WECC Model Validation Working Group Denver, Colorado May 18-19, 2009 Wind Power Plant Wind Power Plant Dynamic Modeling and Dynamic Modeling and Validation Validation Eduard Muljadi [email protected]National Renewable Energy Laboratory Golden, CO Abraham Ellis [email protected]Sandia National Laboratories Albuquerque, NM
29
Embed
WECC Model Validation Working Group Denver, Colorado May 18-19, 2009 Wind Power Plant Dynamic Modeling and Validation Eduard Muljadi [email protected].
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
WECC Model Validation Working Group
Denver, Colorado May 18-19, 2009
Wind Power Plant Wind Power Plant Dynamic Modeling and ValidationDynamic Modeling and Validation
Wind Power Plant (WPP) TopologyWind Power Plant (WPP) Topology
POI or connection to the grid Collector System
Station
Feeders and Laterals (overhead and/or underground)
Individual WTGs
Interconnection Transmission Line
BackgroundBackground((Dynamic Model of a Wind Power Plant)Dynamic Model of a Wind Power Plant)
Dynamic models are needed to study the dynamic behavior of power system. Users include system planners and operators, generation developers, equipment manufacturers, researchers, and consultants.
Wind Power Plant (WPP) models are needed to study the impact of proposed or existing wind power plants on power system and vice versa (i.e. to keep voltage and frequency within acceptable limits).
Models need to reproduce WPP behavior during transient events such as faults/clear events, generation/load tripping, etc.
G1G2
G3
WTGwind turbine generator
newline
lossofline
Resizing
ShortCircuit
Conventional Power Plant
Single Large (40MW to 1000MW+) generator
Prime mover: Steam, Combustion Engine – non-renewable fuel
Controllability: adjustable up to max limit and down to min limit.
Located where convenient for fuel and transmission access.
Generator: Synchronous Generator
Fixed speed – no slip: Flux is controlled via exciter winding. Flux and rotor rotate synchronously.
Located at wind resource, it may be far from the load center.
Generator: Four different types (fixed speed, variable slip, variable speed, full converter)
Type 3 & 4: variable speed with flux oriented controller (FOC) via power converter. Rotor does not have to rotate synchronously.
Differences between Differences between Wind Power Plant and Conventional Power PlantWind Power Plant and Conventional Power Plant
Four basic types, based on the WTG technology:
Type 1 – Fixed-speed, conventional induction generators
Variable Slip WTG Type 2 – Induction generators with variable rotor resistance
Variable Speed WTGs Type 3 – Doubly-fed asynchronous generators with rotor-side converter Type 4 – Asynchronous generators with full converter interface
• WECC developed and adopted guidelines for WPP representation• Based on single-machine representation• Access to guidelines: www.wecc.biz -> Committees -> MVWG -> WGMG
Major components of WPP Equivalent Representation:
• Wind Turbine Generator (WTG) Equivalent and power factor correction (PFC) caps• Pad-mounted Transformer Equivalent• Collector System Equivalent branch.
– For N identical pad-mounted transformers, each with
impedance ZT , the equivalent impedance ZTeq is:
– For 1.5 MVA to 3 MVA, 600V/34.5kV: ZT = 6% on
transformer MVA base; adjustable (fixed) taps
N
ZjXRZ T
TeqTeqTeq on pad-mounted transformer MVA base
on N × pad-mounted transformer MVA base TTeq ZZ OR
Reactive Power LimitsReactive Power Limits
• Type 1 and Type 2 WTGs (induction machines)– At full output and nominal voltage, PF ~ 0.9 under-
excited => Qmin = Qmax = Qgen = -½ Prated
– MSCs at WTG terminals maintain PF near unity at nominal voltage => Qcap = ½ Prated
– Example:
~100 MW WPP, Type 1 WTG• Pgen = Prated = 100 MW
• Qmin = Qmax = Qgen = -50 Mvar
• Qcap = 50 Mvar
Reactive Power LimitsReactive Power Limits
• Type 3 and Type 4 WTGs– Line-side converter allows for PF adjustment at WTG
terminals; MSCs at WTG terminals are not needed
– If WTG PF is fixed, Qmin = Qmax = Pgen × tan(cos-1(PF))
– If WTG PF range is used for steady-state voltage control, set Qmin and Qmax according to PF range and Pgen
• WTG PF adjusted by plant-level controller. Patents may apply.
– Example:
~
100 MW WPP, Type 3,+/-0.95 reactive range, controlling POI voltage
• Pgen = Prated = 100 MW
• Qmin = -33 Mvar; Qmax = 33 Mvar
• Qgen depends on POI voltagePOI
Dynamic RepresentationDynamic Representation
Power System Dynamic Time ScalesPower System Dynamic Time Scales
Source: Dynamic Simulation Applications Using PSLF – Short Course Note – GE Energy
• Objective of the model? (fault transient or long term dynamic, mechanical or electrical characteristics, power system transient or power quality of the wind power plant).
• Major components (wind turbine type 1, 2, 3, 4, include aerodynamic or use simplified one, positive sequence or 3 phase representation, use complete generator and power electronics or simplified power conversion, include system protection?).
• For each component block, the equations governing the function of the block can be derived (assumptions may be made to simplify power converter, aerodynamic, saturation level, nonlinear circuits).
• Control algorithm will be formulated according to the wind turbine system to be modeled (WTG type 1 is different from WTG type 4 etc, different manufacturers may have unique algorithms).
• Choose the method of calculation and/or the software to be used (PSLF, PSSE for positive sequence representation, PSCAD, PSIM, EMTP if detail of the power electronics switching to be emphasized, MATLAB/Simulink, Mathcad may be considered for different aspects of simulation).
Example of Wind Turbine ModelExample of Wind Turbine Model
WECC Generic Models
• Generic model development in PSSE/PSLF– Complete suite of prototype models implemented
– Type 3 model formally approved for use in WECC; others pending
• Current focus– Model validation & refinement (e.g., freq. response)– Identification of generic model parameters for different manufacturers (at
NREL)
Model Type Type 1 Type 2 Type 3 Type 4Generator wt1g wt2g wt3g wt4gExcitation / Controller wt2e wt3e wt4eTurbine wt1t wt2t wt3t wt4tPitch Controller wt1p wt2p wt3p wt4p
• Prepare the simulation carefully (i.e. the correct information must be used): type of WTG, collector system impedance, transformers, power system network, input parameters to dynamic models, control flags settings set-up, reactive power compensation at the turbine level or at the plant level.
• Initialize the simulation based on pre-fault condition (check v, i, p, q, f, if available).
• Recreate the nature of the faults if possible, otherwise use the recorded data to drive the simulation and compare the measured output to the simulated output (pre-fault, during the fault, post-fault).
• Represent the events for the duration of observation (any changes in wind, how many turbine were taken offline due to the fault?).
• Prepare the data measured to match the designed frequency range of the software used.
• Field data is expensive to monitor, public domain data is limited, difficult to get, and quality of data needs to be scrutinized
– Anticipate errors in the measurement and make the necessary correction– The location of simulation should be measured at the corresponding monitored data.
Method of WTG Model ValidationMethod of WTG Model ValidationComparison against field test measurementComparison against field test measurement
Example of Dynamic ModelExample of Dynamic ModelSimulation versus Field Data (Type 3)Simulation versus Field Data (Type 3)
W
Pad-mounted Transformer Equivalent
91% WTGs stays “on” after the fault.
Collector System
Equivalent
Interconnection Transmission
LinePOI or Connection to
the Transmission System
Station Transformer(s)
W 9% WTGs were dropped of line during the fault.
Two Turbine Representation
Interconnection Transmission
LinePOI or Connection to
the Transmission System
Station Transformer(s)
136 WTGs were represented
9% WTGs were dropped of line during the fault.
Complete Representation (136 turbines)
V and f
0.2
0.4
0.6
0.8
1
1.2
0 0.5 1 1.5 2Time (s)
Volta
ge (p
.u.)
0.95
0.99
1.03
1.07
1.11
1.15
Freq
uenc
y (p
.u.)
Vf
Real Power Comparison
0
20
40
60
80
100
120
140
0 0.5 1 1.5 2 2.5 3 3.5 4
Time (s)
Rea
l Pow
er (M
W)
P-sim-1wtg (MW)P-measured (MW)P-sim-136WTG
Reactive Power Comparison
-60
-40
-20
0
20
40
60
80
0 0.5 1 1.5 2 2.5 3 3.5 4
Time (s)
Rea
ctiv
e Po
wer
(MVA
R)
Q-sim-1wtg (MVAR)Q-measured (MVAR)Q-sim-136WTG
W Wind Turbine Generator Equivalent
InputV and f
A C BSystem Generator
Compare P&Q measured to P&Q simulatedV and f
Regulated Bus
Example of Dynamic ModelExample of Dynamic ModelSimulation versus Field Data (Type 3)Simulation versus Field Data (Type 3)
• Another method to validate new model is to use another model that has been validated against field measurement as a benchmark model.
• Several transient fault scenarios can be performed using both models, and the results can be compared.
• Parameter Tuning– The new model and the benchmark model may have some differences in
implementation, we may have to perform parameter tuning to match the output of the benchmark model.
– However, one should realize that the model may not be able to match the output of the benchmark model in all transient events.
• Parameter Sensitivity– In order to limit the number of parameters that should be tuned, parameter
sensitivity analysis may need to be performed.– In general important parameters are varied one by one and the sensitive
parameters can be tuned to match the bench mark model.
Method of WTG Model ValidationMethod of WTG Model ValidationComparison against other model (Benchmarking)Comparison against other model (Benchmarking)
Example of Model to Model Comparison Example of Model to Model Comparison (Type 2 “Detailed” Model vs Generic Model)(Type 2 “Detailed” Model vs Generic Model)
Terminal Voltage Real Power
Reactive Power Turbine Speed
Parameter SensitivityParameter Sensitivity
The output of the simulation and the measured data can be used to find the total error of the measurement.
Perr = |Pmeas.-Psimulated|
Qerr = |Qmeas.-Qsimulated|
The error and the sensitivity parameter k1 with respect to the error can be computed.
Use the other parameters k1, k2, k3, k4 etc
The parameter sensitivity can be observed from the results.
The trend can be used to drive the changes of the parameters.
Actual Wind Plant
Model with the parameter to be tuned
Input T(k)
+
-
Parameter k
SkT k( ) dT k( )
dk
k
T k( )Sk
T k( )
Parameter Sensitivity - ExampleParameter Sensitivity - Example
Sensitivity of parameter f1 (Sf1) to Perror and Qerror
-0.04
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0 0.5 1 1.5 2 2.5
f1
Sp
_f1
and
Sq
_f1
Sp_f1
Sq_f1
28/2
Response of P output.
Sensitivity of P output to a range of parameters.Qualitatively similar results for other outputs.Note that a lot of parameters have small and/or correlated influence.Sensitivities obtained as a by-product of running the simulation.
Parameter Sensitivity - ExampleParameter Sensitivity - Example
SummarySummary
Wind power plant model is different from conventional synchronous generator modeling in different aspects:
Different types of generators used (level of capability for reactive compensation, voltage controllability, and LVRT are different)
Wind power plant represents hundreds of generators (i.e. the collective behavior of all turbines is more important than the behavior of individual turbines)
Wind power plant covers a very large area. During faults, each generator may have different operating condition with respect to other generators due to diversity within the wind plant.
The simplification or equivalencing wind power plant may compromise the accuracy of the simulation, however, a complete model requires to represent hundreds of turbines (impractical)
In some cases, components of the system needs to be simplified for many different reasons (NDA, complexity, time constant of interest).