Beruflich Dokumente
Kultur Dokumente
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
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.
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:
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
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.
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.
//--------------------------------------------------------------------------------
// 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
}
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.
“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
“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.
“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.
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.
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.
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
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 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.
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.
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
5. Select from the list of generators the Colstrip Units and add to the model
25
After adding the Colstrip ATR models, users can find the information in the model explorer, under:
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: Flat Line" 0.000000 "Simulation" "SET TimeStep 1" "ALWAYS" ""
"002: Chief Joseph Brake Insertion" 1.000000 "Load '40232' 'CH'" "CLOSE" "ALWAYS" ""
"002: Chief Joseph Brake Insertion" 1.500000 "Load '40232' 'CH'" "OPEN" "ALWAYS" ""
"003: Double Palo Verde" 1.000000 "Gen '14931' '1'" "OPEN" "ALWAYS" ""
"003: Double Palo Verde" 1.000000 "Gen '14932' '1'" "OPEN" "ALWAYS" ""
}
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
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.
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.”
1.5.2.1 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.
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.
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.
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.
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. 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
Unit
Large Unit Bus Number ID
HERM 1G 45454 1
FREDONA2 42112 2
LIB 01 44191 1
ROCKYR02 46842 C2
MCN 02 44102 2
KFALLCT1 45448 1
LANCAS G 47568 1
COYO G1 43111 1
GEP G1 47687 1
CHEH G1 47588 1
DWOR3 40365 1
PORTW G1 43905 1
RVR RD C 47216 1
HPP G1 47639 1
GRYHB G1 47596 1
CENTR G1 47740 1
CGS 40063 1
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
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.
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. Numerical instability
2. System instability
3. Model errors
4. Power flow case unsolvable or collapse
5. Contingency definition errors
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.
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.
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)
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.
Undamped or sustained oscillation, regardless of magnitude, should always catch the attention of users.
In most of cases, such oscillations would be considered unstable.
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.
Don Johnson
Senior Planning Engineer
T&D Planning
PGE Company
PGE Company
Confidential
Training
Power System Stability – Outline
2
PGE Company Training
Stability
• Definition of Stability:
3
PGE Company Training
Power System Stability Classification
4
PGE Company Training
Power System Stability
5
PGE Company Training
Transient Stability Analysis Time Frame
6
PGE Company Training
Power Flow Program
7
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)
− 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.
8
PGE Company Training
Numerical Solution Methods
− 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
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
10
PGE Company Training
Numerical Method – Euler’s
ẋ = f(x(t)) = as
Then
• 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
where
1 = ∆t ∗ f(x(t))
2 = ∆t ∗ f(x(t) + 1)
12
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
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
16
PGE Company Training
Transient Stability Models
Model Classes
17
PGE Company Training
Transient Stabilty Models
18
PGE Company Training
Important Input/Output Values for Models
19
PGE Company Training
Transient Stability – Generator Model
20
PGE Company Training
Transient Stability Models - Generator
• Note: Normally used only for “academic studies”; as it is only valid for transients
up to about one second, thus should not be used ….
No exciter or governor
Assumes constant mechanical power
Damping is negligible
21
PGE Company Training
Transient Stability Models - Generator
• Represents a Synchronous machine (either round rotor or salient pole) with stator
and rotor dynamics modeled along with saturation
22
PGE Company Training
Information for Generator Modeling
23
PGE Company Training
WECC Recommended Generator Models
GENERATOR MODELS
IEEE
GE PSLF PTI PSS/E* PowerWorld Simulator Status Comments PTI/GE/PowerWorld Comments
Standard
gentpf GENROU/IEEEVC GENTPF approved 8/11/06 WECC Model
genrou GENROU/IEEEVC GENROU approved 8/11/06 Round rotor generator model, use for thermal generator models
gentpj GENTPJU1 GENTPJ approved 1/23/09 modified gentpf with improved saturation modeling Available in PSS/E version 33.2
gencc GENROU/IEEEVC GENCC approved 8/11/06 Cross Compound generator model
pvd1 PVD1 approved 3/19/14 Distributed Photovoltaic system model
regc_a REGCAU1 REGC_A approved 3/19/14 Generator/converter model for Photovoltaic, Wind type 3/4
wt1g WT1G1 WT1G and WT1G1 approved 1/21/11 Wind Type 1 generic generator model
wt2g WT2G1 WT2G and WT2G1 approved 8/28/09 Wind Type 2 generic generator model 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
gencls not used GENCLS Used to force a signal, or classical generator model 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
25
PGE Company Training
Exciter Models
Excitation System Models
Three Basic Types Rotating Exciters
• DC Excitation Systems • Brushless (No Slip Rings)
• AC Excitation Systems • Brush Type
• Static Excitation Systems
Static Exciters (Power Source)
• Shunt – (Generator Output Voltage)
• Series – (Derived from Generator Output
Voltage & Current)
26
PGE Company Training
Exciter Model - SEXS
27
PGE Company Training
Exciter Model – EXST1
28
PGE Company Training
WECC Excitation Models
esac3a ESAC3A ESAC3A AC3A approved 1/21/11 2005 IEEE standard - updated AC3A In both programs
exac4 EXAC4 EXAC4 AC4A approved 8/11/06 Rotating AC with controlled rectifier (Althyrex) (rare) Differs from IEEE AC4A -- no OEL/UEL inputs
esac4a ESAC4A ESAC4A AC4A approved 1/21/11 2005 IEEE standard - updated AC4A In both programs
esac5a ESAC5A ESAC5A AC5A approved 1/21/11 Simplified brushless exciter In both programs
exac6a ESAC6A EXAC6A AC6A Alternator, noncontrolled rectifier, lead-lag Differs from IEEE AC6A -- no OEL/UEL inputs; speed multiplier, not a new model for PSS/E (model already exists)
esac6a ESAC6A ESAC6A AC6A approved 1/21/11 2005 IEEE standard - updated AC6A In both programs
esac7b AC7B ESAC7B and AC7B AC7B approved 1/21/11 2005 IEEE standard - new In both programs
exac8b ESAC8B EXAC8B ESAC8B approved 8/11/06 Brushless exciter with PID voltage regulator Differs from IEEE AC8B -- no exciter upper limit; added input limits and speed multiplier
esac8b AC8B ESAC8B_GE and AC8B AC8B approved 1/21/11 2005 IEEE standard - updated AC8B In both programs
exbbc BBSEX1 EXBBC and BBSEX1 approved 8/11/06 Static with ABB regulator In both programs
exdc1 IEEEX1 EXDC1 and IEEEX1 DC1A approved 8/11/06 Rotating DC Differs from IEEE DC1A -- no UEL inputs; speed multiplier
esdc1a ESDC1A ESDC1A DC1A approved 1/21/11 2005 IEEE standard - updated DC1A In both programs
exdc2 EXDC2 EXDC2_GE and EXDC2_PTI approved 8/11/06 Rotating DC with terminal fed pilot, alternate feedback
exdc2a EXDC2 EXDC2A and EXDC2_PTI DC2A approved 8/11/06 Rotating DC with terminal fed pilot Differs from IEEE DC2A -- no UEL inputs; speed multiplier
esdc2a ESDC2A ESDC2A DC2A approved 1/21/11 2005 IEEE standard - updated DC2A In both programs
exdc4 IEEET4 EXDC4 and IEEET4 DC3A approved 8/11/06 Rotating, noncontinuous - minor differences between models If Kr = 0, should convert to IEEEX4 (IEEE DC3A). Model added in PSS/E -32.
esdc3a DC3A ESDC3A and DC3A DC3A approved 1/21/11 Rotating, noncontinuous In both programs
esdc4b DC4B ESDC4B DC4B approved 1/21/11 Rotating DC with PID In both programs
exeli EXELI EXELI approved 8/11/06 Static PI transformer fed excitation system
exst1 EXST1 EXST1_GE and EXST1_PTI ST1A approved 8/11/06 Static with double lead/lag Differs from IEEE ST1A -- no OEL/UEL inputs; added Xe Ifd loading; RFB before field current limiter.
esst1a ESST1A ESST1A and ESST1A_GE ST1A approved 1/21/11 In both programs
exst2 EXST2 EXST2 approved 8/11/06 SCPT - lead/lag block (Tc, Tb) added
exst2a ESST2A EXST2A ST2A approved 8/11/06 lead/lag block (Tc, Tb) is included to match the WECC FM Differs from IEEE ST2A -- no UEL inputs; added lead/lag.
esst2a ESST2A ST2A approved 1/21/11 2005 IEEE standard - updated ST2A
exst3 EXST3 EXST3 ST3 approved 8/11/06
exst3a ESST3A EXST3A ST3A approved 8/11/06 Use for GE Generex Differs from IEEE ST2A -- no UEL inputs; fewer time constants.
esst3a ESST3A ST3A approved 1/21/11 2005 IEEE standard - updated ST3A
GE EX2000 bus fed potential source, static compound and Generrex-PPS or -CPS, and
exst4b ESST4B EXST4B ST4B approved 8/11/06 Differs from IEEE ST2A -- no OEL/UEL inputs
SILCOmatic 5 excitation systems, with proportional plus integral (PI) voltage controller
esst4b ESST4B ESST4B ST4B approved 1/21/11 2005 IEEE standard - updated ST4B In both programs
esst5b ST5B ESST5B and ST5B ST5B approved 1/21/11 Variation of ST1A (New IEEE Model) In both programs
esst6b ST6B ESST6B and ST6B ST6B approved 1/21/11 Variation of ST4B with field current limit (New IEEE model) In both programs
esst7b ST7B ESST7B and ST7B ST7B approved 1/21/11 Static with limiters (Alstom) (New IEEE model) In both programs
ieeet1 IEEET1 IEEET1 approved 8/11/06 Old type 1
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 REXSYS REXS approved 8/11/06 General Purpose Rotating Excitation System Model
scrx SCRX SCRX approved 8/11/06 intended for use where negative field current may be a problem
sexs SEXS SEXS_GE and SEXS_PTI 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)
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
oel1 Not converted (277) OEL1 approved 4/27/12 Over excitation limiter model, we have to get the block diagram from GE. Presentation at March 2012 M&VWG meeting, use OEL1. Has required functionality.
uel1 UEL1 Not Used UEL1 approved 4/27/12 Under excitation limiter
uel2 UEL2 Not Used UEL2 approved 4/27/12 Under excitation limiter
wt2e WT2E1 WT2E and WT2E1 approved 8/28/09 Wind Type 2 generic excitation/controller model In PSLF 17 and PSSE32
reec_a REECAU1 REEC_A approved 3/19/14 Renewable energy electrical control model for Wind type 3/4
reec_b REECBU1 REEC_B approved 3/19/14 Renewable energy electrical control model for Photovoltaic
reec_c REECCU1 REEC_C approved 3/18/15 Renewable energy electrical control model for Energy Storage Devices
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
30
PGE Company Training
WECC PSS Models
PSS MODELS
IEEE
GE PSLF PTI PSS/E* PowerWorld Simulator Status Comments PTI/GE/PowerWorld Comments
Standard
wsccst ST2CUT WSCCST and ST2CUT approved 8/11/06 Dual input PSS - Old WSCC model
pss2a PSS2A PSS2A PSS2A, PSS3Bapproved 8/11/06 Dual input PSS (delta P-omega)
ieeest IEEEST IEEEST PSS1A approved 8/11/06 Single input PSS, dual lead lag
psssb PSS2A PSSSB PSS2A, PSS3Bapproved 8/11/06 pss2a + transient stabilizer
pss1a IEEEST PSS1A PSS1A Generic single input PSS - not used in WECC
pss2b PSS2B PSS2B PSS2B approved 8/11/06 Dual input PSS - Extra lead/lag (or rate) block added at end (up to 4 lead/lags total) In both programs
pss3b PSS3B PSS3B PSS3B approved 8/11/06 Thyripol, Unitrol In both programs
pss4b PSS4B PSS4B PSS4B approved 8/11/06 ABB multi-band In both programs
psssh PSSSH Siemens H infinity PSS 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
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 Default Data Variable Description
34
PGE Company Training
IEEEG1 Governor
35
PGE Company Training
HYGOV Governor
• Represents a hydro turbine and governor with straight forward penstock
configurations and electro-hydraulic governors
Variable Simple Francis Kaplan Variable Description
36
PGE Company Training
WECC Used Governor Models
TURBINE/GOVERNOR MODELS
IEEE
GE PSLF PTI PSS/E* PowerWorld Simulator Status Comments PTI/GE/PowerWorld Comments
Standard
gast URGS3T GAST_GE and URGS3T approved 8/11/06
ggov1 GGOV1 GGOV1 approved 8/11/06
h6b H6B approved 8/9/13
hyg3 WSHYGP HYG3 approved 8/11/06
hygov HYGOV HYGOV approved 8/11/06
hygov4 IEEEG3 HYGOV4 approved 8/11/06
hygovr hygovr HYGOVR approved 2008 Added in 2008
ieeeg1 WSIEG1 IEEEG1 and WSIEG1 approved 8/11/06
ieeeg3 IEEEG3 IEEEG3 approved 8/11/06 Use hygov 4 for new models
lcfb1 LCFB1 LCFB1 and LCFB1_PTI approved 8/11/06
pidgov PIDGOV PIDGOV approved 8/11/06 Use hyg3 for new models
tgov1 TGOV1 TGOV1 approved 8/11/06
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.
ggov2 GGOV2 new in GE PSLF We hope to make it a standard model for the next major release.
ggov3 GGOV3 approved 2010 new in GE PSLF
wt1t WT12T1 WT1T and WT12T1 approved 1/21/11 Wind Type 1 generic turbine model
wt1p_b WT12A1 WT1P_B approved 3/19/14 Wind Type 1 & Type 2 Pitch controller model/Pseudo Gov aerodynamics
wt2t WT12T1 WT2T approved 8/28/09 Wind Type 2 generic turbine model
wtgt_a WTDTAU1 WTGT_A approved 3/19/14 Drive train model for Wind type 3/4
wtga_a WTARAU1 WTGA_A approved 3/19/14 Aerodynamic model for Wind type 3
wtgp_a WTPTAU1 WTGPT_A approved 3/19/14 Pitch control model for Wind type 3
wtgq_a WTTQAU1 WTGTRQ_A approved 3/19/14 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
38
PGE Company Training
WECC Composite Load Model - CMPLDW
39
PGE Company Training
CMPLDW – Composite Load Model
40
PGE Company Training
Composite Load Model Diagram
41
PGE Company Training
WECC Load Models
LOAD MODELS
IEEE
GE PSLF PTI PSS/E* PowerWorld Simulator Status Comments PTI/GE/PowerWorld Comments
Standard
alwscc IEELAR WSCC assigned to an area approved 8/11/06 Area load model
blwscc IEELBL WSCC assigned to a bus or load approved 8/11/06 Bus load model
ld1pac ACMTBLU1 LD1PAC approved 8/11/06 Single-phase AC model (performance based model)
Induction machine, represented in load flow as generator. Use to represent motor start-
motor1 CIMTR4 MOTOR1 approved 8/11/06
up. Should use generic wind model for wind machine
motorw CIMWBL MOTORW approved 8/11/06 Induction Motor Model
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
• Because of the small unit size (usually 3 MW or less), and large number of units,
equivalencing is necessary
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 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.
GENERATOR MODELS
IEEE
GE PSLF PTI PSS/E* PowerWorld Simulator Status Comments Modifications/Actions Needed PTI/GE/PowerWorld Comments
Standard
pvd1 PVD1 approved 3/19/14 Distributed Photovoltaic system model
regc_a REGCAU1 REGC_A approved 3/19/14 Generator/converter model for Photovoltaic, Wind type 3/4
wt1g WT1G1 WT1G and WT1G1 approved 1/21/11 Wind Type 1 generic generator model
wt2g WT2G1 WT2G and WT2G1 approved 8/28/09 Wind Type 2 generic generator model In PSLF 17 and PSSE32
This model will be phased out by June
wt3g WT3G1 WT3G and WT3G1 approved 8/28/09 Wind Type 3 generic generator model (GE Technology) 2017 and should be replaced with In PSLF 17 and PSSE32
regc_a.
This model will be phased out by June
wt4g WT4G1 WT4G and WT4G1 approved 8/28/09 Wind Type 4 generic generator model 2017 and should be replaced with In PSLF 17 and PSSE32
regc_a.
TURBINE/GOVERNOR MODELS
IEEE
GE PSLF PTI PSS/E* PowerWorld Simulator Status Comments Modifications/Actions Needed PTI/GE/PowerWorld Comments
Standard
wt1t WT12T1 WT1T and WT12T1 approved 1/21/11 Wind Type 1 generic turbine model
This model will be phased out by June
wt1p WT12A1 WT1P and WT12A1 approved 1/21/11 Wind Type 1 generic Pitch controller model/Pseudo Gov:aerodynamics 2017 and should be replaced with
wt1p_b model.
wt1p_b WT12A1 WT1P_B approved 3/19/14 Wind Type 1 & Type 2 Pitch controller model/Pseudo Gov aerodynamics
wt2t WT12T1 WT2T approved 8/28/09 Wind Type 2 generic turbine model
This model will be phased out by June
wt2p WT12A1 WT2P approved 8/28/09 Wind Type 2 generic Pitch controller model/Pseudo Gov:aerodynamics 2017 and should be replaced with
wt1p_b model.
This model will be phased out by June
wt3t WT3T1 WT3T and WT3T1 approved 8/28/09 Wind Type 3 generic turbine model (GE Technology) 2017 and should be replaced with
wtg*_a models.
This model will be phased out by June
wt3p WT3P1 WT3P and WT3P1 approved 8/28/09 Wind Type 3 generic Pitch controller model 2017 and should be replaced with
wtg*_a models.
transient features are This model will be phased out by June
wt4t inside the WT4E1 WT4T approved 8/28/09 Wind Type 4 generic turbine model 2017 and should be replaced with
model wtgt_a.
wtgt_a WTDTAU1 WTGT_A approved 3/19/14 Drive train model for Wind type 3/4
wtga_a WTARAU1 WTGA_A approved 3/19/14 Aerodynamic model for Wind type 3
wtgp_a WTPTAU1 WTGPT_A approved 3/19/14 Pitch control model for Wind type 3
wtgq_a WTTQAU1 WTGTRQ_A approved 3/19/14 Torque control model for Wind type 3
49
PGE Company Training
Other WECC Models & Relays Used
OTHER MODELS
IEEE
GE PSLF PTI PSS/E* PowerWorld Simulator Status Comments PTI/GE/PowerWorld Comments
Standard
ccomp COMPCC CCOMP and COMPCC approved 11/20/14 Cross & Joint current compensation model
ccomp4 CCOMP4 approved 3/17/2015
Not Used Not Used ATRRELAY approved 3/17/2015 Colstrip Acceleration Trend Relay (ATR)
For 3-terminal version of PDCI:
MTDC_PDCI,
CONV_CELILO_E, 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
CONV_CELILO_N, these models originated as user-written models in GE using EPCL. Note: the PDCI model will be going away as the CELILO converters are
dcmt PDCNSU, PDCSNU approved 8/11/06 new PDCI DC model being replaced. Full documentation describing the IPP model can be found at
CONV_SYLMAR; For IPP
http://www.powerworld.com/files/clientconf2014/06DC%20Line%20Model%20of%20IPP.pdf
model: MTDC_IPP,
CONV_IntMtnPP,
CONV_Adelanto
epcdc CDC6 EPCDC and CDC6 approved 8/11/06 Intermountain DC model
gp1 not converted (4) GP1 Generator Protection relay We don't have a PSS/E model for this, need details
lhfrt FRQTPA LHFRT approved 8/9/13 Low/High frequency ride-through generator protection
lhvrt VTGTPA LHVRT approved 8/9/13 Low/High voltage ride-through generator protection
locti TIOCR1 LOCTI and TIOCR1 approved 8/9/13 Branch overcurrent relay with inverse time characteristic
lsdt1 LDS3BL LSDT1 and (LDS3 assigned to a load) approved 8/11/06 Underfrequency relay
lsdt2 LVS3BL LSDT2 and (LVS3 assigned to a load) approved 8/11/06 Undervoltage relay
lsdt9 LDS3BL LSDT9 and (LDS3 assigned to a load) approved 8/11/06 Underfrequency relay
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
ooslen not converted (11) OOSLEN approved 8/11/06 3 zone out of step relay 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.
repc_a REPCAU1 REPC_A approved 3/19/14 Power Plant Controller for Photovoltaic, Wind type 3/4
scmov SCMOV Series capacitor MOV and bypass 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
stcon not converted (2) STCON Static synchronous condenser 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.
svcwsc CSVGN5, CSVGN6 SVCWSC, CVSGN5 and CVSGN6 Static Var Source model, replace with appropriate generic model
svsmo1 SVSMO1U2 SVSMO1 approved 1/21/11 Generic Static Var Source model (continuous control)
svsmo2 SVSMO2U2 SVSMO2 approved 8/26/11 Generic Static Var Source model (discrete control)
svsmo3 SVSMO3U2 SVSMO3 approved 8/26/11 Generic STATCOM model (continuous control)
msc1 SWSHNT MSC1 and SWSHNT approved 1/21/11 Mechanically Switched Shunt model, links to svsmo models
msr1 approved 3/17/2015
mslr1 Model Spec only was approved 3/17/15.
tiocrs TIOCRS approved 8/9/13 Over-current relay
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
tlin1 not converted (114) TLIN1 approved 8/11/06 under frequency or under voltage line relay 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
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 53.2 MW Bus 2
-4.1 Mvar Note:
Bus 4 Infinite Bus 1) GENCLS is oscillatory
1.000 2) Damping only effects by
slack Inertia “H”
60.5 80
1.15
75
60.4
70
1.1 60.3 65
60
60.2
55
1.05
60.1 50
45
1 60
40
35
59.9
0.95 30
59.8 25
20
0.9 59.7
15
59.6 10
0.85 5
59.5
0
-5
0.8
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10
b
c
d
e
f
g Rotor Angle, No Shift, Gen Bus 2 #1 g
b
c
d
e
f Rotor Angle, No Shift, Gen Bus 4 #1
b
c
d
e
f
g Volt (pu), Bus Bus 2 g
b
c
d
e
f Volt (pu), Bus Bus 4 b
c
d
e
f
g Frequency, Bus Bus 2 g
b
c
d
e
f Frequency, Bus Bus 4
52
PGE Company Training
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 Xq = 0.5
Xdp = 0.2 Xqp = 0.5
Xdpp = 0.18 Xqpp = 0.18
Xl = 0.15 Td0p = 7.0 Tq0p = 0.75
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_Volt (pu) Bus_Frequency Gen_Rotor Angle, No Shift
1.2 105
60.55 100
60.5 95
1.15
60.45 90
60.4
85
1.1 60.35
80
60.3
75
1.05 60.25
70
60.2
60.15 65
1 60
60.1
60.05 55
0.95 60 50
59.95 45
59.9 40
0.9
59.85 35
59.8
30
0.85 59.75
25
59.7
20
0.8 59.65
15
59.6
59.55 10
0.75 5
59.5
59.45 0
0.7 -5
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
b
c
d
e
f
g Volt (pu), Bus Bus 2 g
b
c
d
e
f Volt (pu), Bus Bus 4 b
c
d
e
f
g Frequency, Bus Bus 2 g
b
c
d
e
f Frequency, Bus Bus 4 b
c
d
e
f
g Rotor Angle, No Shift, Gen Bus 2 #1 g
b
c
d
e
f Rotor Angle, No Shift, Gen Bus 4 #1
53
PGE Company Training
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 Vmin = -0.1
Tc = 1.0 Tb = 10.0 Ka = 200.0
Ta = 0.02 Vrmax = 5.0 Vrmin = -5.0
Kc = 0.05 kf = 0.0
Tf = 1.0 Tc1 = 1.0 Tb1 = 1.0
VaMax = 5.0 VaMin = -5.0
Xe = 0.04 Ilr = 2.8 Klr = 5.0
59.55 10
0.75
59.5 5
59.45 0
0.7 -5
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
b
c
d
e
f
g Volt (pu), Bus Bus 2 g
b
c
d
e
f Volt (pu), Bus Bus 4
b
c
d
e
f
g Frequency, Bus Bus 2 g
b
c
d
e
f Frequency, Bus Bus 4 b
c
d
e
f
g Rotor Angle, No Shift, Gen Bus 2 #1 g
b
c
d
e
f Rotor Angle, No Shift, Gen Bus 4 #1
54
PGE Company Training
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
Line Bus1-Bus3-Bus2): Note:
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
60.55 100
60.5 95
1.15
60.45 90
60.4
85
1.1 60.35
80
60.3
75
1.05 60.25
70
60.2
60.15 65
1 60
60.1
60.05 55
0.95 60 50
59.95 45
59.9 40
0.9
59.85 35
59.8
30
0.85 59.75
25
59.7
20
0.8 59.65
15
59.6
59.55 10
0.75 5
59.5
59.45 0
0.7 -5
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
b
c
d
e
f
g Volt (pu), Bus Bus 2 g
b
c
d
e
f Volt (pu), Bus Bus 4 b
c
d
e
f
g Frequency, Bus Bus 2 g
b
c
d
e
f Frequency, Bus Bus 4 b
c
d
e
f
g Rotor Angle, No Shif t, Gen Bus 2 #1 g
b
c
d
e
f Rotor Angle, No Shif t, Gen Bus 4 #1
55
PGE Company Training
Transient Stability – System Behavior
56
PGE Company Training
Transient Stability – Important Factors
• System Operation
– Generator loading
– Voltage levels, power factor
– Transmission system loading
• Nature of Disturbance
– Location, type of fault, and clearing time
– Lines lost due to the fault
– Relay/Breaker failure
– Failure of Remedial Action (RAS)
– “Sympathic” Relaying actions
57
PGE Company Training
Transient Stability – Modeling of Faults
TE’s During
Fault for Different
Types of Faults
58
PGE Company Training
Transient Stability Studies
• Assess Stability Limit Increases for Remedial Actions (Unit tripping, Under-frequency
Load Shedding, Under-voltage Load Shedding, Series Capacitor Bypassing/Insertion,
Capacitor Switching, etc).
59
PGE Company Training
Transient Stability Studies
60
PGE Company Training
Transient Stability Studies
• 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.
61
PGE Company Training
Transient Stabilty Simulation Study
• 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
• 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 Machine values for generation of interest (Pmech, Pelect, Efd, Vterm,
Speed, etc)
• Anything else of interest (Line Flows - P&Q, Apparent Impedance - R&X, etc)
63
PGE Company Training
Transient Stability Studies
Output Plot Observations
What to Observe
• 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
66
PGE Company Training
Voltage Stability
67
PGE Company Training
Power-Voltage Characteristics
68
PGE Company Training
Characteristics of Voltage Instability
69
PGE Company Training
Voltage and Reactive Power Planning
70
PGE Company Training
Q-V Curves
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
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
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
• 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
80
PGE Company Training
Voltage Stability
Reactive Margin for Outages
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: 1.05 pu
− 230 kV: 1.02 pu
− 69 -161 kV: 1.00 pu
82
PGE Company Training
Voltage Stability Summary
83
PGE Company Training
Questions?
85
PGE Company Training
52
This gives a good introduction on history and current development of composite load model Phase 2
project.
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