IEEE Electrification - September 2022 - 35

inverters and synchronous condensers (but no synchronous
generators, such that all of the generation comes
from PVs and wind).
The PSCAD computer simulation study also indicated
that implementing GFM controls on one 30-MVA portion
of a 60-MVA PV-BESS plant stabilized most unstable scenarios
and that putting all 60 MVA of that plant in GFM
mode stabilized even the scenario with zero synchronous
machines. (Note that this study addressed dynamic stability
and resilience to faults and contingencies only; not all
scenarios studied are viable unless other concerns, such
as protection system operations, are addressed.) A related
EMT study by a utility consultant also found significant
benefits to GFM controls on this system (see https://dms.
puc.hawaii.gov/dms/DocumentViewer?pid=A1001001A21F
14B62327F00172). However, no simulation model fully captures
all IBR behavior, so there is great value in validating
such findings experimentally. Field demonstrations of
such scenarios are possible but involve the risk of unexpected
behavior potentially leading to outages.
To help bridge the gap between pure computer simulations
and live field tests, it is possible to connect real hardware
inverters to a computer model running in real time
in such a way that the hardware inverter and the power
system interact dynamically; this is called a PHIL simulation.
The use of a PHIL simulation allows experimental validation
using real hardware without putting customers
and utility equipment at risk; one or more hardware
inverters can be dynamically connected to a real-time
model of a power system to experimentally observe the
dynamics of the inverter-grid system. This is especially
valuable when evaluating whether to operate a power system
in an unprecedented way, such as in an extremely
high-IBR scenario.
PHIL simulation is not intended to replace the modelonly
dynamic simulations that are fundamental to validating
the dynamic stability of power systems. Instead, its
purpose is to improve confidence in computer simulations
by replacing key components (in this case, a PV-BESS plant)
with real hardware to validate selected key results. In the
future, as industry gains confidence in GFM inverter performance
and the ability of dynamic models to capture
that performance, computer simulations (without PHIL)
will be sufficient for most, if not all, interconnections.
This article presents the first PHIL demonstration the
authors are aware of to evaluate the ability of a hardware
GFM inverter to stabilize otherwise unstable operating
scenarios of a transmission system. The demonstration
was conducted using a real-time model of the year 2023
Maui power system (Figure 1), which includes two large
transmission-connected PV-BESS hybrid power plants
expected to be commissioned around 2023: a 60-MVA PVBESS
plant with a 240-MWh battery system and a 15-MVA
PV-BESS plant with a 60-MWh battery system. (These two
large hybrid power plants are the first of several expected
to be installed on the island in the coming years.) The
model also contains four wind plants (two of which have a
short-duration BESS for frequency response only) and a
detailed dynamic representation of the very large capacity
of distribution-connected PVs present on Maui.
Half of the 60-MVA PV-BESS plant (which will be connected
at two separate 30-MVA interconnections) in the
model was replaced with a PHIL connection to a 2.2-MVA
real-hardware BESS inverter capable of operating in GFL or
GFM mode, allowing the impact of GFM operation to be
clearly evaluated. The output of the 2.2-MVA inverter was
scaled up in a manner described in the " A Real-Time
Model of the Maui Power System " section to represent
the full 30-MVA plant section. The other half of the plant
was represented within the real-time simulation by a
detailed dynamic model, also capable of operating in GFL
or GFM mode. The PHIL setup, consisting of the real-time
Maui power system model and the 2.2-MVA hardware
inverter, was used to simulate N-1 loss-of-generation
events and fault events to evaluate the stability of the
system under several operating scenarios with varying
levels of synchronous machines online. The details of the
model, the PHIL experiment setup, and the results are
described in the following sections.
A Real-Time Model of the Maui Power System
The real-time model of the Maui power system was
developed in RSCAD software (an electromagnetic simulation
software developed for real-time simulation on
RTDS computing hardware) starting from Hawaiian Electric's
full 200-bus PSSE model, with most dynamic components
translated from the PSCAD model mentioned
earlier. To run the RSCAD model in real time on an available
four-core RTDS simulator while keeping the key
dynamic components, the network model was reduced
by aggregating network subsections and dynamic models.
This was accomplished by identifying single-port and
two-port subnetworks containing relatively homogeneous
dynamic models (e.g., load and distributed PV) and
replacing each with a reduced network containing a single
instance of each type of dynamic model using the
code available publicly at https://github.com/NREL/PSSE
_Network_Reduction.
More than 170 aggregated behind-the-meter PV models,
representing more than 70% of the generation in the daytime
operating scenario, were aggregated and represented
by a detailed EMT inverter model (available at https://
github.com/NREL/PyPSCAD) at 10 aggregated substation
locations. This inverter model is an averaged-switch
model including all ac-side control details, such as the
PLL, power controllers, and inner current controllers,
which have been shown to be important for capturing the
dynamics of high-IBR systems in previous work by the
authors and others. The dynamics of the dc side circuit are
not represented in detail aside from limiting the available
dc power. Three-phase aggregate models were used to
approximate the dynamics of the distributed generation,
IEEE Electrification Magazine / SEPTEMBER 2022
35
https://dms.puc.hawaii.gov/dms/DocumentViewer?pid=A1001001A21F14B62327F00172 https://dms.puc.hawaii.gov/dms/DocumentViewer?pid=A1001001A21F14B62327F00172 https://dms.puc.hawaii.gov/dms/DocumentViewer?pid=A1001001A21F14B62327F00172 https://www.github.com/NREL/PSSE_Network_Reduction https://www.github.com/NREL/PSSE_Network_Reduction https://github.com/NREL/PyPSCAD https://github.com/NREL/PyPSCAD
IEEE Electrification - September 2022

Table of Contents for the Digital Edition of IEEE Electrification - September 2022
