IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 61, NO.

3, MARCH 2012

1467

TABLE I C OMPARISON B ETWEEN OSTBC/GSC AND OSTBC/MS-GSC IN T ERMS OF P OWER S AVINGS AND BER W ITH Nt = 2, L = 5, AND Lc = 3

[11] V. Tarokh, H. Jafarkhani, and A. R. Calderbank, Space-time block codes from orthogonal designs, IEEE Trans. Inf. Theory, vol. 45, no. 5, pp. 14561467, Jul. 1999. [12] C. Xu and K. S. Kwak, On decoding algorithm and performance of space-time block codes, IEEE Trans. Wireless Commun., vol. 4, no. 3, pp. 825829, May 2005. [13] S. M. Alamouti, A simple transmit diversity technique for wireless communications, IEEE J. Sel. Areas Commun., vol. 16, no. 8, pp. 14511458, Oct. 1998. [14] W. Li, C. Beaulieu, and Y. Chen, Generalized receiver selection combining schemes for Alamouti MIMO systems with MPSK, IEEE Trans. Commun., vol. 57, no. 6, pp. 15991602, Jun. 2009. [15] I. S. Gradshteyn and I. M. Ryzhik, Tables of Integrals, Series and Products, 7th ed. New York: Academic, 2007.

Fig. 3. SER performance for OSTBC/MS-GSC with L = 3, Lc = 2, = 1 b/s/Hz, and T = 5 dB.

Control of a Three-Level Boost Converter Based on a Differential Flatness Approach for Fuel Cell Vehicle Applications
Phatiphat Thounthong, Member, IEEE

OSTBC/GSC system, with the extra incentive of reduced complexity and power savings at the receiver when the predetermined SNR threshold T is optimized. R EFERENCES
[1] M.-S. Alouini and M. K. Simon, Digital Communication Over Fading Channels, 2nd ed. New York: Wiley, 2005. [2] M.-S. Alouini and M. K. Simon, An MGF-based performance analysis of generalized selection combining over Rayleigh fading channels, IEEE Trans. Commun., vol. 48, no. 3, pp. 401415, Mar. 2000. [3] M.-S. Alouini and M. K. Simon, Performance of coherent receivers with hybrid SC/MRC over Nakagami- m fading channels, IEEE Trans. Veh. Technol., vol. 48, no. 4, pp. 11551164, Jul. 1999. [4] B. S. Tan, K. H. Li, and K. C. Teh, Performance analysis of LDPC codes with maximum-ratio combining cascaded with selection combining over Nakagami-m fading, IEEE Trans. Wireless Commun., vol. 10, no. 6, pp. 18861894, Jun. 2011. [5] M. K. Simon and M.-S. Alouini, Performance analysis of generalized selection combining with threshold test per branch (T-GSC), IEEE Trans. Veh. Technol., vol. 51, no. 5, pp. 10181029, Sep. 2002. [6] Y. Ma, X. Dong, and H.-C. Yang, Asymptotic performance of thresholdbased generalized selection combining, IEEE Trans. Veh. Technol., vol. 58, no. 5, pp. 25792585, Jun. 2009. [7] S. W. Kim, D. S. Ha, and J. H. Reed, Minimum selection GSC and adaptive low-power RAKE combining scheme, in Proc. IEEE ISCAS, vol. 4, pp. IV-357IV-360. [8] H.-C. Yang, New results on ordered statistics and analysis of minimumselection generalized selection combining (GSC), IEEE Trans. Wireless Commun., vol. 5, no. 7, pp. 18761885, Jul. 2006. [9] M.-A. Daghfous, R. M. Radaydeh, and M.-S. Alouini, Performance of adaptive MS-GSC in the presence of cochannel interference, IEEE Trans. Veh. Technol., vol. 60, no. 6, pp. 28292837, Jul. 2011. [10] T. M. Duman and A. Ghrayeb, Coding for MIMO Communication Systems. West Sussex, U.K.: Wiley, 2007.

AbstractThis paper presents a high-gain boost converter (three-level converter and transformerless converter) for fuel cell (FC) vehicle applications. An original nonlinear control law based on the atness principle for distributed dc generation is studied. Utilizing the atness property, we propose simple solutions to the system performance and stabilization problems. The design controller parameters are straightforward and autonomous at the operating point. To validate the proposed method, a prototype FC power converter (1.2-kW three-level boost converter) is developed in the laboratory. The proposed control law is implemented with a digital estimate in a dSPACE 1104 controller card. The experimental results from the laboratory using a 1200-W and 46-A polymer electrolyte membrane FC (PEMFC) verify that this is a good control scheme. Index TermsConverters, atness-based control, fuel cells (FCs), nonlinear, power control.

Manuscript received July 13, 2011; revised October 22, 2011; accepted January 1, 2012. Date of publication January 10, 2012; date of current version March 21, 2012. This work was supported in part by a research program in cooperation with King Mongkuts University of Technology North Bangkok with the Institut National Polytechnique de Lorraine (Universit de Lorraine) under the Franco-Thai Cooperation Program in Higher Education and Research Joint Project and in part by the Thai-French Innovation Institute and the Commission on Higher Education, Ministry of Education, Thailand, under Grant 37485. The review of this paper was coordinated by Prof. L. Guvenc. The author is with the Department of Teacher Training in Electrical Engineering and the Renewable Energy Research Centre, King Mongkuts University of Technology North Bangkok, Bangkok 10800, Thailand (e-mail: phtt@ kmutnb.ac.th). 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/TVT.2012.2183628

0018-9545/$31.00 2012 IEEE

1468

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 61, NO. 3, MARCH 2012

I. I NTRODUCTION Energy and environmental issues are becoming a serious issue in the world; accordingly, society must consider renewable- and cleanenergy sources. One potential alternative source of electric power is the fuel cell (FC). FC power generation systems are expected to be employed in a growing number of areas, including portable applications [1], transportation applications [2], [3], and stationary power applications [4]. FCs produce a low dc voltage, and they are always connected to electric power networks through a step-up (boost) converter [5]. Theoretically, conventional boost converters are able to achieve high step-up voltage gain under heavy-duty load conditions. In practice, however, the voltage gain of the boost converter is limited because of the losses associated with the inductor, lter capacitor, main power switch, and rectier diode [6]. In many applications, the use of an isolation transformer can provide an increased output/input voltage conversion ratio because the required and full-bridge topologies can be used. However, there are applications where transformerless energy converter systems could potentially offer signicant advantages, including simplicity, cost, and converter size reduction, particularly in high-powered applications [6]. Recently, multilevel converters (transformerless) have been studied for high-power applications due to their ability to meet the increasing demand for power ratings and higher efciencies when compared with conventional topologies. In particular, multilevel converters for ac/dc and dc/ac conversions [see Fig. 1(a)] have been developed for many years [7][9]. Since Nabae et al. [7] proposed the topology neutralpoint-clamped inverter in 1981, the advantages of multilevel converters have become well known. Later, Portillo et al. [8] proposed a threelevel back-to-back power converter for three-phase wind turbine applications, and Freitas et al. [9] studied a three-level three-leg acdcac converter for single-phase applications. In this paper, a three-level step-up converter that was adapted and modied from the three-level ac/dc converter [transformerless; see Fig. 1(b)] is proposed to solve the problems associated with the aforementioned modied boost converters. The most common method of controlling dc/dc chopper converters still relies on averaged smallsignal models and then the application of linear control theory; the proportionalintegral (PI) controller is an example of this method [8], [10]. The power-electronic converter exhibits nonlinear behavior. To control the voltage, the current, or the power in the converter, a linearized technique is often used to study the stability and to select the controller parameters of the nonlinear converter. In this paper, a nonlinear control algorithm based on the atness property of the system is proposed. Flatness provides a convenient framework for meeting a number of performance specications for the power converter [11]. This paper is organized as follows: The proposed power converter circuit, a brief description of the operating principle, and the average model are presented in Section II. Section III discusses the proof of the differential atness of the proposed FC converter models and the control law and stability. Experimental results are used to validate the proposed control system in Section IV. Finally, this paper ends with concluding remarks in Section V. II. T HREE -L EVEL B OOST C ONVERTER A. Power Circuit FC voltages are not only very low, but they also vary over a wide range. FC power sources are always linked to networks through a boost converter [5]. The input lter inductance of classical boost converters must be large enough to reduce the FC current ripple, resulting in a low power density and a slow transient response.

Fig. 1. Three-level converters. (a) ACDC and dcac three-level converter. (b) Proposed dcdc three-level boost converter and its switching functions for FC applications, where vFC and iFC are the FC voltage and current, respectively. iLoad is the load current. vBus the dc bus voltage.

Remark 1: The constraints to operate an FC are given here. 1) The FC power or current must be kept within an interval (rated value, minimum value, or zero). 2) The FC current must be controlled as a unidirectional current. 3) The FC current slope must be limited to a maximum absolute value to prevent the fuel starvation phenomenon from occurring in an FC stack [6]. 4) The FC ripple current must be small to ensure that it has a minor impact on the FC conditions [6]. A classical boost converter connected with an FC will be limited when the power increases or for higher voltage step-up ratios. To increase the voltage conversion ratio, Grbovi et al. [12] and Shahin et al. c [13] proposed a three-level boost converter (transformerless power converter) for a low-voltage power source, as shown in Fig. 1(b). FC current ripple reduction is a major issue for FC converter design [6]. The proposed conguration was chosen, instead of a conventional boost converter with the smaller inductor needed to achieve comparably low ripple [12]. In addition, it reduces the required semiconductor device voltage rating by a factor of two.

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 61, NO. 3, MARCH 2012

1469

The startup time of an FC system is typically more than 2 min. The startup sequence begins with the water-circulating pump wetting the FC, the solenoid valve opening to supply fuel (hydrogen), and the blower starting to feed reaction air. Once the open circuit voltage of the FC is veried, the converter can be connected to the FC. This startup sequence takes approximately 2 min to complete. To avoid startup problems when the converter is connected to the FC because a high inrush current ows through inductor L, diode D1 , capacitor C1 , capacitor C2 , and diode D4 , a soft-start circuit is proposed, as shown in Fig. 1(b). In comparison with a conventional boost converter, this circuit involves more devices, including four power switches (S1 , S2 , S3 , and S4 ) and two output capacitors (C1 and C2 ). By cascading the capacitor voltage vC1 and the capacitor voltage vC2 , a high output voltage vBus is easily achieved. In addition, there is low voltage stress on the power switch, diode, and output capacitors in comparison with conventional boost converters. Power switches with low voltage ratings are used to reduce the conduction loss, and as a result, the overall efciency is signicantly improved. This allows us to use lower-voltage-rated switches with better switching and conduction performance than the switches rated for the full blocking voltage. Therefore, the converters overall performance, including the cost and efciency, can be signicantly better than two-level converters, particularly when the switching frequency is above 20 kHz or metaloxidesemiconductor eld-effect transistors (MOSFETs) are used. B. Operation Principle The switching functions are generated by pulsewidth modulators PWM1 and PWM2 , as shown in Fig. 2, where TS is the switching period. The modulation signals d1 and d2 (> 0.5) are duty cycles generated by a control circuit (the current control or power control loop presented hereafter). The carrier signals are triangular signals vCar1 and vCar2 shifted by (called the interleaving technique) and running at the same switching frequency fS (= 1/TS ). The proposed converter operating in continuous conduction mode (CCM) is considered. According to Remark 1, the FC must be operated in CCM with a small ripple current to avoid using an input L-C lter. The switching sequences and key waveforms of the converter in CCM are shown in Fig. 2. The output capacitors C1 and C2 are assumed to be large enough so that the output voltages vC1 and vC2 across them are assumed to be constant during the entire switching cycle and become vBus = vC1 + vC2 . C. Average Model From the average model of a traditional boost converter [14], the circuit can be described by the following set of equations: dFC i = vFC rL FC (1d1 ) vC1 (1d2 ) vC2 i dt dC1 v = (1 d1 ) FC iLoad i C1 dt dC2 v = (1 d2 ) FC iLoad i C2 dt L (1) (2) (3)
Fig. 2. Key waveforms and operational modes of the three-level boost converter in CCM at d1 = d2 = 0.8.

The input power pFC is given versus vFC and iFC by the following differential equation: i pFC = vFC FC dFC p dFC i = vFC dt dt .
vFC =Constant

(5) (6)

where rL is the series resistance of inductor L. Note that rL also represents the static losses in the converter. The lter capacitors are assumed to be identical, i.e., C1 = C2 = C, d1 = d2 = d, and vC1 = vC2 = vBus /2. From (1), the differential equation can be written as follows: L dFC i = vFC rL FC (1 d) vBus . i dt (4)

III. P OWER C ONTROL L OOP A. Flatness of the Proposed Converter Based on the atness principle introduced in [11] and [13], the FC power pFC is assumed to be the at output component. Thus, we dene

1470

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 61, NO. 3, MARCH 2012

a at output as y = pFC , a control input variable as u = d, and a state variable as x = iFC . State variable x can be written as x= pFC = (y). vFC (7)

From (4) and (6), control variable u can be calculated from the at output y and its time derivative y, as follows: u = 1+ y L 1 +rL iFC vFC = d = (y). vFC vBus (8)
Fig. 3. Proposed FC power control scheme for the three-level converter.

It is clear that x = (y) and that u = (y). Consequently, the mathematical model of the converter can be assumed to be a at system [11], [13]. B. Control Law and Stability The input power reference for the converter is represented by yREF (= pFCREF ). A linearizing feedback control law that provides exponential asymptotic tracking of the trajectory is given by the following expression [15], [16]: (y yREF )+K11 (yyREF )+K12 (yyREF )d = 0 (9)

where K11 and K12 are the controller parameters. Then y = yREF + K11 (yREF y) + K12 (yREF y)d. (10)

Inserting the term for y from (10) into (8) gives the equation for the closed-loop static-state feedback duty cycle d (called the inverse dynamic equation). Evidently, the control system is stable for K11 , K12 > 0. The FC power is measured by using (5), which is associated with a rst-order lter that is used to reduce the harmonics caused by power-electronic switching. Based on the power-electronic constant switching frequency S (PWM) and the cascade control structure, the outer control loop (here, the input power control) must operate at F (the cutoff frequency of the rst-order a cutoff frequency P lter) S . Once the at outputs are stabilized, the whole system is exponentially stable because all the variables of the system are expressed in terms of the at outputs [17]. C. Balance Voltage in the Output Capacitors Voltage-balancing controller functions are used to eliminate the difference between voltages vC1 and vC2 . Duty cycles d1 and d2 are determined as follows: d1 = d d2 = d + (11)

Fig. 4.

Test bench system.

D. Conclusion of the Control Algorithm The FC power reference pFCREF must be restricted to an interval with a maximum of pFCMax (corresponding to the rated power of the FC) and a minimum of pFCMin (set to 0 W) and be limited in dynamics; these limitations ensure safe operation of the FC with respect to the constraints that are associated with the FC (i.e., the prevention of fuel starvation in an FC stack [18][20]). Thus, to limit transient FC power, a low-pass lter (second order) is employed such that the power demand pFCDEM from the external loop is always limited (desired trajectory planning) as follows: pFCREF (t) = pFCDEM (t) 1 e
t
1

where d is the main duty cycle generated by the power controller [see (8)], and represents a correction term, which is used to balance the voltages. A proportional controller is sufcient as long as the gain is high enough to introduce only a small static error. Therefore, voltagebalancing control can be expressed by the following function: = Kv vBus vC2 2 (12)

t t e 1 1

(13)

where 1 is the control parameter. The nonlinear control law based on the differential atness of the FC power as previously detailed is shown in Fig. 3. IV. E XPERIMENTAL VALIDATION A small-scale test bench was implemented in the laboratory, as shown in Fig. 4. The parameters of the three-level boost converter and the semiconductor components are described in this section. Inductor

where KV is the controller parameter.

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 61, NO. 3, MARCH 2012

1471

Fig. 5. FC power control characteristics of the prototype system at pFCREF = 500 W, vBus = 120 V, vC1 = 60 V, vFC = 34.0 V, and iFC = 14.7 A.

Fig. 6. Converter response during the FC power reference trajectory increases from 200 to 500 W.

L is obtained by means of a ferrite core; its inductance is 223 H, and its series resistance RL1 is 0.14 . Capacitors C1 and C2 are obtained by means of electrolytic capacitors with capacitances of 1000 F each. Switches S1 , S2 , S3 , and S4 are IRFP2640N power MOSFET transistors (250 V, 38 A). The FC system used in this study was a PEMFC system (1.2 kW, 46 A from Ballard Power Systems, Inc.), as shown in Fig. 4. The system was supplied with pure hydrogen from an aluminum tank at a pressure of 150 bar and clean and dry air from a compressor. The FC power control loop, the voltage-balancing control, and the PWM generation algorithm were implemented in the mathematical environment of MATLAB-Simulink using a real-time dSPACE DS1104 card (see Fig. 4), the fourth-order Runge-Kutta integration algorithm, and a sampling time of 20 s. The switching frequency fS was set to 10 kHz. Measurements of the FC current iFC , the FC voltage vFC , the dc bus voltage vBus , and the output voltage vC2 were collected with zero-ux Hall effect sensors. The nonlinear controller gains used were K11 = 141.4 rd s1 and K12 = 10 000 rd2 s2 . As a result, the cutoff frequency P of the closed-loop FC power was equal to 100 rd s1 , which is lower than the cutoff frequency F of the measured power lter of 1000 rd s1 (TF = 1 ms), and therefore, the system was exponentially stable. The voltage-balancing controller gain was KV = 0.6. Parameter 1 for the FC power dynamic (desired trajectory) was 5 s. This value has experimentally been found to be the highest power slope of our FC system at which no fuel starvation occurs. To illustrate the switching characteristics of the proposed converter, the oscilloscope waveforms in Fig. 5 show the steady-state waveforms for FC power regulation at the reference power of 500 W. This gure shows the FC voltage vFC (CH1), the FC current iFC (CH2), the gate drive signal for S2 vGS2 (CH3), and the gate drive signal for S3 vGS3 (CH4). The power-switching devices S2 and S3 (individual interleaved power stages) operate at 10 kHz, which is the same frequency as the classical boost converter design, but the effective input FC current ripple frequency is 20 kHz; as a result, the FC ripple current is reduced. The waveforms that were obtained as the FC power reference trajectory pFCREF increased are shown in Fig. 6. Data show the FC power reference trajectory pFCREF (CH1), the measured FC power pFC (CH2), the FC voltage vFC (CH3), and the FC current iFC (CH4). In the initial state, the FC reference power is equal to 200 W, the

measured FC power is equal to 200 W, the FC voltage is equal to 36.4 V, and the FC current is equal to 5.5 A. At t = 5 s, the FC power reference trajectory slowly increases to a nal constant power of 500 W. The measured FC power perfectly follows its reference, and there is no fuel starvation phenomenon in the FC system because the dynamics of the FC power reference trajectory are limited, as described in (13). Finally, the waveforms obtained as the FC power reference trajectory pFCREF decreased are shown in Fig. 7. Data show the dc bus voltage vBus , the output capacitor voltage vC1 , the output capacitor voltage vC2 , the FC power reference trajectory pFCREF , the measured FC power pFC , the FC voltage vFC , and the FC current iFC . In the initial state, the FC power reference is equal to 500 W, the measured FC power is equal to 500 W, the FC voltage is equal to 33.3 V, the FC current is equal to 15.0 A, the dc bus voltage is equal to 118 V, and the output capacitor voltages (vC1 and vC2 ) are equal to 59.0 V. At t = 10 s, the FC power reference trajectory slowly decreases to a nal constant power of 200 W. Again, the measured FC power perfectly follows its reference. It shows the behavior of the balancing system associated with the power reference variation for the transient state. This result shows that the voltage-balancing control is still efcient for a transient state. V. C ONCLUSION A step-up three-level converter topology (transformerless and of high gain) with a new control approach based on the differential atness theory has been proposed for FC vehicle applications. The proposed dc/dc converter is adapted and modied from a multilevel ac-dc converter studied by Nabae et al. [7] of Japan in 1981. The advantages of a multilevel converter are well known and include increasing the magnitude of the output voltage and reducing the input inductor size (because the input switching frequency increases). In this respect, it permits us to use the same type of switches for double the voltage. These power topologies synthesize the voltage waveform using a number of semiconductor devices connected in a special arrangement and rated at a fraction of the dc bus voltage. A multilevel converter allows the dc bus voltage to be increased without requiring a high-frequency transformer. The main contributions of this paper are the proposal of a three-level converter for FC applications and the presentation of the differential

1472

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 61, NO. 3, MARCH 2012

R EFERENCES
[1] M. Guarnieri, V. Di Noto, and F. Moro, A dynamic circuit model of a small direct methanol fuel cell for portable electronic devices, IEEE Trans. Ind. Electron., vol. 57, no. 6, pp. 18651873, Jun. 2010. [2] A. Khaligh and Z. Li, Battery, ultracapacitor, fuel Cell, and hybrid energy storage systems for electric, hybrid electric, fuel cell, and plug-in hybrid electric vehicles: State of the art, IEEE Trans. Veh. Technol., vol. 59, no. 6, pp. 28062814, Jul. 2010. [3] C. C. Chan, A. Bouscayrol, and K. Chen, Electric, hybrid, and fuel-cell vehicles: Architectures and modeling, IEEE Trans. Veh. Technol., vol. 59, no. 2, pp. 589598, Feb. 2010. [4] U. S. Selamogullari, D. A. Torrey, and S. Salon, A systems approach for a stand-alone residential fuel cell power inverter design, IEEE Trans. Energy Convers., vol. 25, no. 3, pp. 741749, Sep. 2010. [5] W. Na, T. Park, T. Kim, and S. Kwak, Light fuel-cell hybrid electric vehicles based on predictive controllers, IEEE Trans. Veh. Technol., vol. 60, no. 1, pp. 8997, Jan. 2011. [6] P. Thounthong and B. Davat, Study of a multiphase interleaved step up converter for fuel cell high power applications, Energy Convers. Manage., vol. 51, no. 4, pp. 826832, Apr. 2010. [7] A. Nabae, I. Takahashi, and H. Akagi, A neutral-point clamped PWM inverter, IEEE Trans. Ind. Appl., vol. IA-17, no. 5, pp. 518523, Sep. 1981. [8] R. Portillo, M. M. Prats, J. I. Len, J. A. Snchez, J. M. Carrasco, E. Galvn, and L. G. Franquelo, Modeling strategy for back-to-back three-level converters applied to high-power wind turbines, IEEE Trans. Ind. Electron., vol. 53, no. 5, pp. 14831491, Oct. 2006. [9] I. S. de Freitas, C. B. Jacobina, E. R. C. da Silva, and T. M. Oliveira, Single-phase ACDCAC three-level three-leg converter, IEEE Trans. Ind. Electron., vol. 57, no. 12, pp. 40754084, Dec. 2010. [10] E. Tedeschi, M. Carraro, M. Molinas, and P. Mattavelli, Effect of control strategies and power take-off efciency on the power capture from sea waves, IEEE Trans. Energy Convers., vol. 26, no. 4, pp. 10881098, Dec. 2011. [11] M. Zandi, A. Payman, J. P. Martin, S. Pierfederici, B. Davat, and F. Meibody-Tabar, Energy management of a fuel cell/supercapacitor/ battery power source for electric vehicular applications, IEEE Trans. Veh. Technol., vol. 60, no. 2, pp. 433443, Feb. 2011. [12] P. J. Grbovi , P. Delarue, P. Le Moigne, and P. Bartholomeus, A bidirecc tional three-level DC-DC converter for the ultracapacitor applications, IEEE Trans. Ind. Electron., vol. 57, no. 10, pp. 34153430, Oct. 2010. [13] A. Shahin, M. Hinaje, J. P. Martin, S. Pierfederici, S. Ral, and B. Davat, High voltage ratio DCDC converter for fuel-cell applications, IEEE Trans. Ind. Electron., vol. 57, no. 12, pp. 39443955, Dec. 2010. [14] A. M. Rahimi, G. A. Williamson, and A. Emadi, Loop-cancellation technique: A novel nonlinear feedback to overcome the destabilizing effect of constant-power loads, IEEE Trans. Veh. Technol., vol. 59, no. 2, pp. 650 661, Feb. 2010. [15] P. Thounthong, Model based-energy control of a solar power plant with a supercapacitor for grid-independent applications, IEEE Trans. Energy Convers., vol. 26, no. 4, pp. 12101218, Dec. 2011. [16] P. Thounthong, V. Chunkag, P. Sethakul, S. Sikkabut, S. Pierfederici, and B. Davat, Energy management of fuel cell/solar cell/supercapacitor hybrid power source, J. Power Sources, vol. 196, no. 1, pp. 313324, Jan. 2011. [17] P. Thounthong, S. Pierfederici, and B. Davat, Analysis of differential atness-based control for a fuel cell hybrid power source, IEEE Trans. Energy Convers., vol. 25, no. 3, pp. 909920, Sep. 2010. [18] M. Nayeripour, M. Hoseintabar, and T. Niknam, A new method for dynamic performance improvement of a hybrid power system by coordination of converters controller, J. Power Sources, vol. 196, no. 8, pp. 40334043, Apr. 2011. [19] G. Seenumani, H. Peng, and J. Sun, A reference governor-based hierarchical control for failure mode power management of hybrid power systems for all-electric ships, J. Power Sources, vol. 196, no. 8, pp. 1599 1607, Feb. 2011. [20] N. Bizon, A new topology of fuel cell hybrid power source for efcient operation and high reliability, J. Power Sources, vol. 196, no. 6, pp. 32603270, Mar. 2011.

Fig. 7. Converter response during the FC power reference trajectory decreases from 500 to 200 W.

atness-based control of an FC power source. Using the nonlinear control approach based on the atness property, we have proposed a simple solution to the optimization and stabilization problems in a nonlinear power-electronic system. This is an original concept for this type of application. A prototype FC converter of 1.2 kW has been designed, and the solution has been experimentally veried in the laboratory. The prototype has been tested, and the experimental results have been presented and discussed. The results demonstrate good agreement with the theoretical analysis provided in this paper.

ACKNOWLEDGMENT The author would like to thank S. Lekapat (Administrator, the ThaiFrench Innovation Institute), who is in charge of the process of the Franco-Thai on Higher Education and Research Joint Project and of the research funding for this paper, and Dr. P. Sethakul (KMUTNB), Dr. S. Pierfederici, and Dr. B. Davat (Universit de Lorraine) for their valuable comments and suggestions about power electronics and control.