You are on page 1of 8

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 27, NO.

3, JULY 2012

1501

Detailed and Averaged Models for a 401-Level MMCHVDC System


Jaime Peralta, Student Member, IEEE, Hani Saad, Student Member, IEEE, Sbastien Dennetire, Member, IEEE, Jean Mahseredjian, Senior Member, IEEE, and Samuel Nguefeu, Member, IEEE

AbstractVoltage-source-converter (VSC) technologies present a bright opportunity in a variety of elds within the power system industry. New modular multilevel converters (MMCs) are expected to supersede two- and three-level VSC-based technologies for HVDC applications due to their recognized advantages in terms of scalability, performance, and efciency. The computational burden introduced by detailed modeling of MMC-HVDC systems in electromagnetic-transients (EMT)-type programs complicates the study of transients especially when these systems are integrated into a large network. This paper presents a novel average-value model (AVM) for efcient and accurate representation of a detailed MMCHVDC system. It also develops a detailed 401-level MMC-HVDC model for validating the AVM and studies the performance of both models when integrated into a large 400-kV transmission system in Europe. The results show that the AVM is signicantly more efcient while maintaining its accuracy for the dynamic response of the overall system. Index TermsAverage-value model, electromagnetic-transients (EMT) programs, HVDC transmission, modular multilevel converter, voltage-source converter (VSC).

I. INTRODUCTION HE development of power-electronic and controllable devices in power systems is rapidly expanding the eld of applications for voltage-source-converter (VSC)-based HVDC technologies. This is mainly due to the recognized advantages of present VSCHVDC systems in comparison with traditional line-commutated converter (LCC)-based HVDC transmission [1]. HVDC technology based on VSC combines power converters with dc cables (or overhead lines) and lters to transmit power up to more than 1000 MW [2]. In addition to controlling power ow between two ac networks, VSC-HVDC systems can supply weak (low short-circuit ratio) and even passive networks [3]. Conventional two- and three-level VSC-HVDC systems present a faster dynamic response thanks to the pulsewidth-modulation (PWM) control in comparison with the fundamental switching frequency operation of traditional

Manuscript received August 09, 2011; revised December 20, 2011; accepted February 18, 2012. Date of publication April 05, 2012; date of current version June 20, 2012. Paper no. TPWRD-00671-2011. J. Peralta, H. Saad, and J. Mahseredjian are with Ecole Polytechnique de Montreal, Montreal QC H3C 3A7, Canada (e-mail: jaime.peralta@polymtl.ca; hani.saad@polymtl.ca; jean.mahseredjian@polymtl.ca). S. Dennetire and S. Nguefeu are with Rseau de Transport dElectricit (RTE), Paris-La Dfense, Paris 92068, France (e-mail: sebastien. dennetiere@rte-france.com; samuel.nguefeu@rte-france.com). Color versions of one or more of the gures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identier 10.1109/TPWRD.2012.2188911

HVDC [4]. VSC converters are based on insulated-gate bipolar transistors (IGBT) which, combined with PWM techniques, achieve high speed and independent control of active and reactive powers by maintaining stable voltage and frequency [5]. The ac output waveform in a VSC system is determined by the topology of the converter. The topologies suitable for dc power transmission are: two-level, multilevel neutral-point diode-clamped (NPC), and multilevel oating capacitor. Advantages and disadvantages of these three converter topologies can be found in [6]. Additional multilevel topologies developed for industrial medium-voltage (MV) applications are the cascaded multilevel and the hybrid multilevel converters [7]. Recent trends on multilevel converters for HVDC systems include modular multilevel converter (MMC) topology which connects two-level converter modules in cascade to achieve the desired ac voltage [8], [9]. MMC topologies enable using a smaller switching frequency to help reduce converter losses. In addition, lter requirements are eliminated by using a signicant number of levels per phase. Scalability to higher voltages is easily achieved and reliability is improved by increasing the number of submodules (SMs) [10]. Detailed models (DMs) for MMC-HVDC systems include the representation of thousands of semiconductor switches and must normally use small numerical integration timesteps to accurately represent fast switching events. The computational burden introduced by these models highlights the need to develop simplied models that provide similar behavior and dynamic response. These simplied models are known as mean- or average-value models (AVMs) and their purpose is to replicate the average response of switching devices, converters, and controls by using simplied functions and controlled sources [11], [12]. AVMs have been successfully developed for wind generation technologies [13], [14]. However, it is a new trend in high-power systems with few applications presented to date, including twoand three-level VSC-HVDC technologies [15][17]. A simplied model for MMC-HVDC systems was introduced in [18], but it does not provide details on the AVM and its DM version used for validation purposes. Reference [19] develops an efcient methodology for simulating MMC systems in EMT-type programs, but it does not model a detailed MMC, including a large number of levels and integrated into a large-scale transmission system. The main contributions of this paper are the development of a detailed MMC-HVDC representation including 401 levels (line-to-neutral voltage), the development of a novel AVM that efciently and accurately replicates the dynamic behavior of the

0885-8977/$31.00 2012 IEEE

1502

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 27, NO. 3, JULY 2012

DM, and the study of the dynamic performance of both models when integrated into an actual large system. This paper starts in Section II with a description of the detailed MMC-HVDC system. The AVM and its mathematic derivation are introduced in Section III. The comparison of the dynamic performance of the AVM against its DM version is presented in Section IV. II. DETAILED MMC-HVDC MODEL This section describes the detailed MMC-HVDC system developed for validating the proposed AVM. The system is based on the preliminary design of an MMC-HVDC system planned to interconnect the 400-kV systems of France and Spain by 2013. The interconnection will include two independent HVDC links, each one containing two MMC terminals with a rated transmission capacity of 1059 MVA each and a dc voltage of 320 kV. The VSC-HVDC technology, based on MMC, has been selected for this project due to dynamic performance and power-ow control requirements, and the low ac short-circuit ratio at the point of interconnection of the France-Spain system. It is expected that this project will be the most powerful MMC-HVDC link in operation by 2013. A. MMC Submodules The detailed model (DM) developed in this paper considers one 1059 MVA (1000 MW, 350 Mvar) VSC-HVDC link with two MMCs, including 800 SMs per phase (400 SMs per multivalve arm). Fig. 1 shows the MMC topology where each SM (Fig. 2(a)) contains a capacitor and two insulated-gate bipolar transistor (IGBT) switches (S1 and S2). At any instant during normal operation, only one of the two switches (S1 or S2) is ON. As a result, when the switch S1 is ON (S2 is OFF), the voltage and when the switch S2 is ON (S1 is OFF), of the th SM is helps control and the SM voltage is zero. The arm reactor balance circulating currents in the phase arms and limiting fault currents [10]. It has a value of 15% on the system impedance base. The IGBT switches are modeled when using an ideal controlled switch, two nonideal (series and antiparallel) diodes and a snubber circuit [Fig. 2(b)]. The nonideal diodes are modeled as nonlinear resistances using the classical diode function. The switch K1 in Fig. 2(a) is a high-speed bypass switch used to increase safety and reliability of the MMC in case of an SM failure [8]. K2 is a press-pack thyristor used to protect the MMC semiconductors and cables from high fault currents. DC fault currents will ow from the ac to dc side through the free-wheeling diodes which have low capacity to withstand high surge currents. K2 is red during the fault, allowing most of the current to ow through the thyristors and not through the diodes [8]. The detailed model developed in this paper includes 4800 ideal switches and 9600 nonideal diodes per MMC. This number excludes the press-pack thyristors used for protection which are triggered only during dc faults. The ideal switches K1 were not modeled. There are several advantages offered by the availability of the DM: increased accuracy of submodule models (nonlinear, switching losses), capability to account for special switching states, direct representation of unbalanced conditions in the converters, and establishment of a reference model.
Fig. 1. Detailed MMC topology.

Fig. 2. (a) MMC submodule. (b) IGBT valve.

The capacitor is selected with a value so that the ripple of the SM voltages is kept within a range of 10%. To achieve this, the energy stored in each SM should be in the range of 3040 kJ/MVA [20]. The SM capacitance is then estimated as follows: (1) where is the energy per megavolt-ampere (MVA) stored on each MMC, is the nominal capacity of the MMC (1059 MVA), is the number of SMs per multivalve arm (400), is 1.6 kV. For a stored energy of and the nominal value for 30 kJ/MVA, the resulting capacitor value is 10 mF. The complete setup for the new MMC-HVDC interconnection of France and Spain transmission grids is shown in Fig. 3. B. Converter Transformer and DC Cables The transformer is a 1059-MVA, 400/333-kV three-phase transformer with its secondary winding connected in delta to block the zero-sequence voltages generated by the MMC. The is 18%. converter transformer impedance

PERALTA et al.: DETAILED AND AVERAGED MODELS FOR A 401-LEVEL MMCHVDC SYSTEM

1503

Fig. 3. MMCHVDC transmission system.

The MMC-HVDC system includes two 70-km underground transmission cables. The 320-kV single-core cables are modeled using a frequency-dependent model [21]. C. MMC Control The MMC uses a vector control strategy that calculates a voltagetime area across the transformer/converter equivalent reactor which is required to change the current from -frame curthe present value to the reference value. The rent orders to the controller are calculated from preset and powers, and preset ac and dc voltages. The inner controller permits controlling the converter ac voltage that will be used to generate the modulated switching pattern. The ac-frame can be independently tive and reactive currents in the controlled via a proportional-integral (PI) control [5]. The reactive power control includes an ac voltage override block intended to maintain the ac voltage within acceptable limits. The active power control includes a dc voltage limiter that overrides the active power control in order to maintain the dc voltage within an acceptable range. Control limiters are also required to limit the current reference signals from going into the inner controller. D. MMC Modulation Technique Traditional modulation techniques proposed to date for MMCs include phase-disposition modulation (PD-PWM) [22], phase-shift modulation (PS-PWM) [23], space-vector modulation (SV-PWM) [9], and the improved selective harmonic elimination method (SHE) [8]. As the number of levels increases in MMCs, PWM and SHE techniques become cumbersome for EMT-type simulations. Therefore, more efcient staircase-type methods, such as the nearest level control (NLC) technique, have been proposed in [24] and [25]. The models developed in this paper use the NLC technique proposed in [24]. E. SM Capacitor Balancing Control The capacitor voltages ( ) at all SMs must be balanced and kept the same during normal operation. To achieve this, [Fig. 2(a)] must be monitored and the capacithe voltage tors switched ON and OFF based on a balancing control algorithm (BCA). The BCA proposed in [22] measures the capacitor voltages at any instant and sorts them before selecting the upper and lower SM to switch ON. The number of SMs is determined from two time-dependent switching functions and . The combination gives to the total number

of SMs per multivalve arm (400). Reference [23] proposes a different BCA where an SM is switched ON and OFF any time the reference signal crosses one of the triangular signals from the PS-PWM. An alternative BCA that can be directly implemented to each SM by means of a proportional-integral (PI) controller is proposed in [26], but adding individual PI controllers to each SM signicantly increases the simulation time as the number of levels increases. The detailed MMC model developed in this paper uses the BCA presented in [22]. To improve the efciency of the algorithm, the model includes a trigger control that acor change. This avoids tivates the BCA only when switching the SMs at each solution time point. F. Circulating Current Suppression Voltage unbalances between the arm phases of the MMC introduce circulating currents containing a second-harmonic component which not only distorts the arm currents, but also increases the ripple of SM voltages, thus impacting the rating of SM capacitors and switches. Circulating currents can be eliminated by using an active control over the modulated voltage [27], [28] or by adding a parallel capacitor (resonant lter) between the midpoints of the upper and lower arm inductances on each phase [20]. The proposed DM uses the circulating current suppression control (CCSC) proposed in [28]. Fig. 4 shows the effect of the control on the circulating arm current and on SM voltage ripples (showing several SMs) before and after CCSC removal at 1 s. It is observed that the CCSC eliminates the ac (2nd harmonic) component of the circulating current and decreases the voltage ripple from 30% to 8%. III. AVERAGE VALUE MMC-HVDC MODEL In the proposed AVM, the IGBTs are not explicitly modeled and the MMC behavior is represented using controlled voltage and current sources. The controlled sources include the harmonic content from the modulation control (or switching functions) in the ac voltage waveforms. Similar to the DM, the reference voltages are the output voltages obtained from the inner vector control where amplitude and phase are controlled independently. A. AC-Side Representation of the AVM The following equations can be derived from Fig. 1 for each : phase ,

(2)

1504

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 27, NO. 3, JULY 2012

Fig. 5. MMC AVM ac-side representation.

For arm currents in each phase (see also [22]) (9) (10) where the circulating current (ac second harmonic) is given by (11) and . Since the AVM assumes perfectly balanced voltages on all capacitors at any time, the second harmonic ( ) circulating currents are zero. By subtracting (7) from (8) (12)
Fig. 4. CCSC effect, CCSC is removed at 1 s. (a) Circulating current (in kiloamperes). b) SM voltage ripples (in kilovolts). (c) Converter dc voltage (in per unit). (d) Converter ac currents (in per unit), phase a is red, phase-b is blue, and phase c is green.

where (13) Replacing (9) and (13) into (7) gives

(3) (4) By using the same approach for lower arm equations

(14)

(5) (6) (7) (8) where is the total upper arm voltage on each phase in, the voltage (see Fig. 1) cluding the voltage of reactor is the total voltage of all upper SMs and is a function of the number of capacitors turned on as given by (4). In this equation, gives the state of each capacitor. Simthe binary function ilar denitions are applicable for the lower arm identied with the subscript .

(15) The ac-side representation of the AVM is presented in Fig. 5 and only phase- control blocs are shown for convenience. The is generated using the inner controller dereference voltage scribed in Section II-C. The AVM can support any modulation technique, but only the NLC method is used in this paper. Fig. 6 shows the voltage of the AVM for a 21-level MMC example using the NLC modulation method. It should be noted that the magnitudes and angles of voltages determine the ac currents going into (or out of) the MMC. The AVM proposed in this paper incorporates the harmonic content of the switching evens in the ac waveforms of currents and voltages which signicantly improves its accuracy over previously available AVMs using only the fundamental frequency representation.

PERALTA et al.: DETAILED AND AVERAGED MODELS FOR A 401-LEVEL MMCHVDC SYSTEM

1505

Fig. 6. MMC ac voltage v (in per unit), 21 levels, AVM.

Fig. 7. AVM dc-side representation of the MMC.

B. DC-Side Representation of the AVM The dc-side model is derived using the principle of power balance, meaning that the power on the ac side must be equal to the power on the dc side plus converter losses (16) (17) From the denition of the modulation index (18) the power balance of (17) can be rewritten as (19) The converter current loss function is dened as (20) with (21) where is the equivalent resistance of the MMC representing both switching and resistive losses, and is the equivalent dc current including converter losses. The dc current is then derived from (19) and (21) (22) Since converter losses are dependent on the modulation technique and switching frequency, they should be taken into account when estimating the equivalent resistance of the MMC. The value of is selected using the MMC losses from the DM which are close to 1% per converter. A parallel resistance on the dc side of the MMC does not allow modeling current-dependent losses as the latter will depend on the dc voltage only. Therefore, the loss function of (20) is selected to model the MMC losses. The dc side of the MMC AVM is then represented by two current-controlled current sources which depend on the current as shown in Fig. 7.

in Fig. 7 represents a capacitance equivalent The capacitor to the DM. It is derived using the energy conservation principle

(23) Assuming all SMs have the same voltage can be derived from (23) capacitor , the equivalent

(24) It is noted that the control parameters remain unchanged for the AVM in order to ensure fast and accurate dynamic response. During dc faults, all SMs in the detailed MMC are shorted by the thyristor K2 [see Fig. 2(a)] transforming the MMC into a sixpulse bridge diode converter. Therefore, the voltage-controlled sources in the AVM (see Fig. 5) are shorted and the dc capacitor is disconnected in order to mimic the effect of K2 in the DM. A series thyristor is added in the dc-side representation of the AVM (Fig. 7) to force the dc current ow direction from the ac to dc side. IV. DYNAMIC PERFORMANCE OF THE MMCHVDC MODELS The dynamic performance of the system shown in Fig. 3 is studied in this section. This 1059-MVA system is used to interconnect the 400-kV grids of France and Spain. The DM, including 401 levels, is compared against the AVM representation developed in Section III. The selected control strategy considers an active/reactive power-ow controller on the sending (rectier MMC-1) side and a dc voltage/reactive power controller on the receiving (inverter MMC-2) side. The active power ow is set to 1000 MW and the reactive power reference is initially set to zero (unity power factor) at both converters. All model developments and simulations are performed using EMTP-RV [29]. A. AC System Model and Initialization The ac model includes a dynamic equivalent system of the French and Spain grids. The total system model comprises 60 transmission lines of 400 kV and 23 synchronous generators, 12 of which are modeled in detail with their controls. Transmission lines are modeled using the constant parameter (CP) model. Such a complete setup enables simulating electromagnetic and

1506

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 27, NO. 3, JULY 2012

electromechanical (or lower frequency) transients using the same data set and software environment. Initialization of large networks, including HVDC systems, is a key issue for EMT-type solvers. To deal with this limitation in a large power system, the MMC-HVDC link is initially connected to ideal voltage sources and then synchronized to the ac grid when the reference power is reached. The voltage magnitudes and angles of the ideal sources are automatically calculated from the load-ow solution of the ac grid. In the load-ow solution, the MMCs are modeled as PQ constraints, but PV constraints can be alternatively used. The ac grid itself, including synchronous machine controls, is automatically initialized for the time-domain solution. In the following graphs, the solid red line is used for DM waveforms and the dashed blue line is used for AVM waveforms. The numerical integration timestep is 20 s for both models. B. Three-Phase Fault at MMC-2 1.2 s, a three-phase fault At the simulation time point of is applied for 200 ms on the high-voltage (HV) side of the 400/333-kV transformer of MMC-2. The system response, including dc and ac voltages (positive sequence) and active and reactive powers, is presented in Fig. 8. During the fault, the dc overvoltage is limited to 20% by the dc voltage controller. From Fig. 8(c), it is observed that the ac voltage on the rectier side is only slightly impacted by the fault on the inverter side of the system, which conrms the assumption of independent voltage control of each MMC. The active power is reduced to zero during the fault and it recovers 400 ms after a transient overload that is limited by the power control on the rectier side of the system. It is noticed that the reactive power is invariant during the fault and experiences a short transient right after the fault clearing to slowly recover after approximately 600 ms. The AVM provides very accurate results for the ac and dc waveforms during steadystate and transient operation. C. Reversal of Active Power Direction Power reversal is tested by changing the power reference at 1 s from 1000 MW to 500 MW using a 200-ms ramp reference. Fig. 9(a) shows that the system can suddenly reverse the power-ow direction by reversing the power setpoint at MMC-1. The controls allow switching the MMCs from rectier to inverter operation without signicantly altering the ac voltages [Fig. 9(b)]. The control strategy for each MMC remains unchanged during the power-ow reversal, but the ac voltage limiters in the outer controller are relaxed to allow the voltage to vary in a range of 10%. Since the reactive power reference remains unchanged (unity power factor), a voltage drop of 6% is observed at MMC-2 after power reversal. This voltage drop can be compensated by varying the reactive power setpoint in the outer controller. D. DC Pole-to-Pole Fault at MMC-2 DC faults are less frequent in underground cable layouts, but represent an important concern for MMC-HVDC systems. Even

Fig. 8. Three-phase fault at the MMC-2 transformer, HV side, DM: solid red line, AVM: dashed blue line. (a) DC voltage (in per unit), MMC-2. (b) AC voltage (in per unit), MMC-2. (c) AC voltage (in per unit), MMC-1. (d) Active power (in per unit), MMC-2. (e) Reactive power (in per unit), MMC-2.

though K2 in Fig. 2(a) is used to protect and bypass the SM and its diodes, dc faults impose stringent conditions for dc transmission cables. A permanent dc fault between the positive and neg1 s. The implemented ative poles of MMC-2 is applied at protection system assumes that the IBGTs are bypassed by the thyristor K2 [see Fig. 2(a)] 40 s after the fault is applied. The fault current measured on the dc side of MMC-1 is presented in Fig. 10. It can be observed that the fault current is limited to 6 p.u., which is a current that can be tolerated by the thyristors and cables for approximately up to 200 ms until the ac breaker

PERALTA et al.: DETAILED AND AVERAGED MODELS FOR A 401-LEVEL MMCHVDC SYSTEM

1507

Fig. 11. Phase-to-neutral voltages at MMC-1 (in per unit), phase-a: black, phase-b: blue, phase-c: green.

Fig. 9. Active power reversal test, DM: solid red line, AVM: dashed blue line. (a) Active power (in per unit), MMC-1. (b) AC voltage (in per unit), MMC-2.

The AVM approach is much faster than DM. It is also about 7 times faster than simplied equivalent circuit modeling methods, such as in [19]. EMTP-RV [29] uses a sparse matrix solver, and the presence of ideal switches and nonlinear functions in the DM requires refactorization at each switch position change. The mean value number of iterations for nonlinear diodes per timepoint is less than 2. F. Harmonic Content Analysis As the number of levels increase in MMCs, the ac voltage waveforms become almost perfectly sinusoidal functions. Fig. 11 shows the three phase-to-neutral voltages on the ac side of MMC-1 for the DM. The harmonic content is almost negligible for the 401-level DM with a total harmonic distortion (THD) value of 1.24% and 0.13% on the secondary (delta) and primary (wye) sides of the converter transformer, respectively. The AVM presents THD values of 0.28% and 0.12% for the secondary and primary transformer voltages, respectively. These values are below the threshold of 2% typically specied by international standards for maximum harmonic content and ltering requirements. V. CONCLUSION This paper presented the development of a detailed model for an MMC-HVDC system of 401 levels. It also presented a novel average-value model for the same system. Both models have been implemented into EMT-type software and tested in an existing and large transmission network for a new HVDC interconnection project between France and Spain. The detailed model includes all switching devices and, consequently, imposes a higher computational burden. The proposed average-value model has been shown to accurately replicate the dynamic performance of the detailed model. It is at least 370 times more efcient than the detailed model for the same simulation timestep. It is robust and easily scalable from smaller to larger systems expanding the eld of applications and types of studies that can be performed by EMT-type programs. The detailed model remains useful for simulating higher frequency transients, for studying detailed performance conditions in converters, and for calibrating the average value model. REFERENCES
[1] N. Flourentzou, V. G. Agelidis, and G. D. Demetriades, VSC-based HVDC power transmission systems: An overview, IEEE Trans. Power Electron., vol. 24, no. 3, pp. 592602, Mar. 2009.

Fig. 10. DC-fault current contribution from MMC-1 (in per unit), DM: solid red line, AVM: dashed blue line.

TABLE I COMPUTER TIMINGS, SIMULATION OF 3 S

opens. The AVM provides a good representation of fault currents in cables during dc faults, but contrary to ac-side faults, it is less accurate when compared to the DM. E. Simulation Performance The simulation performance test was done on a computer with a 2.66-GHz Intel Core i7-620M processor and 8 GB of RAM. The AVM performs signicantly better in terms of computer speed. A simulation of 3 s, using a timestep of 20 s, can be performed 370 times faster using the AVM without compromising the accuracy of the systems dynamic response. The computer times for the DM and AVM are presented in Table I. The DM remains sufciently precise when the timestep is increased up to 40 s. Since the switching valves are not modeled in the AVM, it can use timesteps slightly higher than 40 s without compromising accuracy and, therefore, allowing further computational speed gains.

1508

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 27, NO. 3, JULY 2012

[2] ABB, It is time to connect, technical description of HVDC Light Technology, Sweden, 2008. [3] C. Du, VSC-HVDC for industrial power systems, Ph.D. dissertation, Elect. Eng. Dept., Chalmers Univ. of Technology, Gteborg, Sweden, 2007. [4] J. Holtz, Pulse-width modulation for electronic power conversion, Proc. IEEE, vol. 82, no. 8, pp. 11941214, Aug. 1994. [5] A. Lindberg, PWM and control of two and three level high power voltage source converters, Licentiate degree, Royal Inst. Technol., Stockholm, Sweden, 1995. [6] B. R. Andersen, L. Xu, and K. T. G. Wong, Topologies for VSC transmission, in Proc. 7th Int. Conf. AC-DC Power Transm., London, U.K., Nov. 2001, pp. 298304. [7] J. Rodrguez, J. S. Lai, and F. Z. Peng, Multilevel inverters: A survey of topologies, controls, and applications, IEEE Trans. Ind. Electron., vol. 49, no. 5, pp. 724738, Aug. 2002. [8] G. Ding, G. Tang, Z. He, and M. Ding, New technologies of Voltage Source Converter (VSC) for HVDC transmission system based on VSC, in Proc. IEEE Power Eng. Soc. Gen. Meeting, Beijing, China, Jun. 2008, pp. 18. [9] A. Lesnicar and R. Marquardt, An innovative modular multilevel converter topology suitable for a wide power range, presented at the IEEE Power Tech. Conf., Bologna, Italy, Jun. 2003. [10] B. Gemmell, J. Dorn, D. Retzmann, and D. Soerangr, Prospects of multilevel VSC technologies for power transmission, in Proc. IEEE Transm. Distrib. Conf. Expo., Milpitas, CA, Apr. 2008, pp. 116. [11] S. R. Sanders, J. M. Noworolski, X. Z. Liu, and G. C. Verghese, Generalized averaging method for power conversion circuits, IEEE Trans. Power Electron., vol. 6, no. 2, pp. 251259, Apr. 1991. [12] S. Chiniforoosh, J. Jatskevich, A. Yazdani, V. Sood, V. Dinavahi, J. A. Martinez, and A. Ramirez, Denitions and applications of dynamic average models for analysis of power systems, IEEE Trans. Power Del., vol. 25, no. 4, pp. 26552669, Oct. 2010. [13] J. Morren, S. W. H. de Haan, P. Bauer, J. Pierik, and J. Bozelie, Comparison of complete and reduced models of a wind turbine with Doubly-fed Induction Generator, in Proc. 10th Eur. Conf. Power Electron. Appl., Toulouse, France, Sep. 2003, pp. 110. [14] J. G. Slootweg, H. Polinder, and W. L. Kling, Representing wind turbine electrical generating systems in fundamental frequency simulations, IEEE Trans. Energy Convers., vol. 18, no. 4, pp. 516524, Dec. 2003. [15] A. Yazdani and R. Iravani, Dynamic model and control of the NPCbased back-to-back HVDC systems, IEEE Trans. Power Del., vol. 21, no. 1, pp. 414424, Jan. 2006. [16] H. Ouquelle, L. A. Dessaint, and S. Casoria, An average value modelbased design of a deadbeat controller for VSC-HVDC transmission link, in Proc. IEEE Power Energy Soc. Gen. Meeting, Calgary, AB, Canada, Jul. 2009, pp. 16. [17] J. Peralta, S. Dennetiere, and J. Mahseredjian, Average-value models for the simulation of VSC-HVDC transmission systems, presented at the CIGRE Int. Symp., Bologna, Italy, Sep. 2011. [18] S. P. Teeuwsen, Simplied dynamic model of a voltage-sourced converter with modular multilevel converter design, in Proc. IEEE Power Eng. Soc., Power Syst. Conf. Expo., Seattle, WA, Mar. 2009, pp. 16. [19] U. N. Gnanarathna, A. M. Gole, and R. P. Jayasinghe, Efcient modeling of modular multilevel HVDC converters (MMC) on electromagnetic transient simulation programs, IEEE Trans. Power Del., vol. 26, no. 1, pp. 316324, Jan. 2011. [20] B. Jacobson, P. Karlsson, G. Asplund, L. Harnefors, and T. Jonsson, VSC-HVDC transmission with cascaded two-level converters, in Proc. CIGRE Conf., Paris, France, Aug. 2010, pp. B4110. [21] A. Morched, B. Gustavsen, and M. Tartibi, A universal model for accurate calculation of electromagnetic transients on overhead lines and underground cables, IEEE Trans. Power Del., vol. 14, no. 3, pp. 10321038, Jul. 1999. [22] M. Saeedifard and R. Iravani, Dynamic performance of a modular multilevel back-to-back HVDC system, IEEE Trans. Power Del., vol. 25, no. 4, pp. 29032912, Oct. 2010. [23] E. Solas, G. Abad, J. A. Barrena, A. Carcar, and S. Aurtenetxea, Modulation of modular multilevel converter for HVDC application, in Proc. 14th Int. Power Electron. Mot. Control Conf., Mondragon, Spain, Sep. 2010, pp. 8489.

[24] Q. Tu and Z. Xu, Impact of sampling frequency on harmonic distortion for modular multilevel converter, IEEE Trans. Power Del., vol. 26, no. 1, pp. 298306, Jan. 2011. [25] K. Li and C. Zhao, New technologies of modular multilevel converter for VSC-HVDC application, presented at the Asia-Pacic Power and Energy Eng. Conf., Baoding, China, Mar. 2010. [26] M. Hagiwara and H. Akagi, PWM control and experiment of modular multilevel converter, in Proc. IEEE Power Electron. Specialists Conf., Tokyo, Japan, Jun. 2008, pp. 154161. [27] A. Antonopoulos, L. Angquist, and H. P. Nee, On dynamics and voltage control of the modular multilevel converter, presented at the 13th Eur. Conf. Power Electron. Appl., Barcelona, Spain, Oct. 2009. [28] Q. Tu, Z. Xu, and L. Xu, Reduced switching-frequency modulation and circulating current suppression for modular multilevel converters, IEEE Trans. Power Del., vol. 23, no. 3, pp. 20092017, Jul. 2011. [29] J. Mahseredjian, S. Dennetire, L. Dub, B. Khodabakhchian, and L. Grin-Lajoie, On a new approach for the simulation of transients in power systems, Elect. Power Syst. Res., vol. 77, no. 11, pp. 15141520, Sep. 2007.

Jaime Peralta (S07) was born in Santiago, Chile, in 1969. He received the B.Sc. degree in electrical engineering from the University of Chile, Santiago, in 1994 and the M.A.Sc. degree in electrical engineering from cole Polytechnique de Montral, Montral, QC, Canada, in 2007, where he is currently pursuing the Ph.D. degree. Currently, he is with Siemens Power Technologies International, Vancouver, BC, Canada.

Hani Saad (S07) received the B.Sc. degree in electrical engineering from cole Polytechnique de Montral, Montral, QC, Canada, in 2007, where he is currently pursuing the Ph.D. degree in electrical engineering. From 2008 to 2010, he was with TechImp Spa and in the Laboratory of Materials Engineering and High Voltages (LIMAT), University of Bologna, Bologna, Italy. His research activities are related to partial-discharge diagnostics in power systems.

Sbastien Dennetire (M04) received the M.A.Sc. degree in electrical engineering from cole Polytechnique de Montral, Montral, QC, Canada, in 2003. From 2002 to 2004, he was with IREQ (Hydro-Qubec), working on research-and-development activities related to the simulation and analysis of electromagnetic transients. From 2004 to 2009, he was with EDF France in the eld of insulation coordination and power system simulations. In 2010, he joined the French transmission system operator Rseau de Transport dElectricit (RTE) in the eld of power system simulation.

Jean Mahseredjian (M87SM08) received the M.A.Sc. and Ph.D. degrees in electrical engineering from cole Polytechnique de Montral, Montral, QC, Canada, in 1985 and 1991, respectively. From 1987 to 2004, he was with IREQ (Hydro-Qubec), Varennes, QC, Canada, working on research-and-development activities related to the simulation and analysis of electromagnetic transients. In 2004, he joined the faculty of electrical engineering at cole Polytechnique de Montral.

Samuel Nguefeu (M04) received the M.A.Sc. and Ph.D. degrees in electrical engineering from Universit Pierre et Marie (Paris VI), Paris, France, in 1991 and 1993, respectively. He was a Consultant for two years before joining Thomson, France, in 1996. From 1999 to 2005, he was with EDF R&D, France, in power systems and power electronics. In 2005, he joined the French transmission system operator Rseau de Transport dElectricit (RTE), where he is currently involved in exible ac transmission systems and HVDC projects.

You might also like