Catastrophic Relay Misoperations and Successful Relay Operation Steve Turner (Beckwith Electric Co., Inc.) Introduction This paper provides detailed technical analysis of several catastrophic relay misoperations and demonstrates how to prevent them from occurring. It also provides an example of using data recorded by a relay during onerous conditions to implement a new protection scheme. An undesired overall differential relay operation occurred at a hydro power plant during an external fault on the low side of the auxiliary transformer. This was an extremely challenging case to analyze since there were no three-line diagrams and it was eventually discovered that one set of CTs was located inside the delta winding of the GSU. An unwanted breaker failure operation occurred that tripped a large generator during high load which resulted in an outage in the adjoining downtown area of a large city. A transformer differential trip protecting the step-up transformer at a process plant occurred due to sympathetic inrush when a large nearby GSU was energized via the interconnecting high voltage transmission line which resulted in an extended outage. An intermittent arcing ground fault on the stator windings of a large generator was not cleared and eventually led to flashover on all three phases; as a result, extensive damage occurred to the large generator. Another intermittent arcing ground fault on the stator windings of a large coal fired generator could not be detected by traditional stator ground fault protection. The ground was on a bushing at the terminals and had arced many times prior to the trip. The customer was able to determine the location of the fault and remedy the situation by implementing a high-speed arcing ground fault protection scheme. Each individual analysis ends with a conclusion which states why the relay misoperated and also provides a recommendation as to the best practice for the particular application. Case 1 Overall Differential Relay Operation at a Hydro Power Plant during External Fault on an Auxiliary Transformer Low Side Synopsis A phase-to-phase fault occurred within the 13.2 kv auxiliary system which was external to the overall differential scheme for generator 1 and resulted in a trip. This section shows why the protection operated and how to prevent it from re-occurring in the future.
Figure 1 shows the one-line diagram for this protection. Generator 1 is rated 200 MW. Note that the winding 1 CTs are internal to the delta-connected GSU transformer winding connected to the generator. 500 kv 10000:5 (wye) WDG1 16 kv Breaking Resistor G1 Zigzag 13.2 kv 10000:5 (wye) 10000:5 (wye) 300:5 (wye) WDG2 WDG3 WDG4 x Phase-to-Phase Fault Figure 1 Hydro Electric Plant Single Line Diagram (Generator 1) Existing Protection Settings
Fault Current Signals Figure 2A shows the oscillography captured by the relay at the time of the trip. Figure 2B shows the current phasors measured by the protective relay. Figure 2A Fault Event Oscillography
Figure 2B Fault Current Phasors Winding 1 CT Connections Figure 3 shows how it is believed that the CTs are connected to the winding 1 current inputs for the differential relay. This has not been confirmed by the customer. Zero-sequence current filtering is enabled to prevent unwanted operations since zerosequence current circulates inside the delta winding during external ground faults within the transmission grid. I C I A I B I C I A /2 A B C I A I B I A - I C I B - I A I C - I B Figure 3 Winding 1 Internal CT Connections Differential Operating Equations Figure 4A shows how the differential protection eliminates zero-sequence current from the winding 1 current inputs.
Figure 4A Winding 1 Zero-Sequence Current Elimination Figure 4B shows how the differential protection compensates for the 30 o phase shift across the GSU. Figure 4B Input Current Compensation Figure 4C shows how the differential protection calculates the differential and restraint current for each phase. Figure 4C Differential and Restraint Current Calculations Non-Fault Balanced Load Current Figure 5A shows balanced load current measured by the protective relay during a normal non-fault load condition.
Figure 5A Current Phasors during Normal Load Flow (amps secondary) Figure 5B shows the resultant differential and restraint currents calculated by the relay. Figure 5B Protective Relay Secondary Metering Figure 5C shows the corresponding current phasors measured by the relay.
Figure 5C Current Phasors measured by Protective Relay (amps secondary) Figure 5D shows the calculated differential and restraint currents which match the values displayed in the relay secondary metering. Note that these values are in per unit. Figure 5D Differential and Restraint Current Calculations (per unit) Fault Current and Trip Figure 6A shows the fault currents measured by the relay when the trip occurred due to the external phase-to-phase fault.
Figure 6A Fault Current Phasors (amps secondary) Figure 6B shows the calculated differential and restraint current when the trip occurred. Figure 6B Differential and Restraint Current (per unit) Figure 7 shows the operating points for each phase mapped onto the differential operating characteristic. All three phases are in the operate region. 0.8 0.6 0.4 0.2 0.7 0.5 0.3 0.1 A & C Bias (P.U.) Figure 7 Differential Operating Characteristic
Fault Current and Non-Trip The solution is to roll the winding 4 current inputs as shown in Figure 8. Figure 8 Correct Relay Setup Figure 9 shows the new fault current measurements for winding 4 due to the correction. Figure 9 Correct Winding 4 Current Measurements (amps secondary) Figure 10 shows the corrected calculated differential and restraint current when the trip occurred. There is now minimal differential current and the relay does not trip. Figure 10 Differential and Restraint Current (per unit) 0.8 0.6 0.4 0.2 0.9 0.7 0.5 0.3 0.1 0.1 A & C 0.3 B 0.5 0.7 0.9 1.1 Bias (P.U.) 0.2 0.4 0.6 0.8 1.0 Figure 11 Corrected Differential Operating Characteristic
1 st Conclusion The differential protection tripped because of an incorrect CT phase compensation setting for winding 4 current inputs. The best practice is to not include the auxiliary transformer inside the zone of overall differential protection as shown in Figure 12. The auxiliary transformer should be protected by a separate differential relay as denoted by the green CTs that overlap with the overall differential protection. 500 kv 10000:5 (wye) WDG1 16 kv Breaking Resistor 10000:5 (wye) G1 Zigzag 13.2 kv 10000:5 (wye) 10000:5 (wye) WDG2 WDG3 WDG4 x Phase-to-Phase Fault Figure 12 Hydro Electric Plant Single Line Diagram (Generator 1) Case 2 Unwanted Breaker Failure Operation - Large Generator Tripped during High Load Period in Downtown Area of Large City Synopsis A large generator located in the downtown area of a large city was offline and a breaker failure trip occurred during a period of high load disrupting service. Figure 13 shows the system topology at the time of the trip. Note that the generator is connected to the transmission grid via an HV breaker. The low-side winding of the GSU drew excitation current since it was energized via the auxiliary station service.
0 Generator Transformer Excitation Current GSU High Side Breaker Auxiliary Station Service Original Breaker Failure Scheme Logic Figure 13 System Operating Conditions 50BF BFI (Input 4) 0 BF Trip Original Protection Settings Fault Current Signals Figure 14A shows the oscillography captured by the relay at the time of the trip. Figure 14B shows the current phasors measured by the protective relay.
Figure 14A Fault Event Oscillography Figure 14B Fault Current Phasors 2 nd Conclusion The breaker failure trip occurred because I C was above the current detector pickup setting and input 4 (BFI) was asserted. The breaker failure function may be used for a unit breaker rather than a generator breaker. It is limited in that it has no fault detector associated with the unit breaker.
Output contact operation would occur if any of the initiate contacts close and the 52b contact indicated a closed breaker after the set time delay. The corresponding logic is shown in Figure 15. t p 52a BF Trip BFI 0 Figure 15 Correct Breaker Failure Logic Case 3 Transformer Differential Trip due to Sympathetic Inrush when nearby Large GSU Energized via Interconnecting High-Voltage Transmission Line Synopsis The transformer differential relay protecting the step-up transformer at a processing plant tripped when a nearby large GSU at a power plant was energized from the high side. The trip was due to sympathetic inrush current flowing through the step-up transformer, as shown in Figure 16. GSU Generator Processing Plant GSU 87 Original Protection Settings Figure 16 System Operating Conditions (arrows indicate direction of Inrush Current)
Fault Current Signals Figure 17A shows the oscillography captured by the relay at the time of the trip. Note that current input IAW1 is almost completely offset and there is some distortion in other current inputs as well. Figure 17A Fault Event Oscillography (Raw waveforms) Figure 17B shows the 2 nd harmonic content of the current inputs at the time of the trip. Figure 17B Fault Event Oscillography (2 nd Harmonic Content)
The 2 nd harmonic differential current present when the trip occurred was as follows: A-Phase = 17% B-Phase = 13% C-Phase = 13% The ratio of harmonic to fundamental differential current used to restrain the transformer differential protection is calculated as follows: If the ratio is greater than the restraint setting, then the transformer differential protection is blocked. The original 2 nd harmonic restraint setting was 20% for the electro-mechanical transformer differential relay. The customer used the same setting for the multifunction numerical relay that replaced the original electro-mechanical relay. We can see from Figure 17B that a setting of 20% was not sensitive enough to detect the sympathetic inrush current flowing through the step-up transformer. 3 rd Conclusion Electro-mechanical relays had a fixed harmonic inhibit level of 20% for several decades. This worked well for many years; then transformer manufacturers started making better transformers that used less material and were designed with smaller tolerances. Therefore, modern laminated steel core transformers will not reliably produce 20% 2 nd harmonic current during inrush. A setting of 11% for the 2 nd harmonic restraint would be the most reliable based upon this particular event. Note that the multifunction numerical relay actually uses the RMS of the 2 nd and 4 th harmonic differential current but that still was not enough to restrain the protection. Case 4 Intermittent Arcing Ground Fault Leads to Flashover on All Three Phases of Large Generator - Extensive Damage Occurred Synopsis An intermittent arcing ground fault on a large machine stator winding was undetected for a long time and eventually evolved into a three-phase fault which caused extensive damage to the generator as shown in Figures 18A and 18B.
Figure 18A Initial Arc Figure 18B Ground Fault Evolving into Three-Phase Fault
4 th Conclusion Conventional protection methods are not reliable due to the intermittent nature of an arcing fault; that is any single arc may not last long enough to operate time-delayed neutral voltage protection such as 59N or 27TN. Case 5 Intermittent Arcing Ground Fault Quickly Cleared by High-Speed Arcing Ground Fault Protection Scheme Synopsis Another intermittent arcing ground fault on a large machine stator winding was quickly detected and cleared in less than 30 cycles as shown in Figure 19. The duration of the first arc was 8 cycles followed by another with duration of 10 cycles; then the high-speed ground fault trip occurred. Figure 19 High Speed Arcing Ground Fault Trip 5 th Conclusion 100% stator ground fault protection is provided by injecting a subharmonic 20 Hz voltage signal into the secondary of the generator neutral grounding transformer through a band-pass filter. The band-pass filter passes only the 20 Hz signal and rejects out-of-band signals. The main advantage of this protection is 100% protection of the stator windings for ground faults including when the machine is off-line (provided that the 20 Hz signal is present). Figure 20 illustrates a typical application. A 20 Hz voltage signal is impressed across the grounding resistor (R N ) by the 20 Hz signal generator. The band-pass filter only passes the 20 Hz signal and rejects out-of-band signals. The voltage across the grounding resistor is also connected across the voltage input (V N ) of the 64S relay. The current input (I N ) of the 64S relay measures the 20 Hz current flowing on the grounded side of the grounding transformer and is stepped down through a CT. It is important to note that the relay does not measure the 20 Hz current flowing through the grounding
resistor. The 20 Hz current increases during ground faults on the stator winding and an overcurrent element tuned to 20 Hz that operates on this current provides the protection. A short time delay on pickup is possible since the 20 Hz signal is decoupled from the fundamental power system (i.e., 50 or 60 Hz). Conventional stator ground fault protection requires a long delay to prevent unwanted tripping for external ground faults in the transmission grid due to capacitive coupling across the windings of the GSU as shown in Figure 21. The 64S protection has two overcurrent elements: one operates on the total current (I N ) and the other operates on the real component of the total current. The overcurrent element that operates on the real component sees a change in the insulation resistance of the stator windings and can detect when the insulation first starts to break down, thus preventing catastrophic damage as what occurred in the fourth case presented in this report. Neutral Grounding Transformer L l K k R N 20 Hz Band Pass Filter 1B1 1A1 1A3 1A4 1B4 Wiring Shielded 20 Hz Generator 4A1 4A3 Bl 1A1 1A2 1A3 2A1 2A3 3A2 3A3 3A1 Supply Voltage DC +V Aux -V Aux 20 Hz CT High Voltage 64S Relay Low Voltage 59N 44 45 V N 52 53 I N Figure 20-20 Hz Injection Grounding Network
Figure 21 Neutral Voltage Induced due to Capacitive Coupling across GSU Another solution is the accelerated stator ground fault tripping scheme that uses sequence voltages V 2 and V 0 measured at the machine terminal, thus avoiding the long time delay required to coordinate with system protection because of the capacitive coupling across the generator step-up transformer (GSU). Figure 22 shows the scheme which is very easy to set. V N + Short Delay 59N1P (Setting) - AND 0 Accelerated Trip V 2 V2P (Setting) - + V 0 V2P (Setting) + - V N + 59N2P (Setting) - Long Delay 0 Delayed Trip 59N1P = 59N2P Figure 22 Accelerated Stator Ground Fault Tripping Scheme Logic Note that the relay must measure line-to-ground phase voltage in order for the relay to be able to calculate the zero-sequence voltage at the machine terminal. Therefore, the VTs must be connected wye - wye.
Figure 23A shows the recommended settings for the neutral overvoltage element 59N1. The pickup setting should cover at least 90 percent of the stator windings. The pickup settings for V 2 and V 0 inhibits are based upon observation of many simulations and real world case studies. Figure 23A 59N1 Settings Figure 23B shows the recommended settings for 59N2. The pickup is the same as 59N1 and uses the time delay required to coordinate with system protection for external ground faults in the transmission system. Figure 23B 59N2 Settings Figure 24 shows the symmetrical components short circuit diagram for a single phase-to-ground fault on the high side of the GSU (i.e., external). There is ample negative-sequence voltage at the machine terminal while no zero-sequence voltage present since there is an open circuit on the generator side of the GSU; therefore, the 59N1 is blocked from tripping.
E G + + F x Z 1G Z T Z 1S E S V 2 F x Z 2G Z T Z 2S V 0 F x Z 0G Z T Z 0S Figure 24 External Ground Fault Final Conclusions This paper provides detailed technical analysis of several catastrophic relay misoperations and demonstrates how to prevent them from occurring. It also provides an example of using data recorded by a relay during onerous conditions to implement a new protection scheme. Each individual analysis ends with a conclusion which states why the relay misoperated and also provides a recommendation as to the best practice for the particular application. As mistakes often occur at the design stage, it is necessary to carefully assess the initial settings. About the Author Steve Turner, IEEE Senior Member, is a Senior Applications Engineer at Beckwith Electric Company. His previous experience includes work as an application engineer with GEC Alstom, and an application engineer in the international market for SEL, focusing on transmission line protection applications. Steve worked for Duke Energy (formerly Progress Energy), where he developed a patent for double-ended fault location on overhead transmission lines. Steve has a BSEE and MSEE from Virginia Tech. He has presented at numerous conferences including Georgia Tech Protective Relay Conference, Western Protective Relay Conference, ECNE and Doble User Groups, as well as various international conferences.