Residual current is one of the means to check the presence of zero sequence current. When ever earth fault take place in a grounded system, zero current flows in system. The presence of this zero sequence current can be sensed by connecting an earth fault relay in residual connection of CTs.
Are you referring to the current transformers used for sensing the condition?
A residual current earth fault relay would be connected to the star point (residual) circuit of a set of 3-phase CTs, as RAgrawal describes. This has limitations as to the sensitivity of the earth fault detection, as the CT ratio is based on the phase currents and the 3 CTs will not be perfectly matched. Bottom line is that you can't set an earth fault relay below say 10% of the rated current with this connection.
A zero sequence CT is one that encloses all 3 phase conductors. This can be selected to sense whatever level of earth fault current you wish, down to the order of 1 A primary.
Residual current is a measured quantity (measured by CT or CT's). It is the vectorial sum of the three phase currents and is the actual earth fault current. The residual current is equal to three times the zero sequence current.
Zero sequence current is a calculated value that you get when you do short circuit analysis of a network.
Some relays can calculate the residual current from the three phase current measurement.
Suggestion: Reference:
1. IEEE Std 100-2000 The Authoritative Dictionary of IEEE Standards Terms, 7th Ed, IEEE Press, 2000
Residual Current (Protective Relaying). The sum of the three-phase currents on a three-phase circuit. The current that flows in the neutral return circuit of the three wye-connected current transformers is the residual current.
Zero Sequence Current Io is:
Io = Iao = Ibo = Ico
Iao, Ibo, Ico, Io are currents forming a set of rotating phasors that are always in phase and always equal in magnitude. If any of them exists, it exists equally in all three phases and never alone in one phase. The CTs neutral for the ground relay has 3Io=Ia+Ib+Ic
where Ia, Ib, and Ic are line (phase) currents.
Slightly off topic, but peterb may be interested to know that we routinely use the residual connection of 400/1 CTs generally around 10P20 or 10P50 class for sensitive earth fault relaying to 2% (8A primary). The tripping time is 10 sec definite time. Some, but not all, are checked by a CT on the transformer neutral. But I am not claiming it is entirely free of interesting funnies!
You're right, Bung, I am interested to hear about your application. I'm also interested to hear that it's not entirely free of "interesting funnies". Those funnies are the reason why traditional schemes don't try to push the envelope too far.
Perhaps what I should have said was "Bottom line is that the minimum practical setting for secure operation is around 10% of the rated current with this connection, using conventional ground relaying". I still feel that any lower setting needs a core balance CT for security, although your 10 second delay does take a lot of the uncertainty out of the equation.
Gordonl - one reference is the GEC Measurements Protective Relaying Application Guide, which is an overall great book if you can get your hand on a copy of it. A strong hint is the fact that the minimum setting current on relays such as the SEL501 and the overcurrent elements of the SEL321-1, SEL551 and SEL587 is 0.5A for the 5A model, 0.1A for the 1A model.
Sensitive ground fault relay elements are certainly available, but I believe that they are primarily intended for use with core balance CTs, NOT in the residual circuit of the phase CTs. An example of this is the GE Multilin 750 relay. The manual for this states, relative to sensitive ground elements, that "The sensitive ground current input can be connected to a zero sequence CT for increased sensitivity and accuracy when physically possible in the system." They do show the possibility of a residual connection, but note the comment re sensitivity & accuracy.
I am presently in the process of specifying a high-resistance grounded 2.4 KV system, where I will be using sensitive ground fault elements of the motor protection relays, with core balance CTs, to give an alarm for a ground fault level of 1A primary - I don't think that I could do that reliably with 300/5A phase CTs.
The biggest problem we have had is using older electro-mech SEF relays that did not have any kind of harmonic restraint. Switch on of a 33/11 transformer in the substation could sometimes cause random tripping of 11kV feeder SEF relays. There is a bit of literature around on this subject, but I don't know where you might find it. We cure this little problem on an ad hoc basis by replacing the e-m relay with a harmonic filtered electronic relay (from RMS in Australia) as it happens. We have around 600 of the older e-m relays in service, and we have to fix up 1 or 2 a year because of this problem - it isn't a major deal really.
Another problem (not related to CT configuration) is the instantaneous reset of most SEF relays with definite time settings. Some conductor down faults result in bursts of current as the conductor fizzes and pops on the ground, so lots of "false starts" on the timer, and you may not get a full 10 sec burst to take the relay to trip! On one particular case, the times between pickup varied randomly between <1 sec and 74 (seventy-four) seconds. The relay never operated - it was tripped by Ops via SCADA.
Was there a particular area in the book where this was discussed, 9.16 they seem to refering to changes in burden of electromechanicals. (I have a copy published by Alstom in 95)
Certainly I agree that a rsidual connection is not advisable for an ungrounded or high resistance grounded system.
The sensitive ground input on the 750 is intended for ungrounded systems, but as a matter of interest the internally calculated neutral current (ground current on three phase system) removes the harmonics and dc offset for the neutral overcurrent settings. Which would go to support the 10% rule of thumb when it isn't used.
