12th Joint European Thermodynamis Conferene Bresia, July 1-5, 2013 EXERGY BASED METHODS FOR ECONOMIC AND ENVIRONMENTAL ANALYSIS APPLIED TO A 320 MW COMBINED CYCLE POWER PLANT Matteo Roo*, Claudia Toro *Politenio di Milano, Department of Energy, via Lambrushini 4, Milan, Italy Università di Roma Sapienza, Department of Mehanial and Aerospae Engineering, via Eudossiana 18, Rome, Italy ABSTRACT In the last deades, the growing sarity of non-renewable resoures led analysts and researhers to sharpen Seond Law analysis methods in order to understand how to minimize the onsumption of natural resoures on the part of energy onversion systems. Thermoeonomis demonstrates to be a proper and promising framework to evaluate and optimize exergeti and eonomi osts of energy systems produts. Understanding the relation between the eonomi ost and its natural resoure ounterpart is likely to be a key fator in future researh ativities. This paper presents an Exergy and Thermoeonomi analysis of a 320 MW Dual Pressure Combined Cyle Plant, aimed to identify the optimal design onfigurations of the system with respet to its speifi objetive funtions: seond law effiieny, eonomi ost and natural resoure onsumption ost of the generated unit of eletri energy. The natural resoure onsumption of the system is omputed aording to the Cumulative Exergy Consumption (CExC) method. The CCPP plant simulations have been performed by using CAMEL-Pro Proess Simulator and the sensitivity study of the plant behaviour and its optimization as a funtion of the seleted parameters have been developed by using the Proper Orthogonal Deomposition proedure. Our results onfirm that the optimal design onfiguration is stritly dependent on the onsidered objetive funtion, and helps to investigate the relationship between the thermodynamis, the eonomis and the resoure onsumption of the system, thus giving a more omprehensive understanding of its performane from different perspetives. INTRODUCTION Combined Cyle Power Plants (CCPPs) are well proven and reliable tehnology for eletriity prodution. CCPPs are widely used in the Italian grid network and their design and optimization are today more and more relevant. This paper presents an original implementation of the Thermoeonomis framework for the optimal design analysis of a 320 MW Dual Pressure CCPP. The purpose of the study is to identify the possible optimal design onfiguration of this system, inluding thermodynami, eonomi and environmental perspetives. As a first step, modelling and simulation of the system is performed. Seondly, eonomi and environmental perspetives are assessed performing exergy based speifi analyses. Exergy analysis (EA) is used to determine the seond law effiieny of the system, whereas Thermoeonomi framework is used to assess both the eonomi and the environmental osts of the produt. In the ase study eletriity is onsidered as the unique produt of the system. Eonomi optimal ost of produt is assessed with the Thermoeonomi analysis (TA-ECO), finding the best trade-off between investment and operative eonomi osts [1]. On the other hand, Thermoeonomi analysis is also used to assess, in an environmental ost perspetive, the primary exergy onsumed in order to produe the system produt (TA-EXER). The primary exergy may represent the natural resoures onsumed [2]. As will be shown, the optimal design onfiguration is stritly dependent on the onsidered objetive ost funtion. The paper shows how hanging the objetive funtion of the analysis (effiieny, monetary ost or primary exergy ost) an influene the optimal design of the system, and proposes a key to understanding the relationship between eonomi and environmental osts of energy systems. CASE STUDY: DUAL PRESSURE COMBINED CYCLE POWER PLANT Plant layout and simulation As a ase study, the Neka CCPP power plant operative data have been used [3]. The main omponents of this ombined plant are two gas turbines, two air ompressors, two HRSGs with a supplementary fired unit (dut burner), one steam turbine and one surfae ondenser with a seawater ooling system. The total output power is 320 MW, 130MW produed by the steam turbine and 190 MW two gas turbines. Table 1. CCPP fixed operative parameters. Parameter u. m. Value Outlet Power MW 160 Gas Turbine Adiabati Effiieny % 87.7 Compressor Adiabati Effiieny % 88.0 Steam Turbine Effiieny % 78.0 Condenser inlet pressure kpa 14 HRSG Low Pressure kpa 1029 HRSG High Pressure kpa 11425 Turbine Inlet Temperature K 1383 Plant Availability h/y 2628 464 1
The power plant was modeled with CAMEL-Pro proess simulator [4]; its layout onsist in two idential bloks, eah generating half of the total output: all simulations were performed aordingly for one single blok with an output net power of 160 MW. The layout of the simulated plant is reported in Figure 1 and the operative onditions (steady state operation being assumed throughout) are reported in Table 1. The low Load Fator (2628 hours per year) reflets the atual operative onditions of the average Italian CCPPs. Table 2. Seleted design variables and their respetive ranges. Proess variables Symbol [u. m.] Min. Max. value value Air pressure ratio β C [-] 10 21 Post-Firing fuel ṁ 14 [kg/s] 1 2 Figure 1. Plant layout. Table 3. Constraints for plant operation. Parameter [u. m.] Devie / Lower Higher Pipe no. limit limit T pp,lp [ C] 16 5 35 T pp,hp [ C] 18 5 35 T ap,lp [ C] 16,30 10 - T ap,hp [ C] 17,18,12 10 - T 40 [ C] 40 100 - EXERGY BASED METHODS ANALYSES AND OPTIMIZATION Thermodynami evaluation: exergy analysis To proeed with the optimization of the plant, two proess variables have been seleted: the gas turbine pressure ratio β C and the dut burner fuel mass flow rate ṁ NG,pf (ṁ 14 ) Their respetive possible ranges of variation are reported in Table 2. Determination of the plant behavior: Proper Orthogonal Deomposition method (POD) The sensitivity study of the plant behavior and the optimization with respet to the seleted proess variables and the objetive funtions have been developed by using the Proper Orthogonal Deomposition mathematial proedure (POD RBF). The POD Method is a is a statistial method that aims to provide a ompat representation of the data by projeting the data set into a lower dimensional spae. The POD-RBF proedure has been previously tested on the optimization of a simple MSF desalination plant [5] and of a single pressure CCGT [6] plant and the very satisfatory results obtained for these plants suggested extending its appliation to more omplex onfigurations and to different proesses. Moreover, the POD enables designers to extrapolate funtions linking the variables to be optimized with the seleted proess variables. More details about POD as well as an introdutory mathematial explanation of its oneptual basis are provided in [7]. In the ase study here presented the objetive funtions are minimized onsidering the onstraints listed in Table 3. As stated in [2], in order to perform the exergy analysis for a generi energy system, it is onvenient to set up its produtive struture, or funtional diagram. Using the physial model of the system as referene and grouping all the energy and material flows for every omponent of the system, and therefore for the whole system, the produtive struture is ompleted aording to the Resoures Produt Wastes (R/P/I) riterion. For the generi j-th system omponent, exergy balane is: E & = E & + E & + E & (1) R, j P, j I, j D, j For every system onsisting of n omponents onneted by m streams, the exergy balane system an be expressed in matrix form by (2), where A is the n m inidene matrix of the system, defined in [2]: A (n m ) E(m 1) = E D,(n 1) (2) For eah omponent of the system, exergy effiieny is defined as the ratio between exergy of produts over exergy of the resoures and it represents a riterion for evaluating the thermodynami performane of the omponent. The objetive of the exergy analysis is to find the ombination of the seleted proess variables β C and ṁ 14 that provides the highest exergy effiieny for the whole system, defined as (3). 465
E& P,tot η II, tot = = E& F,tot & W el,net (m& + m& ) e 6 14 This goal has been reahed applying the POD proedure on simulated plant results, whih allows extrapolating the exergy effiieny (3) as a funtion of the seleted proess variables β C and ṁ 14. Figure 2 shows the exergy effiieny as funtion of these proess variables, normalized within their own range. NG (3) other m-n auxiliary monetary osts equations [1]: some of them depends on the adopted branhings and ost alloation riteria [2], whereas the others are defined by the boundary onditions, suh as market pries. In the ase study the following auxiliary equations were adopted: = = (7) eo,6 eo,14 eo, NG eo,3 (8) eo,40 (9) The speifi ost of the natural gas (7) was omputed on the base of the Italian market average prie, and it was onsidered onstant for the entire lifetime of the plant. With the auxiliary equation (9) it omes out that all the eonomi osts of exiting flue gases are harged to the HRSG, thus on the ost of the produt. Other standard assumptions have been made to distribute osts among internal streams [8]. Figure 2. Overall plant exergy effiieny map. For the plant given in the ase study, the best exergy effiieny results around 44.8 %, orresponding to β C = 14.36 and ṁ 14 = 1.6 kg/s. Thermoeonomi design analysis for Eonomi ost evaluation and optimization (TE-ECO) Aording to produtive struture adopted for exergy analysis, the eonomi ost rate balanes for the i-th plant omponent is: C& + Z = C& + C& & (4) eo, R, i eo, i eo, P, i eo, I, i Table 4. Eonomi parameters. Parameter Symbol [u. m.] Max. value Interest rate i [%] 5 Plant lifetime t [years] 30 Natural gas ost eo,ng [ /Nm 3 ] 0.35 The main parameters of the eonomi analysis are reported in Table 3. To alulate equipment osts as a funtion of the main plant operation parameters, the Frangopoulos apital osting equations have been used: values obtained with these equations ould be onsidered aeptable approximations of real values whih usually are not given by industry as a funtion of omponents parameters [9]. The main result of the thermoeonomi plant optimization, obtained by the appliation of the POD-RBF proedure, is the ombination of the proess variables whih lead to the attainment of the most onvenient ompromise between plant effiieny and eonomi osts. Where Ċ eo,i represent the eonomi ost rate assoiated with eah exergy transfer, and Ż eo,i represents the sum of apital investment, operating and maintenane ost rates for the i-th system omponent. Exergy osting priniple (5) allows to ompute the eonomi ost rate of every j-th material or energy flow entering the i-th system omponent as the produt between its average monetary ost per unit of exergy eo,j (in /kj) and its exergy ontent: C & = E & (5) eo, j eo, j j The omplete thermoeonomi system an be rewritten in matrix form as follows: A (n m) Ceo,(m 1) + Zeo,(n 1) (n 1) (6) If the onsidered system has n omponents and m streams, C eo is the m 1 eonomi ost rates vetor and Z eo is the n 1 investment ost rates vetor of system omponents. In order to lose the equations system, it is therefore neessary to write Figure 3. Eonomi ost map of the system produt. Figure 3 shows the eonomi ost map of eletriity as a funtion of the normalized seleted proess variables β C and ṁ 14. For this plant, the minimum ost results 5.90 /s, orresponding to values of β C = 13.32 and ṁ 14 = 1.84 kg/s. 466
Thermoeonomi design analysis for Environmental ost evaluation and optimization (TA-EXER) Aording to literature [10], the impat of energy systems on environment is mainly related to: natural resoure onsumption in the whole life yle of the system (a) and polluting effet of all the waste emissions in water, atmosphere and soil (b). Exergy is widely aepted as a ommon measure of natural resoures onsumption and it an be therefore used as an indiator for the environmental impat [11, 12]. Indeed, several attempts have been made to ombine exergy analysis and Life Cyle Assessment (LCA) to quantify the natural resoures onsumption (a) of industrial proesses [13, 14]: suh as Cumulative Exergy Consumption (CExC) [15], Thermo - Eologial Cost (TEC) [16], Exergeti Life Cyle Analysis (ELCA) [17], Cumulative Exergy Extration From Natural Environment (CEENE) [18] and so on. All these indiators are development of the embodied energy paradigm, well explained in [19]: they differ from eah other in the definition of the resoure ost fators inluded into the aounting and in the analysis time window. On the other hand, one of the main weakness of exergy analysis is that the exergy of waste emissions hardly reflets the magnitudes of the environmental impat [20]. For this reason, some authors propose to evaluate the waste emissions polluting effet (b) as the additional natural resoure onsumption needed to redue the exergy ontent of the effluents to zero: CExCT [21], Zero-ELCA [22] and Extended Exergy Aounting [14] are examples of this approah. In this paper the authors propose the adoption of the Cumulative Exergy Consumption (CExC) indiator for evaluating the natural resoure onsumption of the system as partial evaluation of the environmental ost of energy systems. This approah aounts only for energy and materials resoures (renewables and non-renewables), as well as human labour, involved in the prodution of a unit of energy or material produts [16]; it does not inludes other externalities. The thermoeonomi ost rate balane (4) for eah i-th plant omponent is rewritten aording to the same produtive struture and input data adopted in 0, replaing eonomi osts rates of Resoures, Produts, Wastes and Plant omponents with their respetive exergeti osts: C& + Z = C& + C& & (10) ex, R, i ex, i ex, P, i ex, I, i Where Ċ ex,i represent the resoure ost rate embodied in eah exergy transfer, and Ż ex,i represents the resoure ost embodied in the i-th system omponent. Exergy osting priniple (11) allows to ompute the resoure ost rate for every j-th material or energy flow entering the i-th system omponent as the produt between its exergy ontent and its CExC (represented here by ex,j, in kj/kj): C & = E & (11) ex, j ex, j j In a dual way of paragraph 0, the omplete thermoeonomi system an be rewritten in matrix form (12). A (n m) Cex,(m 1) + Zex,(n 1) (n 1) (12) Where C ex is the m 1 thermoeonomi osts vetor and Z ex is the n 1 investment ost rates vetor of system omponents. Here, the same rules for branhings and ost alloations adopted in 0 were used, and auxiliary relations neessary to lose the equation system were omputed relying to Simapro 7.3.3 software [23] and Szargut database [15]. = = (13) ex,6 ex,14 ex, NG ex,3 (14) ex,40 (15) CExC of the Eoinvent unit proess Natural gas, high pressure, at onsumer/rer U is 1.069 MJ/MJ and was adopted as speifi exergy ost of the natural gas (13). Like TE-ECO analysis, with the auxiliary equation (15) omes out that all the exergeti osts are harged to the HRSG, thus on the ost of the produt. Exergy ost funtions for plant and O&M osts (Ż ex,i ) have been extrapolated from Eoinvent database as a funtion of the size, weight and the operative parameters of the plant omponents [24]. In ase of data sarity, average value for primary exergy onsumption of European Mahinery and Equipment prodution setor have been extrapolated from European Input-Output tables (year 2003) and result to be 50.21 MJ/kg [25]. Like the investment eonomi ost funtions, the exergy ost funtions ould be onsidered aeptable approximations of real values. Exergy osts of system produt were obtained as a funtion of the seleted proess variables β C and ṁ 14 by applying the POD-RBF proedure. Figure 4. Exergeti ost map of the system produt. Values in x and y axes are normalized between 0 and 1. Figure 4 shows the exergeti ost map of eletriity as a funtion of normalized proess variables β C and ṁ 14. For this plant, the minimum resoures onsumption results to be 434.5 MJ/s, orresponding to values of β C = 14.29 and ṁ 14 = 1.75 kg/s. It is noteworthy that, starting from this thermoeonomi analysis, an extension of the traditional seond law effiieny an be introdued: η CExC, tot = C& E & ex, R,tot P,tot + Z& ex, tot (16) Sine it takes into aount also for the prodution proesses of the fuel and the plant equipment, effiieny defined in (16) is an extended insight of the overall energy onversion proess. For this plant, best CExC-effiieny is about 0.36, obtained in 467
orrespondene with the minimum resoure onsumption as previously desribed. RESULTS AND DISCUSSION Exergy based methods results As previously explained, three different exergy based optimization riteria have been applied to the speifi ase study. For eah identified optimal onfiguration, it is therefore possible to alulate all the orresponding proess variables, summarized in Table 5. As expeted, the three ouples of β C and ṁ 14 variables that define eah optimal onfiguration are different; therefore, also the other operative parameters suh as effiieny, exergeti and eonomi osts differ. Considering a plant availability of 2628 hours per year, the minimization of the eonomi ost of eletriity (Ċ eo,opt ) allows to redue the annual ost of eletriity by 336 k /year with respet to the optimal effiieny onfiguration (η II,opt ), and by 219 k /year with respet to the optimal exergy ost onfiguration (Ċ ex,opt ). On the other hand, onsidering as design point the optimal exergy ost of eletriity (Ċ ex,opt ), the global resoure onsumption of the plant is redued by 84,5 toe/year and 1070 toe/year if ompared respetively with the optimal effiieny (η II,opt ) and eonomi (Ċ eo,opt ) onfigurations. Table 5. Results of CCPP exergy based optimizations. Parameter u. m. η II,opt Ċ eo,opt Ċ ex,opt β C - 14.36 13.32 14.29 ṁ NG,pf kg/s 1.60 1.84 1.75 ṁ NG,tot kg/s 9.24 9.41 9.25 ṁ Air kg/s 282.76 280.47 282.70 η II - 0.448 0.438 0.449 η CExC - 0.358 0.353 0.36 Ċ eo,el /h 21373 21245 21328 Ż eo /h 6149 5712 6106 Ċ eo,ng /h 15305 15586 15315 Ċ ex,el toe/h 37.39 37.76 37.36 Ż ex toe/h 4.27 3.97 4.24 Ċ ex,ng toe/h 33.51 34.12 33.53 To omprehensively evaluate the relationship between the three optimisation funtions it is therefore neessary to investigate the relation between thermodynami, eonomi and environmental osts of the produt. Coupling proedure for global optimization Figure 5 depits a general simplified 2-D representation of this optimization problem: the eonomi optimization of the plant design (Eo,opt) leads to an additional onsumption of resoures ( Ċ ex,p.41 toe/h) while the resoure ost optimization (Ex,opt) auses an inrement of the eonomi ost of the produt ( Ċ eo,p = 83.63 /h). The subsequent question that arises is whether it is possible to link these two aspets. Referring to Table 5 data, and assuming an average oil barrel market prie of 623 /toe (2011) [26] as a proxy for the primary exergy market prie, the CCPP operating in the optimal eonomi ost design absorbs 0.407 toe/h more with respet to the exergy optimal design. This primary energy surplus purhased by soiety at its market prie would results in 253.76 /h whih is grater then the differene of osts between the eonomi optimal design and the exergy optimal design ( Ċ eo,p = 83.63 /h). Therefore, this CCPP plant pays the primary exergy less than the ommerial prie: for the soietal nihe in whih this CCPP operates, it is onvenient to invest (perhaps by means of a speifially aimed inentive poliy) in systems designed to minimize the exergeti primary resoures rather than the eonomi osts: saving resoures osts globally less than produing them. Even if today we still live in world where the objetive funtion is the monetary ost, and therefore the most probable onfiguration hosen at the end of the proess would be the optimal eonomi ost onfiguration (Eo,opt), the urrent analysis open a window over a new hane for evaluating resoure onsumption as an environmental ost for the soiety. The evaluation of the optimal exergy ost onfiguration (Ex,opt) add another set of information to Deision Makers for having a more omprehensive understanding of the overall impat of the total system for the whole soiety. Figure 5. 2-D shemati omparison among eonomi and exergeti ost funtion of produt. CONCLUSION Results of the presented exergy based analyses onfirm the existene of substantial differenes in designing the CCPP plant onsidering the optimization of seond law effiieny, eonomi or exergy osts of system produt. Thermoeonomis proved to be an appropriate framework to evaluate both eonomi and exergy ost of system produt. In partiular, exergy ost evaluation was expanded in order to inlude the embodied exergy of resoures and equipment into the analysis, as shown by eq. (12), aording to CExC method. This improvement gives a better insight of the overall energy onversion proess with respet to a standard exergy analysis. Moreover, it has been proposed a riterion for determining the relation between the eonomi and the environmental osts of the produt linked to the onsumption of resoures (i.e. ost of the primary exergy), giving therefore different perspetives to Deision Makers. In the urrent eonomi asset, the optimal eonomi ost onfiguration will be probably seleted but a number of information an be obtained by omparing the optimal effiieny onfiguration and the optimal exergy ost onfiguration. On the other hand, some drawbaks an be identified and they indiate possible further researh diretions to improve Thermoeonomi analysis from pratial point of view. They fall under two major ategories: need for standardization, refinement and extension of the CExC database, and a more 468
aurate soio-eonomi model to ompute the primary exergy market prie. NOMENCLATURE Symbol Quantity SI Unit Speifi ost /J J/J Ċ Cost rate /s J/s e Speifi exergy J/kg Ė Exergy rate J/s i Interest rate % m Total system streams - ṁ Mass flow rate kg/s n Total plant omponents - T Temperature C t Plant lifetime years Ẇ Work J/s Ż Investment ost rate /s J/s β C Air pressure ratio - η Effiieny - Subsripts adim D eo el ex i I II j P pf R tot REFERENCES Adimensional Destrution Eonomi Eletriity Exergeti Plant omponent no. Waste Seond Law Material / Energy flow no. Produt Post-firing Resoure total [1] A. Bejan, G. Tsatsaronis, M.J. Moran. Thermal Design and Optimization: Wiley, 1996. [2] E. Querol, B. González-Regueral, J.L.P. Benedito. Pratial Approah to Exergy and Thermoeonomi Analyses of Industrial Proesses: Springer, 2012. [3] P. Ahmadi, I. Diner. Thermodynami analysis and thermoeonomi optimization of a dual pressure ombined yle power plant with a supplementary firing unit. Energy Conversion and Management. 2011;52(5):2296-308. [4] Università di Roma "La Sapienza", CAMEL-Pro Users Manual, v.4. wwwturbomahineryit. 2008. [5] R. Melli, E. Siubba, C. Toro, A. Zoli-Porroni. An Improved POD tehnique for the optimization of MSF proesses. International Journal of Thermodynamis. 2012;15(4):231-8. [6] R. Melli, E. Siubba, C. Toro, A. Zoli-Porroni. An example of Thermo-Eonomi optimization of a CCGT by means of the Proper Orthogonal Deomposition Method. Proeedings of the ASME IMECE2012 November 9-15, 2012, Houston, TEXAS, USA. [7] V. Buljak. Proper Orthogonal Deomposition and Radial Basis Funtions for Fast Simulations. Inverse Analyses with Model Redution: Springer Berlin Heidelberg; 2012. p. 85-139. [8] A.a.T. Lazzaretto, G.. On the quest for objetive equations in exergy osting. in ML Ramalingam, JG Lage, V Mei, and JN Chapman (eds), Proeedings 'if the ASME Advaned Energy Syrems Devision, 37, ASME, New York, 197-210. 1997. [9] C.A. Frangopoulos. Comparison of thermoeonomi and thermodynami optimal design of a ombined-yle plant. in: DA Kouremenos (Ed), Proeedings of the International Conferene on the Analysis of Thermal and Energy Systems, Athens. 1991. [10] S.E. Jorgensen, B. Fath. Enylopedia of Eology, Five-Volume Set: Elsevier Siene, 2008. [11] G. Wall. Conditions and tools in the design of energy onversion and management systems of a sustainable soiety. Energy Conversion and Management. 2002;43(9 12):1235-48. [12] M.A. Rosen, I. Diner. On Exergy and Environmental Impat. International Journal of Energy Researh. 1997;21(7):643-54. [13] J.-F. Portha, S. Louret, M.-N. Pons, J.-N. Jaubert. Estimation of the environmental impat of a petrohemial proess using oupled LCA and exergy analysis. Resoures, Conservation and Reyling. 2010;54(5):291-8. [14] E. Siubba. Can an Environmental Indiator valid both at the loal and global sales be derived on a thermodynami basis? Eologial Indiators. 2013;29(0):125-37. [15] J. Szargut, D.R. Morris, F.R. Steward. Exergy analysis of thermal, hemial, and metallurgial proesses: Hemisphere, 1988. [16] J. Szargut. Exergy Method: Tehnial And Eologial Appliations: WIT Press, 2005. [17] R.L. Cornelissen, G.G. Hirs. The value of the exergeti life yle assessment besides the LCA. Energy Conversion and Management. 2002;43(9 12):1417-24. [18] B. De Meester, J. Dewulf, S. Verbeke, A. Janssens, H. Van Langenhove. Exergeti life-yle assessment (ELCA) for resoure onsumption evaluation in the built environment. Building and Environment. 2009;44(1):11-7. [19] S. International Federation of Institutes for Advaned. IFIAS Workshop Report, energy analysis and eonomis. Resoures and Energy. 1978;1(2):151-204. [20] R.U. Ayres, L.W. Ayres, K. Martinás. Exergy, waste aounting, and life-yle analysis. Energy. 1998;23(5):355-63. [21] P. Zhu, X. Feng, R.J. Shen. An Extension to the Cumulative Exergy Consumption Applied to Environmental Impat Analysis of Industrial Proesses. Proess Safety and Environmental Protetion. 2005;83(3):257-61. [22] G. Hirs. Thermodynamis applied. Where? Why? Energy. 2003;28(13):1303-13. [23] Simapro 7.3.3, site: http://www.pre.nl/simapro/. Pré Consultants. [24] X. Feng, G. Zhong, P. Zhu, Z. Gu. Cumulative exergy analysis of heat exhanger prodution and heat exhange proesses. Energy & fuels. 2004;18(4):1194-8. [25] S. Suh. Handbook of Input-output Analysis Eonomis in Industrial Eology: Springer London, Limited, 2009. [26] B. Petroleum. Statistial Review of World Energy. 2012. 469