurization rates less than 2 psig/second.

Wouldn't it be prudent to have similar PRD and relay set pressures?

A goose neck on the PRD discharge also seems like a good idea.

 

[prc]

Goose neck is not a good idea. It delays the pressure transmission. We have to remember that the pressure build up after a fault is not uniform or static. The dynamic pressure moves like a wave and hit the nearest tank surface. So the limitations of PRD and gas detector relay.
PRD settings are normally between 0.42 to 0.7 kG/cm2.Too low setting is not good for PRD.

 

[Bambie]

prc,

For environmental impact and fire potential mitigation the PRD manufacturers offer a directional shield and discharge pipe that direct oil flow to the berm. My calculations indicate that the PRD flow capacity is halved at the very least.

If the encumbered PRD can only respond to slow rates of pressure rise that the gas relay cannot detect and moderate rates from decay heat following successful trips by the much quicker gas relay, then it's purpose is really to protect the gas relay and should be set-up accordingly.

The brittle micarta diaphragm is superior because it responds quicker with higher capacity.

Perhaps the only advantage of a PRD is not mitigating overpressure but avoiding the consequence of rupture; oil contamination that increases the likelihood of a successive fault.

 

[ScottyUK]

Don't under-estimate the effect of the swan-neck on the ability of the bursting disc to relieve pressure during the fault. The directly-mounted PRD may well out-perform a similar-sized bursting disc. If you make the bursting disc and swan-neck sufficiently large then of course it will out-perform a commercially-available PRD, but experience says that bursting discs on transformers are rarely larger than 8" or so even on large transformers.

Personally I agree with prc, if the fault is big enough then the tank is going to rupture regardless of whether you install a PRD or bursting disc. The only difference is whether the rupture occurs with some of the pressure relieved to atmosphere or not. When we lost a transformer in a massive fault one of the questions we asked over and over again in the immediate aftermath was "Why didn't the PRD prevent the tank rupture?". We estimated at the time that the area of the apertures created by the tank rupture were at least 50x that of the Qualitrol 208 PRD and could well have been double that.

 

[prc]

To my understanding gas detector relay should detect pressure rise first and isolate the transformer. Then only (or if fault is sudden and severe together) PRD will come in to play. Purpose of PRD is not to protect gas relay. Discharge pipe is not going to affect the operation setting of PRD.

 

[ScottyUK]

Bambie,

Is the gas detector relay of this general type:

https://library.e.abb.com/public/7e...3cc/1ZUA5663-210_r2_Sudden Pressure Relay.pdf

? If not, do you have a datasheet you can link to. I'm assuming that you're not talking about a conventional Buchholz relay.

 

[Bambie]

ScottyUK,

I have attached the characteristic curve for the original gas detector relay (top) and the replacement (bottom).

 

[Bambie]

ScottyUK,

And yes, the old and new gas detector relays are BUCHHOLZ designs.

 

[Bambie]

I found a TVA transformer overpressure calculation (attached).
I would be interested in seeing this or a similar detailed calculation.

 

[prc]

Bambie, There are three type of relays- (1)Buchholz (gas operated- used with conservator with breathers) (2) Gas operated relay - Mounted over the tank (used with nitrogen sealed units) operated by collected fault gases, (3) SPR (shown by Scotty) mainly used in sealed units, operated by rate of rise of pressure inside tank, mounted on side on tank or in the nitrogen gas space

SPR is expected to do the same function as Buchholz. The characteristic curves put up by you seems of SPR type relay. So please mention correctly the type & make of relay to understand the issue.

The calculation is simply basic. Unfortunately actual PRD working is not based on the volume of gas generated. As I mentioned earlier, dynamic forces travel at lightning speed and hit the nearest side and then spread. Engineers have developed much better calculations and analysis for fault gas generated pressures, rupturing strength etc. Still I will take these with lot of doubt. Please refer IEEE draft standard PC57.156. D4.1-2015 Guide for tank rupture mitigation of liquid filled transformers. This draft is prepared based on a number papers and studies published during last 30 years.

 

[Bambie]

prc,

Thanks for the IEEE reference.

The characteristic curves are for the original Model 12 Westinghouse Buchholtz design and the newer ABB Model 12 replacement.

Both manuals state that pressure excursions beyond 5 psig cause loss of calibration and damage to the bellows.

I have difficulty verifying this change to overpressure protection for the transformer (and gas detector relay) because the relief valve Vendor cannot provide:

1) relief capacity with a directional shield and discharge pipe attached.
2) relief capacity for pressure differentials other than 15 psig
3) relief capacity for transformer oil (or a conversion factor to apply to air scfm)
4) any test data confirming that a relief valve replacement for a micarta diaphragm with the same inlet diameter has a higher capacity and a quicker response time.

 

[prc]

As per my understanding Buchholz relay has no bellows, only floats or buckets operated by the gas collected(alarm) and velocity of oil flow to conservator. (trip) I believe what you are referring is SPR for which calibration may get affected by pressure. But SPR is for detecting such pressure rises before PRD release and why to worry to save it when transformer itself is getting damaged.

There is no need or possible that PRD will release all the gas generated. From the disturbance records, what I have seen, is PRD is operated after several ms after electrical relays trip the transformer. I don't think any one can honestly answer the questions raised by you and there is no relevance to actual site situation during a severe fault. IEEE standard mentioned above may clear some of your doubts.

 

[ScottyUK]

Hi prc,

Here's the ABB manual for the relay. It is definitely not a conventional Buchholz relay, although it seems to serve a similar purpose to the float chamber in a Buchholz design, i.e. a device for capturing gas evolved slowly from a relatively gentle fault rather than a gross failure which would operate the Buchholz surge trip or activate a PRD. I haven't seen one of these before.

 

[prc]

Thank you Scotty for that manual. It is from ABB, Alamo, USA. I was also not aware of it. It is a combination of two relays (2)& (3) mentioned in my earlier post. I knew that both these relays are used in US, but first time to see a combo. I am curious to know whether any gas operated relay is additionally used in the conservator pipe in addition to this relay on tank top.
Since this is not used in the pipe line to conservator, but over the tank, 1.5 meter below conservator oil level, we need not be concerned about over pressure. The manual is also not saying anything about PRD setting matching with setting of this relay.