Advantages of the Virtual Synchronous Machine

regulation for integrating Low-Inertia Variable

Renewable Generation in Transmission Systems
Fabio Bignucolo, Riccardo Stecca, Massimiliano Coppo
Department of Industrial Engineering, University of Padova
Padova, Italy
[email protected]

Abstract—The increasing penetration of renewables requires over 50% of the network overall load could result in electrical
Transmission System Operators to manage networks where both grid instability problems [3]-[6]. It is worth noting that this VRG
the rotating inertia and the primary reserve are sensibly reduced level has been already reached in some contries, e.g., in Italy
compared to the past. This issue leads to potential frequency wide during some sunny and windy non-working days (huge amount
oscillations and instability in the case of sudden and unpredicted of static generation combined with a reduced overall electric
load variations. Therefore, finding alternative solutions able to demand).
preserve the network stability, i.e. replacing the inertial
contribution intrinsically provided by traditional synchronous In this scenario, new technical solutions are strictly required
generators, is becoming an urgent issue. In this work, the to effectively prevent: i) lower, variable and heterogeneous
frequency regulation has been investigated by analyzing a grid inertia; ii) wider frequency oscillations, i.e. reduced Frequncy
with high penetration of variable renewable generation, such as Nadir (FN); iii) increased Rate Of Change Of Frequency
photovoltaic plants and wind farms. Studying the stabilizing (ROCOF) in the case of severe network perturbations; iv) higher
contribution of suitably controlled static converters, capable of voltage instability; and v) difficulties in both programming load
supplying synthetic inertia in the case of frequency perturbations, shedding and preserving the ROCOF-based protections
is the main target of the paper. A couple of approaches in terms of effectiveness [3][4]. One of the proposed solutions is
power converter control scheme have been examined in depth and intentionally supplying Synthetic Inertia (SI), i.e. that electronic
compared through dynamic simulations: on one hand, the active
converters are called to provide an active power injection or
power contribution evaluated according to the swing equation
absorption, proportional to the ROCOF measured at their
based inertial response; on the other hand, the innovative control
of the power converter as a virtual synchronous machine. Results
terminals. SI could be supplied by each electronic converter able
show that the latter scheme is able to better counteract the to rapidly modulate its DC power depending on the frequency
frequency oscillation in the time instants immediately after the measurement (e.g., photovoltaic and wind parks, storage
perturbing event, whereas both of the modeled controllers limit systems or HVDC power converters).
the frequency oscillation width. DIgSILENT PowerFactory has As of today, the SI response is evaluated through the Swing
been used as simulation environment. Equation Based Inertial Response (SEBIR) [5]. This regulating
approach requires the ROCOF measurement, which means that
Keywords—Variable Renewable Generation, Synthetic Inertia,
ROCOF, Frequency Nadir, SEBIR (Swing Equation Based Inertial
delays introduced by the frequency measurement chain limit the
Response), VSM (Virtual Synchronous Machine) SI effectiveness. In other words, a True Inertial (TI) response,
i.e. the one intrinsically supplied by Synchronous Generators
(SGs), may be difficult to be obtained in the time instants
I. INTRODUCTION immediately after the perturbing event. The shortcomings of this
Electricity grids are undergoing a drastic transformation control strategy in the case of high VRG penetration are
process, moving from few centralized synchronous generators discussed in detail in [3]-[5]. Controlling the electronic power
to a huge number of distributed units, mainly based on converter as a Virtual Synchronous Machine (VSM) could be an
Renewable Energy Sources (RESs). This phenomenon, interesting alternative to avoid the SEBIR response intrinsic
although necessary to match the international goals of CO2 delay [7]-[10]. Implementing the VSM approach, relying on a
emissions reduction, could lead to serious problems in terms of more detailed control scheme modelling the electromechanical
network frequency stability. In particular, the growing presence behaviour of an equivalent SG, the electronic converter could be
of Variable Renewable Generation (VRG) implies that able to inject an effective TI contribution in the main grid. More
traditional synchronous units are more and more excluded by the in detail, one of the advantages of this approach is the
market, resulting in a lower and time-variable system inertia. opportunity of emulating a generic SG, with rated charateristics
The reduction of primary control reserve is an additional issue customized according to frequency and voltage regulation
affecting the main grid stability [1][2]. Several studies available requirements. Unrealistic SG rated characteristics (e.g. very
in the literature confirm that increasing the VRG penetration high inertial constant) could be equivantely emulated too.

978-1-5386-2910-9/18/$31.00 ©2018 IEEE

 This work aims at investigating the technical feasibility, synchronous generator, by using different models or specific
benefits and disadvantages of SEBIR and VSM controls. These characteristics of this kind of machine and, in some cases,
regulation strategies have been applied to a Battery Energy adding further properties to improve the inertial response.
Storage System (BESS), i.e. to the power converter interfacing
the electrochemical storage plant with the grid. Frequency From the main network perspective, a power converter
trends, ROCOF, FN and the active power contribution provided regulated through the VSM control scheme behaves as a
by the electronic converter are used as Key Performance balanced positive-sequence sinusoidal voltage generator, behind
Indicators (KPIs) in the comparison among the modelled a filter impedance (usually simplified as a pure inductive
controllers. RMS simulations are carried out in DIgSILENT impedance). This is obtained by controlling both the magnitude
PowerFactory. and the angle of the virtual rotor voltage, compared with
magnitude and angle of the voltage at the Point of Common
The paper is structured as follows: Section II reports a brief Coupling (PCC) with the grid [3][8]. In this way, the power
introduction on SEBIR and VSM control systems, whereas converter operated as VSM emulates many of the advantages of
Section III describes a procedure for suitably tuning the VSM a traditional SM in providing an active power contribution
main settings, making use of the Parameter Estimation tool during a network perturbation. Furthermore, it is important to
embedded in the MATLAB Simulink environment. The realistic underline that: i) there are no limits in setting the rated
case study is introduced in Section IV, where representative characteristics of the equivalent SM emulated by the VSM-
results are reported in detail. Finally, Section V summarizes the based power converter (whereas a real rotating machine is
main conclusions and discusses abount possible future works. affected by several physical constraints); ii) no fast frequency
measurements are required, and ROCOF has not to be computed
II. SEBIR AND VSM CONTROL SCHEMES (i.e. no delays in the power converter response); iii) since the
VSM-based power converter operates as a voltage source
A. SEBIR control scheme connected to the main grid, it can supply fault currents, as well
as unbalanced and harmonic currents to non-linear loads. The
The most common existing control strategy applied to latter benefits are not obtainable by applying a current source
converters is Direct Quadrature Current Injection (DQCI) with approach. On the other hand, the VSM-based controller applied
Phase Locked Loop (PLL) type control [11]. The converter to the BESS electronic power converter requires a particular
control scheme implements an inner loop performing a current attention in preserving the static machine against dangerous
regulation, therefore a current source behaviour at the system over-currents.
rated frequency is expected. SEBIR is a control scheme in which
the SI response of the converter is provided by: i) measuring the
frequency through a PLL; ii) estimating the ROCOF; and iii) III. TUNING OF THE VSM CONTROLLER
computing the active power contribution according to the Swing In Fig. 1, the VSM diagram, comprising its control system
Equation. This approach is the easiest method for providing SI: and measurement devices, is shown. A MATLAB Simulink
in the DQCI architecture, converters are controlled by procedure to optimally tune the control parameters of the static
computing the set-points of both direct and quadrature currents converter operated in VSM mode has been developed for this
(Id and Iq, respectively), once the desired outputs in terms of work. The VSM modelled in MATLAB Simulink is a simplified
active power P and reactive power Q are known. This means frequency domain model representing its dynamic response to
that existing DQCI converters should be updated in both perturbations. The influence of measurement equipments,
hardware and software components to provide an ative power battery model and AVR is negleted, as well as the Current
contribution directly proportional to the ROCOF measure Limiter (CL). Coefficients are computed with the purpose of
(basing on the well-known Swing Equation and considering the obtaining a power converter dynamic response close to the
per-unit synthetic inertia Hsi), if the DC side can operate at behavior of a typical large-scale synchronous generator, with
variable power [5]. Unfortunately, this technique needs some known rated characteristics. In this way, it will be possible to
time to complete the ROCOF measurement and close the control substitute the traditional SGs with RES power plants and ESSs,
loop. Thus, a real TI response is not obtainable and consequently controlled as a VSM.
benefits are negligible in the first time instants after the network
perturbation. The 5 steps composing the tunig procedure are:
1. The frequency response of the grid is evaluated through
B. VSM control scheme dynamic simulations in DIgSILENT Powerfactory. A high
VSM is a complex control method which implements the
electromechanical synchronous machine model in the power
converter main drive. Differently from the SEBIR method, BATTERY
which simply provides a power response proportional to the MODEL
ROCOF, in the VSM approach a virtual rotor frequency is PLL
f

computed [4][8]. A few control techniques have been developed SoC

P
for the VSM, which are able to emulate the behaviour of the Wattmeter
Q GOVERNOR
Pt
Pmr VSM
Idc
synchronous generator, in such a way that the response of the ROTOR Inverter
Voltmeter
inverter is close to be equal to the one of a real synchronous
Eexc Elim EMULATOR Pmi
V AVR CL
generator with analogous characteristics. The main target of this Ampmeter
I
control technique is to correctly simulate the dynamics of a Fig. 1: Block scheme of the VSM used in the grid.
 PowerFactory (information about the model is reported in the

DIgSILEN
80,0
Synchronous generator next section), whereas the MATLAB Simulink model considers
VSM
only the primary control and its load transfer function.
60,0
Interactions among reactive power and frequency perturbations
Active power [MW]

have been neglected in the simplified network representation, as

40,0 well as current limitations in transmission lines and connected
plants. In addition, the DC side of the electronic power converter
20,0
is modelled as an ideal voltage source in the MATLAB Simulink
equivalent scheme.

0,00
Even if some approximations affect the simplified model, it
demonstrates to be reliable in tuning the VSM control
parameters. For example, results of a simulation in DIgSILENT
-20,0
-10,0 0,00 10,0 20,0 30,0 [s] 40,0
PowerFactory are depicted in Fig. 2 (VSM rated power equal to
2% of the overall network demand). The active power injected
Time [s]
by the static converter operated in VSM mode (blue line, in
Fig. 2: Comparison among active power behaviours of the electronic power [MW]) is close to the active power injection of a synchronous
converter operated in VSM mode (blue line, in [MW]) and a synchronous generator with the same rated power (red trend, in [MW]). The
generator (red line, in [MW]), in the case the VSM rated power is 2% of the
overall network demand.
error obtained in this case is equal to 2.1% of the machine rated
power. Since the Parameter Estimation tool is a deterministic
technique, potentially subjected to local minima issues, this
penetration of RES plants is considered (details are reported in could be one of the causes of this acceptable error among the
the next section). The disconnection of a traditional power plant depicted power trends.
is considered as perturbating event.
2. A simplified frequency domain model of the network

EXT GRID
Ext Grid

SG
~
(i.e. representing its dynamic response to perturbations) is
modelled in MATLAB Simulink. The transfer-function blocks
are suitably tuned to reproduce a frequency response similar to

Trf 0.69kV -..

TR WG B
the one in the original network model.

Overhead Line..
Line 380kV
3. The dynamic response of a standard synchronous

TR 220/19kV (B) RES

255 MVA 220/19 kV

255 MVA 220/19 kV

255 MVA 220/19 kV
Trafo 220/19kV (A)

TR 220/19kV
generator during the perturbating event, in terms of active power
injection trend, is analysed in DIgSILENT Powerfactory.
Trf 0.69kV -..
TR WG A
Several values of rotating inertia and SG rated power can be
Trf 0.69kV -..
TR WG C

considered.

250MVA 220/132kV
TR 220/132kV (1)
4. Similarly to point 2, a simplified frequency domain
version of the VSM control is represented in MATLAB
Simulink and connected to the equivalent network parametrized
at point 2. The static converter rated power is equal to the rated
power of the standard synchronous generator considered at the
previous point. Making use of the Parameter Estimation tool
Fig. 3: Extended representation of the High Voltage transmission network used
(embedded in MATLAB Simulink), the VSM control as case study.
parameters are tuned to obtain a dynamic response of the
electronic power converter close to the one of the synchronous
Bu

machine.
PV Pa..

210 MVA 132/15.75 kV

PV 3

TR 132/15.75kV (1)
~
LV(1)

Overhead Line 13..

255 MVA 13..

Line 132kV 3-A

5. The VSM control parameters evaluated in MATLAB

TR RES 1
255 MVA 13..
TR RES 3
DFIG..

DFIG..
WG 2
G

G
~

Simulink are imported in the DIgSILENT Powerfactory model.

210 MVA 132/15.75 kV
TR 132/15.75kV (3)

Overhead Line 13..

Line 132kV C-D
LV
PV 2
~

Then, the dynamic response of the power converter is studied in

210MVA 15.75kV Gas
Trf 0.69kV..
Overhead Line..

TR WG 1
Line 132kV A

SG 2
SG
~
Bus Gen 2

DIgSILENT Powerfactory in comparison with the rotating

Overhead Line 132kV
Line 132kV E-D

machine dynamic behaviour obtained at point 3.

Since the aim of the MATLAB Simulink model is studying
the dynamics of the network for the tuning of controllers, it is
210 MVA 132/20 kV

simplified respect to the extended representation in DIgSILENT

TR 132/20kV (A)

210 MVA 132/20 kV

TR 132/20kV (C)
210 MVA 132/20 kV
TR 132/20kV (B)

PowerFactory. In particular, only the power droop governor and

the rotor emulation system are represented in that compact
TR 20/11kV (Gen 6)
30MVA 20/11kV

TR 20/11kV (Gen 4-5)

representation. Oppositely, the DIgSILENT PowerFactory

30MVA 20/11kV

30MVA 20/11kV
TR 20/11kV (6)

model includes also the voltage regulation performed by

controlling the reactive power exchange at the generator
terminals. The network is modelled in detail in DIgSILENT
Fig. 4: Details of the High Voltage and Medium Voltage modelled grid.
 50,1
[Hz]

45 MVA ..
TR 11/0..
50.000 Hz
50,0

Frequency [Hz]
49,9 49.890 Hz

49,8
Ri SEBIR..

Ri VSM..

Ri VSM..

Ri VSM..

Ri VSM..

Ri VSM..

Ri VSM..

Ri VSM..
Ri SEB..

Ri SEB..

X = 30,000 s
49,7

49,6
-10,0 0,00 10,0 20,0 30,0 [s] 40,0
Time [s]
Fig. 5: Representation of the storage systems composed by several parallel- Fig. 6: Frequency trend in the case RES plants are out of service (thermoelectric
connected elementary units with different control schemes and rated power. power stations supply the entire network demand).

50,1

DIgSILE
Traditional 100% Traditional 40%
IV. CASE STUDY AND RESULTS [Hz] VRG 0% VRG 60%
The studied network is a plausible model of the electricity 50,0
50.000 Hz
transmission grid of Sicily (Fig. 3 and Fig. 4), i.e. the largest

Frequency [Hz]
island located in the south of Italy. The represented system is 49.890 Hz
interconnected with the remaining Italian peninsula by means of 49,9 49.879 Hz
few High Voltage (HV) transmission lines.
The VRG increasing penetration scenario is considered by 49,8
placing both a photovoltaic plant and a wind power park (doubly
fed induction generator technology) in parallel to each

X = 30,000 s
thermoelectric plant. Each RES plant has the same rated power 49,7
as the conventional power station connected in parallel, but for
each node only a generating system is in service. Different
49,6
scenarios are then obtained by connecting no more than one -10,0 0,00 10,0 20,0 30,0 [s] 40,0
generating plant to each generating node of the network. The Time [s]
mix among photovoltaic and wind plants is compliant with the Fig. 7: Frequency perturbation in the case the network demand is exclusively
present regional evolution of RES exploitation, reported in supplied by traditional generators (blue trend) and considering that VRGs cover
[12][13]. With these hypotheses, the study aims at 60% of the produced power (red line).
characterizing the frequency perturbation in the case a generator
50,1

DIgSILE
connected to the peninsular portion of the transmission system BESS: no f support
trips. The generation loss is approximately equal to 10% of the [Hz] BESS: SEBIR mode
BESS: VSM mode
Sicilian overall active power demand. Once the overall share of 50,0
50.000 Hz
RES is set, the selection of the plants operated by either source
is made randomly, since the actual position of each one does not
Frequency [Hz]

49.884 Hz
impact significantly the frequency response of the system. 49,9 49.879 Hz

Finally, two types of BESS operating modes are considered:

static power converters are controlled in SEBIR mode or in 49,8
VSM mode. The equivalent representation of the DC side of the
BESS units is independent from the control mode applied to the
X = 30,000 s

power converters. Referring to Fig. 5, for simulation 49,7

convenience, several independent units are connected in parallel
to study BESSs with different rated power Sn (ranging from 49,6
0.5% to 3.5% of the regional network overall load). In the -5,00 4,00 13,0 22,0 31,0 [s] 40,0
following, the BESS size is considered equal to 0.5%, 2.0% e Time [s]
3.5% of the network demand. No more than one elementary unit
Fig. 8: Frequency perturbation in the case 60% of the network demand is
is authorized to inject power at the common bus. In normal supplied by VRGs: no frequency support provided by the storage unit (black
conditions, the aggregate storage system supplies no power to trend), BESS operating in SEBIR mode (orange line) and BESS operating in
the network. VSM mode (blue line).
 TABLE I: CHARACTERIZATION OF THE SEBIR SUPPORT IN TERMS OF TABLE III: CHARACTERIZATION OF THE SEBIR SUPPORT IN TERMS OF
CONTAINMENT OF THE FREQUENCY OSCILLATION ROCOF, AT DIFFERENT TIME INSTANT AFTER THE PERTURBING EVENT

Sn of Frequency Time to reach Regime value of Sn of ROCOF at 0.1 s ROCOF at 0.5 s ROCOF at 1.0 s
SEBIR [%] Nadir (FN) [Hz] FN [s] frequency[Hz] SEBIR [%] [mHz/s] [mHz/s] [mHz/s]
0 49.6242 3.3444 49.8791 0 29.209 120.519 231.091

0.5 49.6337 3.2838 49.8791 0.5 29.231 120.531 229.917

2.0 49.6595 3.1774 49.8789 2 29.326 120.484 226.687

3.5 49.6819 2.9804 49.8787 3.5 29.478 120.454 223.821

TABLE II: CHARACTERIZATION OF THE VSM SUPPORT IN TERMS OF TABLE IV: CHARACTERIZATION OF THE VSM SUPPORT IN TERMS OF
CONTAINMENT OF THE FREQUENCY OSCILLATION ROCOF, AT DIFFERENT TIME INSTANT AFTER THE PERTURBING EVENT

Sn of VSM Frequency Time to reach Regime value of Sn of VSM ROCOF at 0.1 s ROCOF at 0.5 s ROCOF at 1.0 s
[%] Nadir (FN) [Hz] FN [s] frequency[Hz] [%] [mHz/s] [mHz/s] [mHz/s]
0 49.6242 3.3444 49.8791 0 29.209 120.519 231.091

0.5 49.6353 3.3664 49.8803 0.5 28.956 118.583 226.323

2.0 49.6604 3.2595 49.8841 2 28.525 116.205 219.807

3.5 49.6925 2.9992 49.8903 3.5 28.073 114.318 213.482

49.7 C. Stabilizing contributions in SEBIR/VSM operating modes

49.69
Frequency Nadir [Hz]

The frequency behaviour is significantly modified if the

49.68 BESS contributes in stabilizing the network frequency.
49.67 Referring to Fig. 8, the black frequency trend is obtained when
49.66 the BESS is out of service (similarly to the red trend in Fig. 7).
49.65
49.64 VSM
Controlling the electronic power converter in the SEBIR
SEBIR mode (with converter rated power equal to 2.0% of the power
49.63
demand) results in a mitigation of the frequency oscillation
49.62
width (higher FN), even if an increased ROCOF appears in the
0.0 1.0 2.0 3.0 4.0
first time instants after the perturbating event. The delay in the
BESS rated power/Overall network demand [%] frequency measurement, and consequently in the ROCOF
Fig. 9: Influence of the BESS converter rated power in limiting the maximum computation, is the most likely cause.
width of the network frequency oscillation (i.e. the Frequency Nadir). If the VSM operating mode is applied to the same electronic
power converter, a similar benefit in terms of FN is observed. It
A. Traditional rotating genertors is important to note how the VSM mode is able to reduce the
ROCOF during the first part of the frequency perturbation, i.e.
The transmission network supplied exclusively by
problems introduced by the VRG diffusion are partially
traditional rotating units (synchronous generators) is considered
mitigated.
as base case. The perturbating event appears at t = 0 s and causes
the frequency perturbation depicted in Fig. 6. The first A more detailed comparison among SEBIR mode and VSM
oscillation width is 324 mHz, the FN is reached 4.1 s after the mode is possible by analising the results in terms of frequency
perturbating event and the frequency returns in the range 49.8- behavior, reported in Table I and Table II (referred to the SEBIR
50.2 Hz after about 7.8 s. Since both the secondary and the mode and the VSM mode, respectively). Several BESS rated
tertiary frequency regulation are not implemented in the model, sizes Sn have been considered. Both the tables report:
the frequency reaches a regime value of 49.89 Hz after 30 s.
• The FN, in [Hz], i.e. the maximum width of the
frequency oscillation after the perturing event;
B. Grid with high penetration of VRG
When a significant part of thermoelectric power stations is • The time required to reach the FN, in [s];
disconnected to host RES plants (60% of the overall network • The regime value of the frequency, in [Hz]. This result is
load in the considered case study), a sensibly different frequency evaluated as the system frequency 30 s after the
perturbation trend is obtained. The oscillation width increases as perturbation occurrence.
well as the ROCOF in the time instants immediately after the
perturbing event (red line in Fig. 7). In addition, it should be It is confirmed that increasing the rated power of the BESS
noted that the ROCOF increses due to both a larger frequency converter, the frequency oscillation is reduced up to about 0.06
oscillation and a reduced time required to reach the FN. A Hz, as graphically described in Fig. 9. The FN value is weekly
slightly lower regime frequency is obtained (49.87 Hz). influenced by the control mode applied to the BESS, but it is
reached faster in the case the SEBIR mode is applied. The VSM behavior immediately after the perturbing event (in
mode contributes in limiting the frequency deviation in regime particular in the first 0.5 s).
conditions.
Future works could be focused on modelling other primary
Another interesting parameter to be analysed is the ROCOF sources (e.g., supercapacitors or the electrostatic energy
at different time instants after the perturbing event. Results are intrinsically stored in HVDC systems) and their impact in terms
reported in Table III and Table IV for the SEBIR operating mode of frequency stability when the SEBIR approach or the VSM
and the VSM operating mode, respectively. It is confirmed that operating mode are considered. Furthermore, the VSM tuning
controlling the electronic power converter in the SEBIR mode procedure could be improved to further reduce the error in the
results in a higher ROCOF immediately after the perturbation, electronic converter dynamic response in comparison with a
and this trend is observed for all the considered BESS sizes. At traditional SG. More complex parameter estimation techniques
t = 0.5 s, the ROCOF with the SEBIR mode returns more or less could be implemented instead of using the tool currently
equal to the case in which the BESS is out of service. Benefits embedded in the MATLAB environment.
in terms of ROCOF reduction appear for t > 0.5 s. Differently,
controlling the electronic converter with the VSM approach REFERENCES
allows an interesting advantage in terms of ROCOF reduction
[1] F. Bignucolo, R. Caldon, M. Pettinà, F. Pasut, "Renewables contributing
immediately after the perturbing event. Moreover, the VSM to Primary Control Reserve: The role of Battery Energy Storage
mode is able to reduce the ROCOF in a more effective way in Systems," 2017 IEEE International Conference on Environment and
comparison with the SEBIR operating mode. Electrical Engineering and 2017 IEEE Industrial and Commercial Power
Systems Europe (EEEIC/I&CPS Europe), Milan, 2017, pp. 1-6.
[2] F. Bignucolo, R. Caldon, M. Coppo, F. Pasut, M. Pettinà, “Integration of
V. CONCLUSIONS Lithium-Ion Battery Storage Systems in Hydroelectric Plants for
In this paper, the role of synthetic inertia in improving the Supplying Primary Control Reserve,” Energies, Vol. 10, n. 98, 2017.
grid stability during perturbing events has been analysed. Both [3] ENTSO-E, “High Penetration of Power Electronic Interfaced Power
the SEBIR and VSM control modes have been modelled and Sources (HPoPEIPS),” 2017.
applied to a storage system. A detailed grid with high [4] T. Ackermann, T. Prevost, V. Vittal, A.J. Roscoe, J. Matevosyan, N.
Miller, “Paving the Way: A Future Without Inertia Is Closer Than You
penetration of VRG has been implemented as plausible case Think,” IEEE Power and Energy Magazine, Vol. 15, pp. 61-69, 2017.
study, carrying on dynamical simulations in the software
[5] M. Yu, A.J. Roscoe, A. Dyśko, C.D. Booth, R. Ierna, J. Zhu, H. Urdal,
environment DIgSILENT Powerfactory. “Effects of Swing Equation-Based Inertial Response (SEBIR) Control on
Penetration Limits of Non-Synchronous Generation in the GB Power
A novel procedure for optimally tuning the VSM control System,” International Conference on Renewable Power Generation
scheme parameters has been developed. During a frequency (RPG 2015), Beijing, 2015, pp. 1-6.
perturbation, the algorithm aims at obtaining a VSM dynamic [6] M. Yu, A.J. Roscoe, A. Dyśko, C.D. Booth, R. Ierna, J. Zhu, H. Urdal,
response similar to the active power behavior of a traditional SG “Instantaneous penetration level limits of nonsynchronous devices in the
with the same rated power. To make use of the MATLAB British power system,” IET Renewable Power Generation, vol. 11, no. 8,
Simulink environment and the embedded Parameter Estimation pp. 1211-1217, 2017.
tool, simplified models have been realized for both the main grid [7] H.P. Beck, R. Hesse, “Virtual Synchronous Machine,” 2007 9th
and the VSM control scheme. International Conference on Electrical Power Quality and Utilisation,
Barcelona, 2007, pp. 1-6.
Once the VSM parameters are computed, this control mode [8] Q.C. Zhong, G. Weiss, “Synchronverters: Inverters That Mimic
is applied to an electrochemical storage system interfaced with Synchronous Generators,” IEEE Transactions on Industrial Electronics,
Vol. 58, pp. 1259-1267, 2011.
the network by means of an electronic converter. The impact of
this additional unit in terms of frequency perturbation trend and [9] M. Yu, A.J. Roscoe, A. Dyśko, C.D. Booth, R. Ierna, J. Zhu, H. Urdal,
“Effects of VSM Convertor Control on Penetration Limits of Non-
ROCOF has been analysed, in comparison with the SEBIR Synchronous Generation in the GB Power System," 15th Wind
operating mode. The benefits of the VSM approach, which does Integration Workshop, Vienna, Austria, 2016.
not require a fast frequency measurement to compute the active [10] M. Yu, A.J. Roscoe, A. Dyśko, C.D. Booth, R. Ierna, J. Zhu, H. Urdal, “A
power to be injected in the grid, are confirmed by dynamics VSM (Virtual Synchronous Machine) Convertor Control Model Suitable
simulations. In particular, results show that: for RMS Studies for Resolving System Operator/Owner Challenges,"
15th Wind Integration Workshop, Vienna, Austria, 2016.
• Both the SEBIR operating mode and the VSM operating [11] F. Bignucolo, A. Raciti, R. Caldon, "Coordinating active and reactive
mode have a positive impact in limiting the depth of energy balances in islanded networks supported by renewables and
network frequency perturbations (i.e. increasing the FN); BESS," 3rd Renewable Power Generation Conference (RPG 2014),
Naples, 2014, pp. 1-6.
• Only the VSM operating mode is able to contain the [12] GSE, “Rapporto Statistico 2016: Solare Fotovoltaico,” 2016, available in
ROCOF immediately after the perturbing event. Since Italian at www.gse.it.
the SEBIR regulation requires the frequency [13] GSE, “Rapporto Statistico 2016: Energia da Fonti Rinnovabili in Italia,”
measurement to compute the ROCOF, delays in the 2016, available in Italian at www.gse.it.
measurement chain affect the electronic converter [14] ENTSO-E, “Operational Handbook - Policy 1,” 2004.
[15] ENTSO-E, “European Power System 2040