Modelling of Wind Ge Genera neratio tion n for fo r Fault Fault Leve Levell Studies Markus Pöller/DI Pöller/DIgSILEN gSILENT T GmbH
Fault Level Studies
• Relevant indices (according to IEC60909): – – – – –
Initial AC short circuit current Initial peak short circuit current AC-break short circuit current Peak break current Equivalent thermal short circuit current
Ik’’ ip Ib ip Ith
• Calculation of maximum short circuit currents for bus bar and circuit breaker rating. • Calculation of voltage sags due to grid faults. • Calculation of minimum fault levels for selectivity studies.
Methods for Short Circuit Calculation
X‘‘,X‘
U‘‘,U‘
R
Methods for Short Circuit Calculation
Top Envelope DC-Component iDC
Methods for Short Circuit Calculation
Planning Conditions
Operational Condition s Online- short circuit calculation
Simplified method (IEC, ANSI, ...) Reduced data set
Complete Method Complete data set
Method 1: Equivalent voltage source at fault location
Initial symmetrical (sub-transient) short-circuit current I”SC (Ikss) κ
ip
µ
Ib
m, n Ith
Method 2.1: Superposition Method
I"k, Uki
Method 2.2: Solution of differential Eq.
ik(t)
Wind Generator Response to Grid Faults
• First generation of WTGs (ASM): – contribution to Ik‘‘ and ip (Thevenin equivalent). – Disconnection within Tb-relevant time scales.
• Second generation of WTGs (DFIG with crow bar): – Linear contribution to Ik‘‘ and ip (Thevenin equivalent). – Contribution to Ib and ib hard to predict (crow bar protection).
• Third generation of WTGs (DFIG without crow bar, fully rated converter): – Linear contribution to Ik‘‘ and ip (Thevenin equivalent). – Controlled response in Ib-relevant time scales, according to grid code requirements.
Requir ed steady state response during grid faults (SDLWindV/Germany)
Additional reactive current: dIq
Imax=1p.u. Reactive currents prioritized
Voltage sag dU
Required WTG response (according to SDLWindV/Germany)
Voltage deviation/ add. reactive current
60ms Voltage step Add. Reactive current
30ms
Wind Generator Modelli ng for Fault Level Stud ies - Existing Approaches -
• Thevenin equivalent, likewise synchronous machines (“Equivalent synchronous generator model”, definition of x’’ and x’) – Can be integrated into standard short circuit analysis tools. – Accurate representation for faults close to WTG. – Very low accuracy for estimating remote contribution.
• Short circuit analysis using time domain simulation and dynamic models. – Complex model setup required. – Relatively long calculation times (“e.g. calculation of fault levels at all bus bars and terminals”). – Dynamic models not necessarily made for fault level studies. Accuracy for predicting subtransient time scales sometimes very poor.
Wind Generator Modelli ng for Fault Level Stud ies - Proposed Approach -
• Requirement: Simple and fast approach but sufficiently accurate for remote contribution • Subtransient time scale (relevant for Ik’’ and ip): – Linear model representation (classical representation) – Parameters: x’’ (or Ik’’ in case of solid fault)
• Transient time scale (relevant for Ib and ib) – Nonlinear model representation modelling steady state response to grid faults (reactive current contribution). – Parameters: K-factor, maximum current.
Wind Generator Modelli ng for Fault Level Stud ies - Proposed Approach -
• Iterative algorithm required for modelling controlled response
Ik‘‘
Method, integrated in DIgSILENT PowerFactory (complete method)
• Calculation of Ik’’ using linear method (with or without load flow initialisation) • Calculation of Ik’ using current iteration (considering reactive current priority when using load flow initialisation) • Calculation of ip using Ik’’ and IEC60909 method B or C (user’s choice) for decaying DC-component • Calculation of Ib using Ik’’ and Ik’ and assuming a transient short circuit time constant: − T / T '' I b = I k '+ I k ' '− I k ' e b
(
)
Method, integrated in DIgSILENT PowerFactory (complete method)
• Calculation of ib (peak break current) using Ib and IEC60909 method B or C (user’s choice) for decaying DC-component (at Tb) • Calculation of Ith using m and n factors according to IEC60909. The heating factor of the AC-component uses Ik’ instead of Ik’’.
Method, integrated in DIgSILENT PowerFactory Discussion
• Proposed method is easy to use and fast. • Simple current iteration method can be applied for calculating transient fault currents. – No additional factorisation of the Y-matrix required – Converges well, typically within 5 to 10 iterations -> around 5-10 times more time required as for a conventional IEC-type short circuit calculation
• Additional model parameters are limited to K and Imax • Additional simplification: – Using du=|du|=|ushc-uldf | instead of du=|ushc|-|uldf | provides conservative estimate of max. fault current contribution if short circuits shall be calculated without preceding load flow.
Example – Linear Method T N E L I S g I D
Externes Netz
0,326 kA 0,326 kA 0,000 kA
Substation/BB2 Substation/BB1
102,517 102,517 0,932 0,390
Ikss=2,588 ip=6,788 Iks=2,006 Ib=2,054 ib=5,327 20kV Station/BB
0,000 kA 0,000 kA 0,000 kA
0,326 kA 0,326 kA 0,000 kA
Synchronm.. r T
1,794 kA 1,794 kA 1,794 kA
Ikss=2,59 ip=6,79 Iks=2,01 Ib=2,05 ib=5,33
~ G
0,000 kA 0,000 kA 0,000 kA
0,802 kA 0,213 kA 0,261 kA
2,460 0,123 -5,890
Cable
Ikss 0,802 kA Iks 0,213 kA Ib 0,261 kA
Cable(2)
0,205 kA 0,053 kA 0,000 kA
0,201 kA 0,053 kA 0,000 kA
) 3 ( f r T
Ikss 5,95 kA Iks 1,55 kA
0,202 0,293 -128,689
5,954 kA 1,551 kA 0,000 kA
G4
DIgSILENT
0,213 0,309 -130,248
0,197 kA 0,053 kA 0,000 kA
f r T
) 1 ( f r T
5,840 kA 1,544 kA 0,000 kA
LV(2..
1 2 0 3 0 7 7 4 , 1 , , 5 3 0 -
Cable(1)
0,198 kA 0,053 kA 0,000 kA
) 2 ( f r T
5,954 kA 1,551 kA 0,000 kA
LV(3..
3,281 0,164 -5,756
2,917 0,146 -5,827
Cable(3)
5,749 kA 1,538 kA 0,000 kA
LV(1.. 5,840 kA 1,544 kA 0,000 kA
G3
0,222 0,322 -131,363
5,719 kA 1,536 kA 0,000 kA
LV 5,749 kA 1,538 kA 0,000 kA
G2
0,225 0,326 -131,712
Ikss 5,72 kA Iks 1,54 kA
5,719 kA 1,536 kA 0,000 kA
G1
Wind Generation Modelling for Fault Level Analysis
Projekt: Demo
Example – Iterative Method T N E L I S g I D
Externes Netz
0,326 kA 0,326 kA 0,000 kA
Substation/BB2 Substation/BB1
102,517 102,517 0,932 0,390
Ikss=2,588 ip=6,788 Iks=2,025 Ib=2,071 ib=5,351 20kV Station/BB
0,000 kA 0,000 kA 0,000 kA
0,326 kA 0,326 kA 0,000 kA
Synchronm.. r T
1,794 kA 1,794 kA 1,794 kA
Ikss=2,59 ip=6,79 Iks=2,02 Ib=2,07 ib=5,35
~ G
0,000 kA 0,000 kA 0,000 kA
0,802 kA 0,231 kA 0,278 kA
2,460 0,123 -5,890
Cable
Ikss 0,802 kA Iks 0,231 kA Ib 0,278 kA
Cable(2)
0,205 kA 0,058 kA 0,000 kA
0,201 kA 0,058 kA 0,000 kA
) 3 ( f r T
Ikss 5,95 kA Iks 1,67 kA
0,202 0,293 -128,689
5,954 kA 1,673 kA 0,000 kA
G4
DIgSILENT
0,213 0,309 -130,248
0,197 kA 0,058 kA 0,000 kA
f r T
) 1 ( f r T
5,840 kA 1,673 kA 0,000 kA
LV(2..
1 2 0 3 0 7 7 4 , 1 , , 5 3 0 -
Cable(1)
0,198 kA 0,058 kA 0,000 kA
) 2 ( f r T
5,954 kA 1,673 kA 0,000 kA
LV(3..
3,281 0,164 -5,756
2,917 0,146 -5,827
Cable(3)
5,749 kA 1,673 kA 0,000 kA
LV(1.. 5,840 kA 1,673 kA 0,000 kA
G3
0,222 0,322 -131,363
5,719 kA 1,673 kA 0,000 kA
LV 5,749 kA 1,673 kA 0,000 kA
G2
0,225 0,326 -131,712
Ikss 5,72 kA Iks 1,67 kA
5,719 kA 1,673 kA 0,000 kA
G1
Wind Generation Modelling for Fault Level Analysis
Projekt: Demo
Modelli ng of Wind Generation for Fault L evel Studies Summary
• International short circuit standards, like IEC60909 or ANSI C37 don‘t provide any guidelines for the modelling of wind generation. • Unlike conventional synchronous or asynchronous machines, converter driven generators have a controlled response within the time scales that are relevant for fault level studies. • The proposed method combines elements of IEC60909 and G74 with an iterative approach. • WTG model definition only requires two additional parameters, K and imax for modelling the reactive current response of WTGs. • The proposed iterative method is based on a fast current iteration, which typically requires 5 to 10 iterations (and no re-factorisation of matrices during iteration) The
osed method is available in the n
ion of DIgSILENT
Thank You
Markus Pöller
[email protected]
DIgSILENT GmbH Heinrich-Hertz-Str. 9 72810 Gomaringen www.digsilent.de
