Apps for Low-Voltage Network Scenarios

• 1: Background on Network-Protector Relaying
• 2: Network Backfeeds for an Open Feeder
• 3: Network-Protector Operations with a Primary-Side Blown Fuse
• 3.1: Background on Network-Protector Operations with a Primary-Side Blown Fuse
• 4: Crossed or Rolled Phases on the Primary

1 - Background on Network-Protector Relaying

Historic Perspective for Network Protector Relaying

With reference to Figure, the first non-dedicated feeder applications, where the two-unit spot network was fed from overhead multi-grounded neutral feeders were made in the 1960’s when the only relays available for network protectors were the electromechanical type, such as the Westinghouse CN-33, and the General Electric CHN. Studies were performed simulating the most common type of fault on the cable circuit to the spot network transformers, the single-line-to-ground fault with the blown fuse in just the faulted phases. These studies showed that the standard CN-33 relay or the GE CHN relay would not reliably detect the SLG fault with the blown fuse in the faulted phase. Important parameters affecting the relay response were the X to R ratio of the network transformers, and the loading on the spot network paralleling bus at the time of the SLG fault with blown fuse. And if the relay does not detect the SLG fault with blown fuse in the faulted phase, then the fuses in the network protector associated with the faulted phase can blow. But in a two-unit spot network, the fuse associate with the faulted phase in both protectors could blow, creating a single phase condition for the spot network, thereby defeating the purpose of the spot network which is to provide reliable service to the load for faults on the primary system.

Manufacturers and utility engineers looked at what could be done to the standard electromechanical network relays to allow them to reliably detect the SLG fault with blown fuse in the faulted phase. With the Westinghouse CN-33 relay, it was determined that by changing the connections to the relay current coils from the protector current transformers, a trip characteristic was developed that would reliably detect the SLG fault with blown fuse. The relay with the revised connections from the protector CT’s to the relay current coils was referred to as the “watt-var” relay. Figure 1 shows the CN-33 relay standard (“watt”) trip characteristics under balanced three-phase conditions.

Although the “watt-var” connections to the CN-33 relay would reliably detect the SLG fault with blown fuse, there were some short comings to the connections. Figure 2 shows the CN-33 relay trip characteristics under balanced three-phase conditions when the watt-var connections are made. First, the protector would trip for leading power factor network loads if the power factor were less than 86%. Second, if the relay with “watt-var”connections were used in a dedicated feeder network, where he backfeed could be capacitive if the feeder breaker were opened in absence of a fault, the “watt-var” relay would not detect the capacitive backfeed.

Although the characteristics shown in Figure 1 and in Figure 2 apply to the Westinghouse electro-mechanical network protector relays, there was a similar arrangement for the General Electric electro-mechanical relays that allowed them to reliably detect the SLG fault on the primary feeder with a blown fuse in the faulted phase.

In today’s microprocessor relays for network protectors, some relays allow adjusting the angles of the trip curve relative to the reference phasor. Considered in these apps are three trip algorithms that can be used for a microprocessor relay.

Microprocessor Relay Trip Algorithms

Three trip algorithms are included in the apps for network microprocessor relays. The first is the positive-sequence directional overcurrent trip characteristics. The other two are power based trip algorithms.

Positive-Sequence Directional Overcurrent Trip Characteristic

The positive-sequence directional overcurrent trip characteristic is shown in Figure 3. The relay extracts the positive-sequence component of the network line-to-ground voltage, shown as V_1N in Figure 3. This is the reference for the relay trip characteristic. The relay also calculates the positive-sequence component of the three phase currents in the network protector, designated as I₁. The angle between the positive-sequence current, I₁, in the network protector and the positive-sequence component of the network line to ground voltage is θ_1rly as shown in Figure 3.

The relay trip curve consists of two segments, one in quadrant 2 identified as TC2, and the other in quadrant 3 and 4 identified as TC1. The positive-sequence current at 180 degrees required to make the relay trip contact is equal to the protector current transformer rating times the 180 degree trip setting in percent of CT rating, divided by 100. Both TC2 and TC1 pass thru this point.

The angle of TC2 is defined by angle θ_SH2, and the angle of TC1 is defined by angle θ_SH1, where the positive direction for both angles is counter-clockwise as shown in Figure 3. In general, the current in the protector associated with the faulted primary feeder lies in quadrant 2, and the positive-sequence current in the protector associated with the unfaulted feeder lies in quadrant 3 or quadrants 3 and 4.

In Figure 3, current I_1MAG is the magnitude of the current in the protector associated with the faulted feeder as calculated by the apps. The apps then calculate the shift angle θ_SH2, at which the positive-sequence current having magnitude I_1MAG will just intercept trip curve 2. For the relay in the protector to reliably detect the fault, the shift angle for TC2 should be less than θ_SH2 to allow for tolerances and errors in the relay.

Similarly, the apps calculate for the network protector associated with the unfaulted primary feeder the angle of trip curve segment TC1, angle θ_SH1 where the positive-sequence current will just intercept trip curve segment TC1. So that the protector associated with the unfaulted primary feeder does not trip, the shift angle for TC1, must be less than the calculated value of θ_SH1.

For the Eaton MPCV relay, when the watt trip characteristic is selected, angle θ_SH2 is fixed at -5 degrees, and angle θ_SH1 is fixed at +5 degrees. For the MPCV relay when the watt-var trip characteristic is selected, angle θ_SH2 is at -60 degrees, and angle θ_SH1 is fixed at +5 degrees. The apps give angle θ_SH2 and angle θ_SH1, from which it can be determined if the MPCV relay in the protector connected to the faulted feeder, and in the protector connected to the unfaulted feeder will respond correctly.

The DigitalGrid network relay can be programed to have the positive-sequence directional overcurrent trip characteristic as shown in Figure 3. With this relay, settings can be made, referred to as “trip tilt angle” and “trim angle” that allow the user in effect to set shift angles θ_SH2 and θ_SH1 to any desired value. The apps provide the information needed for the user to set the DigitalGrid network relay so that the protector associated with the faulted feeder will trip, and the network protector associated with the unfaulted feeder will not trip.

Power Based Trip Algorithm Using Network Actual Line-to-Ground Voltage Magnitudes

The first power-based trip characteristic is shown in Figure 4, where the horizontal axis is the real power flow in the network protector, and the vertical axis is the reactive flow in the network protector. The trip curve also consists of two straight line segments, one in the second quadrant, segment TC2 having shift angle θ_SH2, and one in the third and fourth quadrant, TC1 having shift angle θ_SH1. Both segments pass through the point having a P value of P_SET, and a Q value of zero as shown. P_SET is calculated using the nominal line-to-ground voltage, typically either 120-volts or 277 volts, and the relay reverse current trip setting in percent of current transformer rating, RCT_%, and the current transformer primary current rating, CT.

The apps calculate in each network protector for the simultaneous fault the net three-phase power flow, P_NET, which is the sum of the power flows in the three phases of the protector, and it calculates the net three-phase reactive flow in the protector, Q_NET, which is the sum of the reactive flows in the three phases of the network protector. If this point lies on or to the left of the relays sensitive trip curve shown in Figure 4, the sensitive trip criteria is satisfied.

In the network protector associated with the faulted HV feeder, the P-Q point lies in either the first or second quadrant in Figure 4. And in the network protector associated with the unfaulted HV feeder, the P-Q point lies in the fourth or third quadrant in Figure 4.

Given P_NET, Q_NET, and P_SET, for the protector associated with the faulted HV feeder, the apps calculate the shift angle θ_SH2 such that the P-Q point just lies on the trip curve segment TC2. For the relay in the protector associated with the faulted primary feeder, the actual shift angle must be less than calculated angle θ_SH2 to assure that the P-Q point will lie in the trip region. Similarly, in the network protector associated with the unfaulted HV feeder, the apps calculate shift angle θ_SH1 for trip curve segment TC1 shown in Figure 4 where the P-Q point just lies on the trip curve. To prevent the protector associated with the unfaulted HV feeder from tripping, the shift angle for TC1 for the protector on the unfaulted feeder must be less than the calculated value of angle θ_SH1 in Figure 4 to assure the protector does not trip.

Power Based Trip Algorithm Using Network Nominal Line-to-Ground Voltage Magnitude

The second power-based trip algorithm in the apps is very similar to the first, in that the P-Q point is calculated using for each phase the nominal line-to-ground voltage, the phase current, and the angle between the actual network line-to-ground voltage and the line current. This is done for each phase, and the net P and Q for the three phases, designated P_NETnom and Q_NETnom is calculated. This method for calculating the P-Q point is described in the instruction book for the SEL 632-1 network relay.

In the protector associated with the faulted HV feeder, the P_NETnom-Q_NETnom point normally lies in the first or second quadrant, and in the protector supplied from the unfaulted primary feeder the P_NETnom-Q_NETnom point normally lies in the fourth or third quadrant.

For the network protector associated with the faulted feeder, the apps calculate the shift angle, θ_TC2, for trip curve segment TC2 where the PQ point will just lie on the trip curve. For the relay to reliably detect the fault, the shift angle for TC2 should be less than the calculated value for θ_TC2, to allow for tolerances and uncertainties in the system data used for the analysis.

For the network protector associated with the unfaulted primary feeder, the apps calculate the shift angle for trip curve segment TC1, angle θ_TC1, where the P-Q point just lies on trip curve segment TC1. To assure that the protector associated with the unfaulted feeder does not trip, the shift angle for TC1 should be less than the calculated value for θ_TC1, to account for tolerances and uncertainties in the system data used for the calculation of the currents and voltages in the network protectors.

Other Trip Algorithms for Microprocessor Network Protector Relays

These apps for simulating simultaneous fault and blown fuse conditions solve for the currents and voltages in the network protectors in the two-unit spot networks. The apps then use these currents and voltages to evaluate the response of microprocessor network protector relays that have the trip algorithms discussed in the preceding sections. In particular, the apps find the trip curve shift angle for the protector on the faulted feeder to assure that it will reliably detect the fault. And the apps determine the trip curve angle at which the protector associated with the unfaulted primary feeder will trip.

Since the apps give the protector currents and voltages, the user can use these to evaluate the performance of other microprocessor network protector relays if the relay manufacturer provides the algorithm that his relay uses to make the trip decision. The phase voltage and current magnitudes, along with their angles, can be used to evaluate the relays performance.

2 - Network Backfeeds for an Open Feeder

Purpose: This app can be used to model the expected currents and voltages at a back feeding network protector onto a medium-voltage feeder which has locked out at the substation, and with all other protectors on the feeder opened, including the case where the MV feeder has locked out because of a MV fault. The app can be used to examine and determine network protector relay trip characteristics and settings to assure that the protector opens on backfeed in these conditions. The app can also be used to examine system conditions and predict system behaviors in a case where the protector fails to open on backfeed, such as because of a mechanical failure. Under these conditions, network protector fuses may blow, which can be simulated with the app.

Background: Network protectors which are properly maintained and “set” are designed to open when a medium voltage (MV) feeder at the substation locks out, either with or without a fault on the MV feeder. They do this by detecting the backfeed from the secondary, and then open, preventing backfeed onto the MV feeder or into a cable fault. If the network protector fails to open on backfeed, either because of improper settings or other issues, this can result in varied conditions depending on factors such as the type of fault, system voltage, the stiffness of the LV network at the backfeed location, and the amount of cable charging on the primary circuit. This app enables exploration of various scenarios associated with backfeed through a network protector onto an open faulted or non faulted feeder. The fault types considered on the primary feeder are the single line-to-ground and the double line-to-ground faults. One app also simulates backfeed to an open feeder breaker with no fault on the primary feeder. The effects of blown fuses in the backfeeding protector can be examined by specifying the protector phase that has the blown fuse. The app applies to systems where the network transformers have the delta connected HV windings and the wye connected LV windings.

If there is a backfeed to a single line-to-ground fault when the primary cable charging kvar is high, overvoltages may occur in the secondary network at the backfeed location, that can damage customer loads or electronic components in the network protector. The apps allow examining the effect of shunt reactors placed on the primary feeder for limiting the voltage rise in the secondary network to acceptable levels.

In the network diagram below, hover over network elements to see detailed results.

Results

Results applicable to three different relay characterics are shown. For more background, see here.

Positive-Sequence Directional Relay Response

For this trip characteristic, the positive-sequence component of the network voltages, V_1N, is the reference at zero degrees. The sensitive trip setting, RCT_% is the positive-sequence current at 180°, in percent of CT rating, required to make the trip contact. The position of the sensitive trip curve in quadrant 2 is defined by shift angle θ_SH2, and in quadrant 3 and 4 by shift angle θ_SH1, with the positive direction being counter clockwise for both shift angles. See Figure 1.

To ensure that NWP 2 trips, its relay shift angle θ_SH2 must be less than the angle θ_SH given above.

To ensure that NWP 1 does not trip, its relay shift angle θ_SH1 must be less than the angle θ_SH given above.

In some microprocessor relays settings can be made such that angle θ_SH2 and angle θ_SH1 are adjustable and different. In the Eaton MPCV relay, angle θ_SH2 is fixed at -5 degrees, and angle θ_SH1 is fixed at +5 degrees, which gives the “gull wing” trip characteristic.

Power-Based Relay Responses

The first power-based relay response uses net real and net reactive power to determine tripping. Positive reactive power means a flow into an inductive load. The trip curve consists of two straight-line segments, one in quadrant 2, having shift angle θ_SH2, and the second in quadrant 3-4, having shift angle θ_SH1, with the positive direction for both shift angles being counter clockwise. See Figure 2.

If the point defined by P_NET and Q_NET lies on or to the left of the sensitive trip curve in Figure 2, the sensitive trip characteristic is satisfied.

To ensure that NWP 2 trips, its relay shift angle θ_SH2 must be less than the angle θ_SH given above.

To ensure that NWP 1 does not trip, its relay shift angle θ_SH1 must be less than the angle θ_SH given above.

The second power-based relay response calculates P_NET-nominal and Q_NET-nominal using the nominal voltage magnitude instead of actual voltages. The tripping criteria is the same after that. This is the response used by the SEL 632 relay.

Other Notes

This app can be used to show the following:

1. If the primary feeder cable charging is modest, as in most 15-kV class systems, the backfeed currents in the network protector are usually not high enough to blow the fuses in the backfeeding network protector should it fail to open. Further the phase-to-ground voltages at the backfeeding network protector are not excessively high, and damage to customer load is not expected.

2. If the primary feeder cable charging is high, as encountered in systems operating at 23 kV, 27 kV, 33 kV, and 34.5 kV, excessive overvoltages can occur at the backfeeding network protector which can damage customer loads. However, by applying a shunt reactor on the primary feeder, whose zero-sequence impedance is less than its positive-sequence impedance, the overvoltages at the backfeeding network protector are prevented. This is indeed what some utilities do on their 23 kV, 27 kV, and 33 kV primary feeders.

3. The two parameters which have the greatest effect on the voltages at the backfeeding network protector are the stiffness of the LV network at the backfeed location, and the amount of cable charging on the primary circuit. Backfeed from a weak location in the network to a primary feeder with high cable charging can produce excessive overvoltages in the secondary system.

Here are some cases to try:

• 13.8-kV case with no fault
• 26-kV case with SLG-fault case and no reactor
• 26-kV case with SLG-fault case and a compensating reactor (no overvoltages)

3 - Network-Protector Operations with a Primary-Side Blown Fuse

This app calculates the system response for a fault on the cable side of the fuse for underground cable #2 that supplies network transformer T2. The fuse(s) in the faulted phase(s) are blown, but the fuses in unfaulted phases are not blown. For more background, see here.

Purpose: This app can be used to model the expected behavior of a spot network that is fed from non-dedicated network feeders, such as overhead feeders with radial load, when a fault occurs on the system. In these cases, the fault may result in a single phase tap fuse blowing (where the underground to the spot network takes off from the overhead system, for example). It is important that the network protector relay settings are established such that the protector will open on the faulted feeder in these conditions. This app enables the user to model this type of fault and determine appropriate network protector relay settings to assure that the protector will open.

Background: In a network system with dedicated feeders, the only protective devices on the primary feeder are circuit breakers at the substation. When a fault occurs on the primary feeder, the circuit breaker at the substation opens. Following this, the network relays in the network protector associated with the faulted feeder will detect the fault and open. This isolates the fault on the primary feeder from the load served from the secondary network, and the customers served from the secondary network do not experience a power outage.

In some cases, utilities will supply spot networks from overhead primary feeders. In these cases, the spot network may be supplied from fused taps from the overhead feeder. If a fault occurs on the cable circuit (the underground portion feeding the spot network), this could result in the fuse on the faulted phase blowing, rather than the substation breaker opening. If the protector relay does not have the correct trip characteristic and is not set properly, this condition could result in in the protector not opening, the blowing of network protector fuses, and a single phase condition for the load supplied by the spot network.

In the network diagram below, hover over network elements to see detailed results.

Results

Results applicable to three different relay characterics are shown. For more background, see here.

Positive-Sequence Directional Relay Response

For this trip characteristic, the positive-sequence component of the network voltages, V_1N, is the reference at zero degrees. The sensitive trip setting, RCT_% is the positive-sequence current at 180°, in percent of CT rating, required to make the trip contact. The position of the sensitive trip curve in quadrant 2 is defined by shift angle θ_SH2, and in quadrant 3 and 4 by shift angle θ_SH1, with the positive direction being counter clockwise for both shift angles. See Figure 1.

To ensure that NWP 2 trips, its relay shift angle θ_SH2 must be less than the angle θ_SH given above.

To ensure that NWP 1 does not trip, its relay shift angle θ_SH1 must be less than the angle θ_SH given above.

In some microprocessor relays settings can be made such that angle θ_SH2 and angle θ_SH1 are adjustable and different. In the Eaton MPCV relay, angle θ_SH2 is fixed at -5 degrees, and angle θ_SH1 is fixed at +5 degrees, which gives the “gull wing” trip characteristic.

Power-Based Relay Responses

The first power-based relay response uses net real and net reactive power to determine tripping. Positive reactive power means a flow into an inductive load. The trip curve consists of two straight-line segments, one in quadrant 2, having shift angle θ_SH2, and the second in quadrant 3-4, having shift angle θ_SH1, with the positive direction for both shift angles being counter clockwise. See Figure 2.

If the point defined by P_NET and Q_NET lies on or to the left of the sensitive trip curve in Figure 2, the sensitive trip characteristic is satisfied.

To ensure that NWP 2 trips, its relay shift angle θ_SH2 must be less than the angle θ_SH given above.

To ensure that NWP 1 does not trip, its relay shift angle θ_SH1 must be less than the angle θ_SH given above.

The second power-based relay response calculates P_NET-nominal and Q_NET-nominal using the nominal voltage magnitude instead of actual voltages. The tripping criteria is the same after that. This is the response used by the SEL 632 relay.

Other Notes

Here are some cases to try:

• wye-wye SLG-fault case
• delta-wye SLG-fault case
• delta-wye LL-fault case

3.1 - Background on Network-Protector Operations with a Primary-Side Blown Fuse

Figure 1 is a single-line diagram of a two-unit spot network where there are two feeders protected with phase and ground relays at the substation. Normally, in secondary network systems the only protective devices on the primary feeder are three-pole circuit breakers at the substation. When a fault occurs on the primary feeder, the circuit breaker at the substation opens. Following this, the network relays in the network protector associated with the faulted feeder will detect the fault, and the network protectors fed from the faulted feeder open. This then isolates the fault on the primary feeder from the load served from the secondary network, and the customers served from the secondary network do not experience a power outage.

In some situations, utilities have applied spot networks fed from overhead primary feeder in suburban area, for important loads such as airports, computer centers, hospitals and critical manufacturing facilities. The utility will supply the spot network from taps taken from the overhead feeder, which for most utilities is a three-phase four-wire multi-grounded neutral feeder, which has distribution transformers whose primary windings are connected from phase-to-neutral, and between phases.

The network transformers in the spot network, shown as a two-unit spot network in Figure 1, are fed through cable circuits, most always made with single-conductor cables with a flat-strap or multi-wire concentric neutral. If a fault occurs on the cable circuit, and if the cable circuit were connected to the overhead feeder without fuses, a fault on the cable circuit would result in tripping of the feeder breaker at the substation, and cause an outage to the customers served radially from the overhead feeder. To prevent this from happening, the cable circuits for the spot network are connected to the overhead feeder through fuses, either distribution fuse cutouts or a separate switch and fuses.

A similar situation can exist in a dedicated primary feeder network, if network transformers in the spot network are connected to the primary feeder through a three-phase switch with primary fuses. In such applications, the ampere rating of the fuse is typically less than the pickup of the phase relays for the primary feeder at the substation, and the fuses provide more sensitive protection for incipient faults in the HV or LV windings of the network transformers.

Whether the two-unit spot network in Figure 1 is fed from non-dedicated overhead primary feeders with radially connected distribution load, or from network dedicated primary feeders, when a fault occurs on the tap circuit to the network transformer on Feeder 2, all three phases do not open on the primary. For example, if a single line-to-ground fault occurs on the cable circuit, only the fuse in the faulted phase will blow. The fuses in the two unfaulted phases will not blow. It is important that the network protector relay in the protector fed from the faulted feeder with blown fuse will detect the fault, and open the network protector on the faulted feeder. If the relay does not detect the fault, and the protector does not open, then network protector fuses associated with the faulted phases will open. Furthermore, if the spot network is a two-unit spot network, and the protector doesn’t open, then fuses in both network protectors can blow, creating a single-phase condition for the load supplied from the spot network.

With reference to Figure 1, in the app the kVA rating of each network transformer in the spot network is specified, along with its impedance, X to R ratio of the impedance, and the rated voltage on the HV and LV side of the network transformer. Although the single line shows a two-unit spot network, it can be used to find the response in spot networks with three or four network transformers. This is accomplished by doubling or tripling the kVA rating of the network transformer supplied from primary feeder cable UG1 in Figure 1.

The other issue that impacts the system response for this scenario are the winding connections for the network transformers, either delta-wye or wye-wye. If the spot network is fed from non-dedicated multi-grounded neutral overhead primary feeders with line-to-neutral connected distribution transformers, the network transformers in the spot network should have the wye-wye winding connections. If the delta-wye connections are selected and a single line-to-ground fault (SLG) occurs on the overhead primary feeder, the voltage on the unfaulted primary phases rise up to full phase-to-phase voltage, and a 73% overvoltage is applied to the load supplied from the line-to-neutral connected distribution transformers on the unfaulted phase. On a 120-volt basis, the 120 volt load have 208 volts applied to them, and they will be damaged if a protector fails to open, or if the protector has time delay tripping.

If the primary feeder from the substation in Figure 1 are network dedicated primary feeders where all network transformers have their primary windings connected in delta, then the network transformers in the two-unit spot network can have the delta connected primary windings.

The app considers three fault types on the cable circuit to network transformer 2. The first is the single line-to-ground fault on any one phase, with a blown fuse in any one phase, which normally would be the same phase with the fault. The second is the ungrounded line-to-line fault between any two phases, with blown fuses in any two phases as specified in the app. However, with an ungrounded line-to-line fault, one fuse may blow before the other, so the app allows simulating the ungrounded line-to-line fault with a blown fuse in either one of the faulted phase, as specified by the user of the app. The third fault type is the double line-to-ground (DLG) fault on any two phases, with a blown fuse in any two phases, where the blown fuses would be in the two faulted phases. But with the DLG, the fuse in one faulted phase could blow first, so the app allows simulating the DLG fault with a blown fuse in just one of the faulted phases.

4 - Crossed or Rolled Phases on the Primary

Purpose: This app models system conditions and network protector behavior associated with a spot network in the situation where work was performed on the MV feeder, and phases were inadvertently rolled or crossed on the MV feeder prior to re-energizing the feeder.

Background: In a situation where phases are rolled or crossed on an energized primary feeder with the network protectors open, the network protectors will not auto close after the primary feeder has been energized. With some network protector relays, with the protector open and rolled or crossed phases on the primary feeder, the network relay will make its trip contact. However, when entire networks are dropped, and the system is then re-energized by simultaneous closing of all feeder breakers at the substation, the response of the network relays in protectors on the feeder with the rolled or crossed phases, and on the other feeder with correct phasing is needed. In some cases, protectors may be locked in the closed position as part of the network pick up strategy. This app can help engineers predict system conditions in the case of rolled or crossed phases when a de-energized network is picked up with rolled or crossed phases on one primary feeder.

In the network diagram below, hover over network elements to see detailed results.

Results

Results applicable to three different relay characterics are shown. For more background, see here.

Positive-Sequence Directional Relay Response

For this trip characteristic, the positive-sequence component of the network voltages, V_1N, is the reference at zero degrees. The sensitive trip setting, RCT_% is the positive-sequence current at 180°, in percent of CT rating, required to make the trip contact. The position of the sensitive trip curve in quadrant 2 is defined by shift angle θ_SH2, and in quadrant 3 and 4 by shift angle θ_SH1, with the positive direction being counter clockwise for both shift angles. See Figure 1.

To ensure that NWP 2 trips, its relay shift angle θ_SH2 must be less than the angle θ_SH given above.

To ensure that NWP 1 does not trip, its relay shift angle θ_SH1 must be less than the angle θ_SH given above.

In some microprocessor relays settings can be made such that angle θ_SH2 and angle θ_SH1 are adjustable and different. In the Eaton MPCV relay, angle θ_SH2 is fixed at -5 degrees, and angle θ_SH1 is fixed at +5 degrees, which gives the “gull wing” trip characteristic.

Power-Based Relay Responses

The first power-based relay response uses net real and net reactive power to determine tripping. Positive reactive power means a flow into an inductive load. The trip curve consists of two straight-line segments, one in quadrant 2, having shift angle θ_SH2, and the second in quadrant 3-4, having shift angle θ_SH1, with the positive direction for both shift angles being counter clockwise. See Figure 2.

If the point defined by P_NET and Q_NET lies on or to the left of the sensitive trip curve in Figure 2, the sensitive trip characteristic is satisfied.

To ensure that NWP 2 trips, its relay shift angle θ_SH2 must be less than the angle θ_SH given above.

To ensure that NWP 1 does not trip, its relay shift angle θ_SH1 must be less than the angle θ_SH given above.

The second power-based relay response calculates P_NET-nominal and Q_NET-nominal using the nominal voltage magnitude instead of actual voltages. The tripping criteria is the same after that. This is the response used by the SEL 632 relay. Note that there are some numerical issues with this approach if the voltages on any phases are nearly zero. This happens with the crossed phases with a delta-wye transformer.

Other Notes

If work is done on a primary feeder to replace a section of cable between two manholes, with all network protectors open, but if phases are rolled or crossed, when the feeder is re-energized by closing the feeder breaker at the substation, the network protectors will not auto close.

However, one way that the system could be energized with rolled or crossed phases in one feeder is if the system was de-energized (network dropped) and work had been done on a primary cable. If the system were then re-energized by simultaneous closing of all feeder breakers at the substation, the response of the network relays in protectors on the feeder with the rolled or crossed phases, and on the other feeder with correct phasing is needed.

Here are some cases to try:

• delta-wye crossed phases
• wye-wye crossed phases
• delta-wye rolled phases
• wye-wye rolled phases