Delft University of Technology

Electrical Power Systems of the Future

EE4545

Assignment 4: Stability and control of grid side

converters

Group 4

Authors:
Diego Irazoqui 5838282
Fernando Canales 5672848
Michael Schweizer 5899508

May 23, 2023

Introduction
The current electrical energy systems is built on big fossil and nuclear power plants with synchronous machines
(SM). Those SM are by there natural physics able to provide several services which enhance the grids stability.
SM can provide inertia, variable reactive power, supply high amounts of active power during a failure and hence
can supply a high fault current and are able to synchronise themselves. Caused by the need of sustainable and
clean energy the energy production moves away from those big power plants to distributed renewable energy
sources like wind, solar, hydro and biomass. While hydro and biomass power plants are still using grid connected
SM the majority of power will be generated by wind turbines and solar farms which are connected via converters
to the grid. Most of those converters in the past where designed as grid following converters which relay on a
stiff grid and just supply a through phase-lock loop (PLL) controlled current signal to the grid. Furthermore
this means they cant provide hardly any of those services SM can.
Since the part of power generation which is connected by such converters will rise those converters need to
become able to provide the described services. Luckily the behaviour of these converters is mainly determined
by the implemented control principles. Conversely, this means that this behaviour can also meet the previously
described requirements of the electrical network through new control concepts. All new converters which are
able to do so are described as grid forming converters and are realised with voltage source converters (VSC).
Impelmeted priciples are: Virtual Synchronous Machine (VSM), Power Synchronisation Control (PSC or Sync),
Droop-Based Approaches (Droop) and Direct Voltage Control (DVC). [1, 2]
To get to know such control principles and experience how they react to faults with different controller settings
a modified version of the IEEE 9 bus network, depicted in Figure 1, is provided as a Power factory file. The
two Synchronous generators SG1 and SG3 where replaced by the wind farms WT_1 and WT_2 which can be
connected to the grid via different converters.
798.9
S1-A -785.6 -14.4 S2-A 14.5
-153.5 29.5 200.0 25.1 3.4 -92.6

S1
1.205 1.183 0.042 0.138 391.6
1.03

389.8
1.03
A 395.5
1.04
146.4
S2
154.3 146.2
-1497.1 0.0 300.0 398.2 800.0 -0.0 0.0 25.5 25.0 -65.0
215.1 0.0 50.0 -111.6 100.0 -325.0 0.0 107.9 10.0 -25.4
2.240 0.000 0.451 0.613 1.177 0.474 0.000 0.164 0.040 0.103
WT_trafo_S1

15.0
S1-C
Trf SG S1

Trf SG S2
S2-B
6.4
96.0

46.4
0.0

-394.8 Load A LoadStep -25.1 Load S2

0 Load S1 73.1 -170.1 0

C
0.590 1 0.260
381.8
1.00
1500.0 -0.0
392.9
146.3
B 65.1
228.1
1269.543
-0.0
0.000 1.03
150.2 800.0 -0.0
Compensation_A 30.2
2.554
FSC_BB_S1 SG S1_BB 100.0 -320.8 -405.2 -774.9 800.0
147.7 70.1 100.0
0.7 15.7
1.185 0.471
0.634 1.177 1.219
16.2
SG S2 BB
1.00 1.00 1.03
170.9 4.3 1 -0.0
29.0
S3-B
16.4
S3-C

65.1
Load C 30.2
1500.0 Load B 2.554
228.1 409.1 787.9
1269.543 Compensation_C -180.2 1.182
0.671 -23.3

SG
S3 384.9 ~
SG
~
1.01
154.5 -1497.0
153.5
300.0
50.0
0.0
-0.0
SG S2
WT_1 SG S1 2.257 0.456 0.000 47.8
WT_trafo_S3

Trf SG S3
96.8

0.0

Load S3 0

1500.0 -0.0
296.3 0.0
1279.368 0.000
0.7 1500.0 15.5
1.00 296.3 0.98

FSC_BB_S3
171.3 1279.368
WT_2 4.5

SG S3_BB
SG
~ SG S3

Figure 1: Overview of the used Test system based on the IEEE 9 bus network

1 Optimisation of Fast voltage control gain for a short circuit

In the fist part The influence of changing control parameters, filter parameters and Compensation Device (CD)
parameters to the response after during a short circuit (SC) shall be analysed. In order to do so a three-pole
SC is simulated at load bus A of Figure 1. The SC occurs after 1 second of simulation and is cleared 0.15s later.

1.1 a) influence of Ku on the voltage profile

To see the influence of the Fast voltage control gain K( u) its value is varied between 0.1 and 1 and the reaction
to the SC of the voltage magnitude profile and the output power and rotor speed of the only remaining SM
in the grid (SG 2) are analysed. The response of voltage, rotor speed of SG 2, frequencies and power of SG 2
are depicted for Ku = 0.2 and Ku = 0.9 in Figure 2 and Figure 3. In the top left of both figures the voltage
magnitude of all 3 generator buses and the load bus A are shown. The most severe voltage drops for all values
of Ku occurs at load bus A where the SC happen. With rising electrical distance (rising impedance) to the SC
the voltage drop becomes smaller. This behaviour is logical since the high fault current leads to a high voltage
drop along the lines. But if you look closer at the lengths of the different lines it gets obvious that the line from
A to S1 is with 29.9 Ω a lot longer than the line between A ans S2. But since the SG S2 supplies the major part
of the fault current the voltage drop difference is not as big as it would be for a grid with only SM. A higher
Ku has also an impact on the magnitude of the voltage drops. As the numbers and colors in Figure 4 show the
voltage drops keeps getting smaller with a higher value for Ku . The same counts for the estimated settling time

1
of the voltage magnitude. Since all the signal swing about the same amount after the clearance of the SC only
one time is estimated for all of them.

p.u. Hz
1 0.6

0.8 0.4

0.2
0.6
0
0.4
-0.2
0.2
0 2 4 6 8 10 -0.4
s

A: Voltage, Magnitude S2: Voltage, Magnitude 0 2 4 6 8 s 10

S1: Voltage, Magnitude S3: Voltage, Magnitude SG S2: Speed Deviation

200
51
150
Hz

50.5 100

50
50
0
0 2 4 6 8 s 10
0 2 4 6 8 s 10
S1: Electrical Frequency S3: Electrical Frequency
S2: Electrical Frequency SG S2: Electrical Frequency SG S2: Active Power [MW] SG S2: Reactive Power [Mvar]

Figure 2: Bus voltage, frequency rotor speed deviation and power of G2 at Ku = 0.2

p.u. Hz
1 0.6

0.4
0.8
0.2
0.6
0
0.4
-0.2
0.2
0 2 4 6 8 10 -0.4
s

A: Voltage, Magnitude S2: Voltage, Magnitude 0 2 4 6 8 s 10

S1: Voltage, Magnitude S3: Voltage, Magnitude SG S2: Speed Deviation

200
Hz
50.3 150
50.2
100
50.1

50 50

49.9
0
0 2 4 6 8 s 10
0 2 4 6 8 s 10
S1: Electrical Frequency S3: Electrical Frequency
S2: Electrical Frequency SG S2: Electrical Frequency SG S2: Active Power [MW] SG S2: Reactive Power [Mvar]

Figure 3: Bus voltage, frequency rotor speed deviation and power of G2 at Ku = 0.9

Ku 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

V mag A [pu] -0.8418 -0.8038 -0.7838 -0.7723 -0.7646 -0.7588 -0.7548 -0.7532 -0.7521
V mag s1 [pu] -0.6122 -0.5241 -0.4784 -0.4512 -0.433 -0.4196 -0.4099 -0.4077 -0.4052
V mag s2 [pu] -0.6985 -0.6431 -0.6148 -0.5989 -0.5886 -0.5801 -0.575 -0.5712 -0.5694
V mag s3 [pu] -0.5003 -0.4021 -0.3553 -0.3275 -0.3093 -0.2955 -0.2861 -0.2797 -0.2753
V mag se�ling it me [s] 4 4 3 3.3 3 2.9 2.8 2.8 2.7
S1 f [Hz] 50.31 50.93 50.16 50.13 50.1 50.1 50.09 50.08 50.08
S2 f [Hz] 50.32 50.76 50.2 50.17 50.15 50.13 50.12 50.11 50.11
S3 f [Hz] 50.29 50.88 50.17 50.14 50.12 50.11 50.1 50.09 50.08
SG S2 f [Hz] 50.47 50.59 50.37 50.33 50.32 50.3 50.3 50.29 50.28
G2P div + 74.93 87.93 74.63 75.93 76.93 76.93 78.13 78.43 78.93
G2P div - -61.194 -68.347 -52.95 -52.25 -39.01 -51.69 -51.65 -51.84 -52.69
G2Q div + 157.87 152.77 150.17 148.57 147.17 146.27 145.47 145.47 145.17
G2Q div - -49.32 -41.01 -42.16 -40.31 -39.489 -35.99 -36.485 -36.46 -38.93
G2 omega div [Hz] 0.5184 0.6197 0.4378 0.4179 0.4053 0.3968 0.3918 0.3856 0.384
G2 se�eling it me [s] 8 8 7 6.7 6 6 7 6 6

Figure 4: Overshoot undershoot and settling time over Ku colored in green for favorable and red for unfavorable

The influence of Ku on the behaviour of the remaining SM in the grid. While the transient behaviour of the
rotor speed deviation becomes significantly smaller with rising Ku , the power behaves different. In the displayed
signals for the reactive power it is clearly visible that it has an upwards spike when the SC and a downward
spike when it is cleared again. Both of those spikes have their biggest values at Ku = 0.2. There optimum value
of Ku for the active and reactive power is rather hard to find since the reactive power behaves differently to the

2
active power. While the rective power has its smallest transient swinging at Ku = 0.9 the active power has its
smallest upwards swing at Ku = 0.3 and its smallest downwards swing at Ku = 0.6.
Subsequently, it can be said that all signals except for the active power show a more favourable response to
the SC at large values of Ku . Looking at Figure 4 again Ku = 0.8 seems to be the most suitable value.

1.2 b) Influence of Compensation device and behaviour of frequency

To compensate reactive power of other grid devices a CD is connected to bus A and C respectively. Both of
those devices consume about 320 Mvar of reactive power during operation. To see there influence their set
point of 300 Mvar is reduced in steps of 50 Mvar. In order to avoid the influence of the imbalanced due to the
missing reactive power on the response to the SC the simulation is run for 20 s and only the last 10s are shown.
While the undisturbed voltage of the load bus A rises slightly compared with every 50 Mvar less at the DC
the generator bus voltages stay almost at the same level. Due to missing apparent power on the load side, the
undisturbed frequency also rises to a level of 50.3 Hz when the reactive power of the CD is zero. The response
to the SC at zero Mvar in the CD’s is displayed in Figure 5. Compared to the response with 300 Mvar power
of each CD only the voltages drop of bus A is higher. This means that the CD mainly keep the voltage of the
load buses at a higher and hence more preferable level during normal operation but don’t affect the reaction to
the fault significantly.

0.6
1
p.u. Hz
0.8 0.4

0.6
0.2

0.4
0

11 12 13 14 15 16 17 18 19 s

A: Voltage, Magnitude S2: Voltage, Magnitude 11 12 13 14 15 16 17 18 19 s

S1: Voltage, Magnitude S3: Voltage, Magnitude SG S2: Speed Deviation

250
50.5
Hz
200

50.4 150

50.3 100

50
50.2
0
11 12 13 14 15 16 17 18 19 s
11 12 13 14 15 16 17 18 19 s
S1: Electrical Frequency S3: Electrical Frequency
S2: Electrical Frequency SG S2: Electrical Frequency SG S2: Active Power [MW] SG S2: Reactive Power [Mvar]

Figure 5: Bus voltage, frequency rotor speed deviation and power of G2 at Ku = 0.9 with the CD’s at 0 Mvar.

Proceeding the analysis of the response to the SC the frequency singals shall be analysed. Those are depicted
in the lower left of Figure 2 and Figure 3. The first think which is visible is that the shown frequencies only
defer from each other from the start of the SC for about 3.5 s. The strongest reaction for all values of Ku except
Ku = 0.2 can be seen at the frequency of the SM SG S2. At Ku = 0.2 this is different because all production
buses have relatively high upward spikes in the beginning before they start to swing and settle back to the
original level of 50 Hz.
With the help of Figure 6 the behaviour of the frequency in the beginning shall be analysed further. With the
start of the SC all the frequencies rise but while the frequencies close to the SM keep rising during the whole
time the frequency at the Converters of the windturbines tend to settle at a certain level. Additionally the
frequency at bus S1 even rises more when the SC is cleared opposed to the behaviour of all other frequencies.
The plateau where the frequency tends to settle during the SC is depending on Ku is presented in Figure 4 and
is the highest at Ku = 0.2 and the lowest at Ku = 0.9. The difference to the values at Ku = 0.8 is very small
such that this is also a good set point for the frequency.

1.3 c) Influence of Damping filter

After analysing the influence of the CD’s the influence of the controller parameters shall be analysed further.
Therefore the damping and transient current control Washout filter proportional constant Kw is varied between
its default value of 0.03 and 0.3. while increasing this values the first differences are seen at Kw = 0.11 when
the spike visible at around 1.22 s in left plot of Figure 7 starts to form.
The mentioned spike keeps getting bigger until at Kw = 0.29 a additional downwards spike in the moment
of the SC forms. This second spike is significantly bigger an can be seen in the left graph of Figure 7. At
Kw = 0.3 the frequency gets a lot of even bigger spikes up and down directly after the SC. Subsequently that
means all values of Kw below 0.9 are fine for the frequency response to a SC. The other signals are not affected

3
 50.3
Hz
50.2

50.1

50

49.9

1 1.2 1.4 1.6 1.8 2 2.2 s

S1: Frequency S3: Frequency S2: Frequency SG S2: Frequency

Figure 6: Zoom on the frequency during the first 2.4 s after the short circuit

50.5
50.3
Hz
Hz Kw = 0.27
50
50.2
49.5
50.1 Kw = 0.29
49
50
48.5
49.9
48
1 1.2 1.4 s 1.6 1 1.2 1.4 s 1.6
S1: Frequency S2: Frequency S3: Frequency SG S2: Frequency

Figure 7: Frequency response to the SC with Ku = 0.8 and Kw = 0.27(left) and Kw = 0.29(right)

significantly by a change of Kw . Just during the SC the voltage swings more and the active power sees a small
oscillation directly after the SC.

2 Grid forming control principles

As already mentioned in the Table and further explained in [1, 2] there are different control principles to ensure
the desired network quality while using a grid forming con converter. To compare the performance of three
of those converters during different events, a new grid is used to carry out such tests. This grid Is almost the
same as the one shown in Figure 1 but has three additional generating devices at bus S1 and S3. Those devices
simulate wind farms which are connected with Virtual Synchronous Machine (VSM), Droop-Based Approaches
(Droop), or Power Synchronisation Control (Sync). This enables the user to switch between all these control
principles just by activating the converter type that should be analysed. Additionally to the three different
controllers, three different grid events are configured as well. First a three-phase short circuit at bus S1. Sec-
ond a load increase from 800 MW to 1050 MW at load bus A. Finally, the third event is represented by the
disconnection of a load at bus S3.

2.1 d) Reaction of the different controllers to a short circuit

As aforementioned, the first case of study considers a three phase short circuit at bus S1. Each VSC technology
was tested under such disturbances. In addition to this, different virtual impedance setting was implemented.
The simulations for each type of converter are shown in the following subsections. The corresponding tables
show the peak values of each component parameters, while the figures show a typical trend for the PEC1 and
the SG2 in each case. It is a representative trend for every simulation, as there is not much variation on the
curve, but the peak values of it.

It can be concluded that the network frequency, the settling time after the fault is cleared, and the peak cur-
rent in each converter are the key parameters to evaluate the performance under different virtual impedance’s
setting. The higher the virtual impedance is, the lower fault current peak and the lower settling time are
achieved. However, it also generates the higher peak in the electrical frequency oscillations.

Furthermore, implementing a higher X/R ratio improves the peak current limitation and settling time as
well, in detrimental of the settling time. Lowering the X/R ratio generates achieves the opposite: lowers the

4
oscillations of the electrical frequency, while allowing the fault current to reach higher values and the settling
time to be longer. To find the optimal setup, a deeper anaylsis should be done.

There is no significant difference in the behaviour of each type of converter under a short circuit condition.

2.1.1 VSM
The results for the VDM controller type are shown in Table 1, Figure 8 and Figure 9.

Table 1: Short circuit event - VSM results

Component Measurement Units 0.001+j0.001 0.006+j0.006 0.06+j0.06 0.006+j0.06

SG2 Active Power (P) MW 124.9 126.6 143.2 130
SG2 Reactive Power (Q) MVAr 176.9 176.5 176.3 182.2
SG2 Electrical Frequency (f) Hz 50.41 50.42 51.35 51.21
PEC1 Active Power (P) MW 1557 1559 1550 1524
PEC1 Reactive Power (Q) MVAr 1568 1553 1444 1282
PEC1 Current (I) kA 1480 1470 1422 1363
PEC3 Active Power (P) MW 1541 1541 1536 1518
PEC3 Reactive Power (Q) MVAr 1180 1182 1200 1117
PEC3 Current (I) kA 1477 1468 1421 1363
Grid Settling time (Ts) s 4.332 3.872 3.053 2.812

Current, I Active and reactive power, P and Q

200
kA 150
Mvar
1400 MW 150
100
1350
100

1300 50 50

1250
0 0
0 2 4 6 8 s 10
0 2 4 6 8 s 10
Dccontrol1: m:I:bus1 SynC1: m:I1:bus1
VSM1: Positive-Sequence Current, Magnitude Droop1: m:I1:bus1 SG S2: Positive-sequence, active power SG S2: Positive-sequence, reactive power

Active and reactive power, P and Q Electrical Frequency, f [Hz]

1500 51.5
Hz
1500
Mvar 51
MW
1000
50.5
1000
50
500

500 49.5
0 2 4 6 8 s 10
0 2 4 6 8 s 10
Dccontrol1: m:P:bus1 VSM1: Active Power SynC1: m:P:bus1 Droop1: m:P:bus1 S1: Electrical Frequency S3: Electrical Frequency
Dccontrol1: m:Q:bus1 VSM1: Reactive Power SynC1: m:Q:bus1 Droop1: m:Q:bus1 S2: Electrical Frequency SG S2: Electrical Frequency

Figure 8: PEC1 variables - Virtual impedance: Figure 9: SG2 variables - Virtual impedance: 0.06+j0.06
0.06+j0.06

2.1.2 VDM
The results for the VSM controller type are shown in Table 2, Figure 10 and Figure 10.

Table 2: Short circuit event - VDM results

Component Measurement Units 0.001+j0.001 0.006+j0.006 0.06+j0.06 0.006+j0.06

SG2 Active Power (P) MW 147 149.2 157.5 153.3
SG2 Reactive Power (Q) MVAr 180.6 177.2 180.3 182.2
SG2 Electrical Frequency (f) Hz 50.51 50.5 51.2 50.99
PEC1 Active Power (P) MW 1517 1513 1500 1507
PEC1 Reactive Power (Q) MVAr 1388 1387 1433 1341
PEC1 Current (I) kA 1425 1424 1414 1363
PEC3 Active Power (P) MW 1502 1500 1500 1500
PEC3 Reactive Power (Q) MVAr 1139 1139 1197 1111
PEC3 Current (I) kA 1425 1424 1414 1363
Grid Settiling time (Ts) s 3.897 3.654 3.266 3.150

5
 Current, I Active and reactive power, P and Q

kA 150 Mvar
1400
MW 150
100
1350
100

1300 50
50

1250 0
0
0 2 4 6 8 s 10
0 2 4 6 8 s 10
Dccontrol1: m:I:bus1 SynC1: m:I1:bus1
VSM1: m:I1:bus1 Droop1: Positive-Sequence Current, Magnitude SG S2: Positive-sequence, active power SG S2: Positive-sequence, reactive power

Active and reactive power, P and Q Electrical Frequency, f [Hz]

1500 1500 Hz
Mvar 50.4
MW
1000 50.2
1000
50
500

500 49.8
0 2 4 6 8 s 10
0 2 4 6 8 s 10
Dccontrol1: m:P:bus1 VSM1: m:P:bus1 SynC1: m:P:bus1 Droop1: Active Power S1: Electrical Frequency S3: Electrical Frequency
Dccontrol1: m:Q:bus1 VSM1: m:Q:bus1 SynC1: m:Q:bus1 Droop1: Reactive Power S2: Electrical Frequency SG S2: Electrical Frequency

Figure 10: PEC1 variables - Virtual impedance: Figure 11: SG2 variables - Virtual impedance: 0.06+j0.06
0.06+j0.06

2.1.3 Sync
The results for the Synchroconverter controller type are shown in Table 1, Figure 12 and Figure 13.

Table 3: Short circuit event - SYNC results

Component Measurement Units 0.001+j0.001 0.006+j0.006 0.06+j0.06 0.006+j0.06

SG2 Active Power (P) MW 123.3 125.4 141.3 125.7
SG2 Reactive Power (Q) MVAr 173.7 177.2 180.3 182.2
SG2 Electrical Frequency (f) Hz 50.43 50.53 51.36 51.21
PEC1 Active Power (P) MW 1568 1558 1500 1542
PEC1 Reactive Power (Q) MVAr 1468 1454 1433 1294
PEC1 Current (I) kA 1435 1423 1414 1363
PEC3 Active Power (P) MW 1543 1536 1500 1520
PEC3 Reactive Power (Q) MVAr 1148 1156 1201 1115
PEC3 Current (I) kA 1437 1423 1414 1363
Grid Settiling time (Ts) s 3.96 3.865 3.531 3.015

Current, I Active and reactive power, P and Q

200
1350 MW Mvar
150
100
kA
100
1300
50
50
1250
0 0
0 2 4 6 8 s 10
0 2 4 6 8 s 10
Dccontrol1: m:I:bus1 SynC1: Positive-Sequence Current, Magnitude
VSM1: m:I1:bus1 Droop1: m:I1:bus1 SG S2: Positive-sequence, active power SG S2: Positive-sequence, reactive power

Active and reactive power, P and Q Electrical Frequency, f [Hz]

Hz
1500 Mvar 51
MW 1000
50.5
1000
50
500

500 49.5
0 2 4 6 8 s 10
0 2 4 6 8 s 10
Dccontrol1: m:P:bus1 VSM1: m:P:bus1 SynC1: Active Power Droop1: m:P:bus1 S1: Electrical Frequency S3: Electrical Frequency
Dccontrol1: m:Q:bus1 VSM1: m:Q:bus1 SynC1: Reactive Power Droop1: m:Q:bus1 S2: Electrical Frequency SG S2: Electrical Frequency

Figure 12: PEC1 variables - Virtual impedance: Figure 13: SG2 variables and frequency - Virtual impedance:
0.006+j0.06 0.006+j0.06

2.2 e) Influence of the model control

To analyze the impact using proportional over-current limitation or a constant model control, the virtual
impedance was set at 0.006+j0.006 pu and simulations for each type of converter were done. The outcome for

6
each converter does not present significant differences. With respect to the control technique, under a short
circuit situation it can be said that using constant model control is a better solution, as it effectively limits the
maximum instant current. As can be seen in the figures below, using the proportional over-current limitation
leads to a instantaneous spike in the converter’s current, that could lead to the damage of the equipment.

Current, I Current, I

kA 1450
kA
1400
1400

1350 1350

1300 1300

1250 1250

0 2 4 6 8 s 10 0 2 4 6 8 s 10
Dccontrol1: m:I:bus1 SynC1: Positive-Sequence Current, Magnitude Dccontrol1: m:I:bus1 SynC1: Positive-Sequence Current, Magnitude
VSM1: m:I1:bus1 Droop1: m:I1:bus1 VSM1: m:I1:bus1 Droop1: m:I1:bus1

Active and reactive power, P and Q Active and reactive power, P and Q

1500 1500
1500 1500
Mvar Mvar
MW MW
1000 1000
1000 1000

500 500

500 500
0 2 4 6 8 s 10 0 2 4 6 8 s 10
Dccontrol1: m:P:bus1 VSM1: m:P:bus1 SynC1: Active Power Droop1: m:P:bus1 Dccontrol1: m:P:bus1 VSM1: m:P:bus1 SynC1: Active Power Droop1: m:P:bus1
Dccontrol1: m:Q:bus1 VSM1: m:Q:bus1 SynC1: Reactive Power Droop1: m:Q:bus1 Dccontrol1: m:Q:bus1 VSM1: m:Q:bus1 SynC1: Reactive Power Droop1: m:Q:bus1

Figure 14: PEC1 with proportional over-current Figure 15: PEC1 with constant over-current limitation
limitation

2.3 f ) Control parameter tuning Case 2 & 3

In this subsection the response of 2 different dynamic events are simulated and compared. These events are
Case 2: a load increase of 250 M W in bus A and Case 3: a load trip event in Load S3 (300 MW) connected to
bus S3 as shown in Figure 1.
Firstly, we compare 3 control schemes. These schemes are the following: Virtual Synchronous Machine
(VSM), Synchronverter (SynC), and Droop control. The resulting output of the simulation using the default
settings are shown in Figures 16 - 21. Clearly from the behaviour seen in the plots we can conclude that:
1. From Figure 16 and Figure 17: The effect of the event in Case 3 is higher. The Droop control scheme
minimises the frequency swinging amplitude and deviation after re-stabilizing in both cases.
2. From Figure 18 and Figure 19: In Case 2, the output power of both PECs (P1 and P3) in all control
strategy cases decreases 19. In Case 3, the output of both converters, except in the case of droop control,
decrease 18. The output power of both PECs seem to mirror each other except in the case of droop control
where they seem to stabilize very promptly and acting together. The droop control scheme has a lower
oscillating amplitude and power deviation.
3. From Figure 20 and Figure 21: The effects on the behaviour of the synchronous generator are bigger in the
case of a load decrease. There is no difference in the settling time of the generator power. The oscillating
amplitude in the power swinging of the generator seem to be equal in all control schemes, however there
is a lower deviation in the case of droop control.

Figure 16: Frequency swinging Load Increase

Figure 17: Frequency swinging Load Trip

7
 Figure 19: Power swinging Converters 1 & 3
Figure 18: Power swinging Converters 1 & 3

Figure 20: Power swinging Generator SG2

Figure 21: Power swinging Generator SG2

Given that in the previous plots we’ve concluded that the best dynamic performance is achieved using droop
control and the worst effects on the grid occur in Case 3, for the following part only plots regarding Droop
coefficient (mp) for Case 3 are shown in Figure 22 and Figure 23. However, all plots for the tuning of the other
two controls parameter, acceleration time (T a), and all cases are included in Appendix A.
Clearly from the behaviour seen in the plots we can conclude that:
1. The best performance in terms of frequency and power deviation/swinging were obtained using the lowest
values of mp. (Faster PECs operation)
2. In contrast to droop control, in VSM and SynC controls, upon lowering T a below a certain threshold the
frequency overshoots but apparently settles up faster.

Figure 22: Parameter tuning (mp) effect on fre- Figure 23: Parameter tuning (mp) effect on SG2 power swing-
quency swinging ing

(ALL PLOTS FROM THIS SECTION ARE ALSO ATTACHED IN FULL SIZE IN AP-
PENDIX FOR BETTER RESOLUTION)

8
References
Arghir, C., Jouini, T., & Dörfler, F. (2018, September). Grid-forming control for power converters based on
matching of synchronous machines. Automatica, 95 , 273–282. Retrieved from [Link]
[Link].2018.05.037 doi: 10.1016/[Link].2018.05.037
TU-Delft. (2023). Electrical power systems of the future - reader 4.

9
A Appendix
In this appendix all plots from the parameter tuning evaluated in subsection 2.3 are shown. These plots serve
only as graphical support to the conclusions reached in this subsection.
Figures concerning the effects of the event in Case 2 and parameter tuning for all control strategies in this
case are shown in Figures 24 - 35.

Figure 24: Frequency swinging Case 2

Figure 25: PECs power swinging Case 2

10
Figure 26: SG2 power swinging Case 2

11
 Figure 27: Parameter tuning (mp) effect on frequency swinging Case 2

Figure 28: Parameter tuning (mp) effect on PECs power swinging Case 2

12
Figure 29: Parameter tuning (mp) effect on SG2 power swinging Case 2

13
 Figure 30: Parameter tuning SynC (Ta) effect on frequency swinging Case 2

Figure 31: Parameter tuning SynC (Ta) effect on PECs power swinging Case 2

14
Figure 32: Parameter tuning SynC (Ta) effect on SG2 power swinging Case 2

15
 Figure 33: Parameter tuning VSM (Ta) effect on frequency swinging Case 2

Figure 34: Parameter tuning VSM (Ta) effect on PECs power swinging Case 2

16
Figure 35: Parameter tuning VSM (Ta) effect on SG2 power swinging Case 2

17
 Figures concerning the effects of the event in Case 3 and parameter tuning for all control strategies in this
case are shown in Figures 36 - 47.

Figure 36: Frequency swinging Case 3

Figure 37: PECs power swinging Case 3

18
Figure 38: SG2 power swinging Case 3

19
 Figure 39: Parameter tuning (mp) effect on frequency swinging Case 3

Figure 40: Parameter tuning (mp) effect on PECs power swinging Case 3

20
Figure 41: Parameter tuning (mp) effect on SG2 power swinging Case 3

21
 Figure 42: Parameter tuning SynC (Ta) effect on frequency swinging Case 3

Figure 43: Parameter tuning SynC (Ta) effect on PECs power swinging Case 3

22
Figure 44: Parameter tuning SynC (Ta) effect on SG2 power swinging Case 3

23
 Figure 45: Parameter tuning VSM (Ta) effect on frequency swinging Case 3

Figure 46: Parameter tuning VSM (Ta) effect on PECs power swinging Case 3

Figure 47: Parameter tuning VSM (Ta) effect on SG2 power swinging Case 3

24