optimization power systems pinch analysis

uan

iversal an

i Tek

S)urce

(1) the size of the most expensive RE generator, (2) the size of generator with the most abundant REsources available during the time interval with large electricity surplus and (3) the size of both the most

the d

2011).Different types of renewable energy generators have been

installed in HPS to produce electricity to be supplied to the loads.The high uctuations in time and output of many RE sourceshowever makes them harder to be utilised efciently in large

articular load may2009). In order toods such as thed articial intelli-

g the HPS includeHOGA (Bernal-

und et al., 2009),ptimisation Modely et al., 2013). A

graphical approach for optimal HPS was introduced by Borowy andSalameh (1996) who proposed a methodology to calculate theoptimum size of a battery bank and the photovoltaic (PV) array in ahybrid wind-PV system. The minimum cost of the systemwas usedto calculate the optimum conguration for a given load and adesired loss of power supply probability (LPSP). The optimumsizing is achieved by constructing the curve that represents therelationship between the number of PV modules and batteries.Kaldellis et al. (2009) developed an optimum sizing methodology

* Corresponding author. Tel.: 60 7 5536025; fax: 60 7 5588166.

Contents lists availab

Journal of Clean

.e ls

Journal of Cleaner Production xxx (2013) 1e10E-mail addresses: [email protected], [email protected] (S.R.WanAlwi).gas emissions andmitigating global warming are becoming sociallyand economically pressing for nations across the globe(Georgakellos, 2012). Effective measures to prevent climate changeinclude mitigating emissions from the power generation systems(Battaglini et al., 2009) and to accelerate the implementation ofrenewable energy (RE) sources as clean alternatives to fossil fuels inpower generation and hybrid power systems (HPS). In the long run,application of RE sources can prove to be a smart economic strategyas it can provide an effective safeguard to the changing climatewhile enhancing energy security and efciency (Purvins et al.,

larger HPS sizes, while supply uctuations for a poccur due to smaller HPS sizes (Hocaoglu et al.,obtain an optimum HPS, various sizing methsimulation, graphical, iterative, probabilistic angence techniques can be implemented.

Software tools that are available for designinHybrid Optimization by Genetic Algorithm eAgustn and Dufo-Lpez, 2009), energyPRO (LRETScreen (Redpath et al., 2011) and Hybrid Ofor Electric Renewables e HOMER (Goodbodsources, climate change and environmental emissions have becomethe key drivers to sustainable development. Reducing greenhouse

therefore vital to ensure a cost-effective utilisation of RE sources atthe desired conditions. Higher investment cost results from theKeywords:Power pinch analysis (PoPA)Hybrid power systems (HPS)Renewable energySizingOptimisationManagement

1. Introduction

The growing global concerns on0959-6526/$ e see front matter 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.jclepro.2013.12.028

Please cite this article in press as: MohammaCleaner Production (2013), http://dx.doi.orgexpensive and abundant RE sources available during the time interval with large electricity surplus. Theresults show that the rst option yields the minimum capital and operating costs and results in thelowest payback period for a given set of electricity targets.

2013 Elsevier Ltd. All rights reserved.

epletion of energy re-

power networks (Grbe et al., 2012). This can signicantly affect thesystems performance because electricity should be produced andsupplied at the timewhen it is needed. An optimal sizing method is27 November 2013Accepted 9 December 2013Available online xxx designers to choose the best alternative for their systems. The scenarios considered are the reduction ofReceived in revised form achieve the technical and economical feasibility of the HPS. Power Pinch Analysis (PoPA) method hasbeen applied to set the guidelines for proper HPS sizing. Different scenarios for RE generators allow theOptimal sizing of hybrid power systems

Nor Erniza Mohammad Rozali a, Sharifah Radah WJir Jaromr Klemes b, Mohammad Yusri Hassan c

a Process Systems Engineering Centre (PROSPECT), Faculty of Chemical Engineering, UnbCentre for Process Integration and Intensication e CPI2, Research Institute of ChemicTechnology, University of Pannonia, Egyetem u. 10, H-8200 Veszprm, HungarycCentre of Electrical Energy Systems (CEES), Faculty of Electrical Engineering, Universit

a r t i c l e i n f o

Article history:Received 3 September 2013

a b s t r a c t

Hybrid Power Systems (HPrenewable energy (RE) so

journal homepage: wwwAll rights reserved.

d Rozali, N.E., et al., Optimal s/10.1016/j.jclepro.2013.12.028sing power pinch analysis

Alwi a,*, Zainuddin Abdul Manan a,

iti Teknologi Malaysia, 81310 Johor Bahru, Malaysiad Process Engineering e M}UKI, Faculty of Information

nologi Malaysia, 81310 UTM Johor Bahru, Johor, Malaysia

consist of different renewable generators, which produce electricity froms required by the load. An optimal sizing method is the key factor to

le at ScienceDirect

er Production

evier .com/locate/ jc leproizing of hybrid power systems using power pinch analysis, Journal of

demands in ascending order, while Column 2 gives the duration

the AC and DC electricity are listed separately.

Table 1Limiting power sources for Illustrative Case Study.

Power source Time, h Timeinterval, h

Power sourcerating, kW

Electricitygeneration,kWh

AC DC From To

Wind 2 10 8 80 640Biomass 0 24 24 70 1680

Solar 8 18 10 60 600

Table 2Limiting power demands for Illustrative Case Study.

Power demand appliances Time, h Timeinterval, h

Powerdemandrating, kW

Electricityconsumption,kWh

AC DC From To

Appliance 1 0 24 24 30 720Appliance 2 8 18 10 50 500

Appliance 3 0 24 24 20 480Appliance 4 8 18 10 50 500

l of Cleaner Production xxx (2013) 1e10for stand-alone PV-battery systems according to the PV panelsnumber against battery maximum size curve. Application of themethod on case studies shows that the minimum energy paybackperiod is achieved whilst providing 100% energy autonomy forremote consumers. The concept of design space is applied byBandyopadhyay (2011) to establish the optimum sizing of genera-tors and storage for isolated power systems. Identication of thedesign space is done by constructing the sizing curve, which rep-resents the minimum storage capacity for a given generator rating.

Application of iterative optimisation technique for HPS optimi-sation has been carried out by Kaabeche et al. (2011). They rec-ommended an optimal sizing model based on an iterativetechnique to optimise the capacity of different components inhybrid PV/wind power generation systems using a battery bank. Atwo-stage iterative approach for distributed generation (DG) sizingwas given by Rotaru et al. (2012). The time-dependent evolution ofgeneration and load are taken into account in determining thepseudo-optimal DG sizing without violating any of the systemconstraints under any operating condition. Mohamed and Khatib(2013) recently proposed an optimisation method based on itera-tive simulation to optimally size a PV/wind/diesel generators withbattery storage. The optimal sizes obtained are closely matchedwith the results calculated by HOMER software.

The probabilistic approach was presented by Tina and Gagliano(2011). The authors evaluate the long-term performance of hybridsolar-wind power systems using the probability density function(PDF) based on the convolution technique. Ould Bilal et al. (2013)incorporated the uctuating nature of the resources as well asthe loads by using the probabilistic technique. The presentedmethod eliminates the need for time-series data in optimising thehybrid PV-wind-battery system in order to reach the bestcompromise between annual cost system (ACS) and the loss ofpower supply probability (LPSP).

Rajkumar et al. (2011) applied the articial intelligence methodnamely Neuro-Fuzzy in optimising the HPS. The PV and wind sys-tems are modelled with the Adaptive Neuro-Fuzzy System (ANFIS)and the results showed that low excess energy is achieved.Nasiraghdam and Jadid (2012) recommended a novel multi-objective articial bee colony algorithm to investigate the distri-bution system reconguration and the optimal sizing of a hybridenergy system. The key parameters considered as the optimisationobjectives include the total power loss, total electrical energy costand total emission. The optimal capacity of individual componentsin a stand-alone hybrid generation system is decided using Adap-tive Genetic Algorithm by Chen (2013). The proposed method ap-pears to be useful in locating the global optimum for largenonlinear systems.

In this paper, the Pinch Analysis technique is applied to opti-mally size an HPS e see e.g. Klemes and Varbanov (2013). PinchAnalysis has been, and is still widely applied for the optimal tar-geting and design for various resource networks. This is demon-strated by the recent publications e.g. heat (Torres et al., 2013),mass (Tay and Ng, 2012), water (Shenoy and Shenoy, 2013), carbon(Munir et al., 2012), property (Saw et al., 2011) and gas (Lou et al.,2013). Those papers declared that the Pinch Analysis has gainedgeneral acceptance by the public on its usefulness due to its simpleinsightful approaches that are either based on graphical or nu-merical techniques. The recent extension of Pinch Analysis for thedesign of power systems is employed in this paper. Power PinchAnalysis (PoPA) technique introduced by Wan Alwi et al. (2012)helps designers to determine the minimum targets for out-sourced electricity as well as the amount of excess electricity. The(PoPA) technique has been further extended by Mohammad Rozaliet al. (2013a) to include the losses analysis associated with power

N.E. Mohammad Rozali et al. / Journa2conversion, transfer and storage. The previous studies have been

Please cite this article in press as: Mohammad Rozali, N.E., et al., OptimalCleaner Production (2013), http://dx.doi.org/10.1016/j.jclepro.2013.12.028between two adjacent time-intervals.2) The total sum of ratings for power sources and power demands

for each time interval are given in columns 3 and 4. These valuescan be obtained from the Power Cascade Table e PCT(Mohammad Rozali et al., 2013b). The sources and demands forbroadened in the current paper to set the feasible limits for the sizeof RE generators in the HPS and to determine the battery capacity.Three scenarios have been considered to allow the user to decidethe choice of investment paths consisting of several combinationsof RE technologies.

2. Methodology

This section describes the step-wise procedure to obtain theoptimal sizing of an HPS. The Modied Storage Cascade Table (SCT)previously developed byMohammad Rozali et al. (2013a) is appliedfor electricity targeting and allocation in the system before furtherdetailed design is carried out to establish the optimal sizing ofgenerators and storage systems. In order to obtain the cost-effectiveHPSwith theminimum electricity targets, an Illustrative Case Studyis used to demonstrate the sizing method. The studied systemconsists of wind turbine, biomass generator and PV modules as thepower producer while the lead-acid battery functions as the powerstorage system. The sizing procedure is implemented as follows;

Step 1: Based on the meteorological data and the load demandsof a location, the limiting power data is extracted (MohammadRozali et al., 2013b). The maximum capacity for all RE generatorsis initially assumed without considering the demand proles. Thetotal electricity generation (source) is obtained by assuming that allthe RE sources available for the given sample day are converted toelectricity after the generators efciency is taken into account.Tables 1 and 2 tabulate the limiting power sources and demands forthe Illustrative Case Study. The maximum sizes for the RE genera-tors are 80 kW, 70 kW and 60 kW for wind, biomass and PV.

Step 2: The Modied SCT (Mohammad Rozali et al., 2013a) isused to obtain the electricity targets for the system. The step-wiseconstruction of the Modied SCT is done as follows (see alsoTable 3a);

1) Column 1 lists the time interval for power sources and powerAppliance 5 8 20 12 40 480

sizing of hybrid power systems using power pinch analysis, Journal of

l ofTime interval duration(1)

4) The sources are sent directly to the demands accordingly for theAC and DC. The surpluses and decits for the AC and DC elec-tricity between time intervals are calculated by using Equation(2), and listed in Column 7.

Electricity surplus=deficit X

Electricity source

X

Electricity demand (2)

Equation 2 should be applied separately for the AC and DCelectricity. A positive value indicates an electricity surplus while anegative value, electricity decit.

Table 3b is constructed as follows;

1) The electricity decit would be satised by converting the3) The quantities of the electricity sources and demands betweentime intervals are obtained using Equation (1), and listed inColumns 5 and 6.

XElectricity source=demand

XPower rating

Time, h Timeinterval, h

Power sourcerating, kW

Power demandrating, kW

AC DC AC DC

02 70 0 0 50

26 150 0 0 50

82 150 60 140 50

108 70 60 140 50

182 70 0 40 50

204 70 0 0 50

24Table 3aModied Storage Cascade Table for Illustrative Case Study.

1 2 3 4P P

N.E. Mohammad Rozali et al. / Journaelectricity surplus. Column 8 gives the converted AC and DCelectricity surplus, which is obtained from Equation (3). For theDC electricity surplus, similar equation as Equation (3) (replacethe AC to DC electricity surplus and the rectier to inverter ef-ciency) can be used if the amount of surplus is less than the ACdecit. If the amount of DC surplus is higher than the AC decit,only the exact amount of the required AC load is converted fromthe available DC surplus. Equation (4) is derived to calculate theamount of DC surplus to be converted to AC if this case occurs.The AC electricity surplus is completely converted to DC becauseany excess DC can be directly sent to the storage system.

Amount of converted AC electricity to DC

AC electricity surplus Rectifier efficiency (3)

Amount of DC electricity surplus to be converted to AC

Amount of AC deficitInverter efficiency

(4)

Please cite this article in press as: Mohammad Rozali, N.E., et al., Optimal sCleaner Production (2013), http://dx.doi.org/10.1016/j.jclepro.2013.12.028be calculated using Equation (6), as follows:

DC Electricity to be discharged

Converted DC surplus AC deficitRectifier efficiency

(6)2) Column 9 shows the DC electricity available for storage afterload utilisation, which is obtained using Equation (5).

Charging=Discharging quantity DCs=d ACconverted DCconverted (5)

whereDCs/d DC electricity surplus/decit; ACconverted amount of DC

converted from AC electricity surplus; DCconverted amount of DCelectricity surplus that will be converted to AC to satisfy the AC loaddemand.

The positive value represents the charging quantity while thenegative value indicates the discharging quantity for the DC decit.

3) The quantity of the electricity discharged from the battery tosatisfy the AC decit is calculated using Equation (6), and listedin Column 10. Taking into account of the positive/negative signs,the amount of DC electricity to be discharged from battery can

5 6 7P

Electricitysource, kWh

PElectricity

demand, kWhElectricity surplus/decit,kWh

AC DC AC DC AC DC

140 0 0 100 140 100

900 0 0 300 900 300

300 120 280 100 20 20

560 480 1120 400 560 80

140 0 80 100 60 100

280 0 0 200 280 200

Cleaner Production xxx (2013) 1e10 3If the battery capacity is less than the DC discharge requirementto satisfy the AC decit, the battery will be discharged to its depthof discharge (DoD). The DoD of the lead-acid battery used is typi-cally about 80% of its maximum capacity (Komor and Glassmire,2012). Equation (7) is applied in this scenario to calculate theamount of DC electricity available from battery to satisfy the ACdemand decit.

DC electricity available from battery

Bt11 s T hI hd (7)

whereBt1 battery capacity at previous time interval [kWh];

s hourly self-discharge rate [0.00004/h]; t time [h]; T timeinterval [h]; hI inverter efciency (0.95); hd discharging ef-ciency (0.9).

4) Based on the values in Columns 9 and 10, the cumulative storagecapacity (Column 11) is calculated by including the battery

izing of hybrid power systems using power pinch analysis, Journal of

12 13 14

Operation

hOutsourced electricity, kWh Battery

capacity,Outsourced electricity, kWh

AC DC AC DC

l of Cleaner Production xxx (2013) 1e10electricity demand, external electricity may be purchased fromthe grid. Column 12 lists the net electricity decit, which in-dicates the outsourced AC and DC electricity requirements. Thegrid supplies the AC electricity. The DC electricity calculated isdivided with the rectier efciency to give the required amountefciency (self-discharge rate, charging and discharging) viaEquation (8).

Bt Bt11 s T Ct hc Dt=hd (8)

whereBt battery capacity [kWh]; Ct charging quantity [kWh];

Dt discharging quantity [kWh]; s hourly self-discharge rate[0.00004/h]; t time [h]; T time interval [h]; hc charging ef-ciency (0.9); hd discharging efciency (0.9).

If the battery has been discharged to its DoD (e.g. between 10and 18 h), the electricity cascade for the next time interval resumesat zero.

5) When the amount of storage is still insufcient to satisfy the

57.00 0 0 0 0

266.00 0 66.00 0 59.40Table 3bModied Storage Cascade Table for Illustrative Case Study.

8 9 10 11

Converted surplus,kWh

Charging/dischargingquantity (DC), kWh

Discharge for ACdecit, kWh

Start up

Batterycapacity, kW

ACeDC DCeAC

0

133.00 0 33.00 0 29.70

855.00 0 555.00 0 529.19

19.00 0 39.00 0 564.25

0 76.00 0 482.41 0

N.E. Mohammad Rozali et al. / Journa4of the outsourced AC electricity to satisfy the DC demand.Equation (9) is used to obtain the kW instantaneous externalpower demand.

Outsourced power rating Outsourced electricityTime interval

(9)

Procedures 1 to 6 are carried out to obtain the electricity targetsfor startup operation (Day 1). During startup at the time t 0 h, nopower is available in the battery. After power is generated andutilised during the startup period, the excess power stored in thebattery at t 24 h is brought to the next day to reduce the requiredoutsourced electricity.

6) For the next day (normal 24 h) operation, the battery capacity att 0 is taken from the electricity stored during startup, at t 24(59.40 kWh). Equation (8) is used to calculate the battery ca-pacity. The battery capacities are listed in Column 13.

The key electricity targets that need to be extracted from thetable are; (i) The minimum outsourced electricity supply (MOES),(ii) Available excess electricity for the next day (AEEND) and (iii)The storage capacity.

Please cite this article in press as: Mohammad Rozali, N.E., et al., OptimalCleaner Production (2013), http://dx.doi.org/10.1016/j.jclepro.2013.12.028Results from the Modied SCT show that 59.40 kWh of elec-tricity (AEEND) is in excess at the end of the rst day (Column 11,t 24 h). This value is higher than the total amount of MOESrequired for startup, which are 1.59 kWh and 43.00 kWh for AC andDC demands (Column 12). The difference between the AEEND andMOES (59.40 e (1.59/0.95) 43.00 14.73 kWh) indicates theamount of electricity, which would be wasted without being storedor sent to the load. This amount of electricity is not transferred tothe next day because continuously cascading the excess AEEND tothe next day would accumulate the energy inside the storage sys-tem (Wan Alwi et al., 2012) e see Column 13. This scenario offersthe opportunity to reduce the initially installed capacity of REgenerators.

Step 3: Reduce the size of the selected RE generator. Differentscenarios can be considered in selecting the best RE generator to beresized. The established targets in Step 2 can guide the users todecide on the best scenario in order to achieve the minimum costfor the designed system. The following scenarios that have beenanalysed and the results compared among one other:

(i) Reduce the size of the most expensive RE generator

kWh

59.400 0

89.100 0

0 0588.57

0 0

0 0623.63

0 0

1.59 057.52

0 0

0 43.009.74

0 0

0 069.14

0 0(ii) Reduce the size of generator with the most abundant REsources available during the time interval with large elec-tricity surplus

(iii) Reduce the size of both the most expensive and abundant REsources available during the time interval with large elec-tricity surplus.

The excess electricity targets are plotted against their corre-sponding RE generator sizes for each scenario. The size of thegenerator without electricity wastage has been selected in order tooptimise the utilisation of RE sources.

(i) Scenario 1: Reduce the size of the most expensive REgenerator

Table 4Power plant capital and operating costs (U. S. Energy Information Administration,2010).

Capital cost, $/kW Fixed operating and maintenancecost, $/kW.yrs

Wind 2438.00 28.07Biomass 3860.00 100.50Solar 6050.00 26.04


-10

-5

0

5

10

15

20

55 60 65 70 75 80

Exce

ss e

lect

rici

ty, kW

h

Generator size, kW

SolarWind

8

10

12

14

0.3

0.4

0.5

0.6

0.7

0.8

sp

eed,

m

/s

tion,

kW

/m2

Solar

N.E. Mohammad Rozali et al. / Journal of Cleaner Production xxx (2013) 1e10 5-15

Fig. 1. Excess electricity versus generator size for Illustrative Case Study (Scenarios 1and 2).

10

15

20

hTable 4 shows the cost for the PV, wind and biomass technolo-gies considered in this study. Among the REs, PV facility is selectedto be resized because of its high costs compared to the other REs.Fig.1 illustrates the plot of excess electricity versus various PV sizes,below its maximum size (60 kW). From the graph, the optimumsize for PV array which gives zero excess electricity is 58.47 kW. Theoptimal conguration for the new system is therefore 80 kW for thewind turbine, 70 kW for the biomass generator and 58.47 kW for PVpanels. Inserting these RE generators size into theModied SCT, thecapacity of 701.87 kWh is obtained for the lead-acid battery afterconsideration of 80% depth of dischargee DoD (Notton et al., 2011).

-15

-10

-5

0

5

55 60 65 70 75 80 85

Exce

ss e

lect

rici

ty, k

W

Generator size, kW

SolarWind

Fig. 2. Excess electricity versus generator size for Illustrative Case Study (Scenario 3).

Table 5Economic evaluation of each Scenario for Illustrative Case Study.

Scenario 1 Scenario 2 Scenario 3

Generator size (S), kW PV e 58.47 PV e 60 PV e 59.07Wind e 80 Wind e 77.61 Wind e 79.07Biomass e 70 Biomass e 70 Biomass e 70

Daily electricitygeneration(EG), kWh

PV e 584.68 PV e 600 PV e 590.68Wind e 640 Wind e 620.85 Wind e 632.54Biomass e 1680 Biomass e 1680 Biomass e 1680

Total annual operatingand maintenancecost, $/kW

10,803 10,776 10,793

Net capital investment, $ 818,974 822,405 820,329Net annual savings, $/y 116,422 116,282 116,368Payback period, y 7.03 7.07 7.05

Please cite this article in press as: Mohammad Rozali, N.E., et al., Optimal sCleaner Production (2013), http://dx.doi.org/10.1016/j.jclepro.2013.12.028(ii) Scenario 2: Reduce the size of generator with the mostabundant RE sources available during time interval withlarge electricity surplus

Based on Tables 3a and b, it can be observed that a large ACelectricity surplus exists between 2 and 10 h (Column 7). Thisobservation is strongly supported by the huge amount of storage atthe same time interval (Columns 11 and 13). The main AC sourceduring that period is identied, which is the wind energy. The sizeof wind turbine is therefore reduced from its maximum capacity,80 kW. Fig. 1 shows the variation of excess electricity with thechange in wind turbine size. The required size for the wind turbineof 77.61 kW is expected to efciently utilise all the RE sourceswithout any excess electricity. The lead-acid battery capacity in thiscase is 684.85 kWh.

(iii) Scenario 3: Reduce the size of both the most expensive andabundant RE sources available during the time interval withlarge electricity surplus

Both PV arrays and wind turbine capacities are reduced equally(e.g. 1 kW reduction in PV capacity and 1 kW in wind). Repeatingthe same procedure as for the previous scenarios, different sizes forPV arrays and wind turbine are combined, below their maximumcapacities. The resulting excess electricity for each size combinationis plotted and shown in Fig. 2. The PV array and the wind turbine

0

2

4

6

-0.1

0

0.1

0.2

-1 4 9 14 19 24

Win

d

Inso

la

Time, h

Wind

Fig. 3. Average hourly solar insolation and wind prole (Feroldi et al., 2013).capacity when there is no excess in electricity occur at 59.07 kWand 79.07 kW. For this condition, the system requires 70 kW ofbiomass generator and 695.25 kWh storage capacity.

Step 4: Calculate the payback period to decide the best invest-ment path between the three scenarios. Equation (10) gives thepayback period for the design.

Payback period Net capital investmentNet annual savings

(10)

Table 6Parameters of RE technologies.

2.50 MW PV panelsTotal area, m2 20,000Efciency 0.1644.00 MW wind turbineSwept area, m2 3904Air density, kg/m3 1.225Efciency 0.95


0.12 $/kWh/0.377 RM/kWh (Tenaga Nasional Berhad, 2013), S is theRE capacity or size and OM is the annualised operating and main-tenance cost obtained from Table 4.

The industrial pricing and tariff currently implemented inMalaysia is used in the calculations. Even though the tariff rate isdifferent for different consumption ranges and types of industries, asingle rate (low voltage industrial tariff for overall monthly con-sumption of more than 200 kWh/month) is applied in this study.Table 5 lists the payback period for the projects selected from eachscenario.

As can be seen in Table 5, Scenario 1 gives the shortest paybackperiod (7.03 y) compared to the other two solutions. The totalannual expenditure on this alternative is higher than the otheroptions because the OM for the other two generators are higherthan that of solar. However, this RE technology results in thehighest annual savings and the capital cost is also much lower. Theconguration from Scenario 1 is therefore chosen as the mostoptimal design for this demand trend.

3. Case Study

The electricity system of a chemical plant is taken as the CaseStudy. Fig. 3 presents the hourly average solar insolation and windspeed for a typical day at the plant site (Feroldi et al., 2013). Thestorage scheme applied in the system is the lead-acid battery with90% charging/discharging efciency (Zhou et al., 2008) and self-

Table 7Limiting power sources for Case Study.

Time, h Power source rating, kW Power demand rating, kW

From To AC DC AC DC

0 1 1575.00 0 903.96 311.011 2 1500.00 0 803.52 276.482 3 1462.50 0 669.60 230.403 4 1375.00 0 502.20 172.804 5 1300.00 0 435.24 149.765 6 1275.00 0 368.28 126.726 7 1250.00 37.50 468.29 116.717 8 1175.00 400.00 540.34 134.668 9 1050.00 825.00 1765.10 439.909 10 950.00 1200.00 2017.26 502.7410 11 925.00 1525.00 2017.26 502.7411 12 875.00 1725.00 1584.99 395.0112 13 975.00 1750.00 1260.79 314.2113 14 1050.00 1725.00 1584.99 395.0114 15 1225.00 1525.00 1945.22 484.7915 16 1350.00 1200.00 1945.22 484.7916 17 1525.00 825.00 1621.01 403.9917 18 1675.00 400.00 1801.13 448.8818 19 1700.00 37.50 1693.06 421.9419 20 1737.50 0 1071.36 368.6420 21 1775.00 0 970.92 334.0821 22 1750.00 0 1640.52 564.4822 23 1725.00 0 1540.08 529.9223 24 1700.00 0 1540.08 529.92

N.E. Mohammad Rozali et al. / Journal of Cleaner Production xxx (2013) 1e106The net annual savings is calculated using Equation (11), whichis the difference between the cost of electricity production and thecommercial electricity tariff.

Net annual savings X

i

EG D TE S OM (11)

wherei is the number of RE technologies, EG is the total daily electricity

generation for each RE source, D are days for an annualoperation (assumed as 365 d), TE is the tariff rate for electricity e

Table 8

Limiting power demands for Case Study.

Time,h

Power demand rating, kW

AC

Motors Pumps Workshopmachines

Overheadcranes

Ventilation Furnace Boilers Dustcollectingequipme

0 571.66 167.06 72.90 36.45 23.81 15.31 10.69 3.651 508.14 148.5 64.80 32.40 21.17 13.61 9.50 3.242 423.45 123.75 54.00 27.00 17.64 11.34 7.92 2.703 317.59 92.81 40.50 20.25 13.23 8.51 5.94 2.034 275.25 80.44 35.10 17.55 11.47 7.37 5.15 1.765 232.90 68.06 29.70 14.85 9.70 6.24 4.36 1.496 275.24 80.44 35.10 17.55 11.47 7.37 5.15 1.767 317.59 92.81 40.50 20.25 13.23 8.51 5.94 2.038 1037.45 303.19 132.30 66.15 43.22 27.78 19.40 6.629 1185.66 346.5 151.20 75.60 49.39 31.75 22.18 7.5610 1185.66 346.5 151.20 75.60 49.39 31.75 22.18 7.5611 931.59 272.25 118.80 59.40 38.81 24.95 17.42 5.9412 741.04 216.56 94.50 47.25 30.87 19.85 13.86 4.7313 931.59 272.25 118.80 59.40 38.81 24.95 17.42 5.9414 1143.32 334.13 145.80 72.90 47.63 30.62 21.38 7.2915 1143.32 334.13 145.80 72.90 47.63 30.62 21.38 7.2916 952.76 278.44 121.50 60.75 39.69 25.52 17.82 6.0817 1058.63 309.38 135.00 67.50 44.10 28.35 19.80 6.7518 995.11 290.81 126.90 63.45 41.45 26.65 18.61 6.3519 677.52 198.00 86.40 43.20 28.22 18.14 12.67 4.3220 614.00 179.44 78.30 39.15 25.58 16.44 11.48 3.9221 1037.45 303.19 132.30 66.15 43.22 27.78 19.40 6.6222 973.94 284.63 124.20 62.10 40.57 26.08 18.22 6.2123 973.94 284.63 124.20 62.10 40.57 26.08 18.22 6.21

Please cite this article in press as: Mohammad Rozali, N.E., et al., OptimalCleaner Production (2013), http://dx.doi.org/10.1016/j.jclepro.2013.12.028discharge rate of 0.004%/h (Pickard et al., 2009).Step 1: Referring to Fig. 3, the hourly power output from the solar

and wind energy are calculated using Equations (12) and (13)(Nelson et al., 2006). The specication data for PV facility and windturbine used for installation at the plant site are tabulated in Table 6.

PPVt ItAhPV (12)

where I(t) is the insolation data at time t (kW/m2), A is the area ofPV panels (m2), and hPV is the overall efciency of the PV panels andthe DC/DC converter.

DC

nt

Lift Compressors Air-conditioners

Lights Conveyors Refrigerationsystems

Others

2.43 107.53 87.60 68.65 12.15 5.95 29.162.16 95.58 77.87 61.02 10.80 5.29 25.921.80 79.65 64.89 50.85 9.00 4.41 21.601.35 59.74 48.67 38.14 6.75 3.31 16.201.17 51.77 42.18 33.05 5.85 2.87 14.040.99 43.81 35.69 27.97 4.95 2.43 11.881.17 51.77 42.18 33.05 5.85 2.87 14.041.35 59.74 48.67 38.14 6.75 3.31 16.204.41 195.14 158.98 124.58 22.05 10.80 52.925.04 223.02 181.70 142.38 25.20 12.35 60.485.04 223.02 181.70 142.38 25.20 12.35 60.483.96 175.23 142.76 111.87 19.80 9.70 47.523.15 139.39 113.56 88.99 15.75 7.72 37.803.96 175.23 142.76 111.87 19.80 9.70 47.524.86 215.06 175.20 137.30 24.30 11.91 58.324.86 215.06 175.20 137.30 24.30 11.91 58.324.05 179.21 146.00 11,441 20.25 9.92 48.604.50 199.13 162.23 127.13 22.50 11.03 54.004.23 187.18 152.49 119.50 21.15 10.36 50.762.88 127.44 103.82 81.36 14.40 7.06 34.562.61 115.49 94.09 73.73 13.05 6.39 31.324.41 195.14 158.98 124.58 22.05 10.80 52.924.14 183.20 149.25 116.96 20.70 10.14 49.68

4.14 183.20 149.25 116.96 20.70 10.14 49.68


Table 9aModied Storage Cascade Table for Case Study between time 0 and 11 h.

1 2 3 4 5 6 7


PPower source rating,

kW

PPower demand rating,

kW

PElectricity source,

kWh

PElectricity demand,

kWhElectricity surplus/decit,kWh

AC DC AC DC AC DC AC DC AC DC

01 251.23 0 903.96 311.04 251.23 0 903.96 311.04 652.74 311.04

11 283.96 0 803.52 276.48 283.96 0 803.52 276.48 519.57 276.48

21 283.96 0 669.60 230.40 283.96 0 669.60 230.40 385.65 230.40

31 168.30 0 502.20 172.80 168.30 0 502.20 172.80 333.90 172.80

41 338.20 0 435.24 149.76 338.20 0 435.24 149.76 97.05 149.76

51 338.20 0 368.28 126.72 338.20 0 368.28 126.72 30.08 126.72

61 623.85 0 468.29 116.71 623.85 0 468.29 116.71 155.56 116.71

71 1037.08 0 540.34 134.66 1037.08 0 540.34 134.66 496.74 134.66

81 1037.08 328.00 1765.10 439.90 1037.08 328.00 1765.10 439.90 728.02 111.90

91 920.52 951.20 2017.26 502.74 920.52 951.20 2017.26 502.74 1096.74 448.46

101 1207.24 1049.60 2017.26 502.74 1207.24 1049.60 2017.26 502.74 810.02 546.86

111 1298.89 1935.20 1584.99 395.01 1298.89 1935.20 1584.99 395.01 2,86.10 1540.19

N.E. Mohammad Rozali et al. / Journal of Cleaner Production xxx (2013) 1e10 7PWindt 1=2rAvt3Cp (13)

where r is the air density (kg/m3), A is the swept area of the rotor(m2), v(t) is the wind speed at time t (m/s), Cp is the efciency of thewind turbine.

The extracted limiting power data for the Case Study is given in

Tables 7and 8.

Table 9bModied Storage Cascade Table for Case Study between time 12 and 24 h.

1 2 3 4 5


PPower source rating,

kW

PPower demand rating,

kW

P

k

AC DC AC DC A

121 3454.88 2263.20 1260.79 314.21 3

131 2629.71 2460.00 1584.99 395.01 2

141 3925.39 2197.60 1945.22 484.79 3

151 3454.88 1213.60 1945.22 484.79 3

161 2629.71 1115.20 1621.01 403.99 2

171 1298.89 1148.00 1801.13 448.88 1

181 1495.88 393.60 1693.06 421.94 1

191 267.26 164.00 1071.36 368.64 2

201 515.62 0.00 970.92 334.08 5

211 595.50 0.00 1640.52 564.48 5

221 1711.84 0.00 1540.08 529.92 1

231 2073.26 0.00 1540.08 529.92 2

24

Please cite this article in press as: Mohammad Rozali, N.E., et al., Optimal sCleaner Production (2013), http://dx.doi.org/10.1016/j.jclepro.2013.12.028Step 2: The Modied SCT for the Case Study is represented byTables 9aed. It can be observed that the amount of AEEND atstartup (9712.53 kWh) is more than the quantity required for theoutsourced electricity (4984.14 kWh). This clearly shows that theinstalled capacities of PV and wind generators are too big and noteconomical. The excess 4728.38 kWh should be eliminated byaltering the current systems conguration in order to decrease thecapital cost of the design.6 7

Electricity source,Wh

PElectricity demand,

kWhElectricity surplus/decit,kWh

C DC AC DC AC DC

454.88 2263.20 1260.79 314.21 2194.09 1948.99

629.71 2460.00 1584.99 395.01 1044.72 2064.99

925.39 2197.60 1945.22 484.79 1980.18 1712.82

454.88 1213.60 1945.22 484.79 1509.67 728.82

629.71 1115.20 1621.01 403.99 1008.70 711.21

298.89 1148.00 1801.13 448.88 502.24 699.13

495.88 393.60 1693.06 421.94 197.18 28.34

67.26 164.00 1071.36 368.64 804.10 204.64

15.62 0.00 9,70.92 334.08 455.30 334.08

95.50 0.00 1640.52 564.48 1045.02 564.48

711.84 0.00 1540.08 529.92 171.76 529.92

073.26 0.00 1540.08 529.92 533.18 529.92


Table 9cModied Storage Cascade Table for Case Study between time 0 and 11 h.

1 8 9 10 11 12 13 14

Time, h Converted surplus,kWh

Charging/Dischargingquantity (DC), kWh


Start up Operation

Batterycapacity,kWh

Outsourced electricity,kWh

Batterycapacity,

kWh


ACeDC DCeAC AC DC AC DC

00

9858.80

1 0 0 0 00

652.74 311.04 8749.37 0 0

2 0 0 0 00

519.57 276.48 7834.15 0 0

3 0 0 0 00

385.65 230.40 7126.79 0 0

4 0 0 0 00

333.90 172.80 6543.98 0 0

5 0 0 0 00

97.05 149.76 6263.81 0 0

6 0 0 0 00

30.09 126.72 6087.57 0 0

7 147.78 0 31.07 027.96

0 0 6115.29 0 0

8 471.91 0 337.24 0331.48

0 0 6115.05 0 0

9 0 0 111.90 177.100

550.92 0 5138.99 0 0

10 0 426.04 0 00

670.70 0 5923.22 0 0

11 0 519.52 0 00

290.50 0 6262.76 0 0

N.E. Mohammad Rozali et al. / Journal of Cleaner Production xxx (2013) 1e108Step 3: It is assumed that the PV and wind turbine in thisCase Study is purchased from the same supplier as in theprevious Illustrative Case Study. The type and cost of the tech-nologies are therefore identical and applied in the latercalculations.Table 9dModied Storage Cascade Table for Case Study between time 12 and 24 h.

1 8 9 10 11

Time, h Converted surplus,kWh

Charging/Dischargingquantity (DC), kWh


Start u

BatterykWh

ACeDC DCeAC

120 301.16 1239.03 0 1115.1

132084.39 0 4033.38 0 4745.1

14992.48 0 3057.47 0 7496.6

151881.17 0 3593.98 0 10,730

161434.18 0 2163.00 0 12,677

17958.26 0 1669.48 0 14,179

180 528.67 170.46 0 14,332

190 0 28.34 207.56 14,069

200 0 204.64 846.53 12,900

210 0 334.08 479.26 11,996

220 0 564.48 1100.02 10,146

23163.17 0 366.75 0 9738.9

24506.52 0 23.40 0 9712.5

Please cite this article in press as: Mohammad Rozali, N.E., et al., OptimalCleaner Production (2013), http://dx.doi.org/10.1016/j.jclepro.2013.12.028(i) Scenario 1: Reduce the size of the most expensive REgenerator

As the most expensive RE technology, PV system is given thepriority to be resized. Various PV sizes below 2.5 MW with their12 13 14

p Operation

capacity, Outsourced electricity,kWh

Batterycapacity,kWh


AC DC AC DC

3 0 0 7377.64 0 0

2 0 0 11,007.38 0 0

6 0 0 13,758.66 0 0

.94 0 0 16,992.70 0 0

.21 0 0 18,938.71 0 0

.23 0 0 20,440.48 0 0

.07 0 0 20,593.08 0 0

.39 0 0 20,330.15 0 0

.98 0 0 19,161.48 0 0

.75 0 0 18,257.00 0 0

.82 0 0 16,406.82 0 0

1 0 0 15,998.67 0 0

3 0 0 15,972.03 0 0


interval, the charging quantity is also the highest because of there isno decit. Further resizing is made to the wind turbine as the onlyAC source for the system. Fig. 4 provides the excess electricity thatis obtained as a result of varying the wind turbine sizes below itsmaximum capacity of 4.00 MW. In order to utilise the RE sourceswithout any electricity excess, a wind turbine of 3.37 MW capacityhas been chosen for installation with 2.5 MW PV arrays.

(iii) Scenario 3: Reduce the size of both the most expensive andabundant RE sources available during time interval with0

1000

2000

3000

4000

5000

6000

1.5 2 2.5 3 3.5 4cess

ele

ctri

city

, MW

h

SolarWind

N.E. Mohammad Rozali et al. / Journal of Cleaner Production xxx (2013) 1e10 9-4000

-3000

-2000

-1000Ex Generator size, MW

Fig. 4. Excess electricity versus generator size for Case Study (Scenarios 1 and 2).

1

2

3

4

5

6

lect

rici

ty, M

Whequivalent excess electricity are given in Fig. 4. The plot shows thatalmost 1 MW unit can be reduced from the initial maximum PVcapacity in order to avoid any electricity excess. The optimal PV sizewith efcient utilisation of all the RE sources without excessoccurrence is found to be 1.67 MW. Applying the Modied SCTmethod on an HPS comprising PV panels with 1.67 MW poweroutput and wind turbine with 4.00 MW power output shows that13.53 MW of storage capacity is required for the design.

(ii) Scenario 2: Reduce the size of generator with the mostabundant RE sources available during time interval withlarge electricity surplus

The electricity surpluses at all time intervals are screened (seeColumn 8, Tables 9c, d). It can be seen that the highest surplusoccurs between 12 and 13 h, from the AC source. During this time

period of 11.78 y as compared to 13.87 y and 12.99 y for Scenarios 2

-3

-2

-1

02 2.5 3 3.5 4E

xce

ss e

Generator size, MW

Fig. 5. Excess electricity versus generator size for Case Study (Scenario 3).

Table 10Economic evaluation of each Scenario for Case Study.

Scenario 1

Generator size (S), MW PV e 1.67Wind e 4.00

Daily electricity generation (EG), kWh PV e 10,142.0Wind e 31,84

Total annual operating and maintenance cost, $/kW 155,663Net capital investment, $ 19,831,300Net annual savings, $/y 1,683,209Payback period, y 11.78

Please cite this article in press as: Mohammad Rozali, N.E., et al., Optimal sCleaner Production (2013), http://dx.doi.org/10.1016/j.jclepro.2013.12.028and 3.Note that Scenario 3 gives the intermediate results between

Scenarios 1 and 2 for both the Illustrative Example and the CaseStudy. In order to save time, the designer is recommended to omitScenario 3 during analysis.

4. Conclusion

An optimal sizing method for a hybrid power system has beenpresented. In order to target the components conguration, theinsight-based Pinch Analysis technique has been utilised as analternative to complex mathematical model. The PoPA method isapplied to determine the amount of excess electricity that has beenapplied as the indicator to decide on the optimal utilisation of REsources. For more ways on how to assess the environmental im-pacts of HPS, readers are referred to e.g. Cucek et al. (2012). Threepossible scenarios have been considered as the solution to obtainthe optimal sizing. Two case studies presented show that the mostcost-effective solution is by reducing the size of the most expensiveRE technology. In some countries the situation is further compli-cated by peak and off-peak grid electricity prices. Nevertheless, thedescribed methodology can be adapted to those cases as well. Thedifferent tariff rates for the peak and off-peak periods e see e.g.

Scenario 2 Scenario 3

PV e 2.5 PV e 2.17Wind e 3.37 Wind e 3.67

8 PV e 15,219.20 PV e 13,195.661.32 Wind e 26,854.95 Wind e 29,195.29

159,696 159,39423,341,060 22,055,5891,683,152 1,697,329large electricity surplus

The capacities of PV and wind generators are evenly reduceduntil the same AEEND and MOES are achieved. The amount ofexcess electricity with each size reduction is shown in Fig. 5. It isfound that about 0.33 MW capacity reductions in both RE genera-tors is required to ensure an optimal use of the REs while main-taining the systems reliability. The optimal conguration obtainedis 2.17 MW for PV panels, 3.67 MW for the wind turbine and14.49 MW for the lead-acid battery.

Step 4: Table 10 summarises the results of the economicassessment on the three designs under the different scenarios. Theresults are parallel with the ndings in the Illustrative Case Study.The rst Scenario appears to be the most promising solution thatcan lead to the optimal design of the HPS. As the capital cost of PVfacility is about two-fold higher than the capital cost for the othertwo technologies, reducing its capacity is seen to be the mostworthwhile strategy. This solution yields the minimum payback13.87 12.99


(Tenaga Nasional Berhad, 2013) will be considered in our futurework.

Kaldellis, J., Simotas, M., Zarakis, D., Kondili, E., 2009. Optimum autonomousphotovoltaic solution for the Greek islands on the basis of energy pay-backanalysis. J. Clean. Prod. 17, 1311e1323.

Klemes, J., Varbanov, P., 2013. Process intensication and integration: an assess-ment. Clean Technol. Environ. Pol. 15, 417e422.

N.E. Mohammad Rozali et al. / Journal of Cleaner Production xxx (2013) 1e1010Acknowledgements

The authors would like to thank the Universiti TeknologiMalaysia (UTM) and Ministry of Higher Education (MOHE) ofMalaysia for providing the nancial support through the ResearchUniversity Grant under the Vote No. Q.J130000.2544.03H44. Theauthors also acknowledge the nancial support from the HungarianState and the European Union under project TAMOP-4.2.2.A-11/1/KONV-2012-0072 d Design and optimisation of the modernisa-tion and efcient operation of energy-supply and the utilisationsystems using renewable energy sources and ICTs. These grantsprovided the opportunity to complete this research.

Abbreviations

AC alternating currentACS annual cost systemAEEND available excess electricity for next dayANFIS Adaptive Neuro-Fuzzy SystemDC direct currentDG distributed generationHOGA Hybrid Optimization by Genetic AlgorithmHOMER Hybrid Optimisation Model for Electric RenewablesHPS hybrid power systemsLPSP loss of power supply probabilityMOES minimum outsourced electricity supplyPDF probability density functionPoPA Power Pinch AnalysisPV photovoltaicRE renewable energySCT Storage Cascade Table

References

Bandyopadhyay, S., 2011. Design and optimization of isolated energy systemsthrough pinch analysis. Asia-Pac. J. Chem. Eng. 6, 518e526.

Battaglini, A., Lilliestam, J., Haas, A., Patt, A., 2009. Development of SuperSmartGrids for a more efcient utilisation of electricity from renewable sources.J. Clean. Prod. 17, 911e918.

Bernal-Agustn, J.L., Dufo-Lpez, R., 2009. Simulation and optimization of stand-alone hybrid renewable energy systems. Renew. Sustain. Energ. Rev. 13,2111e2118.

Borowy, B.S., Salameh, Z.M., 1996. Methodology for optimally sizing the combina-tion of a battery bank and PV array in a wind/PV hybrid system. Energ. Conv.IEEE Trans. 11, 367e375.

Chen, H.-C., 2013. Optimum capacity determination of stand-alone hybrid genera-tion system considering cost and reliability. Appl. Energ. 103, 155e164.

Feroldi, D., Degliuomini, L.N., Basualdo, M., 2013. Energy management of a hybridsystem based on wind-solar power sources and bioethanol. Chem. Eng. Res.Des. 91 (8), 1440e1455.

Georgakellos, D.A., 2012. Climate change external cost appraisal of electricity gen-eration systems from a life cycle perspective: the case of Greece. J. Clean. Prod.32, 124e140.

Goodbody, C., Walsh, E., McDonnell, K.P., Owende, P., 2013. Regional integration ofrenewable energy systems in Ireland e the role of hybrid energy systems forsmall communities. Int. J. Elect. Power Energ. Syst. 44, 713e720.

Grbe, P., Magyar, A., Hangos, K.M., 2012. Reduction of power losses with smartgrids fueled with renewable sources and applying EV batteries. J. Clean. Prod.34, 125e137.

Hocaoglu, F.O., Gerek, .N., Kurban, M., 2009. A novel hybrid (windephotovoltaic)system sizing procedure. Sol Energ. 83, 2019e2028.

Kaabeche, A., Belhamel, M., Ibtiouen, R., 2011. Sizing optimization of grid-independent hybrid photovoltaic/wind power generation system. Energy 36,1214e1222.Please cite this article in press as: Mohammad Rozali, N.E., et al., OptimalCleaner Production (2013), http://dx.doi.org/10.1016/j.jclepro.2013.12.028Komor, P., Glassmire, J., 2012. Electricity Storage and Renewables for Island Power ea Guide for Decision Makers. International Renewable Energy Agency (IRENA),Bonn, Germany.

Lou, J., Liao, Z., Jiang, B., Wang, J., Yang, Y., 2013. Pinch sliding approach for targetinghydrogen and water networks with different types of purier. Ind. Eng. Chem.Res. 52, 8538e8549.

Lund, H., Salgi, G., Elmegaard, B., Andersen, A.N., 2009. Optimal operation strategiesof compressed air energy storage (CAES) on electricity spot markets withuctuating prices. Appl. Therm. Eng. 29, 799e806.

Mohamed, A., Khatib, T., 2013. Optimal sizing of a PV/wind/diesel hybrid energysystem for Malaysia. In: Industrial Technology (ICIT), 2013 IEEE InternationalConference on. IEEE, pp. 752e757.

Mohammad Rozali, N.E., Wan Alwi, S.R., Abdul Manan, Z., Klemes, J.J., Hassan, M.Y.,2013a. Process integration of hybrid power systems with energy losses con-siderations. Energy 55, 38e45.

Mohammad Rozali, N.E., Wan Alwi, S.R., Manan, Z.A., Klemes, J.J., Hassan, M.Y.,2013b. Process integration techniques for optimal design of hybrid powersystems. Appl. Ther. Eng. 61, 26e35.

Munir, S.M., Manan, Z.A., Wan Alwi, S.R., 2012. Holistic carbon planning for in-dustrial parks: a waste-to-resources process integration approach. J. Clean.Prod. 33, 74e85.

Nasiraghdam, H., Jadid, S., 2012. Optimal hybrid PV/WT/FC sizing and distributionsystem reconguration using multi-objective articial bee colony (MOABC)algorithm. Sol Energ. 86, 3057e3071.

Nelson, D.B., Nehrir, M.H., Wang, C., 2006. Unit sizing and cost analysis of stand-alone hybrid wind/PV/fuel cell power generation systems. Renew. Energ. 31,1641e1656.

Notton, G., Lazarov, V., Stoyanov, L., 2011. Analysis of pumped hydroelectric storagefor a wind/PV system for grid integration. Ecol. Eng. Environ. Prot., 64e73.

Ould Bilal, B., Sambou, V., Ndiaye, P., Kebe, C., Ndongo, M., 2013. Multi-objectivedesign of PV-wind-batteries hybrid systems by minimizing the annualized costsystem and the loss of power supply probability (LPSP). In: Industrial Tech-nology (ICIT), 2013 IEEE International Conference on. IEEE, pp. 861e868.

Pickard, W.F., Shen, A.Q., Hansing, N.J., 2009. Parking the power: strategies andphysical limitations for bulk energy storage in supplyedemand matching on agrid whose input power is provided by intermittent sources. Renew. Sustain.Energ. Rev. 13, 1934e1945.

Purvins, A., Wilkening, H., Fulli, G., Tzimas, E., Celli, G., Mocci, S., Pilo, F., Tedde, S.,2011. A European supergrid for renewable energy: local impacts and far-reaching challenges. J. Clean. Prod. 19, 1909e1916.

Rajkumar, R.K., Ramachandaramurthy, V.K., Yong, B.L., Chia, D.B., 2011. Techno-economical optimization of hybrid pv/wind/battery system using Neuro-Fuzzy.Energy 36, 5148e5153.

Redpath, D.A.G., McIlveen-Wright, D., Kattakayam, T., Hewitt, N.J., Karlowski, J.,Bardi, U., 2011. Battery powered electric vehicles charged via solar photovoltaicarrays developed for light agricultural duties in remote hilly areas in theSouthern Mediterranean region. J. Clean. Prod. 19, 2034e2048.

Rotaru, F., Chicco, G., Grigoras, G., Cartina, G., 2012. Two-stage distributed genera-tion optimal sizing with clustering-based node selection. Int. J. Elect. PowerEnerg. Syst. 40, 120e129.

Saw, S., Lee, L., Lim, M., Foo, D., Chew, I., Tan, R., Klemes, J., 2011. An extendedgraphical targeting technique for direct reuse/recycle in concentration andproperty-based resource conservation networks. Clean Technol. Environ. Policy13, 347e357.

Shenoy, A.U., Shenoy, U.V., 2013. Targeting and design of CWNs (cooling waternetworks). Energy 55, 1033e1043.

Tay, D.H.S., Ng, D.K.S., 2012. Multiple-cascade automated targeting for synthesis of agasication-based integrated biorenery. J. Clean. Prod. 34, 38e48.

Tenaga Nasional Berhad, 2013. Industrial Pricing & Tariff. www.tnb.com.my/business/for-industrial/pricing-tariff.html (accessed 26.11.13.).

Tina, G., Gagliano, S., 2011. Probabilistic analysis of weather data for a hybrid solar/wind energy system. Int. J. Energ. Res. 35, 221e232.

Torres, C.M., Gadalla, M., Mateo-Sanz, J.M., Jimnez, L., 2013. An automated envi-ronmental and economic evaluation methodology for the optimization of a sourwater stripping plant. J. Clean. Prod. 44, 56e68.

Wan Alwi, S.R., Mohammad Rozali, N.E., Abdul-Manan, Z., Klemes, J.J., 2012.A process integration targeting method for hybrid power systems. Energy 44,6e10.

Zhou, W., Yang, H., Fang, Z., 2008. Battery behavior prediction and battery workingstates analysis of a hybrid solarewind power generation system. Renew. Energ.33, 1413e1423.

Cucek, L., Klemes, J.J., Kravanja, Z., 2012. A Review of Footprint analysis tools formonitoring impacts on sustainability. J. Clean. Prod. 34, 9e20.sizing of hybrid power systems using power pinch analysis, Journal of

Optimal sizing of hybrid power systems using power pinch analysis1 Introduction2 Methodology3 Case Study4 ConclusionAcknowledgementsAbbreviationsReferences

optimization power systems pinch analysis

Documents