economics of REF

[baj]

At what kVA and what transformer design (dry, oil etc.) does it become sensible to use REF to protect secondaries?
In the past, my company (UK based but many plants around the world) used oil xformers outdoors with cables to LV switchboards, and always used REF. Now, trend is to dry transformers bus-barred directly to switchboards. As transformers are now relatively cheap, and risk to secondary connections is less than with cables, is REF still sensible? FYI, typical transformer ratings are 1MVA to 2,5MVA, in 10 to 20kV range, REF is costing c.GBP1000 per transformer. We always have 2 fully rated xformers feeding each switchboard, hence 100% redundancy built in.
I appreciate this is probably not 'black or white' but welcome opinions.

 

[mc5w]

What the heck is REF?

 

[RRaghunath]

REF protection is for the winding of the transformer and especially the star connected windings (the conventional earth fault protection in the residual current circuit of the three phase CTs is restricted by default, any way).
It is normal to provide this protection for transformers whose neutral is earthed through a resistor (i.e. not solidly grounded) even if the transformer happens to be low kVA rated, from the risk perception point of view.

For solidly earthed systems, the current trend in India is to provide REF protection for transformers rated 2MVA and above.

 

[stevenal]

REF = restricted earth fault protection. You answered a question about testing such a system recently. thread238-117555

 

[ScottyUK]

I'd agree with rraghunath that REF is uncommon on small solidly-earthed transformers. Typical UK practice on a solidly earthed distribution-level transformer would, off the top of my head, be overcurrent + high set, earth fault + high set, winding / oil temp trips, PRD, Buchholz if fitted with a conservator or sudden pressure relay if hermetic.

Larger solidly-earthed transformers justify the additional cost of providing an REF scheme, particularly as it can be piggybacked on the transformer differential CTs. transformer differential is itself a protection rarely seen on small transformers.

£1000 sounds like a bargain if you are fitting four dedicated CTs to the transformer, buying the relay, installing cabling, commissioning it...



----------------------------------

If we learn from our mistakes,
I'm getting a great education!

 

[RalphChristie]

I have never seen REF-protection on transformers smaller than 10MVA. Even our resistance-earthed transformers (used on some of our mining equipment, 3MVA, 11/6.6kV) do not have it. Not that there is anything wrong with it. The more types of protection, the more kinds of faults it can detect.

As a rule of thumb, I've heard to use at least 10% of the transformer's cost for protection. Normally, on older diff and REF-schemes, class-X CTs were specified. Class-X CTs are much more expensive than normal protection CTs, resulting in a high-cost scheme. With the new kind of relays you can use normal protection CTs, and except maybe for the relay, the schemes are not so expensive anymore.

Another thing to consider is the cost, importance and the accessibility of a transformer. Big transformers take a long time to rewire and are not always easily accessible, which can justify the cost of more protection schemes.

Ralph

 

[baj]

Thank you all for the interesting posts.
I realise now that I didn't state that we always use solid earthed neutrals on our projects.
I believe that class X CTs are still required (and we always spec them).
Part of the reason I started looking into REF protection was because I had a couple of REF mal-ops on my last project. I have never seen a genuine trip on REF.
At risk of being contentious, if a genuine fault occured that REF detected, the transformer is out of service anyway, and I'm not sure that the more sensistive protection of REF would benefit me much. For a high value transformer maybe, but 1-2.5MVA dry transformers are not high value (relatively).
On my projects, it could be that with 16 or 18 'cheap' transformers, it would be more econonomical to not use REF (saving at least GBP 16,000-20,000)which would more than cover the cost of a new transformer/rewind in case of failure.
As ever, I'm interested to hear opinions. Thanks.

 

[RRaghunath]

I agree with the contention, especially considering that the transformer neutrals are solidly grounded.

I myself know the varying practice with regard to motor differential protection. In some countries all motors exceeding 1000kW rating are specified to have differential protection and in some other countries even 3500kW rated motors are not required to have. That way the protections are region/country specific. The 10% rule with regard to cost of protection (vis-a-vis the equipment cost) is also one of the basic rules I learnt in the early days of my career and I think is a good guide always.

With regard to malfunctioning of REF protection, it could be some thing to do with the sizing of stabilising resistor in the REF relay circuit and is not common (I haven't come across any such maloperations).

Yes, CTs need to be Class X CTs for the high impedance type REF protection (high impedance type is more sensitive as well as fast and preferred for REF protection even today).

 

[RalphChristie]

What type of REF-protection do you use?

The reason for using REF-protection is mainly to protect transformers against earth-faults, and especially earth-faults near the neutral-point. From a HV phase-view this current can be very low (too low for an overcurrent relay), but if seen from the neutral, the current can be very high. (faulty winding acts like an autotransformer)
For diff-protection on star-connected transformers there are a large part of the winding not protected:
10% diff-setting = protect 58% of the winding
20% diff setting = protect 41% of the winding
30% diff setting = protect 28% of the winding
40% diff setting = protect 17% of the winding
With REF-protection you can protect more than 90% of the winding. Another good feature of REF-protection is the fast-operating time, (like diff-protection it need not to coordinate with downstream devices) resulting in less damage if there is a fault.

On all the REF-schemes I've worked with, (high impedance) there were never any maloperations. If it tripped, there was either a fault in the zone or there was a fault in the setup of the scheme. I have never had any problems with stabilising resistors, but I've found sometimes wrong CT-polarities, and especially a wrong polarity at the neutral-CT. What actually happen then is that the scheme becomes stable for faults in the zone, but trips for faults outside the zone.

Regards
Ralph

 

[ScottyUK]

Hi Ralph,

The reversed neutral CT is a pig to find isn't it...



----------------------------------

If we learn from our mistakes,
I'm getting a great education!

 

[RalphChristie]

Scotty, how true...even more if the CT/CTs are not easily accessible...

 

[GerH]

You might consider what REF is going to achieve. I would say it has the possibility to reduce the extent of damage in the event of a fault. I have seen REF and DIFF faults that resulted in easy to repair damage.

If it is a small tranformer say 1 or 2MVA then the replacement might be on more or less immedaite delivery and most people are going to fit a new unit so the benefit is doubtful. On the other if you had a single unit that could not be easily replaced e.g. non-standard voltage or the larger sizes that are only made to order on 16 or more weeks delivery the REF protection cost could be insignificant compared to downtime avoided.