See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/289011538

A modified CIGRE HVDC benchmark model for 60 Hz applications

Article · November 2014

DOI: 10.1109/NAPS.2014.6965411

CITATION READS
1 173

3 authors, including:

Hamidreza Zareipour
The University of Calgary
170 PUBLICATIONS   7,682 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

On Capacity Value of Energy-Limited Resources in Capacity Markets View project

Innovative Approaches to Advanced Optimization in Electricity Markets View project

All content following this page was uploaded by Hamidreza Zareipour on 23 August 2019.

The user has requested enhancement of the downloaded file.

 A Modified CIGRE HVDC Benchmark Model
for 60 Hz Applications
James Schwartz∗† , Ebrahim Rahimi‡† , and Hamid Zareipour†
∗ AltaLinkManagement Ltd., Calgary, AB, Canada
† Electrical & Computer Engineering Dept., Schulich School of Engineering, University of Calgary, Calgary, AB, Canada
‡ Alberta Electric System Operator, Calgary, AB, Canada

Abstract—This paper proposes a 60 Hz equivalent to the during disturbances, where both AC systems are operating
CIGRE benchmark model (CBM) for HVDC control stud- at 60 Hz instead of 50 Hz. The required modeling changes
ies. Component configuration and system impedances of the are discussed in detail in this paper. The PSCAD/EMTDC
PSCAD/EMTDC implementation of the CBM are modified to
arrive at the 60 Hz equivalent model. These parameter changes implementation of the CBM is used for all simulations.
are discussed and validation is given. Simulations are conducted Section II discusses the parameter changes required of the
in steady state and compared to that of the CBM. The response CBM to modify it for application at 60 Hz. The impedance
of the proposed equivalent to a current order step-response and changes made to the CBM are validated in Section III with the
AC faults are compared to that of the CBM. The performance aid of harmonic impedance analysis. Section IV compares both
comparisons match closely and this paper concludes that the
proposed 60 Hz equivalent is adequate for use when running models via steady state simulations. Section V further validates
simulations that require an HVDC benchmark model be con- the modified CBM by comparing the response of both models
nected to a representation of a 60 Hz AC network. to disturbances including current order step response, balanced
AC faults, and unbalanced AC faults. Section VI discusses
I. I NTRODUCTION the adequacy of the proposed 60 Hz equivalent as a means of
The CIGRE benchmark model (CBM) for line commutated representing the performance the original CBM.
converter (LCC) HVDC systems is well documented [1], [2]
II. CBM PARAMETER M ODIFICATIONS
and widely used in power system literature. The CBM is used
for a variety of studies including the following: 1) comparing The proposed parameter changes in the original CBM con-
simulation algorithms across different software platforms [3], sist of two types: 1) PSCAD component configuration values
[4], 2) studying interactions between AC and DC systems referring to fundamental frequency and 2) impedances in AC
including commutation failure [5]–[11], 3) investigating new and DC systems. The component configuration modifications
HVDC control strategies [12]–[16], and 4) investigating new are achieved by simply changing the fundamental frequency
HVDC converter topologies [17]. from 50 Hz to 60 Hz in the following components of the
The CBM represents an HVDC cable system which con- PSCAD implementation of the CBM:
nects two asynchronous, 50 Hz, AC systems. Therefore the • Three-phase voltage source models (2)

components of the AC and DC systems are designed for 50 • Three-phase, two-winding transformer models (4)

Hz operation. Using the CBM in the analysis of 60 Hz AC • Six-pulse thyristor bridge models (4)

systems, such as the large interconnections in North America, • Inverter minimum gamma over one cycle (1)

can be challenging because of the difference in fundamental • RMS multi-meters (4)

frequency. The impedance changes are more involved and are summa-
The motivation for a 60 Hz compatible version of the rized below:
CBM is simple, i.e., to have a convenient tool which can • AC source impedances; refer to Section II-A for details.
be connected to 60 Hz AC systems or components under • AC filter and shunt capacitor impedances; refer to Section
study which are modeled in electromagnetic transient (EMT) II-B for details.
software and then being able to simulate the integrated system. • Smoothing reactor and HVDC cable impedances; refer to
Having a 60 Hz equivalent CBM would avoid AC network II-C for details
conversion from 60 Hz to 50 Hz for the sole reason of being The converter transformer impedances are entered in per-
able to use an HVDC benchmark model, as experienced by unit and therefore do not require modification. The nominal
the authors of [6]. voltages of the AC sources and converter transformers are not
Reference [18] appears to use a 60 Hz CBM for the study of modified. Non-linear saturation of the converter transformers
harmonic impedance at inter-harmonics but does not provide is enabled for all simulations. No scaling of the saturation
details of the model changes. parameters is performed because the data is entered in per-unit
This paper proposes modeling changes to the CBM to arrive based on air core reactance, knee voltage, and magnetizing
at a 60 Hz equivalent. The goal with the proposed equivalent current. The following subsections discuss the AC and DC
is to have similar HVDC link performance in steady state and system impedance modifications in detail.
A. AC Source Impedances 500

450
original CBM (50 Hz)
The values of short-circuit ratio (SCR) and damping angles modified CBM (60 Hz)
specified in [1] should be maintained in the modified CBM so 400

that HVDC performance is similar to the original CBM. The 350

definition of SCR [19] is given by (1): 300

|Z+| [Ω]
250

SCM V A V2 2
Vrated 200
SCR = = rated = p (1) 150
Pd Pd |Zth | 2
Pd Rth + (ω1 Lth )2
100

where SCM V A is the short-circuit MVA of the AC system. 50

Pd is the rated HVDC link capacity. Zth is the Thevenin 0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49
impedance seen from the converter AC bus, evaluated at harmonic rank
fundamental frequency. Expressing Zth in rectangular form,
as shown in (1), permits manipulation of the key parameter Fig. 1. Harmonic impedance scan at rectifier AC bus
for this paper, i.e., the system fundamental frequency ω1 .
Let a constant k be defined to represent the change in 90

fundamental frequency required for the modified CBM: original CBM (50 Hz)
80
modified CBM (60 Hz)
70

fnew fold fnew fnew

ωnew = = = ωold = kωold (2) 60
2π 2π fold fold

|Z+| [Ω]
50

where all ω and f values are fundamental frequencies of their 40

respective system. For the development of the modified CBM, 30
k = ω1new /ω1old = 60/50 = 1.20. Therefore in order to
20
keep SCR constant, Lth in (1) must be replaced with Lth /k
to counteract the replacement of ω1old with ω1new = kω1old . 10

Therefore, all inductances in the source impedance are divided 0

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49
by k = 1.2, and resistances are not changed. It follows that harmonic rank
the damping angles will also be preserved.
Fig. 2. Harmonic impedance scan at inverter AC bus
B. AC Filters and Shunt Capacitors
The procedure used in section II-A is also used to modify III. VALIDATION OF I MPEDANCE C HANGES
the AC filter and shunt capacitor impedances as follows:
To validate the AC-side impedance changes proposed in
• Inductances divided by k = 1.2.
Section II, a harmonic impedance scan is performed at each
• Capacitances divided by k = 1.2.
converter AC bus in both the original CBM and the modified
• Resistances not changed.
CBM. For harmonic ranks from 0 to 50, the magnitude
Since a general capacitor impedance is given by of positive sequence impedance is recorded. The harmonic
Z = 1/(jωC), all capacitances in the modified CBM must be impedance comparison for rectifier and inverter is shown in
divided by k = 1.2 for the same reason as with the inductances Fig. 1 and Fig. 2, respectively.
in Section II-A. The AC harmonic impedance comparison shows agreement
between the original and modified CBM, which is expected.
C. Smoothing Reactor and HVDC Cable The harmonic impedance scans show a pronounced harmonic
A unique feature of the CBM is the DC-side resonance resonance at second harmonic which agrees with [1], [2]. A
at fundamental frequency and AC-side resonance at second Thevenin impedance calculation, not shown in this paper, of
harmonic [1], [2]. Reference [20] discusses how the HVDC the smoothing reactor in parallel with the cable capacitance
cable can be replaced by an equivalent overhead transmission shows a peak impedance at fundamental frequency in both
line model, in which this DC-side resonance at fundamental the original and modified CBM.
frequency is preserved. The parameters of the smoothing reac-
tor, cable resistance, and cable capacitance must be considered IV. S TEADY S TATE C OMPARISON
here because the fundamental frequency is being changed ac- The original and modified CBMs are simulated in
cording to (2). In similar fashion to the AC filters as discussed PSCAD/EMTDC and brought to steady state with a current
in Section II-B, the values of the smoothing inductance and order of 1.0 pu. Various AC and DC variables were measured
cable capacitance are each divided by k = 1.2 in the modified and recorded for comparison. The results are shown in Table
CBM. The value of the cable resistance is not changed. I, recorded at a simulation time of t = 2.0s.
 TABLE I (a) direct current measured at rectifier
S TEADY S TATE S IMULATION R ESULTS 2
CMR_50Hz [pu]
1.5 CMR_60Hz [pu]
Original CBM Modified CBM error [%]
V acrec [pu, rms] 1.013 1.014 0.10 1
P acrec [M W ] 1013.72 1015.28 0.15
Qacrec [M var] -46.38 -47.78 3.02 0.5
V acinv [kVrms ] 0.991 0.992 0.10
0
P acinv [M W ] 959.78 960.7 0.10 0 5 10 15 20 25
Qacinv [M var] 76.7 78.41 2.23
αrec [◦ ] 20.12 20.15 0.15 (b) rectifier firing angle
100
γinv [◦ ] 15.17 15.17 0
ALFA_REC_50Hz [°]
V drec [kV ] 497.83 498.28 0.09
ALFA_REC_60Hz [°]
Id[kA] 2 2 0
V dinv [kV ] 488.74 489.29 0.11 50

0
The reader should note that some of the variables in Table 0 5 10 15 20 25
I are not truly constant during steady state. For example, the number of cycles after current order increase request

firing angle, αrec , is observed in both the original CBM and

Fig. 3. Rectifier channels during current order step-response
modified CBM to vary with a range of 0.3◦ during what is con-
sidered steady state. This will affect the bus voltage and power
measurement when compared at a particular time step. Id and of the AC system voltage occurs (1/50) − (1/60) = 3.33ms
γinv have zero percent error because the rectifier and inverter faster than in the 50 Hz system.
are in constant current (CC) and constant extinction angle The selected output channels for plotting in the next sub-
(CEA) control modes, respectively. The proportional-integral sections are the following:
(PI) controllers ensure that the controlled value matches the
• rectifier measured direct-current (CMR), Idrec
setpoint value with zero error during steady state.
• rectifier firing angle (ALFA REC), αrec
The steady state results comparison in Table I show agree-
• inverter measured direct-current (CMI), Idinv
ment between both models with all errors being less than 3.0%,
• inverter extinction angle (GAMA INV), γinv
and the majority of errors being less than 0.2%. Therefore
comparison of EMT performance may proceed with reason- The reason for this is because in the CBM the rectifier is in CC
able confidence in the base electrical model. control mode while the inverter can be in either CEA or CC
control mode. Therefore, by analyzing the output channels for
V. E LECTROMAGNETIC T RANSIENT C OMPARISON Id, α, and γ we should be observing the most critical channels
Most HVDC projects are delivered to manufacturers who for EMT performance comparisons.
must procure the equipment according to a performance-
based technical specification [21]. In light of this, the EMT A. Direct Current Step-Response
comparisons given in the following subsections are compared The PSCAD/EMTDC implementation of the CBM does
from the perspective of recovery duration after disturbances not contain a lag/smoothing block between the current order
are cleared. from the user and the PI controller used for calculation of
The disturbances analyzed are a current order step response, the converter firing angles. Therefore, a sudden change in the
three-phase AC faults, and 1-phase AC faults. All of the current order can result in a sudden change in firing angle.
AC faults are applied on the source-side of the converter This can result in a sudden increase of direct current which,
transformers. For all disturbances, the original CBM (50 Hz) if large enough, will cause commutation failure (CF) of the
and the modified CBM (60 Hz) are simulated separately and inverter [5]. The smoothing reactors help limit the rate of rise
the relevant channels are recorded. MATLAB is then used to of Id, however not enough by which to completely eliminate
combine and plot related channels on a common horizontal CF.
axis of cycles elapsed after the disturbance is applied, i.e., each For this comparison, both simulations are brought to steady
plot in the following subsections shows the same variable as state with Id = 0.8pu. A current order increase of 0.2 pu
simulated by the original CBM and the modified CBM. was requested at the first zero-crossing of the rectifier phase
Output channels from the original CBM and modified CBM A voltage, after t = 1.4s. Thus, the channels from the both
are best compared on a common axis of number of cycles models start at the same point-on-wave. Output channels from
instead of a common time axis because the two systems have the rectifier and inverter are shown in Fig. 3 and Fig. 4,
different fundamental frequencies. If channels are compared respectively.
on a common time axis, the channels from the 60 Hz CBM The step change is requested at cycle zero. αrec and γinv are
will appear as an advanced version, in time, of those from the observed to decrease slightly to allow the initial increase of Id.
50 Hz CBM. The time-shift between these two channels is not The inverter CEA controller acts to reduce αinv in an effort to
constant, but increases with time according to the difference in mitigate the decrease of γinv . This is not successful because
fundamental frequencies, i.e., in the 60 Hz system each cycle the increasing Id results in an increasing inverter overlap
 (a) direct current measured at inverter (a) direct current measured at inverter
3 3
CMI_50Hz [pu] CMI_50Hz [pu]
2 CMI_60Hz [pu] 2 CMI_60Hz [pu]

1 1

0 0

−1 −1
0 5 10 15 20 25 0 5 10 15 20 25

(b) inverter extinction angle (b) inverter extinction angle

80 100
GAMA_INV_50Hz [°] GAMA_INV_50Hz [°]
60 GAMA_INV_60Hz [°] GAMA_INV_60Hz [°]

40 50

20

0 0
0 5 10 15 20 25 0 5 10 15 20 25
number of cycles after current order increase request number of cycles after fault application

Fig. 4. Inverter channels during current order step-response Fig. 6. Inverter channels during a 3-phase, 4-cycle, AC rectifier fault

(a) direct current measured at rectifier (a) direct current measured at rectifier
2 2
CMR_50Hz [pu]
1.5 CMR_60Hz [pu]
1

1
0
CMR_50Hz [pu] 0.5
CMR_60Hz [pu]
−1 0
0 5 10 15 20 25 0 5 10 15 20 25

(b) rectifier firing angle (b) rectifier firing angle

100 150
ALFA_REC_50Hz [°] ALFA_REC_50Hz [°]
ALFA_REC_60Hz [°] ALFA_REC_60Hz [°]
100
50

50

0 0
0 5 10 15 20 25 0 5 10 15 20 25
number of cycles after fault application number of cycles after fault application

Fig. 5. Rectifier channels during a 3-phase, 4-cycle, AC rectifier fault Fig. 7. Rectifier channels during a 3-phase, 4-cycle, AC inverter fault

angle, µinv , which results in γinv continuously reducing until shown by γinv reaching zero in Fig. 6b. After this occurs, the
CF occurs as shown in Fig. 4b at approximately 1 cycle. response of both models is the same but the channels from
The reader should recall that the sum of angles for a given the original CBM are advanced from those of the modified
converter must be 180◦ at any time [5], i.e., α+µ+γ = 180◦ . CBM. The reason why CF occurs faster in the original CBM
The important point to note from the comparisons in Fig. is attributed to the controller parameters not being changed in
3 and Fig. 4 is that the output channels are similar and they the modified CBM, i.e., time-constants and gains used in the
reach the same values for the new steady state. inverter CEA and CC controllers. Despite this transient shift
B. Rectifier Balanced AC Fault in the output channels for a few cycles, both models reach the
same Id values with the same duration, i.e., approximately 20
A 3-phase, 4-cycle, rectifier AC fault is applied to both
cycles after fault application, as shown by Fig. 5a and Fig. 6a.
models with Id = 1.0pu. The rectifier and inverter channels
for both models are shown in Fig. 5 and Fig. 6, respectively.
After fault clearing, HVDC recovery begins and CF occurs
C. Inverter Balanced AC Fault
at approximately 12.5 cycles after fault application: γinv is
observed to drop to zero when the inverter control switches A 3-phase, 4-cycle, inverter AC fault is applied to both
from CC to CEA mode. This CF resolves itself and the HVDC models with Id = 1.0pu. The rectifier and inverter channels
begins recovery to the pre-fault current order of 1.0 pu. for both models are shown in Fig. 7 and Fig. 8, respectively.
The comparison of the channels from both models match The inverter balanced AC fault causes the AC and DC
until approximately 12.5 cycles after fault application. CF voltage to drop to zero and the inverter experiences CF.
occurs faster in the original CBM than the modified CBM, as After the AC fault is cleared, both models start recovery and
 (a) direct current measured at inverter (a) direct current measured at inverter
3 1.5
CMI_50Hz [pu]
2 CMI_60Hz [pu]
1
1
0.5
0 CMI_50Hz [pu]
CMI_60Hz [pu]
−1 0
0 5 10 15 20 25 0 5 10 15 20 25

(b) inverter extinction angle (b) inverter extinction angle

60 60
GAMA_INV_50Hz [°]
40 40
GAMA_INV_60Hz [°]

20 20
GAMA_INV_50Hz [°]
GAMA_INV_60Hz [°]
0 0
0 5 10 15 20 25 0 5 10 15 20 25
number of cycles after fault application number of cycles after fault application

Fig. 8. Inverter channels during a 3-phase, 4-cycle, AC inverter fault Fig. 10. Inverter channels during a 1-phase, 4-cycle, AC rectifier fault

(a) direct current measured at rectifier

1.5 αrec reaching its minimum value of 5 degrees. The inverter
takes over control of the current, as supervised by the voltage-
1
dependent current-order limit (VDCOL). This can be observed
0.5
by the significant increase in γinv , and Idinv being controlled
CMR_50Hz [pu] to approximately 0.6 pu. After the fault is cleared, voltage
CMR_60Hz [pu]
0 at the rectifier AC bus returns to normal and the limitation
0 5 10 15 20 25
on αrec is released, i.e., the rectifier resumes control of the
(b) rectifier firing angle
40 current at the pre-fault current order of 1.0 pu.
The comparison of the channels from both models match
30
until approximately 6 cycles after fault application. γinv in the
20 original CBM drops very low at the 6 cycle mark whereas that
10 ALFA_REC_50Hz [°] in the modified CBM does not follow this excursion. Despite
ALFA_REC_60Hz [°]
0
this transient difference, all channel values in Fig. 9 and Fig.
0 5 10 15 20 25
number of cycles after fault application
10 match once again at 15 cycles after fault application. This
transient difference is attributed to no controller modifications
Fig. 9. Rectifier channels during a 1-phase, 4-cycle, AC rectifier fault being made in the modified CBM.
E. Inverter Unbalanced AC Fault
experience a second CF. This is observed by γinv reaching A phase-A, 4-cycle, inverter AC fault is applied to both
zero in both models for a second occurrence; refer to Fig. 8b. models with Id = 1.0pu. The rectifier and inverter channels
The comparison of the channels from both models match for both models are shown in Fig. 11 and Fig. 12, respectively.
until approximately 7 cycles after fault application. The second As with the balanced inverter fault discussed in Section V-C,
CF occurs faster in the original CBM than the modified CBM. the drop in inverter AC bus voltage at fault application results
As in Section V-B, this is attributed to the inverter controller in CF. After the fault is cleared, the inverter AC voltage returns
parameters not being changed in the modified CBM. to nominal and the HVDC link recovers from the fault without
All output channels shown in Fig. 7 and Fig. 8 match any subsequent CFs. This is unlike the case of a balanced AC
again at approximately 20 cycles after fault application. The fault at the inverter which results in a second CF during the
difference in the output channels between 7 cycles and 20 recovery, as discussed in V-C.
cycles after fault application is more pronounced than that The channels from both models are similar throughout the
observed in Section V-B. The αrec and γinv channels appear simulation. No transient difference is observed in Fig. 11 and
to have a larger difference than the Id channels. Fig. 12 as with the previous EMT simulations in this paper.
D. Rectifier Unbalanced AC Fault VI. D ISCUSSION
A phase-A, 4-cycle, rectifier AC fault is applied to both The steady state comparisons in Section IV indicate the
models with Id = 1.0pu. The rectifier and inverter channels modified CBM resembles that of the original within 3.0%
for both models are shown in Fig. 9 and Fig. 10, respectively. error. The EMT simulations in Section V show that the mod-
Because of the bolted fault placed on phase A at the rectifier ified CBM response to typical disturbances closely matches
AC bus, the rectifier loses control of the current as shown by that of the original when compared on a per-cycle basis. The
 (a) direct current measured at rectifier on a common per-cycle axis and shows agreement. The 60 Hz
2
CMR_50Hz [pu]
model proposed in this paper has performance that is similar
1.5 CMR_60Hz [pu] to the CBM and is adequate for simulations which require a
1 60 Hz version of the well-known benchmark model.
0.5 R EFERENCES
0
0 5 10 15 20 25
[1] M. Szechtman, T. Wess, and C. V. Thio, “First benchmark model for
HVDC control studies,” Electra, vol. 135, no. 4, pp. 56–73, April 1991.
(b) rectifier firing angle [2] ——, “A benchmark model for HVDC system studies,” in AC and DC
150 Power Transmission, 1991., International Conference on, Sep 1991, pp.
ALFA_REC_50Hz [°] 374–378.
100
ALFA_REC_60Hz [°] [3] M. Faruque, Y. Zhang, and V. Dinavahi, “Detailed modeling
of CIGRE HVDC benchmark system using PSCAD/EMTDC and
50 PSB/SIMULINK,” Power Delivery, IEEE Transactions on, vol. 21, no. 1,
pp. 378–387, Jan 2006.
[4] C. Liu, A. Bose, and P. Tian, “Modeling and analysis of HVDC converter
0
0 5 10 15 20 25 by three-phase dynamic phasor,” Power Delivery, IEEE Transactions on,
number of cycles after fault application vol. 29, no. 1, pp. 3–12, Feb 2014.
[5] C. V. Thio, J. Davies, and K. Kent, “Commutation failures in HVDC
Fig. 11. Rectifier channels during a 1-phase, 4-cycle, AC inverter fault transmission systems,” Power Delivery, IEEE Transactions on, vol. 11,
no. 2, pp. 946–957, Apr 1996.
[6] K. Meah and A. Sadrul Ula, “Simulation study of the CIGRE HVDC
(a) direct current measured at inverter benchmark model with the WSCC nine-bus power system network,” in
3 Power Systems Conference and Exposition, 2009. PSCE ’09. IEEE/PES,
CMI_50Hz [pu] March 2009, pp. 1–5.
2 CMI_60Hz [pu]
[7] E. Rahimi, A. Gole, J. Davies, I. Fernando, and K. Kent, “Commutation
1 failure analysis in multi-infeed HVDC systems,” Power Delivery, IEEE
Transactions on, vol. 26, no. 1, pp. 378–384, Jan 2011.
0 [8] X. Chen, A. Gole, M. Han, and C. Liu, “Influence of the MIIF index
on operation of multi-infeed HVDC systems,” in Electrical Power and
−1
0 5 10 15 20 25 Energy Conference (EPEC), 2011 IEEE, Oct 2011, pp. 216–221.
[9] C. Guo, Y. Zhang, A. Gole, and C. Zhao, “Analysis of dual-infeed
(b) inverter extinction angle HVDC with LCC-HVDC and VSC-HVDC,” Power Delivery, IEEE
80
GAMA_INV_50Hz [°]
Transactions on, vol. 27, no. 3, pp. 1529–1537, July 2012.
60 GAMA_INV_60Hz [°]
[10] Y. Liu and Z. Chen, “A flexible power control method of VSC-HVDC
link for the enhancement of effective short-circuit ratio in a hybrid multi-
40 infeed HVDC system,” Power Systems, IEEE Transactions on, vol. 28,
no. 2, pp. 1568–1581, May 2013.
20
[11] X. Chen, A. Gole, and M. Han, “Analysis of mixed inverter/rectifier
0 multi-infeed HVDC systems,” Power Delivery, IEEE Transactions on,
0 5 10 15 20 25 vol. 27, no. 3, pp. 1565–1573, July 2012.
number of cycles after fault application
[12] A. Daneshpooy, A. Gole, D. Chapman, and J. Davies, “Fuzzy logic
control for HVDC transmission,” Power Delivery, IEEE Transactions
Fig. 12. Inverter channels during a 1-phase, 4-cycle, AC inverter fault on, vol. 12, no. 4, pp. 1690–1697, Oct 1997.
[13] J. Bauman and M. Kazerani, “Commutation failure reduction in HVDC
systems using adaptive fuzzy logic controller,” Power Systems, IEEE
Transactions on, vol. 22, no. 4, pp. 1995–2002, Nov 2007.
comparisons match during the fault. Some transient differences [14] Y. Z. Sun, L. Peng, F. Ma, G. J. Li, and P. F. Lv, “Design a fuzzy
are observed after fault clearing. The same steady state is controller to minimize the effect of HVDC commutation failure on
reached by both models with the same duration, i.e., number power system,” Power Systems, IEEE Transactions on, vol. 23, no. 1,
pp. 100–107, Feb 2008.
of cycles. The performance of the modified CBM could be [15] B. Das, N. Watson, and Y. Liu, “Simulation study of conventional and
made closer to that of the original if the converter control time- hybrid HVDC rectifier based on CIGRE benchmark model using PLL-
constants and gains are studied and modified. Given the simple less synchronisation scheme,” in Power Engineering and Optimization
Conference (PEOCO), 2011 5th International, June 2011, pp. 312–317.
modifications indicated in Section II, the performance of the [16] D. Povh, P. Thepparat, and D. Westermann, “Further development of
modified CBM as reported in this paper meets the goal stated HVDC control,” in PowerTech, 2011 IEEE Trondheim, June 2011, pp.
in Section I, i.e., a convenient tool with similar performance 1–8.
[17] Y. Li, F. Liu, L. Luo, C. Rehtanz, and Y. Cao, “Enhancement of
to the original CBM. commutation reliability of an HVDC inverter by means of an inductive
filtering method,” Power Electronics, IEEE Transactions on, vol. 28,
VII. C ONCLUSION no. 11, pp. 4917–4929, Nov 2013.
This paper presents simple component and impedance [18] X. Jiang and A. Gole, “A frequency scanning method for the identifica-
tion of harmonic instabilities in HVDC systems,” Power Delivery, IEEE
changes to the PSCAD/EMTDC implementation of the CIGRE Transactions on, vol. 10, no. 4, pp. 1875–1881, Oct 1995.
Benchmark Model (CBM) which effectively convert it to a 60 [19] P. Kundur, Power Sytsem Stability and Control. McGraw-Hill, 1994.
Hz equivalent. The impedance modifications are validated with [20] M. Szechtman et al., “The CIGRE HVDC benchmark model - a new
proposal with revised parameters,” Electra, vol. 157, no. 2, pp. 61–66,
the aid of harmonic impedance scans. Steady state quantities December 1994.
between both models are shown to have similar values. The [21] J. Jardini, A. Gole et al., “TB 563: Modelling and simulation studies to
response of both models to a current order step-response and be performed during the lifecycle of HVDC systems,” CIGRE Working
Group B4.38, Tech. Rep., 2013.
AC faults are compared. This comparison is visually analyzed

View publication stats