1 ColumbiaGrid Transient Stability Manual Bo Gong, PhD gong@columbiagrid.org 2 3 In March 2016, ColumbiaGrid developed the first draft of this transient stability manual with the collaboration of member utilities. This manual serves two purposes: First, it provides a reference for engineers who are not familiar with transient stability to perform transient stability simulation with procedures. Second, it explains most of the simulation options that were accepted by ColumbiaGrid members and participants on how to set up PowerWorld for dynamic simulation. By adopting these options, all users can seamlessly compare and exchange information and results. The first part of the document is the procedure for performing a dynamic simulation with the focus on using PowerWorld simulator, which is adopted by almost all ColumbiaGrid member utilities and planning participants. The procedure lists most key steps an engineer may choose to follow to verify the case and model information, and evaluate the simulation results. It is not intended to cover all fundamental topics and important issues about transient stability. Most of the basic information about theory can be found in a variety of references. This manual, on the other hand, was developed to be a quick reference for engineers to follow in a step by step manner. Each section is relatively independent so that more experienced users can skip some of the previous sections. The second part of the document lists some of the references that ColumbiaGrid members provided. The topics include an introduction to transient stability and special model descriptions. The intention of having the second part is to provide some reference for users to quickly find information on some important transient stability related issues. We expect that both the first and second parts will be continuously updated. After each revision, the document will be made available in the ColumbiaGrid website for downloading. 4 Part 1: Dynamic Simulation Procedure 1.1 Transient stability base cases 1.2 Initial case checking 1.3 Dynamic simulation options 1.4 Simulation initialization and standard tests 1.5 Add contingencies 1.6 Define plots 1.7 Perform simulation 1.8 Evaluate results 5 1.1 Transient stability base cases Each year WECC publishes a number of base cases for power flow and transient stability studies, taking into account the latest planned system upgrades available that year. Generally speaking, these base cases may include: a. Five operating cases b. Three scenario cases c. One five-year summer planning case d. One five-year winter planning case e. One 10-year planning case (alternate winter and summer) WECC collaborates with area or planning coordinators to collect and update both power flow and stability information for each base case. The planned schedule for the base case preparation and review process for 2016 is showed in Table 1.1. Table 1.1 WECC schedule of base case preparation After collecting the data from each area coordinator, WECC develops base cases and sends them out for review. After review, base cases are finalized and posted on the WECC base case website. The actual posted date is normally later than the proposed schedule shown in table 1.1. The post date and status of the base case can be found on the WECC website, shown in Figure 1.1. 6 Figure 1.1 WECC base case website with status and post date To use a base case in PowerWorld, users should only download it in GE PSLF EPC format. The procedure for downloading a base case is as follows: 1. Go to WECC home page: https://www.wecc.biz/Pages/home.aspx, select “Program Areas” tap from the upper left corner. In the dropdown menu, click “Reliability Planning & Performance Analysis” and then click “Planning Services”. 7 2. Clicked the “base cases” link on the left under Planning Services. To download a base case, users are required to log into WECC website using his user name and password. The login button is on the upper right corner. After login, the base cases can be found and downloaded by clicking the year it is prepared. 8 3. PowerWorld users should download a GE PSLF EPC format base case. Please check the status of the base case and make sure it is “Approved/Final” before downloading it. Clicking the case name, a pop-up window will appear. All base case files are normally encapsulated in ZIP format. Users need to unzip it for use. 9 4. After unzipping the base case file, users can check in the folder to verify that it has power flow raw data file in GE PSLF format ( file named as *.epc) and dynamic data (file named as *.dyd). 10 1.2 Initial Case Checking 1.2.1 Power Flow Case Checking After loading a raw format power flow case in PowerWorld, users should make sure the case can be solved correctly. This is especially important for cases being converted from other formats, as some power flow data entries in one software may be missing in another and set to some default values. If a power flow case that fails to be solved is used for transient stability simulation, it may cause a lot of initialization problems and lead to erroneous simulation results. To accurately solve a power flow case, several options can be set depending upon the conditions of each case. For example, users should first check the power flow solution option. One option particularly important is the Generator VAR Limits option. By default this option is not checked when PowerWorld reads a case from GE PSLF format. Without checking this option, a case may end up using many more iterations to solve or may not reach a solution. After solving a power flow case, users should go through the messages in the log window and make sure there are no other problems marked by the software. If any problems are identified, they should be fixed before users proceed to the next step. 11 Users are also encouraged to check the limit monitoring results before proceeding to the stability simulation. A shortcut to the limit monitoring button is shown in a red circle in the next figure. 12 In the limit monitoring window, it is worth checking bus voltages and line flows that deviate significantly from the nominal range: for voltage (0.9 – 1.1 pu), for line flow (>100%). Extremely low or high voltage may indicate problematic generator voltage set points or switching device status that may lead to initialization problems. Similarly, overloads of transmission lines can also indicate generator or load model problems that need to be fixed before proceeding to the transient simulation. 1.2.2 Dynamic Data Checking After loading stability data in DYD format, users should read through messages in the log files to make sure data are is loaded correctly. A detailed procedure to verify stability simulation data will be discussed in section 1.4. 13 1.3 Dynamic simulation options PowerWorld simulator allows users to select many different solution options for transient stability simulation and reporting. All these options can be adjusted easily depending on the application or preference of the user. In order to coordinate transient stability simulation using PowerWorld simulator across all ColumbiaGrid member utilities, a workshop was hosted by ColumbiaGrid in August, 2015. During the workshop, members reached consensus on adopting common dynamic simulation options for transient stability simulation in the Northwest region for more consistent information sharing. All dynamic simulation options, including monitored transient limits, are listed in Table 1.2 in PowerWorld Aux file format. Users can copy and paste this text into a notepad and save as an aux file. Before each transient stability simulation, this file can be loaded to set the options automatically. In order to update dynamic options, the dynamic simulation options aux file must be loaded in “Edit Mode” only. Adynamic simulation options file can also be downloaded from the ColumbiaGrid website: http://www.columbiagrid.org/download.cfm?DVID=4095. Upon downloading, the file should be saved in an extension name *.aux. Table 1.2 Dynamic Simulation Options for Transient Stability //-------------------------------------------------------------------------------// THE FOLLOWING ARE THE TRANSIENT STABILITY OPTIONS //-------------------------------------------------------------------------------DATA (TRANSIENT_OPTIONS, [MaxItr,ConvergenceTol,ConvergenceTol:1,ExpDirectory,TSOStorageOption, TSOUpdateDisplayNTimeStep,TSOTransferOnManualTimeStep,TSOTransferOnRunUntil, TSOTimeStepUpdateTransferToPF,TSOTimeStepUpdateResults, TSOShowResultPageWhenDone,TSOBusIDFormat,TSOMVABaseForInputDisplay, TSOValidationAllowUnSupportedModel,TSOMaxAngleDifference, TSOInfiniteBusModeling,TSOAngleReferenceOption,TSOInitRefAngleAtZero, TSOAngleRefGenNum,TSOAngleRefGenID,TSOFastValvingOption, TSOFastValvingParameter,TSODefaultLoadModel,TSOGroupResultsBy, TSOResultsUseAreaZoneFilters,TSOSaveResultsForOpenDevices, TSOSaveResultsTimeStepsPerSave,TSOManualTimeSteps,TSOManualRunUntilTime, TSOSaveMinMaxValues,TSOSaveMinMaxValuesTime,TSUseAreaZone,TSEveryResult, TSEveryResult:1,TSEveryResult:2,TSEveryResult:3,TSEveryResult:4, TSEveryResult:5,TSEveryResult:6,TSEveryResult:7,TSEveryResult:8, TSEveryResult:9,TSEveryResult:10,TSEveryResult:11,TSEveryResult:12, TSEveryResult:13,TSEveryResult:14,TSEveryResult:15,TSEveryResult:16, TSOSynGenLowFreqHz,TSOSynGenLowFreqSec,TSOSynGenLowFreqAction, TSOSynGenHighFreqHz,TSOSynGenHighFreqSec,TSOSynGenHighFreqAction, TSOSynGenAngleDeg,TSOSynGenAngleSec,TSOSynGenAngleAction, TSOSynGenCBDelayCycles,TSOSynGenOnlyNoRelay,TSStoreResultsInRAM, TSSaveResultsToHardDrive,TSOMinDelt,TSOInitLimitViolation,TSOTransferOnEvent, TSORunProportional,TSORunProportionalMult,TSOForceSolution, TSOUseVoltageExtrapolation,TSOIgnoreSpeedInSwing,SaturationModel, IntegrationMethod,TSExciterParamCalc,TSMachSatIgnore,IncludePDCI, TSOStartLimitMonitoringValues,TSOStartLimitMonitoringValuesAfterLastEventTime, TSOStartLimitMonitoringValuesTime,TSBusFreqMeasT,TSOWhereResultEvents, TSOWhereResultEvents:1,TSOWhereResultEvents:2]) { 25 0.001 1.000 "" "NO" 30 "YES" "YES" "NO " "NO " "YES" "Name(Number)" "Device" "Error" 1080.000 "None" "Average" "NO " 30000 "1" "Frequency" 0.100 "PI, QZ" "Object/Field" "NO " "NO " 1 1 0.000 "After Last Event" 0.000 "NO " "YES" "YES" "NO " "NO " "YES" "NO " "NO " "YES" "YES" "NO " "6" "YES" "YES" "NO " "NO " "NO " "NO " 57.600 2.000 "Log Warning" 62.400 2.000 "Log Warning" 180.000 0.000 "Log Warning" 0.000 "YES" "NO " "YES" 4.000 "Abort" "YES" "NO " 5.000 0.000 "YES" "NO " "Quadratic" "RK2" "GE Approach" "Flip Values" "YES" "After Last Event" 0.000 0.000 0.050 "Both Log and Event" "Both Log and Event" "Both Log and Event" } 14 //-------------------------------------------------------------------------------// THE FOLLOWING ARE THE TRANSIENT LIMIT MONITORING //-------------------------------------------------------------------------------DATA (TSLIMITMONITOR, [LSName,Active,Category,Abort,TSTdelay,CTGViol,ObjectType,VariableName,FilterName, LimViolValue,Duration,Side,UnitsType,ConditionCaseAbs,UseStopValue,StopValue, UseStartValue,StartValue,UseStopValue:1,StopValue:1,UseStartValue:1, StartValue:1],AUXDEF,YES) { "WECC Category B Voltage Dip Non-Load Bus" "YES" "" "Log" 0 100 "Bus" "TSBusVPU" "Non-Load Only" -30 0 "Lower" "Percent Deviation" "NO " "NO " 0 "NO " 0 "NO " 0 "NO " 0 "WECC Category B Voltage Dip Load Bus" "YES" "" "Log" 0 100 "Bus" "TSBusVPU" "Load Only" -25 0 "Lower" "Percent Deviation" "NO " "NO " 0 "NO " 0 "NO " 0 "NO " 0 "WECC Category B Voltage Dip Load Bus Duration" "YES" "" "Log" 0 100 "Bus" "TSBusVPU" "Load Only" -20 0.333 "Lower" "Percent Deviation" "NO " "NO " 0 "NO " 0 "NO " 0 "NO " 0 "WECC Category B Frequency" "YES" "" "Log" 0 100 "Bus" "Frequency" "Load Only" 59.6 0.1 "Lower" "Actual" "NO " "NO " 0 "NO " 0 "NO " 0 "NO " 0 "WECC Category C Voltage Dip Any Bus" "YES" "" "Log" 0 100 "Bus" "TSBusVPU" "" -30 0 "Lower" "Percent Deviation" "NO " "NO " 0 "NO " 0 "NO " 0 "NO " 0 "WECC Category C Voltage Dip Any Bus Duration" "YES" "" "Log" 0 100 "Bus" "TSBusVPU" "Load Only" -20 0.667 "Lower" "Percent Deviation" "NO " "NO " 0 "NO " 0 "NO " 0 "NO " 0 "WECC Category C Frequency" "YES" "" "Log" 0 100 "Bus" "Frequency" "Load Only" 59 0.1 "Lower" "Actual" "NO " "NO " 0 "NO " 0 "NO " 0 "NO " 0 } In the next few sections, each option will be discussed in detail. 1.3.1 Simulation time The first adjustable options for dynamic simulation are listed under Simulation Control tab. Users can specify simulation time values here for each contingency, shown in a screenshot. There are no fixed values for each of the time selections. A typical starting time is set at 0.00 seconds. End time can be decided depending on individual contingencies. For certain contingencies where its events, such as reclosing, delayed relay actions or system adjustment, are triggered consecutively in a longer period, 15 end times should be long enough to cover the last event and subsequent system restoration period. In events where large a amount of generation is lost, restoration of generation and demand balance driven by governor response normally takes a relatively longer period. In general, the simulation end time should always be long enough to capture the final stable condition where all generators are ramping up or down to a steady state values. On the other hand, an end time should not normally exceed 30 seconds, as many dynamic behaviors that occur after this time frame are not modeled in a transient stability program. A typical simulation time step of 0.5 or 0.25 cycles should be used. A larger time step (>=1 cycle) should be avoided as it may invalidate many transient behaviors with a time constant less than a few cycles. In cases where a certain simulation with 0.5 cycle time step may show numerical instability, a smaller time step of 0.25 cycles can be used. Please notice that a smaller simulation time step will lead to longer computation times and larger storage space for data. For example, a simulation with 0.25 cycle time step will normally take more than twice the time compared to a simulation with time step at 0.5 cycle. 1.3.2 General Options “MVA base for input/Display of Generator Values” should be set to “Use Individual Generator MVA Base” as all existing generator values are computed based on their individual generator MVA base. All the other options in this tab are kept at their default values. These can be changed at the users discretion and will not impact the simulation result. For “Automatic Update and Transfer Results to Power Flow Options”, allowing results to be transferred to Power Flow more frequently may slow down the simulation. 16 1.3.3 Power System Model: Common “Power System Values” and “Network Equations Solution Options” are left at their default values in PowerWorld. “Infinite Bus Modeling” should use “No infinite buses” for a large system simulation. The other option “Model the power flow slack buses as infinite buses” should be only used for small or test systems. “Handling of Initial Limit Violations” should be set to “Abort”. This means whenever there are some initial limit violations found during the initialization stage, the simulation will be aborted. This is an important step for users to check initial limit violations before any simulation is performed. Initial limit violations may imply severe base case errors that lead to problematic oscillatory behavior or instability. All these violations should be carefully reviewed and corrected before any simulation is performed. “Load Modeling” sets default load models (loads without an explicit dynamic load model) in a dynamic simulation. As a convention, Constant Current P, Constant Impedance Q are chosen as the default values. Such a selection will not impact any load with an explicit stability model, such as composite load model. Therefore, this selection will have no impact on the load modeling in the Northwest region, but more or less other regions. “Minimum Per Unit Voltage for Constant Power Models” and “Constant Current Models” are threshold values used by PowerWorld to scale down a load when its bus voltage drops below these values. This scaling is adopted to avoid numerical problems (unsolved power flow) during a simulation. The default 17 values are kept. Decreasing these values may cause power flow to be more difficult to solve, and therefore may more easily crash a simulation or cause earlier termination. “Integration Method” should be set to the default value “Second Order Runge-Kutta”, as this method has proven to provide better numerical stability than “Euler” method. However, if a user wants to have a better comparison between PowerWorld and GE PSLF or Siemens PSS/E, he can choose to use the “Euler” method. The simulation results may have a slight difference in the magnitude of one or two time steps. 1.3.4 Power System Model: Compatibility Option “Exciter Saturation Model” option was determined during the Workshop as “Quadratic (GE Approach)” to be consistent with model imported from GE DYD format. “Exciter Automatic Parameters” option was determined to use “VR = Zero Approach” to be consistent with how GE PSLF models excitation system and determine KE values. “Machine Saturation for S12 < S10” option was determined to use “Flip Values” to accommodate likely errors where saturation factors are placed in wrong positions in the parameter list (S12 < S10). “Saturation when One SE is Zero” option was determined to use “Treat as Always Zero” to totally ignore saturation when the parameter is missing. All the other options are kept to default values as convention to handle model accurately. 1.3.5 Result Option 18 Result options can be adjusted by users depending on the scenarios and purpose of the study. In general, members decided to have PowerWorld check for results after the last event. This excludes any extremal conditions such as extreme low or high voltages/frequency during certain events. However, for some other events, voltage or frequency may stay low or high for a short period after the fault is cleared, or they may jump in an opposite direction immediately for a few cycles after fault clearing. To exclude such short transient response from the result reporting, users can customize the time period by using a different option. Average of generator angles will be used as the angle reference. Events including transition, model trip and relay trip will be both logged and triggered. 1.3.6 Generic Limit Monitors Option 19 Based on the discussion during the workshop, synchronous generator limit violation without relays will be monitored only for reviewing purposes. Generic pickup values and pickup time are used for Absolute Angle Deviation, Over Speed and Under Speed. Users can adjust these values if other values can better reflect their system protection scheme. Since these relays are not explicitly modeled, violations of these limit monitor settings should not be used for the purpose of tripping existing generators. 1.3.7 Transient Limit Monitors Option Currently, a new set of WECC transient limit monitors has been proposed and are under review. Before it is finalized, the existing WECC transient limit monitoring criteria will be used for transient stability simulation. These include WECC category B and C voltage and frequency monitors. Upon the approval of the new criteria, ColumbiaGrid will update this manual accordingly. 20 1.4 Simulation initialization and Standard tests Initialization and test simulation is an important step to identify and fix potential model errors. In general, stability model problems need to be corrected before any transient stability simulation is performed. Any model errors can be propagated and exaggerated along the simulation process and eventually invalidate the whole simulation results. On the other hand, model errors are known to be difficult to identify and correct. The large set of model types and model parameters makes this effort even more challenging. In recent years, new components such as renewable devices have been added into the system with aging devices being replaced and upgraded constantly. However, most of the models for these devices are still the same ones that were originally developed several decades ago. Compared to this trend, the progress to correct existing errors is relatively slow. The WECC MVWG has spent a significant amount of time to identify errors and communicate with the owners for potential fixes. Consequently, considerable engineering judgement is necessary to decide how to handle simulation model problems. In this section, we will suggest some ways to identify and fix potential problems, specifically fit to the PowerWorld Transient Stability Program. It should be mentioned that such methods neither assume a complete solution to address any stability problems, nor does it give a fixed procedure that engineers must follow. An engineer is encouraged to utilize his own judgement whether some or all of the methods described here should be adopted for his specific case. 1.4.1 Validation Tool from PowerWorld PowerWorld incorporates a validation tool to help users identify and fix potential model errors before any simulations. The tool is easy to use and well documented for issues being found. If such an issue can be corrected, PowerWorld will also provide information on the corrected parameter value. As the first step of data checking, users are always encouraged to use the validation tool to do the first round of screening. Based on our experience, this validation tool primarily focuses on errors in three categories: generic errors that may prevent simulator from running correctly, parameters outside their generic range, and mismatchs between certain parameters with power flow data. The first category includes errors such as: Generator without a model Machine model type mismatch (Generator vs Exciter vs Governor, etc) Model parameter time constants too small (may cause numerical problem) Model parameter gains too large (may cause numerical problem) The second category includes errors such as: Parameter values exceeding their generic range Several related parameters doe not satisfy their mathematical logic 21 The third category includes parameters for certain types of models whose parameters require a match to its power flow data. These models include some SVC models and wind turbine models. Validation tools also provide a one-click solution to auto-correct all the errors and some of the warnings. After the fix, users should always check the message window and verify solutions are valid from PowerWorld. It is worth mentioning that identification and correction of errors with the PowerWorld Validation tool only provides very preliminary checking of common errors. The main objective for this checking and fix is to allow simulator to run without potential crashing. It does not provide a complete solution to the model errors, nor does it guarantee the solutions are accurate. Therefore, users are always encouraged to fix model errors with their best knowledge of the equipment before using the validation tools to give generic solutions. Also, after using the validation tool, users should continue to work on identifying and fixing the potential model problems. The details will be discussed in the next few sections. The following figure shows the message window of validation results. The button to perform validation and auto-correction is circled in red. 1.4.2 Colstrip ATR model In 2015, a Colstrip Acceleration Trend Relay (ATR) model was developed with the collaboration of PowerWorld and Northwestern Energy. This relay allows to trip partially or wholly the Colstrip generation units from the grid when the generators’ rotor speed ramping up too fast for a certain amount of time. Such a quick increase of rotor speed is normally caused by a sudden loss of large amount of load in the system elsewhere. To protect the Colstrip units, which have large inertia 22 constants and can only gradually reduce its generation with a conventional governor control, the relay was designed to cut the generation in a much faster way. ATR was designed with complicated logic to achieve the goal of faster response while still maintaining relatively stable service to accommodate normal system disturbances. ATR action can lead to significant change in system conditions during the transient stability time frame, as it allows to drop more than 1000 MW generation in a short period (less than a second). It is therefore important to have ATR correctly modeled in any transient stability simulation software. Currently, only PowerWorld has an ATR model finalized. When using a database that has been converted from GE PSLF, users need to manually add the ATR model to the database. One way to insert the ATR model into the database is to use the following segment of aux command. Users can copy and paste the aux command into a text file and load into PowerWorld. //Insert ATR Model DATA (RELAYMODEL_ATRRELAY, [BusNum,GenID,WhoAmI,WhoAmI:1,WhoAmI:2,TSDeviceStatus,u1vsu2,u3vsu4,smvslg, lgvs2sm,Monitor,u1delay,u2delay,u3delay,u4delay,auxdelay]) { 623501 "1 " "Gen '623502' '1'" "Gen '623503' '1'" "Gen '623504' '1'" "Active" 0 0 0 0 0 0.05 0.05 0.0667 0.0667 0.1427 } Users can also manually add the ATR model using the following procedure: 1. In the bus view, users can click the Colstrip GN1 unit (circled in red) and open the generation information window 23 2. In the opened dialog, selected the stability tab (circled in red) 3. Selected the “Other Models” tab 24 4. Click “Insert” Button, in the opened dialog, select “ATRRELAY” 5. Select from the list of generators the Colstrip Units and add to the model 25 6. After adding Colstrip ATR model, it should look like After adding the Colstrip ATR models, users can find the information in the model explorer, under: Transient Stability Generator Other Models Relay Model 26 1.4.3 Test Runs The second step to check the stability data is to perform some standard simulation tests and verify the results to be consistent with expectations. The standard simulations we discuss here include three tests of the Western Interconnection system adopted by several utilities: No Fault Test Chief Joseph Braking Test Double Palo Verde Test The no fault test, sometimes referred to as a “flat run test”, is the most basic transient stability simulation test normally performed for every system. As its name implies, the no fault test will simulate the system for a period without any disturbance. If stability models are initialized correctly and the system is well damped, all trajectories of simulated quantities versus time should be flat. That’s the reason why it is also called as “flat run test”. To add the three contingencies to the simulator, users can utilize the following aux commands: //-------------------------------------------------------------------------------// THE FOLLOWING DESCRIBES THE TRANSIENT CONTINGENCIES //-------------------------------------------------------------------------------DATA (TSCONTINGENCY, [TSCTGName,Category,StartTime,EndTime,UseCyclesForTimeStep,TimeStep,CTGSkip, CTGProc,CTGSolved,ReasonNotSolved,CTGViol,TSTotalLoadMWTripped, TSTotalGenMWTripped,PLVisible,PLColor,PLThickness,SODashed,SymbolType],AUXDEF,YES) { "001: Flat Line" "" 0.000 60.000 "YES" 0.500 "NO " "YES" "NO " "Running at 6.0000" 0 0.000 0.000 "YES" -1 Default "Default" "Default" "002: Chief Joseph Brake Insertion" "" 0.000 60.000 "YES" 0.500 "NO " "YES" "NO " "Paused at 28.7500" 0 0.000 0.000 "YES" -1 Default "Default" "Default" "003: Double Palo Verde" "" 0.000 60.000 "YES" 0.250 "NO " "NO " "NO " "" 0 0.000 0.000 "YES" -1 Default "Default" "Default" } DATA (TSCONTINGENCYELEMENT, [TSCTGName,TSTimeInSeconds,WhoAmI,TSEventString,Enabled,FilterName],AUXDEF,YES) 27 { "001: "002: "002: "003: "003: } Flat Line" 0.000000 "Simulation" "SET TimeStep 1" "ALWAYS" "" Chief Joseph Brake Insertion" 1.000000 "Load '40232' 'CH'" "CLOSE" "ALWAYS" "" Chief Joseph Brake Insertion" 1.500000 "Load '40232' 'CH'" "OPEN" "ALWAYS" "" Double Palo Verde" 1.000000 "Gen '14931' '1'" "OPEN" "ALWAYS" "" Double Palo Verde" 1.000000 "Gen '14932' '1'" "OPEN" "ALWAYS" "" 1.4.3.1 No Fault Test Users can also manually insert a no fault simulation using the windows “Add” button. After adding a contingency, users will only need to change the contingency name, the end time and time step. No events will need to be specified for this no fault simulation. The first step of the no fault run is to change the end time to 0 second, and click the “run transient stability” button. This step will check for initialization problems. Initialization of a transient stability simulation means the simulator uses the power flow data (voltage, power, etc) to compute backwards the internal state values for each model. At the initialization stage, if all three elements (model parameters, model structure and the power flow setting) are accurate and modeled consistently with each other as it is supposed to be, all internal states should be at their nominal values and, more importantly, the derivative of the states should be at 0. Using the state derivative value is a good way to verify if some of the above three elements may have potential problems. As shown in the next figure, in States/Manual Control All States table, users can sort the (absolute value of) derivative of states after initialization. In general, derivatives larger than 1.0 should be marked and the corresponding states should be reviewed before users proceed to simulation of system events. Exceptions may exist for systems where certain models and certain states may have large derivatives for known reasons. 28 After fixing all initialization problems, users can proceed to run a no fault simulation. It is suggested that users should perform the simulation for several short periods, for example, every 2-5 seconds. After each period, users can pause and check the simulation plot as well as state derivatives. As shown in the following figure, in the first 2 second simulation, users may already observe some oscillatory behavior which means the simulation is not flat. Also, some state derivatives may grow quite fast during the simulation. If these are observed, users should stop and go back to fix the problem before repeating the no fault simulation until a satisfactory flat result is obtained. Some oscillatory behavior may eventually damp out by itself if not fixed, but this behavior may trigger some other unexpected dynamic behaviors to invalidate the simulation result. For this reason, it is worthwhile to fix all related problems and make sure the simulation is flat from the beginning to the end of the no fault simulation. 29 1.4.3.2 Chief Joseph Braking Test The second standard test is the Chief Joseph Braking test. In this test, an artificial load of 1400 MW will be switched on and off at Chief Joseph 230 kV substation for 30 cycles. This test will check the system response to a sudden change (increasing and decreasing) of load/generation balance. To add the artificial load into the system, a segment of aux code is used as follows: DATA (LOAD, [BusNum,BusName,AreaName,ZoneName,LoadID,LoadStatus,LoadMW,LoadMVR,LoadMVA, LoadSMW,LoadSMVR]) { 40232 "CHIEF J2" "NORTHWEST" "Central Washington" "CH" "Open" 1400.00 0.00 1400.00 0.00 } 1400.00 A typical response from the Chief Joseph Braking test is shown as follows. The system is considered as having satisfactory response if it is stable and the frequency returns back to around 60 Hz. 1.4.3.3 Double Palo Verde Test The third standard test is the double Palo Verde test. In this test, both Palo Verde units are tripped to test the system frequency response to loss of a large amount of generation. This simulation normally should last for 60 seconds to fully capture governor response and other slow transient phenomena. A typical simulation result looks like the plot as follows. The system frequency should drop after the two generators are tripped, but should subsequently return back to a stable state. Due to the droop setting of speed governors, the frequency will not return back to 60 Hz, but around 59.9 Hz. 30 1.4.3.4 Other Stability Tests The above three tests should not be considered as a complete set of tests for transient stability cases. Depending on the application, various other tests should be performed to further verify the data. These tests may include a 3 phase fault test with normal clearing, a stuck breaker test with remote clearing and/or reclosing, or a critical clearing time test, etc. 31 1.5 Add Contingencies Users can add contingencies either manually, or using aux files. In this section, we will discuss the procedure for manually adding or changing stability contingencies, with several steps to check contingencies for their accuracy. 1.5.1 Manually Add Stability Contingencies First, users can insert a contingency using the “add” button. The contingency will be given a default name such as “My Transient Contingency X”. Users can change the name of the contingency using the “Rename” Button. Both buttons are shown in red circle. After inserting the contingency, users can specify the events in the contingency by clicking the insert button at the bottom, circled in red in the figure below. A window will be pop up with the options to specify the event. To specify a contingency event, one can first select the objects from the list in the mid-left side. An object can be Branch/Transformer, Bus, Generator, etc. With the object type selected, users can choose (on the right side) the element according to their bus numbers or bus name. On the bottom left, the action type should be selected. The action can be applying a fault, clearing a fault, open or close a device, etc. The bottom right part further specifies the action parameter details. The time of the action should be specified in seconds in the blank above the selection of actions. 32 For example, if a user wanted to insert a single line to ground fault at the middle of 115 kV line from Aspen to S. Joseph at 1.1 second, his selection will be: 33 Users can continue to add events for a contingency scenario until all events are added. After a contingency definition is finalized, it can be exported into an aux file and passed to other people easily. To save contingencies in an aux file, one can click the “Save All Settings To” button at the bottom left corner of the simulator, select “Save Auxillary”, specify the file name in the popup window, and select “Save Transient Stability Events.” 1.5.2 Consistency Checking of Stability Contingencies Sometimes users may find a dynamic simulation produces an unexpected unstable result. A lot of unexpected instability may actually be caused by errors in the contingency definition rather than modeling problems. Therefore, it is always worthwhile to check the contingency definition and make sure it is accurately modelled. In this section, we will summarize some potential problems and make some suggestions for modeling events in a contingency. 1.5.2.1 Simulation time not long enough to capture the last event A common mistake for defining a contingency is that the simulation time for this contingency is too short to fully capture the system response after the last event. For example, a 10 second simulation can typically capture good enough information on a three phase fault contingency with normal clearing. However, 10 seconds may be too short a time frame to simulate large frequency events where generations or load are tripped in significant amounts due to the fact that a large amount of governor response can take much longer to reach a stable condition. Therefore, these types of contingencies normally need a longer time period to simulate. 34 Some other contingencies may include delayed clearing, reclosing events, RAS or SPS actions. Those events may be triggered seconds or minutes after previous events. A simulation time should give enough time for the system to fully respond to the last event. Other type of factors that may impact the simulation time choice could be switching devices with a longer delay time. For example, a 10 second simulation time won’t be able to capture the system dynamics for a capacitor with a 15 seconds switching delay. Since these switching actions may not be explicitly known before the simulation, users are suggested to start with longer simulation times for a new contingency. 1.5.2.2 A fault was never cleared Some users may forget to clear a fault in a contingency definition. If so, the fault will stay permanently in the system and may eventually lead to some extreme conditions where the simulation shows an unstable result. Such mistakes can happen easily for a bus fault where all remote breakers will be opened with various time delays. Due to topology changes or a breaker status change, some breakers may fail to be opened in the contingency definition resulting in the fault not being cleared. Users are always encouraged to check the topology of the base case with the contingency definition to verify that all faults have been cleared successfully. Certain techniques may be helpful to model contingencies. For example, whenever a fault has been changed, such as a stuck breaker changes a three phase fault into a single line to ground fault, users can always first clear the three phase fault, and then add a new single line to ground fault with the new location or impedance. By doing so, he can guarantee that faults are added and cleared in pairs to prevent a fault from remaining on the system through the entire simulation. 1.5.2.3 Generation, load or branches are not tripped When a contingency is trying to open a generator, a load or a branch which has already been opened, this event will normally being ignored. This can happen easily with topology constantly changing. If so, it might lead to unexpected instability, for example, when a RAS action is designed to trip a certain amount of load to mitigate generation loss while these loads are offline. Fortunately, these events are normally marked by the simulators in the log file. It is suggested that usersreview the log file and check for these issues. 1.5.2.4 Long fault clearing time One common reason for a system going unstable unexpectedly is the fault being cleared too late. System upgrades, topology changes, devices online or offline may easily change the critical clearing time at some fault locations. Under these conditions, an existing fault clearing time may need to be reevaluated with studies. It is possible that existing contingency definitions have not been updated with the correct clearing times. 35 1.6 Define Plots Ifusers plan to perform dynamic simulation for a large number of contingencies, it would be more convenient to first define the plot channels. This helps to save users a lot of time by generating the plots immediately after each simulation. Otherwise, users will need to load the individual result files back into the simulator which could be very time consuming. It is suggested that users save the simulation results into a *.tsr file for future reference. In this section, we will go through the basic steps of defining the result storage and plots. 1.6.1 Result Storage Users can specify the result storage options in the simulation window by selecting “Result Storage” in the left panel. In the option “where to save/store results”. Users should check the “Save Results to Hard Drive” option to enable simulation results to be saved as *.tsr files. Also we recommend that users uncheck the “Store Results to RAM” option for automated simulation procedure to save some memory. In order to reduce disk space usage, users may choose to adjust the number in “Save Results Every n Timesteps”. If n=1 is used, every time step of the simulation will be saved in the result file. This file will provide the best precision of simulation results but will use a lot of hard disk space. Using a larger 36 number can reduce the file size significantly while still maintaining a certain degree of precision. Users should always use an odd number in this option, such as 1, 3, 5, 7, etc. This is shown in the following Figure. Using an even step for plot or storage may lose the information of simulated oscillatory behavior. In the “Save to Hard Drive Options” Tab, users can specify the hard disk location for storage of the *.tsr and *.aux files. There is no need to give a name for each contingency result file as it will be automatically named using the contingency name. Currently, PowerWorld only allows storage of channels with the Area/Zone filter. Individual device quantities cannot be specified. This may lead to a large amount of data being stored. Users can also select the type of information to save, including generators, buses, 37 loads, switched shunts, branches, line shunts, DC lines, Multi-terminal DC, MTDC converters, Areas, Zones, Substations, Interfaces, Injection Groups, System, Measurement Objects. 1.6.2 Plot Designer After selecting the storage option, users can define the plot channels by using the Plot Designer. In Plot Designer, users should first select the device type. Then a list of quantities associated with that type of device can be selected for each device as shown in the following Figure. It is recommended to put same type of quantities in a single plot for consistency of the plot axis ranges. For example, if users want to plot both generator rotor speeds and terminal voltages for a group of generators, they can add two plots: one for rotor speeds and one for terminal voltages. A single plot can also contain multiple sub-plots. After defining the plot, users can change the plot name on the right panel, clicking the “Rename Plot” button. 38 The best way to generate plots is to generate them immediately after each simulation. This saves a significant amount of time for simulator to load back the output files just for generating the plots. To specify this, users need to select each plot and choose from the drop down menu for “When” under “Auto-Save an Image File of the Plot” as “After each contingency”. Also users can specify the file type with pixel requirement. By doing so, the simulator will automatically create all plots and save to the specified folder for each contingency. 39 Under the “Title Block” tab, users can add titles to each of the plots. A title can be information that a reader can use to easily identify the plot, such as the contingency name, case name or other information. Users can also insert their company’s logo into the plot. This can be specified at the bottom of the window. In the last tab, “Plot Series List”, users can change an individual channel’s plot properties. Users can select different types of line, thickness, color, etc. 1.6.3 Generic Plot Channels In the ColumbiaGrid transient stability workshop, ColumbiaGrid members decided to use the following information as the generic plot channels for simulation. These channels include: 1. Real power, reactive power, mechanical power, rotor speed, terminal voltage, field voltage, field current, and rotor angle of generators 2. Voltage and frequency for buses 40 3. MW flow for interfaces The generators for consideration includes three types: 1. Large Thermal Units in ColumbiaGrid footprint Large Unit HERM 1G FREDONA2 LIB 01 ROCKYR02 MCN 02 KFALLCT1 LANCAS G COYO G1 GEP G1 CHEH G1 DWOR3 PORTW G1 RVR RD C HPP G1 GRYHB G1 CENTR G1 CGS Bus Number 45454 42112 44191 46842 44102 45448 47568 43111 47687 47588 40365 43905 47216 47639 47596 47740 40063 Unit ID 1 2 1 C2 2 1 1 1 1 1 1 1 1 1 1 1 1 2. Large Hyrdo Units in ColumbiaGrid footprint Hydro Unit MNTFRM G LO BAKER UP BAKER PELTON PRIEST01 LWG 0102 DETROIT HUNGHR12 NOXON12 CHELAN BOUNDG51 LGS 0102 LMN 0102 ICE H1-1 Bus Number Unit ID 47675 1 42121 1 42124 1 43407 1 46170 1 44231 1 40344 1 40555 1 48285 1 46803 A1 46464 51 44211 1 44251 1 40559 1 41 Hydro Unit Bus Number Unit ID WANAPM01 46180 1 CHJ 0304 44142 3 TDA 0304 44042 3 BON 0304 44002 3 JDA 1112 44076 11 COULEE22 40296 1 3. Large Generation Units in Neighboring system Neighbor Gen Bus Number Unit ID Colstrip GN 1 623501 1 Colstrip GN4 623504 1 Colstrip GN3 623503 1 Colstrip GN2 623502 1 PALOVRD1 14931 1 REV 13G1 50644 1 BRIDGER1 65386 1 DIABLO 1 36411 1 INTERM1G 26039 1 SJUAN_G1 10318 1 COMAN_3 70777 C3 GENES 39 54490 3 The high voltage buses for plotting include: HV bus ASHE BELL BPA BELL S1 MIDWAY MIDWAY MCNARY MCNRY S1 JDA PH1 WAUTOMA TDA PH3 CHIEF J2 MONROE RAVER CENTR P1 Bus Number 40061 40091 40086 30060 30970 40723 41351 44086 41138 44057 40232 40749 40869 47741 KV 500 500 230 500 230 500 230 500 500 230 230 500 500 500 42 HV bus Bus Number ALLSTON 40045 KEELER 40601 OSTRNDER 40809 MARION 40699 COULEE 40287 GARRISON 40459 SUMMER L 41043 MALIN 40687 GRIZZLY 40489 HEMINWAY 60155 ALVEY 40051 OLYMPIA 40797 BOUNDRYE 40145 FRANKLIN 40443 N LEWIST 48255 WALAWALA 45327 RESTON 40883 CAPTJACK 45035 KV 500 500 500 500 500 500 500 500 500 500 500 500 230 230 230 230 230 500 The Interface for plotting includes: Interface COI IDAHO - NW MONTANA - NW NW - CANADA PDCI WEST OF HATWAI WEST OF CASCADES - NORTH WEST OF CASCADES - SOUTH ALBERTA - BRITISH COLUMBIA NORTH OF JOHNDAY MID POINT - SUMMER LAKE SOUTH OF ALLSTON NORTH OF HANFORD WEST OF SLATT WEST OF MCNARY 43 1.7 Perform Simulation Users have the option to perform the dynamic simulation in PowerWorld for contingencies either manually, or in an automated way. This option is set as “One contingency at a Time”, and “Multiple Contingencies”. Users can select from these two options depending on their application. 1.7.1 One Contingency at a Time (manual simulation) The option to run simulation for one contingency at a time or multiple contingencies can be selected from the lower left part of the window (in red). The user interface for “One Contingency at a time” looks like: If users select a contingency from the drop down menu on the top (circled in red), the simulator will automatically stop after this contingency simulation. By choosing this option, users have more flexibility to pause during a simulation and review the results. This option should be used in the debugging mode for contingencies that may cause problems. 1.7.2 Multiple Contingencies (automated simulation) When the option “Multiple Contingency” is selected, the user interface will look like the following. In the main window, all contingencies will be listed with their properties such as: start time, end time, and time steps. Also, if a simulation has been performed previously for a contingency and their result summary has been saved, several useful information columns will be shown in this window: 44 If a simulation was performed successfully to the specified end time, the “Solved” property of the contingency will be marked as “Yes”. On the other hand, if a simulation crashes before reaching the end time, “Solved” will be “No”. A reason for the crash will also be shown in the next column. This helps to easily identify contingencies that don’t solve in a large pool of contingencies. Also, PowerWorld simulator automatically generates the number of violations, generation & load tripped during the simulation, and islanded system information on this page for quick review. To perform the automated simulation for only a portion of a whole list of contingencies, the user can change the contingency property “Skip” to “yes” for all contingencies to be skipped. 45 1.8 Evaluate Results After finishing the simulation and obtaining plots and output file, users should review the results for unexpected dynamic behaviors. In general, a few issues should always be marked down if seen. Other issues may be considered problematic depending on the application of the transient stability study. In this section, we will discuss some general guidelines for evaluating transient stability results. 1.8.1 Simulation terminated before the end time If a simulation is terminated by simulator before the specified end time, either crashing the software or not, such a simulation should always be marked as problematic. Following is a list of potential reasons that may contribute to premature termination of a simulation: 1. 2. 3. 4. 5. Numerical instability System instability Model errors Power flow case unsolvable or collapse Contingency definition errors 1.8.1.1 Numerical instability It may look strange to users that a well modeled system can still experience a crashing simulation. If this happens, it may be caused by numerical instability. Numerical instability can be triggered by the numerical integration algorithms utilized by the simulator. Algorithms commonly used in commercial software include the Euler method, Runge-Kutta method, trapezoidal integration, etc. Though all these methods demonstrate efficiency for solving large scale dynamic problems, implicitly, they all are subject to numerical instability issues. In practice, such numerical instability can be associated with extra small time constants or extra-large gains in some device models. With the latest trend of adding more electronics related devices such as FACTS, solar PV, wind, or storage, all these converter based devices respond extremely fast and induce quite small time constants. When modeled without consideration of the simulator’s capability to handle small time constants, such devices may cause numerical instability. Numerical instability is normally shown in the simulation plots as very high frequency sustained oscillation. For debugging purposes, disabling certain devices can efficiently restore the simulation back to normal. In practice, reducing simulation time steps may resolve the numerical instability problem but leads to longer simulation time. It is suggested that models that are identified with numerical problems be reported to the simulator vendors (GE, Siemens, PowerWorld) so they can develop better modeling techniques to handle such issues. 1.8.1.2 System instability The most common reason for a simulation to crash is system instability. When a simulation of a bulk system goes unstable, lots of device models may operate outside their normal range. In extreme conditions, an input/output value of a device can trigger issues that crash the simulation, e.g., divided by 0 errors. 46 Sometimes, it is relatively hard to distinguish numerical instability with system instability. Under both conditions systems are experiencing instability behavior. However, numerical instability is normally caused by one (or a few) device models with small time constants or large gains, and it doesn’t reflect what could happen in reality. Also they can be fixed easily by disabling or correction of these models. For system instability, it is a system condition that can happen in reality with lower frequency oscillation. The solution to system instability normally requires mitigation procedures or actual system additions or upgrades. Due to the fact that there are so many possible reasons for an unstable simulation result, the solution to an unstable simulation can be varied. Any unexpected instability scenario should be evaluated in more detail and a solution sought with coordinated planning and operation efforts. 1.8.1.3 Model errors Another likely source of unexpected instability or software crash is model errors. In this section, model errors refer to both model parameter errors as well as model programming errors. Poorly written models and/or bad parameters can both crash a simulator, while often escaping the attention of users. In reality, these types of problems are among the most difficult ones to be found and fixed. To help users track down these types of errors, each simulator has tried to provide as much information as possible to locate these sources of error. In PowerWorld, the simulator will log most problems it finds during the simulation process. Therefore, users should always check the log when some unexpected crash happens. Normally, the last events before the crash may provide information on which model might be responsible for the problem. Often, a model itself does not contribute alone to the crash. The mismatch between the stability model and power flow may be the reason. For example, during a simulation, terminal voltage for some generator may go too high or too low, exceeding the normal operating range of the units, resulting in the generator model exhibiting unexpected behavior. Users should also check the power flow condition when reviewing a problematic stability model. WECC Model Validation Working Group (MVWG) has several task forces to review and fix model errors detected in the system. Power Plant validation task force (PPMVDTF) has produced a list of potential model errors and is working with owners to get these errors fixed. The list of errors can be requested by WECC members through Kent Bolton who coordinates WECC MVWG (Kent@wecc.biz) 1.8.1.4 Power flow case unsolvable or collapse Sometimes a simulation will crash due to transmission network problems, such as voltage collapse resulting from severe disturbances. These problems can normally be identified by reviewing the power flow conditions. To view the problem, users can plot bus voltages around the area of interest and observe if the voltages are experiencing constant decreasing or a sudden drop due to system events, especially after the fault is 47 cleared. A comparison between the pre-fault voltage and the post-fault voltage may also provide better information on whether the system is experiencing voltage collapse. For high voltage systems where the nominal KV is above 230 kV, voltage levels remaining below 0.7 pu will be considered as collapse. Even voltage levels that drop below 0.85 pu may be considered as collapse in some cases. For lower voltage systems with a nominal voltage below 115 kV, somewhere below 0.6 pu would represent a high probability of voltage collapse occurring. Converter based devices such as wind turbines, solar PV, storage, etc have almost no inertia. During contingency events, these devices can drastically increase or decrease their output in a short period and lead to collapse in a weak system. Fortunately, such unsolvable power flows are always logged by the simulator for users to check. If a device model constantly causes power flow solution problems, it should be reported to the simulator vendor for updating. 1.8.1.5 Contingency definition errors As described in 1.5.2, some contingencies may result in instability due to errors in the contingency definition which fail to clear the fault and leave it permanently on the system. This also includes inaccurate modeling of RAS or special protection systems in the contingency descriptions. Such problems can be identified by careful review of the system topology and comparison with the contingency definition. 48 1.8.2 Unstable simulation A stable simulation may look unstable and vice versa. In this section, we will discuss some typical stable vs unstable simulation scenarios. It is worth mentioning that dynamic simulation is only a tool to evaluate system response. Any issues or concerns identified by the simulation should be examined carefully until a reasonable explanation can be derived to interpret the results. Interpretation should reflect the anticipated performance of physical devices for the simulated events. Any unexpected issues or abnormal behaviors reported by the simulation results should be considered as potential problems that may happen in reality. Additional review or examination of these problems is required. 1.8.2.1 Undamped or sustained oscillation Undamped or sustained oscillation, regardless of magnitude, should always catch the attention of users. In most of cases, such oscillations would be considered unstable. 1.8.2.2 Constant Device Switching Like sustained oscillation, if a control device, including an exciter, governor, FACTS, capacitor/shunt, HVDC, load, etc, is experiencing endless switching during a simulation, it is an indication that some switching criteria cannot be satisfied by the system conditions. Most likely, it is a good indication that the system is unstable. 49 1.8.2.3 Pole slipping or machines oscillate against each other Sometimes during a fault, a machine or a group of machines close to the fault location may accelerate much faster than the rest of the machines in the system. If the fault is not cleared in time, these machines may accelerate to a much higher speed and then, when the fault is cleared, they go out-ofstep. Considering the fact that most generator protection relays (eg, out-of-step relays) are not modeled in the transient stability data, the simulation may show that all the machine angles eventually settle down to a stable rotor angle which is greater than 360 (electrical) degrees from the other machines. Even though the simulation plots of these machines appear quite stable, this is a typical scenario of instability where some machines have actually gone out-of-step. 1.8.3 Stable Simulation If the post fault system plots are all flat, it is a good indication of a stable simulation. Even if the simulated trajectories are not flat, it may also be considered as stable simulation. For example, the rotor angles in the following plot are not flat but moving together. The relative angle difference between rotor angles are quite stable. The reason for these rotor angles moving is that, after a fault, the system supply/demand balance can no longer be maintained. Fault events may cause generation or load tripping during the fault with other generation in the system attempting to pick up the difference until the droop control set point is reached. At this point, the system will be stable at a frequency slightly below or above 60 Hz. The synchronous machine speeds will be similarly slightly below or above 60Hz. The rotor angles therefore keep increasing or decreasing. 1.8.4 Generation/Load tripping An important result to monitor from dynamic simulations is generation and load tripping. PowerWorld automatically summarizes the generation and load tripping information in the summary page of the 50 transient stability tool, as shown below. If the generation and load tripping amounts are not as expected, users should review the simulation log files to track down all unexpected generation or load tripping involved. It is worth mentioning that any generation or load tripping resulting from a RAS or SPS action, where the tripping is part of the contingency definition, is not counted in the tripping summary page. 1.8.5 WECC limit violation As described in Section 1.3.7, WECC limit violations will be automatically recorded if limit monitors are set up prior to the simulation. The results can be tracked in the Simulation->Violation tab and Transient Limit Monitors -> Monitored Violations tab. 51 Part 2: Reference Documents 2.1 Introduction to Power System Stability ColumbiaGrid would like to thank Mr. Don Johnson from Portland General Electric for generously providing his training presentation for “introduction to power system stability”. The slides serve as a good reference for understanding the basic concepts of transient stability. Transmission Planning Power System Stability Don Johnson Senior Planning Engineer T&D Planning PGE Company PGE Company Confidential Training Power System Stability – Outline Transient Stability – Questions to be Answered − − − − − What is Transient Stability? How it is different from “Power Flow” Analysis? What models to use? Why is it important? How to conduct studies? Elements of Transient Stability − Stability Fundamentals − Different Types of Power System Stability − Transient Stability Models − What are the important models and parameters? Running Power System Stability Studies − Transient Stability Studies − Voltage Stability – PV & QV Studies 2 PGE Company Training Stability • Definition of Stability: Power system stability is the ability of an electric power system, for a given initial operating condition, to regain a state of operating equilibrium after being subjected to a physical disturbance, with most system variables bounded so that practically the entire system remains intact. Source: IEEE/CIGRE Joint Task Force on Stability Terms and Definitions, “Definition and Classification of Power System Stability”, IEEE Transactions on Power Systems, 2004 3 PGE Company Training Power System Stability Classification 4 PGE Company Training Power System Stability Definitions of Three Types of Stability • Rotor Angle Stability – Ability of synchronous machines of an interconnected power system to remain in synchronism after being subjected to a disturbance. • Voltage Stability – Ability of a power system to maintain steady voltages at all buses in the system after being subjected to a disturbance from a given initial operating condition. • Frequency Stability – Ability of a power system to maintain steady frequency following a severe system upset resulting in a significant imbalance between generation and load. 5 PGE Company Training Transient Stability Analysis Time Frame 6 PGE Company Training Power Flow Program • How does Transient Stability Program compare to the Power Flow Program? • First, need to understand what the power flow program accomplishes • A power flow program is used to determine a “steady-state” operating condition for a power system, (assumes “steady-state” condition both before and after a disturbance) − Goal is to solve a set of algebraic equations of the form: g(y) = 0 {y variables are bus voltage and angle} 7 − Models used reflect the steady-state assumptions (i.e. generator PV buses, constant power loads, LTC transformers, etc) − Assumes a constant generator/load balance where the frequency is constant PGE Company Training Transient Stability Program • A transient stability program is used to determine how the system responses from an initial stable operating point, how the power system responses through time because of a disturbance, and returns to a new “steady-state” operating point. − Important point – Loads, generation, voltages, and frequency can change with time! − Goal is to solve a set of differential and algebraic equations of the form: dx/dt = f(x,y) {y variables are bus voltage and angle} g(x,y) = 0 {x variables are dynamic state variables} f - primarily represents the generator dynamics g - primarily represents the bus power balance equations − Assumes the system starts from a steady-state, and returns to a new steady-state − Frequency, voltages, generation and loads are “state” variables; thus can change with time − Models reflect the transient stability time frame (up to dozens of seconds) Some values assumed to be slow enough to be constant (LTC tap changers, AGC action, etc.) Others values are still fast enough to treat as algebraic (synchronous machine stator dynamics, voltage source converter dynamics, etc) − 8 In order to solve the complexity of the differential equations, numerical methods are used Requires an initial value of x0 be known to determine initial state variables, f(x) = 0 Need to determine x(t) for future time. PGE Company Training Numerical Solution Methods • Numerical solution methods do not generate exact solutions: they are an approximation and thus introduce some error − Assumes time advances in discrete increments, called a step size (also known as time step), ∆t − − Speed versus accuracy tradeoff: a smaller ∆t gives a better solution, but it takes longer to compute Numeric roundoff error due to finite computer word size • Key issue is the derivative of x, f(x) depends on x, the value that is trying to be determined − A solution exists as long as f(x) is continuously differentiable • There are a wide variety of numerical solution approaches (Powerworld has a choice of two); both require information about solution at one point, x(t) − Forward Euler − Runge-Kutta 9 PGE Company Training Numerical Solution Methods - Errors • At each time step the total round-off error is the sum of the local round-off at the current time and the propagated error from each step 1, 2, … , k-1 • An algorithm with the desirable property that local round-off error decays with increasing number of steps is considered to be numerically stable • Otherwise, the algorithm is numerically unstable • Numerically unstable algorithms can still provide good performance if appropriate time steps are used − This is especially true when coupled with algebraic equations 10 PGE Company Training Numerical Method – Euler’s • One of the techniques for numerically integrating differential equations is known as Euler’s Method (sometimes referred as the Forward Euler’s Method) − Key idea is to approximate the derivative of x: ẋ = f(x(t)) = as Then x(t +∆t) ≈ x(t) + ∆t ∗ ( t ) • In general, the smaller the ∆t (time step), the more accurate the solution. However, it also takes more time steps 11 PGE Company Training Numerical Method – 2nd Order Runge-Kutta • Runge-Kutta method improves on Euler’s method by evaluating f(x) at selected points over the time step • Simplest method is the second order method in which: x(t +∆t) = x(t) + ∗ ( 1 + 2) where • 12 1 = ∆t ∗f(x(t)) 2 = ∆t ∗f(x(t) + 1) is what comes from Euler’s; 2 improves on this by reevaluating at the estimated end of the time step and then average the two 1 PGE Company Training 2nd Order Runge-Kutta Versus Euler’s • 2nd Order Runge-Kutta method requires twice the function evaluations per iteration, but give better results • With 2nd Order Runge-Kutta method the error tends to vary with the cube of the step size, compared with the square of the step size for Euler’s • Thus, the smaller error allows for larger step sizes compared to Euler’s • One thing to remember, the models use time constants and thus it is required that the time step ∆t used is smaller than the smallest time constant in the stability models used 13 PGE Company Training Transient Stability Models • Transient Stability Studies are Highly Dependent on the Models Used − Should use the best models for the values available for the model used (However: Preoccupation with detail is a SUREWAY to end up with misleading results). Should always consider “what have I assumed” and “how may it impact my studies“? − Models available include: Generator models Excitation System models Power System Stabilizer models Governor models Dynamic Load models SVC models DC terminals and DC Line (If present) Relay models • Requires knowledge of both the mechanical and electrical properties of the combined system 14 PGE Company Training Generator Modeling • For dynamic simulation calculations used in the transient stability program, the connection of the generator to the power system network is modeled as the Norton Generator Equivalent: • Important to note, this equivalent, is the only time that the “generator” subtransient reactance is used. This generator “reactance” is not used in the power flow program. 15 PGE Company Training Generator Unit Stability Models Physical Structure of Power System Components 16 PGE Company Training Transient Stability Models Model Classes 17 PGE Company Training Transient Stabilty Models Turbine - Generator Modeling • Complete Generating Unit Model has several classes of models assigned to them − Governor (Mechanical Representation – Boiler/Turbine Characteristics) − Machine (Generator Physical Electrical Characteristics for Stator & Rotor) − Exciter (Excitation Source/Voltage Controller) − Power System Stabilizer (External Excitation Stabilization) • Other Misc. Turbine-Generator Models − Excitation Limiters (OEL, URAL, etc) − External Voltage Compensation (Remote Voltage Regulation) − Generator Protection Relay Models (Loss of Excitation, Volts/Hz, etc) − Turbine Load Controllers 18 PGE Company Training Important Input/Output Values for Models 19 PGE Company Training Transient Stability – Generator Model Generators - Important Parameters Needed • Generator/Turbine (Mechanical part) − Mechanical Characteristics Inertia Constant (Combined for Turbine & Generator if on the same shaft) • Generator (Electrical part) − Electrical Characteristics Reactances, Time Constants, Saturation Stator Winding and Voltage Capabilities Field Winding Capabilities Step-up Transformer Capabilities 20 PGE Company Training Transient Stability Models - Generator • Simplest Model Is the Classical Generator Model – GENCLS • Represents a Synchronous machine with constant voltage behind a transient reactance. (Assumes constant excitation and mechanical power) • Note: Normally used only for “academic studies”; as it is only valid for transients up to about one second, thus should not be used …. Variable Default Data Variable Description No exciter or governor Assumes constant mechanical power Damping is negligible 21 PGE Company Training Transient Stability Models - Generator • Recommended detailed synchronous machine model - GENTPJ • Represents a Synchronous machine (either round rotor or salient pole) with stator and rotor dynamics modeled along with saturation • WECC recommended model for synchronous generators …. Variable 22 Default Data Variable Description PGE Company Training Information for Generator Modeling Important Generator Capabilities to Know 23 PGE Company Training WECC Recommended Generator Models GENERATOR MODELS GE PSLF PTI PSS/E* PowerWorld Simulator gentpf genrou gentpj gencc pvd1 regc_a wt1g wt2g GENROU/IEEEVC GENROU/IEEEVC GENTPJU1 GENROU/IEEEVC REGCAU1 WT1G1 WT2G1 GENTPF GENROU GENTPJ GENCC PVD1 REGC_A WT1G and WT1G1 WT2G and WT2G1 gencls not used GENCLS IEEE Standard Status approved approved approved approved approved approved approved approved Comments 8/11/06 8/11/06 1/23/09 8/11/06 3/19/14 3/19/14 1/21/11 8/28/09 WECC Model Round rotor generator model, use for thermal generator models modified gentpf with improved saturation modeling Cross Compound generator model Distributed Photovoltaic system model Generator/converter model for Photovoltaic, Wind type 3/4 Wind Type 1 generic generator model Wind Type 2 generic generator model Used to force a signal, or classical generator model PTI/GE/PowerWorld Comments Available in PSS/E version 33.2 In PSLF 17 and PSSE32 We have a GENCLS model. The PSLF model gencls does get converted to the PSS/E model GENCLS. [Forcing signal (playback) feature not needed in library datasets.] Where different variants of the same model exist, the preferred version for submittal to WECC are highlighted in green These models currently are not converted from PSLF to PSS/E. 24 PGE Company Training Excitation Systems Purpose of Excitation System • Functions of the Excitation System − − Provide direct current to the synchronous generation field winding Perform control and protective functions essential to the satisfactory operation of the power system • Performance of the excitation system is determined by: − Generator Considerations Supply and adjust field current as the generator output varies within its continuous capability Rotor insulation failure due to high field voltage Rotor heating due to high field current Stator heating due to high VAR loading Heating due to excessive flux (Volts/Hz) − Power System considerations – contributes to effective control of system voltage and improvement of system stability. 25 PGE Company Training Exciter Models Excitation System Models Three Basic Types • DC Excitation Systems • AC Excitation Systems • Static Excitation Systems Rotating Exciters • • Brushless (No Slip Rings) Brush Type Static Exciters (Power Source) • • 26 Shunt – (Generator Output Voltage) Series – (Derived from Generator Output Voltage & Current) PGE Company Training Exciter Model - SEXS SEXS Excitation Model – Simplified Excitation Model • Use when no information is known about an exciter Variable 27 Default Data Variable Description PGE Company Training Exciter Model – EXST1 EXST1 Excitation Model – IEEE ST1 Excitation Model • Represents a static controlled-rectifier excitation system with an AC power source fed from the generator terminals Variable 28 Default Data Variable Description PGE Company Training WECC Excitation Models EXCITATION SYSTEM MODELS (Volt/Var Control Models) GE PSLF PTI PSS/E* PowerWorld Simulator exac1 esac1a exac1a exac2 esac2a exac3 EXAC1 ESAC1A EXAC1A EXAC2 ESAC2A EXAC3 EXAC1 ESAC1A EXAC1A EXAC2 ESAC2A EXAC3 IEEE Standard AC1A AC1A AC2A Status approved approved approved approved approved 8/11/06 1/21/11 8/11/06 8/11/06 1/21/11 Comments PTI/GE/PowerWorld Comments Brushless AC 2005 IEEE standard - updated AC1A with OEL/UEL inputs exac1 with altered rate feedback source HIR Brushless 2005 IEEE standard - updated AC2A Not used in WECC database Differs from IEEE AC1A -- does not have OEL/UEL inputs and multiplies output by speed. Differs from IEEE AC3A -- no OEL/UEL inputs; different field current limit; speed multiplier, PSS/E Model same as IEEE AC3A model exac3a ESAC3A EXAC3A AC3A approved 8/11/06 GE Alterrex (rare) esac3a exac4 esac4a esac5a exac6a esac6a esac7b exac8b esac8b exbbc exdc1 esdc1a exdc2 exdc2a esdc2a exdc4 esdc3a esdc4b exeli exst1 esst1a exst2 exst2a esst2a exst3 exst3a esst3a ESAC3A EXAC4 ESAC4A ESAC5A ESAC6A ESAC6A AC7B ESAC8B AC8B BBSEX1 IEEEX1 ESDC1A EXDC2 EXDC2 ESDC2A IEEET4 DC3A DC4B EXELI EXST1 ESST1A EXST2 ESST2A ESAC3A EXAC4 ESAC4A ESAC5A EXAC6A ESAC6A ESAC7B and AC7B EXAC8B ESAC8B_GE and AC8B EXBBC and BBSEX1 EXDC1 and IEEEX1 ESDC1A EXDC2_GE and EXDC2_PTI EXDC2A and EXDC2_PTI ESDC2A EXDC4 and IEEET4 ESDC3A and DC3A ESDC4B EXELI EXST1_GE and EXST1_PTI ESST1A and ESST1A_GE EXST2 EXST2A ESST2A EXST3 EXST3A ESST3A AC3A AC4A AC4A AC5A AC6A AC6A AC7B ESAC8B AC8B approved approved approved approved 2005 IEEE standard - updated AC3A Rotating AC with controlled rectifier (Althyrex) (rare) 2005 IEEE standard - updated AC4A Simplified brushless exciter Alternator, noncontrolled rectifier, lead-lag 2005 IEEE standard - updated AC6A 2005 IEEE standard - new Brushless exciter with PID voltage regulator 2005 IEEE standard - updated AC8B Static with ABB regulator Rotating DC 2005 IEEE standard - updated DC1A Rotating DC with terminal fed pilot, alternate feedback Rotating DC with terminal fed pilot 2005 IEEE standard - updated DC2A Rotating, noncontinuous - minor differences between models Rotating, noncontinuous Rotating DC with PID Static PI transformer fed excitation system Static with double lead/lag EXST3 ESST3A DC1A DC1A DC2A DC2A DC3A DC3A DC4B ST1A ST1A ST2A ST2A ST3 ST3A ST3A approved approved approved approved approved approved approved approved approved approved approved approved approved approved approved approved approved approved approved approved approved approved 1/21/11 8/11/06 1/21/11 1/21/11 1/21/11 1/21/11 8/11/06 1/21/11 8/11/06 8/11/06 1/21/11 8/11/06 8/11/06 1/21/11 8/11/06 1/21/11 1/21/11 8/11/06 8/11/06 1/21/11 8/11/06 8/11/06 1/21/11 8/11/06 8/11/06 1/21/11 In both programs Differs from IEEE AC2A -- no OEL/UEL inputs; different field current limit; speed multiplier In both programs In both programs In both programs Differs from IEEE AC4A -- no OEL/UEL inputs In both programs In both programs Differs from IEEE AC6A -- no OEL/UEL inputs; speed multiplier, not a new model for PSS/E (model already exists) In both programs In both programs Differs from IEEE AC8B -- no exciter upper limit; added input limits and speed multiplier In both programs In both programs Differs from IEEE DC1A -- no UEL inputs; speed multiplier In both programs Differs from IEEE DC2A -- no UEL inputs; speed multiplier In both programs If Kr = 0, should convert to IEEEX4 (IEEE DC3A). Model added in PSS/E -32. In both programs In both programs Differs from IEEE ST1A -- no OEL/UEL inputs; added Xe Ifd loading; RFB before field current limiter. In both programs SCPT - lead/lag block (Tc, Tb) added lead/lag block (Tc, Tb) is included to match the WECC FM 2005 IEEE standard - updated ST2A Use for GE Generex 2005 IEEE standard - updated ST3A Differs from IEEE ST2A -- no UEL inputs; added lead/lag. Differs from IEEE ST2A -- no UEL inputs; fewer time constants. exst4b ESST4B EXST4B ST4B approved 8/11/06 GE EX2000 bus fed potential source, static compound and Generrex-PPS or -CPS, and Differs from IEEE ST2A -- no OEL/UEL inputs SILCOmatic 5 excitation systems, with proportional plus integral (PI) voltage controller esst4b esst5b esst6b esst7b ieeet1 ESST4B ST5B ST6B ST7B IEEET1 ESST4B ESST5B and ST5B ESST6B and ST6B ESST7B and ST7B IEEET1 ST4B ST5B ST6B ST7B approved approved approved approved approved 2005 IEEE standard - updated ST4B Variation of ST1A (New IEEE Model) Variation of ST4B with field current limit (New IEEE model) Static with limiters (Alstom) (New IEEE model) Old type 1 In both programs In both programs In both programs In both programs 1/21/11 1/21/11 1/21/11 1/21/11 8/11/06 pfqrg Not used PFQRG Power factor / Reactive power regulator The output of this model feeds into an exciter as the stabilizer input, thus this model can not be used in conjunction with another stabilizer rexs scrx sexs REXSYS SCRX SEXS REXS SCRX SEXS_GE and SEXS_PTI approved 8/11/06 approved 8/11/06 General Purpose Rotating Excitation System Model intended for use where negative field current may be a problem for use where details of the actual excitation system are unknown and/or unspecified PSS/E has a SEXS (simplified excitation system) model (which is similar to the PSLF sexs model but without the PI control block) oel1 Not converted (277) OEL1 approved 4/27/12 Over excitation limiter Please note that this is not an IEEE standard model. GE developed this model for WECC use. If we have to provide a corresponding PSS/E model, we have to get the block diagram from GE. Presentation at March 2012 M&VWG meeting, use OEL1. Has required functionality. uel1 uel2 wt2e reec_a reec_b reec_c UEL1 UEL2 WT2E1 REECAU1 REECBU1 REECCU1 Not Used Not Used WT2E and WT2E1 REEC_A REEC_B REEC_C approved approved approved approved approved approved Under excitation limiter Under excitation limiter Wind Type 2 generic excitation/controller model Renewable energy electrical control model for Wind type 3/4 Renewable energy electrical control model for Photovoltaic Renewable energy electrical control model for Energy Storage Devices UEL1 UEL2 4/27/12 4/27/12 8/28/09 3/19/14 3/19/14 3/18/15 In PSLF 17 and PSSE32 Where different variants of the same model exist, the preferred version for submittal to WECC are highlighted in green These models currently are not converted from PSLF to PSS/E. 29 PGE Company Training Power System Stabilizer Model – PSS2A • Used to model the Power System Stabilizer (PSS) with dual inputs that is used to provide a supplementary signal via the exciter to provide damping for power swings Variable 30 Default Data Variable Description PGE Company Training WECC PSS Models PSS MODELS GE PSLF PTI PSS/E* PowerWorld Simulator wsccst pss2a ieeest psssb pss1a pss2b pss3b pss4b psssh ST2CUT PSS2A IEEEST PSS2A IEEEST PSS2B PSS3B PSS4B WSCCST and ST2CUT PSS2A IEEEST PSSSB PSS1A PSS2B PSS3B PSS4B PSSSH IEEE Standard Status approved PSS2A, PSS3Bapproved PSS1A approved PSS2A, PSS3Bapproved PSS1A PSS2B approved PSS3B approved PSS4B approved Comments 8/11/06 8/11/06 8/11/06 8/11/06 8/11/06 8/11/06 8/11/06 Dual input PSS - Old WSCC model Dual input PSS (delta P-omega) Single input PSS, dual lead lag pss2a + transient stabilizer Generic single input PSS - not used in WECC Dual input PSS - Extra lead/lag (or rate) block added at end (up to 4 lead/lags total) Thyripol, Unitrol ABB multi-band Siemens H infinity PSS PTI/GE/PowerWorld Comments In both programs In both programs In both programs In PSLF 17 Where different variants of the same model exist, the preferred version for submittal to WECC are highlighted in green These models currently are not converted from PSLF to PSS/E. 31 PGE Company Training Break 32 PGE Company Training Governor Models Purpose of Governor • Function of the Governor − Primary Control is to provide relationship between power and speed of rotating machinery Rotating speed of generator-turbine is measured, and compared to a set point Difference between measurement and set point turned into a mechanical opening or closing of valves or gates Change of flow of steam, water, gas into the turbine increases or decreases rotational energy Turbine and generator increase or decrease rotational speed − Secondary Controls Can Exist – Plant Level – Unit load control – Grid Level – Load frequency control 33 PGE Company Training GGOV1 Governor Model • Represents a Proportional Integral/Derivative (PID) controlled governor (gas turbines, diesel engines, steam turbines, and simple hydro turbines) Variable 34 Default Data Variable Description PGE Company Training IEEEG1 Governor • IEEE Large Steam Turbine/Governor model Variable 35 Default Data Variable Description PGE Company Training HYGOV Governor • Represents a hydro turbine and governor with straight forward penstock configurations and electro-hydraulic governors Variable Simple Francis Kaplan 36 Variable Description PGE Company Training WECC Used Governor Models TURBINE/GOVERNOR MODELS GE PSLF gast ggov1 h6b hyg3 hygov hygov4 hygovr ieeeg1 ieeeg3 lcfb1 pidgov tgov1 PTI PSS/E* URGS3T GGOV1 WSHYGP HYGOV IEEEG3 hygovr WSIEG1 IEEEG3 LCFB1 PIDGOV TGOV1 ggov2 ggov3 wt1t wt1p_b wt2t wtgt_a wtga_a wtgp_a wtgq_a PowerWorld Simulator GAST_GE and URGS3T GGOV1 H6B HYG3 HYGOV HYGOV4 HYGOVR IEEEG1 and WSIEG1 IEEEG3 LCFB1 and LCFB1_PTI PIDGOV TGOV1 IEEE Standard Status approved approved approved approved approved approved approved approved approved approved approved approved 8/11/06 8/11/06 8/9/13 8/11/06 8/11/06 8/11/06 2008 8/11/06 8/11/06 8/11/06 8/11/06 8/11/06 GGOV2 WT12T1 WT12A1 WT12T1 WTDTAU1 WTARAU1 WTPTAU1 WTTQAU1 GGOV3 WT1T and WT12T1 WT1P_B WT2T WTGT_A WTGA_A WTGPT_A WTGTRQ_A Comments Added in 2008 Use hygov 4 for new models Use hyg3 for new models new in GE PSLF approved approved approved approved approved approved approved approved 2010 1/21/11 3/19/14 8/28/09 3/19/14 3/19/14 3/19/14 3/19/14 PTI/GE/PowerWorld Comments We have the new GGOV2 model in a user written format. We will see if this can be given to users as a user model in the next point release. We hope to make it a standard model for the next major release. new in GE PSLF Wind Type 1 generic turbine model Wind Type 1 & Type 2 Pitch controller model/Pseudo Gov aerodynamics Wind Type 2 generic turbine model Drive train model for Wind type 3/4 Aerodynamic model for Wind type 3 Pitch control model for Wind type 3 Torque control model for Wind type 3 Where different variants of the same model exist, the preferred version for submittal to WECC are highlighted in green These models currently are not converted from PSLF to PSS/E. 37 PGE Company Training Load Models • Load falls into two categories in transient stability − Static Load Model A function of voltage and/or frequency Discharge Lighting (Fluorescent Lights) Voltage Dependent − Dynamic Load Models Induction Motors • Load Characteristic Models end up being combinations of all of the above. “Complex” load models include all of them in various proportions • Latest “Composite Load Model” includes modeling of the equivalent distribution system, different types of static and dynamic loads, and air conditioning load 38 PGE Company Training WECC Composite Load Model - CMPLDW CMPLDW Model after initialization 39 PGE Company Training CMPLDW – Composite Load Model Variable Default Data Variable Description Note: Input parameters though Mtypd must be included. If At any point after that, the remaining parameters are omitted, Default values shown will be used. 40 PGE Company Training Composite Load Model Diagram 41 PGE Company Training WECC Load Models LOAD MODELS IEEE Standard GE PSLF PTI PSS/E* PowerWorld Simulator Status Comments alwscc blwscc IEELAR IEELBL WSCC assigned to an area WSCC assigned to a bus or load approved 8/11/06 approved 8/11/06 Area load model Bus load model cmpldw CMLDBLU1 CMPLDW and CMPLDWNF (with a separate Distribution Equivalent Model) approved 1/25/13 Composite Load Model ld1pac ACMTBLU1 LD1PAC approved 8/11/06 motor1 CIMTR4 MOTOR1 approved 8/11/06 motorw CIMWBL MOTORW approved 8/11/06 Single-phase AC model (performance based model) Induction machine, represented in load flow as generator. Use to represent motor startup. Should use generic wind model for wind machine Induction Motor Model PTI/GE/PowerWorld Comments Where different variants of the same model exist, the preferred version for submittal to WECC are highlighted in green These models currently are not converted from PSLF to PSS/E. 42 PGE Company Training Wind Plant Modeling • Wind Plant Modeling requires special handling • Because of the small unit size (usually 3 MW or less), and large number of units, equivalencing is necessary • Example: Tucannon Wind farm has 116, 2.3 MW Wind generators 43 PGE Company Training Power Flow Representation Wind Plants • Equivalent needs to model an equivalent generator and associated power factor correction capacitors to model total generating capability and reactive compensation • The equivalent generator step-up transformer (pad-mounted) represents the sum of all WTG step-up transformers • The equivalent collector system represents the aggregate branch effects of the WTG collector system. Should approximate both real power losses and voltage drop to the “average” WTG in the wind plant. 44 PGE Company Training Equivalent Collector System Determination (Example) 45 PGE Company Training Different Types of Wind Generators 46 PGE Company Training Wind Stability Models 47 PGE Company Training Renewable Energy Models Wind, Photovoltaic, Battery Energy Storage 48 PGE Company Training Renewable Energy WECC Modeling Recommendations Where different variants of the same model exist, the preferred version for submittal to WECC are highlighted in green These models currently are not converted from PSLF to PSS/E. EXCITATION SYSTEM MODELS (Volt/Var Control Models) IEEE Standard GE PSLF PTI PSS/E* PowerWorld Simulator Status Comments Modifications/Actions Needed PTI/GE/PowerWorld Comments wt2e WT2E1 WT2E and WT2E1 approved 8/28/09 Wind Type 2 generic excitation/controller model wt3e WT3E1 WT3E and WT3E1 approved 8/28/09 Wind Type 3 generic excitation/controller model (GE Technology) wt4e WT4E1 WT4E and WT4E1 approved 8/28/09 Wind Type 4 generic excitation/controller model In PSLF 17 and PSSE32 This model will be phased out by June 2017 and should be replaced with In PSLF 17 and PSSE32 reec_a. This model will be phased out by June 2017 and should be replaced with In PSLF 17 and PSSE32 reec_a. reec_a reec_b reec_c REECAU1 REECBU1 REECCU1 REEC_A REEC_B REEC_C approved 3/19/14 approved 3/19/14 approved 3/18/15 Renewable energy electrical control model for Wind type 3/4 Renewable energy electrical control model for Photovoltaic Renewable energy electrical control model for Energy Storage Devices GENERATOR MODELS GE PSLF IEEE Standard PTI PSS/E* PowerWorld Simulator pvd1 regc_a wt1g wt2g REGCAU1 WT1G1 WT2G1 PVD1 REGC_A WT1G and WT1G1 WT2G and WT2G1 approved approved approved approved Status Comments wt3g WT3G1 WT3G and WT3G1 approved 8/28/09 Wind Type 3 generic generator model (GE Technology) wt4g WT4G1 WT4G and WT4G1 approved 8/28/09 Wind Type 4 generic generator model 3/19/14 3/19/14 1/21/11 8/28/09 Distributed Photovoltaic system model Generator/converter model for Photovoltaic, Wind type 3/4 Wind Type 1 generic generator model Wind Type 2 generic generator model Modifications/Actions Needed PTI/GE/PowerWorld Comments In PSLF 17 and PSSE32 This model will be phased out by June 2017 and should be replaced with In PSLF 17 and PSSE32 regc_a. This model will be phased out by June 2017 and should be replaced with In PSLF 17 and PSSE32 regc_a. TURBINE/GOVERNOR MODELS PTI PSS/E* PowerWorld Simulator Status Comments wt1t WT12T1 WT1T and WT12T1 approved 1/21/11 Wind Type 1 generic turbine model wt1p WT12A1 WT1P and WT12A1 approved 1/21/11 Wind Type 1 generic Pitch controller model/Pseudo Gov:aerodynamics wt1p_b wt2t WT12A1 WT12T1 WT1P_B WT2T approved 3/19/14 approved 8/28/09 Wind Type 1 & Type 2 Pitch controller model/Pseudo Gov aerodynamics Wind Type 2 generic turbine model wt2p WT12A1 WT2P approved 8/28/09 Wind Type 2 generic Pitch controller model/Pseudo Gov:aerodynamics wt3t WT3T1 WT3T and WT3T1 approved 8/28/09 Wind Type 3 generic turbine model (GE Technology) WT3P1 WT3P and WT3P1 approved 8/28/09 Wind Type 3 generic Pitch controller model WT4T approved 8/28/09 Wind Type 4 generic turbine model WTGT_A WTGA_A WTGPT_A WTGTRQ_A approved approved approved approved Drive train model for Wind type 3/4 Aerodynamic model for Wind type 3 Pitch control model for Wind type 3 Torque control model for Wind type 3 wt3p wt4t wtgt_a wtga_a wtgp_a wtgq_a 49 IEEE Standard GE PSLF transient features are inside the WT4E1 model WTDTAU1 WTARAU1 WTPTAU1 WTTQAU1 3/19/14 3/19/14 3/19/14 3/19/14 Modifications/Actions Needed PTI/GE/PowerWorld Comments This model will be phased out by June 2017 and should be replaced with wt1p_b model. This model will be phased out by June 2017 and should be replaced with wt1p_b model. This model will be phased out by June 2017 and should be replaced with wtg*_a models. This model will be phased out by June 2017 and should be replaced with wtg*_a models. This model will be phased out by June 2017 and should be replaced with wtgt_a. PGE Company Training Other WECC Models & Relays Used • SVC/Shunt Switching Models, DC Line Models, Relay Models OTHER MODELS epcdc gp1 lhfrt lhvrt locti lsdt1 lsdt2 lsdt9 IEEE PowerWorld Simulator Standard CCOMP and COMPCC CCOMP4 Not Used ATRRELAY For 3-terminal version of PDCI: MTDC_PDCI, CONV_CELILO_E, CONV_CELILO_N, PDCNSU, PDCSNU CONV_SYLMAR; For IPP model: MTDC_IPP, CONV_IntMtnPP, CONV_Adelanto CDC6 EPCDC and CDC6 not converted (4) GP1 FRQTPA LHFRT VTGTPA LHVRT TIOCR1 LOCTI and TIOCR1 LDS3BL LSDT1 and (LDS3 assigned to a load) LVS3BL LSDT2 and (LVS3 assigned to a load) LDS3BL LSDT9 and (LDS3 assigned to a load) ooslen not converted (11) OOSLEN approved 8/11/06 3 zone out of step relay repc_a scmov REPCAU1 REPC_A SCMOV approved 3/19/14 Power Plant Controller for Photovoltaic, Wind type 3/4 Series capacitor MOV and bypass model stcon not converted (2) STCON svcwsc svsmo1 svsmo2 svsmo3 msc1 msr1 mslr1 tiocrs CSVGN5, CSVGN6 SVSMO1U2 SVSMO2U2 SVSMO3U2 SWSHNT SVCWSC, CVSGN5 and CVSGN6 SVSMO1 SVSMO2 SVSMO3 MSC1 and SWSHNT tlin1 vwscc GE PSLF ccomp ccomp4 Not Used dcmt PTI PSS/E* Status Comments COMPCC approved 11/20/14 approved 3/17/2015 approved 3/17/2015 Cross & Joint current compensation model Colstrip Acceleration Trend Relay (ATR) approved 8/11/06 new PDCI DC model approved 8/11/06 Intermountain DC model Generator Protection relay Low/High frequency ride-through generator protection Low/High voltage ride-through generator protection Branch overcurrent relay with inverse time characteristic Underfrequency relay Undervoltage relay Underfrequency relay approved approved approved approved approved approved 8/9/13 8/9/13 8/9/13 8/11/06 8/11/06 8/11/06 Static synchronous condenser approved 1/21/11 approved 8/26/11 approved 8/26/11 approved 1/21/11 approved 3/17/2015 PTI/GE/PowerWorld Comments We have just developed two new models (north to south and south to north) for the PDCI. GE needs details for data conversion to PSLF. All of these models originated as user-written models in GE using EPCL. Note: the PDCI model will be going away as the CELILO converters are being replaced. Full documentation describing the IPP model can be found at http://www.powerworld.com/files/clientconf2014/06DC%20Line%20Model%20of%20IPP.pdf We don't have a PSS/E model for this, need details We don't convert this. The reason is not because we don't have a model. PSS/E has a double circle or lens out-of step line relay model called 'CIROS1' (please note that like any other relay model, this also is a generic line-relay model not representing any particular manufacturer). The reason that the data is not converted is probably because the data requirements of the PSLF 'ooslen' model do not match the data requirements of the PSS/E 'CIROS1' model. However, this does not prevent the PSS/E users to create a DYR data record and include the CIROS1 model for every occurrence of the PSLF 'ooslen' model. In PSLF We don't convert this. This model, per our notes from the previous M&V meetings, was not to be used in WECC. This also is a generic model not representing any particular manufacturer. PSS/E also has two generic static condenser models - the CSTATT (use of this requires a generator model in load flow), and the CSTCNT (use of this requires a FACTS device model in load flow). We can not convert the PSLF STCON to PSS/E CSTATT or the CSTCNT models because the data requirements are different. Static Var Source model, replace with appropriate generic model Generic Static Var Source model (continuous control) Generic Static Var Source model (discrete control) Generic STATCOM model (continuous control) Mechanically Switched Shunt model, links to svsmo models TIOCRS approved 8/9/13 Model Spec only was approved 3/17/15. Over-current relay not converted (114) TLIN1 approved 8/11/06 under frequency or under voltage line relay CSVGN5 VWSCC approved 8/11/06 Static Var Source model We don't convert this, because PSS/E does not have the under frequency or under voltage line relay model. Our consulting group has a user written model and we can include it in PSS/E. We will add this in our list of task to do. As an interim solution we can check if we can make this available as a user written model before it becomes a PSS/E standard model. However, given the fact that this also is a generic model, the data requirements of the PSLF 'tlin1' may not match the data requirements of the PSS/E model, and hence we may not be able to convert from the PSLF to the corresponding PSS/E model. Nonetheless, a model can be made available for WECC PSS/E users. Where different variants of the same model exist, the preferred version for submittal to WECC are highlighted in green These models currently are not converted from PSLF to PSS/E. 50 PGE Company Training Demonstration of Model Importance • Consider the following system model to analyze a Generator connected at Bus 4 to an Infinite bus modeled at Bus 2 (All impedances on 100 MVA Base): 51 PGE Company Training Demonstration of Modeling Importance In Regards to Stability - GENCLS • Take the classical model of generator (GENCLS) only and see how it responds due to a fault (Example: 3-ph fault @ Bus 3, clear in 5 cycles, Open Line Bus1-Bus3-Bus2): Bus 1 Bus 2 53.2 MW -4.1 Mvar Bus 4 Note: 1) GENCLS is oscillatory 2) Damping only effects by Inertia “H” Infinite Bus 1.000 slack 125.0 MW 18.35 Deg 27.8 Mvar Bus 3 71.5 MW 6.1 Mvar 1.050 pu 11.71 Deg 1.030 pu Model GENCLS - 100 MVA Base H = 3.0 D = 0.0 Ra = 0.0 Xdp = 0.2 Rcomp = 0.0 Xcomp = 0.0 -123.8 MW 12.6 Mvar 70.5 MW -8.4 Mvar 7.84 Deg 1.020 pu 0.00 Deg 1.020 pu Gen_Rotor Angle, No Shift Bus_Frequency Bus_Volt (pu) 90 1.2 85 60.5 80 1.15 75 60.4 70 65 60.3 1.1 60 60.2 55 1.05 50 60.1 45 60 1 40 35 59.9 30 0.95 59.8 25 20 59.7 0.9 15 10 59.6 0.85 5 59.5 0 -5 0.8 0 0.5 1 1.5 2 2.5 3 3.5 b c d e f g 52 4 4.5 5 5.5 Volt (pu), Bus Bus 2 g b c d e f 6 6.5 7 Volt (pu), Bus Bus 4 7.5 8 8.5 9 9.5 10 0 0.5 1 1.5 2 2.5 3 3.5 b c d e f g 4 4.5 5 5.5 Frequency, Bus Bus 2 g b c d e f 6 6.5 7 7.5 Frequency, Bus Bus 4 8 8.5 9 9.5 10 0 0.5 1 1.5 2 2.5 b c d e f g 3 3.5 4 4.5 5 5.5 Rotor Angle, No Shift, Gen Bus 2 #1 g b c d e f 6 6.5 7 7.5 8 Rotor Angle, No Shift, Gen Bus 4 #1 PGE Company Training 8.5 9 9.5 10 Demonstration of Modeling Importance In Regards to Stability - GENTPJ • Use a full generator Model – GENTPJ and see how it responds due to a fault (Example: 3-ph fault @ Bus 3, clear in 5 cycles, Open Line Bus1-Bus3-Bus2): Note: 1) Adding detail does not always make things better 2) Running to only 7 sec fails to show instability Model GENTPJ - 100 MVA Base H = 3.0 D = 0.0 Ra = 0.0 Xd = 1.1 Xdp = 0.2 Xqp = 0.5 Xdpp = 0.18 Xqpp = 0.18 Xl = 0.15 Td0p = 7.0 Td0pp = 0.035 Tq0pp = 0.50 S(1.0) = 0.0 S(1.2) = 0.0 Rcomp = 0.0 Xcomp = 0.0 Accel = 0.4 Kis = 0.0 Bus_Frequency Bus_Volt (pu) Xq = 0.5 Tq0p = 0.75 Gen_Rotor Angle, No Shift 105 1.2 1.15 60.55 100 60.5 95 60.45 90 60.4 1.1 60.35 1.05 60.25 85 80 60.3 75 70 60.2 1 0.95 0.9 60.15 65 60.1 60 60.05 55 60 50 59.95 45 59.9 40 59.85 35 59.8 0.85 59.75 0.8 59.65 30 25 59.7 20 15 59.6 0.75 59.55 10 59.5 5 0 59.45 0.7 -5 0 1 2 3 4 5 6 7 b c d e f g 53 8 9 10 11 Volt (pu), Bus Bus 2 g b c d e f 12 13 14 Volt (pu), Bus Bus 4 15 16 17 18 19 20 0 1 2 3 4 5 6 7 b c d e f g 8 9 10 11 Frequency, Bus Bus 2 g b c d e f 12 13 14 15 Frequency, Bus Bus 4 16 17 18 19 20 0 1 2 3 4 b c d e f g 5 6 7 8 9 10 Rotor Angle, No Shift, Gen Bus 2 #1 g b c d e f 11 12 13 14 15 16 Rotor Angle, No Shift, Gen Bus 4 #1 PGE Company Training 17 18 19 20 Demonstration of Modeling Importance In Regards to Stability – Add EXST1 • Use full generator & Exciter Models – GENTPJ & EXST1 and see how it responds due to a fault (Example: 3-ph fault @ Bus 3, clear in 5 cycles, Open Line Bus1-Bus3-Bus2): Note: 1) Adding an exciter was able to provide stabilization 2) Note the “overshoot” in voltage due to the response of the exciter Model EXST1 - 100 MVA Base Tr = 0.0 Vmax = 0.1 Tc = 1.0 Tb = 10.0 Ta = 0.02 Vrmax = 5.0 Kc = 0.05 kf = 0.0 Tf = 1.0 Tc1 = 1.0 VaMax = 5.0 VaMin = -5.0 Xe = 0.04 Ilr = 2.8 Bus_Volt (pu) Bus_Frequency Tb1 = 1.0 Klr = 5.0 Gen_Rotor Angle, No Shift 110 1.2 1.15 1.1 60.55 105 60.5 100 60.45 95 60.4 90 60.35 85 60.3 80 60.25 1.05 75 60.2 70 60.15 1 65 60.1 60 60.05 0.95 55 60 50 59.95 45 59.9 0.9 40 59.85 35 59.8 0.85 30 59.75 25 59.7 0.8 0.75 59.65 20 59.6 15 59.55 10 5 59.5 0 59.45 0.7 -5 0 1 2 3 4 5 6 7 b c d e f g 54 Vmin = -0.1 Ka = 200.0 Vrmin = -5.0 8 9 10 11 Volt (pu), Bus Bus 2 g b c d e f 12 13 14 Volt (pu), Bus Bus 4 15 16 17 18 19 20 0 1 2 3 4 5 6 7 b c d e f g 8 9 10 11 Frequency, Bus Bus 2 g b c d e f 12 13 14 15 Frequency, Bus Bus 4 16 17 18 19 20 0 1 2 3 4 b c d e f g 5 6 7 8 9 10 Rotor Angle, No Shift, Gen Bus 2 #1 g b c d e f 11 12 13 14 15 16 Rotor Angle, No Shift, Gen Bus 4 #1 PGE Company Training 17 18 19 20 Demonstration of Modeling Importance In Regards to Stability – Add PSS • Use full Generator, Exciter, & PSS Models – GENTPJ, EXST1, PSS2A and see how it responds due to a fault (Example: 3-ph fault @ Bus 3, clear in 5 cycles, Open Note: Line Bus1-Bus3-Bus2): 1) Adding a PSS helps 2) Only a slight improvement due to the PSS not necessarily tuned for optimization Model PSS2A - 100 MVA Base Ics1 = 1 Ics2 = 3 M=5 N=1 Tw1 = 10.0 Tw2 = 10.0 T6 = 0.02 Tw3 = 10.0 Tw4 = 0.0 T7 = 10.0 Ks2 = 1.47 Ks3 = 1.0 T8 = 0.50 T9 = 0.1 Ks1 = 4.0 T1 = 0.16 T2 = 0.02 T3 = 0.16 T4 = 0.02 Vstmax = 0.1 Vstmin = -0.1 A = 1.0 Ta = 0.0 Tb = 0.0 Ks4 = 1.0 Bus_Volt (pu) Bus_Frequency Gen_Rotor Angle, No Shift 1.2 105 1.15 60.55 100 60.5 95 60.45 90 60.4 1.1 85 60.35 80 60.3 75 60.25 1.05 70 60.2 1 0.95 0.9 60.15 65 60.1 60 60.05 55 60 50 59.95 45 59.9 40 59.85 35 59.8 0.85 30 59.75 25 59.7 20 59.65 0.8 15 59.6 10 59.55 0.75 5 59.5 0 59.45 0.7 -5 0 1 2 3 4 5 6 7 b c d e f g 55 8 9 10 11 Volt (pu), Bus Bus 2 g b c d e f 12 13 14 Volt (pu), Bus Bus 4 15 16 17 18 19 20 0 1 2 3 4 5 6 7 b c d e f g 8 9 10 11 Frequency, Bus Bus 2 g b c d e f 12 13 14 15 Frequency, Bus Bus 4 16 17 18 19 20 0 1 2 3 4 b c d e f g 5 6 7 8 9 10 Rotor Angle, No Shif t, Gen Bus 2 #1 g b c d e f 11 12 13 14 15 16 Rotor Angle, No Shif t, Gen Bus 4 #1 PGE Company Training 17 18 19 20 Transient Stability – System Behavior • Power Relationship decreases with system changes • System starts at pt a, with a fault, system moves to pt b, when fault clears system has moved to pt c, when equipment opens moves to pt d, then system moves to maximum of pt e. • To have a stable system, Area from starting condition to fault clearing (Area A1), must be greater than or equal to the area after the fault is cleared (Area A2). 56 PGE Company Training Transient Stability – Important Factors Factors that Influence Transient Stability • System Design – – – – – • System Operation – – – • Generator loading Voltage levels, power factor Transmission system loading Nature of Disturbance – – – – – 57 Generator/Turbine design (Inertia, generator transient reactance, etc) External System Equivalent Reactance (Number of lines, intermediate switching stations, series capacitors, etc) Intermediate voltage support (SVC, Capacitors, etc) Control system response (excitation, governor, SVS, etc) Discrete Switching of equipment (reclosing, generation dropping, load shedding, capacitor switching, series capacitor insertion/bypassing, single-pole clearing, etc) Location, type of fault, and clearing time Lines lost due to the fault Relay/Breaker failure Failure of Remedial Action (RAS) “Sympathic” Relaying actions PGE Company Training Transient Stability – Modeling of Faults Nature of Fault Is Important TE’s During Fault for Different Types of Faults How to Model Fault Impedance Of the Fault Shunt For Different Types of Faults (Requires knowledge of Negative and Zero Sequence Data From Fault Study) 58 PGE Company Training Transient Stability Studies What Can Be Studied With Transient Stability? • Critical Clearing Times for Units • Breaker Closing Angles and Breaker Failure Timing • Assess Stability of Contingencies Due to Faults and Subsequent Loss of Equipment (Unit trips, line trips, capacitor trips, load loss, etc) • Assess Stability of Stressed Line and “Path” Flows • Assess Stability Limit Increases for Remedial Actions (Unit tripping, Under-frequency Load Shedding, Under-voltage Load Shedding, Series Capacitor Bypassing/Insertion, Capacitor Switching, etc). • Outage Limitations (Unit and Line Outages) 59 PGE Company Training Transient Stability Studies Stability Studies – Beginning Data Verification • Review Dynamic data for possible errors − Oscillatory behavior of models − Missing models Net generation with load if no dynamic data is available − Bad Model Data Negative time constants in models Model data suspicious (data outside of normal ranges, or data item relationships unreasonable) Initializing model states out of limits − Always fix ‘bad’ data Look to readjust generator power flow conditions (gen MW, Var limits, Vhold) if generators are initiating “out of limits” 60 PGE Company Training Transient Stability Studies Performing Transient Stability Studies - Runningming Transient Stability Studies – Running • Initialize the power flow with the dynamics data, and get an initial state condition. Should be stable!!!! • Look at output of initial conditions to notice any units that initiate out of limits or have large “State” value changes. This is indicative of problems with the data. These possible data errors should be fixed until a good initial state is observed. − Run “Data Validation” to see what Powerworld would change • Make a no disturbance run to at least 20 seconds to verify the “outputs” of machines, bus voltages, and frequency are not changing. This helps to verify where there may be additional problems with the data. • Should always “Fix” problems before doing any further analysis!!! 61 PGE Company Training Transient Stabilty Simulation Study Performing Transient Stability Studies – Running Disturbances • Should always run for 1 second to verify the system is initially stable • Apply the fault for the time duration (in cycles) that the type of fault requires (relay time and breaker opening time, breaker failure timing, etc). − Use an appropriate fault impedance if running a non 3-phase fault (Obtain Impedance from fault study) • Clear the fault • Open Line(s), Transformer(s), Generator(s), or Load(s) that would trip due to the fault • Apply any Remedial Action Schemes (RAS) at the appropriate timing − − − − Tripping of Generators Capacitor Switching Bypassing/Inserting Series Capacitors Tripping other Lines or Transformers • Run to at least 20 seconds to verify the system is stable after clearing the fault, tripping of equipment, and any RAS actions 62 PGE Company Training Transient Stability Studies – Output Plots Performing Transient Stability Studies – Plotting of Output Data • Should have Rotor Angles of units across the system • Should have Machine values for generation of interest (Pmech, Pelect, Efd, Vterm, Speed, etc) • Bus Voltages and Angles of substations across the system • Bus Frequency of substations across the system • Anything else of interest (Line Flows - P&Q, Apparent Impedance - R&X, etc) • If you think you want to see it, plot it!!! 63 PGE Company Training Transient Stability Studies Output Plot Observations What to Observe • Verify that all units remain synchronized by looking at rotor angles • Look for excessive, low frequency swing bus voltages. Long low voltages may cause motors to stall or undervoltage load shedding causing loss of load. • Undamped oscillations indicate the case may be above a power transfer limit. • Low frequency that may cause loss of load due to underfrequency load shedding (if it is not planned for the outage). • High frequency oscillations in voltage. This may indicate that a unit is unstable somewhere in the system. Look for bus of highest voltage swings for the problem area. 64 PGE Company Training Break 65 PGE Company Training Voltage Stability Definition of Voltage Stability Voltage Stability is defined as the ability of the power system to maintain acceptable voltages at all buses in the system under normal conditions and after experiencing a disturbance. 66 PGE Company Training Voltage Stability Two Major Types of Voltage Instability 67 • Short-term – Onset of voltage collapse due to a disturbance prior to transformer LTC action (timeframe: 0 to 30 seconds). Causes: – Motor dynamics/stalling – Generator over-excitation limits • Long-term – Slow deterioration of voltage ultimately resulting in voltage collapse after an outage (time frame 1 to 60 minutes) – Transformer & Regulator tap changes – Thermostatically controlled loads – Generator current limits PGE Company Training Power-Voltage Characteristics Note: Voltage Declines as Power Transfer (or Load) Increases 68 PGE Company Training Characteristics of Voltage Instability • Is usually associated with a heavily “stressed” power system − Driving force for voltage instability is usually the load • Main factor is the inability of the power system to maintain a proper balance of reactive power and voltage control • Possible outcomes of voltage instability: − Loss of loads − Tripping of lines and other elements leading to cascading outages − Possible loss of synchronism of some generators from line outages or from operating conditions that violate generator field current limits 69 PGE Company Training Voltage and Reactive Power Planning • Most Voltage Stability Problems can be studied with a Power flow program − Exception: Fast collapse with significant amount of induction motor/air conditioning load − Capacitor relay coordination/hunting study where time/step sequence response is needed • Two Power Flow Study Methods − QV Curve Analysis − PV Curve Analysis 70 PGE Company Training Q-V Curves • Q-V curves are useful planning tool to determine reactive requirements for normal and contingency conditions. • Development of a Q-V Curve − Place an unlimited SVC at a selected bus − Allow the SVC to hold a specified voltage at the bus − Set all generators without high-side voltage control to hold its’ own terminal voltage − Have transformer LTC’s and non-automatic switched capacitors fixed − Allow “Vhold” to vary (typically 105% to 70% voltage in 1% steps) − Note the Q output of the SVC to hold the set “Vhold” bus voltage − (Should be a 0 Mvar output at initial bus voltage) − As voltage increases, the SVC is a reactive “source”, as voltage decreases the SVC is a reactive “sink” 71 PGE Company Training Q-V Curve Example 72 PGE Company Training Q-V Curve Characteristics Effects on Local System • SVC as a source − Voltage increase causes local line charging and Mvar output of local capacitors to increase − Lower I2X losses and lower line current − Backs off local generation reactive power output • SVC as a sink − Voltage decrease causes reactive power to be drawn from system − Initial reactive power drawn from local generation increases unless at limits − Voltage decrease causes local line charging and Mvar output of local capacitors to decrease − Higher I2X losses and higher line current − At some point, the system uses more reactive than coming from ties lines and the SVC injects reactive to lower voltage 73 PGE Company Training Q-V Curves – Information Determined • Reactive MVar Margin at a bus • Critical Voltage where voltage collapse occurs • Potential Operating Parameters for capacitors at the bus • Slope of the curve gives indication on dV/dQ, and where reactive could be added • Can be used to determine capacitor sizing 74 PGE Company Training Q-V Analysis - Disadvantages • Must probe to find the bus that has the least amount of reactive margin (“critical bus”) • Does not give a good indication of how far from the voltage collapse point the power system is − Can be accomplished with multiple QV curves and increasing load or transfers until there is no reactive margin left. • Motors may stall below 90% voltage, so any reactive margin below 0.85 pu voltage may not be useful. 75 PGE Company Training P-V Curves • P-V curves are also a useful planning tool to determine reactive requirements for normal and contingency conditions. • Development of a P-V Curve − Note the voltage and load (or power transfer across the system) for a desired bus − Constant power load models used − Gradually increase the load (or power transfer across the system) recording the load (or transfer flow) for each level until the power flow will not solve − Plot the results with voltage as the Y axis and load (or power transfer) on the X axis 76 PGE Company Training P-V Curve 77 PGE Company Training P-V Analysis • Advantages − Provides a visual indication of where the MW Voltage collapse is − Provides a indication of severity of voltage collapse problem Flatness of PV curve indicates over-use of shunt caps Voltage of the collapse point − Easily determine power margin requirements − Provides relationship between generator power reserve and the load that can be served • Disadvantages − Failure to solve may or may not be the “true” collapse point − Control settings in the power flow are more complex Need control settings for each MW base case increment and maybe different control settings for each outage case − Selection of power transfer must be well thought out 78 PGE Company Training Solutions to Improve Voltage Stability • Install/Operate Shunt Capacitor Banks − Least Cost, but Var’s reduce when needed the most • Add dynamic Shunt Compensation in the form of SVC/STATCOM to mitigate transient voltage dips − Provide dynamic response for sudden changes • Add Series Compensation on transmission lines in the problem area − Increases system stability and transfer capability − Var’s increase when needed the most − However, must consider subsynchronous resonance on generators in the vicinity if used (SSR evaluation needed) • Implement Under-Voltage Load Shedding (UVLS) Program • Construct new transmission/generation facilities 79 PGE Company Training Use of Capacitors - Caution • Adding Capacitors causes instability point to move up in voltage − When point is near normal operating voltages, implies too many capacitors have been added − Voltage becomes a poor indicator of Voltage Instability 80 PGE Company Training Voltage Stability Reactive Margin for Outages • Adding a Power Margin for N-1 Outages helps to insure of a safe operating point for system operation 81 PGE Company Training Reactive Power Requirements for Generators • Generation units should generate reactive power in accordance with a voltage schedule (Example): − Above 230 kV: − 230 kV: − 69 -161 kV: 1.05 pu 1.02 pu 1.00 pu • Units should have a reactive power capability to maintain a power factor between 0.95 lagging and 0.97 leading. • Units must be operated with voltage regulators in auto mode (i.e. on AC excitation control) • A Generator should be required to operate up to maximum reactive capability to meet required voltage schedules. • Ensure that reactive reserves are available for the system. 82 PGE Company Training Voltage Stability Summary • Inadequate reactive power supply has been a major factor in most of the recent worldwide blackouts. • The increasing need to operate the transmission system at or near its maximum safe transfer limit has become a primary concern. • Reactive power supply and VAR management is an important ingredient for maintaining healthy power system voltages and facilitating power transfers. • Q-V, P-V, and Transient Stability Analysis should be used together to give the most “robust picture” for Voltage Stability Planning 83 PGE Company Training Questions? 84 PGE Company Training Test of Knowledge 85 PGE Company Training 52 2.2 Composite Load Model Description 2.2.1 Introduction to Composite Load Model A recorded video (1h47m) given by BPA (Mr. Hamodi Hindi) in the most recent (Oct. 2015) NERC LMTF webinar can be downloaded at: https://www.wecc.biz/Administrative/Introduction%20to%20CMPLDW%20102215.zip This gives a good introduction on history and current development of composite load model Phase 2 project. 2.2.2 Data Tool for Composite Load Model Dr. Pavel Etingov at Pacific Northwest National Lab developed an open source tool with user interface that can automatically generate the composite load model stability models based on climate zone, season and hours. The first version can be downloaded from: https://svn.pnl.gov/LoadTool The second version is under testing. You can request a copy from him by sending email to: Pavel.Etingov@pnnl.gov 53