Effective commissioning of bus bar protection systems using a dynamic simulation in the field

Effective commissioning of bus bar protection systems using a dynamic simulation in the field F Fink *, J Köppel, T Hensler *OMICRON electronics GmbH, Austria, florian.fink@omicron.at, ABB AG, Germany, joerg.koeppel@de.abb.com, OMICRON electronics GmbH, Austria, thomas.hensler@omicron.at Keywords: Bus bar protection, protection testing, dynamic simulation, real-time closed-loop simulation Abstract Comprehensive testing of a bus bar protection system during commissioning or factory acceptance tests has always been a quite challenging and exhaustive task. Using a dynamic simulation software which directly controls multiple conventional test sets for simultaneous injection of test currents into all bay units is very efficient, since it does not require any rewiring all the time. Within the software the detailed topology of the whole bus bar is modelled and the injected transient current signals are calculated automatically for all the different fault and operational scenarios. A dynamic simulation does provide much more realistic signals for testing than simple steadystate values and even more advanced effects such as the saturation of CTs can be simulated. Additionally the test cases are much more comprehensive for the end user so that he can better understand and assess the behaviour of the protection relays. Another big challenge for bus bar protection testing is the complex logic within the relays, e.g. for zone selection, dead zone trips, breaker failure functions or other custom relay logic. Therefore a simulation of scenarios, where the test system responds to relay reactions in real-time (closed-loop) would be required. With multiple conventional test sets in the field, which could even be distributed in the substation to inject into all field units, a real-time closed-loop simulator is not suitable. A new iterative approach to closed-loop simulation, which mimics the real-time behaviour by repeatedly integrating all relay reactions using multiple iterative test steps can be used and finally provides the same results. 1 Introduction Bus bars are one of the most critical assets in our electrical power systems. An outage of a bus bar can cause a shut-down of a larger part of a power grid. Therefore, dedicated bus bar protection is applied in most transmission systems and for higher voltage levels. For dedicated bus bar protection the principle of differential protection is used, because this allows for very fast fault clearing times. For a more detailed introduction into bus bar protection see chapter 10 in [1]. Protection for large complex bus bars is quite challenging since it has to consider all the different possible topologies like double bus bars, multiple bus bar sections, bus couplers, transfer buses, etc. Within modern numerical low-impedance bus bar protection relays the topology of the bus bar with all the detailed arrangement of the circuit breakers (CBs), isolators, bus bar sections and current transformers (CTs) is modelled in software or the bus bar zones are configured with extensive logic expressions. This allows the relay to selectively trip only those parts of a bus bar affected by a fault. Selectivity is possible since the protection systems constantly supervises all the isolator and CB positions and measures all currents of all CTs in parallel. For large bus bars a distributed protection system is required, because the location of CTs, CBs and isolators for the individual bays are too far apart. Multiple field units are used, which are connected together, mostly using dedicated fibre optical connections, with a central unit, where the decision on selective tripping is done, which is then forwarded to the field units close to the CBs installed. Besides its main differential protection modern bus bar protection system have implemented additional logic functions to provide additional protection functions such as breaker failure protection, supervision functions and additional customized logic. 2 Commissioning of bus bar protection Commissioning of a bus bar protection system has always been an exhaustive and complex task. Since every bus bar topology is individual, the configuration of the protection relay has to match the protected bus bar exactly. The detailed protection settings have to be determined and applied to the relays either during factory acceptance testing or at commissioning in the field. To verify that the protection system is configured properly, a detailed test of the protection behaviour has to be done. Conventional protection testing, which does focus on testing individual relay elements and tries to verify individual relay 1

settings independently, does not provide enough test coverage for a bus bar protection relay. Since a bus bar protection is a complex protection system, sometimes even installed distributed, a test of the overall system including all the relay elements together is required. Similar to the approach explained in [2], a system-oriented test approach, which does test all parts of the protection relays as they work together, is much more effective for bus bar protection testing. Such a comprehensive overall test can unveil setting errors in the protection relays, which cannot be detected with a simple check of individual relay parameters. And for a distributed protection system it is essential to include the communications channels between the field units and the central unit in the test. 3 Bus bar protection testing Testing of a bus bar protection system should verify, that the relays are configured properly and that all the wiring of the protection devices within the substation was done correctly. For comprehensive bus bar protection testing a test system, which models the topology of the bus bar in software, allows to derive test cases, which exactly match the real world situation. Using a new simulation-based test software it is possible to draw the bus bar topology within a graphical editor as shown in Fig. 1. evaluate and assess the observed behaviour of the protection relay under test. 4 Simultaneous injection To be able to cover all test cases a simultaneous injection of test currents into all CT inputs of the whole protection system is required. For a larger bus bar protection the number of currents can get quite high because for every CT 3-phase current has to be generated so that all different fault types, like LN, LL and LLL faults, can be simulated. Additionally a scalable test system is necessary to cover all the different bus bar configurations with different number of bays and CTs. Therefore a flexible test setup with conventional test sets capable of 3-phase (or 6-phase) current injection is preferred. These test sets can be used for other test and commissioning purposes too. From a new PC-based test software, which can control multiple conventional test sets for simultaneous injection, the whole test can be operated easily. To test differential protection the injected currents have to be synchronized precisely so that no erroneous trips are possible due to small phase shifts in the injected quantities for stable situations. Therefore the test sets have to be time-synchronized. Modern injection test sets are capable of being timesynchronized using IEEE 1588 Precision Time Protocol (PTP), which is generated from a PTP grandmaster clock and provided to the test sets using an Ethernet network connection. A setup similar to the one shown in Fig. 2 is possible then. Fig. 1: Editor to model the bus bar topology in the test software Based on the topology modelled the software can then calculate the currents to inject for the different test cases using a transient network simulation. The simulation will consider all the isolator and CB positions for its calculations, which can be easily changed within the graphical user interface. Additionally the protection engineer, who is executing the tests, can see the modelled topology in the test system and therefore easily comprehend, what is going on during a specific test scenario. Therefore it is much easier for him to Fig. 2: Test setup with multiple time-synchronized test sets Multiple conventional test sets are used to inject currents into all the field units. They are connected to a PTP grandmaster clock and the controlling PC, where the test software is running, using an Ethernet network and a PTP transparent network switch. This setup is possible even for a distributed protection system in the field, where the test sets can be located close to the individual bays, where they are connected 2

to the field units. More details about possibilities for testing of distributed protection system can be found in [3]. 5 Simulation of isolator positions For a comprehensive test of the whole bus bar protection the position of all the isolator contacts have to be changed a lot. During factory acceptance tests and even at commissioning in the field the isolator positions are simulated using the binary input contacts of the relays. For redundancy reasons usually two contacts per isolator are used, a normally open (NO) and a normally closed (NC) contact. Additionally the protection system can supervise the CB positions using the CB auxiliary contacts (52a and 52b), which should be simulated during testing too. In the past for the purpose of simulation of the isolator positions manual switch boxes were used, where every contact could be switched manually from a central location. The construction of such switch boxes has always been a quite cumbersome work and every switch box has to be adapted for every individual bus bar topology. A far better approach is the possibility to simulate the isolator positions directly from the test software used for testing. Therefore either binary outputs of the test sets or additional binary extension devices are used, which are wired to the contact inputs of the protection system. Within the test software the isolators modelled within the topology are associated with the binary outputs (for NO and NC contacts) of such a device once. Then the protection engineer can toggle the isolator simply within the graphical interface of the test software. Since the test software is applying the isolator position from within its modelled topology automatically, the state of the bus bar is guaranteed to match the state used for the simulation of the injected currents. Using this approach even fully automated tests can be done, where the test software will switch isolator positions in between individual test cases automatically. 6 Dynamic simulation Within the bus bar topology modelled in the test software it is possible to define the infeed and load conditions on the different feeders of the bus bar. Using that information, the software can calculate all the currents used for injection during testing in a consistent and realistic way. Both magnitude of load flow and fault currents are close to the real values which would occur in the primary power system. From then on it is possible to define any fault or operational scenario within the test software. The user can set any operational and load state, the software will calculate all simultaneous current automatically and the test can be executed from within the test software easily. Additionally the user can place faults within the topology, on any node of the network. Faults are possible both inside of the protected area, e.g. on a specific bus bar section, or outside on any feeder. Even more sophisticated fault scenarios with multiple faults in the topology, which can occur at different times during the simulation, are possible. Test cases with double faults or evolving faults can be simulated too. Additionally the software is capable of simulating the CT saturation effects for high fault currents after the occurrence of a fault event, as it is shown in Fig. 3. Fig. 3: CT saturation effects at fault inception Overall all of the following important test cases can be simulated using a realistic dynamic network simulation: Stable load flow for different operational states Stability of differential protection for outside faults Stability for outside faults close to a bus bar where CT saturation occurs Faults inside the protected area on selective bus bar sections Fault in the dead zone of a coupling field For a more detailed discussion about the possibilities with dynamic transient simulation for testing of all kinds of protection systems see [4]. 7 Iterative close-loop testing Bus bar protection relays operating in the power system trip the circuit breakers to isolate a fault via the trip signals on the individual CBs. During testing the CBs are usually disconnected and the trip commands are recorded by the test system. For realistic tests the test system should react on this trip commands and simulate the corresponding event, e.g. opening of a breaker, within its dynamic simulation in realtime. Such a real-time closed-loop simulation can be done using a complex and expensive real-time simulator in the lab. Using conventional test sets and a network simulation running on a standard PC, reaction times of a real-time simulator are not possible. If the test setup is distributed and multiple test sets have to be used, which are connected to the controlling computer over longer distances, a real-time simulator is not suitable too. However, an interactive approach is possible, which can mimic the behaviour of a closed-loop simulation using 3

multiple repetitions. Therefore, a fault scenario is simulated repeatedly taking into account all the reactions of all the relays under test, recorded during a previous execution. We assume that the relays behave deterministically (relays are reset properly in between every iteration). When the same scenario is repeated again, the relays will issue its trip commands at approximately the same time (the software allows for some tolerance in the timing of events) and the simulation can now foresee this and simulate the corresponding reaction accordingly. This procedure is iterated until no more new relay reactions are recorded. The final execution of the fault scenario is then exactly the same as if a real-time closed-loop simulator had been used. For dynamic testing of bus bar protection systems there are multiple test scenarios, where this iterative closed-loop approach can be used. Selective trips for faults in one bus bar section can be tested and assessed easily using just one repetition. An overall screenshot of the test software for this scenario is shown in Fig. 6. For example, the following figures show two iterations for a selective bus bar trip on a bus bar protection system with two buses and 4 bays. For the first iteration, the simulation cannot take into account that bay 2-4 trip the breakers. Therefore the fault current persists on all bays and breaker-failure protection will trip the remaining bay on the non-faulted bus bar with a certain time delay. See Fig. 4. Fig. 6: Selective trips for faulted bus Fig. 4: 1 st iteration without simulation of breaker trips For the next iteration, the simulation considers the instantaneous trips for bay 2-4 and opens the breakers after the CB delay time so that the fault is cleared. Now the relays will not activate the breaker-failure protection and the trip signals will reset. See Fig. 5. Additionally the software can suppress the application of the relay reaction within the simulation and simulate a failure of a circuit breaker for a specific bay only. This scenario can be seen in Fig. 7, where only two of the bays tripped correctly and where the breaker-failure protection tripped the remaining bays with a certain time delay so that the fault is cleared ultimately by tripping all bays (three iterations used). Fig. 5: 2 nd iteration with breaker trips The whole procedure is executed from the test software on the controlling PC automatically without any user interaction. It is not restricted to two iterations but is repeated until no more new reactions are seen from the relays within the overall simulation duration. 4

8 Practical examples This new testing approach of using a dynamic simulationbased test for commissioning or factory acceptance testing of bus bar protection systems has already been used multiple times practically. The test setup with its scalable number of test sets is applicable both for centralized bus bar protection system, see Fig. 9, and distributed systems as shown in Fig. 10. Fig. 7: CB failure trip Even more complex scenarios such as faults in the dead zone of a coupling field can be tested. This is shown in Fig. 8, where a fault is placed in the coupling field in between the CB and the CT of the coupler (for a coupling field with a single CT only). The dead zone fault cannot be cleared by the selective instantaneous trips of a single bus bar zone, so that the protection will trip the whole bus bar afterwards. Fig. 9: Commissioning of a bus bar protection in the field Fig. 10: Test of a distributed bus bar protection system Using a new integrated test software on a PC makes bus bar protection testing much more efficient and effective. The whole test can be controlled from a single PC. The test procedure can be prepared in the software already upfront, so that the actual execution of all the necessary test steps can even be done automatically in the field. Fig. 8: Fault in dead zone with delayed trip of whole bus bar For some cases substantial time savings were possible compared to conventional commissioning using manual tools before. Using this new approach errors in the protection settings could be found, which were not detected using a 5

manual approach. For more information about practical examples of bus bar protection tests in the field, see [6, 7]. 9 Conclusions Commissioning and factory acceptance testing of bus bar protection systems using a dynamic simulation has been proved to be effective and efficient. Using this new approach a new quality for bus bar protection testing is possible using conventional injection test sets. Within a new PC based test software a convenient test and easy and comprehensible assessment of the protection behaviour is possible. Using a dynamic transient network simulation even more advanced effects such as the saturation of CTs can be tested. For testing of the detailed behaviour of relay logic a closedloop simulation is necessary. Using an iterative closed-loop approach, which is executed by the test software automatically, the same results as with a real-time close-loop simulators can be obtained with multiple test sets even for a distributed test setup. Test cases for the following important scenarios are possible and can even be prepared for automated execution in the field: Stable load flow and operational states Stability for outside faults (optionally with CT saturation) Selective trips for inside faults Trip of breaker-failure protection Faults in the dead zone of a coupling field References [1] Ziegler, G.: Numerical Differential Protection Principles and Applications, Publicis Publishing, 2nd Edition, 2012 [2] C. Pritchard, D. Costello, K. Zimmerman, Moving the Focus from Relay Element Testing to Protection System Testing, PAC World Conference, Raleigh, 2015 [3] B. Bastigkeit, C. Pritchard, T. Hensler, New possibilities in field testing of distributed protection systems, PAC World Conference, Zagreb, 2014 [4] T. Hensler, C. Pritchard, F. Fink, New Possibilities for Protection Testing using Dynamic Simulations in the Field, MATPOST Conference, Lyon, 2015 [5] T. Hensler, Iterative closed-loop testing of protection devices using a dynamic simulation in the field (in German Iterative Closed-Loop Prüfung von Schutzgeräten mit dynamischer Simulation im Feld ), e & i, Heft 8, Nov./Dec.. 2014, page 24-27 [6] C. Pritchard, T. Hensler, Test and verification of a busbar protection using a simulation-based iterative closed-loop approach in the field, Australian Protection Symposium, Sydney, 2014 [7] F. Fink, T. Hensler, F. Trillenberg, J. Köppel, A system-oriented approach for testing a distributed busbar protection (in German Systemorientierter Ansatz für die Prüfung eines verteilten Sammelschienenschutzes ), netzpraxis, Magazin für Energieversorgung, Heft 12, Dec. 2014, page 40-46 About the authors Dipl.-Ing. (FH) Florian Fink was born 1983 in Bergisch Gladbach / Germany. He received his diploma in Electrical Power Engineering at the University of Applied Science in Cologne in 2009. From 2009 until 2012 he worked as project engineer for Cegelec / Germany and from 2012 to 2013 as planning engineer for InfraServ Knapsack / Germany. Since 2013 he is working for OMICRON electronics in product management as an application engineer for power system protection. Dipl.-Ing. (FH) Jörg Köppel was born 1973 in Erzhausen / Germany. He studied Electrical Power Engineering at the University of Applied Science in Darmstadt / Germany, where he received his diploma in 1997. From 1997 until 2010 he worked for ABB Germany in the area of power system protection and control focused on busbar protection. Since 2010 he is working in project engineering for protection with a main focus on busbar protection. Dipl.-Ing. Thomas Hensler was born in 1968 in Feldkirch / Austria. He received his diploma (Master s Degree) in Computer Science at the Technical University of Vienna in 1995. He joined OMICRON electronics in 1995 where he worked in application software development in the field of testing solutions for protection and measurement systems. Additionally he is responsible for product management for application software for protection testing. 6