Distribution transformer protection settings

[ScottyUK]

At a hypothetical plant with an extensive HV and LV private distribution system it has been noted that the overload protection for the distribution transformers are set 'a little on the high side'. At this plant a hypothetical electrical engineer is mildly concerned that the overload protection is not properly protecting the assets on his system, principally the transformers and the switchgear. The engineer, who has inherited this hypothetical mess from his predecessors, faces opposition to changing relay settings "because that's how they've always been and there's never been a problem". The engineer privately wonders which substation will be the first to burn down.

For those who work in the distribution industry or who operate a private system - ideally a British DNO, but I'm really not all that bothered about location - roughly how far above continuous maximum rating do you set the pickup of your transformer secondary overcurrent protection? Transformer sizes are in the 1MVA - 10 MVA range, secondary voltages being 3,300V and 400V, principal load is motors. I'll not initially share some of the settings from this hypothetical plant because I want to test opinion rather than inject my own.

Thanks!

 

[mgtrp]

We (a generation/transmission/distribution utility, albeit a very small one) typically set at approximately 130% of the transformer rating. Typically, our loads are well below the transformer ratings (so we don't expect overloads), and our peak loading generally coincides with extremely cold temperatures (-40C), so being overly sensitive or overly conservative with the settings isn't of great concern to us.

Based on this and other experience at past employers (also utilitieS), something in the ballpark of 120-150% seems reasonable, especially if you also have winding/oil temperature protection and the ancillary equipment ratings are much greater than the overall transformer itself (e,g., no marginally rated bushings, tap changers, cables, etc).

Cheers,
mgtrp

 

[unclebob]

What mgtrp said is true for me too. Upper limit is often limited to coordinate with the primary relay/fuse. This is somewhat easy when you a have the ''luxury'' of a secondary Main Breaker. Things get harder when no main breaker is used in the switchgear, and the total load is greater than the transformer nominal power, or, when feeding a double-ended switchgear with only one main breaker and a closed tie. I guess 125% would be ok for small transformers (less than 2000 kva), and 110% for bigger ones.

 

[veritas]

From my earliest days in the hypothetical world I have been taught that the trfr should be able to handle 130% for a certain period of time - typically a few hours but that varies wildly based on ambient temperature, the nature of the load, type of cooling, etc.

I was then taught by a protection guru back in the early nineties to set the protection to 150% of rating - reason being the high reset ratio of the electromechanical relays (i.e. consider trfr heavily loaded, downstream feeder fault which is cleared by downstream protection - however, relay does not reset because reset is 90% - so it times out and trips). With the advent of numerical relays worst case reset is around 95% - typical is 97% to 99%.

What also needs to go into the melting pot is the actual relay pickup vs setting consideration. Some relays actually only start doing their job at 120% of pickup settings whilst start already at setting. So if you have a relay of the former kind, then if you set to 130% of trfr rating protection will only start at 1.2 *1.3 = 1.56 * FLA.

I have always held the opinion that there needs to be a clear distinction between overload (OL) and overcurrent (OC). In general OL situations are not the domain of the protection world but rather the system operator. Thus the protection should allow for the maximum loading of the item of plant - often termed emergency ratings in some utilities. Above this is when the protection should wake up.

The only trfr protection really which caters for OL conditions is the oil and winding temp. With these present I am more than happy to set to 130% of FLA as a rule of thumb. Should oil and winding temp be absent and ambient temperatures can be onerous, then this may require a rethink particularly where peak demand is associated with hot weather, i.e. air conditioning load. I would be even more concerned if it is an aged trfr - hypothetically speaking of course!

 

[Parchie]

I keep two important check points when setting the protection of distribution transformers: 1) allow transformer to operate below the transformer damage curve and 2) settings should be just above the inrush requirements of the load served.

 

[electricnewbie123]

I have always set my pickup to be 120% of the transformer rating or the maximum load expected. The reason I say maximum load expected is that some times, the transformers are oversized for current loading scenarios and there are some cost savings associated with cable installations if protection is set to loading requirements. In your case, I assume that the load is almost equal to transformer size and therefore setting of 120% is what I would be looking at.

However, one thing to note is that a check has to be done on transformer damage curves to ensure that the overload protection operates before damaging the transformer. Now, the transformer damage curves are not usually readily available (especially if the transformers are old). In that case, I would use generic transformer models and make some conservative assumptions and cater for age of transformers etc. to 'estimate' the transformer damage curves.

It is a must in my opinion that the transformer protection overload settings are below the damage curves.

 

[ScottyUK]

Thanks folks, the 1.2x Continuous Maximum Rating factor is what I've generally come across the industries I've worked in. The engineer in the OP has found some cases where the setting is about 2.3x CMR and others at around 1.6x and 1.8x CMR, in some instances also exceeding the rating of the secondary breaker and switchboard. I've never come across such high settings before.

As electricnewbie123 guessed it's an old process plant and details such as damage curves aren't available. Evidence of overload was noticed when some of the transformers were de-tanked for repairs, and the DGA records suggest that the transformers have run hot in the past. Most have an oil temperature alarm, but no trip function. Winding temperature indicators aren't fitted.

 

[prc]

As per the IEC transformer loading guide, 60076-7 of 2015, medium sized power transformers shall be suitable for overloading up to 1.5 pu under normal cycling loading and long time emergency loading.(for large units 1.3 and for distribution transformers 1.5/1.8 )So the current setting has to be selected based on the expected overloading and age of transformers. Overloading an old transformer or moisture laden unit is not good and must be restricted to rated power or even below. For such units, a setting of 1.1 or max 1.2 may be appropriate from transformer angle.

 

[waross]

This is probably not what you want to hear,but:
I spent some time with a utility where the standard was to use and abuse small distribution transformers to destruction. The primary fusing was to disconnect failed transformers from the grid and protect the rest of the grid.
When a fuse failed on overload it would be replaced with a higher rating.
I tried without success for several years to change the policy.


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

 

[Parchie]

@waross,
It's not that bad when you look at the bigger picture in a utility environment.
AFAIK, primary fuses are designed to melt around twice its rating (example: 3E clears in 10 minutes when loaded 235%), and this is a compromise between having to replace busted fuse links a number of times versus subjecting the transformer under overload for a reasonable period of time and survive! Besides, pole-mounted units have the luxury of getting lots of cooling air compared to transformer on vaults.

It is a totally different thing when you compare with an industrial setup, where loads are almost at fixed levels. It is a sin to set a tripping limit way beyond each transformer capacity there because of the typical constant load that will ultimately burn the unit out.

 

[Mbrooke]

I agree, where a transformer be tripped on overload varies greatly based on application. Cyclic load transformers (such as pole units) can have primary over load protection starting at 300% FLA without worry, substation transformers with cyclic load at 125 to 135% while others in continuous loading be set around 100%.

 

[electricnewbie123]

If someone sets a transformer overcurrent pickup level to 1.6 times the rating of the transformer, then that person thinks that it is OK to load it up to that level continuously. I can't understand why the transformer manufacturer won't sell such a transformer as a 1.6 times rated transformer!

 

[waross]

At 1.6 times rated load the transformer will run hotter. There is a close correlation between high temperatire and reduced lifetime of a transformer.
In a situation where a pole mounted transformer may be overloaded daily, temperatures above the design temperature may only occur on the hottest days. A cost benefit analysis may show that the cost of replacing the transformer compared to a shorter life argues in favor of a shorter life. The transformer may be changed out due to load growth long before the end of even the theoretically shortened life span.
Marketing and continuous loading at 1.6 rating in ambient temperatures close to the design maximum may take 10 or 15 years off of the original design life time.

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

 

[Mbrooke]

Not only that, but cyclic loading actually prevents reaching temperatures over the design limit.

 

[electricnewbie123]

waross, exactly the reason why they do not state the transformer rating as 1.6 times, and hence the reason why the protection has to prevent loading of transformer to those levels.

Mbrooke, I am actually not 100% sure on what you mean by 'Not only that, but cyclic loading actually prevents reaching temperatures over the design limit.' - could you please elaborate? Thank you

 

[Parchie]

If I may answer that, what mbrooke meant is that transformers with loads that change/ cycle from overload down to load levels below ratings can have enough time to cool down during off-peak times. The transformer temperature doesn't go beyond the specified max temp rise, hence it will likely survive longer than you expected.

 

[waross]

Cyclic loading;
The temperature depends on the length of time of the overload cycle.
Two of the philosophies behind demand metering and demand charges are:
1> The cost of higher capacity equipment to handle sustained demands.
2> The degradation and loss of equipment life due to short lived, relatively high demands.
Old time demand meters had a thermal lag of about 15 or 20 minutes before they completely registered an increase in demand.
The old texts claim that the time was chosen to appoximate the heating curve of a pole mounted distribution transformer.
I suggest that the comments on cyclic loading apply mostly to short term loading of less than about 10 or 15 minutes.
I'm starting to suspect that you are more concerned with the transformers feeding distribution circuits than with the final step-down distribution transformers, Scotty. If that is the case then I appologize for the diversion.

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

 

[ScottyUK]

Bill,

It's a generally interesting discussion, so thanks to all parties.

In the specific case I'm involved with these industrial substations are each typically equipped a two-section or three-section 3.3kV board fed from 11/3.3kV transformers and a two-section or three-section 415V board fed from 3.3/0.415kV transformers hanging off the 3.3kV bars. The load splits into a base load of process drives, plus cyclic stuff like lighting, plus intermittent loads like building services. The big loads are from the process drives, so an overload scenario is most likely to be a result of a transformer tripping off the bus and its partner(s) picking the up the combined load. The overload would probably be a sustained or continuous condition, which is what concerns me about the relay pickups being so high.

 

[waross]

Scotty
My thoughts on the specific issue:
First, What do yo estimate the time to failure to be of the transformer running at a sustained load of 160%?
Will the transformer be expected to fail in hours, days, weeks or more.
If the high setting will lead to rapid failure then the setting may be unwise.
If the setting will probably result in the lifetime of the transformer being shortened by a few years, then to maintain service and production the setting may be justified.
You are in the best position to evaluate the age, condition, ambient temperatures and load profiles.
In the event that a transformer is overloaded due to the failure of a comanion transformer; what will be the response time? what will be the expected actual load on the transformer?
A transformer failure often occurs at maximum loading, So there is a better than even chance that you may encounter a "Worst case".
May the production be curtailed to reduce the load somewhat?
A possible course of action may be to submit a memo outlining:
A> The long term cost of overloading the transformer. (To the expected load level, rather than the 1.6 factor.)
B> The cost of lost production due to a domino effect failure of a second transformer. (Domino effect.)
C> Similar projections with settings that you prefer.
The main thing is a CYA memo to let the bean counters decide and take responsibility for the final settings.
I hope that I'm not wasting your time by mentioning things that you already know.

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

 

[ScottyUK]

Hi Bill,

Very rare I find people here waste my time.

"The main thing is a CYA memo to let the bean counters decide and take responsibility for the final settings."

You're joking - I'll very likely be the one putting my name on the settings. The transformers are in fair condition for their age, but they aren't young. Normally we won't see an overload condition, but our regulations are fairly specific that

Strength and capability of electrical equipment

5. No electrical equipment shall be put into use where its strength and capability may be exceeded in such a way as may give rise to danger.

(Electricity at Work regulations 1989)

The settings presently in the relays allows a scenario where the capability of the transformers and thermal rating of the switchgear can be exceeded indefinitely without any automatic intervention to disconnect the load. As the duty holder for making sure our electrical system is managed appropriately, this situation is a source of concern. My current thoughts are to set the relay to pick up at the lower of: 120% of transformer rated current, or at 100% of the switchgear's rated current.