Optimal size and cost analysis of stand-alone hybrid wind/photovoltaic power-generation systems

Civil Engineering and Environmental Systems

Date Submitted by the Author: Complete List of Authors:

rP Fo
Journal: Manuscript ID: Manuscript Type: n/a Keywords:

Optimal Size and Cost Analysis of Stand-Alone Hybrid Wind/PV Power Generation Systems

URL: http:/mc.manuscriptcentral.com/gcee

ee
Draft

Civil Engineering and Environmental Systems

Original Article

Yazdanpanah Jahromi, Mohammad Ali; University of Sistan and Balushestan, Mechanical Engineering Department Farahat, Said; University of Sistan and baluchestan, Mechanical Engineering Department Barakati, Seyed Masoud; University of Sistan and Balushestan, Power Electronic Engineering hybrid wind/PV systems, multi-objective optimization, sizing method, electricity match rate (EMR), match evaluation method (MEM), management strategy

rR ev ie w On ly

Page 1 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Manuscript # GCEE-2012-0150 (Optimal Size and Cost Analysis of Stand-Alone Hybrid Wind/PV Power Generation Systems) Amendments concerning the comments of Reviewer # 1
We thank the respected reviewer for the comments that he/she raised that helped us to enhance the quality of the paper. These comments are addressed separately below. 1) Comment #1: The models used in this paper have already been presented by the authors
in the Journal of Mathematics and Computer Science TJMCS Vol.5 No. 2 (2012) 134-145. This paper has to be cited and parts of the manuscript submitted to CEES, where the models are described, can be adequately shortened.

Amendments/Replies: Based on this comment of the respected reviewer, there are a few points which should be clearly highlighted. The mentioned previously work has been published in late December 2011. We have submitted the currents work in 18 December 2012 in Civil Engineering and Environmental Systems, and we didn’t know about our previous work publication since then. The mentioned reference is now included in the manuscript as reference [15]. We removed or added some texts to improve the quality of the paper in the revised version of the paper. Examples are: • • • • • • A text has been added to the end of second paragraph in Section 1 to include the mentioned reference. We found that Figures 3 and 4 (of the original paper) is out of the scope of this paper. So we removed these figures. We added a text and an equation to Section 2.2, after equation (8). Sections 2.3, 5.4 and 7 have been added to the article. Two text has been added to the first paragraph in section 8 to clarify how the solutions have been obtained. We added two paragraphs to the end of Section 8. A study of operating hours of diesel generator in optimal configuration is carried out and given in Table 9.

2) Comment #2: The choice of the two technical objectives (IC and CC) is very critical, in fact
due to the statistical variations of the load. A perfect generation, that overlaps the loads yearly profile, doesn’t represent an optimal solution as it introduces an high number of unavailability hours . To overcome this problem a storage system has to be considered in the optimization algorithm.

Amendments/Replies: We have include a diesel generator and a battery as backup and storage systems, respectively, to even out the irregularities. Moreover, a proper management strategy is designed to control the starting and stopping operation times of diesel generator. The mathematical modelling of battery storage system is given in section 2.3. Annual replacement cost (ARC), and annual fuel cost (AFC) is two cost parameters which appear after adding diesel generator and battery to the system. These two cost parameter are included in Equation (19) and their description given in
1
URL: http:/mc.manuscriptcentral.com/gcee

rP Fo

ee

rR

ev

ie

w

On

ly

Civil Engineering and Environmental Systems

Page 2 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

sections 5.3 and 5.4. The operation strategy for the proposed hybrid system is expressed in section 7. 3) Comment #3: The assumption of 10 year in the life time of the project is an evident
limitation considering the big difference between the life time of wind generator and PV modules.

Amendments/Replies: We appreciate the suggestions of the respected reviewer. The life time of the project is assumed to be 20 years. All the hybrid components is assumed to have a life time of 20 years, except the battery life time which considered to have a life time of 5 years. Life time specification for all hybrid components is included in Table 6. 4) Comment #4: It is not clear how the solutions have been generated. The authors combine
a system (wind turbine) with components of a PV system, i.e. PV modules. These solutions are practically useless as an inverter can accept only a certain number of PV modules. It means that the PV optimization requires also the consideration of inverters.

Amendments/Replies: We have added two text to clarify how the solutions have been obtained. The first one, is included in the last section of 4 as:

rP Fo

ee

rR

-

In this work, match evaluation method (MEM) is used for sizing purpose. This is first uses in [15]. St in Equation 11 and 12 is the sum of two or sometimes three parts; NPV.SPV, NWT.SWT, and Nbattery.Sbattery, or NPV.SPV, NWT.SWT, and Sdiesel which respectively denote the energy supply sources, PV modules, WTs, and battery, or PV modules, WTs, and diesel generator. NPV is the number of PV modules, NWT is the number of WTs, and NBattery is the number of battery. The management strategy controls the starting and stopping operation times of diesel generator.

The second one is include in the first paragraph in section 8 as follows: - Load is needed to be matched with different supplies in a way that resultant supply (N1.S1 + N2S2 + … +Nn.Sn) meet the load with high electricity match rate (EMR). Main objective of the proposed optimal algorithm is to find the optimal values of “N1, N2, …, Nn”. At the first paragraph of section 5 (just before section 5.1), we revised the following paragraph: Four main parts are considered in this revision article: wind turbine together with its tower, PV module, battery and other devices. The other devices which are not included in the decision variables, are equipment including inverter, controller, and cables.

5) Comment #5: The analysis needs further investigation, specifically concerning the following
aspects: 2
URL: http:/mc.manuscriptcentral.com/gcee

ev

ie

w

On

ly

Page 3 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1. Which is the criteria for selecting the wind turbines ? Further important parameters have to specified such as cut-in wind speed, cut-out wind speed, rated wind speed. 2. Does the selection consider the pdf wind speed distribution ? 3. Which is the criteria for selecting the PV modules? And what about the inverters ? Further important parameters have to specified such as the thermal coefficient of maximum power, and the thermal coefficient of open circuit voltage. 4. Does the selection consider the impact of ambient temperature on photovoltaic production?

Amendments/Replies: we appreciate the suggestion of the reviewer 1, and we have taken the above comment on-board when revising the manuscript. 1- At the section of 8, we revised the following text: - In order to determine the best values of parameters, evaluating convergence, several executions of the design program have been worked out. Each hybrid system has included one type of PV modules and one type of WTs together with diesel and battery storage systems. The results of these studies, suggest the choice of Kyocera Solar (KC200) PV module and Bornay (Inclin 3000) wind turbine for the proposed hybrid power system, compared to the other configurations. The reason of this selection is that this configuration provides the lowest cost in larger EMR. Other configurations, either have not optimal IC range ( 0 ≤ IC ≤ 0.4 ), or optimal CC range ( CC = +1, −1 ), or have higher cost than the selected configuration. In addition, detailed specification of the WTs are included in the article as Table 4. 2- Thanking this comment of the reviewer, we added the following text in the paper to the second paragraph in Section 2.2: - The Weibull PDF is basely depend on the shape factor, the scale factor, and the wind speed (equation 8). The shape factor will typically range from 1 to 3 [28]. The selection of shape factor is usually based on the experience and multiple observations of sites where wind speed data have been recorded. The η can be calculated using equation (9) [29]. ν η= (9) 1 Γ(1 + ) β where ν is the mean wind speed and Γ is the gamma function. Therefore, β and η are independent of WT specifications and they are constant for all WTs. For wind speed profile which shown in Figure 4, the Wiebull PDF has been plotted in Figure 5, with a shape Factor of 2.5.

3- As noted earlier, several executions of the design program have been done. Larger EMR in lower ACS is the criteria for selecting the mentioned configuration. The reason of this selection is that this configuration provides the lowest cost in larger EMR. Other configurations, either have not optimal IC range ( 0 ≤ IC ≤ 0.4 ), or optimal CC range ( CC = +1, −1 ), or have higher cost than the selected configuration.

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev
3

ie

w

On

ly

Civil Engineering and Environmental Systems

Page 4 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

4- Yes. The model predict the output power of PV panels in different meteorological conditions of ambient temperature and solar radiation. The monthly ambient temperature and solar radiation is added as Figures 1 and 2. Because of large number of inputs and complicated procedure, we prefer to simulate the system based on monthly data (in August). Finally we would like to thank this respected reviewer for his/her comments that helped us to improve the quality of the paper.

rP Fo
URL: http:/mc.manuscriptcentral.com/gcee

ee rR ev ie w On ly
4

Page 5 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Manuscript # GCEE-2012-0150 (Optimal Size and Cost Analysis of Stand-Alone Hybrid Wind/PV Power Generation Systems) Amendments concerning the comments of Reviewer # 2
We thank the respected reviewer for the comments that he/she raised that helped us to enhance the quality of the paper. These comments are addressed separately below. 1) Comment #1: The manuscript needs improvements before it can be considered for
publication. There are some pieces of information it seems that the authors have forgotten to include, the presentation of their results needs to be improved for clarity, and they need a native English speaker to read through the manuscript and correct the language. The paper needs to be reviewed for clarity to tune up recurring grammatical issues.

Amendments/Replies: We added some texts to improve the quality of the article in the revised version of it. Examples are: • Two text has been added to the first paragraph in section 8 to clarify how the solutions have been obtained. - Load is needed to be matched with different supplies in a way that resultant supply (N1.S1 + N2S2 + … +Nn.Sn) meet the load with high electricity match rate (EMR). Main objective of the proposed optimal algorithm is to find the optimal values of “N1, N2, …, Nn” - In order to determine the best values of parameters, evaluating convergence, several executions of the design program have been worked out. Each hybrid system has included one type of PV modules and one type of WTs together with diesel and battery storage systems. We have added a text as follow in the last section of 4 as: - In this work, match evaluation method (MEM) is used for sizing purpose. This is first uses in [15]. St in Equation 11 and 12 is the sum of two or sometimes three parts; NPV.SPV, NWT.SWT, and Nbattery.Sbattery, or NPV.SPV, NWT.SWT, and Sdiesel which respectively denote the energy supply sources, PV modules, WTs, and battery, or PV modules, WTs, and diesel generator. NPV is the number of PV modules, NWT is the number of WTs, and NBattery is the number of battery. The management strategy controls the starting and stopping operation times of diesel generator. Sections 2.3, 5.4 and 7 have been added to the article. We added two paragraphs to the end of Section 8. A study of operating hours of diesel generator in optimal configuration is carried out and given in Table 9. Correct. Thank you. We now have revised the WHOLE manuscript carefully and tried to avoid any grammar or syntax error. We believe that the language is now acceptable for the publication.

rP Fo

ee

rR

ev

ie

•

w

On

ly

• • •

5
URL: http:/mc.manuscriptcentral.com/gcee

Civil Engineering and Environmental Systems

Page 6 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2) Comment #2: The references for the prices of the equipment used should be noted. These
values change with time (and are currently different than what is noted in the paper - in fact I believe the Sky stream turbine's cost quoted is closer to its educational price, rather than MSRP). There is another challenge that Southwest Wind power has recently gone out of business and sold several of these products off. I don't think this should impact this paper being published, as the real focus of the paper is the process not the specifics, but the text should clarify where information has been collected from and mention that these values vary with time.

Amendments/Replies: Correct. The required information is now included in the manuscript, and we have change all the WTs in revised version. The references for the prices and detailed technical characteristics are included in the first paragraph of section 3. Other cost parameters are given in Tables 6 and 7. At the end of section 5, the following text is added: - It is worth mentioning that the price of these components change over time. 3) Comment #3: In this reviewers opinion, Figures 3 and 4 are not needed. This information
is basic background material which doesn't add value to the current paper. The paper should, however, provide the details of the shape factor and scale factor which fit the Zabol data as plotted in Figure 6. The paper is incomplete without this detail.

Amendments/Replies: We appreciate the suggestion of reviewer 2. We removed these figures. We added the following text in the paper to the second paragraph in Section 2.2: - The Weibull PDF is basely depend on the shape factor, the scale factor, and the wind speed (equation 8). The shape factor will typically range from 1 to 3 [28]. The selection of shape factor is usually based on the experience and multiple observations of sites where wind speed data have been recorded. The η can be calculated using equation (9) [29]. ν η= (9) 1 Γ(1 + ) β where ν is the mean wind speed and Γ is the gamma function. Therefore, β and η are independent of WT specifications and they are constant for all WTs. For wind speed profile which shown in Figure 4, the Wiebull PDF has been shown in Figure 5, with a shape Factor of 2.5.

4) Comment #4: Figure 8's description should be modified to note that this is the energy
output by wind speed for one month and also include which month it is. An annual plot would be useful to show as well.

Amendments/Replies: Monthly and yearly plots is included in Figures 7 and 8 for the given WTs in Table 2. The corresponding text in now added under the figures. 5) Comment #5: It is not explained in the text why there are values such as "7247/27" in
Tables 4 and 5. What is the division symbol being used to represent in these values? The results of the optimization should be very clearly explained. 6
URL: http:/mc.manuscriptcentral.com/gcee

rP Fo

ee

rR

ev

ie

w

On

ly

Page 7 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Amendments/Replies: Sorry. We correct them in the revised version of manuscript. 6) Comment #6: Figure 14 is the most important figure in this study, and it is unfortunately
quite difficult to interpret. Perhaps the scale could be changed to zoom in on the results. Arrows overlaying the direction of the optimal solutions would help. Perhaps color coding the top several designs would also make this more intuitive to interpret.

Amendments/Replies: Some ranges that there are no values have removed in 3D Pareto front for the best configuration which is now Figure 15 in revised version. Arrows to overlay the direction of optimal solutions are now included in Figures 12, 13, 14, and 15. 7) Comment #7: An additional result of this study that would be quite interesting to see is a
plot showing the resulting generation and demand over some period of time from the optimized system designs and an evaluation of the amount of backup generation or storage that would be needed to make each scenario work for a stand-alone application (if that is an objective of the work).

Amendments/Replies: In the current paper, in a novel approach a battery and also a diesel generator have been added to different hybrid systems as back-up and storage systems, respectively. A suitable management strategy controls the starting and stopping time of diesel generator which is a critical factor for optimization. The algorithm and the management strategy are hourly basis. Study of operating hours of diesel generator in optimal configuration is carried out. Other simulation results is out of the objectives of this work. Because of large number of inputs and complicated procedure, we prefer to simulate the system based on monthly data (in August). Finally we would like to thank this respected reviewer for his/her comments that helped us to improve the quality of the paper.

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev
7

ie

w

On

ly

Civil Engineering and Environmental Systems

Page 8 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Optimal Size and Cost Analysis of Stand-Alone Hybrid Wind/PV Power Generation Systems Mohammad Ali Yazdanpanah Jahromia∗, Said Farahatb, Seyed Masoud Barakatic
a

Department of Mechanical Engineering, University of Sistan and Baluchestan, Zahedan, Iran

rP Fo

b

Department of Mechanical Engineering, University of Sistan and Baluchestan, Zahedan, Iran

c

Department of Power Electronic Engineering, University of Sistan and Baluchestan,

∗ Corresponding author. Tel.: ++98-917-392-3846; fax: +0-541-244-7092; e-mail: [email protected]

URL: http:/mc.manuscriptcentral.com/gcee

ee

Zahedan, Iran

rR ev ie w On ly

Page 9 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Optimal Size and Cost Analysis of Stand-Alone Hybrid Wind/PV Power Generation Systems
Design of sustainable energy systems for the supply of electricity need correct selection and sizing to reduce investment costs. In this article, a new sizing methodology is developed for stand-alone hybrid wind/PV power systems, using multi-objective optimization algorithms. MOPSO algorithm and NSGA-II are selected related to their match with nature of renewable energy sizing problem. A match evaluation method (MEM) is developed based on renewable energy supply/demand match evaluation criteria, to size the proposed system in lowest cost. As an example of application of this technique, six different wind turbines and also six different PV modules have been considered. The sizing methodology determines a multi-objective design, obtaining the best solutions that the applied algorithm has found simultaneously considering three objectives: inequality coefficient (IC), correlation coefficient (CC), and annualized cost of system (ACS). The optimal number of wind turbines, PV modules and batteries ensuring that the system total cost is minimized while guaranteeing a highly reliable source of load power is obtained. A management strategy has been designed to achieve higher electricity match rate (EMR). Based on the proposed technique, the algorithm developed for different cases, using the climatic condition data of the city Zabol, located in south-east of Iran. Additionally, a study of operating hours of diesel generator in optimal configuration is carried out. Keywords: Hybrid wind/PV systems; multi-objective optimization; sizing method; electricity match rate (EMR); match evaluation method (MEM),

1. Introduction

Nowadays, the renewable energy and estimation of energy production are popular research areas and they should be further investigated. Fast depletion of conventional energy resources, rise in the fuel prices, harmful emissions from the burning of fossil fuels and growing in energy demand have made power generation from conventional energy sources unsustainable. Standalone hybrid power systems are promising

rP Fo
management strategy

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Civil Engineering and Environmental Systems

Page 10 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

alternatives particularly in remote areas as an isolated small power producing units for the supply of power. Energy security under varying weather condition and the corresponding system cost are the two major issues in designing of hybrid power generation systems. Hybrid power systems can address the limitations of reliability, efficiency, cost and emission on individual renewable energy supply options. The two technologies that have seen the most significant growth are wind turbine (WT) and solar photovoltaic (PV). Wind power has recently become the fastest growing renewable energy resource and is projected to lead the growth of the renewable power portfolio in the near term [1]. Solar energy, both as a thermal and an electric options, is well suitable for the built environment. The solar energy distribution is mostly periodic and the wind speed may present stochastic patterns. These both together, present supplementary availability. Hence the combined exploitation of the available wind and solar potential caused reliable power generation. The reliability of hybrid systems are important to both planning and utilization stages. Designing energy systems including solar and wind energies together, to some extent, reduce the depth of the problem [2]. M. Vafaei (2011) showed that hybrid stand-alone power generation systems are usually more reliable and less costly than systems that use only single source of energy [3]. Design and modelling are important aspects of the analysis of hybrid power systems. There are many studies carried out by various researchers in the field of renewable energies. The design of hybrid systems is usually done by searching the configuration and/or control with the lowest total cost throughout the useful life of the installation or pollutant emissions. Optimal sizing of the hybrid components is one of the main issues related with the application of such hybrid alternative energy systems. Proper sizing ensures the load demand supply in all conditions with the lowest total cost. Hence, the economic disadvantages of renewable energy systems compared to

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Page 11 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

conventional energy sources can partially be overcome. By increasing number of hybrid components, the sizing procedure will be more complex. Variation of load demand in different time intervals that may not match with the generation power of renewable energy sources further promotes the complex structure of the sizing procedure. The sizing methodologies for hybrid power systems available in the literature cover a large spectrum of approaches. Energy system reliability is one of the key aspects regarding stochastic nature of renewable sources. One parameter that helps to elucidate the system reliability is loss of power supply probability (LPSP) technique. Optimal configuration was calculated by LPSP technique and minimum annualized cost of system (ACS) in [4-10]. One other sizing method commonly used is based on levelized cost of energy (LCE) as used in [11, 12]. The LCE can be defined as a metric that describes the cost of every unit of energy generated by a project. Some other sizing method, deal with pollutant emission and unmet load. A multi-objective design of hybrid systems by minimizing the total cost, pollutant emission and unmet load is presented in [13]. LunaRubio et al. (2012) reviewed different sizing methodologies developed in the recent years in [14]. Match evaluation method (MEM) which is used in this article is another sizing method [15]. The MEM is based on the coordination criteria between generation

and consumption intervals.

techniques are not able to take into account all the characteristics associated to the posed problem consuming excessive CPU time. Nevertheless, modern optimization methods obtain a set of non-dominated solutions with little computational effort. Most of these methods are based on certain characteristics and behavior of biological, swarm of insects, molecular, and neurobiological systems. Particle swarm optimization (PSO) algorithm and genetic algorithm (GA) are the two of modern stochastic optimization

rP Fo

The design process is generally very complex. The classical optimization

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Civil Engineering and Environmental Systems

Page 12 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

methods that can find Pareto-optimal solution in one single simulation run. Y. S. Zhoaand et al. (2006) proposed an optimized wind/PV hybrid power system using PSO algorithm to have higher capacity and faster search efficiency [16]. J. Dhillon (2009) proposed the non-dominated sorting genetic algorithm (NSGA-II) to simultaneously minimize the total system real power losses in transmission network and cost, by satisfying power balance equation [17]. Using of genetic algorithm (GA) in unit sizing of photovoltaic/wind generator systems is discussed in [18]. O. Erdinc and M. Uzunoglu (2012) reviewed different optimum sizing approaches in literatures [19]. In this article, a new optimum sizing methodology for stand-alone hybrid

systems is developed based on MEM at the lowest investment. The electricity match rate (EMR) technique, which is considered to be the criteria for sizing, is employed in different wind/PV hybrid systems. In this procedure, three objectives are proposed. They are inequality coefficient (IC), correlation coefficient (CC) and annualized cost of system (ACS). IC and CC control the EMR while ACS checks the system cost. IC provides a measure of how well a time series of estimated values compares to a corresponding time series of observed values [20]. In other word, IC provides a relative measure of forecast accuracy in terms of deviation from the perfect forecast [21]. CC measure of how well the predicted values from a forecast model "fit" with the real-life data [22]. IC gives the match magnitude while CC, deals with trend matching. Hence, IC and CC are selected together, to check the EMR between supplies and load demand. These two objectives together, can obtain good match rate for hybrid systems. More coordination between the components increases the system efficiency. When the output power of renewable energy resources cannot meet the load demand, the strategy will be used to start the diesel generator or use the battery power. The optimization algorithm selects the optimal size of hybrid components based on above procedure. The

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Page 13 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

optimum combination of hybrid wind-PV system can make the best compromise between the three considered objectives: IC, CC and ACS. The MEM sizing technique is implemented based on multi-objective particle swarm optimization (MOPSO) algorithm. The results are validated by NSGA-II. Six different kinds of wind turbine (WT) and also six different kinds of PV module, with different output powers and costs are considered for this optimization procedure to investigate the efficiency of the proposed methodology. The simulation is carried out based on the basis of the algorithm developed for different cases using the climatic condition data of the city Zabol, located in south-east of Iran. Using the EMR objective functions, the configuration of the proposed hybrid system which gives the highest EMR requirements can be obtained. The decision variables included in the optimization process are the number of PV module, WT, and battery. The optimization is conducted by an iterative simulation using a system model and real weather and load demand data.

2. Modelling of Hybrid Wind/PV System

In order to predict the performance of a hybrid power system, individual components need to be modelled first, and then the generation power can be evaluated to meet the load demand. The proposed hybrid power generation system consists of WT, PV array, inverter, cables and other accessory devices. Weather data from the city of Zabol obtained from a nearby meteorological station is used for more precise estimation of the local potential of both solar and wind energy. A brief description for modelling of windPV-battery system is presented in forthcoming subsection.

2.1. Mathematical Model of PV Module PV technology is identified as most environment friendly technologies [23]. Simulation of PV array performance has been done by considering the modelling of the maximum

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev ie w On ly

Civil Engineering and Environmental Systems

Page 14 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

power point tracking (MPPT) controller. This model can predict output power of PV panel in different temperatures and various irradiation levels. The PV panel model depending on the solar radiation and the temperature can be calculated as [24]:
I(V)= V 1   . 1 − exp( − ) 1 bVx b  . 1-exp(- )  b Ix

(1)

Vx = s.

V −V oc Ei E .TCV .(T − T N ) + s V . max − s .(V max −V min ).exp( i .ln( max )) E iN E iN V max −V min
Ix = p . Ei .[ I sc + TCi .(T − T N ) ] E iN

(2) (3) (4)

where P is the output power of the photovoltaic panel [W], I(V) is the output current of the photovoltaic panel [A], V is the output voltage of the photovoltaic [V], Isc and Voc are the short-circuit current and the open-circuit voltage at 25 [°C] and 1000 [W/m2], respectively, Vmax, is the maximum open-circuit voltage at 25 [°C] and 1200 [W/m2] (usually Vmax is close to 1.03Voc), Vmin is the minimum open-circuit voltage at 25 [°C] and 200 [W/m2], (usually, Vmin is close to 0.85Voc), T is the solar panel temperature [°C], Ei, is the effective solar irradiation impinging the cell in [W/m2], Tw is 25 [°C] standard test condition (STC), TCi is the temperature coefficient of Voc in [V/C], Ix and Vx is short circuit current and open circuit voltage, respectively, at any given Ei and T ; s is the number of photovoltaic panels in series, p is the number of photovoltaic panels in parallel, b is characteristic constant based on I-V curve. The characteristic constant, b, usually varies from 0.01 to 0.18 and can be calculated using (5) with iterative procedure [25].

rP Fo

P (V ) =

V .Ix 1 − exp( −1 ) b

.[1 − exp(

V 1 − )] . bVx b

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Page 15 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

b n +1 =

V op −V oc I −1 V oc .ln(1 − op .(1 − exp( ))) Isc bn

(5)

For calculating the available energy of PV module at a specific site the following Equation is used [25]:
E PV = Pout (E x ).(SolarWindow ).(TotalDay )

(6)

where Epv is the expected production of photovoltaic energy in [kWh], ”SolarWindow” is the total time hours that sun hit the PV module at an average hourly solar irradiation, the product of “TotalDay” is to change from daily to monthly or yearly quantities, Pout(Ex) is the PV module output power at an average hourly solar irradiation (Ex). Figure 3 shows the P-V and I-V curves for each photovoltaic module given in Table 1 by using available solar radiation and ambient temperature which is given in Figures 1 and 2, respectively, for the selected site.

2.2. Mathematical Model of Wind Turbine

Adjusting the measured wind speed to the hub height, by using the wind speed data at a reference height from the database, is an important phase before calculating the output power of WTs. This can be done through the following expression [26]:

where V2, is the wind speed at the desired height H2, V1 is wind speed measured at known height H1, α is wind shear exponent coefficient which varies with pressure, temperature and time of day. A commonly used value for open land is one-seventh (1/7). The variations in wind speed are best described by the Weibull probability distribution function (PDF), ‘f(v)’, with two parameters, the shape factor ‘β’, and the scale factor ‘η’ [27]. The PDF calculates the probability that the wind speed will be

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR
V2 =(

ev
H2 α ) V1 H1

ie

w

On

(7)

ly

Civil Engineering and Environmental Systems

Page 16 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

occurred between zero and infinity during the entire chosen time period. Note that the PDF curve shape and the height of it provide in some way that the area under the PDF curve is unity. There are various notations for Weibull PDF in articles. In this article, the Weibull PDF is defined as [27]:
f (v ) =

β v β −1 − ( η )β ( ) e η η

v

(8)

where β is the shape factor, η is scale factor, and v is the wind speed. The value of β controls the curve shape and hence is called the shape factor. The larger shape factor indicates a relatively narrow distribution of wind speeds around the average while the lower shape factor indicates a relatively wide distribution of wind speeds around the average. The scale factor (η) defines where the bulk of the distribution lies and how stretched out [25]. The Weibull PDF is basely depend on the shape factor, the scale factor and the wind speed. The shape factor will typically range from 1 to 3 [28]. The selection of shape factor is usually based on the experience and multiple observations of sites where wind speed data have been recorded. The η can be calculated using Equation

(9) [29].

where ν is the mean wind speed and Γ is the gamma function. Therefore, β and η are independent of WT specifications. For wind speed profile which shown in Figure 4, the Wiebull PDF has been plotted in Figure 5, with a shape Factor of 2.5. The wind speed distribution (PDF) is the key information needed to estimate the total kWh produced in a period of time by a WT at a given site. And then, using the WT power curve the annual energy output can be calculated. A power curve is a graph that presents the output power of wind turbine at any wind speed. This curve is a function of the turbine

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR
η= ν

ev
Γ (1 + 1

ie w
) (9)

β

On

ly

Page 17 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

design and normally obtained from the wind turbine manufacturer. The power curves of the WTs which given in Table 2, can be seen in Figure 6. The energy available for a WT at a specific site can be calculated using (10) [30].

EWT = (days )(hours ).∑ Pc f (ν , β ,η )
υ =1

25

(10)

where Ewt is the expected energy production of WT in kWh for a specific site. The product of days by hours, gives the total hours in the period of simulation, Pc is the output power of wind turbine; f(v) is the Weibull PDF for wind speed (ν), β is the shape factor, and η is the scale factor. Monthly (August) and yearly total energy outputs for each WT are presented in Figures 7 and 8, respectively.

2.3. Modelling of Battery Storage System

Stochastic nature of renewable sources make their supply really intermittent and unreliable. This characteristic necessitate the use of energy storage system in renewable power systems. Batteries are the most widely used devices for energy storage. Leadacid batteries are usually used for stand-alone hybrid wind-PV-diesel generation systems [27]. Batteries are required to even out irregularities in the solar and wind power distributions. Surplus electrical energy is stored in a battery bank which supplies power to the load when the total power output of WTs and PVs is insufficient. So the correct battery sizing is critical. There are different models in literatures for battery behaviour simulation. The modelling of battery based on state of charge (SOC) is the most commonly used model. SOC is an important parameter in system assessments
' [31]. Temperature can also affect battery capacity. The available battery capacity ( C bat

[AH]), in a given temperature (T bat [K]), can be calculated using (11) [32].
' '' C bat = C bat .(1 + δ c .(T bat − 298.15)),

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

(11)

Civil Engineering and Environmental Systems

Page 18 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

where δc is temperature coefficient. The value of 0.6% per degree is usually used for δc , unless otherwise specified by the manufacturer. For the proposed hybrid WPDB system, it is supposed that the WTs has the DC outputs. If the cable losses in the system are neglected, the battery current rate at time t can be expressed as (12) [31].
PPV (t ) + P W ind (t ) − PACLoad (t ) V bat (t )

I bat (t ) =

ηinverter − PDCLoad

(12)

where ηinverter is the inverter efficiency which is considered as 92% in this study. The SOC at any hour t is depending on the battery current, the charge or

discharge time and previous state of charge. By all above consideration the battery SOC can be defined as [31]:

where η bat is the battery efficiency, which 90% for charging stage and 100% in discharging process are recommended. σ is the self-discharge rate; 0.2% per day is recommended. When the wind turbine and PV module supply power more than the load demand, the overcharging process is occurred. On the other hand, when the load demand is more than the total output energy of supply sources, the battery SOC may decrease to the minimum level which is defined as SOCmin = 1 - DOD, where DOD is the depth of discharging of battery. In this study, for longevity of battery lifetime, the value of DOD is considered 60%. In order to prevent the batteries against destruction, it is important to control the batteries SOC at the following constrain:

where SOCmax is the maximum state of charge for batteries (SOCmax = 1).

rP Fo

SOC (t + 1) = SOC (t ).(1 −

URL: http:/mc.manuscriptcentral.com/gcee

ee

SOC min ≤ SOC ≤ SOC max

rR

σ .∆t
24

)+

I bat (t ).∆t .ηbat ' C bat

(13)

ev

ie

w

On

ly

(14)

Page 19 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2.4. Load Model
The output power of the proposed hybrid system should meet the power load demand. The hourly load data in August used in this study is shown in Figure 9. This is the Monthly variation of domestic load profile in the region.

3. The Components Characteristics
There are six possible different PV generators and also six possible different WTs. The specifications of these components used to design and to optimize the hybrid wind/PV are presented in Tables 1 and 2, respectively [25]. Detailed technical characteristics of WTs, PVs and battery which is given by the Manufacturers are given in Tables 3, 4 and 5, respectively [25, 33]. Each hybrid system will include one kind of PV as well as one kind of WT with battery and diesel.

[Table 1. Solar module power at standard test condition rating and price] [Table 2. Small wind turbines rating and price]

[Table 3. Solar module specifications at standard test condition rating] [Table 4. Detailed specifications of the wind turbine] [Table 5. Battery characteristic]

[Figure 1. Meteorological conditions of solar radiation in August]

[Figure 2. Ambient temperature in August] [Figure 3. P-V and I-V Curves]

[Figure 4. Meteorological conditions of wind speed in August] [Figure 5. Weibull probability density function (f(v))]

[Figure 6. Wind turbine power curves (The symbols represent data sampled from the power curve graphs given by the manufacturer)]

[Figure 7. Total wind turbine energy outputs by wind speed in August] [Figure 8. Total wind turbine energy outputs by wind speed for one year] [Figure 9. Monthly (August) variation of domestic load profile]

4. Sizing Model Based on Match Evaluation Method (MEM)
There are challenges in term of finding the correct capacity for hybrid power generation

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On ly

Civil Engineering and Environmental Systems

Page 20 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

systems. The maximization of EMR between demand and supply intervals is an important subject in hybrid power systems. For quantifying the magnitude of deviation between two set of data variables, the least squares (LS) approach can be used. The following Equation describes LS [34]:

LS = ∑ (Dt − S t ) 2
t =0

n

(15)

where Dt and St are demand and supply at time t, respectively. The LS in Equation (15) is always a positive quantity. Zero value of LS indicates a perfect match. Spearman's Rank correlation coefficient (CC) is one of the objectives which can describe the correlation between supply and demands. Calculation of this coefficient will always result in a value between -1 and 1. Result of “1” shows the perfect positive match while “-1” indicates perfect negative match. In perfect negative match ( CC = −1 ), as one variable tends to increase the other will decrease at the same rate and vice versa for the perfect positive match ( CC = +1 ). Value of “0” represents no match. The correlation coefficient, CC, can express as [35]:
n

where Dt and St are the load demand and supply at time t, respectively; d and s are the mean demand and supply over time period n, respectively. The CC is used to describe the trend matching between the time series of two data sets. It does not explain the relative match magnitudes of the individual variables. Thus, if the size of a power supply doubled, however the excess supply would be far greater, the CC would stabilize the same. Moreover, if two profiles are perfectly in phase with each another, but of very different magnitudes, would result in perfect correlation, but not a perfect match rate. For a perfect match rate, both phase and magnitude must be considered. Hence, another

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee
CC =
n t =0

rR
∑ (D
t =0 t

ev
n t =0

− d ).(S t − s )

ie w On ly
(16)

∑ (D

t

− d )2 .∑ (S t − s ) 2

Page 21 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

criterion is needed to determine the match magnitude. The inequality coefficient (IC), describes the inequality in the magnitude domain due to three sources: unequal tendency (mean), unequal variation (variance) and imperfect co-variation (co-variance) [35]. Therefore, IC and CC are selected together, to check the EMR between supplies and load demand. The resultant IC can range in value between 0 and 1.The smaller IC denotes the larger match rate. Value of “0” represents a perfect match while “1” shows no match. The IC can be given by the following Equation [35]:
1 n ( Dt − S t ) 2 ∑ n t =0 1 n 1 n ( Dt ) 2 + ∑ ∑ (S t )2 n t =0 n t =0

where Dt and St are demand and supply at time t, respectively. n is the total time period. Values of IC between 0 - 0.4 represents good match and value above 0.5 shows weak match [36]. IC is more important criterion than CC in determining the strength of matching between supplies and demand. However CC is also good but it is not as well as IC. In this work, match evaluation method (MEM) is used for sizing purpose. This is first uses in [15]. St in Equation 11 and 12 is the sum of two or sometimes three parts;

NPV.SPV, NWT.SWT, and Nbattery.Sbattery, or NPV.SPV, NWT.SWT, and Sdiesel which respectively
denote the energy supply sources, PV modules, WTs, and battery, or PV modules, WTs, and diesel generator. NPV is the number of PV modules, NWT is the number of WTs, and

NBattery is the number of battery. The management strategy controls the starting and
stopping operation times of diesel generator.

5. Cost Analysis Based on ACS Concept
A cost analysis of the system is performed for each configuration according to the concept of annualized cost of system (ACS) [4, 6, 37]. An optimum combination of a

rP Fo

IC =

(17)

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Civil Engineering and Environmental Systems

Page 22 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

hybrid wind-PV-battery energy system must satisfy both the reliable and economical requirements. For all configurations, the ACS is composed of the sum of individual annualized capital cost of components (ACC), annualized operation and maintenance cost (AOC), annual replacement cost (ARC), and annual fuel cost (AFC). Four main parts are considered: wind turbine together with its tower, PV module, battery and other devices. The other devices which are not included in the decision variables, are equipment including inverter, controller, and cables. The ACS is expressed by:

5.1. Annualized Capital Cost

The annualized capital cost (ACC) of each component has taken into account the installation cost, and can be calculated using (19).

where Ccap is capital cost of each component, US$; and CRF is capital recovery factor, a ratio that calculate the present value of an annuity (a series of equal annual cash flows).

CRF can define as:
CRF (i , n ) = i (1 + i ) n (1 + i ) n − 1

where n is the component lifetime, year; i is the annual interest rate related to the nominal interest rate (iloan, the rate at which a loan could be obtained) and the annual inflation rate, f, by the Equation given below:
i = i loan − f 1+ f

rP Fo

ACS = A CC (PV +W ind + Tower + Diesel + Battety + Other ) + AOC (PV +W ind + Tower + Battey + Other ) + ARC ( Battery ) + A FC ( Diesel )

(18)

URL: http:/mc.manuscriptcentral.com/gcee

ee

A CC = C cap .CRF (i , n )

rR

ev ie w On ly

(19)

(20)

(21)

Page 23 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

5.2. Operation and Maintenance Cost
The operation and maintenance cost is the maintenance and repair cost of hybrid components. The maintenance cost of each component, which has taken the inflation rate f into account, can be calculated as follows:

A OC = A OC (1) * (1 + f ) n

(22)

where AOC(1) is the maintenance cost of that component for the first year of the project.

5.3. Annual Replacement Cost
The components which have a lifetime less than the project lifetime need to be replaced during the lifetime of the project. The ARC can be calculated from Equation (23).
A RC = C rep * SFF (i , n rep )

Where Crep, is replacement cost of units, US$,

calculate the future value of a series of equal annual cash flows, depends on lifetime of units (nrep) and interest rate (i), as given below.
SFF (i , n rep ) =

5.4. Annual Fuel Cost (AFC)

The cost of fuel for diesel generator is calculated using the following Equation (25).

where T fc is total fuel consumption for lifetime of the project.

The fuel (gas-oil) price is considered 0.16049 $/kWh. The expected CO2 emission is 0.669 kg/kWh. The output power of diesel generator is 500 W. The initial capital cost of different PV modules and WTs, are given in Table 1 and 2, respectively [25]. Other cost parameters used in this paper are shown in Tables 6 and 7 [3, 31]. It is worth mentioning that the price of these components change over time.

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

(23) is sinking fund factor, a ratio to

AFC = T fc *CRF (i , n )

rR

SFF

ev

i (1 + i ) nrep − 1

ie w On ly

(24)

(25)

Civil Engineering and Environmental Systems

Page 24 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

[Table 6. The cost and lifetime aspect for the proposed hybrid components] [Table 7. Cost parameters]

6. Multi-Objective Optimization Procedure Using MOPSO
The sizing of the hybrid wind/PV systems is much more complicated than the single source power generation systems. It is because of more variables and parameters that have to be considered in system optimization. Long-term system performance, economical parameters and EMR objectives must be considered in order to reach the best compromise for both power match rate and cost. PSO and GA are the most suitable algorithms in term of global optimization, stochastic nature of renewable sources, and particular nature of sizing method. Implementation of MOPSO and NSGA-II has been done in various engineering and business applications in recent years. In this article, a multi-objective PSO (MOPSO) algorithm is employed to size the hybrid stand-alone wind/PV power generation systems. The PSO algorithm was originally proposed by Kennedy and Eberhart in 1995 [38]. Generally, PSO is based on a simple concept, short time computation, easy implementation and few memory requirements. It works by maintaining a population of candidate solutions to the problem. Each solution is considered as a particle. The particles move through the search space. In every iteration of the algorithm, the fitness of each candidate solution is evaluated. Best value of fitness achieved so far is remembered as its personal best fitness. The global best fitness and the candidate solution that achieved this fitness are also remembered. The local and global bests are updated in each iteration, in order to achieved better fitness. For the purpose of this article, six different WTs and also six different PV modules, with different characteristics and costs are considered. Backup and storage systems such as diesel generator and battery storage is needed to have better EMR and also even out irregularities. The system configurations will be optimized by

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Page 25 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

employing the MOPSO algorithm, which dynamically searches for the optimal configuration to maximize the EMR and also minimize the ACS. IC, CC and ACS, are the three objective functions of this optimization process which conflicting to each other. By mathematically formulating of multi-objective design problem and applying it to each configuration of hybrid system, the best combination of components (minimizing IC and ACS and also maximizing the CC) can been obtained. The solutions are validated by NSGA-II. These multi-objective optimization algorithms can find Pareto-optimal solution in one single simulation run. With the same number of iteration and population, the MOPSO algorithm has higher speed than NSGA-II. Proper sizing algorithm is the one which can find the optimal size of each

component in each configuration to maximize the EMR between demand and supplies. The number of PV panels, WTs and batteries are design variables. The minimum value (lower limit) of design variables is selected 1 to be sure that there is at least one of each supply in the system and the upper bound of them is set as max(D ) min(S n ) ,where max(D) and

min(Sn) are the maximum and minimum values of demand and supplies over considered
time period, respectively. The input data for the simulation are ambient air temperature, WT installation high, hourly solar irradiation on a horizontal surface, wind speed, and load demand data for one month. By employing the MOPSO algorithm to each configuration, a set of possible solutions (Pareto set) will be obtained. The flow chart of

the optimization process is presented in Figure 10.

[Figure 10. Flow chart of optimization process using MOPSO (or NSGA-II)] The optimization process starts with taking input values of hourly data for WT and PV module output powers, and load demand profile. Then a for loop is run to call MOPSO algorithm (or NSGA-II) several times (n) with objectives of minimizing IC and ACS, and also maximizing CC. The results of optimum capacity generated by

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Civil Engineering and Environmental Systems

Page 26 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

MOPSO algorithm (or NSGA-II) in decimal values are then converted to closest integer values in order to get whole value of each sizing of supply sources. Optimum sizing results along with corresponding IC, CC and ACS values are recorded for each run and stored in arrays, which update in each iteration. When a pre-specified iteration count (

n = n max ) is reached, the algorithm is terminated. The n max = 200 , and a population size
of n pop = 200 , are considered. A set of possible solutions (Pareto set) with relative number of each supply will be finally provided.

7. Operation of Proposed Hybrid System
In order to have continuous power generation, there is need for backup and

storage systems such as diesel generator and battery storage. Diesel generator fuel consumption is an important subject for entire operation cost. Hence, one of the critical factor for optimization is managing the starting and stopping time of diesel generator. The operation strategy could be explained in detail as fallow.

rP Fo

ee

rR

ev

•

If the total power generated from wind turbines (PWT) and PV panels (PPV) is more than the load demand (PL), the excess power is used to charge the batteries. In this case, the sizing optimization will be done with only two supplies.

ie

w

On

•

If the total generated power (PWT + PPV) is less than the load demand, and SOC of batteries is higher than SOCmin, the batteries will supply the extra power. If the batteries SOC are equal or less than SOCmin, the diesel generator will start and supply the power in order to protect the batteries against excessive draining. Surplus power from diesel will charge the batteries as amount as SOCmax. In this case, the calculation of sizing optimization is divided in two categories: PV-wind-battery and PV-wind including one diesel generator. The decision parameters for the optimization algorithm are the numbers of PV modules, wind turbines, and batteries.

ly

URL: http:/mc.manuscriptcentral.com/gcee

Page 27 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The proposed management strategy is presented as a flowchart in Figure 11. The optimized model gives the optimal size for hybrid components based on MEM and minimum ACS. To calculate the CO2 emission and fuel cost of diesel generator, they are calculated correspond with the generating kWh, every time diesel generator started. [Figure 11- Flowchart of operation strategy]

8. Optimization Results and Discussion
The output power of PV arrays and WTs, have been calculated according to the model which described before, by specifications of PV modules and WTs, given in Tables 1 and 2. The sizing of optimal hybrid wind/PV systems was achieved using multiobjectives PSO algorithm and GA approaches. The algorithm has been implemented using data collected of the city Zabol, south-east of Iran, based on the MEM. Load is needed to be matched with different supplies in a way that resultant supply (N1.S1 +

N2S2 + … +Nn.Sn) meet the load with high electricity match rate (EMR). Main objective
of the proposed optimal algorithm is to find the optimal values of “N1, N2, …, Nn”. The obtained solutions for goal functions are presented as optimal Pareto front. The inequality coefficient (IC), correlation coefficient (CC) and annualized cost of system (ACS) were computed, using the developed program for each combination. It can be noted that the increasing of IC implies the decreasing of CC. It is the same for both IC vs. ACS and CC vs. ACS. In order to determine the best values of parameters, evaluating convergence, several executions of the design program have been worked out. Each hybrid system has included one type of PV modules and one type of WTs together with diesel and battery storage systems. The results of these studies, suggest the choice of Kyocera Solar (KC200) PV module and Bornay (Inclin 3000) wind turbine for the proposed hybrid power system, compared to the other configurations. The reason of this selection is that this configuration provides the lowest cost in larger

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Civil Engineering and Environmental Systems

Page 28 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

EMR. Other configurations, either have not optimal IC range ( 0 ≤ IC ≤ 0.4 ), or optimal CC range ( CC = +1, −1 ), or have higher cost than the selected configuration. Owing to

above considerations, the Pareto front has been plotted for IC vs. CC, IC vs. ACS, and also CC vs. ACS, as shown in Figures 12, 13 and 14, respectively. The evolution of the 3D Pareto front can be observed in Figure 15. [Figure 12. 2D Pareto front for the last generation. Inequality coefficient (IC) vs. correlation coefficient (CC)] [Figure 13. 2D Pareto front for the last generation. Inequality coefficient (IC) vs. annualized cost of system (ACS)] [Figure 14. 2D Pareto front for the last generation. Correlation coefficient (CC) vs. annualized cost of system (ACS)]

[Figure 15. 3D Pareto front for the best configuration] The results obtained by the optimization algorithms for the best configuration are summarized in Table 8. They are the optimum combination of hybrid components (PV, WT, and battery) needed to supply the energy to the load demand at the lowest cost possible. The 11 best sizing selections from 30 runs for the best configuration have been obtained. The calculations of the goal functions have also been done for these sizing numbers.

[Table 8. Pareto front/optimal solutions obtained from multi-objective optimization] As mentioned earlier, IC is more important factor than CC and ACS. Table 8 show that the values 6, 2 and 3 for PV modules, WTs, and battery units, respectively, give us better match rate respect to others. But this configuration has the highest ACS which can be very important in practical system installation. One scope of using hybrid renewable energy systems is to use green energies like solar and wind instead of fossil fuels. So the total hours the diesel generator operates in one month (744 hours) can be one of the criteria in the selection of optimal solutions. Table 9 shows the total hours the diesel operates in one month (August).

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Page 29 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

[Table 9. Total hours the diesel operates in one month for each sizing solutions] It can be seen that the sizing number of 4, 1, and 2 for PV module, WT and battery gives the maximum diesel generator operating hours in lowest ACS and number of 6, 2, and 3 for them, gives the minimum one in highest ACS. It depends to the designer to select one of the obtained optimal sizing numbers for the hybrid components, by considering the fuel cost, the necessity of match rate, and the cost consideration

9. Conclusion

This study has been dedicated to determining the optimum stand-alone hybrid wind/PV power generating systems. A new sizing methodology has been developed based on the match evaluation method (MEM), considering resource uncertainties associated with wind speed, solar irradiation and load demand. A diesel generator and a battery is used to even out the irregularities. A management strategy has been designed to achieve higher match rate between supply and demand intervals. The optimization process provides optimum capacity of as many numbers of supplies as required to match with a load demand in lowest investment, so it can handle large scale design problems. This sizing methodology is useful for better energy utilization, eliminating exhaustive search and avoid excessive computation times. This work is undertaken with triple objective function: inequality coefficient (IC), correlation coefficient (CC), and annualized cost of system (ACS). Studies have been done on different configurations of stand-alone hybrid wind/PV systems. Six different wind turbines (WTs) and also six different PV modules, with different characteristics have been considered. Sizing parameters have been determined by the multi-objective particle swarm optimization (MOPSO) algorithm. The results are also validated by NSGA-II. Obtained results suggest the choice of Kyocera Solar (KC200) PV module and Bornay (Inclin 3000) wind turbine for the

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Civil Engineering and Environmental Systems

Page 30 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

proposed hybrid power system, among all configurations. The algorithm has been run for the best configuration. Obtained solutions are non-dominated and they form the Pareto front. Simulation results show that a configuration with 6 PVs, 2 wind turbines, and 3 battery units has a high electricity match rate (EMR) and lower operating hours for the diesel generator with the expense of high ACS. Another configuration with 4 PV modules, 1 wind turbine, and 2 battery units has long diesel operating hours in acceptable match rate, and its ACS is the lowest. The designers can select the best configuration among the Pareto set which fits their desire. It is worth mentioning that the proposed methodology can be effectively employed for any composition of hybrid energy systems in any locations taking account of the meteorological data and the

consumer’s demand.

References

[1] Brian Tarroja, Fabian Mueller, Joshua D.Eichman, and Scott Samuelsen, 2012, "Metrics for evaluating the impacts of intermittent renewable generation on utility loadbalancing," Energy, Elsevier, p. 546 e 562.

[2] S.Abedi, A.Alimardani, G.B.Gharehpetian, G.H.Riahy, and S.H.Hosseinian, 2012, "A comprehensive method for optimal power management and design of hybrid RES-based autonomous energy systems," enewable and Sustainable Energy Reviews, Elsevier, pp.

1577–1587.

[3] Vafaei Mehdi, 2011,"Optimally-Sized Design of a Wind/Diesel/Fuel Cell Hybrid System fo Remote Community " Master of Applied Science Electrical and Computer

Engineering University of Waterloo.

[4] Hongxing Yang, Wei Zhou, Lin Lu a, and Z.Fang, 2008, "Optimal sizing method for stand-alone hybrid solar–wind system with LPSP technology by using genetic algorithm," Solar Energy, ELSEVIER, pp. 354–367. [5] D.B.Nelson, M.H.Nehrir, and C.Wang, 2006, "Unit sizing and cost analysis of standalone hybrid wind/PV/fuel cell power generation systems," Renewable Energy, ELSEVIER, pp. 1641 – 1656.

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Page 31 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

[6] Wei Zhou, 2007, "Simulation and Optimum Design of Hybrid Solar-Wind and SolarWind-Diesel Power Generation Systems,," Doctor of Philosophy, The Hong Kong Polytechnic University. [7] Daming Xu, Longyun Kang, and Binggang Cao, 2006, "The Elitist Non-dominated Sorting GA for Multi-objective Optimization of Standalone Hybrid Wind/PV Power Systems," Applied Sciences. [8]Adel A. Elbaset, July 2011, "Design, Modeling and Control Strategy of PV/FC Hybrid Power System," Electrical Systems, pp. 270-286. [9]Yahia Bakelli, Amar Hadj Arab, and Boubekeur Azoui, 2011, "Optimal sizing of photovoltaic pumping system with water tank storage using LPSP concept," Solar Energy, pp. 288-294.

[10] S.Diaf, G.Notton, M.Belhamel, M.Haddadi, and A.Louche, 2008, "Design and technoeconomical optimization for hybrid PV/wind system under various meteorological conditions," Applied Energy.

[11] Bilala B. O, V. Samboua, C.M.F Kébé, P.A.Ndiaye, and M.Ndongo, 2012, "Methodology to Size an Optimal Stand-Alone PV/wind/diesel/battery System Minimizing the Levelized cost of Energy and the CO2 Emissions " Energy Procedia,

ELSEVIER, pp. 1636 – 1647.

[12] Hongxing Yang, Lin Lu, and Wei Zaou, 2007, "A novel optimization sizing model for hybrid solar-wind power generation system," Solar Energy, ELSEVIER. [13] Rodolfo Dufo-Lopez and Jose´ L.Bernal-Agustin, 2008, "Multi-objective design of PV– wind– diesel– hydrogen– battery systems," Renewable Energy, ELSEVIER, pp. 2559– 2572.

[14] R.Luna-Rubio, M.Trejo-Perea, D.Vargas-Va´zquez, and G.J.Rı´os-Moreno, 2012, "Optimal sizing of renewable hybrids energy systems: A review of methodologies," Solar Energy, ELSEVIER, pp. 1077–1088.

[15] Mohammad Ali Yazdanpanah Jahromi, Said Farahat, and S. M. Barakati, 2012, "A Novel Sizing Methodology Based on Match Evaluation Method for Optimal Sizing of Stand-Alone Hybrid Energy Systems Using NSGA-II," The Journal of Mathematics and Computer Science, vol. 5, pp. 134-145. [16] Y. S. Zhao, J. Zhan, Y. Zhang, D. P. Wang, and B. G. Zou, 2006, "The Optimal Capacity Configuration of an Independent Wind/PV Hybrid Power Supply System Based On Improved PSO Algorithm", Shandong Electric Power Research Institute, Jinan,China.

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Civil Engineering and Environmental Systems

Page 32 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

[17] Dhillon Javed, 2009, "Multi-objective Optimization Of power Dispatch Problem Using NSGA-II," Master of Engineering Power Systems & Electric Drives, Thapar University, Patiala. [18] Eftichios Koutroulis, Dionissia Kolokotsa, Antonis Potirakis, and Kostas Kalaitzakis, 2006, "Methodology for optimal sizing of stand-alone photovoltaic/wind-generator systems using genetic algorithms," Solar Energy, ELSEVIER, pp. 1072–1088. [19] O.Erdinc and M.Uzunoglu, 2012, "Optimum design of hybrid renewable energy systems: Overview of different approaches," Renewable and Sustainable Energy Reviews,ELSEVIER, pp. 1412– 1425. [20] AEMO, 2010, VICTORIAN ANNUAL PLANNING REPORT: Australian Market Operator Limited (AEMO). [21] Neville Henderson, 2010, Report to reliability panel on demand forecast,: AEMO,. [22] Chin-Yuan Hsieh and Chen-Yu Hsieh, 2012, "Climate Change Prediction by Wireless Sensor Technology " International Electrical Engineering Journal (IEEJ),, pp. 620624. Energy

[23] Mohammadi. M, S.H.Hosseinian, and G.B.Gharehpetian, 2012, "GA-based optimal sizing of microgrid and DG units under pool and hybrid electricity markets," Electrical Power and Energy Systems, Elsevier, pp. 83–92.

[24] Eduardo Ivan Ortiz Rivera, 2006, "Modeling and analysis of solar distributed generation," Ph.D.Dissertation, Department of Electrical and Computer Engineering, Michigan State University.

[25] Miguel Rios Rivera, 2008, "Small Wind/Photovoltaic Hybrid Renewable Energy System Optimization," Master of Science, Electrical Engineering, Puerto Rico,

Mayagüez Campus.

[26] Bogdan S.Borowy and Ziyad M.Salameh, 1996, "Methodology for Optimally Sizing the Combination of a Battery Bank and PV Array in a Wind/PV Hybrid System," IEEE

Trans. Energy Conversion, vol. 11, pp. 367-375.

[27] Mukund R.Patel, 2006, Wind and Solar Power Systems:Design, Analysis, and Operation, Second ed.: Taylor & Francis Group. [28] RERScreen International Clean Energy Decision Support Centre, 2004, Textbook, "Wind Energy Project Analysis," ed. Canada, Minister of Natural Resources Canada, p. 168.

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Page 33 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

[29] S. H. Jangamshetti and V. Guruprasada Rau, December 1999, "Site Matching of Wind Turbine Generators: A Case Study," IEEE Transactions on Energy Conversion, vol. 14, pp. 1537-1543. [30] Carlos Antonio Ramos Robles, 2005, "Determination Of Favorable Conditions For The Developement Of a Wind Power Farm In PUERTO RICO," Master Of Science, Electrical Engineering, PUERTO RICO. [31] Wei. Z, 2007, "Simulation and Optimum Design of Hybrid Solar-Wind and SolarWind-Diesel Power Generation Systems," Doctor of Philosophy, The Hong Kong Polytechnic University.

[32] Yang. H, Zhou. W, Lua. L, and Fang. Z, 2008, "Optimal sizing method for standalone hybrid solar–wind system with LPSP technology by using genetic algorithm", Solar Energy, pp. 354–367.

[33] E. T. b. Q. A. Certifications, 2013, Industrial power Battery, Powerbatt, Available: http://www.polluxbattery.com.my/

[34] Scheaffer, Mulekar, and MvClave, 2011, Probability and Statistics for Engineers. [35] Born Francesca Jane, 2001, "Aiding Renewable Energy Integration through Complimentary Demand-Supply Matching," Doctor of Philosophy, Energy Systems Research Unit, University of Strathclyde.

[36] Waqas Sana, 2011, "Development of an Optimisation Algorithm for Auto sizing

Capacity of Renewable and Low Carbon

Department of Mechanical Engineering,, University of Strathclyde Engineering. [37] Heri Suryoatmojo, 2009, Takashi Hiyama, Adel A. Elbaset, and Mochamad Ashari, "Optimal Design of Wind-PV-Diesel-Battery System using Genetic Algorithm". [38] Singiresu S.Rao, 2009, Engineering Optimization Theory and Practice.

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

Energy Systems," Master of Science,

w

On ly

Civil Engineering and Environmental Systems

Page 34 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

All Figures
1.2

1

Insolation (kW/m2)

0.8

Ambient Temperature ( o C)

rP Fo
0.6 0.4 0.2

ee
100

0 0

200 300 400 500 600 Time (Number of Hours In August)

700

800

Figure 1. Meteorological conditions of solar radiation in August

rR

ev ie w On ly
100 200 300 400 500 600 Time (Number of Hours In August) 700 800

40

35

30

25

20

15 0

Figure 2. Ambient temperature in August

URL: http:/mc.manuscriptcentral.com/gcee

Page 35 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

200 Power (W) 150 100 50 0 0 10 20 30 Voltage (V) 40 50 Type1 Type2 Type3 Type4 Type5 Type6 60

Wind Speed (m/s)

rP Fo
8 6 4 2 0

Type1 Type2 Type3 Type4 Type5 Type6 20 30 Voltage (V) 40 50 60

Current (A)

ee

0

10

Figure 3. P-V and I-V Curves

rR

30

ev ie w On ly
100 200 300 400 500 600 Time (Number of Hours In August) 700 800

25

20

15

10

5

0 0

Figure 4. Meteorological conditions of wind speed in August

URL: http:/mc.manuscriptcentral.com/gcee

Civil Engineering and Environmental Systems

Page 36 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

0.12

0.1

Weibull PDF (f(v))

0.08

0.06

rP Fo
0.04 0.02

ee
5
5

0 0

10 15 Wind Speed (m/s)

20

25

Figure 5. Weibull probability density function (f(v))

rR

ev

12
Type1 Type2

10

Type3 Type4 Type5

ie

Power Output (kW)

8

Type6

w

6

On

4

2

ly

0 0

10 Wind Speed (m/s)

15

20

25

Figure 6. Wind turbine power curves (The symbols represent data sampled from the power curve graphs given by the manufacturer)

URL: http:/mc.manuscriptcentral.com/gcee

Page 37 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

800 700 600 Energy (kWh/Month) 500 400 300 200 100
Type1 Type2 Type3 Type4 Type5 Type6

0 0

12000

10000

rP Fo
5
8000 6000 4000 2000 0 0 5

10 Wind Speed (m/s)

15

20

25

Figure 7. Total wind turbine energy outputs by wind speed in August

ee

rR ev ie w On
10 Wind Speed (m/s) 15 20

Type1 Type2 Type3 Type4 Type5 Type6

Energy (kWh/Year)

25

Figure 8. Total wind turbine energy outputs by wind speed for one year

ly

URL: http:/mc.manuscriptcentral.com/gcee

Civil Engineering and Environmental Systems

Page 38 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

0.9

0.8 Load (kW)

rP Fo
0.7 0.6

ee rR
100 200 300 400 500 600 Time (Number of Hours In August ) 700 800

0.5

0.4 0

Figure 9. Monthly (August) variation of domestic load profile

URL: http:/mc.manuscriptcentral.com/gcee

ev

ie w On ly

Page 39 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

rP Fo
Figure 10. Flow chart of optimization process using MOPSO (or NSGA-II)
URL: http:/mc.manuscriptcentral.com/gcee

ee rR ev ie w On ly

Civil Engineering and Environmental Systems

Page 40 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

rP Fo
Figure 11. Operation strategy of proposed hybrid System
URL: http:/mc.manuscriptcentral.com/gcee

ee rR ev ie w On ly

Page 41 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

-0.911

-0.9115 Optimum Direction of Correlation Coefficient (CC)

-0.912

Figure 12. 2D Pareto front for the last generation. Inequality coefficient (IC) vs. correlation coefficient (CC).

rP Fo
-0.9125 -0.913 -0.9135 0.128 0.1285 0.129 0.1295
URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

0.13

0.1305 0.131 0.1315 0.132

Optimum Direction of Inequality Coefficient (IC)

ev

ie w On ly

s

Civil Engineering and Environmental Systems

Page 42 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

4000

3800 Optimum Direction of Annualized Cost of System (ACS)

3600

3400

Figure 13. 2D Pareto front for the last generation. Inequality coefficient (IC) vs. annualized cost of system (ACS).

rP Fo
3200 3000

ee
0.14

2800 0.13

0.135

0.145

0.15

0.155

0.16

0.165

0.17

0.175

Optimum Direction of Inequality Coefficient (IC)

URL: http:/mc.manuscriptcentral.com/gcee

rR

ev

ie w On ly

Page 43 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

3700 3600 Optimum Direction of Annualized Cost of System (ACS) 3500 3400 3300

Optimum Direction of Correlation Coefficient (CC) Figure 14. 2D Pareto front for the last generation. Correlation coefficient (CC) vs. annualized cost of system (ACS).

rP Fo
3200 3100 3000 2900 2800 2700 -0.94 -0.92
URL: http:/mc.manuscriptcentral.com/gcee

ee rR
-0.9 -0.88 -0.86 -0.84 -0.82 -0.8

ev

ie

w On ly

Civil Engineering and Environmental Systems

Page 44 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Optimum Direction of Annualized Cost of System (ACS)

4200 4000 3800 3600 3400 3200 3000 2800

Optimum Direction of Correlation Coefficient (CC)

rP Fo
-0.8

0.35 0.3 -0.9 0.2 -1 0.25 Optimum Direction of Inequality Coefficient (IC)

Figure 15. 3D Pareto front for the best configuration

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev ie w On ly

Page 45 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Table 1. Solar module power at standard test condition rating and price PV Panels Parameters Type of PV PV MSRP Watt at 1000 Product module (US$/Unit) W/m2 1 Kyocera Solar (KC200) 800.00 200 2 3 4 5 6 BP Solar (SX 170B) Evergreen (Spruce ES-170) Evergreen (Spruce ES-180) Evergreen (Spruce ES-190) Solar World (SW-165) 728.97 731.00 774.00 817.00 709.97 170 170 180 190 165

efficiency 0.20 0.17 0.17 0.18 0.19 0.17

Table 2. Small wind turbines rating and price Small Wind Turbine Parameters Type of wind Turbine MSRP Tower price Product turbine (US$/Unit) (US$/Unit) 1 Solacity (Eoltec) 25,200 1,968 2 3 4 5 6 Kestrel Wind (Kestrel 3000) Bornay (Inclin 6000) Bornay (Inclin 3000) Bergey (BWC Excel-R) 8,400 36,000 10,070 6,028 23,000 1,968 2,396 1,968 1,968 2,396 Abundant Renewable Energy (ARE442)

Table 3. Solar module specifications at standard test condition rating Product Nominal Nominal Short-circuit Open-circuit voltage at voltage current current at [V] STC [V] [ A] STC [A] (Voc) (ISC) KC200 26.3 7.6 8.2 32.9 SX170B 35.4 4.8 5.0 43.6 Spruce ES-190 25.3 6.7 7.6 32.4 Spruce ES-180 25.9 6.9 7.8 32.6 Spruce ES-190 26.7 7.1 8.1 32.8 SW-165 35.3 4.6 5.1 43.9

Table 4. Detailed specifications of the wind turbines Product Rated power Cut-in wind Cut-out wind capacity (W) speed (m/s) speed (m/s) Eoltec 6,000 2.7 None Kestrel 3000 3,000 2.8 None ARE442 10,000 2.5 11.2 Inclin 6000 6,000 3.5 15.0 Inclin 3000 3,000 3.5 14.0 BWC Excel-R 7,500 3.1 None

rP Fo
12 60

Watt at 12.5 (m/s) 6000 3000 10000 6000 3000 8100

Table 5. Battery characteristic Volt Rated Capacity (Ah)

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

Temperature coefficient of Isc [%/°C] (TCi) 0.003 0.065 0.060 0.060 0.060 0.060

Temperature coefficient of Voc [%/°C] (TCV) -0.123 -0.160 -0.340 -0.340 -0.340 -0.350

ev
Width 0.166

Dimensions (m) Length 0.258 Height 0.206

ie

w

On
Rated wind speed (m/s) 11.5 11.0 11.2 12.0 12.0 13.8

Seller Solacity Kestrel ARE Bornay Bornay AltE Store

ly
22.40

Weight (kg)

Civil Engineering and Environmental Systems

Page 46 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Table 6. The cost and lifetime aspect for the proposed hybrid components Components Initial capital cost Maintenance cost in the (US$/kW) first year (US$/kW) Wind turbines and (Table 2) 95.00 their towers PV modules (Table 1) 65.00 Battery 124.12 25.00 Diesel 500.00 1000.00 Other components 900.00 90.00

Replacement cost (US$/kAh) Null Null 124.12 Null Null

Life time (year) 20.00 20.00 5.00 20.00 20.00

Table 7. Cost parameters Interest rate, Inflation rate, f iloan(%) (%) 5.000000 2.000000

Carbon tax (US$/kWh) 0.010485

Fuel cost (US$/kWh) 0.160490

rP Fo

Table 8. Pareto front/optimal solutions obtained from multi-objective optimization Solution Optimization Algorithm Made in MOPSO and NSGA-II NPV NWIND NBattery IC CC ACS 1 4 2 2 0.2106 0.8804 3810.56 2 4 1 2 0.2399 0.8596 3134.86 3 4 2 3 0.2102 0.8814 3878.82 4 5 1 2 0.2180 0.8892 3284.92 5 5 2 2 0.1947 0.9080 3960.63 6 5 2 3 0.1942 0.9096 4028.89 7 6 2 2 0.1857 0.9160 4110.69 8 6 2 3 0.1850 0.9171 4178.95 9 6 1 2 0.1998 0.9181 3434.98 10 7 1 2 0.1875 0.9270 3585.06 11 7 1 3 0.1860 0.9286 3653.32

Table 9. Total hours the diesel operates in one month for each sizing solutions Diesel Operating Hours For One Month Solution NPV NWIND NBattery (Total Month Hours In August = 744) 1 4 2 2 378 2 4 1 2 421 3 4 2 3 367 4 5 1 2 395 5 5 2 2 342 6 5 2 3 334 7 6 2 2 301 8 6 2 3 299 9 6 1 2 363 10 7 1 2 327 11 7 1 3 321

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR

ev

ie

w

On

ly

Page 47 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

rP Fo
Meteorological conditions of solar radiation in August 176x155mm (96 x 96 DPI)

URL: http:/mc.manuscriptcentral.com/gcee

ee rR ev ie w On ly

Civil Engineering and Environmental Systems

Page 48 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

rP Fo

P-V and I-V Curves 361x166mm (96 x 96 DPI)

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR ev ie w On ly

Page 49 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

rP Fo
Meteorological conditions of wind speed in August 176x161mm (96 x 96 DPI)

URL: http:/mc.manuscriptcentral.com/gcee

ee rR ev ie w On ly

Civil Engineering and Environmental Systems

Page 50 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

rP Fo
Meteorological conditions of wind speed in August 176x161mm (96 x 96 DPI)

URL: http:/mc.manuscriptcentral.com/gcee

ee rR ev ie w On ly

Page 51 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Wind turbine power curves (The symbols represent data sampled from the power curve graphs given by the manufacturer) 361x159mm (96 x 96 DPI)

rP Fo

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR ev ie w On ly

Civil Engineering and Environmental Systems

Page 52 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

rP Fo

Total wind turbine energy outputs by wind speed in August 361x159mm (96 x 96 DPI)

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR ev ie w On ly

Page 53 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

rP Fo

Total wind turbine energy outputs by wind speed for one year 361x159mm (96 x 96 DPI)

URL: http:/mc.manuscriptcentral.com/gcee

ee

rR ev ie w On ly

Civil Engineering and Environmental Systems

Page 54 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

rP Fo
Monthly (August) variation of domestic load profile 176x161mm (96 x 96 DPI)

URL: http:/mc.manuscriptcentral.com/gcee

ee rR ev ie w On ly

Page 55 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

rP Fo
Flow chart of optimization process using MOPSO (or NSGA-II) 136x194mm (180 x 180 DPI)

URL: http:/mc.manuscriptcentral.com/gcee

ee rR ev ie w On ly

Civil Engineering and Environmental Systems

Page 56 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The link ed image cannot be display ed. The file may hav e been mov ed, renamed, or deleted. Verify that the link points to the correct file and location.

rP Fo
Operation strategy of proposed hybrid System 151x174mm (180 x 180 DPI)

URL: http:/mc.manuscriptcentral.com/gcee

ee rR ev ie w On ly

Page 57 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2D Pareto front for the last generation. Inequality coefficient (IC) vs. correlation coefficient (CC). 187x155mm (96 x 96 DPI)

rP Fo
URL: http:/mc.manuscriptcentral.com/gcee

ee rR ev ie w On ly

Civil Engineering and Environmental Systems

Page 58 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2D Pareto front for the last generation. Inequality coefficient (IC) vs. annualized cost of system (ACS). 208x155mm (96 x 96 DPI)

rP Fo
URL: http:/mc.manuscriptcentral.com/gcee

ee rR ev ie w On ly

Page 59 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2D Pareto front for the last generation. Correlation coefficient (CC) vs. annualized cost of system (ACS). 176x155mm (96 x 96 DPI)

rP Fo
URL: http:/mc.manuscriptcentral.com/gcee

ee rR ev ie w On ly

Civil Engineering and Environmental Systems

Page 60 of 61

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

rP Fo
URL: http:/mc.manuscriptcentral.com/gcee

ee
3D Pareto front for the best configuration 211x155mm (96 x 96 DPI)

rR ev ie w On ly

Page 61 of 61

Civil Engineering and Environmental Systems

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

rP Fo
URL: http:/mc.manuscriptcentral.com/gcee

ee
3D Pareto front for the best configuration 211x155mm (96 x 96 DPI)

rR ev ie w On ly