grid Connected Fuel Cell

8th International Conference on Power Electronics - ECCE Asia

May 30-June 3, 2011, The Shilla Jeju, Korea

978-1-61284-957-7/11/$26.00 ©2011 IEEE

[ThB2-5]

Abstract-- In general, two power conversion stages are required when low-voltage unregulated fuel cell (FC) output is conditioned to generate AC power. In this paper, the boost-inverter topology that achieves both boosting and inversion functions in a single-stage is used as a building block to develop a single-phase grid-connected FC-system which offers high conversion efficiency, low-cost and compactness. The proposed system incorporates additional battery-based energy storage and a DC-DC bi-directional converter to support the slow dynamics of the FC. The single-phase boost-inverter is voltage-mode controlled and the DC-DC bi-directional converter is current-mode controlled. The low-frequency current ripple is supplied by the battery which minimizes the effects of such ripple being drawn directly from the FC itself. Moreover, this system can operate either in a grid-connected or stand-alone mode. In the grid-connected mode, the boost-inverter is able to control the active (P) and reactive (Q) power using an algorithm based on a Second Order Generalized Integrator (SOGI) which provides a fast signal conditioning for single-phase systems. Analysis, simulation and experimental results from a laboratory prototype are presented to confirm the validity of the proposed system.

Index Terms-- Boost-inverter, fuel cell, grid-connected inverter, power conditioning system (PCS), PQ control.

I. INTRODUCTION Alternative energy sources based on photovoltaic (PV)

and fuel cell (FC) systems need to be conditioned which may include energy conversion or energy storage. Especially, FC system must be supported through additional energy storage unit to achieve high quality supply of power [1]-[3]. When such systems are used to power AC loads or to be connected with the electricity grid, an inversion stage is also required.

The typical output voltage of low power FC is low and variable with respect to the load current. For instance, based on the current-voltage characteristics of a 72-cell Proton Exchange Membrane FC (PEMFC) power module, the voltage varies between 39 to 69 VDC depending upon the level of the output current in Fig. 1 [4]. Moreover, the slow dynamics due to the natural electrochemical reactions required for the balance of enthalpy must be taken into account when designing the

FC system [5], [6]. This is crucial, especially when the power drawn from the FC exceeds the maximum permissible power, as in this case, the FC module may not only fail to supply the required power to the load but also cease to operate or be damaged [7]-[9]. Therefore, the power converter needs to ensure that the required power remains within the limit of the maximum availability [8], [9].

A two-stage fuel cell power conditioning system to deliver AC power has been commonly considered and studied in numerous technical papers [2], [3], [7]-[12]. A system with reduced conversion stages is proposed in [13]. The double loop control scheme of this topology has also been proposed for better performance even during transient conditions [14]. The two-stage FC power conditioning system encounters drawbacks such as being bulky, costly and inefficient due to its cascaded power stages. In order to alleviate these drawbacks, a topology that is suitable for AC loads and is powered from DC sources able to boost and invert the voltage at the same time has been proposed in [15].

The objective of this paper is to propose and report selected experimental results of a grid-connected FC system using only a single energy conversion stage. In particular, the proposed system based on the boost-inverter with a back-up energy storage unit, solves the

Grid-Connected Fuel Cell System Based on a Boost-Inverter with a Battery Back-Up Unit

M. Jang, M. Ciobotaru, and V. G. Agelidis School of Electrical Engineering and Telecommunications

The University of New South Wales Sydney, NSW 2052, Australia

This work is supported by an Australian Postgraduate Award from the Australian Government and an Engineering Research Award from the University of New South Wales.

Fig. 1. The characteristics of the 72-cell PEMFC system (Horizon H-1000, 1.0kW): voltage-current and power-current characteristics. At rated power output voltage and current are 43VDC and 23.5A respectively. The FC system output voltage ranges from 39 VDC to 69 VDC.

previously mentioned issues (e.g. the low and variable output voltage of the FC, its slow dynamics, and current harmonics on the FC side). This single-stage including boosting and inversion functions provides a high power conversion efficiency, reduced converter size and low cost. The single-phase grid-connected FC system can operate either in grid-connected or stand-alone mode. The boost-inverter is also able to control the active (P) and reactive (Q) power in the grid-connected mode [16].

The paper is organized as follows. In Section II, the proposed grid-connected FC system is introduced including the converter topology, the control algorithm of the boost-inverter and the back-up unit. The P and Q control algorithm for a single-phase FC system is also illustrated in Section II. In Section III, analysis and simulation results are presented to verify the performance of the system. Experimental results are presented from 1kW laboratory prototype in Section IV. Finally, the conclusions are presented in Section V.

II. PROPOSED FULL CELL ENERGY SYSTEM

A. Description of the FC system The proposed grid-connected FC system consists of

two power converters: the boost-inverter and the bi-directional back-up unit as shown in Fig. 2(a). Fig. 2(b) shows the laboratory setup for the proposed FC system. The boost-inverter is supplied by the FC and the back-up unit, which are both connected to the same unregulated DC bus while the output side is connected to the load and grid through an inductor. The system incorporates a current mode controlled bi-directional boost converter with battery-based energy storage to support the FC power generation and a voltage controlled boost-inverter providing both boosting and inversion functions simultaneously.

The FC system should dynamically adjust to varying input voltage while maintaining constant power operation. Voltage and current limits, which should be provided from FC stack manufacturers, need to be imposed at the input of the converter to protect the FC from damage due to excessive loading and transients. Moreover, the power has to be ramped up and down so that the FC can react appropriately, avoiding transients and extending its life time. The converter also has to meet

the maximum ripple current requirements of the FC [5]. In grid-connected mode, the system is also providing

active and reactive power control. A key concept of the PQ control in the inductive coupled voltage sources is the use of a grid compatible frequency and voltage droops [16]. Therefore, the active and reactive power is controlled by the small variations of the voltage phase and magnitude. The control of the inverter requires a fast signal conditioning for single-phase systems based on a Second Order Generalized Integrator (SOGI) algorithm [16].

B. Boost-inverter The boost-inverter consists of two bi-directional boost

converters and their outputs are connected in series as shown in Fig. 2(a). Each boost converter generates a DC bias with deliberate AC output voltage (a DC-biased sinusoidal waveform as an output), so that each converter generates a unipolar voltage greater than the FC voltage with a variable duty cycle. Each converter output and the combined outputs are described by

sin21

11 AVV dc (1)

)180sin(21

22 AVV dc (2)

2121 22,sin AAAwhenAVVV ooo (3)

2o

indcA

VV

where Vdc is the DC offset voltage of each boost converter and have to be greater than 0.5Ao+Vin.

From (3) it can be observed that the output voltage Vo contains only the AC component. This concept has been discussed in numerous papers [13], [14]. The boost-inverter is based on the voltage mode control.

In this paper, a double-loop control scheme is chosen for the boost-inverter control being the most appropriate method to control the individual boost converters covering the wide range of operating points. This control method is based on the averaged continuous-time model of the boost topology and has several advantages with special conditions that may not be provided by the sliding mode control, such as nonlinear loads, abrupt load

(a) (b) Fig. 2. Grid-connected FC system. (a) General structure of the system. (b) Laboratory prototype.

variations and transient short circuit situations. Using this control method the inverter maintains a stable operating condition by means of limiting the inductor current. Because of this ability to keep the system under control even in these situations, the inverter achieves a very reliable operation [14].

The reference voltage of the boost-inverter is provided from the PQ control algorithm being able to control the active and reactive power. The voltages across C1 and C2 are controlled to track the voltage references using Proportional-Resonant (PR) controllers. Compared with the conventional PI controller, the PR controller has the ability to minimize the drawbacks of the PI controller such as lack of tracking a sinusoidal reference with zero steady-state error and poor disturbance rejection capability [18], [19].

The currents through L1 and L2 are controlled by PR controllers to achieve a stable operation in special conditions such as nonlinear loads and transients.

The control block diagram for the boost-inverter is shown in Fig. 3. The output voltage reference is divided to generate the two individual output voltage references of the two boost converters with the DC-bias (Vdc). The DC-bias can be obtained by adding the input voltage (Vin) to the half of the peak output amplitude. Vdc is also used to minimize the output voltages of the converters and the switching losses in the variable input voltage condition.

The output voltage reference is determined by

)sin()(. tdVVV opppprefo .

tanddVVAwhen oppppo

where Vpp is the peak value of the typical grid voltage, dVpp is a small variation of the output voltage reference

affecting to the reactive power, o is grid fundamental angular frequency, and is phase difference between Vo and Eg relating with the active power. Then V1.ref and V2.ref are calculated by (1) and (2). The PR controller in Fig. 4 is presented and is described by [18], [19]

22 22)(

oc

cips ss

skksG

(6)

where kp and ki are the proportional and integral gains of PR controller respectively, and c is cut-off angular frequency. The transfer function in (6) shows a non ideal PR controller which has lower gain and wider bandwidth than ideal PR controller at the resonant frequency.

C. Backup unit The functions of the back-up unit are divided into two

parts. Firstly, the back-up unit is designed to support the slow response of the FC. Secondly, in order to protect the FC system, the back-up unit provides low-frequency AC

Fig. 5. Backup unit control block diagram

Fig. 3. Boost inverter control block diagram

Fig. 4. PR controller block diagram including cutoff frequency

current that is required from the boost-inverter operation. The back-up unit comprises a current-mode controlled bi-directional boost converter and a battery as the energy storage unit. For instance, when a 1kW load is added from a no-load condition, the back-up unit immediately provides the 1kW power from the battery to the load as shown in Table I. On the other hand, when the load is disconnected suddenly, the surplus power from the FC could be recovered and stored into the battery to increase the overall efficiency of the energy system. Two generic 12 V lead-acid batteries are introduced in this unit for energy storage to deal with the need to provide fast response and a relatively low cost solution.

The back-up unit controller is designed to control the output current of the back-up unit in Fig. 5. The reference of ILb1 is determined by Idc through a high-pass filter and the demanded current Idemand that is relating the load change. The AC component of the current reference is dealing with eliminating AC ripple current into the FC power module while the DC component is dealing with slow dynamics of the FC.

D. Control of the grid-connected boost-inverter Fig. 6 illustrates the equivalent circuit of the grid-

connected FC energy system consisting of two AC sources, AC inductor between the two AC sources and the load. The boost-inverter output voltage (including the FC and back-up unit) is indicated as Vo and Eg is the grid voltage. The active and reactive powers at the Point of Common Coupling (PCC) are expressed by [16], [17].

)sin()( gfo

og

LLVE

P

(7)

)cos()()(

2

gfo

og

gfo

g

LLVE

LLE

Q

(8)

where Lf is the filter inductance and Lg is the grid inductance.

From (7) and (8) the phase shift between Vo and Eg causes the active power flow and the reactive power is due to voltage differences, Eg-Vo. Therefore, to control the power flow between the converter and the grid, the FC system must be able to vary its output voltage in amplitude and phase with respect to the grid voltage [17].

TABLE I. BACK-UP UNIT OPERATIONS

P3 Increase (P1+P2 P3)

P3 Decrease (P1 P2+P3)

Normal (P1=P3)

Discharge

Charge

Normal

Charge

Normal

Normal

Fig. 6. Equivalent circuit of the grid-connected FC energy system

(a)

(b)

(c)

Fig. 7. Vector diagram for the active and reactive power control. (a) When reactive power reference is zero. (b) Active and reactive powers are controlled simultaneously. (c) When active power reference is zero.

Fig. 7 shows the basic idea of how to control the power between the grid and the boost inverter with different vector diagrams. According to these vector diagrams, power flow, active power and reactive power should be controlled by the phase angle, and the inverter voltage amplitude, Vo. For instance, when the reactive power reference is zero Fig. 7 (a) shows active power controlling with small variations of and dVpp. If active and reactive powers need to be controlled simultaneously, Fig. 7 (b) is the key how to control them in same time. On the other hand Fig. 7 (c) illustrates only dVpp is controlled for reactive power while the magnitude of Vo equals to Eg.

Fig. 7, (7) and (8) are illustrating that the system is sensitive with small changes of the phase and the magnitude. Therefore the grid-connected FC system as parallel operation of voltage source inverters requires a precise control. Grid compatible frequency and voltage droops were introduced to control active and reactive power. The droops control for the boost inverter requires the fast acquisition of P and Q. The measurement of P and Q at the PCC is obtained based on the following expressions [16]:

)(21

ggggmeas ivivP

(9)

)(21

ggggmeas ivivQ

(10)

where vg and vg are the instantaneous orthogonal voltages at PCC, and ig and ig are the instantaneous orthogonal currents at PCC. The orthogonal voltage and current systems are obtained using a SOGI based algorithm which provides a fast signal conditioning for single-phase systems, as given in [16].

III. SIMULATION AND EXPERIMENTAL RESULTS The proposed FC system shown in Fig. 2 has been

designed, simulated and tested to validate its overall performance. The simulations have been done using Simulink /MATLAB and PLECS blockset to validate the

analytical results. The output voltage of the system was equal to 230VAC while the input voltage varied between 43 to 69 VDC. The parameters of the proposed FC energy system for the simulation and the prototype are summarized in Table II.

The simulation results show the operations of the boost-inverter and the back-up unit. In particular, Fig. 9(a) illustrates the output voltages of the boost-inverter (V1, V2 and Vo) and Fig. 9(b) shows the grid voltage and grid current at the PCC. The input currents of each boost converter flowing through the inductors L1 and L2 are shown in Fig. 9(c). Figs. 9(d), (e) and (f) illustrate the waveforms of the dc total output current Idc (which is equal to the inverter input current), the FC output current Ifc, and the output current ILb2 of the backup unit respectively. Figs. 9(e) and (f) also illustrate how the back-up unit supports the FC power in transients when the load is increased at 0.15sec. When full-load is required from the no-load operating point, the entire power is provided by the back-up unit to the load as shown in Fig. 9(f). Then, the power drawn from the battery starts decreasing moderately allowing gentle step-up to deliver power which should increase up to meet the demanded load power. Moreover, the back-up unit protects the FC from potential damage by eliminating the

Fig. 8. Voltage reference generation block diagram with P and Q control algorithm

TABLE II. SPECIFICATIONS FOR THE FC SYSTEM

FC output voltage 39 – 69 VDC (72 Cell PEMFC) AC output voltage 230 VAC RMS, 50 Hz

Switching frequency 20 kHz Rated power 1.0kW (43V at 23.5A) Power stack SEMISTACK-IGBT

Controller TMS320F28335 Voltage transducers LEM LV25-P Current transducers LEM HAL50s

Energy storage Two 12V lead acid batteries L1=L2 200μH

Lb1=Lb2 20μH C1=C2=Cb 30μF

Lf 3mH

ripple current due to the boost operation. The high frequency output ripple current of the FC can be canceled by a passive filter placed between the FC and the boost-inverter. Fig. 9(g) shows the duty cycles (d1 and d2) of each boost converter that are varying between approximately 0.15 and 0.85. Fig. 9(h) is the ( ot+ ) which is synchronized with the grid and including as

shown in Fig. 8. The active and reactive power control performances are illustrated with the references and the measured values in Fig. 9 (i) and (j). In addition, Fig. 9 (k) and (l) show the variation of the phase shift between Vo and Eg to control the active power and the variation of the voltage differences, Eg-Vo, relating with reactive power control respectively.

(a) (g)

(b) (h)

(c) (i)

(d) (j)

(e) (k)

(f) (l)

Fig. 9. Simulation results of the proposed FC system. (a) Input and output voltages of the boost-inverter. (b) Grid voltage and current. (c) Current waveforms of L1 and L2. (d) Boost inverter input current. (e) FC output current. (f) Output current of the back-up unit. (g) Duty cycles of the boost-inverter. (h) Theta of the inverter. (i) Active power and its reference. (j) Reactive power and its reference. (k) Small variation of the phase for active power control. (l) Small variation of the voltage amplitude to control reactive power.

Experimental results presented in Fig. 10 show the performance of the boost-inverter operation with the load changing between no-load and 1kW-resistive load. Specifically, Fig. 10(a) illustrates the input and output voltages of the boost-inverter. The current waveforms of the two different inductors are shown in Fig. 10(b). Fig. 10(c) illustrates voltage waveforms when the inverter starts working with full load. FC energy system output voltage and current are shown in Fig. 10(d).

Fig. 10(e) shows that the performance of the back-up unit is able to eliminate the low-frequency current that

may affect the FC with damages or malfunctions while the back-up unit is charging the battery from the FC. As it can be seen, the FC power changed from approximately 500W to 850W in 18 seconds which is a moderate amount of time. Fig. 10(f) also illustrates how the back-up unit supports the FC, since the FC is expected to have a slow power up operation.

IV. CONCLUSIONS A single-phase single power stage grid-connected FC

system based on the boost-inverter topology with a back-

(a) (b)

(c) (d)

Pfc

Vin

Ifc

Back-up unit start

[100W/div][5V/div][5A/div]

[10s/div]

Pfc

VinIfc

Load increase

Steady state

1A/Sec

[100W/div][5V/div][5A/div][2s/div]

(e) (f)

Fig.10. Experimental results (a) Output voltages of the boost-inverter (V1, V2 and Vo) and dc input voltage, Vin (b) Current waveforms of L1 and L2. (c) Voltage waveforms of the boost inverter when the inverter starts with full load. (d) Voltage and current of the boost inverter when full load is connected. (e) Back-up unit starts to eliminate low-frequency current to the FC and charges the battery. (f) When the load is increased, the back-up unit supports the ramp-up operation for the FC.

up battery-based energy storage unit is proposed in this paper. The simulation results and selected laboratory tests verify the operation characteristics of the proposed FC system. In summary, the proposed FC system has a number of attractive features, such as single power stage with high efficiency, simplified topology, low cost and stand-alone as well as grid-connected operation. Moreover, in the grid-connected mode, the single-phase FC system is able to control the active and reactive power using an algorithm based on SOGI.

REFERENCES [1] S.B. Kjaer, J.K. Pedersen, F. Blaabjerg, “A review of

single-phase grid-connected inverters for photovoltaic modules,” IEEE Trans. on Industry Applications, vol. 41 no. 5, pp. 1292 – 1306, Sept.-Oct. 2005.

[2] J.-S. Lai, “Power conditioning circuit topologies” IEEE Industrial Electronics Magazine, Vol. 3, no. 2, pp. 24-34, June 2009

[3] M.E. Schenck, J.-S. Lai and K. Stanton, “Fuel cell and power conditioning system interactions,” Proc. of the APEC 2005. vol. 1, pp. 114 – 120, 6-10 March 2005.

[4] H-Series PEMFC System user guide, Horizon Fuel Cell Technologies, http://www.horizonfuelcell.com

[5] J. Anzicek and M. Thompson, “DC-DC boost converter design for Kettering University’s GEM fuel cell vehicle,” in Electrical Insulation Conference and Electrical Manufacturing Expo, 2005. Proceedings, Oct. 2005, pp. 307 – 316.

[6] J.M. Correa, F.A. Farret, J.R. Gomes, and M. G. Simoes, "Simulation of fuel cell stacks using a computer-controlled power rectifier with the purposes of actual high power injection applications," IEEE Trans. on Industry Applications, vol. 39 no. 4, pp. 1136-1142, July/Aug. 2003.

[7] J. Lee, J. Jo, S. Choi, and S. Han, “A 10-kW SOFC low-voltage battery hybrid power conditioning system for residential Use,” IEEE Trans. Power Electron., vol. 21, no. 2, pp. 575-585, June 2006.

[8] P. Sethakul, S. Rael, B. Davat, and P. Thounthong, “Fuel cell high-power applications,” IEEE Industrial Electronics Magazine, vol. 3, no. 1, pp. 32-46, March 2009.

[9] M. W. Ellis, M. R. Von Spakovsky, and D. J. Nelson, “Fuel cell systems: efficient, flexible energy conversion for the 21st century,” Proc. of the IEEE, vol. 39, no. 12, pp. 1808-1818, Dec. 2001.

[10] C. Pan, C. Lai, “A high efficiency high step-up converter with low switch voltage stress for fuel cell system applications,” IEEE Trans. Industrial Electron., Accepted for future publication, 2009.

[11] X. Yu, M. R. Starke, L. M. Tolbert, and B. Ozpineci; “Fuel cell power conditioning for electric power applications: a summary,” IET-Electric Power Appl., vol. 1, no. 5, pp. 643-656, Sept. 2007.

[12] A. Vazquez-Blanco, C. Aguilar-Castillo, F. Canales-Abarca and J. Arau-Roffiel, “Two-stage and integrated fuel cell power conditioner: Performance comparison,” Proc. of IEEE APEC 2009. vol 15-19, pp. 452–458, Feb. 2009

[13] R. O. Cáceres and I. Barbi, “A boost dc-ac converter: Analysis, design, and experimentation,” IEEE Trans. Power Electron., vol. 14, no. 1, pp.134-141, Jan. 1999.

[14] P. Sanchis, A. Ursea, E. Gubía, and L. Marroyo, “Boost dc-ac inverter: A new control strategy,” IEEE Trans. Power Electron., vol. 20, no. 2, pp.343-353, Mar. 2005.

[15] M. Jang and V. G. Agelidis, “A minimum power-processing stage fuel cell energy system based on aboost-inverter with a bi-directional back-up battery storage,” IEEE Trans. Power Electron., accepted on Sep. 2010.

[16] B. Burger and A. Engler, "Fast signal conditioning in single phase systems", EPE, Graz, Aug. 2001.

[17] B. Bouneb, D. M. Grant, A. Cruden and J. R. McDonald, “Grid connected inverter suitable for economic residential fuel cell operation,”in European Conference on Power Electronics and Applications, 2005, pp. p1-p10.

[18] R. Teodorescu, F. Blaabjerg, M. Liserre and P.C. Loh, “Proportional-resonant controllers and filters for grid-connected voltage-source converters,” in IEE Proceedings - Electric Power Applications, vol. 153, iss. 5, pp. 750-762, 2006.

[19] H. Cha, T. Vu and J. Kim, “Design and control of proportional-resonant controller based photovoltaic power conditioning system,” Proc. of the ECCE 2009, pp. 2198-2205, Sep. 2009.