Robust control of hybrid fuel cell/energy storage distributed power generation system in weak grid...

EUROPEAN TRANSACTIONS ON ELECTRICAL POWEREuro. Trans. Electr. Power 2011; 21:522–540Published online 20 May 2010 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/etep.459

Robust control of hybrid fuel cell/energy storage distributedpower generation system in weak grid under balanced

and unbalanced voltage sag

*C

TecyE-

Co

A. Hajizadeh1*,y, M. A. Golkar1 and L. Norum2

1Electrical and Computer Engineering Department, K.N.Toosi University of Technology, Tehran, Iran2Electrical Power Engineering Department, Norwegian University of Science and Technology, Trondheim,

Norway

SUMMARY

The operation and control of fuel cell (FC) distributed generation (DG) systems combined with energy storageare studied in this paper. For this purpose, modeling of hybrid FC/energy storage DG system is developed.Dynamic models for proton exchange membrane fuel cell (PEMFC), supercapacitor, DC–DC converters, andgrid connected inverter are described. Then a PI controller for the FC converter (boost) and a sliding modecontroller for supercapacitor converter (buck-boost converter) are designed to regulate their duty cycles basedon the requested current. Moreover, a robust current control strategy for voltage source converter based onpositive and negative symmetrical components is presented for treating the unbalance and balance voltage sagconditions. In order to control the power between dc power sources and stabilize the dc-link power duringnormal and fault conditions, a self-tuning fuzzy control strategy has been designed. This controller determinesthe supercapacitor power that should be generated according to the amount of available energy in dc-link. Thehybrid system is studied under unbalance and balance voltage sag condition. Simulation results are given toshow the response of whole system under voltage sag and illustrate the performance including active powercontrol and voltage sag ride-through capability of the proposed control strategy. Copyright # 2010 John Wiley& Sons, Ltd.

key words: control; distributed generation; energy storage; PEM fuel cell; fuzzy; sliding mode; robust;voltage sag

1. INTRODUCTION

The progress in deregulation of the electricity market and utility restructuring makes that the tight

constraints are imposed on the construction of new transmission lines. In this case, the interest in

distributed generation (DG) systems installed near load is increasing [1]. Fuel cell (FC) is one of the

attractive DG technologies. Environmental friendliness, practically noise free operation, and high

efficiency make FCs a very sound competitor on the future power systems. The FCs also produce water

and heat by combining hydrogen with oxygen, and have 50% efficiency for only electricity but could

reach 85% in the case of combined heat and power (CHP). Also, the FCs can be used for various

applications such as stationary sites (buildings, hospitals, domestic utility, etc.), transportation (FC

vehicle), portable power (laptop, cell phone), and distributed power [2]. Many of the FC power systems

are connected to the grid via power electronic converters to improve the system integrity, reliability,

and efficiency [3,4]. So it is important to design control strategy to keep the safe operation of these

systems during normal and faulty conditions. The grid-connected power electronic converters

are highly sensitive to voltage disturbances. This makes it is necessary to reduce the effects of voltage

orrespondence to: A. Hajizadeh, Electrical and Computer Engineering Department, K.N.Toosi University of

hnology, Seyedkhandan, Dr. Shariati Ave, P.O. Box 16315-1355, Tehran, Iran.

mail: [email protected]

pyright # 2010 John Wiley & Sons, Ltd.

ROBUST CONTROL OF HYBRID FUEL CELL 523

disturbances on their operations [5]. A voltage sag is a drop in voltage with duration between one half-

cycle and 1 minute [6], which is, in most cases, caused by a short-circuit fault. The operation of DG

units under voltage sag has not received much attention in the past research. Moreover, many grid

operators demand the immediate shutdown of DG in case of grid disturbances as a prerequisite for grid

connection.

As the power generated by DG units increases, the disconnection of DG units stresses the utility

grid and could cause power unbalance; which may turn into instability. So, the interaction between

DG units and the grid during the voltage sag is very important and it must be considered when

designing the proper control strategy. In this paper a control strategy is proposed for FC power

systems combined with energy storage under voltage sag conditions. Up to now, a considerable

study has not been presented for operation of FC DG systems under voltage sags. Much research

has concentrated only on control of grid connected voltage source converter (VSC) [7–10].

Authors in Ref. [11] have presented a control strategy for hybrid FC/battery energy DG system

with voltage sag ride-through capability. But it has been supposed that the voltage sag ride-through

can be achieved only by increasing the current rating of the VSC and the role of energy storage has

not been considered. During the voltage sag, a decrease in voltage magnitude at the grid-connected

converter is occurs. In this case, in order to avoid overloading of the converter, current controllers

limit the power of DG units that can be supplied to the grid during voltage sag. For FC DG systems,

the power limitation can be a problem due to the slow dynamics of FC. So it needs to combine FC

with energy storage to respond the rapid changes of power demand. The comparison between

energy storages shows that the use of supercapacitor is suitable as short term energy storage [12].

Short-term storage is the storage that maintains energy reserves sufficient to provide rated power

from a few seconds up to a minute. The two basic applications are; for use in uninterruptible

power supply (UPS) and for power system stabilization. Due to the lower power density than

supercapacitors, the batteries cannot charge or discharge fast enough during the voltage sag.

Additionally, the batteries have a slow charging time, limited by a charging current; in contrast,

the supercapacitor can be charged in a short time, depending on the availability of a high charging

power from the main source.

In most of today’s DG applications, batteries are placed as an assistant power source [13] to deliver

the power for long time. So it is important to study the operation and the behavior of FC DG system

combined with supercapacitor energy storage under voltage sag conditions.

In this paper a control strategy has been proposed for hybrid FC/energy storage power generation

system in weak grid during voltage sag conditions. The term weak grid usually refers to systems where

the voltage level is not constant. First, dynamic modeling of each component of the hybrid system is

described. Then control structure of the hybrid power generation system is investigated. Simulation

results prove the effectiveness of the proposed control strategy.

2. DESCRIPTION OF HYBRID FUEL CELL/ENERGY STORAGE DISTRIBUTED

GENERATION SYSTEM

The structure of hybrid FC/energy storage DG system that investigated in this paper is shown in

Figure 1. As shown, it consists of FC, energy storage, DC–DC power converters, grid connected

converter, and transformer and output filter. The whole hybrid DG system is considered to study its

interaction between utility grid during voltage sag.

In this section, mathematical equations and modeling of each component for the proposed hybrid

system is presented.

2.1 Modeling and control of fuel cell subsystem

FCs are static energy conversion devices that convert the chemical energy of fuel directly into electrical

energy. The model of FC power plant used in this study is based on the dynamic proton exchange

membrane fuel cell (PEMFC) stack model developed in Ref. [14]. The performance of FC is affected

by several operating variables, as discussed in the following. This model is based on simulating the

Copyright # 2010 John Wiley & Sons, Ltd. Euro. Trans. Electr. Power 2011; 21:522–540

DOI: 10.1002/etep

Figure 1. Proposed structure of hybrid fuel cell/energy storage distributed generation system.

524 A. HAJIZADEH, M. A. GOLKAR AND L. NORUM

relationship between output voltage and partial pressure of hydrogen, oxygen, and water. The Nernst’s

equation and Ohm’s law determine the average voltage magnitude of the FC stack. The following

equations model the voltage of the FC stack:

Vfc ¼ N0ðE0Þ þRT

2Flog

PH2P0:5

O2

PH2O

! !� RintI (1)

PH2, PH2O, and PO2

are determined by the following differential equations:

PH2

:¼ � 1

tH2

PH2þ 1

KH2

ðqinH2

� 2KrIfcÞ� �

PH2O

:¼ � 1

tH2OPH2O þ 2

KH2OKrIfc

� �PO2

:¼ � 1

tO2

PO2þ 1

KO2

ðqinO2

� KrIfcÞ� � (2)

The Kr constant is defined by the relation between the rate of reactant hydrogen and the FC current:

qrH2

¼ N0Ifc

2F¼ 2KrIfc (3)

Moreover, a simple model of reformer that generates hydrogen through methane has been

considered. The model is second-order transfer function. The mathematical form of the model can be

written as follows [15]:

qH2

qmethane

¼ CV

t1t2s2 þ ðt1 þ t2Þsþ 1(4)

The FC cannot immediately respond to power demand during start-up or sudden load changes due to

its slow dynamics. Usually to connect a FC to an external power system, it is necessary to boost the FC

voltage or to increase the number of cells. The role of the DC–DC boost converter is to increase the FC

voltage, to control the FC power, and to regulate the voltage. Figure 2 shows the DC–DC converter

Figure 2. Boost DC/DC converter model.


DOI: 10.1002/etep


model. This boost converter is described by the following two nonlinear state space averaged

equations [13]:

diL

dt¼ �RL

L� 1 � d

L

� �VC þ 1

LU

dVC

dt¼ 1 � d

CiL � iO

C

(5)

In order to design control strategy for FC power plant, two parameters should be considered and

regulated. These parameters are hydrogen flow according to output power and FC current. According

to Equations (2) and (3), to control hydrogen flow from the FC, a feedback from the stack current is

considered. Equation (3) shows that the reacting fuel quantity,qrH2

, is directly proportional to the output

current, Ifc, the factor Kr being a cell constant. Hence, the desired utilization is translated to

corresponding output current demand:

qinH2

¼ 2Kr

Uf;opt

Idemand (6)

A proportional integral (PI) controller is used to control the flow rate methane in the reformer.

Oxygen flow is determined using the hydrogen–oxygen flow ratio rH–O. The control strategy for

reformer is illustrated in Figure 3.

In the proposed control structure, choosing the control system parameters affects the system

performance. So it is important to design PI controller properly. Another important parameter that must

be controlled properly is FC current. For this purpose, a boost DC–DC converter is selected for FC

converter. The current mode control of DC–DC converter has been used to regulate the FC current. A

typical range of Uf is 80–90% [4], which ensures that the operational limits mentioned above are

observed. The corresponding limitation for the demand current is then calculated as following equation:

0:8qinH2

2Kr

¼ Ifc min � Ifc ref � Ifc max ¼0:9qin

H2

2Kr

(7)

To obtain transfer function of FC current loop and apply classical control analysis and design

methods (such as Nyquist criterion, Bode plots) in converter controls, the following transfer function

can be investigated based on the state space signal models of boos DC–DC converter model [3,16]:

TiFC�dðsÞ ¼ iFCðsÞ

dðsÞ ¼ 1

s2 þ RL

Lsþ 1�D2

LC

sVCO

L

� �þ ILOð1 � DÞ

LC

� �(8)

The control structure for FC current with has been shown in Figure 4. In this structure, the transfer

function of PWM block can be modeled as:

TPWM ¼ 1

KPWM

(9)

2.2 Modeling and control of supercapacitor subsystem

Supercapacitors store electrical energy by accumulating charge on two parallel electrodes separated by

a dielectric material. The capacity represents the relationship between the electric charge stored in the

Figure 3. Reformer controller layout.


DOI: 10.1002/etep

Figure 4. The block diagram of fuel cell converter control strategy.


capacitor and voltage between the two electrodes of the capacitor. The classical equivalent circuit of

the supercapacitor is shown in Figure 5 [17]. The SOC is the only state variable of the energy storage

system:

dSOC

dt¼ IES

Qm

(10)

Since the EPR models leakage effects and influences long-term energy storage performance of the

supercapacitor, only the ESR will be taken into account.

The supercapacitor energy storage connects to the DC-link by using bidirectional converter. This

converter topology is slightly more complex than the boost converter since includes one additional

switch. But it has the advantage of allowing bidirectional power flow, which means energy can flow

from the energy source to the load and back from the load to the energy source. This feature is very

convenient for energy storage devices like batteries and supercapacitors, because it allows to

recharge the device after each time their energy is used. The schematic of buck boost converter

is shown in Figure 6. The average model of such a converter could is defined by the following

equations [18]:

diL

dt¼ ð1 � dÞ 1

LvO þ d

Vin

L

dvO

dt¼ ð1 � dÞ 1

CO

iL � iO

CO

(11)

The control input for the above model is the function d, called duty cycle function.

Figure 5. Equivalent model for the supercapacitor bank.

Figure 6. Buck-boost converter model.


DOI: 10.1002/etep


Two important objectives must be satisfied by proper control of buck-boost converter: regulation of

supercapacitor voltage and providing the requested power from supercapacitor. For this purpose the

following equation is considered:

L disc

dt¼ dvsc � ð1 � dÞvdc � Rscisc

Cdcdvdc

dt¼ ð1 � dÞisc � iO

isc ¼ �Cscdvsc

dt

(12)

So, according to the above nonlinear equations and need for fast response during the transient power,

a nonlinear sliding mode control (SMC) [18] is designed for control of buck-boost converter. To design

the SMC, first the sliding surface should be defined. For this purpose the following surface as a function

is considered:

S ¼ ðisc � isc�refÞ þ ðiO � iO�refÞ (13)

According to this criterion for sliding surface, the requested current (power) that should be produced or

absorbed by supercapacitor during voltage sag is determined. Moreover, the voltage of supercapacitor is

regulated. In Figure 7 the details of the proposed control structure for buck-boost converter is shown.

2.3 Modeling and control of grid connected voltage source converter

A three-phase equivalent circuit of VSC is shown in Figure 8. To reduce harmonics, LCL filter is connected

between the converter and the grid [19]. The dynamic model of the three-phase VSC is represented in:

di1k

dt¼ �R1

L1

i1k þ1

L1

ðvik � vckÞ

di2k

dt¼ �R2

L2

i2k þ1

L1

ðvck � vskÞ

CfdVck

dt¼ i1k � i2k

(14)

where k¼ {a, b, c}.

Figure 7. Proposed control structure for buck-boost converter.

Figure 8. Three-phase dc/ac voltage source converter.


DOI: 10.1002/etep


Control of grid connected VSC is an important problem during voltage disturbances. It needs fast

current controllers to track the current references according to change in active and reactive power

during the fault. The current controller used in this paper consists of two vector current controllers

based on SMC that regulate the positive and negative sequence currents separately and are

implemented in two different rotating coordinate systems. The need for regulating the positive and

negative sequence currents is related for treating the unbalance voltage sag.

A simplified scheme for the proposed control strategy is shown in Figure 9. The three phase grid

currents and voltages are sampled and transformed into its positive and negative sequence components.

The positive and negative sequence of dq-components are then used along the with the reference

current signals to produce the reference voltage signals for the PWM regulator. A sequence separation

method (SSM) is needed to extract positive and negative sequences. Delayed signal cancellation

method (DSC) is probably the best suited SSM [10]; but produces transient oscillations at the start and

end period of voltage sag [5]. The layout of this method is illustrated in Figure 10. The abc system is

first transformed into stationary ab reference frame using Clark’s transformation, and then it is delayed

for T/4. The positive and negative sequences can be calculated by adding or subtracting the present

real-time signal with delayed signal in the following way:

vapðtÞ

vbpðtÞ

vanðtÞ

vbnðtÞ

0BB@

1CCA ¼ 1

2

1 0 0 � 1

0 1 1 0

1 0 0 1

0 1 � 1 0

2664

3775

vaðtÞvbðtÞ

va t � T4

� �vb t � T

4

� �0BB@

1CCA (15)

Figure 9. Block diagram of current control strategy.

Figure 10. Separation positive and negative sequence by ‘‘signal delay cancelation’’ method.


DOI: 10.1002/etep


According to the proposed control strategy, the purpose of the current controller is to synthesize a

voltage correction vector so that the current error vector can be kept to a minimum value.

In this paper, the current controller has been implemented by using SMC technique due to its

robustness and overshoot-free fast tracking capability [20].

The SMC is a nonlinear control approach which complies with the nonlinear characteristic of a

power electronic converter. Such control technique is robust even against the plant parametric variation

and can compensate the modeling approximations. Also, it is characterized by a good dynamic

response. In addition, the SMC is simple to implement.

According to the dynamic model of inverter given in (14), the state space equations of the system can

be written as (16):

X:ðtÞ ¼ AXðtÞ þ BU þ EVg

YðtÞ ¼ CXðtÞ (16)

where the state variable is X, the control input U, and grid voltage Vg and given in (17–19):

X ¼ ip2d i

p2q in2d in2q

h iT

; (17)

U ¼ Upd Up

q Und Un

q

h iT

; (18)

Vg ¼ Vpgd Vp

gq Vngd Vn

gq

h iT

(19)

For the control plant given in (16), the SMC law can be derived as follows. To let the current output Y

track the reference input iref, a sliding mode manifold can be chosen in the form of (20):

Sdqp

Sdqn

� �¼ i

p2dq

in2dq

� ��

ip;ref2dq

in;ref2dq

" #; (20)

where ip;ref2dq and i

n;ref2dq are the specified current vector commands for the positive and negative sequence

of dq-components.

The sliding mode can be reached if the control input U(t) is designed to be the solution of (21):

d

dtSðtÞ ¼ 0 (21)

The control law satisfies (21) is called equivalent control and is given by (22):

UeqðtÞ ¼ ðCBÞ�1½iref � CAXðtÞ � CEVgðtÞ� (22)

In order to generate proper current references, consider the complex apparent power from the grid:

Sg ¼ ðvpsdq ejvt þ vn

sdq e�jvtÞðip2dq ejvt þ in2dq e�jvtÞ� ¼ðPþ P2c cosð2vtÞ þ P2s sinð2vtÞÞþjðQþ Q2c cosð2vtÞ þ Q2s sinð2vtÞÞ

(23)

By expanding (23), the following expression in matrix form can be written:

P ¼ ðvpsdi

p2d þ vp

sqip2q þ vn

sdin2d þ vn

sqin2qÞ

Q ¼ ðvpsqi

p2d � v

psdi

p2q þ vn

sqin2d � vn

sdin2qÞ

Pc2 ¼ ðvpsdi

n2d þ vp

sqin2q þ vn

sdip2d þ vn

sqip2qÞ

Ps2 ¼ ðvpsdi

n2q � vp

sqin2d � vn

sdip2q þ vn

sqip2dÞ

Qc2 ¼ ðvpsqi

n2d � v

psdi

n2q � vn

sdip2q þ vn

sqip2dÞ

Qs2 ¼ ðvpsdi

n2d þ vp

sqin2q � vn

sdip2d � vn

sqip2qÞ

; (24)

where P and Q are the constant active and reactive power, respectively, while the subscripts Ps2 and Pc2

represent the second harmonic sine and cosine component of the active power. These are the oscillating


DOI: 10.1002/etep


active powers due to the unbalance in the grid voltages. During generating the reference currents, the

oscillating reactive powers (Q2c, Q2s) cannot be included in the calculation. Therefore to simplify the

calculation and work with an invertible matrix (4� 4), oscillating reactive power is not controlled and

will flow through the system [5].

Hence, the reference currents can be calculated as follows:

ip�2d

ip�2q

in�2d

in�2q

2664

3775 ¼

vpsd vp

sq vnsd vn

sq

vpsq � v

psd vn

sq � vnsd

vnsq � vn

sd � vpsq vn

sd

vnsd vn

sq vpsd vp

sq

2664

3775�1

P�

Q�

�DPs2

�DPc2

2664

3775; (25)

DP ¼ 2 � ðR1 þ R2Þ � ððip2dÞ2 þ ðip2qÞ

2 þ ðin2dÞ2 þ ðin2qÞ

2Þ; (26)

DPc2 ¼ 2 � ðR1 þ R2Þ � ðip2din2d þ i

p2qi

n2qÞ þ 2 � v� L� ðip2di

n2q � i

p2qi

n2dÞ; (27)

DPs2 ¼ 2 � ðR1 þ R2Þ � ðip2din2q � i

p2qi

n2dÞ þ 2 � v� ðL1 þ L2Þ � ð�i

p2di

n2d � i

p2qi

n2qÞ (28)

This algorithm calculates current references by setting active and reactive power references (P�, Q�),

and by forcing the oscillating active power demanded by the filter to be delivered from the grid

(P�2c ¼ �DP2c; P�

2s ¼ �DP2s). Then, no oscillating active power flows between the dc link and the filter.

2.4 Weak grid model

The term weak grid usually refers to systems where the voltage level is not constant. In other words, the

week grid impedance is significant and cannot be ignored in the analysis or development tasks. The

single line diagram of the simulated scaled distribution network model is shown in Figure 11 [21]. It is

supposed that the weak grid system consists of three identical p sections of series inductors and shunt

capacitors. The hybrid DG system is connected at the PCC to the network.

2.5 Load model

The load model that has been considered here is an R–L load. The three-phase parallel R–L load block

implements a three-phase balanced load as a parallel combination of R–L elements. At the specified

frequency, the load exhibits a constant impedance. The active and reactive powers absorbed by the load

are proportional to the square of the applied voltage.

3. POWER CONTROL OF HYBRID FUEL CELL/ENERGY STORAGE DURING VOLTAGE

SAG

The term, ‘‘power control,’’ refers to the design of the higher-level control algorithm that determines

the proper power level to be generated, and its split between FC and energy storage dc power sources.

Figure 11. Weak grid model.


DOI: 10.1002/etep


This makes that the dc link power be stabilized and regulated the dc link voltage consequently. During the

voltage sag condition, a decrease in voltage amplitude occurs at the converter terminal. It makes an

increasing of the output currents of inverter. In this case, the power control strategy should limit the power at

the dc side of VSC to avoid overloading of the converter. This will thus limit the power that the DG unit can

supply to the grid during voltage sag, resulting the dc-link voltage will increase. To avoid a too high dc-link

voltage, the power balance between inverter power and DG power must be satisfied. One existing method to

solve these issues is to install energy storages which absorb power from dc link. In Figure 12 the flow of

power during different operating conditions in dc link is shown. If the losses in both the FC converter and

supercapacitor converter are neglected, the following differential equation for dc link is given:

Cdcvdc

dvdc

dt¼ PFC þ PSC � Pgrid ¼ vFCiFC þ vSCiSC � Pgrid; (29)

where PFC is the FC power, PSC is the supercapacitor power, and Pgrid is the grid power.

According to Equation (29), in order to regulate the dc link voltage it is necessary to keep the power

balance in dc link. In this equation, the change in grid power is considered as disturbance during the

voltage sag. Moreover it is important to consider the dynamic limitations of FC power to meet the

balance of power in dc link. In this case, the FC power could not change rapidly and the FC controller

with DC–DC converter should regulate the operating point of FC. But the amount of power that should

be absorbed by supercapacitor to balance the power in dc link is very important and it depends on the dc

link energy. The dc link energy measurement is carried out by means of the following calculation:

EdcðkÞ ¼1

2

� �CdcV

2dcðkÞ (30)

In this paper, a power flow control structure has been developed for hybrid power sources during

voltage sag. It is based on self-tuning fuzzy logic control (FLC) strategy that determines the

supercapacitor power according to the following inputs:

eðkÞ ¼ Edc�refðkÞ � EdcðkÞDeðkÞ ¼ eðkÞ � eðk � 1Þ ; (31)

where Edc-ref is the reference dc link energy which calculated by reference dc link voltage.

Hence, it is essential to design robust and stable control strategy to guarantee the stability of the dc

link of hybrid system. For this purpose, a robust self-tuning fuzzy PI control structure has been

developed [22]. The block diagram of the fuzzy PI controller is shown in Figure 13.

By using the scaling factors (SFs) Ge, GDe, the quantities e and De are converted to normalized eN

and DeN. These normalized quantities eN and DeN are crisp in nature and therefore need to be first

converted to their corresponding fuzzy variables. After fuzzification, the fuzzified inputs are given to

the fuzzy inference mechanism which, depending on the given fuzzy rule base, gives the normalized

incremental change in control output (DuN). The output DuN is converted into actual incremental

change in control output (Du) by using the SF Gu. For the implementation the fuzzy inference engine,

Figure 12. Power flow in hybrid fuel cell/energy storage power system.


DOI: 10.1002/etep

Figure 13. The block diagram of the self-tuning fuzzy PI controller.


the ‘‘min’’ operator for connecting multiple antecedents in a rule, the ‘‘min’’ implication operator, and

the ‘‘max’’ aggregation operator have been used [23]. Actually, the output DuN from the inference

mechanism is fuzzy in nature, hence, to determine the crisp output, the defuzzification stage is applied.

The centroid defuzzification Scheme [22] has been used here for obtaining the output Du. Finally, the

actual value of the controller output (u) is computed by:

uðkÞ ¼ uðk � 1Þ þ DuðkÞ (32)

The relationships between the SFs and the input and output variables of the self-tuning FLC are as

follows:

eN ¼ Gee

DeN ¼ GDeDe

Du ¼ ðaGuÞDuN

(33)

In this scheme, the FLC is tuned on-line (while the controller is in operation) by dynamically

adjusting its output scale factor by a gain updating factor (a). The value of a is determined from a rule

base defined on e and De and derived from the knowledge of control engineering. Generally, selection

of suitable values for Ge, GDe and Gu are made based on the knowledge about the process to be

controlled and sometimes through trial and error to achieve the best possible control performance. This

is so because, unlike conventional non-fuzzy controllers, there is no well-defined method for selecting

appropriate values of SFs for FLC.

Each fuzzy control rule in the controller rule base is of the form:

‘‘If e is E and De is DE, then Du is DU’’where E, DE, and DU are the fuzzy sets corresponding to error,

change in error, and the incremental change in the control output, respectively. In this work, for both the

inputs (e and De) and the output (Du), seven fuzzy subsets have been used. These are: positive big (PB),

positive medium (PM), positive small (PS), zero (ZE), negative small (NS), negative medium (NM), and

negative big (NB). For each of these fuzzy sets, triangular membership function (MF) has been used [22].

From this figure it is observed that the triangles are symmetric with equal base having 50% overlap

with neighboring MFs. As each of the two inputs has seven fuzzy sets, there are altogether 49 control

rules in the FLC. The rule base for computing the output Du is shown in Table I which is a widely used

rule base designed with a two dimensional phase plane. The control rules in Table I are built based on

the characteristics of the step response. Moreover, the gain updating factor (a) is calculated using fuzzy

rules of the form:

If e is E and De is DE then a is a.


DOI: 10.1002/etep

Table I. Rule base for computing the output Du.

De/e NB NM NS ZE PS PM PB

NB NB NB NB NM NS NS ZENM NB NM NM NM NS ZE PSNS NB NM NS NS ZE PS PMZE NB NM NS ZE PS PM PBPS NM NS ZE PS PS PM PBPM NS ZE PS PM PM PM PBPB ZE PS PS PM PB PB PB


From Figure 9 it is observed that the value of a is computed from the normalized values of e and De by a

fuzzy rule base. The MFs used for e and De are exactly same as those used in FLC. Moreover, the same

fuzzy operators also been used in this case. The MFs for the factor a are defined in the domain [0,1]. As

each of the two inputs (e and De) to the fuzzy rule base (corresponding to a) has seven fuzzified variables,

the rule base has 49 rules for computing the value of a. Table II shows the rule base for computing a. This

rule base has been designed to improve the control performance under large disturbances such as three-

phase short circuit on the transmission lines, a sudden loss of generating unit or a large loss of load, etc. For

example, immediately after a large disturbance, e may be small but De will be sufficiently large (they will

be of same sign) and, for this case, a is supposed to be large to increase the gain. Therefore, under these

circumstances, the appropriate rules are ‘‘IFe is PS and De is PM THEN is B’’ or ‘‘IFe is NS and De is NM

THEN a is B.’’ On the other hand, for steady state conditions (i.e., e�0 and De�0), controller gain should

be very small (e.g., IFe is ZE andDe is ZE THEN a is ZE) to avoid chattering problem around the set point.

Further justification for using the rule base in Table II can be found in Ref. [24].

4. SIMULATION RESULTS

In order to show the effectiveness of proposed control strategy, the simulation model of the proposed

hybrid DG system has been built in Matlab/Simulink environment. The parameters of the hybrid FC/

energy storage DG system in this study are given in Table III. In this case study, the output power of the

FC power system during normal conditioning is considered to 50 kW. Moreover the supercapacitor

bank has is capable of sustaining the extra load of 20 kW during the voltage disturbances as short term

energy storage. So it is supposed that initial voltage for supercapacitor is 72 V. The average current to

be supplied by the supercapacitor bank corresponds to the average of the maximum and minimum

current demands from the supercapacitor bank, that is:

imax ¼ P

Vmin

¼ 20 000

48¼ 416:6A

imin ¼ P

Vmax

¼ 20 000

96¼ 208:3A

iavg ¼ ð416:6 þ 208:3Þ2

¼ 312:5A

(34)

Table II. Rule base for computing the output a.

De/e NB NM NS ZE PS PM PB

NB VB VB B SB S S ZENM VB VB B B MB S VSNS VB VB B VB VS S VSZE S SB MB ZE MB SB SS VS S VS VB B MB VBPM VS S MB B B VB VBPB ZE S SB B VB VB VB

V, very; S, small; M, medium; B, big; N, negative; P, positive; ZE, zero.


DOI: 10.1002/etep

Table III. Parameters of the hybrid DG system.

Fuel cell system parametersFaraday’s constant (F) 96 484 600 C/kmolHydrogen time constant (tH2

) 26.1 sHydrogen valve molar constant (KH2

) 8.43� 10�4

kr constant¼N0/4F 9.9497� 10�7

No load voltage (E0) 0.6 VNumber of cells (N0) 384Oxygen time constant (tO2

) 2.91 sOxygen valve molar constant (KO2

) 2.52� 10�3

FC internal resistance (r) 0.126VFC absolute temperature (T) 343 KUniversal gas constant (R) 8314.47 J/(kmol K)Utilization factor (Uf) 0.8Water time constant (tH2O) 78.3 sWater valve molar constant (KH2O) 2.81� 10�4

Boost DC–DC converter parametersRated voltage (V) 200/650 VRated power 50 kWNominal duty cycle 0.6

Supercapacitor bank parametersMaximum allowable terminal voltage 96 VMinimum allowable terminal voltage 48 VOperating terminal voltage 72 VSupercapacitor capacitance (Csc) 165 F

Voltage source converter parametersNominal AC voltage 400 VNominal phase current 125 ANominal DC voltage 650 VDC-link capacitance 550mFR1 1.6 mVL1 0.52 mHR2 0.6 mVL2 0.2 mHCf 137.83mFfs 50 Hz

Weak grid parametersRd 0.05VLd 2.05 mHCd 46mF


For the proposed application, Maxwell Technologies Boost-cap1 BMOD0165 type supercapacitor

is used. The specification of the supercapacitor is given in Table I (http://www.maxwell.com/pdf/uc/

datasheets/20090428_48V_series_1009365.6.pdf). In order to meet the power demand, four

supercapacitor modules are used, where two strings of two modules connected in series. The total

capacitance (Ctotal) and resistance (Rtotal) may be obtained as:

Ctotal ¼ Cmodulenp

ns¼ 165 � 2

2¼ 165 F

Rtotal ¼ Rmodulenp

ns¼ 6:1 � 10�3 � 2

2¼ 6:1 mV

; (35)

where np and ns are the number of parallel and series connected capacitor, respectively.

The system was tested under several different operating conditions to investigate its power

management and load-following capabilities as well as its stability during unbalanced and balanced

voltage sags. The unbalance voltage sag that implemented for simulation is voltage sag type C with

60% magnitude.

4.1 Operation under normal conditions

A certain amount of power may be scheduled to be delivered to a load center from the utility grid with

the rest to be supplied by the hybrid FC system. Therefore, a proper controller must be designed to


DOI: 10.1002/etep

Figure 14. Active and reactive powers delivered to the grid from hybrid DG system and the correspondingreference powers.


ensure that only scheduled power is delivered from the grid and that the FC system follows the

remainder of load demand. Figure 14 shows active and reactive powers delivered to the grid from

hybrid DG system and the corresponding scheduled reference powers. Both show that the control

strategy quickly tracks the references with little overshoot. Similarly Figure 15 depicts the FC and

supercapacitor powers during normal operation. As shown, when the reference of active power

changes, the supercapacitor power also changes to deliever part of power during transient to stabilize

DC-link power. Moreover, according to slow dynamics of FC, its generated power changes slowly.

4.2 Operation under unbalance voltage sag

The proposed control strategy has been examined in case of unbalanced voltage sags. A volatg sag,

resulting from two-phase fault (type C), is applied at the grid side. The voltage sag starts at 0.7 seconds

for duration of 0.1 seconds with voltage drop of 60%. The grid voltage during voltage sag type C has

been shown in Figure 16a. As illustrated in Figure 16b, the grid currents increase to above 1.5 p.u. In

this case, the grid currents are limited by current controllers to avoid overloading of the converter.

Also after voltage sag, the current controllers adapt fast according to the new current references to

shape the grid currents. In Figure 16c, the average and instantaneous active power during normal and

Figure 15. Fuel cell and supercapacitor powers.


DOI: 10.1002/etep

Figure 16. Simulation results during voltage sag type C.


voltage sag conditions have been presented. As shown during voltage sag the injected active power to

the grid is decreased and there are small oscillations on instantaneous active power. These oscillations

are related to the proposed current control strategy that considered in this paper. It makes that the

oscillating active power that demands by the filter to be delivered from the grid and no oscillating active

power flows between the dc link and the filter. Moreover, there are large oscillations on reactive power

that is shown in Figure 16d. But the average of reactive power is 0. Because the reference of reactive

power that considered for this simulation is 0, it is supposed that there is no voltage regulation in

proposed control strategy. The DC-link voltage is shown in Figure 16e. During the voltage sag, there is

an increasing on dc link voltage but it is not much more than 10% of nominal value. In these conditions,

to stabilize the dc-link power, the fuzzy controller manages the power flow between FC and

supercapacitor. Hence the reference power of the energy storage bank changes for decreasing the input

power on DC-link during voltage disturbance and the proposed control strategy makes the operating

mode of the energy storage bank goes to the charging mode. In Figure 16e and g, the FC output and

energy storage power are presented. Because of the slow variations of the FC, the output power of the

FC power plant remains constant during the voltage sag and for stabilizing the amount of power on DC-

link; the energy storage absorbs the extra power. According to Figure 16h, it is clear that the output

voltage of FC system has not changed during the occurring the voltage disturbances in distribution

systems and it is very important on the life time of FC system. On the other hand, during the voltage

sag, the energy storage energy storage makes to improve the performance of the hybrid system and the

generated power oscillations during the voltage sag compensate by the supercapacitor energy storage.

In this condition, the proposed power flow control strategy finds the suitable operating point for

the energy storage and FC power sources.

4.3 Operation under balanced voltage sag

For the evaluation of proposed control strategy, the hybrid system also has been studied under balanced

voltage sag with 40%. The duration of voltage sag is considered 0.1 seconds and the response of whole


DOI: 10.1002/etep

Figure 17. Simulation results during balance voltage sag.


system is shown in Figure 17. As shown in Figure 17, in this case all requirements for voltage sag ride-

through have been satisfied by the proposed control strategy and it is able to act well under different

voltage sag conditions. According to the simulation result that presented about the response of hybrid

FC/energy storage DG system, it is determined that the need for integration of energy storage with FC

power source is necessary. In fact, the energy storage makes to keep the stability of DG system under

different voltage disturbances. Moreover, by designing robust control strategy for DC–DC converters

and grid connected inverter, the proper operation of hybrid system is guaranteed during voltage sag in

distribution system. One of the main advantages of proposed control strategy is its robustness under

different voltage sag condition. Another result that may achieved from the simulation results of the

proposed control strategy is the calculation of the necessary amount of energy storage that is need for

voltage sag ride-through of FC DG systems. For this purpose, the system is runned for different

magnitude of voltage sag for both unbalanced and balanced voltage sag.

5. CONCLUSION

This paper presents the dynamic modeling and control of hybrid FC/energy storage DG system under

different operating conditions. For this purpose, complete model of hybrid system is presented and then

by designing control strategy for each component, the power control problem of the proposed system is

studied under unbalanced and balanced voltage sag. To control of the dc link power and regulate of DC-

link voltage, self-tuning fuzzy controller is presented to determine the necessary energy storage power

according to the amount of energy in DC-link. Moreover, robust current control strategy has been

developed to control of positive and negative sequence of dq-components. Simulation results show that

the proposed control strategy is able to tolerate under various voltage sags and keep the system

performances like active power control and stability of DC-link.


DOI: 10.1002/etep


6. LIST OF SYMBOLS

Fuel cell parameters

CV c

Copyright #

onversion factor

E0 v
oltage associated with the reaction free energy
F F
araday’s constant
Ifc c
urrent of the fuel cell stack
N0 n
umber of cells connected in series
qmethane m
ethane flow rate
qinH2

m
olar flow of hydrogen
qinO2

m
olar flow of oxygen
R u
niversal gas constant
Rint in
ternal resistance of fuel cell stack
rH–O h
ydrogen–oxygen flow ratio
T f
uel cell temperature
Uf u
tilization factor
t1, t2 r
eformer time constants
DC–DC converter parameters

d d
uty cycle of the switching device
D n
ominal duty cycle (chosen at 0.7)
iL in
ductor current
iO o
utput current
ILO n
ominal current of inductor
KPWM a
mplitude of the PWM saw tooth carrier signal
U in
put voltage
VC o
utput capacitor voltage
VCO n
ominal voltage of output capacitor
Supercapacitor parameters

C t
otal capacitance of supercapacitor
EPR e
quivalent parallel resistance representing the self-discharging losses
ESR e
quivalent series resistance representing the charging and discharging resistance
IES o
utput current of supercapacitor
isc s
upercapacitor current
Qm m
aximum energy of supercapacitor
SOC s
tate of charge of supercapacitor
vsc s
upercapacitor voltage
DC-link parameters

Edc D
C-link energy
PFC f
uel cell power
Pgrid g
rid power
PSC s
upercapacitor power
vdc D
C-link voltage
Voltage source converter parameters

idq d
q components of current
indq np

egative sequence of dq currents

idq p
ositive sequence of dq currents
i1 i
nductor current (1)
2010 John Wiley & Sons, Ltd. Euro. Trans. Electr. Power 2011; 21:522–540

DOI: 10.1002/etep


i2 i

Copyrig

nductor current (2)

Pg o
utput active power of inverter
Pc2g s
econd harmonic cosine component of the active power
Ps2g s
econd harmonic sine component of the active power
Qg o
utput reactive power of inverter
vc c
apacitor voltage
vi i
nverter output voltage
vs g
rid voltage
vdq d
q components of voltage
vab a
b components of voltage
vpdq p
ositive sequence of dq voltages
vndq n
egative sequence of dq voltages
vnab n
egative sequence of ab voltages
vpab p
ositive sequence of ab voltages
REFERENCES

1. Bayod-Rujula AA. Future development of the electricity systems with distributed generation. Energy 2009; 34:377–

383 10.1016/j.energy.2008.12.008.

2. Ellis MW, VonSpakovsky MR, Nelson DJ. Fuel Cell Systems: Efficient, Flexible Energy Conversion for the 21st

century. in Proceedings of the IEEE, 2001; 89: 1808–1818.

3. Wang C, Nehrir NH, Gao H. Control of PEM fuel cell distributed generation systems. IEEE Transaction on Energy

Conversion 2006; 21:586–595. 10.1109/TEC.2005.860404.

4. Golkar MA, Hajizadeh A. Control strategy of hybrid fuel cell/battery distributed generation system for grid-

connected operation. Journal of Zhejiang University Science A 2009; 10:488–496. 10.1631/jzus.A0820151.

5. Saccomando G, Svensson J. Transient Operation of Grid Connected Voltage Source Converter under Unbalanced

Voltage Conditions. in Proceedings of the IEEE Industry Application Society Annual Meeting, 2001; 4: 2419–2424.

6. Bollen MH. Understanding Power Quality Problems: Voltage Sags and Interruptions, IEEE Press: New York, 1999.

7. Rendersa B, Ryckaert WR, De Gusseme K, Stockman K, Vandevelde L. Improving the voltage dip immunity of

converter-connected distributed generation units. Renewable Energy 2008; 33:1011–1018. 10.1016/

j.renene.2007.05.023.

8. Rodrıguez P, Timbus AV, Teodorescu R, Liserre M, Blaabjerg F. Independent PQ control for distributed power

generation systems under grid faults. in Proceedings of the 32nd Annual Conference IEEE IECON, 2006;

5185–5190.

9. Magueed FA, Sannino A, Svensson J. Transient performance of voltage source converter under unbalanced voltage

dips. in Proceedings of the IEEE PESC, Aachen, Germany, 2004; 1163–1168.

10. Sannino A, Bollen MH, Svensson J. Voltage tolerance testing of three-phase voltage source converters. IEEE

Transaction on Power Delivery 2005; 20:1633–1639. 10.1109/TPWRD.2004.833881.

11. Hajizadeh A, Golkar MA. Control of Hybrid Fuel Cell/Battery Distributed Power Generation System with

Voltage Sag Ride-Through Capability. 2nd IEEE International Conference on Power and Energy (PECon 08),

2008; 463–467. 10.1109/PECON.2008.4762510.

12. Sahay K, Dwivedi B. Energy storage technology for performance enhancement of power systems. Electrical Power

Quality & Utilization Magazine 2009; 4:1–12.

13. Hajizadeh A, Golkar MA. Intelligent power management strategy of hybrid distributed generation system.

International Journal of Electrical Power and Energy Systems 2007; 29:783–795. 10.1016/j.ijepes.2007.06.025.

14. El-Sharkh MY, Rahman A, Alam MY, Byrne PC, Sakla AA, Thomas T. A dynamic model for a stand-alone

PEM fuel cell power plant for residential applications. Journal of Power Sources 2004; 138:199–204. 10.1016/

j.jpowsour.2004.06.037.

15. Uzunoglu M, Alam MS. Dynamic modeling, design, and simulation of a combined PEM fuel cell and ultracapacitor

system for stand-alone residential applications. IEEE Transactions on Energy Conversion 2006; 21:767–775.

10.1109/TEC.2006.875468.

16. Mohan N, Undeland TM, Robbins WP. Power Electronics: Converters, Applications, and Design, Wiley: New York,

2003.

17. Kisacikoglu MC, Uzunoglu M, Alam MS. Load sharing using fuzzy logic control in a fuel cell/ultracapacitor hybrid

vehicle. International Journal of Hydrogen Energy 2009; 34:1497–1507. 10.1016/j.ijhydene.2008.11.035.

18. Yuri B, Shtessel YB, Zinober ASL, Shkolnikov A. Sliding mode control of boost and buck-boost power converters

using method of stable system centre. Automatica 2003; 39:1061–1067. 10.1016/S0005-1098(03)00068-2.

19. Teodorescu R, Blaabjerg F, Borup U, Liserre M. A new control structure for grid-connected LCL PV inverters with

zero steady-state error and selective harmonic compensation. 2004; 1: 580–586. 10.1109/APEC.2004.1295865.

20. Slotine JJ, Li W. Applied Nonlinear Control, Prentice-Hall: Englewood Cliffs, NJ, 1991.

ht # 2010 John Wiley & Sons, Ltd. Euro. Trans. Electr. Power 2011; 21:522–540

DOI: 10.1002/etep


21. Magueed F, Awad H. Voltage Compensation in Weak Grids Using Distributed Generation with Voltage Source

Converter as a Front End. International Conference on Power Electronics and Drives Systems, 2005; 234–239.

10.1109/PEDS.2005.1619691.

22. Mudi RK, Pal NR. A robust self-tuning scheme for PI- and PD-type fuzzy controllers. IEEE Transaction on Fuzzy

Systems 1999; 7:2–16. 10.1109/91.746295.

23. Jamshidi M. Fuzzy Logic and Control, Software and Hardware Applications, Ch 1–4, University of New Mexico,

PTR Prentice-Hall, Inc.: USA, 1993.

24. Pal K, Mudi RK, Pal NR. A new scheme for fuzzy rule-based system identification and its application to self-

tuning fuzzy controllers. IEEE Transaction on Systems, Man and Cybernetic B: Cybernetics 2002; 32:470–482.

1083-4419(02)04375-3.


DOI: 10.1002/etep

Robust control of hybrid fuel cell/energy storage distributed power generation system in weak grid...

Documents

Transcript of Robust control of hybrid fuel cell/energy storage distributed power generation system in weak grid...