EEE8049 Protection of Distribution Networks – Taster Material Unit protection; some initial concepts As we have seen, with overcurrent protection clearance times tend to rise with fault levels as we move closer to source transformers. On the higher voltage side of such transformers the prospective fault levels are generally higher than on the lower voltage side, but the fault currents referred from the lower voltage side are lower. Hence the fault currents referred from the lower voltage side have proportionately less effect on the higher voltage system. On the higher voltage distribution systems (typically 33kV) quite often two or more circuits normally operate in parallel, although full load can generally be supplied for the loss of one circuit (redundancy). Sometimes directional overcurrent protection is used to achieve fault discrimination with such circuits normally operated in parallel. However, as we move further back up the supply system we eventually come to the transmission system where very quick fault clearance is required on such a system to avoid problems with generator and system stability. Such considerations preclude the rather leisurely clearance times associated with IDMT overcurrent protection. Additionally, on the transmission system, which comprises a mesh of circuits, power flow direction in a circuit may change on a diurnal to seasonal basis, with abnormal flows occurring in response to maintenance or fault outages elsewhere on the system. This effectively precludes directional overcurrent protection and duplicate, fast main protection (preferably using different measurement principles) is generally used. Where fitted, IDMT overcurrent protection is used as a third, “last ditch,” protection to try to save transmission (and subtransmission – 132kV) circuits from irreparable damage for catastrophic failures, such as the failure of a breaker to trip. The concept is that fault current contributions from generation all over the system merge to form larger fault currents in the vicinity of the fault. Thus, IDMT overcurrent protection with uniform settings at various locations should result in other circuit breakers close to the fault tripping ahead of more remote circuit breakers. However, stability of parts of the supply system would almost certainly be lost and widescale loss of supplies would be likely. So, if IDMT protection is unsuitable for transmission and sub-transmission circuits it follows that other protection principles must be used. Generally these are based on either unit protection principles, or impedance measurement techniques (Distance or Mho protection). The concept on which unit protection schemes are based is appealing in its simplicity: establish a protected zone delineated by CTs and compare the current entering the zone with the corresponding current leaving the zone. Providing they match the protected zone is healthy. If there is a discrepancy there must be a fault, surely? As with many things in life the answer tends to be a little more complicated, and there can be legitimate discrepancies, so the question can be recast as to how much discrepancy do we allow before declaring a fault? A couple of examples will illustrate the point. Suppose we measure the currents entering and leaving a 33kV cable connected to a resistance earthed system. If the cable circuit is energised from one end, with an open circuit breaker at the other end, we would expect to measure the cable charging currents (= μF / km x circuit km x 10-6 x (33 / √3) x 103 x102π Amps at 50Hz) at the source end, but at the open end the current would be zero – a legitimate discrepancy! For an earth fault elsewhere on the same 33kV system, our cable protected by a unit protection scheme looking at earth fault components would see a virtual earth fault current of up to three time the magnitude of the normal cable charging current – a worse legitimate discrepancy! If the circuit breakers at both ends of the circuit are closed load current will flow, but with these two discrepancies superimposed on the load current. Now consider a 30MVA 132 / 11kV YNdyn0 30%Z transformer with a tapping range extending from 110% to 80% of the nominal 132kV winding by 1.67% taps, as shown in Fig. 1. The transformer has a delta winding merely to largely decouple the (solidly earthed) 132kV neutral zero sequence network from the resistance earthed 11kV neutral zero sequence network, and so under normal operation the delta winding can be ignored. If we fit 200/1A CTs to measure the phase currents at 132kV and 2400/1A CTs to measure the corresponding phase currents at 11kV we would expect there to be very little discrepancy between corresponding CTs (other than the very small transformer magnetising current) under normal operation, providing the transformer is operating at nominal ratio. However, it is normal to have an automatic voltage control scheme maintaining the 11kV output voltage as the transformer is progressively loaded. The internal voltage drop within the transformer is a function of the load current, its 11kV power factor and the transformer impedance, but the essential point is that the transformer can be operating a long way off nominal ratio. This off nominal operation will result in a legitimate discrepancy between the corresponding CTs. Now assume that the transformer is heavily loaded, and thus operating significantly off nominal ratio. Then, if a through fault occurs (not within the transformer, but on the downstream 11kV network) the large flow of fault current through the transformer will result in a much larger, but legitimate discrepancy between the corresponding CTs. Figure 1 132/11kV 30MVA transformer, equipped with 132kV and 11kV Restricted Earth Fault relays and transformer overall differential protection relays Although this is bad enough there is a still worse condition. Consider what happens when our transformer is energised from the 132kV network. There is potentially a huge transient magnetising inrush current of up to 7 x the normal load current, but with no 11kV output current (assuming the 11kV breaker is open). This large, but transient, legitimate discrepancy is bigger than a 11kV phase to phase fault within the transformer. One possible solution to this problem would be to use a fixed time delay in conjunction with the transformer overall differential protection relays in order that the transient discrepancy could decay. However the whole point of using unit protection schemes is that operation should be fast (for application to the transmission system) as a simple decision must be made as to whether or not a fault exists within the zone of the unit protection. So adding significant time delay to unit protection is not an acceptable solution! Thus we see that the original, simple concept of establishing a protected zone delineated by CTs and comparing the current entering the zone with the corresponding current leaving the zone needs modifications to cater for various types of discrepancy. There are various methods of dealing with the legitimate discrepancies outlined above, plus other measurement problems, and these will be discussed later. The thing that distinguishes the different types of unit protection scheme is that each has a specific application, with an associated set of legitimate discrepancies and measurement errors. Different techniques to deal with such discrepancies and measurement errors are used by the various types of unit protection scheme. discrepancies assuming perfect measurement CTs. So far we have dealt with However, these CTs can themselves introduce other errors which result in further discrepancies.