A Modular Multiport Power Electronic Transformer With ... - IEEE Xplore

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 62, NO. 5, MAY 2015

3213

A Modular Multiport Power Electronic Transformer With Integrated Split Battery Energy Storage for Versatile Ultrafast EV Charging Stations Michail Vasiladiotis, Student Member, IEEE, and Alfred Rufer, Fellow, IEEE Abstract—This paper proposes a power converter architecture for the implementation of an ultrafast charging station for electric vehicles (EVs). The versatile converter topology is based on the concept of the power electronic transformer. For the direct transformerless coupling to the medium-voltage grid, a cascaded H-bridge (CHB) converter is utilized. On the level of each submodule, integrated split battery energy storage elements play the role of power buffers, reducing thus the influence of the charging station on the distribution grid. The power interface between the stationary split storage stage and the EV batteries is performed through the use of parallel-connected dual-half-bridge dc/dc converters, shifting the isolation requirements to the medium-frequency range. By choosing several different submodule configurations for the parallel connection, a multiport output concept is achieved, implying the ability to charge several EVs simultaneously without the use of additional high-power chargers. All possible charging station operating modes among with the designed necessary control functions are analyzed. The state-of-charge self-balancing mode of the delta-connected CHB converter is also introduced. Finally, the development of a downscaled laboratory prototype is described, and preliminary experimental results are provided. Index Terms—Cascaded H-bridge (CHB) converter, dc ultrafast charging, dual half-bridge (DHB), electric vehicles (EVs), isolated dc/dc converter, multiport converter, power electronic transformer (PET), split battery energy storage, state of charge (SoC) balancing.

I. I NTRODUCTION

E

NVIRONMENTAL awareness and related concerns have given rise to a high interest toward electrical mobility based on battery energy storage during the recent years. Such an action aims mainly at the significant reduction of CO2 and other pollutant emissions. In order to ensure the widespread utilization of such electric vehicles (EVs), however, significant effort is still needed regarding their competitive market Manuscript received June 13, 2014; revised September 10, 2014; accepted October 17, 2014. Date of publication November 5, 2014; date of current version April 8, 2015. This work was supported by EOS Holding, Switzerland, in the framework of the Ultra-Fast Charging of Electric Vehicles project. The authors are with the Laboratory of Industrial Electronics, École Polytechnique Fédérale de Lausanne, 1015 Lausanne, Switzerland (e-mail: [email protected]; [email protected]). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TIE.2014.2367237

launching over conventional combustion engine vehicles. The main impediment when it comes to EVs concerns the limited battery autonomy in conjunction with the long charging times. Most often, onboard chargers are utilized [1]–[3], fed by the domestic or public grid. When the charging power level increases, however, off-board chargers are preferred which are usually installed outside residential areas in charging stations conceptually similar to the gas-filling ones [4]. A solution to the autonomy problem can be given through the utilization of the so-called battery-swap stations, where the discharged EV battery is directly replaced by a fully charged one. The latter has recently stimulated research interest [5]–[7]. Some of the issues that arise, however, concern the different EV battery type compatibility as well as the charged battery availability of the station. A second attractive solution regards the fast and ultrafast EV charging referring to less than 30 and 10 min, respectively. The latter has been driven by the late high expectations regarding the development of batteries using materials that can withstand very high charging rates [8], [9] as well as the development of more advanced battery charging techniques [10]. Both industry and academia have been pushed toward the fast and ultrafast EV charging concept exploration. The main focus has been initially laid on the infrastructure as well as grid impact [11], [12]. In order not to overload the grid in such a power demanding application, the use of stationary energy storage elements has been proposed [13]–[16]. Advanced and emerging converter technologies, such as power electronic transformers (PETs), are also expected to play a key role in the development of the future charging stations, both in terms of flexibility and efficiency [11], [17]–[21]. The Electric Power Research Institute claims an up to 8% overall system efficiency increase through their Utility Direct Fast Charger Technology [19]. In [20], a novel multiport converter architecture has been proposed for the implementation of medium-voltage ultrafast EV charging stations (UFEVCSs), based on such a concept. The costly and bulky low-frequency transformer is avoided through the utilization of a cascaded H-bridge (CHB) multilevel converter. The latter also offers the possibility to significantly reduce the filtering components on the grid side since the injected currents exhibit very low harmonic content. Such converters have been already utilized in railway traction [22]–[24], network coupling, and back-to-back loop power flow control [25], [26], as well as microgrid applications [27], [28].

0278-0046 © 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

3214


unit. Section V describes the development of a downscaled laboratory prototype and presents some preliminary experimental results, leading to Section VI which concludes this paper.

II. G LOBAL S YSTEM D ESCRIPTION

Fig. 1. Proposed versatile EV charging station converter architecture, based on the concept of M2 PET-CHB converter with integrated battery energy storage and parallel-connected isolated dc/dc converters.

The proposed charging station implementation of [20] is shown in Fig. 1. The CHB converter offers another advantage, i.e., the possibility to integrate battery energy storage elements on the level of each submodule in a split manner. As mentioned in the previous paragraph, this storage stage plays the role of a power buffer, reducing therefore the instantaneous active power extracted from the grid. As in any PET application, mediumfrequency transformer-based dc/dc converters are connected in parallel. In this case, such an action achieves the needed high output currents as well as meeting the EV battery galvanic isolation standards. Compared to the similar converter structure of [21], the proposed converter features two main advantages. Initially, no central high-capacity battery is needed due to the split accumulation concept. This leads to redundancy as well as more straightforward battery management system (BMS) designs. In addition and since a high number of submodules are needed to block the medium voltage on the grid side, different configurations can be chosen on the parallel-connection level in order to achieve a multiport output capable of charging different vehicles simultaneously. The proposed converter architecture is therefore referred to as modular multiport power electronic transformer (M2 PET). In such a case, no additional highpower chargers are needed, leading to higher efficiencies due to the equal power split between the existing isolated converters. However, the converter control becomes more complicated, due to the need for the handling of all power flow directions. The objective of this paper is to present the most important findings of the proposed converter topology investigations up to date as well as to establish and validate a global control system capable of handling all possible EV charging station operating modes. It is organized as follows. Section II introduces the global system components. Section III presents and analyzes the control system for the EV charging station architecture and validates it by means of simulation results. Section IV discusses a novel proposed SoC self-balancing operating mode of the delta-connected CHB-battery energy storage system (BESS)

The proposed versatile EV charging station, which is based on the concept of M2 PET with integrated split battery energy storage, is illustrated in detail in Fig. 1. It consists of three conversion stages: 1) the transformerless active front end (CHB converter); 2) the stationary batteries with the associated interface solution (split storage); and 3) the medium-frequency transformer-based dc/dc converters (isolation stage). The threephase CHB converter can be utilized either in star or in delta configurations. The chosen implementation of the whole conversion chain is depicted in Fig. 2. From this, it can be seen that the CHB converter submodule, which is implemented by means of a full bridge, acts as a single-phase controlled rectifier. Due to the latter, a low-frequency current component of the second harmonic grid frequency appears on its dc link. This poses an impediment for the direct placement of the stationary batteries on the level of each submodule. This second-order harmonic does not contribute to the average battery state of charge (SoC) and only causes resistive losses and temperature variations. Therefore, it is desired to be eliminated. In [20], two different interface solutions were mentioned, namely, the direct passive interface by means of resonant and high-order filtering elements as well as the indirect active interface (IAI) based on additional power electronics circuits. The IAI has been chosen for this paper and is shown in Fig. 2. It is implemented as a standard nonisolated buck converter, which offers an effective active power decoupling between the submodule and the battery buffer as well as an additional degree of freedom for the control of the system. Regarding the implementation of the power interface between the CHB converter and the EV batteries (isolation stage), the basic requirements that it should fulfill are the following: 1) very low charging current ripple; 2) current and voltage control capability; and 3) galvanic isolation. Soft switching over a wide operating range is also favorable since it implies the reduction of passive components through an increase in the switching frequency without affecting significantly the converter efficiency. In works such as [11] and [20], the straightforward solution of a full-bridge phase-shift controlled pulse-width modulation (PWM) converter is discussed. However, several drawbacks of such a topology have been highlighted, such as the unequal thermal stress of the two primary converter legs, as well as the incapability of bidirectional power flow without additional external circuitry. Therefore, Vasiladiotis et al. [29] have investigated the more appropriate solution of the dual half-bridge (DHB) converter proposed by Li et al. [30]. The DHB exhibits significant advantages and fulfills all the aforementioned requirements. In addition, it offers the ability of transferring power from the EV battery to the grid, i.e., vehicle-to-grid (V2G) operation. This topology completes the whole conversion chain depicted in Fig. 2. By phase shifting the transformer secondary voltage in regard to the primary, a

VASILADIOTIS AND RUFER: M2 PET WITH INTEGRATED SPLIT BATTERY ENERGY STORAGE FOR UFEVCSs

Fig. 2.

Implementation of a whole conversion chain for the proposed topology.

Fig. 3.

Overall control block diagram for the M2 PET-based UFEVCS system.

specific amount of active power can be transferred between the two transformer sides. It is noted that the superscript/subscript k ∈ {a, b, c} (star) or {ab, bc, ca} (delta) refers to the kth CHB converter branch and i ∈ {1, . . . , N } refers to the ith submodule of the respective branch. According to the active power direction illustrated in Fig. 2, the following relation holds for the (ki)th conversion chain: ki ki ki PEV = Psm − Pbat =

k Pbr ki + ΔPsm N

ki . − Pbat

(1)

Finally, it is evident that the global system architecture is bidirectional in terms of power flow. Therefore, it can also provide ancillary services, such as frequency control as well as reactive power compensation.

III. G LOBAL S YSTEM C ONTROL AND S IMULATION The proposed global control system for the M2 PET-based UFEVCS is illustrated in Fig. 3, together with all associated feedback variables and control signals. In the following, the functionality of the most important blocks is explained. Finally,

3215

the blocks included in dashed rectangles are executed either in the case of either star- or delta-configured CHB converters. A. Control of the CHB-BESS System The grid voltages ugk and branch currents ik are measured and are used for the synchronization with the three-phase grid as well as vector current control in a rotating reference frame (RRF). The control of the converter inner quantities is then split in two subproblems, the voltage and power control. The latter implies the individual power control between the submodules of a branch as well as across the three converter branches, without disturbing the grid symmetry. This is essential in this application, where each charging port will demand different amounts of power to be handled. The individual power control can be also used in order to balance the SoCs of the intermediate batteries, which will be unbalanced due to the different port power demands. 1) Submodule Capacitor Voltage Control: The IAI control is responsible for stabilizing the submodule capacitor voltage as well as regulating the battery currents explicitly. This is achieved by means of a cascaded set of outer-voltage/innercurrent PI controllers. This is a well-established technique, which is not repeated here. However, the interested reader

3216


might refer to [31] for a respective control block diagram as well as controller gain tuning procedures. 2) Individual Submodule Power Control: It acts by multiplying the whole branch voltage reference uk by a power ratio factor βki calculated as βki

P ki = N sm = ki Psm

k Pbr N

ki + ΔPsm k Pbr

with

N

βki = 1. (2)

i=1

i=1

3) Individual Branch Power Control: For the storage unit in the star configuration, the degree of freedom to achieve this is the injection of a common-mode voltage, whose phase φcm and amplitude u ˆcm are calculated for a three-phase balanced grid as b a 2ΔPbr /ΔPbr +1 √ φcm = φi − θ = φi − arctan 3 u ˆcm =

a 2ΔPbr î cos θ

(3)

where î and φi are the grid phase current amplitude and angle, k denotes the branch balancing power. respectively, and ΔPbr For the delta-connected storage unit, the same goal can be reached by injecting a circulating current between the three branches that will not affect the grid current and whose phase φcirc and amplitude îcirc are obtained accordingly as ab bc 2ΔPbr /ΔPbr +1 √ φcirc = − arccot 3 îcirc = √

ab 2ΔPbr 3ˆ u cos φcirc − π6

(4)

where, similarly, u ˆ represents the converter phase voltage amplitude. It is noted that the CHB active front end stage gives a significant degree of freedom for choosing the number of submodules feeding each charging port. Hence, a parallel connection of the isolated modules between different phases gives the additional advantage of maintaining the symmetry of the three-phase grid currents during the EV charging without the need for commonmode voltage/circulating current injection. 4) SoC Balancing Control: The block diagrams for the branch and submodule SoC balancing controllers are shown in Fig. 4. The branch SoC control forces the kth branch mean SoC k value SoC br to follow the mean SoC value of all three branches SoC 3br . These two quantities are defined as k SoC br

N 1 ki = SoCsm N i=1 a

SoC 3br

(5) b

c

ab bc ca SoC br + SoC br + SoC br . = 3

(6)

The submodule SoC control forces accordingly the kith subk ki module value SoCsm to follow the branch mean value SoC br . The first-order transfer functions relating the branch and sub-

Fig. 4. Block diagrams for (a) branch and (b) submodule balancing of battery SoCs.

module active powers with the respective SoCs can be modeled in the Laplace domain as Gbr S (s) =

100 sm s · N Ebat,n

and

Gsm S (s) =

100 sm s · Ebat,n

(7)

sm denotes the where the SoC is a percentage (%) and Ebat,n submodule battery nominal energy expressed in (W · s). For the calculation of the proportional control gains kpbr and kpsm , certain limitations exist, related to semiconductor voltage/current ratings as well as operation without submodule overmodulation, respectively. Such expressions can be utilized in real time, in order to achieve reduced balancing closed-loop control rising times, according to operating point changes without threatening the safe operation of the converter. Therefore, a gain-scheduling adaptive control algorithm is at hand and has been implemented as well. It is finally noted that the reader is referred to [32] for a detailed description of the control for CHB converters with integrated split BESS, including design principles, SoC control gain limitations, and operation under grid unbalances.

B. Control of DHB Although several techniques have been proposed for achieving fast and ultrafast charging of batteries, this paper focuses on the constant current (CC) phase of a traditional CC–constant voltage charging process. This will reach the EV battery typically up to a point of about 80% of its SoC. Therefore, the DHB converters should be able to provide a desired current value, implying the necessity of an accurate control method derivation. The DHB converter exhibits nonlinearities and dynamics of high order, due to the existence of several passive elements. In this paper, a discrete-time control design is utilized, based on the use of nonparametric models [33], [34]. The required nonparametric model of the single-input–single-output (SISO) system, considering the phase shift φki as input and the current Ioki as output, can be derived from a duty-cycle-dependent analytical average system model or through the use of an identification process. An additional challenge associated with the DHB control regards the second-order submodule current harmonic, which will be seen at the output current Ioki if no additional action is taken. The latter can be compensated for by adding a resonant term to the controller, tuned at this desired


3217

TABLE I S IMULATION /D OWNSCALED P ROTOTYPE PARAMETERS

frequency (100 Hz in the studied case). The overall controller transfer function in the z-domain is obtained as [35] K(z, ρ) =

b1 z −1 + b2 z −2 ρ1 + ρ2 z −1 + ρ . 3 1 − z −1 1 + a1 z −1 + a2 z −2

(8)

The second term represents a discrete-time resonant controller, where ωh is the desired frequency and ζ = 3/ωh is the damping ratio. The coefficients a1−2 and b1−2 are determined by ζωh ζωh 2 η , b2 = α + α η−β b1 = 1 − α β + ωb ωb α1 = − 2αβ, α2 = α2 (9)

where ωb = ωh 1−ζ 2 for ζ < 1, α = e−ζωh Ts , β = cos(ωb Ts ), η = sin(ωb Ts ), and Ts denotes the controller sampling time. Except for the resonant controllers, which compensate for the second harmonic locally on the level of each DHB, the parallel connection between different phases permits a further reduction, as mentioned in a previous paragraph. A second challenge comes from the fact that both the submodule voltage Usm and the EV battery voltage Uo are experiencing significant variations (due to the submodule second current harmonic and EV battery SoC, respectively). This limits the converter zero voltage switching (ZVS) region, particularly at light loads. However, it can be overcome through an acceptable change in the duty cycle D in both primary and secondary half bridges [29], [36]–[38]. The latter can be achieved by detecting the converter operating point in real time and choosing the respective duty-cycle value. For a detailed description of the current control design and the ZVS switching regions of the DHB converter for the specific application, the reader is referred to [29]. C. Simulation Model and Results In order to verify the effectiveness of the proposed control system, a simulation model of the whole PET-based EV charging station has been built. Since the PWM effects are of no significant interest compared to the evolution of magnitudes with high time constants, the switching-averaged power converter models have been considered for all conversion stages. The details for the parameters of the studied system as well as the charging station configuration are given in Table I.

Fig. 5. Simulation results of the M2 PET-based UFEVCS for a hypothetical power profile, exploiting all operation modes.

Fig. 5 shows the simulation results for the different operating modes of the converter architecture utilizing a hypothetical power profile. In Mode I, the converter behaves as an ultrafast EV charger. It is assumed that the three out of four EV charging ports are used, whereas the fourth is in idle mode. The three EVs arrive at the station with a time difference of 1 min and are charged within 10 min each. The three ports charging the EVs will cause a discharge to the stationary intermediate batteries, whereas the fourth port will be charging the respective storage stage. In Mode II, the DHBs are not operating, and the converter structure extracts active power from the grid, in order to recharge the power buffers. Accordingly, in Mode III, the charging station provides active power to the grid by discharging the intermediate batteries. Finally, in Mode IV, no active power is exchanged with the grid or the EVs, and the CHB converter operates with its well-known functionality of a

3218


Fig. 6. Detailed versions of the three-phase grid voltage and current quantities for (a) Mode I, (b) Mode II, (c) Mode III, and (d) Mode IV corresponding to Fig. 5.

STATCOM. At the end of the power profile, all battery SoCs are practically the same due to the SoC balancing controllers that are in action throughout the whole converter operation. Fig. 6 shows some details of the three-phase grid voltage and current quantities for different operation modes. It is clear that the control system is capable of keeping the current symmetry unaffected, even during asymmetric load conditions, e.g., when not all EV charging ports are in use and some intermediate batteries are charging while others are discharging. The active power evolution of the four different charging ports is illustrated in Fig. 7. It is clear that the submodule SoC balancing power ΔPsm plays a major role in the asymmetric power distribution between the stationary batteries. For example, if a charging port is not used, the grid power fed to it will be minimal, ensuring a small SoC variation in the respective batteries. On the other hand, the ports that need a larger amount of charging power will absorb more grid power, leading to a need for less power coming from the stationary batteries. The effect of the gain-scheduling controller for the vertical SoC balancing action is depicted in Fig. 8. It is clear that, when the SoC deviation exceeds a predefined value, the rise time of the closed-loop control system increases. This leads to a respective decrease of the control gain kpsm in order to avoid the saturation of the control action, which would cause submodule overmodulation in this case. It is noted that the gain kpsm has to be the same for all submodules of a branch, in order ki = 0 and that the grid current controller to ensure that ΣΔPsm is not affected. IV. S ELF -B ALANCING O PERATION OF THE D ELTA -CHB In the previous section, the star-connected CHB converter has been considered for the simulation studies. As mentioned

Fig. 7. Simulation results for the active power distribution within the four charging ports.

Fig. 8. Gain-scheduling control behavior of the vertical SoC balancing controller.

in Section II, however, the choice of a delta configuration exists as well. In general, there is a duality between the two aforementioned configurations [32]. Indeed, the delta connection implies the requirement of blocking the whole line-to√ line grid voltage, which is larger by a factor of 3 compared to the star-connected case (if no common-mode voltage is considered). However, the rated branch currents are decreased by the same factor. In addition, the branch balancing is achieved by a circulating current between the three branches without disturbing the grid phase currents as an equivalent to the common-mode voltage injection in the star-configured case.


3219

Fig. 9. Delta-connected active front end CHB converter with integrated split battery energy storage.

The choice between a star or delta connection is related to the design process and parameters such as cost, efficiency, etc. Therefore, no straightforward answer exists for the studied system, particularly since no available data exist yet for such a real ultrafast charging station design. The only apparent advantage of the delta-connected CHBBESS unit over the star-connected one comes through an additional attractive feature, which is now introduced in the framework of the proposed application. Since the circulating current flows among the three connected branches, it can be used to balance the intermediate SoCs without affecting the grid currents and, therefore, the grid power. This leads to an additional operation Mode V, namely, the self-balancing mode. Such an action is not possible in the star-connected CHB, where, even if a common-mode voltage ucm is imposed, a grid current is also needed to create the necessary three branchflowing power. In order to highlight the specific operation mode, the deltaconnected CHB-BESS unit of Fig. 9 has been simulated. It is assumed that the charging station is not operating, e.g., during the night, but the intermediate battery SoCs have significant differences as a result of the unequal port power distribution of a specific day. It is further assumed that the converter cannot be balanced during a battery charging or discharging operation through the grid (Mode II or III). The simulation results for the self-balancing operation are given in Fig. 10. The circulating current injection is defined such as to sequentially discharge one branch and use the resulting power to charge its adjacent. At the end of the self-balancing mode, the submodule SoC deviations are converging to zero. It is noted that, since only two branches are used at any time instant, this creates a periodical unbalance at the branch SoC values. The latter does not cause problems if the specified time intervals are always the same and are executed in a perfectly circular manner. This is also verified in Fig. 10, where the branch SoC deviation value is zero at the end of this 30-min operating mode.

Fig. 10. Time-domain behavior of the delta-connected CHB converter in SoC self-balancing operation.

A detail of the inner converter magnitudes is given in Fig. 11. During this time interval, branch ab is discharging into bc, whereas branch ca does not absorb or inject any power. The vertical SoC balancing controllers are in action, leading to different submodule power values reflected in the different submodule capacitor voltage ripples. The battery voltages are significantly different due to their SoC variations, but the mean submodule voltages are the same. The latter proves the advantage of using the IAI concept over a passive filtering solution. Finally, only a circulating current exists flowing among the three branches, since the grid currents are controlled to be zero. V. E XPERIMENTAL T ESTS In this section, the development of a reduced-scale laboratory prototype is described, and some preliminary experimental

3220


Fig. 12.

Experimental setup.

Fig. 11. Detail of delta-connected CHB converter magnitudes during a discharge of branch ab and charging of branch bc in SoC self-balancing operation.

results are presented. Regarding the CHB-BESS active front end stage, a single branch has been built in the framework of the studied project. It consists of eight submodules, which are interfaced to split battery energy storage systems by means of individual IAIs. Both the submodules and the IAIs are mounted in a rack and are connected to backplanes, which carry all gate signals, power supplies, and measurements. More specifically, the submodule voltages and the battery voltages/currents are measured for control and security purposes. The utilized battery packs are of LiFePO4 25.6 V/3 Ah technology [39] and are placed in the same rack. As far as the isolation stage is concerned, a number of four dual half bridges have been also built, which are connected to the first four respective submodules of the branch. The transformers are constructed using ETD49 cores by EPCOS, and the windings consist of Litz wire. The printed circuit boards (PCBs) are also backplane connected, and their voltage and current outputs are sensed as well. The converter control is carried out in a custom hardware platform, namely, the MMCbox [40], whose design and features resemble that of [41]. A modularized concept is followed similarly to the power stage, where all hardware components are connected to a backplane and communicate by means of a parallel bus. The communication supervision is performed by a central microcontroller-based unit, which is in charge of all high-level control task execution and duty-cycle calculation. Two slave field-programmable gate array (FPGA)-based boards (CHB-BESS and isolation stages, respectively) are then responsible for producing the firing signals and retrieving the necessary measurements from the converter, i.e., the submodule voltages as well as the intermediate battery and DHB output currents/voltages. In addition, they ensure the converter protection, in cases where overvalues are detected. An additional FPGA-based board deals with the retrieval of external converter

Fig. 13. Experimental results for a discharging procedure of the batteries.

measurements, such as grid voltages, as well as the control of relays used for converter precharging and discharging operations. The whole implemented laboratory setup is illustrated in Fig. 12. Fig. 13 shows the results for a split storage stage discharge, corresponding to operation Mode III. The 17-level high-resolution converter waveform leads to a very smooth sinusoidal grid current waveform. Since the measurements are performed through a digital oscilloscope, only the four upper submodule capacitor voltages and split battery currents are depicted because of the channel number limitation. The voltages are very well balanced due to the IAI cascaded voltage/current control. The battery current waveforms coincide as well and contain only the switching-related ripple. The results for another experimental test concerning the parallel connection of the four DHB modules are illustrated in Fig. 14, where the EV battery is emulated by a variable voltage source. It is shown that the interleaving of four channels with 0.5 duty cycle each leads to a completely ripple-free total battery current IoΣ .


3221

R EFERENCES

Fig. 14. Experimental results for the parallel connection of the four DHB modules.

It is noted that the initial purpose of the prototype is to test the basic control/modulation functions and converter operating modes as well as provide a step-by-step experimentation of the so far developed elements. Moreover and due to several restrictions, such as the existence of only one CHB-BESS branch, the lack of SoC estimation provided by the embedded BMS, etc., the implementation of the global system control and SoC balancing algorithms that have been proposed and described in the previous paragraph are not in the scope of this paper. For more experimental results of the developed versatile laboratory prototype in the framework of similar modular multilevel converter-based applications, including submodule integration with BESS elements, the interested reader is referred to works such as [29], [31], and [40]. VI. C ONCLUSION This paper has proposed a multiport PET-based concept for the realization of multifunctional medium-voltage UFEVCSs. All system components have been tailored for the specific application and chosen accordingly. A global system control structure has been described, which is capable of handling all different power flow directions as well as capacitor voltage and battery SoC unbalances. The operating modes of the system have been presented and simulated, verifying the versatility of the introduced converter structure. An additional advantageous SoC self-balancing operating mode of the delta-connected CHB-BESS unit over the star-connected one has been proposed and discussed. Preliminary results from a downscaled developed prototype have finally supported the theoretical investigations. ACKNOWLEDGMENT The authors would like to thank Dr. B. Bahrani for his help on the dual-half-bridge current control design as well as N. Cherix for his valuable contribution to the prototype development.

[1] S. Haghbin, S. Lundmark, M. Alaküla, and O. Carlson, “Grid-connected integrated battery chargers in vehicle applications: Review and new solution,” IEEE Trans. Ind. Electron., vol. 60, no. 2, pp. 459–473, Feb. 2013. [2] J.-Y. Lee and H.-J. Chae, “6.6-kW onboard charger design using DCM PFC converter with harmonic modulation technique and two-stage dc/dc converter,” IEEE Trans. Ind. Electron., vol. 61, no. 3, pp. 1243–1252, Mar. 2014. [3] Z. Amjadi and S. S. Williamson, “Power-electronics-based solutions for plug-in hybrid electric vehicle energy storage and management systems,” IEEE Trans. Ind. Electron., vol. 57, no. 2, pp. 608–616, Feb. 2010. [4] M. Yilmaz and P. T. Krein, “Review of battery charger topologies, charging power levels, infrastructure for plug-in electric and hybrid vehicles,” IEEE Trans. Power Electron., vol. 28, no. 5, pp. 2151–2169, May 2013. [5] A. Kuperman, U. Levy, A. Zafransky, and A. Savernin, “Battery charger for electric vehicle traction battery switch station,” IEEE Trans. Ind. Electron., vol. 60, no. 12, pp. 5391–5399, Dec. 2013. [6] M. Takagi et al., “Economic value of PV energy storage using batteries of battery-switch stations,” IEEE Trans. Sustain. Energy, vol. 4, no. 1, pp. 164–173, Jan. 2013. [7] Q. Dai, T. Cai, S. Duan, and F. Zhao, “Stochastic modeling and forecasting of load demand for electric bus battery-swap station,” IEEE Trans. Power Del., vol. 29, no. 4, pp. 1909–1917, Aug. 2014. [8] B. Kang and G. Ceder, “Battery materials for ultrafast charging and discharging,” Nature, vol. 458, no. 7235, pp. 190–193, Mar. 2009. [9] H. Zhang, X. Yu, and P. V. Braun, “Three-dimensional bicontinuous ultrafast-charge and -discharge bulk battery electrodes,” Nat. Nanotechnol., vol. 6, no. 5, pp. 277–281, May 2011. [10] S.-J. Huang, B.-G. Huang, and F.-S. Pai, “Fast charge strategy based on the characterization and evaluation of LiFePO4 batteries,” IEEE Trans. Power Electron., vol. 28, no. 4, pp. 1555–1562, Apr. 2013. [11] D. Aggeler, F. Canales, H. Zelaya-De La Parra, N. Butcher, and O. Apeldoorn, “Ultra-fast dc-charge infrastructures for EV-mobility and future smart grids,” in Proc. IEEE ISGT, Oct. 11–13, 2010, pp. 1–8. [12] K. Yunus, H. Zelaya-De La Parra, and M. Reza, “Distribution grid impact of plug-in electric vehicles charging at fast charging stations using stochastic charging model,” in Proc. 14th EPE Conf. Appl., Aug. 30– Sep. 1, 2011, pp. 1–11. [13] S. Bai, D. Yu, and S. Lukic, “Optimum design of an EV/PHEV charging station with dc bus and storage system,” in Proc. IEEE ECCE, Sep. 12–16, 2010, pp. 1178–1184. [14] G. Joos, M. de Freige, and M. Dubois, “Design and simulation of a fast charging station for PHEV/EV batteries,” in Proc. IEEE EPEC, Aug. 25–27, 2010, pp. 1–5. [15] H. Hõimoja et al., “Towards ultrafast charging solutions of electric vehicles,” in Proc. Cigré Session 44, Paris, France, Aug. 26–31, 2012, pp. 1–8. [16] S. Bai and S. M. Lukic, “Unified active filter and energy storage system for an MW electric vehicle charging station,” IEEE Trans. Power Electron., vol. 28, no. 12, pp. 5793–5803, Dec. 2013. [17] G. Waltrich, J. L. Duarte, and M. A. M. Hendrix, “Multiport converter for fast charging of electrical vehicle battery,” IEEE Trans. Ind. Appl., vol. 48, no. 6, pp. 2129–2139, Nov./Dec. 2012. [18] H. Hõimoja, M. Vasiladiotis, and A. Rufer, “Power interfaces and storage selection for an ultrafast EV charging station,” in Proc. IET 6th Int. Conf. PEMD, Mar. 27–29, 2012, pp. 1–6. [19] A. Maitra, “Utility direct technology boosts efficiency of fast charging for electric vehicles,” How2power Today Newslett. Apr. 2012. [20] M. Vasiladiotis, A. Rufer, and A. Béguin, “Modular converter architecture for medium voltage ultra fast EV charging stations: Global system considerations,” in Proc. IEEE IEVC, Mar. 4–8, 2012, pp. 1–7. [21] S. Wang, R. Crosier, and Y. Chu, “Investigating the power architectures and circuit topologies for megawatt superfast electric vehicle charging stations with enhanced grid support functionality,” in Proc. IEEE IEVC, Mar. 4–8, 2012, pp. 1–8. [22] A. Rufer, N. Schibli, and C. Briguet, “A direct coupled 4-quadrant multilevel converter for 16 2/3 Hz traction systems,” in Proc. 6th Int. Conf. PEVD, Sep. 23–25, 1996, pp. 448–453. [23] C. Zhao et al., “Power electronic traction transformer-medium voltage prototype,” IEEE Trans. Ind. Electron., vol. 61, no. 7, pp. 3257–3268, Jul. 2014. [24] D. Dujic et al., “Power electronic traction transformer-low voltage prototype,” IEEE Trans. Power Electron., vol. 28, no. 12, pp. 5522–5534, Dec. 2013. [25] F. Iov et al., “UNIFLEX-PM—A key-enabling technology for future European electricity networks,” Eur. Power Electron. Drives Assoc. J., vol. 19, no. 4, pp. 6–16, Oct.–Dec. 2009.

3222


[26] H. Akagi and R. Kitada, “Control and design of a modular multilevel cascade BTB system using bidirectional isolated dc/dc converters,” IEEE Trans. Power Electron., vol. 26, no. 9, pp. 2457–2464, Sep. 2011. [27] T. Zhao, G. Wang, S. Bhattacharya, and A. Q. Huang, “Voltage and power balance control for a cascaded H-bridge converter-based solid-state transformer,” IEEE Trans. Power Electron., vol. 28, no. 4, pp. 1523–1532, Apr. 2013. [28] X. She, A. Q. Huang, and R. Burgos, “Review of solid-state transformer technologies and their application in power distribution systems,” IEEE J. Emerging Sel. Topics Power Electron., vol. 1, no. 3, pp. 186–198, Sep. 2013. [29] M. Vasiladiotis, B. Bahrani, N. Burger, and A. Rufer, “Modular converter architecture for medium voltage ultra fast EV charging stations: Dual halfbridge-based isolation stage,” in Proc. IEEE IPEC, May 18–21, 2014, pp. 1–8. [30] H. Li, F. Z. Peng, and J. S. Lawler, “A natural ZVS medium-power bidirectional dc-dc converter with minimum number of devices,” IEEE Trans. Ind. Appl., vol. 39, no. 2, pp. 525–535, Mar./Apr. 2003. [31] M. Vasiladiotis and A. Rufer, “Analysis and control of modular multilevel converters with integrated battery energy storage,” IEEE Trans. Power Electron., vol. 30, no. 1, pp. 163–175, Jan. 2015. [32] M. Vasiladiotis and A. Rufer, “Balancing control actions for cascaded H-bridge converters with integrated battery energy storage,” in Proc. 15th EPE Conf. Appl., Sep. 2–6, 2013, pp. 1–10. [33] A. Karimi and G. Galdos, “Fixed-order H∞ controller design for nonparametric models by convex optimization,” Automatica, vol. 46, no. 8, pp. 1388–1394, Aug. 2010. [34] G. Galdos, A. Karimi, and R. Longchamp, “H∞ controller design for spectral MIMO models by convex optimization,” J. Process Control, vol. 20, no. 10, pp. 1175–1182, Dec. 2010. [35] B. Bahrani, M. Saeedifard, A. Karimi, and A. Rufer, “A multivariable design methodology for voltage control of a single-DG-unit microgrid,” IEEE Trans. Ind. Informat., vol. 9, no. 2, pp. 589–599, May 2013. [36] Z. Wang and H. Li, “Optimized operating mode of current-fed dual half bridges dc-dc converters for energy storage applications,” in Proc. IEEE ECCE, Sep. 20–24, 2009, pp. 731–737. [37] H. Daneshpajooh, A. Bakhshai, and P. Jain, “Optimizing dual half bridge converter for full range soft switching and high efficiency,” in Proc. IEEE ECCE, Sep. 17–22, 2011, pp. 1296–1301. [38] X. Liu, H. Li, and Z. Wang, “A fuel cell power conditioning system with low-frequency ripple-free input current using a control-oriented power pulsation decoupling strategy,” IEEE Trans. Power Electron., vol. 29, no. 1, pp. 159–169, Jan. 2014. [39] Powerizer Product Catalog. [Online]. Available: www.batteryspace.com

[40] M. Vasiladiotis, N. Cherix, and A. Rufer, “Accurate capacitor voltage ripple estimation and current control considerations for grid-connected modular multilevel converters,” IEEE Trans. Power Electron., vol. 29, no. 9, pp. 4568–4579, Sep. 2014. [41] N. Cherix, S. Delalay, P. Barrade, and A. Rufer, “Fail-safe modular control platform for power electronic applications in R&D environments,” in Proc. 15th EPE Conf. Appl., Sep. 2–6, 2013, pp. 1–10.

Michail Vasiladiotis (S’09) was born in Athens, Greece, in 1986. He received the Diploma in electrical and computer engineering from the National Technical University of Athens, Athens, Greece, in 2009. Since 2010, he has been working toward the Ph.D. degree in the Laboratory of Industrial Electronics, École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland. His research interests include multilevel conversion systems for high-power applications, modern control methods for power converters, as well as power electronic interfaces for ultrafast EV charging.

Alfred Rufer (M’95–SM’01–F’06) received the M.S. degree from the Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland, in 1976. In 1978, he joined ABB, Turgi, Switzerland, where he was involved in the fields of power electronics and control, such as high-power variable-frequency converters for drives, and where he was a Group Leader involved with power electronic development beginning in 1985. In 1993, he became an Assistant Professor at EPFL, where, since 1996, he has been a Full Professor and the Head of the Industrial Electronics Laboratory (LEI). LEI is active in power electronics used in energy conversion and energy storage and in the modeling and simulation of systems, including control strategies and control circuits. He has authored or coauthored many publications on power electronics and applications, such as multilevel converters and various energy-storage systems. He is the holder of several patents.