This material is based upon work supported by the PNNL’s Laboratory Directed Research and

Development (LDRD) program, and U.S. Department of Energy’s Office of Energy Efficiency and
Renewable Energy (EERE) under Solar Energy Technologies Office (SETO) Award Number 38637.

Model Benchmarking of a Droop-Controlled, Grid-Forming Inverter

Model (REGFM_A1)

Wei Du
Staff Research Engineer
Pacific Northwest National Laboratory

WECC Model Validation Subcommittee Annual Meeting

May 25, 2023

OUTLINE
➢ Timeline of REGFM_A1 Model
➢ Review the Basic Concept of Grid-Forming Droop Control
➢ Comparison with Field Test Results
➢ Comparison with CERTS Microgrid field test results
➢ Comparison with OEM (SMA) field test results

➢ REGFM_A1 Model Benchmarking Results

➢ Single-GFM Infinite-Bus system
➢ Two-GFM islanded system

2
Timeline for the REGFM_A1 Model

GFM vendor
Model spec SMA provided Updated model Model
approved detailed spec to include benchmarking
(GFMDRP_A) suggestions Vflag completed
2021.12 2022.1 2022.9 2023.5

2021.12 2022.5 2023.2

NREC sent the Updated model Model renamed
model spec to spec to include to REGFM_A1.
GFM OEMs Qmax/Qmin Software
control block vendors started
to implement

3
Model Specification of a Droop-Controlled, Grid-Forming Inverter (REGFM_A1)
• The model includes the voltage source representation, P-f and Q-V droop control, P/Q limiting, and fault current limiting
• Most of the control blocks came from the CERTS Microgrid Project[1,2]
• SMA suggested to add the Qmax/Qmin control block, and the Vflag=0 option
XL V δV

P, Q, I φ
Edroop δdroop I < Imax
I Imax E δE
V δV+jXLImax φ

Fault current limiting

Voltage source behind impedance

Pinv ω δdroop VFlag Emax

-
mp ω0 1/s Vset 0
Edroop
+ ++ Qinv +
Suggested by SMA
Pset mq - Emin
+ ω +
Emax
+ Vref + Edroop
0 ω0 0 1 kpv
Pmax Qmax - ++
kppmax +- kpqmax ++ Vinv Emin
+- ++ Emax
0 0
kiqmax/s kiv/s
kipmax/s Qf +
Pinv + Emin
+
+ kiqmax/s
kipmax/s
0
0 Qmin +- +
+ Suggested by SMA
kpqmax
Pmin +- kppmax ++
0
0
P-f droop and P Limiting Q-V droop and Q Limiting

[1] Lasseter, Robert H., et al. "CERTS microgrid laboratory test bed." IEEE Transactions on Power Delivery 26.1 (2010): 325-332.
[2] Du, Wei, Robert H. Lasseter, and Amrit S. Khalsa. "Survivability of autonomous microgrid during overload events." IEEE Transactions on Smart Grid 10, no. 4 (2018): 3515-3524.
4
Basic Concept of a Droop-Controlled, Grid-Forming Inverter
• A grid-forming inverter behaves as a controllable voltage source behind impedance
• Two ideal voltage sources cannot be paralleled. The coupling reactance XL is very important for controller design
➢ If XL is well designed (e.g., 5%-20%): P∝δ, Q∝E

 p =  ( − 0 )dt

• Droop Control: Parallel multiple voltage sources in a system

➢ P vs. f droop ensures the phase angles of multiple voltage sources are synchronized
➢ Q vs. V droop avoids large circulating vars between voltage sources

P vs. f droop Q vs. V droop

δP P ω δP (Negative feedback control)

5
Comparison with the CERTS Microgrid Field Test Results

6
CERTS/AEP Microgrid Testbed
• AEP/CERTS testbed: one of the earliest inverter-based microgrids in the world, funded by DOE
• Principle Investigator: Prof. Bob Lasseter from University of Wisconsin-Madison
• The CERTS Microgrid Program has been running for almost 20 years
A 100% Grid-Forming-Inverter-based testbed

Sources

Loads

Static Switch

60 kW Tecogen Inverter-coupled
IC engine-generator
http://certs.lbl.gov/certs-der-pubs.html
CERTS/AEP Testbed
7
[1] Lasseter, R.H., Eto, J.H., Schenkman, B., Stevens, J., Vollkommer, H., Klapp, D., Linton, E., Hurtado, H. and Roy, J., 2010. CERTS microgrid laboratory test bed. IEEE Transactions on Power Delivery, 26(1)
Under-Frequency Load Shedding Testing (All-GFM-based System) CERTS/AEP Test Site

Frequency Feeder A
➢ After loss of the 58 kW ESS, the total 220 kW load exceeds the 193 kW Relay
maximum generation of A1 and B1
Inverter A1 Load Bank 3 Inverter A2 Load Bank 4
➢ Load Bank 4 is tripped in 0.5 s by the frequency relay Feeder B

➢ The overload mitigation control helps to trigger under-frequency load shedding ESS
Energy Storage Load Bank 5
when the entire system is overloaded

Pset ω0
P + Δω + + ω
- mp +
+
0
Pmax
kppmax+kipmax/s
+-
P +
+
Pmin +-
kppmax+kipmax/s
0

EMT and phasor simulation results EMT

Phasor
Field test results from CERTS/AEP testbed

[1] Wei Du, Francis K. Tuffner, Kevin P. Schneider, Robert Lasseter, et al., “Modeling of Grid-Forming and Grid-Following Inverters for Dynamic Simulation of Large-Scale Distribution Systems”.
IEEE Transactions on Power Delivery, 2020. 8
Under-Frequency Load Shedding (GFM & Machine Mixed System) Frequency Feeder A
Relay

• The loss of ESS results in the overload of the entire microgrid Inverter A1 Load Bank 3 Load Bank 4
• Droop curve becomes vertical because of Pmax control, triggering under- Feeder B
frequency load shedding
ESS
• GridLAB-D simulation, PSCAD simulation, and field test results match well Energy Storage Generator B1 Load Bank 5
with each other
CERTS/AEP Testbed

EMT
EMT and phasor simulation results
Phasor
Field test results from CERTS/AEP testbed

[1] Du, Wei, Robert H. Lasseter, and Amrit S. Khalsa. "Survivability of autonomous microgrid during overload events." IEEE Transactions on Smart Grid 10, no. 4 (2018):
3515-3524.
Comparison with the SMA GFM Field Test Results

10
 Comparison between the SMA Field Test Results and the PSLF Simulation Results
• PSLF simulation results match the SMA hardware testing results
• Case study was performed on the micro-WECC system for frequency regulation
• IBR penetration level: 73%, 10% headroom
• Both the simulation and hardware testing show that droop-controlled GFM can
significantly improve the system primary frequency response

73% penetration of GFMs in the micro-WECC system SMA Hardware Test Results on a Small System[1]

Droop GFM Model

VSM GFM
Droop GFM

(Simulation credit: Dmitry, BPA)

[1] A. Knobloch et al., "Synchronous energy storage system with inertia capabilities for angle, voltage and frequency stabilization in power grids," 11th Solar & Storage Power System Integration Workshop (SIW 2021), 2021, pp. 71-78 11
Comparison between the SMA Field Test Results and the PSLF Simulation Results
• The GFM unit behaves as a controllable voltage source behind impedance, so it increases the output power
almost instantaneously after the disturbance
• The synchronous generator’s output power is clamped so its speed does not change too much

PSLF Simulation Results of Micro-WECC System (Credit: Dmitry, BPA) SMA Hardware Test Results on a Microgrid[1]

12
REGFM_A1 Model Benchmarking Results

13
Model Specification of a Droop-Controlled, Grid-Forming Inverter (REGFM_A1)
• The model includes the voltage source representation, P-f and Q-V droop control, P/Q limiting, and fault current limiting
• Most of the control blocks came from the CERTS Microgrid Project[1,2]
• SMA suggested to add the Qmax/Qmin control block, and the Vflag=0 option
XL V δV

P, Q, I φ
Edroop δdroop I < Imax
I Imax E δE
V δV+jXLImax φ

Fault current limiting

Voltage source behind impedance

Pinv ω δdroop VFlag Emax

-
mp ω0 1/s Vset 0
Edroop
+ ++ Qinv +
Suggested by SMA
Pset mq - Emin
+ ω +
Emax
+ Vref + Edroop
0 ω0 0 1 kpv
Pmax Qmax - ++
kppmax +- kpqmax ++ Vinv Emin
+- ++ Emax
0 0
kiqmax/s kiv/s
kipmax/s Qf +
Pinv + Emin
+
+ kiqmax/s
kipmax/s
0
0 Qmin +- +
+ Suggested by SMA
kpqmax
Pmin +- kppmax ++
0
0
P-f droop and P Limiting Q-V droop and Q Limiting

[1] Lasseter, Robert H., et al. "CERTS microgrid laboratory test bed." IEEE Transactions on Power Delivery 26.1 (2010): 325-332.
[2] Du, Wei, Robert H. Lasseter, and Amrit S. Khalsa. "Survivability of autonomous microgrid during overload events." IEEE Transactions on Smart Grid 10, no. 4 (2018): 3515-3524.
14
Single-GFM-Infinite-Bus System
• 0.05 pu Step Increase in Voltage
• VFlag=0

15
Single-GFM-Infinite-Bus System
• 0.05 pu Step Increase in Voltage
• VFlag=1

16
Single-GFM-Infinite-Bus System
• 0.05 pu Step Decrease in Voltage
• VFlag=0

17
Single-GFM-Infinite-Bus System
• 0.05 pu Step Decrease in Voltage
• VFlag=1

18
Single-GFM-Infinite-Bus System
• Frequency step up from 60 Hz to 60.2 Hz
• VFlag=0

19
Single-GFM-Infinite-Bus System
• Frequency step up from 60 Hz to 60.2 Hz
• VFlag=1

20
Single-GFM-Infinite-Bus System
• Frequency step down from 60 Hz to 59.8 Hz
• VFlag=0

21
Single-GFM-Infinite-Bus System
• Frequency step down from 60 Hz to 59.8 Hz
• VFlag=1

22
Single-GFM-Infinite-Bus System
• 0.1 s Short-Circuit Fault
• VFlag=0

23
Single-GFM-Infinite-Bus System
• 0.1 s Short-Circuit Fault
• VFlag=1

24
 Z1 Z2
Two-GFM Islanded System
XLA1 K XLA2
• Step Increase in Load (EA1,δA1) (EA2,δA2)
• VFlag=0 A1 Load 1 Load 2 A2

A Two-GFM Islanded System

Response of GFM2 25
Response of GFM1
 Z1 Z2
Two-GFM Islanded System
XLA1 K XLA2
• Step Increase in Load (EA1,δA1) (EA2,δA2)
• VFlag=1 A1 Load 1 Load 2 A2

A Two-GFM Islanded System

Response of GFM1 Response of GFM2 26

Conclusions
• Model spec approved in December 2021
• Model spec received detailed suggestions from a GFM OEM
• Simulations results compare well with the field test results
• Model benchmarking completed and all models match very well

GFM vendor
Model spec SMA provided Updated model Model
approved detailed spec to include benchmarking
(GFMDRP_A) suggestions Vflag completed
2021.12 2022.1 2022.9 2023.5

2021.12 2022.5 2023.2

NREC sent the Updated model Model renamed
model spec to spec to include to REGFM_A1.
GFM OEMs Qmax/Qmin Software
control block vendors started
to implement

27
I’d like to make a motion to finally approve this REGFM_A1 model

28
Thank you
Wei Du
[email protected]

29
Backup Slides

30
Single-GFM-Infinite-Bus System
• SCR=2
• VFlag=0

31
Single-GFM-Infinite-Bus System
• SCR=1
• VFlag=0

32