THE modular multilevel converter (MMC), first presented

Transcription

1 IECON215-Yokohama November 9-12, 215 Performance of the Modular Multilevel Converter With Redundant Submodules Noman Ahmed, Lennart Ängquist, Antonios Antonopoulos, Lennart Harnefors, Staffan Norrga, Hans-Peter Nee. Department of Electrical Energy Conversion KTH Royal Institute of Technology, Stockholm, Sweden. Abstract The modular multilevel converter (MMC) is the state-of-the-art voltage-source converter (VSC) topology used for various power-conversion applications. In the MMC, submodule failures can occur due to various reasons. Therefore, additional submodules called the redundant submodules are included in the arms of the MMC to fulfill the fault-safe operation requirement. has not been widely covered in the published literature. This paper investigates the performance of the MMC with redundant submodules in the arms. Two different control strategies are used and compared for integrating redundant submodules. The response of the MMC to a submodule failure for the two strategies is also studied. Moreover, the operation of the MMC with redundant submodules is validated experimentally using the converter prototype in the laboratory. Index Terms High-voltage dc (HVDC) transmission, modular multilevel converter (MMC), redundant submodules, voltagesource converter (VSC). I. INTRODUCTION THE modular multilevel converter (MMC), first presented in [1], has arguably become the most promising converter topology for various medium and high power-conversion applications [2], such as high-voltage dc (HVDC) transmission [3], [4], variable-speed drives [5], electric railway supplies [6], propulsion system of electric ships [7], and grid connection of energy storage systems [8]. Compared with conventional twoand three-level voltage-source converters (VSCs), the MMC allows reduction of the switching frequency down towards the fundamental frequency, whereas the harmonic content of the output voltage is still kept low owing to a large number of levels. Other benefits are high scalability due to modular design and no requirement for a common dc-link capacitor [9], [1]. The research on the operation and control of the MMC includes various modulation techniques such as carrier-based pulse-width modulation (PWM) [11], [12], selective harmonic elimination PWM (SHE-PWM) [13], and nearest level control (NLC) modulation [14]. The balancing control of the submodule capacitor voltages within the arm is a key issue. Several alternative voltage-balancing methods were presented in [15], [16]. Other important research aspects include control of circulating current and capacitor voltage ripple [17], [18] In order to understand the steady-state and dynamic behavior of the MMC, several mathematical models have been developed [19] [21]. Simulation models play a very important role in investing the performance of the MMC for control and protection purposes. Equivalent simulation models of MMCs were presented in [22] [24]. In the MMC, submodule failures can occur due to various reasons [4]. Fault-tolerant operation of the MMC requires that in case of a submodule failure, the converter must continue its operation without disturbing its performance. To accomplish this, additional submodules called redundant submodules, are integrated in the arms of the MMC [1]. This results in an increase in number of submodules in the arm. When a submodule fails, the faulty submodule is shorted out and the converter continues its operation, without any interruption. Despite its importance, the performance of the MMC with redundant submodules and its response to submodule failures has not been widely covered in the existing literature. To date, a few studies on the behavior of the MMC with redundant submodules have been published [25], [26]. Reference [27] proposes a method for non-interruptible energy transfer of the MMC, in case of a submodule failure. This paper investigates the performance of the MMC when redundant submodules are included in the arms. Two different strategies for integrating the redundant submodules in the arm are implemented and compared. Unlike [25] and [26], which describe most steady-state properties of the approach to use redundant submodules, the present work provides a detailed analysis of the transients occurring due to submodule failure. A computationally efficient detailed equivalent model (DEM) model [24] of the MMC developed in PSCAD is used for the investigation. The operation of the MMC with redundant submodules is also validated experimentally using a threephase 1-kVA MMC prototype in the laboratory. II. MMC MODELING The circuit schematic of a three-phase MMC is outlined in Fig. 1. It consists of six arms, each constituted by N cascaded submodules. Each submodule, representing a controllable voltage source, consists of a half-bridge with a dc storage capacitor. The number of submodules in the arm can be adjusted based upon the submodule voltage and the dclink voltage. The arm inductance L is needed to limit fault and parasitic currents [4]. The arm resistance R models the resistive losses in the arm. A. DEM The DEM proposed and validated in [24] is used to study the performance of MMC with redundant submodules. The DEM as shown in Fig. 2, is based on the equivalent branch model of an arm of the MMC implemented in PSCAD. The source voltage represents the sum of the capacitor voltages of /15/$ IEEE 3922

2 TABLE I MMC PARAMETERS Parameter Rated power Rated output voltage (L-L) Rated dc-link voltage (V d ) Submodule capacitance (C) Arm inductance (L) Arm resistance (R) Carrier frequency (f carr) Value 88 MVA 4 kv 64 kv 525 μh 95.5 mh 1 Ω 5kHz B. DEM Without Redunndant Submodules To study the effect of redundant submodules on the performance of the MMC, a 21-level MMC with N = 2 submodules per arm (without redundant submodules) is simulated as a reference case. The voltage of each submodule capacitor is set to Fig. 1. Equivalent circuit diagram of a three-phase half-bridge MMC. v c = V d N, (1) where V d is the dc-link voltage. The average value of the submodule capacitor voltages should be maintained to this value. In order to limit the submodule capacitor voltage, the capacitance of the submodule capacitors is selected such that it corresponds to a stored energy of 39.5 kj per MVA of the converter rating [4]. The arm capacitance is given by C arm = C N. (2) Table I shows the parameters used in the DEM for the simulated MMC. Fig. 2. Equivalent branch model of an arm of the half-bridge MMC. the inserted submodules in the arm during the simulation timestep. The submodule capacitor voltages are kept in a vector which is updated depending on the branch current and the switching vector signal. Describing the arm of the MMC in this way allows its computationally efficient modeling for any number of submodules in the arm. The arm model is also capable of representing the blocked-mode of the half-bridge. The internal control of the MMC is based on the openloop approach using estimation of stored energy [28]. The open-loop control method provides fast dynamic performance and is less complicated to implement than most other highperformance control methods proposed for the MMC [12]. The open-loop controller generates six insertion indices governing each arm of the MMC. The insertion indices are utilized by the modulator to produce switching signals for the submodules using PWM. C. DEM With Redundant submodules To study the operation of the MMC with redundant submodules, two additional submodules are considered in each arm of the reference case MMC. The arm of the MMC thus comprises a total number of submodules N+R = 22 (1% redundancy). At any instant, if F is the number of faulty submodules in the arm, the total number of active submodules in the arm can be obtained as N+R-F, wheref varies from zero to R. The equivalent circuit of the MMC with redundant submodules is showninfig.3. The parameters used for the reference case are also adapted to the MMC configuration with redundant submodules. However, compared to (2) the arm capacitance is changed and can now be obtained using C C arm = N + R F. (3) Hence, with redundant submodules in the arm, the arm capacitance decreases. It also varies with the number of faulty submodules. III. CONTROL STRATEGIES A. Constant Total Capacitor Voltage Strategy (Strategy 1) In this strategy the average total capacitor voltage (sum of the capacitor voltages of the healthy submodules) across the 3923

3 based on the instantaneous value of the capacitor voltages and direction of the arm current as discussed in [21]. Considering above, the operation of the MMC during each time-step for this strategy can be summarized as The total voltage of the healthy capacitors in each arm remains unchanged. Actual number of healthy submodules N+R-F is detected. The open-loop control system in each arm will be given a reference for the total energy of the healthy capacitors which changes depending on the number because the resulting arm capacitance will change accordingly. The open-loop control for each arm calculates the desired insertion index. The modulator for each arm obtains the desired insertion index as well as the information of the number of healthy submodules. The modulator for each arm accordingly inserts n arm times N+R-F submodules using a suitable sorting algorithm. Fig. 3. Equivalent circuit diagram of athree-phase half-bridge MMC with redundant submodules. arm is kept constant. In this study this value is selected to be equal to the dc-link voltage. Hence, the individual submodule capacitor voltage is given by vc Σ v c = N + R F = V d N + R F. (4) This indicates that the individual submodule capacitor voltages are lower compared to the reference case. However, submodule capacitor voltages increase with the number of faulty submodules in the arm as the total capacitor voltage in the arm remains constant. This control strategy produces a N+R-F+1-level arm voltage, which implies a higher number of levels than the MMC configuration without redundant submodules. Moreover, the number of levels in the arm voltage also varies with the number of faulty submodules. The average energy stored in each arm of the converter will be E arm = 1 V (N + R F )C[ 2 N + R F ]2. (5) This energy is lower, compared to the energy stored in the arm of the MMC without redundant submodules. However, this energy will increase with the number of faulty submodules. The insertion index n arm for each arm is given by n arm = v arm vc Σ, (6) and provided to the modulator which calculates the number of inserted submodules as the fraction of N+R-F that are instantly available. The individuals among the N+R-F submodules, that will be inserted, are determined by the selection mechanism d B. Constant Submodule Capacitor Voltages Strategy (Strategy 2) In this strategy the individual submodule capacitor voltages v c in the arm are kept constant [26] to the value given by (1). The average total capacitor voltage (sum of the capacitor voltages of the healthy submodules) across the arm is thus given by vc Σ =(N + R F ) V d N. (7) Clearly, the average total capacitor voltage is greater than the dc-link voltage. However, it decreases with the number of faulty submodules in the arm. The average energy stored in each arm of the MMC in this case will be E arm = 1 2 (N + R F )C(V d N )2. (8) It is obvious that the stored energy in the arm is increased, compared to the reference case. Though, this energy decreases with the increase in number of faulty submodules in the arm. The generation of the insertion indices and the selection mechanism of submodules in the arm work in the same manner as in the case of the constant total capacitor voltage scheme. Fig. 4 shows the insertion indices obtained using (6) for the two strategies. It can be observed from the figure that strategy 2 generates a lower insertion index. Hence, among the N+R-F submodules in the arm, the submodules are selected in such a way to generate an N+1-level arm voltage. Thus, the number of levels in the arm voltage remains constant irrespective of the number of redundant and faulty submodules in the arm. Also, all submodules in the arm are treated equally with reference to the balancing control of the capacitor voltage. The operation of the MMC during each simulation time-step for this strategy can be summarized as The reference voltage for each submodule remains unchanged. Actual number of healthy submodules N+R-F is detected. 3924

4 1.1 Strategy 1 Strategy 2 (a).8 v arm,u v arm,l n arm Fig. 4. Comparison of insertion indices for the two strategies. 1 i u i l i diff Current (ka).5.5 Fig. 5. Schematic diagram of the simulated test circuit in PSCAD. Depending on the number of healthy submodules in the arm, the total voltage of the healthy submodule capacitors changes as vc Σ =(N + R F )v c. The open-loop control system in each arm will be given a reference for the total energy of the healthy capacitors which changes depending on the number because the resulting arm capacitance will change accordingly. The open-loop control for each arm calculates the desired insertion index. The modulator for each arm obtains the desired insertion index as well as the information of the number of healthy submodules. The modulator for each arm accordingly inserts n arm N+R-F times submodules using a suitable sorting algorithm. IV. STUDY RESULTS The two strategies discussed in the previous sections are compared to the reference case using a test circuit where an 88-MVA converter is fed from a stiff dc-link and provides power to the ac-network. The ac-network has a line-to-line voltage of 4-kV rms, and is connected to the converter through a 9 MVA, 4/4 kv, transformer, with delta connection on the converter side. The test circuit is outlined in Fig. 5. A. MMC Without Redundant Submodules The performance of the MMC having 2 submodules per arm (without redundant submodules) is shown in Fig. 6. The 21-level upper and lower arm voltages of the converter are shown in Fig. 6(a). The arm and circulating currents of one phase-leg are shown in Fig. 6. As can be seen from the figure, the circulating current does not contain second harmonic ripple. This also shows that the open-loop approach is successful in eliminating the second harmonic ripple from the circulating current. Fig. 6 shows the total capacitor voltage of the upper and lower arms of the same phase-leg. The total capacitor voltage ripple was found to be 7.8%. Fig. 6 shows the submodule capacitor voltages in the upper arm of one phase-leg Fig. 6. Simulation results without redundant submodules in the arms. (a) Upper and lower arm voltages. Arm and circulating currents. Total capacitor voltages of upper and lower arms. Submodule capacitor voltages of upper arm B. MMC with Redundant Submodules (Strategy 1) (N = 2, R = 2) using the constant total capacitor voltage strategy is shown in Fig. 7. At t = 1 s, a submodule failure occurs (F = 1) in the upper arm, reducing the number of redundant submodules in that arm to 1. Before the submodule failure, both the upper and lower arms have a 23-level arm voltage. As a result of the submodule failure, the upper arm voltage levels reduce to 22, whereas the lower arm voltage levels remain intact as shown in Fig. 7(a). A short transient in the upper and lower arm currents is observed at the submodule failure instant. However, except this transient, the arm currents remain balanced as shown in Fig. 7. Compared to the reference case the addition of the redundant submodules in the arm causes a slight increase in the total capacitor voltage ripple as shown in Fig. 7. The total capacitor voltage ripple decreases with the number of faulty submodules. However, the average total capacitor voltage of healthy submodules in the arm remains constant. The average submodule capacitor voltages in the arm with the faulty submodule increase by 4.8% as illustrated in Fig. 7. However, for a typical MMC used for HVDC transmission having hundreds of submodules per arm, the increase in submodule capacitor voltages will be negligible. v Σ cu v Σ cl 3925

5 (a) (a) v arm,u v arm,l v arm,u v arm,l i u i l i diff 1 i u i l i diff Current (ka).5 Current (ka) v Σ cu v Σ cl 7 65 v Σ cu v Σ cl Fig. 7. Simulation results for the constant total capacitor voltage strategy. (a) Upper and lower arm voltages. Arm and circulating currents. Total capacitor voltages of upper and lower arms. Submodule capacitor voltages of upper arm. C. MMC with Redundant Submodules (Strategy 2) (N = 2, R = 2) using the constant submodule capacitor voltage strategy is shown in Fig. 8. This strategy produces a 21-level arm voltage. Also, in this case a submodule failure occurs in the upper arm at t = 1 s. The submodule failure has no impact on the output levels of the arm voltage as shown in Fig. 8(a). Fig. 8 shows that no transient is experienced by the arm and circulating currents. As shown in Fig. 8, the total capacitor voltage of the arm decreases, when the submodule failure occurs. The total capacitor voltage of the lower arm remains unaffected. Fig. 8 shows that the submodule failure has no effect on the submodule capacitor voltages. It can also be observed from the figure that compared to the reference case, this strategy reduces the submodule capacitor voltage ripple. V. EXPERIMENTAL VERIFICATION is also validated experimentally using a three-phase, 1-kVA MMC prototype in the laboratory. The prototype has 5 submodules per arm including 1 redundant submodule (N =4,R = 1). On the ac-side each phase of the prototype is connected to an RL load. The parameters for the MMC prototype used for experimental verification are summarized in Table II. Same parameters are now adapted for the simulation model as well. Fig. 8. Simulation results for constant submodule voltage strategy. (a) Upper and lower arm voltages. Arm and circulating currents. Total capacitor voltages of upper and lower arms. Submodule capacitor voltages of upper arm. TABLE II EXPERIMENTAL VALUES FOR MMC PROTOTYPE Parameter Rated power of the prototype Peak value of internal emf reference (e s) Input dc voltage (V d ) Submodule capacitance (C) Arm inductance (L) Arm resistance (R) Load resistance per phase Load inductance per phase Carrier frequency (f carr) Value 1 kva 225 V 5 V 3.3 nf 4.5 mh 1 Ω 12 Ω 1 mh 1kHz Direct modulation control [12] is used to obtain the insertion indices for both the prototype and the simulation model. The direct modulation approach is simply based on PWM using sinusoidal insertion indices. In this approach, the insertion indices do not compensate for the voltage variations in the submodule capacitors in the arm. As shown in the previous section that constant total capacitor voltage strategy involves more internal dynamics than constant submodule capacitor voltage strategy. Hence, a submodule failure condition for this strategy was tested experimentally. The redundant submodule in the upper arm of one phase-leg is permanently bypassed. In response to that, a 25% increase in the submodule capacitor voltages of the remaining submodules was observed, as shown in 9(a). It 3926

6 Voltage (V) Voltage (V) Experimental Results (a) Simulation Results Fig. 9. Experimental and simulation results when a submodule in the upper arm is bypassed. (a,b) Submodule capacitor voltages in the upper arm. (c,d) Submodule capacitor voltages in the lower arm. can be observed from 9 that except a short transient, the submodule capacitor voltages in the lower arm of the same phase-leg remain unaffected. Simulation results obtained for the same control strategy show an almost identical behavior as illustrated in 9and. VI. CONCLUSION This paper has investigated the performance of the modular multilevel converter with redundant submodules in the arms. Two different control strategies are used to integrate the redundant submodules in the arms of the MMC. Pros and conds of both strategies were compared using the detailed equivalent model of the MMC. Hence, based on the requirement for a particular application, the suitable strategy can be selected. REFERENCES [1] A. Lesnicar and R. Marquardt, An innovative modular multilevel converter topology suitable for a wide power range, in Proc. IEEE Power Tech Conf., Bologna, Italy, Jun. 23. [2] S. Debnath, J. Qin, B. Bahrani, M. Saeedifard, and P. Barbosa, Operation, control, and applications of the modular multilevel converter: A review, IEEE Trans. Power Electron., vol. 3, no. 1, pp , Jan [3] S. Allebrod, R. Hamerski, and R. Marquardt, New transformerless, scalable modular multilevel converters for HVDC-transmission, in Proc. IEEE Power Electron. Specialists Conf., 28., Jun. 28. [4] B. Jacobson, P. Karlsson, G. Asplund, and T. Jonsson, VSC-HVDC transission with cascaded two-level converters, in Cigrè Session, Osaka, Japan, Nov. 28. [5] M. Hagiwara, K. Nishimura, and H. Akagi, A medium-voltage motor drive with a modular multilevel PWM inverter, IEEE Trans. Power Electron., vol. 25, no. 7, pp , Jul. 21. [6] M. Winkelnkemper, A. Korn, and P. Steimer, A modular direct converter for transformerless rail interties, in Proc. IEEE International Symp. on Ind. Electron., Jul. 21. [7] M. Spichartz, V. Staudt, and A. Steimel, Modular multilevel converter for propulsion system of electric ships, in Proc. IEEE Electric Ship Technologies Symp., Apr [8] A. Hillers and J. Biela, Optimal design of the modular multilevel converter for an energy storage system based on split batteries, in Proc. 15th Power Electron. and Appl. Conf. (EPE), Sep [9] N. Ahmed, S. Norrga, H.-P. Nee, A. Haider, D. Van Hertem, L. Zhang, and L. Harnefors, HVDC supergrids with modular multilevel converters the power transmission backbone of the future, in Proc. 9th Int. Multi-Conf. on Syst., Signals and Devices (SSD), Mar [1] B. Gemmell, J. Dorn, D. Retzmann, and D. Soerangr, Prospects of multilevel VSC technologies for power transmission, in Proc. IEEE Transm. and Dist. Conf. and Expo., Apr. 28. [11] A. Hassanpoor, S. Norrga, H. Nee, and L. Ängquist, Evaluation of different carrier-based PWM methods for modular multilevel converters for HVDC application, in Proc. 38th Annual Conf. IEEE Ind. Electron. Society (IECON), Oct [12] D. Siemaszko, A. Antonopoulos, K. Ilves, M. Vasiladiotis, L. Ängquist, and H.-P. Nee, Evaluation of control and modulation methods for modular multilevel converters, in Proc. IEEE Power Electron. Conf. (IPEC), Jun. 21. [13] G. Konstantinou, M. Ciobotaru, and V. Agelidis, Selective harmonic elimination pulse-width modulation of modular multilevel converters, IET Power Electron., vol. 6, no. 1, Jan [14] 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 , Jan [15] M. Hagiwara, R. Maeda, and H. Akagi, Control and analysis of the modular multilevel cascade converter based on double-star choppercells (MMCC-DSCC), IEEE Trans. Power Electron., vol. 26, no. 6, pp , Jun [16] H. Akagi, Classification, terminology, and application of the modular multilevel cascade converter (MMCC), IEEE Trans. Power Electron., vol. 26, no. 11, pp , Nov [17] Y. Li and F. Wang, Arm inductance selection principle for modular multilevel converters with circulating current suppressing control, in Proc. IEEE 28th App. Power Electron. Conf. and Expo. (APEC), Mar [18] X. She and A. Huang, Circulating current control of double-star chopper-cell modular multilevel converter for HVDC system, in Proc. 38th Annual Conf. IEEE Ind. Electron. Society (IECON), Oct [19] L. Harnefors, A. Antonopoulos, S. Norrga, L. Ängquist, and H.-P. Nee, Dynamic analysis of modular multilevel converters, IEEE Trans. Ind. Electron., vol. 6, no. 7, pp , Jul [2] Q. Song, W. Liu, X. Li, H. Rao, S. Xu, and L. Li, A steady-state analysis method for a modular multilevel converter, IEEE Trans. Power Electron., vol. 28, no. 8, pp , Aug [21] A. Antonopoulos, L. Ängquist, and H.-P. Nee, On dynamics and voltage control of the modular multilevel converter, in Proc. 13th European Power Electron. and Appl. Conf. (EPE), Sep. 29. [22] U. Gnanarathna, A. Gole, and R. Jayasinghe, Efficient modeling of modular multilevel hvdc converters (MMC) on electromagnetic transient simulation programs, IEEE Trans. Power Del., vol. 26, no. 1, pp , Jan [23] N. Ahmed, L. Ängquist, S. Norrga, A. Antonopoulos, L. Harnefors, and H.-P. Nee, A computationally efficient continuous model for the modular multilevel converter, IEEE Journal of Emerging and Selected Topics in Power Electron., vol. 2, no. 4, pp , Dec [24] N. Ahmed, L. Ängquist, S. Mehmood, A. Antonopoulos, L. Harnefors, S. Norrga, and H. Nee, Efficient modeling of an MMC-based multiterminal dc system employing hybrid HVDC breakers, IEEE Trans. Power Del., vol. PP, no. 99, pp. 1 1, 215. [25] G. Konstantinou, M. Ciobotaru, and V. Agelidis, Effect of redundant sub-module utilization on modular multilevel converters, in Proc. IEEE International Conf. on Ind. Tech. (ICIT), Mar [26] G. Konstantinou, J. Pou, S. Ceballos, and V. Agelidis, Active redundant submodule configuration in modular multilevel converters, IEEE Trans. Power Del., vol. 28, no. 4, pp , Oct [27] G. T. Son, H.-J. Lee, T. S. Nam, Y.-H. Chung, U.-H. Lee, S.-T. Baek, K. Hur, and J.-W. Park, Design and control of a modular multilevel HVDC converter with redundant power modules for noninterruptible energy transfer, IEEE Trans. Power Del., vol. 27, no. 3, Jul [28] L. Ängquist, A. Antonopoulos, D. Siemaszko, K. Ilves, M. Vasiladiotis, and H.-P. Nee, Open-loop control of modular multilevel converters using estimation of stored energy, IEEE Trans. Ind. Appl., vol. 47, no. 6, pp , Nov