lable at ScienceDirect

Renewable Energy 34 (2009) 1973–1978

Contents lists avai

Renewable Energy

journal homepage: www.elsevier .com/locate/renene

Performance of a PV–wind hybrid system for hydrogen production

Kamaruzzaman Sopian a,*, Mohd Zamri Ibrahim b, Wan Ramli Wan Daud a, Mohd Yusof Othman a,Baharuddin Yatim a, Nowshad Amin a

a Solar Energy Research Institute, Universiti Kebangsaan Malaysia (National University of Malaysia), 43600 Bangi Selangor, Malaysiab Department of Engineering Sciences, Faculty of Science and Technology, University Malaysia Terengganu, Malaysia

a r t i c l e i n f o

Article history:Available online 1 January 2009

Keywords:PV–wind hybrid systemHydrogen productionPEM electrolyserSystem performance

* Corresponding author.E-mail address: ksopian@eng.ukm.my (K. Sopian).

0960-1481/$ – see front matter � 2008 Elsevier Ltd.doi:10.1016/j.renene.2008.12.010

a b s t r a c t

This paper describes the performance of an integrated PV–wind hydrogen energy production system. Thesystem consists of photovoltaic array, wind turbine, PEM electrolyser, battery bank, hydrogen storagetank, and an automatic control system for battery charging and discharging conditions. The systemproduced 130–140 ml/min of hydrogen, for an average global solar radiation and wind speed rangingbetween 200 and 800 W/m2 and 2.0 and 5.0 m/s respectively. A mathematical model for each componentin the system was developed and compared to the experimental results.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

The last few years have seen the development of renewableenergy (RE) systems based mainly on wind and solar power. Suchsystems are especially relevant to off-network communities,remote areas and island, where particularly no grid connectivityis available. These RE systems rely on highly transient energysources and exhibit strong short-term and seasonal variations intheir energy outputs. They thus need to store the energyproduced in period of low demand in order to stabilize theoutput when the demand is high. While batteries are mostcommonly used for this purpose, they typically lose 1–5% of theirenergy content per hour and thus can only store energy for shortperiods of time [1,2]. There are presently no practical meansavailable for long-term storage of excess electrical energyproduced by the RE sources.

Few studies are related to the use of hydrogen for the electricalenergy produced by RE sources such as wind and solar power [3–6].Hydrogen is produced by an electrolyser powered by the excesselectrical energy from the RE source. The hydrogen can then beused to feed an energy conversion device (such as a fuel cell or aninternal combustion engine), which will act as a secondary powersource in periods of high demand. Such a system is set up forremote areas applications such as communication stations. The REsources at HRI are a wind turbine (WT) capable of generating

All rights reserved.

a maximum electrical power of 10 kW and photovoltaic (PV) cells of1 kW maximum power. The excess energy available from the REsources is directed to an electrolyser. The hydrogen produced isthen stored in a pressurized tank. This hydrogen is then fed toa proton exchange membrane fuel cell (PEMFC) system [6] thatwould be used as a load-leveling electrical system when unfavor-able weather conditions arise.

The objectives of these studies include the development andexperimental testing of the best methods for producing hydrogenthrough hybrid RE source especially in tropical climate conditions.Here, we present an overview of the hybrid RE hydrogen produc-tion system as well as the performance of the system whileoperating in different conditions particularly in fluctuating windspeed and solar radiation.

2. Description of the system and experimental setup

The system (Fig. 1) is located at Kuala Terengganu, East Coast ofPeninsular Malaysia (4� 13.5570N, 103� 26.0480E). The system isdesigned for an automatically controlled energy supplied to elec-trolyser based on state-of-charge (SOC) battery in the production ofhydrogen. The PEM electrolyser has been operated on constantcurrent flow with 40 A. The 1 kW photovoltaic modules consist of12 amorphous silicon US-64 modules electrically connected into 2series and 6 parallel to operate nominally at 24 VDC.

The 1 kW wind turbine energy system model from BergeyWindpower Co. has been used to couple with photovoltaicmodules. The wind turbine system is composed of a permanentmagnet WT that can deliver a maximum output power of 1000 W

Photovoltaic

Wind Turbine

Hydrogenpurification

Hydrogen storage tank

Battery Bank

Oxygen release to ambient

Deionized water

Tap water

PEM Electrolyser

Power Controller

Fig. 1. Schematic diagram of the standalone hybrid PV–wind hydrogen production system.

K. Sopian et al. / Renewable Energy 34 (2009) 1973–19781974

at 11 m/s of wind speed. The voltage produced by these sources isregulated and converted to 24 V on a power controlled center. Aset of deep-discharge batteries bank with a capacity of 1000 Ahconnected in a series/parallel configuration acts as a bufferbetween the PEM electrolyser and the power sources. 1 kW PEMelectrolyser custom made consists of 10 cells stack with the MEAarea approximately 100 cm2, which can deliver up to 0.5 m3/h ofpurified hydrogen at ambient pressure. In order to prevent thebatteries from overdischarge conditions the low voltage detectorhas been installed between the battery bank and the PEMelectrolyser.

0 500

1

2

3

4

Win

dsp

eed

(m

/s)

0 10 20 3025

30

35

Tem

peratu

re (°C

)

0 10 20 300

50

100

150

Measured data o

Po

wer (W

att)

wind speed

Ambient temp.

Photovoltaic power

Fig. 2. Wind speed, global radiation, power output from

3. System’s model

A complete model of the hybrid PV–wind hydrogen productionsystem was developed in order to design and validate an adequatesystem operation. The model was built using sub-models for eachindividual component.

3.1. Photovoltaic modules

The basic mathematical model was used to calculate themaximum power output from the photovoltaic modules. The

100 1500

200

400

600

800

1000

So

lar R

ad

iatio

n (W

/m

2)

40 50 6020

30

40

Tem

peratu

re(°C

)

40 50 60

n 15/08/06

0

50

100

150 Po

wer (W

att)

Global radiation

Photovoltaic temp.

Wind turbine power

wind turbine and photovoltaic on August 15th 2006.

0 10 20 30 40 50 6024

24.5

25

Vo

ltag

e (V

)

15

20

25 Cu

rren

t (am

p)

0 10 20 30 40 50 60400

450

500

550

600P

ow

er (W

att)

0.12

0.125

0.13

0.135

0.14F

lo

wrate (l/m

in

))

0 10 20 30 40 50 60

20

40

60

80

Efficien

cy (%

)

20

30

40

Electro

efficien

cy (%

)

0 10 20 30 40 50 60

6

7

8

Measured data on 15/08/06

Efficien

cy (%

)

6.5

7

7. 5

8 Efficien

cy (%

)

Hybrid System efficiency h2prod. efisiensi

Hydrogen flowrateElectrolyser Power

Electrolyser Current

Electrolyser Voltage

Voltage efficiency

Electrolyser efficiencyFaraday efficiency

Fig. 3. Electrolyser current, voltage, power, hydrogen flow rate and efficiency on August 15th 2006.

K. Sopian et al. / Renewable Energy 34 (2009) 1973–1978 1975

power output of the photovoltaic module is based on the maximumpoint for current and voltage as follows:

Imp ¼GT

GT;ref

hImp;ref þ mIsc

�TC � TC;ref

�i(1)

Vmp ¼ Vmp;ref þ mVoc

�TC � TC;ref

�(2)

Pmak ¼ ImpVmp (3)

where Imp and Vmp are the maximum current (ampere) and voltage

5 10 15

20

40

60

80

Efficien

cy (%

)

5 10 150

2

4

6

Efficien

cy (%

)

5 10 150

10

20

30

40

Time (2:00 - 5:55

Farad

ay efficien

cy(%

)

Hybrid system efficiency (exp

h2prod. efficiency (exp.)

Voltage efficiency

Faraday efficiency (exp.)

Electrolyser efficiency

Faraday efficiency hybrid syst

Faraday efficiency wind turbine (teory)

Fig. 4. Comparison of system efficien

(V) of the photovoltaic module respectively. Meanwhile, Pmak is thephotovoltaic module’s maximum power output (watt).

3.2. Wind turbine

The power output from the wind turbine based on Betz Law isgiven below.

Pout ¼12

rAV3ðhmechanicalCPÞ (4)

20 25 30 35

20 25 30 35

20 25 30 35 pm) on 15/08/06

.)

(exp.)

(exp.)

em (teory) Faraday efficiencyphotovoltaic (teory)

cy experimental and theoretical.

0 50 100 150

0 50 100 150

0 50 100 150

02468

10

Win

dsp

eed

(m

/s)

02004006008001000 P

ow

er (W

att)

25

30

35

Tem

peratu

re (°C

)

20

30

40

50

Tem

peratu

re (°C

)

0

50

1000

1500

Measured data on 18/11/06

Rad

iatio

n (W

/m

2)

0

200

400

600

Po

wer (W

att)

wind powerwind speed

Ambient temp. Photovoltaic temp.

Global RadiationPhotovoltaic Power

Fig. 5. Wind speed, global radiation, power output from wind turbine and photovoltaic on November 18th 2006.

K. Sopian et al. / Renewable Energy 34 (2009) 1973–19781976

where CP is the power coefficient, A is the blade swept area (m2), Vis the average wind speed (m/s) and r is the air density (kg/m3).

3.3. Batteries model

The batteries bank is the element linking together eachcomponent of the hybrid PV–wind hydrogen production system.Since it is connected in series with the power center and PEMelectrolyser and it acts as an energy buffer, the current flowing intoor from the batteries is defined as

IBatðtÞ ¼ IPVðtÞ þ IWTðtÞ � IElectrolyserðtÞ (5)

where IPV(t) is the photovoltaic module’s current (ampere), IWT(t) isthe wind turbine’s current (ampere), and IElectrolyser(t) is the current

23

23.5

24

24.5

Vo

ltag

e (V

)

80 85 90

100

200

300

Po

wer (W

att)

80 85 90

0

0.05

0.1

0.15

0.2

Flo

wrate (l/m

in

)

80 85 90

Electrolyser power

Electrolyser Voltage

h2 prod. from pv (theo

h2 prod. hibrid (teori)

h2 prod. hybrid (eks)

Faraday

Time (1:00 - 5:50pm

Fig. 6. Electrolyser current, voltage, power, hydrogen

supplied to electrolyser (ampere). This current is positive when thebatteries are charging and negative otherwise. Knowing thecurrent, it is possible to deduce the voltage by

VBatðtÞ ¼ ð1þ atÞVBat;0 þ RiIBatðtÞ þ KiQRðtÞ (6)

where a is the self-discharge rate (Hz), VBat,0 is the open circuitvoltage at time 0 (V), Ri is the internal resistance (U), Ki is thepolarization coefficient (U/h), QR(t) is the rate of accumulatedcharge.

The total energy stored in the batteries is given by

EBatðtÞ ¼ EBat;0 þ1

3600

ZIBatðtÞ dt (7)

95 100 1050

5

10

15 Cu

rren

t (am

p)

95 100 1050.13

0.135

0.14

Flo

wrate (l/m

in

)

95 100 1055

10

15

20

Efficien

cy (%

)

Hydrogen flow rate

Electrolyser Current

h2 prod. from wt (theory)

ry)

efisiensi (eks)

) on 18/11/06

flow rate and efficiency on November 18th 2006.

80 85 90 95 100 1050

20

40

60

80

Efficien

cy(%

)

80 85 90 95 100 105

0

5

10

Efficien

cy(%

)

80 85 90 95 100 1050

10

20

30

40

50

Efficien

cy(%

)

Electrolyser efficiency (exp)Faraday efficiency (exp.)

Voltage efficiency (exp.)

Hybrid System efficiency

h2 prod. efficiency

Faraday efficiency fromphotovoltaic (teory)

Faraday efficiency from windturbine (teory)

Faraday efficiency fromhybrid system (teory)

Time (1:00 - 5:50pm) on 18/11/06

Fig. 7. Comparison of system efficiency experimental and theoretical.

K. Sopian et al. / Renewable Energy 34 (2009) 1973–1978 1977

where EBat,0 is the batteries’ initial energy (Ah). Substituting (5) into(7), it is possible to express the batteries’ energy as follows:

EBat ¼ EBat;0 þIPV þ IWT � IElectrolyser

3600s(8)

Finally, this energy can be expressed as a state-of-charge usingthe following equation:

SOCBat ¼ 100� EBat

EBat;max(9)

where EBat,max is the total capacity of the batteries (Ah).

3.4. PEM electrolyser

This part of the system can be represented by a current sourcefrom the batteries bank feeding to the PEM electrolyser. The PEMelectrolyser is modeled as an ideal device to produce the hydrogenbased on the first Law of Faraday as given by

H2 prod ¼ hFncðRITtÞ

Fpz(10)

where hF is the Faraday efficiency, nc is the number of PEM elec-trolyser cell stacks, R is the gas constant (J/mol K), I is the currentsupplied to electrolyser (ampere), T is the temperature in kelvin(273þ �C), t is the period of time current supplied to electrolyser(second), F is the Faraday constant, 96 485 C/mol, p is the ambientpressure, 1�105 Pa (1 Pa¼ 1 J/m3), z is the excess number of elec-tron which is 2 for hydrogen and 4 for oxygen. Meanwhile theFaraday efficiency can be defined as

hF ¼1:482Vsel

iexperiment

itheory(11)

where 1.482 is the thermoneutral voltage (ampere), iexperiment

(ampere) is the actual current supplied to PEM electrolyser, Vsel isthe voltage which is applied to each cell stacks and itheory is thetheoretical current supplied to PEM electrolyser.

4. System operation and performance

The performance of the system was evaluated by in-situmeasurement techniques during a few days in monsoon seasonwhich occurred in the months of September–March and non-monsoon season of 2006 (April–August). Relevant data for one dayin non-monsoon season (August 15) are presented in Figs. 2–4.Fig. 2 shows the evolution of the global radiation, wind speed,ambient temperature, photovoltaic temperature, and power outputfrom wind turbine and photovoltaic as a function of time. Theweather conditions were sunny in the afternoon and slightly cloudyin the late afternoon. The average and maximum global radiationwas 130 W/m2 and 500 W/m2 respectively.

The measurements started at 2:00 p.m. The average powerproduced from the wind turbine was 5.739 W with the energyaverage 0.137 kWh-days. Meanwhile, the average power andenergy output from the solar modules was 47 W and 1.1 kWh-daysrespectively. Therefore, the total amount of energy charged tobattery was 1.2 kWh-days. The average of photovoltaic modules andambient temperature recorded was 25.5–34.8 �C and 29.3–31.5 �Crespectively.

The hydrogen production process started once the DC sourcefrom battery was connected to the anode and cathode plate of thePEM electrolyser. The initial values of parameter such as voltage,current and power of the electrolyser while starting the experimentcan be seen in Fig. 4. It was shown that at the beginning the voltageand current for electrolyser were 25 V and 23 A respectively.

It was observed that the voltage decreased continuously alongthe experimental carried out, except at time 2:30–3:15 p.m. wherethe voltage shown is almost constant due to high wind speed asa result of more power produced for charging the battery. Thecurrent from the battery supplied to electrolyser was foundconstant and slightly decreased as the voltage was dropped asa function of time and insolation as well as wind conditions untilthe late afternoon.

A similar evolution was observed on the power of electrolyser asthe power is a product of voltage and current. The flow rateof hydrogen shown was constant as the current supplied to

K. Sopian et al. / Renewable Energy 34 (2009) 1973–19781978

electrolyser is in stable condition. The Faraday efficiency of thesystem was 23–33% with the average current supplied was 14–21 Aand cumulative current was 300–480 Ah with hydrogen generatedwas 136–137 ml/min.

Electrolyser efficiency was 25–40% with average power within300–500 W. Voltage efficiency was 60–65% with the electrolyservoltage average was 22–24 V and the hydrogen production effi-ciency was shown constant at 7%. Overall, the system operated withthe amount of energy supplied to electrolyser was 8–11 kWh withthe system efficiency about 4–6%. In addition, the energy gener-ated from solar modules and wind turbine was 1–4 kWh and0.58–5.49 kWh respectively.

After completely running the experiment on the non-monsoonseason, some continuous test runs were performed to see theperformance of the system operated on the monsoon season.November 18, 2006 was chosen for analyzing the data due to verypromises of wind availability with high wind speed. The averagewind speed was about 1.58–4.72 m/s at 8:00 a.m.–12:50 p.m. in themorning. It was observed that the average wind speed increased to1.16–6.62 m/s at 1:00–5:50 p.m. in the late afternoon. During thismonsoon season, the high wind speed had occurred in the nighttime with the average of 1.42–5.29 m/s. These conditions hadcontributed a very good power generated with 245.17 W up to540 W maximum. The average of energy produced was 3.28 kWh-days with the maximum generated was 136.68 W.

Global solar radiation varies between 100 and 550 W/m2 in themorning and slightly decreased in the afternoon due to cloudyweather conditions. The average photovoltaic and ambienttemperature was 22–40 �C and 26–31 �C respectively. Theseconditions lead to less power generated from photovoltaicmodules. The daily behavior of system operation and performancefor a typical day in monsoon season are displayed in Figs. 5–7. Totalamount of energy flow to electrolyser was 0.23 kWh-days with theaverage of the current was 228.78 Ah.

In addition, the average of voltage and current was 23.76 V and9.53 A respectively. Power to electrolyser was 226.64 W withhydrogen flow rate was 134 ml/min. Faraday, electrolyser, and

voltage efficiency was 15.34%, 59.92%, and 62.37% respectively. Dueto the windiest weather and less solar radiation during thisexperimental period, the efficiency of hybrid system decreased to3.16%. Nevertheless, hydrogen generation efficiency shown slightlyincreased from 7.0% to 7.78%.

5. Conclusion

In general, both electrolyser and hybrid system efficiencies werevery encouraging when compared with those obtained in similarexperiences. From the experience with this study, it may beconcluded that there are no insurmountable technical problemsassociated with hydrogen production by hybrid PV–wind electro-lyser. Field observations show that hybrid systems are feasible andreliable enough, and required less maintenance.

Electrolyser technology appears to be mature enough for hybridsystem application. On the other hand, the electrochemical effectsof operation for periods with an intermittent power source can beovercome by using a battery bank as a buffer prior supplied toelectrolyser. Although the electrolyser was demonstrated to bereliable and its operation satisfactory, improvement on the stacksassemble needs to be done to avoid hydrogen leaking which leadsto less amount of hydrogen collected.

References

[1] Nayar CV. Stand alone wind/diesel/battery hybrid energy systems. Wind Eng1997;21(1):13–9.

[2] Wichert B, Nayar CV, Lawrence WB. Photovoltaic–diesel hybrid energy systemsfor off-grid rural electrification. Int J Renewable Energy Eng 1999;1(1):7–17.

[3] Vosen SR, Keller JO. Hybrid energy storage systems for stand-alone electricpower systems: optimization of system performance and cost through controlstrategies. Int J Hydrogen Energy 1999;24:1139–56.

[4] Menzl F, Wenske M, Lehmann J. In: Hydrogen energy progress XII, vol. 1; 1998.p. 757–65.

[5] Bargthels H, Brocke WA, Bonhoff K. PHOEBUS Julich: an autonomous energysupply comprising photovoltaic, electrolytic hydrogen, fuel cell. In: Hydrogenenergy progress XI, vol. 2; 1996. p. 1005–15.

[6] Ballard GEH. Engines of changes. In: Hydrogen energy progress XI, vol. 2; 1996.p. 1357–71.