overloading a tranny

[Designer69]

do you guys know what the criteria is for overloading a transformer?

what are the limitations, what is acceptable or industry practice etc.

IE, we have some transformers rated at 300KVA and there are times during the winter where the estimated maximum demand is 678KVA so double the rating and these are working fine.

I am trying to understand where I can find the required criteria. thanks!

 

[zog]

The load carrying capability of a transformer is directly dependant on its thermal capability. The thermal ratings for a transformer can be found on the nameplate. The maximum allowable temperature is the ambient temperature plus the rated maximum temperature rise. A common thumb rule for life expectancy of transformers is as follows:

Class “A” insulation - 8C above max temp rating= ? life expectancy
Class “B” insulation - 12C above max temp rating= ? life expectancy

 

[augie47]

zog may correct me, but it appears to me that this question is answered entirely differently from a NEC standpoint than from a utility standpoint.
For work covered by the NEC, 450.3 has some fairly specific rules which limit overloading by virtue of the OCP device whereas utility companies allow considerable overloading based on tables they use and their vast experience in that area.

 

[zog]

zog may correct me, but it appears to me that this question is answered entirely differently from a NEC standpoint than from a utility standpoint.
For work covered by the NEC, 450.3 has some fairly specific rules which limit overloading by virtue of the OCP device whereas utility companies allow considerable overloading based on tables they use and their vast experience in that area.

Nope, I agree, just wanted to leave the NEC stuff to you guys. I bet you can fry eggs on that baby.

 

[iwire]

I agree with the other posters that there are criteria to go by but if it is an NEC application you have a code violation that should be corrected.

 

[G._S._Ohm]

"rated at 300KVA and there are times during the winter where the estimated maximum demand is 678KVA

8C above max temp rating= ? life expectancy"


Say TA = 40 C and Tins = 90 C with 300 kva so 90-40/300 = 0.17 C per kva of thermal resistance.
At 678 kvA it'd be delta T of 113 so 40 + 113 = 153C.
The life at 153 C is 1/79 of the life at 90 C.

The life at 90 C is L so at a "burn rate" of 1 it'd take L years to use up that life.
You're "burning off" the life of the transformer.

If the L at B = 1 is 20 years and you've already used up 10 years of the life, and then you ran at the high burn rate, you'd get 10.1 years instead of 20 out of this transformer.
If you ran at B = 79 you'd get 20/79 years of life [about 3 months] from this unit.

If your 'frmr values differ significantly from my default 'frmr values, so will the life.

BTW, the speed of many chemical reactions doubles at every 10 C of temp. rise and so the life of insulation halves at about every 10 C temp rise.

 

[BJ Conner]

Rought figure on loading

Rought figure on loading

do you guys know what the criteria is for overloading a transformer?

what are the limitations, what is acceptable or industry practice etc.

IE, we have some transformers rated at 300KVA and there are times during the winter where the estimated maximum demand is 678KVA so double the rating and these are working fine.

I am trying to understand where I can find the required criteria. thanks!

From an old Westinghouse book:
Load can be increased 1% per degree below 30 degrees C.
For every degree above 30 degrees C the load is decreased 1.5 %.
I know it works for the lower end. Transformer used for car heaters can be overloaded. Twenty below Fahrenhite is -30 C.
30-(-30) is 60 degrees so you could load it 160%.
People have been know to buy fans at the hardware store and cool transformers with them on a temporary basis.
IF your transformer is one of those oil filled padmounts you can use RB (Rain Bird ) cooling on the fins. Power companies usually don't like it.

 

[broadgage]

As others post, it is heat that kills transformers and not load as such.
I have seen an oil filled transformer loaded to 150% of the manufactuerers loading and it only reached an oil temperature of about 50 centigrade. This was however outdoors in a high wind and sub zero temperatures.
Except in such exceptionly cold and windy conditionss, I dont think that a transformer would survive long at over twice the rated load.
The O/P refers however to "estimated maximum load" which may suggest either an unduly pessimistic estimate, or a maximum reached only briefly.

Utilities regularly overload transformers and normally get away with it, but the NEC restricts this in case of customer owned transformers.

 

[zog]

It is not so much the tempature but rather the rise in tempature between the core and ambient that is the main concern. The difference in tempature from the internal to external part of the windings causes uneven heating and expansion of the windings and results in damaged insulation. That is why the limits on transformers are max rise.

 

[steve66]

do you guys know what the criteria is for overloading a transformer?

what are the limitations, what is acceptable or industry practice etc.

IE, we have some transformers rated at 300KVA and there are times during the winter where the estimated maximum demand is 678KVA so double the rating and these are working fine.

I am trying to understand where I can find the required criteria. thanks!

Just curious: Is that an actual measured demand, or a calculated demand? Outside transformers, or indoors?

 

[jim dungar]

Also, what is the length of time that the overload will exist?, How long between overloading? What is the normal loading?

It is heat versus time that kills a transformer. A 2.5x overload lasting 1 hour and occurring once per week, will have a different effect than a 1.25x overload lasting 24/7 for 6 months.

 

[Jraef]

Just a side note:

"Overloading a tranny" has a totally different meaning around here in San Francisco...

 

[G._S._Ohm]

"A 2.5x overload lasting 1 hour and occurring once per week, will have a different effect than a 1.25x overload lasting 24/7 for 6 months."

A 2.5 overload is
2^2.5 = normal aging sped up 5.7x
with 1.25 it's
2^1.25 = 2.4x

That's the problem: it's an exponential increase.

For just a 10% increase in aging rate it's
2^X = 1.1, so X = 0.14

Just a side note:

"Overloading a tranny" has a totally different meaning around here in San Francisco...

OMG.

I heard that one of these people of uncertain gender killed him/her/other self after saying "I'm all dressed up but I have no place to go."

http://en.wikipedia.org/wiki/Gender_identity_disorder

And going "one way" costs 5x that of going "the other way".

 

[jim dungar]

I was trying to point out that the magnitude of the overload may have very little effect based on the relative length of time it exists. A utility gets away with their extreme overloading during peak periods because there is a much longer period when their units are below full load. A transformer that is loaded 2.5X for only 4% of it life, will not have its expected life severely impacted if it sees 20% loading for 50% of the time and 75% loading for the remainder of its life.

You really need to know the actually loading profile before you can make judgment on life expectancy. Temporally raising the temperature of insulation by 10?C does not mean its life expectancy is automatically cut in half.

 

[chris kennedy]

Just a side note:

"Overloading a tranny" has a totally different meaning around here in San Francisco...

Good stuff.

 

[mbeatty]

Just a side note:

"Overloading a tranny" has a totally different meaning around here in San Francisco...

Now that was funny!!!! :grin:

 

[Designer69]

excellent info, thanks!

I will look into the duration of the OL condition and share.

by the way, this is a utility application. Should have mentioned that...

 

[G._S._Ohm]

This should work. You can add as many RIDs and TARs as you want as long as they don't total more than DLOL. TARs and DLOL is in years.

speed x time = distance
rate x time at that rate = length
Rate of insulation degradation [RID] x time at that rate [TAR]
should be equal to or less than Design Length of life [DLOL]

if insulation life halves for every 10 C increase
and Ta and Tc is amb. And conductor temp., respectively
then RID=2^((Tc-Ta)/10)

40 >enter Ta
100 >enter Tc
64 =calc'd RID

If Rca is the thermal resistance, conductor to ambient
then Tc-Ta = kvA/Rca

50 >enter kvA
60 =calc'd Tc-Ta
1.2 =calc'd Rca

=(RID1 x TAR1) + (RID2 x TAR2) = DLOL

20 >enter DLOL
0.25 >enter TAR1
64 =RID1 from above
1 >enter TAR2
1 >enter RID2
17 =actual length of life
^ it must be less than DLOL

 

[Open Neutral]

It is not so much the tempature but rather the rise in tempature between the core and ambient that is the main concern. The difference in tempature from the internal to external part of the windings causes uneven heating and expansion of the windings and results in damaged insulation. That is why the limits on transformers are max rise.

But won't oil-filled units have far more even heat distribution? I know substation limits for WMATA is temp on the 2MW xfmrs' oil temp; by the end of rush-hour they'll be hitting the red line.

 

[zog]

But won't oil-filled units have far more even heat distribution? I know substation limits for WMATA is temp on the 2MW xfmrs' oil temp; by the end of rush-hour they'll be hitting the red line.

Depeneds if they have pumps or natual circulation. Oil filled typically have several oil temp gauges monitoring different parts of the transformer. It also depends on the core design.