Restricted Earth Fault Protection 4

[electromott]

I am involved in a UK project that has a 11kV/433V sub-station, but have come across conflicting information regarding REF protection. Are there any hard and fast rules when REF protection should be used on the LV side of a transformer and when it shouldn't?

If it is/isn't used what are the pro's and con's?

 

[RalphChristie]

Electromott

There are two types of REF-protection:

High impedance REF - voltage operated.
The main disadvantage of high-impedance protection is the need for dedicated CTs (normally class X), thus higher cost.
Low impedance REF (bias or unbias) - current operated.
Main disadvantage of circulating current protection using low impedance relays is through fault instability due to CT saturation.

A simple overcurrent and earth fault relay will not provide adequate protection for winding earth faults. Even with a biased differential relay, the biasing de-sensitises the relay such that it is not effective for certain earth faults within the winding. This is especially so if the transformer is resistance or impedance earthed, where the current available on an internal fault is disproportionally low. In these circumstances it is often necessary to add some form of separate earth fault protection. The degree of earth fault protection is very much improved by the application of REF-systems. Both windings of a transformer can be protected separately with restricted earth fault, thereby providing high speed protection against earth faults over virtually the whole of the transformer winding.

Choice:
The degree of protection provided depends to a great extent upon the size and importance of the unit been protected. I have not seen it on units smaller than 10MVA, but it seems to be used in the UK on much smaller units.
A further important factor is the economic aspect or cost aspect. (Rule of thump: 10% of transformer cost for protection)
The views of the designer could also be a factor to a lesser extend.

Regards
Ralph

 

[stevenal]

One point to add.
With some microprocessor transformer relays, REF can applied at essentially zero cost; just enable it in the device. (assuming enough Cts are present)

 

[Rodmcm]

If you have a standard start/delta winding then the only protection of an earth fault on the 433V mains is from the 11kV side which will be slow, hence the inclusion of a Ct on the LV side of the transformer and a ref protection speeds up the clearance times. Generally we have found that for transformers under 1MVA no one bothers but above this they seem to. I agree with stevenal that you usually get this option free in good electronic relays, a simple single CT on the transformer doesn't cost much either

 

[xabproject]

Dear electromott,

REF is about balancing the residual current of the phase CTs against the current of neutral CT. As one approaches the distribution level, the current transformer ratios become 'numerically' higher leading to 'less effective' translation of current changes/unbalances. At the same time the unbalance in the phases becomes more pronounced at the distribution level and therefore a residual current almost always exists, and together with the inability of the CTs to reflect small current changes/unbalances effectively, the REF even when applied has to be adequately desensitized to avoid spurious trips, defeating the very purpose of it's application as a sensitive protection! A set of dedicated Residually connected CTs and/or Neutral Current CTs on the LV side provide adequate protection.

Therefore you may not find dedicated REF on the 433V side of the transformer.

I agree with RalphChristie that the considerations are more commercial than technical. It is very rarely seen that a 11/0.433kV Distribution Transformer is protected with REF on the LV side. But as stevenal points out, the numerical type transformer relays come with many protection functions/features and a dedicated REF can be applied at 'essentially' zero cost provided there are adequate number of CTs. Well, this is may be called as the modern trend!

best regards,

 

[RalphChristie]

For some info on REF-schemes: (copy and paste whole address)

http://www.areva-td.com/servlet/Con...uct_Detail&c=AGF_Product&cid=1089880316992&am

p;rid=&lid=en&pid=1017999014820&tab=Chapters&id=1056536208254

download chapter 16 (Transformer and transformer feeder protection)


An example of stevenal's recomendation:

http://www.esac.com/Downloads/SEL387E Brochure.pdf

 

[electromott]

Thanks xabproject

But what do you mean when you say approaching the Distribution Level?

 

[xabproject]

Dear electromott,
I meant distribution level 'voltage', typically the 433V.
best regards,

 

[jghrist]

Until recently at least, REF has been uncommon for utility distribution substation power transformers in the US. These are generally delta-solidly grounded wye with substantial Ø-grd fault current on the low side. The amount of winding unprotected by transformer differential is low (is any unprotected?) with solid grounding. With REF being essentially free on modern relays, should we start applying the protection? How much danger is there of through fault instability due to CT saturation? Normally there is a CT in the transformer neutral that is used for ground overcurrent protection on the low side. Would it compromise the REF protection to use the neutral CT for both overcurrent and REF?

 

[RalphChristie]

jghrist,

Read this paper:

http://www.selinc.com/techpprs/6092.pdf

(By the way, there is some good info at the SEL-site:

http://www.selinc.com/techpprs.htm

)


Now to answer some of your questions:
The amount of winding unprotected by transformer differential is low (is any unprotected?) with solid grounding.
It depends on the differential setting.
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 up to 90% of the winding. REF-schemes protect transformers especially against 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)

With REF being essentially free on modern relays, should we start applying the protection?
I can't see why not. If you can apply diff-protection, why not one additional CT for REF-protection also? That is if you have a relay with all these functions. If not, I would not add it, (for a distribution transformer anyway) but it depends on the size of the unit and the availability of a new unit, if the old one fails. Normally distribution transformers are easily obtainable, so one can replace a failed unit fairly easy.

How much danger is there of through fault instability due to CT saturation?
On this one I am not sure. Maybe someone else in this forum can answer this, but I think the possibility is higher than high-impedance REF-schemes.

Would it compromise the REF protection to use the neutral CT for both overcurrent and REF?
In some of the relays you can use only one CT in the neutral for ground overcurrent protection (we call it standby earth fault protection) and REF-protection. You can even use just three CTs on each side of the transformer (total of six) to do metering, overcurrent protection and differential protection. However, in my opinion, it is not a good idea. If, for some reason, you loose the relay, you have no protection at all. I would use more than one relay to do all the functions.

Regards
Ralph

 

[jghrist]

Ralph,

Are your unprotected percentages for a solidly grounded winding? Do you have any references that show how to calculate the unprotected percentage?

Figure 5 in the SEL paper you linked would seem to indicate a much higher percentage of the winding covered, but I have a problem with this curve from an intuitive sense. It shows the neutral current approaching a very high value as the faulted % of winding approaches zero and the primary current approaching about 20% of nominal. It is intuitively obvious to me, however, that when the fault is at the grounded neutral point of the winding (0%), there will be no primary or neutral current except for load and magnetizing current. I would think the currents would both approach zero for a no-load condition.

 

[stevenal]

Jim,

Think flux linkages. Short out the the first turn of the secondary winding nearest the neutral point and high current will flow in this turn, since the flux will be linking to this turn. Assuming this fault is to ground, high neutral current will be measured by the neutral CT. The secondary # of turns has been reduced to one in this case, so high current will not show up on the primary CT.

For the same reason, you cannot tap a multi-ratio CT more than once. No connecting one device to X1-X2 and another to X4-X5 because most of the flux will link to the lower number of turns.

 

[jghrist]

stevenal,

Now I see! OK, I'm convinced that applying REF is a good thing even in a solidly grounded Dyn transformer if there are provisions to prevent tripping on residual CT currents for a heavy external 3Ø fault. The SEL-387 provides this by requiring the zero-seq current to be a set fraction of the pos-seq current. You can get more security by applying an inverse time curve.

I don't feel too bad about not applying REF earlier. For solidly grounded windings, the differential covers most of the winding and for faults close to the neutral, the neutral current is high enough to trip a 51G relay connected to the neutral CT. The 51G has to coordinate with downstream 51 relays, but for high fault currents, it would still be fairly fast. According to Fig 16.3 in the AREVA PRAG (

http://www.areva-td.com/servlet/Con...lid=en&tab=Chapters&id=1056536208254#spare2),

the neutral current for a fault near the neutral would exceed the maximum Ø-g through fault current (100% distance from neutral).

 

[RalphChristie]

Jim

You are correct, this percentages are for a star winding earthed through a resistance.
I found the calculations in:Protection of Electricity Distribution Networks, 2nd Edition by Juan Gers.
Only after you replied, I noted the percentages were for resistance earthed systems.

For a solidly star winding the percentage winding unprotected (for earth faults) with a differential relay will be much less. There will be a small percentage not covered, I always think of such faults to be similar to the working priciple of an autotransformer.

On distribution level, I won't bother to much with REF protection. Like I said in a earlier post, I have not seen it been used on transformers smaller than 10MVA.
But, if you use diff protection on a transformer, a REF-scheme will just add a few more elements. (even on older schemes) Diff and REF-schemes share many times CTs, wires, etc.

The biggest advantage REF-schemes have over the 51G-relay connected in the neutral is the clearing time for a fault. But on this level it is maybe unimportant.

Regards
Ralph