Download as a PDF - Takeoff Projects

622
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 23, NO. 2, JUNE 2008
An Integrated Hybrid Power Supply for Distributed
Generation Applications Fed by Nonconventional
Energy Sources
Sachin Jain and Vivek Agarwal, Senior Member, IEEE
Abstract—A new, hybrid integrated topology, fed by photovoltaic (PV) and fuel cell (FC) sources and suitable for distributed
generation applications, is proposed. It works as an uninterruptible power source that is able to feed a certain minimum amount of
power into the grid under all conditions. PV is used as the primary
source of power operating near maximum power point (MPP), with
the FC section (block), acting as a current source, feeding only the
deficit power. The unique “integrated” approach obviates the need
for dedicated communication between the two sources for coordination and eliminates the use of a separate, conventional dc/dc
boost converter stage required for PV power processing, resulting
in a reduction of the number of devices, components, and sensors.
Presence of the FC source in parallel (with the PV source) improves
the quality of power fed into the grid by minimizing the voltage
dips in the PV output. Another desirable feature is that even a small
amount of PV power (e.g., during low insolation), can be fed into
the grid. On the other hand, excess power is diverted for auxiliary
functions like electrolysis, resulting in an optimal use of the energy sources. The other advantages of the proposed system include
low cost, compact structure, and high reliability, which render the
system suitable for modular assemblies and “plug-n-play” type applications. All the analytical, simulation, and experimental results
of this research are presented.
Index Terms—Buck-boost, distributed generation, fuel cell,
grid-connected, hybrid, maximum power point tracking (MPPT),
photovoltaic.
I. INTRODUCTION
N THE PAST, centralized power generation was promoted.
The power generation units were generally built away from
the populated areas but close to the sites where the fuel (i.e., fossil fuel) was available. This kept the transportation cost (of the
fuel) to a minimum and eliminated the possibility of pollution
in populated areas. Such schemes remained quite popular until
recently despite drawbacks such as Ohmic (i2 R) losses (due to
transmission of electricity through cables over long distances),
voltage regulation problems, power quality issues, and expansion limitations. With the power demand increasing consistently,
a stage has come when these centralized power generation units
can be stressed no further. As a result, the focus has shifted to
generation (and consumption) of electric power “locally” leading to “distributed power generation systems” (DGS) [1]–[4].
At the same time, increased awareness about the importance
of a clean environment and the quickly vanishing fossil fuels
I
Manuscript received July 6, 2007; revised October 29, 2007. Paper no.
TEC-00237-2007.
The authors are with the Applied Power Electronics Laboratory., Department
of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai
400 076, India (e-mail: [email protected]; [email protected]).
Digital Object Identifier 10.1109/TEC.2008.918631
Fig. 1. Various HDGS configurations. (a) Conventional, multistage topology
using two H-bridge inverters [4], [6]. (b) Modified topology with only one
H-bridge inverter [4]. (c) Proposed topology. λ denotes solar insolation (Suns).
have given impetus to the idea of local power generation using
nonconventional energy (NCE) sources (e.g., photovoltaic (PV)
cells, fuel cells (FC), wind energy, etc.), which may suit a particular region and provide power at various load centers along
the main power grid. Most of these sources are pollution-free
and abundant. Unfortunately, they are not so reliable. For example, the PV source is not available during the nights or during
cloudy conditions. Wind energy may or may not be available.
Other sources, such as fuel cells may be more reliable, but have
monetary issues associated with them. Because of this, two or
more NCE sources are required to ensure a reliable and costeffective power solution. Such integration of different types of
energy sources into a DG system is called a hybrid distributed
generation system (HDGS) [4]–[20].
A combination of PV and FC sources forms a good pair
with promising features [6] for HDGS applications. Of course,
the slow response of the FC needs to be compensated with an
ultracapacitor or a battery (ultracapacitor is preferable due to its
high energy density) [4], [11]. A brief survey of the literature
dealing with HDGS systems is presented next.
Among the earlier work, Tam and Rahman [5] have proposed
an HDGS configuration shown in Fig. 1(a). It consists of two
inverters, operating in parallel, whose outputs are tied to the
grid through a single, multiwinding, step-up transformer. The
0885-8969/$25.00 © 2008 IEEE
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on November 28, 2008 at 02:02 from IEEE Xplore. Restrictions apply.
JAIN AND AGARWAL: AN INTEGRATED HYBRID POWER SUPPLY FOR DISTRIBUTED GENERATION APPLICATIONS
drawback with this otherwise elegant scheme is that it does not
utilize the available sources efficiently as maximum power point
tracking (MPPT) is not implemented. Later, Ro and Rahman [6]
improved upon this system by introducing a two-loop controller
with MPPT. Tao et al., [12] have proposed a multiinput, bidirectional dc–dc converter configuration involving a combination of
dc-link and magnetic coupling. The configuration offers high
boosting capability and galvanic isolation. However, it consists
of multiple power processing stages with an additional dc/ac inverter stage for feeding ac loads. Thus, it requires a large number
of devices.
Monai et al., [13] have proposed another HDGS configuration involving PV, FC, and battery sources that can meet the
fluctuating requirements at the load end. The authors have proposed a modified Euler-type moving average prediction model
for proper sharing of load among the various sources. The
proposed system is a good solution for high power standalone or utility applications. Agbossou et al. [14] have proposed a hybrid system in which the excess energy generated
by the renewable sources is used for producing hydrogen.
Rajashekara [15] proposes another useful system, with PV and
FC sources, for space applications. This system uses the PV
power optimally by diverting the excess power for production
of hydrogen through electrolysis. Another HDGS system with
wind, PV, and battery sources has been discussed by Valenciaga
et al., [7], where they propose a control technique that not only
maintains the load demand but also the state-of-charge of the
battery. Blaabjerg et al., [16] have given an elaborate review
of the various control and grid synchronization techniques used
in HDGS systems. Some other useful HDGS configurations,
based on FC, PV, and wind sources, have also been reported
[17]–[20].
Recently, a PV- and FC-fed hybrid topology has been proposed [4], which makes use of a dedicated boost type dc–dc
converter for each of the two sources, with a common inverter
stage [Fig. 1(b)] for grid connection. The dc–dc converter stages
are meant for boosting the low PV and FC voltages. The PV side
converter also takes care of MPPT. One of the problems with
this system is that, at very low insolation levels, the PV side
dc–dc converter must be cut off to prevent its inefficient operation [21]. Thus, at small power levels, generation by the PV
source remains unutilized. This may be acceptable for highpower applications, but could be a matter of concern for residential or medium-power installations [22]. If the function of
the PV side dc–dc converter is merged with the inversion stage,
as shown in Fig. 1(c), even a small fraction of power generated
during low insolation can be utilized.
Most HDGS configurations discussed earlier use an H-bridge
inverter topology for interfacing with the grid, which either
needs a line frequency bulky transformer at the output or high
dc link voltage at the inverter input. This paper proposes an
integrated solution for PV/FC-based HDGS using a new configuration depicted in Fig. 1(c). The proposed system uses an
inverter with boosting capability, which eliminates the requirement of high dc voltage at the inverter input, thereby saving
the cost of high-voltage buffer capacitor. Various other features,
working principle, control strategy and simulation, and exper-
623
Fig. 2. Proposed Configuration. (a) Block diagram showing the basic concept.
(b) Detailed view of (a) along with the electrolysis application.
imental results for the proposed configuration are described in
the subsequent sections of this paper.
II. BASIC CONCEPT, OPERATING MODES, AND SALIENT
FEATURES OF THE PROPOSED SYSTEM
The proposed topology is built around a buck-boost inverter
topology capable of inversion (dc–ac), boosting and bucking
the voltage and MPPT. The basic idea behind the proposed integrated configuration is shown in Fig. 2(a). A detailed view of
Fig. 2(a) is shown in Fig. 2(b) along with an example (electrolysis) application. A combination of PV and FC sources feeds the
configuration. While the PV source directly feeds the inverter
through a buffer capacitor, CPV , the FC source is interfaced
through a buck-boost type dc–dc converter, as shown in the figure. An extra block is added across CPV to divert the excess
power generated by the PV source.
The proposed system is designed to meet a certain minimum
active power demand (Preq ) from the grid side. PV is the main
source, which is continuously made to track the MPP, while
feeding the required amount of power into the grid. The FC
source, with buckboost type dc–dc converter, acts as a current
source in parallel with the PV source. It is only used to supplement the PV source during low or zero insolation. Thus, FC
supplies only the deficit power into the grid. On the other hand,
any “excess power” generated by the PV source is conditioned
and diverted to an auxiliary application such as electrolysis, to
produce hydrogen, which can be stored for later use by the FC
source. This results in an optimal utilization of the available
sources, rendering a highly economical system [14].
The aforesaid description leads to the following three modes
in which the proposed system operates:
1) Mode-I: Only PV mode (only PV provides power).
2) Mode-II: Hybrid mode (both PV and FC provide power).
3) Mode-III: Only FC mode (only FC provides power).
These operating modes are summarized in Table I.
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on November 28, 2008 at 02:02 from IEEE Xplore. Restrictions apply.
624
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 23, NO. 2, JUNE 2008
TABLE I
OPERATING MODES OF THE PROPOSED HDGS SYSTEM
Fig. 4. Inductor current waveform in discontinuous conduction mode showing
the three constituent intervals.
TABLE II
STATE OF THE DEVICES, GOVERNING EQUATIONS AND ACTIVE CURRENT PATHS
DURING THE POSITIVE HALF-CYCLE OF THE GRID VOLTAGE
Fig. 3. Circuit schematic of the proposed integrated configuration for hybrid
distributed generation system. xi denotes the corresponding state variable.
Due to its unique hybrid integrated nature, the proposed configuration offers several desirable features as outlined next:
1) It obviates the requirement of a boost converter stage for
conditioning the PV power, as shown in Fig. 1(c).
2) Out of the two capacitors Cdc and CPV [Fig. 1(a)], only
one is required [Fig. 1(c)].
3) Presence of FC in parallel reduces the fluctuations in the
PV voltage due to changing environmental conditions.
This, in turn, reduces the fluctuations in the power fed into
the grid. Consequently, the grid voltage profile improves
[23].
4) Elimination of dips and surges in the PV voltage increases
the speed of MPPT.
5) In two-stage systems [4], usually, both PV array and dc
capacitor (Cdc ) voltages are sensed for MPPT and power
control, respectively. In the proposed system, Cdc (or
CPV ) appears right across the PV terminals due to elimination of the dedicated boost stage on the PV side [Fig. 1(b)].
Hence, only one sensor is adequate.
6) It eliminates the requirement of extra hardware for communication and coordination between various sources [24]
to generate and use the available energy optimally.
7) The special configuration, in which the PV and FC sources
are connected, ensures that the FC section works as a
current source irrespective of the voltage magnitude at its
output. This facilitates an appropriate adjustment of the
PV voltage for PV’s operation close to MPP.
8) Proposed configuration is a compact, low-cost, and reliable solution for HDG applications.
III. CIRCUIT OPERATION AND ANALYSIS
Fig. 3. shows the complete circuit schematic of the proposed
configuration along with the electrolysis application. With refer-
ence to the general description of the proposed topology [Fig. 2]
in Section II, the following three sections can be identified in
Fig. 3:
1) Inverter section for grid interfacing.
2) A buck-boost type dc–dc converter for conditioning the
FC power.
3) A buck type dc–dc converter for conditioning and diverting excess PV power for electrolysis application.
The inverter consists of two buck-boost converters connected
back to back [25], as shown in Fig. 3. It uses two pairs of controllable devices. Each pair (SWpa and SWpb or SWn a and
SWn b ) is operated for one half-cycle of the grid voltage. One
of the switches, SWpb (or SWn b ) in the pair is kept ON for
the entire positive (or negative) half-cycle of the grid voltage,
while the other, SWpa (or SWn a ) operates at high switching frequency with sine triangular pulse width modulation (SPWM).
The converters are made to operate in discontinuous Current
mode (DCM) to feed high-quality current waveforms into the
grid. The converter is designed to operate in critical conduction
mode (boundary of continuous current mode (CCM) and DCM)
around the peak of the grid voltage, where, twice the rated average power is being delivered to the grid. The critical conduction
mode and DCM operation are shown in Fig. 4.
When SWpa is ON, the inductor “L” stores energy. When
SWpa is OFF, this stored energy is transferred into the capacitor
“Cf ”. Since the ON/OFF times are modulated in a sinusoidal
manner, current fed into the grid is also sinusoidal. The state
equations governing the operation of the inverter during the
positive half-cycle of the grid voltage are shown in Table II
along with the corresponding active current paths. The operation
and analysis during the negative half-cycle of the grid voltage
is analogous.
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on November 28, 2008 at 02:02 from IEEE Xplore. Restrictions apply.
JAIN AND AGARWAL: AN INTEGRATED HYBRID POWER SUPPLY FOR DISTRIBUTED GENERATION APPLICATIONS
625
As the inverter operates in DCM, there is a complete energy transfer from the buck-boost inductor to the grid during
each high-frequency switching cycle. Assuming the switching
frequency to be an integer multiple of the grid fundamental frequency (i.e., fs (= 1/Ts ) = 2n × fg (1/Tg )), each cycle of the
grid voltage can be divided into “2n” parts or intervals. During
each such interval, a definite amount of energy is being transferred. Assuming grid voltage to be constant during each switching cycle, the duty ratio, DON (k) for the kth(k = 1, 2, . . . n)
switching cycle is given by
DON(k ) = M ×sin(π × (k/n)),
where, M =Vsin /Vtri (1)
where Vsin and Vtri are the amplitudes of the rectified sine wave
and triangular wave, respectively. Energy “Ek (inv) ” transferred
in the kth switching interval with modulation index “M ” is
given by
2
2 VPV
L
k
× Ts ×M × sin π×
(2)
Ek (inv) = ×
2
L
n
where VPV is the average voltage across the PV array. Thus, the
total energy transferred during a fundamental cycle of the grid
voltage is given by
E(inv) =
2n
k =1
Ek(inv) =
n × (T s ×M × V PV )2
.
2×L
(3)
The FC side buck-boost converter can be operated in DCM
or CCM, depending on the control technique used. Due to the
inherent nature of the buck-boost converter’s operation, the input
source never sees the output capacitor CPV . In other words,
the input energy is first stored in inductor Lbb , which is then
transferred on to the output capacitor. The energy drawn from
the FC source during one fundamental cycle of the grid voltage
is given by
Ek (bb) = VFC ×I FC ×T g
(4)
where, VFC and IFC are the average values of FC voltage and
FC current, respectively. Since IFC = DON(FC) ×I L(bb) and VFC
is a function of IFC , (4) may be rewritten as follows:
Ek (bb) = f (DON(FC) , I L (bb) , T g )
(5)
where IFC and IL (bb) are the average values of the FC source
current and FC side buck-boost inductor current, respectively,
and DON (FC) is the turn “ON” interval [Fig. 4] of the FC side
buck-boost converter.
The buck converter is added to the system (as shown in
Figs. 2(b) and 3) to condition the excess power (Pex ) generated
by the PV source for electrolysis process. The buck converter
has the advantage of having high current at the output. This helps
in increasing the production of hydrogen by electrolysis [14].
The electrolysis load is represented by a resistive load (REL ) in
Fig. 3.
IV. MPPT ALGORITHM AND CONTROL STRATEGY
The modulation index, M of the grid-connected SPWM inverter is adjusted to enable the desired power flow. The amount
Fig. 5.
I–V and P–V characteristics of the PV array with β curve.
of power depends on the MPP, which is being tracked. Thus,
the value of M is computed by the MPPT algorithm used.
Various MPPT schemes are available. Unfortunately, popular
MPPT techniques, like hill climbing or incremental conductance
method cannot be applied because they provide slow tracking as
they use a fixed and small incremental change in the modulation
index to track MPP. During mode II, when FC is also feeding
power into the system, small incremental changes in the modulation index will not be able to balance the input and output
power across CPV , resulting in the rise of voltage across CPV .
This is equivalent to shifting the PV operating point toward open
circuit condition (OCC) resulting in underutilization of the PV
source. Also, as the change in the modulation index can be implemented only at the start of the fundamental cycle of the grid
voltage, the probability of underutilizing the PV power becomes
high using slow MPPT algorithms. Therefore, it is imperative
to use a fast MPPT scheme, which can counter the drawbacks
of the slow schemes.
A. MPPT Scheme Used in the Proposed Topology
A fast MPPT scheme, called the “β” method [26], was suitably modified and used for the given application. The scheme
is based on the observation that the value of an intermediate
variable “β”, defined only at MPP condition, varies with in a
narrow band (βm ax − βm in ) as the MPP varies from PM PP(m ax)
to PM PP(m in) over the full insolation and temperature range
(λm ax , Tm ax to λm in , Tm in ), as shown in Fig. 5. β is a subset
of β , which is applicable to any point on the P–V curve, including MPP. β is obtained [26] by using the MPP condition,
∂P/dV = 0 and is given by,
β = ln(−I o ×c) = ln
Ipv
−c × V pv
Vpv
(6)
where Io is the reverse saturation current of the diode. Therefore,
by tracking β using large iterative steps, the operating point can
be quickly brought into a narrow band of MPP [Fig. 5]. Also,
controlling “β” indirectly controls the operating voltage (OV)
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on November 28, 2008 at 02:02 from IEEE Xplore. Restrictions apply.
626
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 23, NO. 2, JUNE 2008
Fig. 7. Simplified block diagram of the proposed HDGS configuration showing the control variables. The arrows indicate the power flow direction along
with the notations.
insolation conditions, which helps in deciding the intershuffling
of operation between modes I–III.
B. Control of Converters and Intershuffling Between Modes
Fig. 6. Flow-chart of the MPPT algorithm used in the proposed HDGS scheme.
M m in and M m a x denote minimum and maximum values of the modulation
index.
of the PV array, which helps in balancing the power across
capacitor CPV .
It is important to note that problems may arise during the lowinsolation phase when the generated PV power is very small and
the value of “M”, as computed by the MPP algorithm, is small.
Under such conditions, as FC continues to feed the deficit power,
this will result in a voltage increase across CPV due to imbalance
between the input and output power. This may lead to shifting
of OV of the PV array toward the OCC. The array will remain
at OCC even when high or normal insolation is restored and the
system will continue to draw the required power from FC. To
avoid such a situation, the algorithm has been suitably modified
to ensure that the PV array voltage is always less than or equal to
a critical voltage (Vcrit ), which is less than the PV array’s VOC .
Whenever the array voltage goes beyond Vcrit (which occurs
only when the power drawn from the FC is not fed into the grid
due to the low value of M) and array power is less than the
minimum power Pm in , the modulation index is set to a fixed
value (Mcon ) by the algorithm. This ensures that VPV ≤ Vcrit .
The complete flowchart of the proposed MPP tracking algorithm
is shown in Fig. 6. In the given flowchart, the average values of
VPV and IPV are processed to obtain the current value of β .
Then, PPV is calculated and its value is compared with the set
minimum value. The value of β is also checked for whether
its magnitude lies within the MPP range. If the PV power is
less and β lies within the range, it corresponds to the minimum
insolation condition. This initiates mode III operation with a set,
fixed value of “M”.
An additional advantage of the β method is that β and
β crit values themselves serve as an indicator of low- or zero-
The power transferred to the grid depends upon the value of M
and VPV during mode-1 and upon M, VPV and DON (FC) during
modes II and III. Equations (3) and (5) show that the amount
of power transferred through the inverter is determined by the
values of VPV and M. The control scheme adjusts the value of
m according to the MPPT requirement, which, in turn, adjusts
VPV . The power supply from FC (controlled through DON (FC))
is not affected by this because the FC source is not a stiff dc
source. The Basic power flow scheme and the control variables
used in the proposed HDGS configuration are shown in Fig. 7.
Excess power, Pex generated by the PV source is diverted for
electrolysis application through a buck converter. Hydrogen, the
end product of electrolysis, is stored and may be used by the
fuel cell to feed Pdef at a later time.
The overall control scheme comprises of three independent
control loops, which work in coordination with each other. These
are:
1) SPWM control of inverter along with MPPT.
2) FC side dc–dc converter control.
3) Excess power control (control of dc–dc buck converter).
Fig. 8 shows the complete control strategy and the truth
table for the logic controller used in the proposed system. When
Preq > PPV , the logic controller turns ON the control for extracting Pdef from the FC and turns OFF the control diverting
Pex for electrolysis and vice versa. The individual controllers
corresponding to the excess or deficit power are controlled by
the logic controller and are not turned on simultaneously. Details
of the two controllers are described next.
1) SPWM Control of Inverter Along With MPPT: The inverter is operated using the SPWM technique to feed the sinusoidal current into the grid during all the three operating modes
[Fig. 8(a)]. This control loop ensures the operation of the PV
system at an optimum voltage corresponding to the MPP of the
PV array in modes I and II. As described earlier, the MPPT algorithm also predicts the condition for entering Mode-III (only
FC mode) based on the values of β , VPV , and PPV . In this
operating mode, the controller shifts to a fixed modulation index, Mcon , which maintains a critical voltage “Vcrit ” across the
PV array. This is achieved by incorporating this feature in the
MPPT algorithm itself. The value of Mcon can be calculated as
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on November 28, 2008 at 02:02 from IEEE Xplore. Restrictions apply.
JAIN AND AGARWAL: AN INTEGRATED HYBRID POWER SUPPLY FOR DISTRIBUTED GENERATION APPLICATIONS
627
the inductor (Lbb ) current. This is the desired reference current,
Iref L(bb) that should flow through Lbb and is given by
Iref L(bb) =[2 × P def ×(V FC +V PV )]/(V FC ×V PV ).
(10)
A fast hysteresis current controller makes the buck-boost inductor current track this reference by controlling DON (FC), which
results in the FC source supplying Pdef .
3) Excess Power Control: The logic controller activates the
control of the buck converter, when PV power is more than
the required power. Fig. 8(c) shows the control scheme. In the
control block of the buck converter, the average output power,
Po is computed using the sensed grid current (assuming constant
Vg ). Po is compared with the reference power (= PPV − Preq )
to calculate the error. The error is then fed into a PI controller,
which modulates the duty ratio, DElec of the buck converter
Fig. 8. Complete control strategy used in the proposed system. (a) Inverter
control. β is computed with in the MPPT block; (b) FC side buck-boost converter control. (c) Buck converter control. (d) Truth table of the logic controllers.
follows. Equating the input and output energy of the converter
in steady state for the kth division, assuming unity power factor operation (i.e., grid voltage and current in phase and both
sinusoidal) yields
VPV (crit)×I p k ×(D ON( k ) ×T s )
2 π×k
= Vg × Ig × Ts × sin
2
n
(7)
where, Vg and Ig are the peak values of the grid voltage and
current, respectively, while Ipk is the peak value of idc . Simplifying (7) near the peak of grid voltage (i.e., k = n/2), using
(1), and solving for “M”(= Mcon ) gives
2
(8)
Mcon = (4 × Preq × L)/(VPV
(crit) × Ts )
During mode-III, the possibility of reverse current flowing into
the PV array is prevented due to the presence of a protection
diode, DPr [Fig. 3].
2) FC side dc–dc converter control: Fig. 8(b) shows the
control block used for the FC side dc–dc converter. If PV power
falls below Preq , the logic controller activates the control of
the buck-boost converter to feed deficit power from the FC
source. It is observed from (5) that the power drawn from the FC
source can be controlled by modulating DON (FC), irrespective
of the voltage across CPV . Effectively, the converter acts as a
current source [Appendix A], feeding current, corresponding to
the deficit power, Pdef . The deficit power that should be fed into
the grid can be calculated as
Pdef = Preq − PPV .
(9)
Another way to determine Pdef is by subtracting Preq from the
average output power, Po . However, a PI controller is required in
this case to correctly maintain the required power reference and
DON (FC). This is because Po may already contain a component
supplied by the FC. The presence of PI, however, slows down
the response. Therefore, in this research, (9) has been used to
determine Pdef . The ratio Pdef /VPV provides the desired value
of iFC , which is “(1 − DON (FC))” times the average value of
V. DESIGN PROCEDURE
Design of the proposed HDGS system involves the design
of the three power converters described in the previous section.
It includes the determination of the values of capacitors and
inductors, which is presented in this section, and the ratings of
the various power devices used in the system, which is obtained
with simulations in Section VI. Table III summarizes the design
values of the various components.
A. Design of Inverter
It involves the design of buck-boost inductor L, capacitor Cf
and inductor Lf [25]. Design values of the parameters are
2
L ≤ (VPV
× Ts )/(4 × P PV (rated) )
(11)
Cf = (P PV (rated) × Ts )/(V g × ∆Vg )
(12)
Lf = ((2 × π × fc )2 × Cf )−1
(13)
CPV =
2 × PPV (rated)
4 × (2 × π × fg ) × V PV ×∆v PV
(14)
where ∆Vg and ∆VPV are the allowed maximum ripple in the
grid and PV voltage, respectively, PPV (rated) is the maximum
rated power extracted from the PV array and fc is the lowest
cutoff frequency determined by the circuit parameters of the
inverter configuration. l can have any value satisfying (11), but
is optimally chosen such that the converters operate in critical
conduction mode at the peak of the grid voltage.
B. Design of FC Side dc-dc Converter
The FC side dc–dc converter is required to supply the deficit
power to the grid. It acts as a voltage-fed controlled current
source. Thus, a proper design of inductor Lbb is critical. As the
FC supplies the deficit power under the control of a hysteresis
controller, the design of Lbb should ensure that the converter
operates in CCM. Let ∆I be the allowed ripple in the inductor
current. Then, the values of inductor and capacitor are given by
Lbb = (T s ×V FC ×V PV )/((V FC +V PV ) × ∆I) (15)
CFC = (T s ×P req )/((VFC + VPV ) × ∆VFC )
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on November 28, 2008 at 02:02 from IEEE Xplore. Restrictions apply.
(16)
628
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 23, NO. 2, JUNE 2008
TABLE III
DESIGN VALUES OF THE COMPONENTS AND RATINGS OF THE DEVICES USED IN THE SIMULATION MODEL
where VFC and ∆VFC are the average value of and the allowed
ripple in the fuel cell voltage, respectively.
C. Buck Converter Design
Design value of the inductor used in the buck converter is
given by
Lbuck =[T s ×(V PV −V El ) × V El ]/(V PV ×∆I buck )
(17)
where, VEl is the output voltage across the electrolysis load and
∆Ibuck is the allowed ripple in the buck inductor current.
VI. SIMULATION AND HARDWARE RESULTS
The proposed configuration, along with the control scheme
[Fig. 8], is simulated using the MATLAB/SIMULINK. Design
values of the components obtained in Section V [Table III] were
used. PV and FC sources are realized in the MATLAB Function with the help of their governing equations [26]–[28]. FC
steady state response is considered with the assumption that
the FC source is coupled with an ultracapacitor to improve
its transient response. The inverter and the boost converter are
simulated in the SIMULINK using the state equations given
in Table II. Specifications of the PV and FC sources, used
in the simulations, are as follows: PV source: Open circuit
voltage, Vo c ≈ 125 V ; short-circuit current, Isc = 7.6 A. FC
source: Vo c = 69 V ; Isc = 42 A. The system is required to feed
500 W (=Preq ) into the grid. The inverter system is designed
for 1 kW power. The ratings of the various devices used in
the system were ascertained by simulations and are listed in
Table III.
Fig. 9 shows the simulation results of the basic HDGS configuration without excess power control. Results corresponding to
low, medium, and high values of λ are shown. It is observed that
beginning at t = 0 s, when λ is negligible, FC supplies the entire
Preq (=500 W). Around t = 1.2 s onwards, λ value picks up and
the PV power increases. The power drawn from the FC automatically decreases, such that PPV (t) + PFC (t) = Preq . As the
PV source operates near MPP, an optimal utilization of the two
sources is ensured. At t = 3 s onwards, λ value improves further
and the entire Preq is supplied by the PV. PFC reduces to zero.
Any excess power generated by the PV is also fed into the grid,
because there is no excess power diversion. It is observed that
the proposed system is able to supply the required power (Preq )
under all conditions and for all the three modes as shown in the
figure. In addition, there is a smooth transition between the various modes (denoted by T1 and T2 in Fig. 9) while maintaining
Preq . Fig. 9 also shows the variation of other parameters like
voltage, current, modulation index (M) of SPWM, etc., during
transition and operation in various modes.
Fig. 9. Simulation results of the integrated hybrid configuration showing transition from mode III to mode II and then to mode I. T1 and T2 denote the
transition between mode III to mode II and mode II to mode I respectively.
Fig. 10. Simulation results of the integrated hybrid configuration operating in
electrolysis mode (mode I to mode III and then to mode I). T1 and T2 denote
the transition between mode I to mode III and mode III to mode I respectively.
Fig. 10 shows simulation results obtained by incorporating
excess power control. The excess PV power (Pex ) is diverted for
electrolysis application through a buck converter. This prevents
any grid voltage surge in mode-I due to pumping. of excess
power into the grid. Further, Preq . is always fed into the grid
with PV array operating near MPP. A small-step increment in
the environmental conditions near 7.9 s shows corresponding
increase in diverted power in steady state with the PV array
operating at MPP.
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on November 28, 2008 at 02:02 from IEEE Xplore. Restrictions apply.
JAIN AND AGARWAL: AN INTEGRATED HYBRID POWER SUPPLY FOR DISTRIBUTED GENERATION APPLICATIONS
629
Fig. 12. Experimental plots showing the current, voltage, and power waveforms on the PV side for Mode II.
Fig. 11. Simulation results. Performance comparison of the proposed HDGS
system with and without an FC source in parallel with the PV source.
To demonstrate the usefulness of placing an FC source in parallel with the PV source, simulations were performed with and
without the FC source in the circuit, for a step change in insolation. Fig. 11 shows the results. It is observed that the presence
of FC eliminates the transient dips in the PV array voltage and
current, which, in turn, improves the MPPT speed and efficiency
of the system. Further, the grid current total harmonic distortion
(THD) during the transients is lower in the presence of the FC.
An experimental prototype of a 300 W system was built
to verify the simulation results and other theoretical claims.
Due to the nonavailability of the fuel cell source, a DC source
was used in the experiments, which can be justified by the
fact that the performance of an FC source, with a reformer
and an ultracapacitor, is close to a stiff dc source [4]. Component values used in the prototype are: CPV = 2000 µF;
L = 310 µH; Cf = 2.2 µF; Lf = 3.25 mH; Lbb = 0.7mH
and Lbuck = 2 mH. Controllable power devices used are MOSFETs IRFP460, while the power diodes used are CSD20060.
Other specifications include VPV ≈ 90 V, VFC = 50 V and
Vg = 140 V. To implement the control scheme depicted in
Fig. 8, TMS320LF2407 digital signal processor was used. A
DSP controller was preferred to enable the computation of
the average values of the sensed PV voltage and current using
fast fourier transform and implementation of the MPPT algorithm. Also, it was used for computation of Po , Iref L(bb) , etc.
LEM current sensors are used for sensing the instantaneous
values of iPV , iL (bb) and igrid . The PI and hysteresis controllers were implemented using JFET input-stage operational
amplifiers TL084 having high input impedance, bandwidth, and
slew rate.
Fig. 12 depicts the PV side waveforms for mode II. Fig. 13
shows the current and voltage waveforms of the dc source used
in place of the FC source. Fig. 14 shows the voltage and current
waveforms of the inverter section across Cf and inductor “L”
respectively. Fig. 15 gives the grid-side experimental waveforms
of voltage, current, and power. It is observed that the grid current
Fig. 13. Experimental waveforms (Mode II) of the dc source current, voltage,
and power. A dc source was used in place of the FC stack.
Fig. 14. Experimental waveforms (Mode II) of the buck-boost inductor (L)
current of the inverter section and voltage across capacitor C f .
Fig. 15. Experimental results (Mode II) showing the waveforms of grid voltage, grid current and power fed into the grid.
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on November 28, 2008 at 02:02 from IEEE Xplore. Restrictions apply.
630
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 23, NO. 2, JUNE 2008
TABLE IV
COMPARISON OF THE RESPONSE OF THE PROPOSED HDGS SYSTEM WITH AND WITHOUT FC SOURCE
TABLE V
TABLE SHOWING LIST OF THE COMPONENTS USED IN HYBRID POWER SUPPLY (HPS) FOR COST ESTIMATION
THD is on the higher side. This can be attributed to a slightly
distorted grid voltage.
VII. CONCLUSION
A compact topology, suitable for grid-connected applications
has been proposed. Its working principle, analysis, and design
procedure have been presented. The topology is fed by a hybrid combination of PV and FC sources. PV is the main source,
while FC serves as an auxiliary source to compensate for the
uncertainties of the PV source. The presence of FC source improves the quality of power (grid current THD, grid voltage
profile, etc.) fed into the grid and decreases the time taken to
reach the MPP. Table IV compares the system performance with
and without the FC block in the system. A good feature of the
proposed configuration is that the PV source is directly coupled
with the inverter (and not through a dedicated dc–dc converter)
and the FC block acts as a current source. Considering that the
FC is not a stiff dc source, this facilitates PV operation at MPP
over a wide range of solar insolation, leading to an optimal utilization of the energy sources. The efficiency of the proposed
system in mode-1 is higher (around 85% to 90%) than mode
2 and 3 (around 80% to 85%). A laboratory prototype of the
proposed system has shown encouraging results in terms of efficiency, complexity, reliability, EMI concerns, and other features.
Table V compares the proposed system and some of the existing
HDGS configurations with respect to various parameters and
features.
APPENDIX
The state-space-averaged [29] equations for the integrated
system, consisting of FC side dc–dc converter and the inverter,
corresponding to the positive half-cycle of the grid voltage are
[Fig. 3]
Lbb ẋ6 = DON(FC) xFC −DOFF(FC) x5
(A1)
CFC ẋFC = iFC −DON(FC) x6
(A2)
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on November 28, 2008 at 02:02 from IEEE Xplore. Restrictions apply.
JAIN AND AGARWAL: AN INTEGRATED HYBRID POWER SUPPLY FOR DISTRIBUTED GENERATION APPLICATIONS
Fig. 16. Equivalent large signal model of the proposed configuration in continuous time domain during positive half cycle of the grid voltage.
CPV ẋ5 = iPV −DON(k ) x4p +DOFF(FC) x6
(A3)
Lẋ4p = DON(k ) x5 −DOFF(k ) x3
(A4)
Cf ẋ3 = −x1 +DOFF(k ) x4p
(A5)
Lf ẋ4 = vg −x3
(A6)
Continuous time-domain large-signal equivalent circuit corresponding to (A1)–(A6) is shown in Fig. 16. It is observed that
the FC source in conjunction with the buck-boost converter acts
as a current source (L1 tends to be large).
REFERENCES
[1] J. Kabouris and G. C. Contaxis, “Optimum expansion planning of an
unconventional generation system operating in parallel with a large scale
network,” IEEE Trans. Energy Convers., vol. 6, no. 3, pp. 394–400, Sep.
1991.
[2] P. Chiradeja and R. Ramakumar, “An approach to quantify the technical
benefits of distributed generation,” IEEE Trans. Energy Convers., vol. 19,
no. 4, pp. 764–773, Dec. 2004.
[3] Y. H. Kim and S. S. Kim, “An electrical modeling and fuzzy logic control
of a fuel cell generation system,” IEEE Trans. Energy Convers., vol. 14,
no. 2, pp. 239–244, Jun. 1999.
[4] K. N. Reddy and V. Agarwal, “Utility interactive hybrid distributed generation scheme with compensation feature,” IEEE Trans. Energy Convers.,
vol. 22, no. 3, pp. 666–673, Sep. 2007.
[5] K. S. Tam and S. Rahman, “System performance improvement provided
by a power conditioning subsystem for central station photovoltaic fuel
cell power plant,” IEEE Trans. Energy Convers., vol. 3, no. 1, pp. 64–70.
[6] R. Kyoungsoo and S. Rahman, “Two loop controller for maximizing performance of a grid-connected, photovoltaic fuel cell hybrid power plant,”
IEEE Trans. Energy Convers., vol. 13, no. 3, pp. 276–281, Sep. 1998.
[7] F. Valenciaga and P. F. Puleston, “Supervisor control for a stand-alone
hybrid generation system using wind and photovoltaic energy,” IEEE
Trans. Energy Convers., vol. 20, no. 2, pp. 398–405, Jun. 2005.
[8] N. Kato, K. Kurozumi, N. Susuld, and S. Muroyama, “Hybrid powersupply system composed of photovoltaic and fuel-cell systems,” in
Telecommun. Energy Conf., Jpn, Oct., 2001, pp. 631–635.
[9] T. Senjyu, T. Nakaji, K. Uezato, and T. Funabashi, “A hybrid power system
using alternative energy facilities in isolated island,” IEEE Trans. Energy
Convers., vol. 20, no. 2, pp. 406–414, Jun. 2005.
[10] Y. M. Chen, C. S. Cheng, and H. C. Wu, “Grid-connected hybrid PV/wind
power generation system with improved dc bus voltage regulation strategy,” in Proc. IEEE APEC, 2006, pp. 1089–1094.
[11] W. Gao, “Performance comparison of a fuel cell-battery hybrid power
train and a fuel cell-ultra capacitor hybrid power train,” IEEE Trans. Veh.
Technol., vol. 54, no. 3, pp. 846–855, May 2005.
[12] H. Tao, A. Kotsopoulos, J. L. Duarte, and M. A. M. Hendrix, “Multi-input
bidirectional dc–dc converter combining dc-link and magneticcoupling for
fuel cell systems,” in Proc. IEEE Conf. Ind. Appl., 2003, vol. 3, pp. 1136–
1142.
[13] T. Monai, I. Takano, H. Nishikawa, and Y. Sawada, “A collaborative
operation method between new energy-type dispersed power supply and
EDLC,” IEEE Trans. Energy Convers., vol. 19, no. 3, pp. 590–598, Sep.
2004.
[14] K. Agbossou, M. Kolhe, J. Hamelin, and T. K. Bose, “Performance of a
stand-alone renewable energy system based on energy storage as hydrogen,” IEEE Trans. Energy Convers., vol. 19, no. 3, pp. 633–640, Sep.
2004.
[15] K. Rajashekara, “Hybrid fuel-cell strategies for clean power generation,”
IEEE Trans. Ind. Appl., vol. 41, no. 3, pp. 682–689, May/Jun. 2005.
631
[16] F. Blaabjerg, R. Teodorescu, M. Liserre, and A. V. Timbus, “Overview
of control and grid synchronization for distributed power generation systems,” IEEE Trans. Ind. Electron., vol. 53, no. 5, pp. 1398–1409, Oct.
2006.
[17] A. Di Napoli, F. Crescimbini, S. Rodo, and L. Solero, “Multiple-input
dc–dc power converter for fuel-cell powered hybrid vehicles,” in Proc.
IEEE IAS, 2002, vol. 3, pp. 1578–1585.
[18] J. Zhenhua and R. A. Dougal, “A hybrid fuel cell power supply with rapid
dynamic response and high peak-power capacity,” in Proc. IEEE APEC,
2006, pp. 1250–1255.
[19] D. Das, R. Esmaili, L. Xu, and D. Nichols, “An optimal design of a
grid connected hybrid wind/photovoltaic/fuel cell system for distributed
energy production,” in Proc. IEEE IECON, 2005, pp. 2499–2504.
[20] S. Nakahara, M. Kobayashi, Y. Sawada, I. Takano, and H. Nishikawa, “A
collaborative operation study of hybrid new energy type dispersed power
supply system,” in Proc. IEEE PESC, 2006, pp. 999–1004.
[21] M. Jantsch and C. W. G. Verhoeve, “AC PV module inverters with full
sine wave burst operation mode for improved efficiency of grid connected
systems at low irradiance,” presented at the 14th EC PVSEC, Barcelona,
Spain, 1997.
[22] F. Blaabjerg, C. Zhe, and S. B. Kjaer, “Power electronics as efficient
interface in dispersed power generation systems,” IEEE Trans. Power
Electron., vol. 19, no. 5, pp. 1184–1194, Sep. 2004.
[23] A. Woyte, V.V. Thong, R. Belmans, and J. Nijs, “Voltage fluctuations
on distribution level introduced by photovoltaic systems,” IEEE Trans.
Energy Convers., vol. 21, no. 1, pp. 202–209, Mar. 2006.
[24] K. J. P. Macken, K. Vanthournout, J. Van Den Keybus, G. Deconinck,
and R. J. M. Belmans, “Distributed control of renewable generation units
with integrated active filter,” IEEE Trans. Power Electron., vol. 19, no. 5,
pp. 1353–1360, Sep. 2004.
[25] S. Jain and V. Agarwal, “A single-stage grid connected inverter topology
for solar PV systems with maximum power point tracking,” IEEE Trans.
Power Electron., vol. 22, no. 5, pp. 1928–1940, Sep. 2007.
[26] S. Jain and V. Agarwal, “A new algorithm for rapid tracking of approximate
maximum power point in photovoltaic systems,” IEEE Power Electron.
Lett., vol. 2, no. 1, pp. 16–19, Mar. 2004.
[27] 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. Ind.
Appl., vol. 39, no. 4, pp. 1136–1142, Jul/Aug. 2003.
[28] G. Walker, “Evaluating MPPT converter topologies using a MATLAB PV
model,” J. Electr. Electron. Eng., Aust. IE Aust., vol. 21, no. 1, pp. 49–56,
2001.
[29] D. M. Mitchell, dc-dc Switching Regulator Analysis, 3rd ed. New York:
McGraw-Hill, 1998.
Sachin Jain received the Bachelor’s degree from B.I.T. Engineering College, Pt.
Ravishankar Shukla University, Bhilai, India, in 2000, and the Master’s degree
in integrated power systems from V.N.I.T., Nagpur, India, in 2002, currently
he is working toward the Ph.D. degree in electrical engineering from Indian
Institute of Technology Powai, Mumbai, India.
His current research interests include power electronics applications in nonconventional energy conditioning, power quality, and distributed generation.
Vivek Agarwal (S’92–M’93–SM’01) received the Bachelor’s degree in physics
from St. Stephen’s College, Delhi University, New Delhi, India, the integrated
Master’s degree in electrical engineering from Indian Institute of Science,
Bangalore, India, and the Ph.D. degree from the Department of Electrical and
Computer Engineering, University of Victoria, Victoria, BC, Canada.
He briefly was with Statpower Technologies, Burnaby, Canada as a Research Engineer. In 1995, he joined the Department of Electrical Engineering,
Indian Institute of Technology Powai, Mumbai, India, where he is currently a
Professor. His current research interests include power electronics with special
emphasis on the modeling and simulation of new power converter configurations, intelligent and hybrid control of power electronic systems, power quality
issues, electromagnetic interference/electromagnetic compatibility issues, and
conditioning of energy from nonconventional sources.
Prof. Agarwal is a Fellow of the Institution of Electronics and Telecommunication Engineers (IETE) and a Life Member of the Indian Society for Technical
Education (ISTE).
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on November 28, 2008 at 02:02 from IEEE Xplore. Restrictions apply.