WIXLIB

HIGH-POWERED RESEARCHFORDecember 2011 Issue TH1 Modeling Current Transformer (CT)Saturation for Detailed Protection StudiesE REAL WORDecember 20116 The Wound Rotor (WR) and SquirrelCage (SQ) Induction Machine Modelsin the PSCAD Library9 New PT Connections11 Meet the Team!12 PSCAD 2012 Training SessionsLDModeling Current Transformer (CT)Saturation for Detailed Protection StudiesDr. Dharshana Muthumuni, Lisa Ruchkall, and Dr. Rohitha Jayasinghe, Manitoba HVDC Research CentreThe PSCAD Master Library contains a numberof models that could be used for detailedprotection system analysis. The most importantof these models are the current transformer models.Current transformer saturation is associated withmany protection problems encountered in powersystems. CT saturation is a complex phenomenonand accurate modeling in a simulation environment ischallenging. The models available in PSCAD are basedon the advanced mathematical models proposed byJiles and Atherton and commonly referred to as the‘Jiles-Atherton theory of Ferromagnetic hysteresis.’These developed models have been extensively testedwith field recordings in collaborations with CT, relayand protection equipment vendors.This case is used to illustrate the effects of CTsaturation. The key parameters that impact CTsaturation are discussed in an effort to guide PSCAD users who may use this model for their simulations.General Theory The magnetic characteristic of theCT is shown in Figure 1 (hysteresis not shown). In thelinear region, the CT will behave almost like an idealratio changer. That is, the CT secondary current isan identical but scaled down replica of the primarycurrent. However, If the CT saturates, more currentis required to magnetize the core and as a result thesecondary current (IS) available as inputs to the relayFigure 1 IM –Flux curvemay not be an identical scaled down replica of theactual primary current (IP). This can lead to protectionissues and should be given due consideration.System Overview The AC system shown inFigure 2 consists of two 230 kV, 60 Hz Thevaninvoltage sources, a 75 MVA load and three 230 kVtransmission lines (125 km, 75 km and 200 km).A single phase (Phase A) to ground fault is appliedbetween the first two transmission lines.

Current Transformer Saturation CT saturationcan be explained using the simplified equivalentcircuit shown in Figure 3.This representation is valid since the CT is essentially ashort circuit (in series) from the primary power systemnetwork perspective.In the linear region of operation, magnetizing current(IM1) is very small and hence IP – IM is approximatelyequal to IP. Thus IS would be a scaled down version(by a factor of N). If the CT saturates, the magnetizingcurrent increases (IM2). As a result, only a part of IPis available for transformation to the secondary.In this simulation case, a fault is simulated on thetransmission line and the CT is at the end of theline (where Iabc is measured).Key Parameters Impacting CT SaturationThe following can have a significant impact onCT saturation and should be given due considerationin a simulation study:Note: Figure 4 shows how the CT model is used inPSCAD . Note also that the input primary current (Iabc)is in kilo amps and the secondary current (Isabc) is inamps. The CTs are modeled as (independent) blocksand do not have to be connected to the electric circuit.1) DC offset in the primary side fault current.2) Remnant flux on the CT prior to the fault (if any).3) Secondary side impedance including those ofthe relay, connecting wires and CT secondaryimpedance - this parameter plays a major role inthe level of saturation the CT will be subjected to.Illustrative simulation results are presented in thefollowing sections to highlight some of the key points.Figure 2 AC system modelFigure 3 Schematic representation of CT (left) and the simplifiedequivalent circuit (right)2 PULSE THE MANITOBA HVDC RESEARCH CENTRE JOURNALFigure 4 CT PSCAD implementation

800Impact of DC offset in the primary fault current.Case 1 – Impact of DC offset in the primary faultcurrent: The point on the voltage wave form atthe instant of the fault determines the level of theDC offset in the primary current. The maximum DCoffset will occur when the fault is applied at a voltageminimum (t 0.49167 s). The results in Figure 5a occurwhen the DC offset is significant.The simulation results in Figure 5b demonstrate asituation when there is no DC offset (fault is applied att 0.4876 s, voltage maximum). As can be seen, the CTdoes not go into saturation and only a small amountof magnetizing current is required to magnetize thecore. Therefore, the secondary current is an exact butscaled down replica of the primary current.As can be seen, the DC offset causes the flux (Figure5b) to be driven down and into saturation. The CThas saturated after about two cycles. The reductionof the secondary current is evident from Figure 5a.In Figure 5a also note the remnant flux in the CT coreonce the fault is cleared. Effects of remnant flux willbe discussed in Case 2.IabcVoltage 00-1.50-2.000.5000.5500.600Figure 5a DC offset 0.7500.8000.8500-20-30DECEMBER 0.450Figure 5b No DC offset 100-150-20020Sec. Current (A)-20-30Ia secPri. Current (kA)-100.4500-100Flux density (T)Pri. Current (kA)Flux density (T)0.850Sec. Current (A)00.5010020104.0Vabc-200Ia secPri. Current (kA)Sec. Current (A)20200Voltage (kV)200150100500-50-100-150-200Main : GraphsFlux density (T)Voltage (kV)Main : GraphsVabcB

Note: the B-H loop trajectory of the CT during thefault is shown in Figure 6. The formation of the minorB-H loops and hysteresis are accurately modeledbased on the Jiles-Atherton theory of ferromagnetichysteresis. Such detailed representations of the CTbehavior are necessary for detailed protection systemanalysis, such as:Case 2 – Reclosing the line while the fault isstill present (auto reclose): In the first fault event,saturation had taken approximately two cycles. If theline is reclosed while the fault is still present, the CTmay saturate much faster due to the presence of theremnant flux. This is demonstrated in Figure 7. As canbe seen, the saturation occurs in approximately half acycle, possibly before the relay had time to respond. CT response during auto reclose Protection schemes with CTs operating in parallel– 6 CTs in parallel in a three-phase transformerdifferential scheme– 3 CTs in parallel in an earth fault relay scheme– CT connections in generator protectionsFigure 7 Case 2 resultsFigure 6 B-H loop trajectory4 PULSE THE MANITOBA HVDC RESEARCH CENTRE JOURNAL

The response of one CT may impact the behaviourof the other and should be appropriately modeled Main : bc2.01.0kA0.0-1.0-2.0-3.0-4.0-5.0B0.00-0.50F lux denNote: The CT model used in the example case isthat of a single CT operating independently ofany other CTs in the protection scheme. In general,a protection scheme may have a number of CTsoperating in parallel. The response of one CT mayimpact the behaviour of the other and thus, shouldbe appropriately modeled. Contact the PSCAD Technical Support Team for more informationon specific models.Ia sec10.0ACase 3 – Effect of the secondary impedance: The CTsecondary side burden impedance has a significantimpact on CT saturation. In Case 1, the burden was setto 2.5 Ω. The results shown in Figure 8 were obtainedby reducing the burden to 0.5 Ω in the simulation.As can be seen from the results, the flux is not drivendown as far. It also takes a longer time for the CTto become saturated (approximately 6 7000.7500.8000.850Figure 8 Case 3 resultsDECEMBER 20115

The Wound Rotor (WR) and Squirrel Cage (SQ)Induction Machine Models in the PSCAD LibraryDr. Dharshana Muthumuni, Manitoba HVDC Research CentreThe induction machine models are two ofthe most widely used components from thePSCAD Master Library. The PSCAD library hastwo induction motor models:The equivalent circuit of a double cage design(as implemented in PSCAD ) squirrel cage machineis shown below in Figure 2.RSXLSXMR1) Squirrel cage induction machine modelrepresenting a double cage design – (SQ).2) Wound rotor induction machine model – (WR).XLC2RC2Figure 2 SQ cage (double cage) equivalent circuitMathematically, the SQ cage machine can berepresented by the WR machine. The WR modelcould also be used to represent a double cage SQmachine. The two examples below will describerelevant data entry considerations and also compareresults for validation purposes.The simple simulation case shown below is used tohighlight the key points. A SQ machine and a WRmachine with comparable parameters (identicalratings and data) are connected to the same supplysource (60 Hz, 0.69 00TTIN2TIN1TIN2Figure 1 Circuit diagram (WR – top, SQ cage – bottom)6 PULSE THE MANITOBA HVDC RESEARCH CENTRE JOURNALRSXLSXLRXMRRTo use the SQ cage machine model to representa single cage machine:WTimedBreakerLogicOpen@t0The equivalent circuit of a wound rotor machinemodel (single rotor winding) is shown in Figure 3.Figure 3 WR IM (single winding) equivalent circuitExample 1: Modeling a single cage inductionmachine The SQ cage machine model or the WRmachine model may be used to represent a singlecage (SQ) machine (IM study 01.pscx).IMRC1XMThe Technical Support Team receives questionsfrom our users as to which model they should useto represent a set of specific data provided by theequipment vendor. The goal of this technical noteis to provide the necessary information to helpPSCAD users when faced with such questions.– Make the ‘second cage resistance’ (RC2) andthe ‘second cage unsaturated reactance’ (XLC2)relatively large (compared to the other leakageinductances/resistances). In this case, the valuesused are RC2 5 PU and XLC2 5PU, which is muchlarger than RC1 0.0507PU and XMR 0.091 PU.– Give the SQ cage ‘rotor unsaturated mutualreactance’ (XMR) the same value as the WR‘rotor leakage reactance’ (XLR).

Figure 4 SQ cage and WR setup configurationFigure 4 shows the data entry for the SQ cage (left)and WR models (right). By using equivalent values,both models show comparable behaviour (Figure 5)and represent a single cage machine design.Simulation 0.500.751.001.251.501.752.00Figure 5 Simulation results (IM study 06 a.pscx)The simulation results shown in Figure 5 show thatthe speed (W – WR, W2 – SQ cage) and torque (T – WR,T2 – SQ cage) of both machines are identical. Thus, anyone of the induction machine models may be used torepresent a single cage induction machine.Example 2: Modeling a double cage inductionmachine. The WR machine model can be setup to represent a double cage SQ cage machine(IM study 02.pscx).In the WR model, select the “No. of Rotor SquirrelCages 1”, as shown in Figure 6. The equivalentcircuit representation is as in Figure 7.Figure 6 WR configurationRSXLSXMRRLRXLC1RRRC1XMFigure 7 WR IM (No. of rotor squirrel cages 1) equivalent circuitNDOE VCEMBER 2010197

Figure 8 WR setup configurationFigure 8 shows the data entry for the WR model.With the appropriate data, the SQ cage and WRmachine models will give identical results (Figure 9).A SQ cage and WR machinemodel deliver equivalent resultswhen configured properly Figure 9 Simulation results (IM study 06 b.pscx)Conclusions As can be seen from the results,a SQ cage and WR machine model deliver equivalentresults when configured properly. Hence, a SQcage machine model can be accurately representedusing a WR machine model. PSCAD users areencouraged to use the WRIM model due to itsmore conventional and straight forwardparameter configuration.PSCAD Cases:IM study 01.pscx and IM study 02.pscx8 PULSE THE MANITOBA HVDC RESEARCH CENTRE JOURNAL