at SciVerse ScienceDirect

Environmental Modelling & Software 30 (2012) 71e80

Contents lists available

Environmental Modelling & Software

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

Modeling simple trigeneration systems for the distribution of environmentalloads

Monica Carvalho a,1,*, Miguel A. Lozano a, Luis M. Serra a, Volker Wohlgemuth b

aGroup of Thermal Engineering and Energy Systems (GITSE), Aragon Institute of Engineering Research (I3A), Department of Mechanical Engineering, University of Zaragoza,CPS de Ingenieros (Edf. Agustin de Betancourt), C/Maris de Luna, s/n, 50018, Zaragoza, SpainbHTW Berlin, University of Applied Sciences, Industrial Environmental Informatics Unit, Wilhelminenhofstr. 75A, 12459 Berlin, Germany

a r t i c l e i n f o

Article history:Received 4 May 2011Received in revised form1 November 2011Accepted 12 November 2011Available online 9 December 2011

Keywords:TrigenerationCO2

Material flow networksEnvironmental costsUmberto softwareEnvironmental Management InformationSystem

* Corresponding author. Present address: MIRARCORamsey Lake Road, Sudbury, ON P3E 2C6, Canada.fax: þ1 705 675 4838.

E-mail addresses: mcarvalho@mirarco.org (M. C(M.A. Lozano), serra@unizar.es (L.M. Serra), volke(V. Wohlgemuth).

1 Tel.: þ34 976 76 1000x5258; fax: þ34 976 76 261

1364-8152/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.envsoft.2011.11.005

a b s t r a c t

Integration of thermoeconomics and Life Cycle Analysis was carried out within the framework of anEnvironmental Management Information System. This combined approach identified where environ-mental loads were generated and tracked environmental loads throughout the system, allowing fora more precise understanding of operational activities. A trigeneration system was modeled, providingelectricity, heat, and cooling to a building. The trigeneration system consists of a cogeneration module,auxiliary boiler, absorption chiller and electrical chiller. The trigeneration system model is flexible, as itallows electricity from/to the electric grid to be purchased/sold, and part of the cogenerated heat to bewasted. Umberto software is specifically designed to analyze the distribution of material and energyresources throughout a productive system. The software is based on Petri networks, double-entrybookkeeping and cost accounting, allowing the setup of complex systems and also a combined material,energy and inventory calculation. An assistant was built to include the tracking of emissions through theapplication of algebra and rules similar to those used in thermoeconomic analysis. It is possible toevaluate the environmental impact in terms of the consumption of natural resources and generation ofemissions in the system, from the input of natural resources to the output of the final products. Networkparameters were used to calculate the emissions associated with the operation of the system. The issueof allocating environmental loads was introduced and two scenarios for each operational mode werecompared: the trigeneration system vs. a conventional energy supply system in which electricity wasproduced in a representative coal power plant. In this case the trigeneration system operated withsignificant reduction of the CO2 emitted into the atmosphere.

� 2011 Elsevier Ltd. All rights reserved.

Software availability

Umberto software: Developed by ifu Hamburg GmbH. Contactaddress: ifu Hamburg GmbH, Max-Brauer-Allee 50, D-22765Hamburg, Germany. Tel.:þ49 404800090; fax:þ49 404800922.info@ifu.com, http://www.umberto.de/en/LINGO software: Developed by LINDO systems Inc. Contactaddress: 1415 North Dayton St, Chicago, IL, 60622, United States.Tel.: þ1 312 988 7422. info@lindo.com, www.lindo.com

, Laurentian University, 935Tel.: þ1 705 675 1151x6613;

arvalho), mlozano@unizar.esr.wohlgemuth@htw-berlin.de

6.

All rights reserved.

SimaPro software: Developed by PRé Consultants. Contactaddress: Printerweg 18, 3821 AD Amersfoort, The Netherlands.Tel.: þ31 33 450 4010. support@pre-sustainability.com, http://www.pre-sustainability.com/

1. Introduction

Residential and tertiary sectors are responsible formore than 40%of final energy consumption in the European Community (DirectiveCOM 2002/91/EC, 2002). The tertiary sector includes different typesof buildings (hospitals, schools, hotels, etc.) with a great variety ofuses and energy services (heating, cooling, and electricity). Europeanresearch projects, such as CHOSE (2001), TRIGEMED (2003), andSummerheat (2008), concluded that there is a significant technicalpotential for the implementation of trigeneration in the residentialand tertiary sector of countries in the Mediterranean area. In these

Nomenclature

AB auxiliary boilerAC absorption chillerC operation mode (followed by subscript)CM cogeneration moduleCOP coefficient of performanceEd electricity demandEp electricity purchased from the gridEr electricity input to electrical chillerEs electricity sold to the grid (cogenerated)EM environmental loads (followed by subscript)ExC example of operation mode (followed by subscript)EC electrical chillerFc fuel for the cogeneration module (natural gas)Fa fuel for the auxiliary boiler (fuel oil)L loss nodeMFN Material flow networksP Purchase nodepep price of purchased electricity

pes price of electricity sold to the gridpfa price of fuel oil (for auxiliary boiler)pfc price of natural gas (for cogeneration module)Q heat distribution nodeQa heat produced by auxiliary boilerQc heat produced by cogeneration moduleQcc heat internally consumedQl wasted heatQr heat input to absorption chillerQd heating demandR cooling nodeRd cooling demandRe cooling produced by electrical chillerRq cooling produced by absorption chillerrql cost of waste heatS sale nodeWc electricity produced by cogeneration moduleWcc electricity internally consumeda efficiency (cogeneration module)h efficiency (boiler)

M. Carvalho et al. / Environmental Modelling & Software 30 (2012) 71e8072

countries the need for heating is restricted to few winter months,limiting the application of cogeneration systems. There is, however,a significant need for cooling during the summer period. One solu-tion is the use of absorption chillers for cooling. By combiningcogeneration with absorption chillers (trigeneration), the energydemand can be extended into the summer months to match coolingloads. Advantages of trigeneration systems in buildings have beendemonstrated in literature, as the improved use of fuel is associatedwith economic savings and sparing of the environment, as less fuel isconsumed and consequently less pollution is generated (Maglorieet al., 2002; Chicco and Mancarella, 2008).

As sustainability-related issues such as energy consumption andenvironmental impact become a more integrated part of opera-tional and long-term planning decisions in energy systems, simu-lation modeling and analysis tools are needed to aid in the decisionmaking process.

Thermoeconomics is a potent tool for energy analysis and hasbeen used to support the design, synthesis and operation of energysystems by providing crucial information not available throughconventional analyses. Thermoeconomics combines economic andthermodynamic analysis with the purpose of revealing opportunitiesof energy and cost savings when designing and operating energyconversion systems (El-Sayed and Evans, 1970; Gaggioli, 1983, El-Sayed and Gaggioli, 1989; Lozano and Valero, 1993; Serra et al.,2009). The basic concept of thermoeconomic analysis is the energycost, understood as the amount of energy resources consumed forobtaining a piece of equipment, a flow or a commodity. Hence, thecost of a flow in a plant represents the amount of resources that havetobe supplied to theoverall systemtoproduce thisflow.Unit costs areused by thermoeconomic cost accounting theories for rational priceassessment, and allow us to follow the cost formation processthroughout the system, from energy resources to final products.Thermoeconomic costs canbeexpressed inmonetaryorenergyunits.

The necessity of considering the environment as an additionaldesign factor is an increasing demand due to the uprise of envi-ronmental conscience and the requirements to reduce the envi-ronmental impact of modern society. Life Cycle Analysis (LCA)provides a more global perspective of environmental loads and hasthe potential to fulfill the need for an adequate design tool forenergy supply systems (Guinée, 2002). LCA is an objective processto evaluate the environmental loads associated with a product,process, or activity (Awasthi and Chauhan, 2011; Carvalho et al.,

2011a,b; Turner et al., 2011). LCA also identifies and quantifies theuse of mass and energy as well as the emissions to the environ-ment, determining the impact of the use of resources and emis-sions, allowing for evaluation and implementation of strategies ofenvironmental improvement. The life cycle or cradle-to-graveimpacts include those resulting from extraction of raw materials,fabrication of the product, transportation or distribution of theproduct to the consumer, use of the product by the consumer, anddisposal or recovery of the product after its useful life. Herein, theenvironmental cost of a flow represents the amount of environ-mental loads that have been generated in the overall system toproduce this flow.

Thermoeconomic and LCA techniques are both based on thepremise that all of the resources required for producing a good orservice need to be accounted for. Thermoeconomics is usuallyapplied to energy conversion systems and the limits of the systemare those of the associated plant. However, there is no constraintthat impedes widening the limits of analysis to include the wellor the mine from where the natural resources were extracted.Thus, merging thermoeconomics and LCA methodologiesprovides a global perspective of a complex system via an inte-grated analysis of energy, economics and environment. Generally,the analyzed system in LCA is treated as a black box from whichonly its inputs and outputs are measurable, without furtherknowledge of the inner structure. Applying the philosophy ofthermoeconomics to energy systems opens this black box andunravels the process of environmental burden formation, which iswhere the importance of combining thermoeconomics with LCAlies (Carvalho, 2011). Similar to the cost formation process inthermoeconomics, it will be possible to evaluate the process offormation of the environmental impact linked with consumptionof natural resources and distribution of environmental loadsthroughout the system e i.e., from the input of natural resourcesto the output of final products.

Integration of thermoeconomics and LCAwas carried out withinthe framework of an Environmental Management InformationSystem (EMIS). EMIS are designed to detect, evaluate and preventa wide range of environmental dangers and stresses. In moreconcrete terms, EMIS consist of computer programs that supportmanagement by collecting, documenting and evaluating all rele-vant data about an enterprise’s interaction with its environmentand plan, initiate and control all activities related to environmental

M. Carvalho et al. / Environmental Modelling & Software 30 (2012) 71e80 73

protection (Pokorný, 2006; Wohlgemuth et al., 2006; Díez andMcIntosh, 2009.). Combination of different techniques or analysesis not uncommon, addressing the shortcomings of individualanalyses and allowing for a more complete result, such as thecombination of Geographical Information System analysis, SystemDynamics model and 3D visualization for sustainability assess-ments (Xu and Coors, 2011) or the combination of PrincipaleAgentmodels with Linear Programming in order to deal with the designof environmental regulation (Viaggi et al., 2009).

Umberto software (2006) is an EMIS and has been specificallydesigned for analyzing the distribution of material and energyresources throughout a productive system. Umberto can be used asan instrument for analyzing complex networks of materials andenergy flows, stocks, as well as material and energy transformations.

This study modeled a simple trigeneration system with theUmberto software, incorporating environmental information onthe consumption of resources. Even if a network can provide dataon the level of material and energy flows, it is necessary to extendthe boundaries because life cycle assessments require a represen-tation of the entire life cycle of products and services, including rawmaterial extraction, distribution, use phase, and waste disposal(Möller, 2010). Such an extension is interesting for companies,as the concept of resource productivity is a new administrativeapproach to deal with sustainability challenges (Porter and van derLinde, 1995).

The combined methodology allowed consumers of electricity,heat, and coolingproducedby trigeneration systems to know the unitenvironmental loads (equivalent to the thermoeconomic unit costs)associated with the consumption of each energy service. Incorpo-rating economic and environmental criteria into the design andplanning process should strive toward the following: (1) increasedefficiency in the use of energy and materials; (2) reduction of energycost of final products; and (3) reduction of environmental burden.

The allocation of environmental loads inmultiproduct processesis a much debated and critical issue. This is the case for trigenera-tion plants, and in general, for any energy plant producing morethan one useful energy flow. Through a detailed examination of theoperation modes of a trigeneration system, an allocation proposalis presented, providing better insight on the characteristics ofa trigeneration system (exposing the distribution of environmentalloads throughout the trigeneration system). The environmentalallocation proposal provided energy services with fewer environ-mental loads than those associated with separate production,2

which is very important on a consumer-oriented basis. If theconsumers of the trigeneration system assess that allocation wasfair, their buy-in is more likely to occur. A fair apportionment willcontribute to the acceptance of the more complex but more effi-cient trigeneration systems by consumers, which is essential for thesuccess of such systems when they are oriented to multiple users.

2. Simple trigeneration system

Trigeneration is the combined production of heat, cooling andelectricity from the same source of energy. The benefits of trigen-eration arise from the comprehensive integration of the processesand technologies used to supply energy in all necessary forms tomeet defined energy demand profiles. Trigeneration technologieshave socio-economic and environmental benefits that relate totheir efficient use of energy resources, typically resulting in thereduction of operation costs and in environmental benefits(reduced carbon emissions).

2 Electricity purchased from the grid, heat produced by a boiler, and coolingproduced by an electric chiller driven by purchased electricity.

The technology behind trigeneration is fundamentally based onthe coupling of a cogeneration module with an absorption chiller.The simple trigeneration system analyzed herein is complementedby the usual equipment present in a conventional plant: a hotwaterboiler and an electrical chiller (Lozano et al., 2009a). Both tech-nologies are used to guarantee supply and also to avoid oversizingthe cogeneration module and the associated absorption chiller. Theidea is that the cogeneration module, jointly with the absorptionchiller, satisfies the average thermal demand for heat and cooling,while the conventional units (boiler and electrical chiller) areutilized in an auxiliary way to make up for the demand peaks.Therefore, supply is guaranteed and the installation is reliable,since the existence of conventional equipment assures the satis-faction of the thermal demand (Lozano et al., 2009b).

Trigeneration systems have wide range of applications: singleresidential applications (Wang et al., 2008), industry (Colonna andGabrielli, 2003), or buildings (Marimón et al., 2011).

2.1. Structure and operation

The purpose of the trigeneration system (Fig. 1) was to meet thedemand of different energy services (electricity, Ed; heating, Qd;and cooling, Rd) of a consumer center. Fig. 1 shows a diagram of theanalyzed trigeneration system, with internal and product flows.

The simple trigeneration system consists of the followingproductive units: a cogeneration module CM (providing heat, Qc,and electricity, Wc), an auxiliary boiler AB (providing heat, Qa), anabsorption chiller AC (providing cooling, Rq, and driven by heat, Qr)and an electrical chiller EC (providing cooling, Re, and driven byelectricity, Er).

Table 1 presents the technical parameters of the simple trigen-eration system. Efficiency coefficients were obtained from equip-ment catalogs and consultations with manufactures. For eachequipment, Pnom is the nominal capacity (power of its mainproduct). Taking the cogeneration module as an example, elec-tricity is the main product and Pnom¼Wc nom¼ 350 kW. To produceWc kW of electricity, Fc¼ (1/aw)Wc kW of natural gas will beconsumed, recovering Qc¼ (aq/aw)Wc kWof heat. It was consideredthat the efficiency coefficients were constant and independentfrom the production P� Pnom of the equipment.

Demands were always met either by the productive units of thetrigeneration system or with the help of purchased electricity fromthe electric grid (Ep). The possibilities also existed that a fraction(Ql> 0) of the cogenerated heat could bewasted, and the electricity

Fig. 1. Simple trigeneration system with flows.

Table 1Technical parameters of the simple trigeneration system.

Transition/equipment Efficiency coefficient Nominal capacity (kW)

Cogeneration module awhWc/Fc¼ 0.35 Wc nom¼ 350aqhQc/Fc¼ 0.40

Auxiliary boiler hqhQa/Fa¼ 0.80 Qa nom¼ 400Absorption chiller COPqh Rq/Qr¼ 0.625 Rq nom¼ 250Electrical chiller COPeh Re/Er¼ 5.0 Re nom¼ 250

M. Carvalho et al. / Environmental Modelling & Software 30 (2012) 71e8074

could be sold to themarket (Es). Fc and Fa refer to the fuel utilized bythe cogeneration module and the auxiliary boiler, respectively.Environmental loads were generated in the cogeneration module,which operates on natural gas and in the auxiliary boiler, whichoperate on fuel oil. When electricity was purchased or sold from/tothe electric grid, environmental loads were also considered.

The minimal operation CO2 emissions of the system wereevaluated by a linear programming model, as explained in section2.3. The model was applied to very different demands of energyservices and the results show how different the operation modesfor a simple trigeneration system can be. A greater sophistication ofthe optimization model, using non linear production restrictionstaking into account how ambient conditions affect capacity andefficiency coefficients and binary variables limiting both theminimum load of the equipments and the on/off status, wouldprovide more precise results but, generally, the achieved conclu-sions would still prevail.

The approach presented in this paper can be used for wholesystem design by considering, in the optimization model and anal-ysis methodology, the life cycle environmental loads of the system:a) the environmental loads embedded in the selected equipment forthe plant and b) the environmental loads associated with the oper-ation of the plant. In this paper, only the second part (b) was studied:the minimization of the environmental loads associated with theoperation of a previously dimensioned trigeneration system. Mini-mization of environmental loads associated with the whole systemshould also be considered as an obligatory prior step of project anddesignmethodologies for new trigeneration systems (Carvalho et al.,2011a,b).

2.2. Environmental loads of fuels and electricity

SimaPro (2008) is a specialized LCA tool, one of the most widelyused LCA softwares, used by major industries and consultants,through to research institutes and universities, and was utilized tocalculate the impact associated with the operation of the system(consumption of resources). This was possible because SimaProincludes several inventory databases with thousands of processesand several well-known impact assessment methods. The systeminteracted with the economic environment (market) through thepurchase of natural gas, fuel oil, and electricity from the grid, aswell as through the sale of cogenerated electricity to the grid.

LCA analyzes the environmental impacts associated witha process or product from ‘the cradle to the grave’, which beginswith the gathering of raw materials from the earth to create theproduct/service and ends at the point when all materials arereturned to the earth (SAIC, 2006).

2.2.1. Natural gasNatural gas was characterized by utilizing the related emissions

of combustion of natural gas, from the IDEMAT database (IDEMAT,2001), and the total aggregated system inventory for a natural gasconsumer in Spain, from the Ecoinvent database (Ecoinvent, 2007).The CO2 emissions associated with the consumption of natural gasin Spain were obtained by utilizing SimaPro, calculated asEMfc¼ 0.272 kg CO2 per kWh of consumed natural gas.

2.2.2. Fuel oilFuel oil was characterized by an inventory module (extraction,

production at refinery and transportation from refinery to anaverage European end user) and related emissions of controlledburning, from the Ecoinvent database. The CO2 emissions associ-ated with the consumption of fuel oil were obtained by utilizingSimaPro, calculated as EMfa¼ 0.305 kg CO2 per kWh of consumedfuel oil.

2.2.3. ElectricityThe CO2 emissions associated with the local electricity were

also calculated by SimaPro, utilizing the Ecoinvent database andconsidering that all electricity originated from a single-fuel repre-sentative coal power plant (EMep¼ 1.020 kg CO2/kWh).

The evaluation of CO2 emissions in a representative coal powerplant was utilized as an example here, in order to illustrate theapplication of the proposedmethodology. Carvalho (2011) analyzedother environmental indicators as well as other types of powerplants, and different national electricity mixes. Using coal powerplants, different operation modes of the system appear whenvarying the energy services demands. This allowed for a clearer andcomplete validation of the proposed methodology.

2.3. Optimal operation modes

A linear programming model was solved in order to obtain theoptimal operation mode from an environmental viewpoint. Theenvironmental analysis considered that the only significant vari-able environmental loads were electricity, natural gas and fuel oil.No environmental burden was associated with the waste ofcogenerated heat, i.e., EMql¼ 0. The objective function to be mini-mized was the operation variable emissions (in kg CO2/h):

Operation variable emissions ¼ EMfcFcþEMfaFaþEMepEp

�EMesEsþEMqlQl (1)

Cogenerated electricity sold to the grid was considered to have thesame environmental load as that of electricity purchased from thegrid (EMes¼ EMep). The concept of avoided emissions is presentedas the emissions avoided elsewhere with the production of elec-tricity by the cogeneration module, consequently avoiding thepurchase of electricity from the grid.

Equation (1) was subject to restrictions of capacity limit andequipment efficiency as well as balance equations, previouslypresented in Lozano et al. (2009a). Results were obtained byutilizing the modeling language and optimizer Lingo (2008). Lingois a commercial software package for solving optimization prob-lems that uses the branch and bound solver to enforce any integerrestrictions contained in a model. The advanced capabilities ofLingo such as cut generation, tree reordering, advanced heuristicand presolve strategies were used as needed.

Given the energy demands to be satisfied, according to thedifferent operation modes, Lingo solved the previous model anddetermined the feasible operation mode with the minimum oper-ation variable emissions.

The resulting feasible operation states could be classified intonine different operation modes, based on the values of purchasedelectricity (Ep), sold electricity (Es), auxiliary heat (Qa) and wasteheat (Ql). These operationmodes corresponded to different demandof the energy services of the consumer center and are shown inTable 2.

A summary of results (demand, flows, and hourly cost) obtainedwith Lingo for four examples ExC1, ExC3, ExC7 and ExC9 that cor-responded to different operation modes (C1, C3, C7 and C9) is pre-sented in Table 3. For each different example, the minimum

Table 3Energy flows and variable CO2 emissions.

ExC1 ExC3 ExC7 ExC9

Ed kW 400 400 200 200Qd kW 400 100 600 100Rd kW 400 100 100 100

Ep kW 100 50 0 0Es kW 0 0 130 150Fc kW 1000 1000 1000 1000Fa kW 300 0 250 0Wc kW 350 350 350 350Qc kW 400 400 400 400Wcc kW 350 350 220 200Er kW 50 0 20 0Ql kW 0 140 0 140Qcc kW 400 260 400 260Qa kW 240 0 200 0Qr kW 240 160 0 160Rq kW 150 100 0 100Re kW 250 0 100 0

CO2 emissions kg CO2/h 465.50 323.00 215.65 119.00Operation mode C1 C3 C7 C9

M. Carvalho et al. / Environmental Modelling & Software 30 (2012) 71e80 75

emissions of satisfying the energy service demand of the consumercenter was reached in a different operationmode, which exchangedenergy flows with the environment and utilized the productivecapacity of the installed equipment. Table 3 shows the energy flowsobtained and the operation variable CO2 emissions.

3. Environmental management information system

3.1. Umberto software for material and energy flow analysis

Software support for Material Flow Networks (MFN) must meeta number of requirements. For instance, the software must becapable of modeling complex production processes from differentfields of application, such as chemistry or engineering, and mustoffer flexible data management for updating and extending themodel. Different ways of data interpretation and display must alsobe supported. The Umberto software, which entered the market inthe mid-1990s, was the first suite of programs that tried to meetthese goals. Umberto software is a powerful and user-friendlymaterial flow analysis tool. It runs under the Microsoft Windowsfamily of operating systems and meets software ergonomic stan-dards (Wohlgemuth et al., 2006).

The inclusion of environmental information on the usage andconsumption of resources into Umberto allowed for the evaluationof the distribution of the environmental loads associated with eachflow of the system. Additionally, integrated models of energy flowsfacilitated a better understanding of the assignment of environ-mental and economic costs to the internal and final products of thetrigeneration system.

The modeling of material flows in multi-stage productionsystems initially focused on absolute flows of companies andsupply chains. The objective is to trace the material flows withina company or between different companies within a value chain. InMFN, the term material refers to substances and energy, meaningthere is virtually no distinction between substances and energy.Based on the concept of material flow networks, the powerfulcalculation algorithm of Umberto allows for the determination ofall material and energy flows in the system under study.

According to Wohlgemuth et al. (2006), the most attractivefeature of MFN is the possibility to combine the compilation of eco-balances for an industrial plant with an analysis of material flowsassociated with given products or services. An eco-balance refers tothe consumption of energy and resources and the pollution causedby the production cycle of a product. An advantage of the MFNapproach resides in its gradual modeling approach, starting froma very basic model of few processes with simple specifications, themodel can be extended step by step to include further processes,sites, more complex specifications, costs, etc. (Viere et al., 2010).

Umberto software allows for the visualization of processes, unitsand flows, carrying out mass and energy balances and analyzingfrom an environmental point of view the emissions generated. PetriNets, double-entry bookkeeping, and cost accounting are the basis ofUmberto software, allowing the setup of complex systems and also

Table 2Operation modes.

Yes purchased electricity

No sold electricity

Yes auxiliary boiler C1No waste heatNo auxiliary boiler C2No waste heatNo auxiliary boiler C3Yes waste heat

a combined material, energy and inventory calculation. Umbertomodels consist of places, transitions, and arrows (directed graphs).

3.1.1. PlacesAn important function of places is that they delimit the system

from its environment; they are points of contact with the world. Theinputs of the simple trigeneration system (Fig. 4, light gray circles)were the consumption of fuel by the cogeneration module (Fc) andauxiliary boiler (Fa), and the electricity purchased from the grid (Ep).

The outputs of the system (dark gray circles in Fig. 4) were thedemands of electricity (Ed), heat (Qd), and cooling (Rd). Freedomwas available to the consumer to decide how the system operated:wasting heat permitted the operation of the cogeneration moduleto match the demand of the consumer center and the sale ofsurplus cogenerated electricity permitted to realize profit. There-fore two more outputs of the system were waste heat (Ql) and thecogenerated electricity sold to the grid (Es).

The place “Environmental loads” accounted for the environ-mental loads originating from the consumption of natural gas andfuel oil, and from the purchase or sale of electricity. The two“Environmental loads” outputs seen in Fig. 2 are duplicate places. Ifan arrow leads to a place far away, the graphical display mightbecome incomprehensible. Therefore the “Environmental loads”place was duplicated and the copy was positioned in the vicinity oftransition P. All emissions go into the atmosphere, but Umbertosoftware tracks the contribution of each transition to account for itsshare of emissions.

3.1.2. TransitionsEach piece of equipment was modeled as a transition. A slightly

more complex but more flexible method to specify transitions was

No purchased electricity No purchased electricity

No sold electricity Yes sold electricity

C4 C7

C5 C8

C6 C9

Fig. 2. Umberto model of simple trigeneration system.

M. Carvalho et al. / Environmental Modelling & Software 30 (2012) 71e8076

applied, utilizing expressions to describe the relationships betweeninput and output flows of a transition, using the technical param-eters shown in Table 1. To guarantee that the network could becalculated in both directions, the user-defined functions alsoconsidered the inverse form.

Environmental load values were incorporated into Umberto asnet parameters, which are parameters valid not only locally fora single transition but for an entire network or for subnets. Netparameters can be used in the functions of the transition specifi-cations in the same way as transition parameters.

Branching and merging points S (Sale node), L (Waste heatnode), P (Purchase node), Q (Heat distribution node), and R (Coolingnode) were also modeled as transitions. Branching and mergingpoints can be interpreted as decision points, in which possibilitiesare reflected. Point S refers to the possibility of selling cogeneratedelectricity to the grid; point L refers to the possibility of wastingpart of the cogenerated heat; point P refers to the possibility ofpurchasing electricity from the grid; point Q refers to the possibilityof operating the auxiliary boiler, and point R adds the contributionsof the chillers to satisfy the cooling demand.

3.2. Environmental cost accounting

As previously mentioned, thermoeconomic analysis (costaccounting) can be combined with LCA, and both techniques canbe integrated into an EMIS. In the particular case of Umberto, itwas necessary to build an assistant to integrate the philosophy/methodology utilized in energy cost analysis (thermoeconomics)with the evaluation of environmental loads. The concept of costcan involve different magnitudes, as for example, environmentalloads. Environmental costs can be understood as a category of cost(according to the consumption of natural resources and genera-tion of environmental loads in order to obtain a flow).

For the implementation of the environmental allocationmethodbased on thermoeconomics, an assistant was created in Umbertosoftware. The assistant performed calculations of environmentalloads of internal flows and products after network calculation. Theinput monitor of Umberto software was used to model the systemwith greater flexibility, allowing modifications in operation modesto be carried out easily. The flows that defined each operationmodewere established in the input model, and then Umberto softwarecalculated the remaining flows.

Umberto offers several interfaces to other programs, one ofwhich allows the specification of transitions. In combination withcomplex algorithms it can embed sub-models into MFN that havebeen developed outside of Umberto. This can be done througha script written in any of the languages supporting Microsoft’sActive Scripting Architecture. The new functions and extensionswere implemented within the menu structure, utilizing structurallanguage XML with code/logic J#. XML was chosen because it ishuman-legible and reasonably clear, easy to create, and the user hasthe advantage of acting independently of software (data can bemoved through software upgrades and even to different softwareproducts). J# was used basically for its text highlighting abilitiesand because it uses Microsoft’s Common Language Runtime. Theassistant was an application that collected data of the calculatedflows to carry out cost accounting. The assistant was necessarybecause Umberto calculates flows and costs simultaneously, andthe implementation of thermoeconomic equations required theflows to be previously calculated.

The assistant was validated with economic costs (correctlyreproducing thermoeconomic cost results published in Lozano et al.(2009a,c). By changing ‘market prices’ to ‘environmental loads’, theassistant turned to an environmental perspective, giving the assis-tant flexibility to support calculations regarding environmental loadsor economic costs.

Balances were formulated in each transition and externalresources used in the production process were valued by the envi-ronmental burden caused. Applying the condition of cost conser-vation to each transition yields:

Equipment

CM : EMfcFc ¼ EMwcWc þ EMqcQc (2)

AB : EMfaFa ¼ EMqaQa (3)

AC : EMqrQr ¼ EMrqRq (4)

EC : EMerEr ¼ EMreRe (5)

Branching and merging points

S : EMwcWc ¼ EMwccWcc þ EMesEs (6)

P : EMwccWcc þ EMepEp ¼ EMerEr þ EMedEd (7)

L : EMqcQc þ EMqlQl ¼ EMqccQcc (8)

R : EMrqRq þ EMreRe ¼ EMrdRd (9)

Q : EMqccQcc þ EMqaQa ¼ EMqrQr þ EMqdQd (10)

Considering that the operation state of the plant was known,then all energy flows, environmental loads for fuel and electricityand the environmental load entailing waste heat were also known.Here it was considered that EMql¼ 0 because the objective was toassess all environmental loads to useful final products. EMql¼ 0does not mean an absence of environmental impact, but that therewas no environmental burden associated with the act of wastingheat. Consequently, there were 12 unit environmental loads ofinternal flows and final products to be calculated: EMwc, EMwcc,EMer, EMed, EMqc, EMqcc, EMqa, EMqr, EMqd, EMrq, EMre, and EMrd. As

Fig. 3. Control volume accounting for the interaction of the cogeneration module withthe environment.

M. Carvalho et al. / Environmental Modelling & Software 30 (2012) 71e80 77

the system was described using 9 equations with 12 unknowns, 3auxiliary equations were needed.

It was considered that the unit environmental load of severalflows obtained from a homogeneous flow was the same. Applyingthis rule to branching points P and Q, two auxiliary equations wereobtained:

P : EMer ¼ EMed (11)

Q : EMqr ¼ EMqd (12)

This means that the environmental loads assessed to the elec-tricity internally consumed in the electrical chiller (Er) and theelectricity provided to the consumer center (Ed) are the same. Theenvironmental loads assessed to the heat internally consumed inthe absorption chiller (Qr) and the heat provided to the consumercenter (Qd) follow the same rule.

The third auxiliary equation must define how the environmentalloads generated in the cogeneration module are attributed to itsproducts: heat and electricity. Allocation is a very important issuewhen apportioning environmental loads tomultiproduct systems, toensure each party is credited with their appropriate share. Researchon allocation of emissions and environmental burdens will allowthe environmental benefits of co- and tri-generation technologies(adequately designed and operated) to be better understood andexploited (Rosen, 2008; Abusoglu and Kanoglu, 2009; Carvalho,2011).

Umberto software supports the consideration of differentapproaches to the allocation issue, and even supports the use ofscripts to attach complex rules ormodels to a transition.However, theissue of allocation was not addressed in depth in this paper. A deepdiscussion on the allocation of environmental loads is presented inCarvalho (2011). Different allocationmethods of environmental loadsto electricity and heat products (third auxiliary equation for theanalyzed system) were found in literature (Phylipsen et al., 1998).

Fig. 4. Conventio

However, the main issue found during the utilization of such simplemethods focused on the immediate products of the cogenerationmodule,Qc andWc, not accounting for possible different destinationsor uses of Qc and Wc nor for the interaction of the cogenerationmodule with its productive environment.

Since traditional solutions to the problemwere unsatisfactory, theauthors proposed an innovative allocation method in Carvalho et al.(2010), which was implemented in Umberto software through theaforementioned assistant, allocating emissions to the consumedproducts of the cogeneration module (Wcc and Qcc), in proportion tothe emissions generated by their alternative production:

EMqcc

EMwcc¼ EMqa

EMep(13)

EMep being the environmental loads corresponding to electricitypurchased from the electric grid, and EMqa the environmentalloads associated with the heat produced in the auxiliary boiler(EMqa¼ EMfa/hq). Equation (13) could be applied directly to alloperationmodes, as it was previously established that EMep¼ EMes.

The choice of such control volume allows the benefits of sellingelectricity and the inefficiency of wasting heat to be both distrib-uted between heat and electricity internally consumed.

When considering different equipment, activities, and optionsincluded in the trigeneration system, the assignment of unit costsshould rather consider the products of the cogeneration modulethat were consumed (Wcc and Qcc). In this way, as shown in Fig. 3,the emissions associated with the operation of the cogenerationmodule, the sale of electricity to the grid, and waste heat(EMfcFc� EMesEsþ EMqlQl) were distributed between heat andelectricity internally consumed (EMwccWcc and EMqccQcc).

4. Results and discussion

Based on the flow quantities entered in the input monitor(energy demands and environmental loads associated withoperation), the system calculated the flows of the entire networkusing the transition specifications. For comparison, a referencesystem (Fig. 4) was created, in which all energy demands weresatisfied in a conventional way (electricity was purchased fromthe grid for Ed and Rd, and Qd was satisfied by an auxiliary boiler).Table 4 shows the emissions for the trigeneration and conven-tional systems.

Umberto software provided the overall amount of emissionsassociated with the operation of the trigeneration system. It wasverified that trigeneration technology reduced significantly thekg of CO2 emitted into the atmosphere. Table 5 shows the unitenvironmental loads associated with the operation of the simpletrigeneration system for all operation examples.

nal system.

Table 6Energy resources and CO2 emissions assessed to final products.

Ep (kWh) Fc (kWh) Fa (kWh) CO2 (kg CO2)

ExC1 Ed 88.89 622.72 e 260.05Qd e 187.15 187.50 108.09Rd 11.11 190.13 112.50 97.36Total 100.00 1000.00 300.00 465.50

ExC3 Ed 50.00 700.56 e 241.55Qd e 115.17 e 31.33Rd e 184.27 e 50.12Total 50.00 1000.00 e 323.00

ExC7 Ed e 400.32 e 108.89Qd e 299.44 250.00 157.70Rd e 40.03 e 10.89Es e 260.21 e �61.83Total e 1000.00 250.00 215.65

ExC9 Ed e 400.32 e 108.89Qd e 115.17 e 31.32Rd e 184.27 e 50.12Es e 300.24 e �71.33Total e 1000.00 e 119.00

Table 4CO2 emissions of trigeneration system vs. conventional system.

ExC1 ExC3 ExC7 ExC9

Conventional kg CO2/h 642.12 466.53 453.18 262.53Trigeneration kg CO2/h 465.50 323.00 215.65 119.00

M. Carvalho et al. / Environmental Modelling & Software 30 (2012) 71e8078

Comparison of emissions between examples ExC1 and ExC3 gaveindications on what occurred when some heat was wasted. EMwcand EMwcc increased their value in ExC3, reflecting the inefficiencyof wasting heat, and consequently, EMed had a higher value. EMqdpresented a lower value in ExC3, promoting the consumption ofwaste heat.

The comparison between examples ExC1 and ExC7 gave indi-cations on the behavior of the systemwhen electricity was sold tothe grid. In ExC7, EMwc presented a higher value and EMwcc pre-sented a lower value than in ExC1. The sale of electricity withlower emissions (but evaluated as having higher emissions)consequently lowered the cost of EMwcc. The benefits of the sale ofelectricity were positively reflected on the values of EMqc andEMqcc, and ultimately on the final emissions of Ed, Qd and Rd,which were lower.

In ExC9, EMed increased reflecting the waste of heat, but withsale of electricity, EMed was still environmentally sounder thanEMep. The sale of electricity benefited all final energy services,resulting in lower emission values when comparing ExC3 andExC9. Internal flows too, were lower in ExC9. When comparingExC7 and ExC9, it could be seen that EMqc had a lower value,resulting in a lower value for EMqd which should promoteconsumption of otherwise wasted heat. EMwc and EMwcc hadincreased values which were translated into higher emissions forEMed and EMrd. The benefits as well as the penalties of the systemwere reflected in all energy services produced in the cogenerationmodule.

Table 6 shows the breakdown of energy resources and CO2emissions assessed to final products (electricity, heat and coolingdemands). Umberto provided the emissions associated with theconsumption of each energy service, and the origin of the emis-sions, tracking the contributions of electricity purchased from thegrid, and fuels of the cogeneration module and auxiliary boiler.

Umberto software uses diagrams to displaymaterials, energy andcost flows. Figs. 5e7 show the breakdown of the emissions associ-ated with the consumption of energy services in case ExC1. A scalingvariant (Maximum flow) was utilized for visualization purposes, sothe maximum width in the network was limited to each energyservice. All other widths were calculated proportionally.

Table 5CO2 emissions (kg CO2/kWh) associated with the operation of the trigenerationsystem.

ExC1 ExC3 ExC7 ExC9

EMed 0.6502 0.6597 0.3773 0.4004EMqd 0.2702 0.2273 0.2211 0.1497EMrd 0.2434 0.3638 0.0755 0.2395

EMes e e 1.0200 1.0200EMwc 0.5445 0.6083 0.6160 0.6660EMqc 0.2035 0.1478 0.1410 0.0973EMwcc 0.5445 0.6083 0.3773 0.4004EMer 0.6502 e 0.3773 e

EMqcc 0.2035 0.2273 0.1410 0.1497EMqa 0.3813 e 0.3813 e

EMqr 0.2702 0.2273 e 0.1497EMrq 0.4323 0.3638 e 0.2395EMre 0.1300 e 0.0755 e

Fig. 5 shows the contributors to the environmental loads of theelectricity demand (Ed), which is satisfied by the generation ofelectricity in the cogeneration module (operating with fuel Fc),and is complemented by the purchase of electricity from thegrid (Ep).

Fig. 6 shows the contributors to the environmental loads of theheat demand (Qd). The cogeneration module generates heat(operating with fuel Fc) and is complemented by heat produced bythe auxiliary boiler Qa (which operates with fuel Fa).

Fig. 7 shows the contributors to the environmental loads of thecooling demand Rd. Electricity purchased Ep from the grid andelectricity generated Wc by the cogeneration module contribute tooperate the electrical chiller. The auxiliary boiler operates on heatproduced by the cogeneration module (Qc) plus heat produced bythe auxiliary boiler (Qa).

It is important to highlight the fact that the integration proposedin this paper (LCA and thermoeconomics into an EMIS) supportsconsideration of any environmental indicator, i.e., Eco-indicator99 points, ozone layer depletion values (kgeq CFC-11 into air),

Fig. 5. Emissions associated with the consumption of Ed in example ExC1.

Fig. 6. Emissions associated with the consumption of Qd in example ExC1.

M. Carvalho et al. / Environmental Modelling & Software 30 (2012) 71e80 79

ecotoxicity values (kgeq triethylene glycol into water and soil), etc.Information shown in Tables 5 and 6 and Figs. 5e7 allows for thestudy of the distribution of consumed resources, of the costformation process, and the generation of environmental loadsthroughout the productive system.

This study emphasized the importance of integrating environ-mental information (obtained with the application of LCA tech-niques) and the philosophy of thermoeconomic analysis inEnvironmental Information Management Systems. Such integrationwas easily accomplished through an assistant implemented inUmberto. The assistant facilitated the registration and trackingof environmental impacts generated in each piece of equipmentas well as the assessment of environmental loads to each outputand internal flow, with the possibility of considering differentapproaches to the allocation issue. This combined approach allowedfor a more precise understanding of operational activities.

International markets are increasingly seeking an indication ofenvironmental performance and the carbon emissions associatedwith products and services. The responsibility exists to ensure long

Fig. 7. Emissions associated with the consumption of Rd in example ExC1.

term ecologically sustainable production, particularly in changingclimatic conditions. From an environmental viewpoint, the purposeof installing a trigeneration system was to provide a reduction inthe amount of environmental loads produced to attend a consumercenter, which was achieved here.

5. Summary and conclusions

A trigeneration systemwas modeled in Umberto software, withthe flexibility to purchase/sell electricity from/to the electric grid.The possibilities of wasting part of the cogenerated heat andoperating an auxiliary boiler also existed. The network was calcu-lated for specific energy service demands, identifying whereemissions were generated, and tracking the emissions throughoutthe system. This study case concentrated on the issue of climatechange and therefore considered the emissions of CO2. Flow anal-ysis of individual production steps specific to operation made itpossible to study the operational activities more precisely.

This paper demonstrates that LCA can and has been formallycombined with thermoeconomic analysis. The result is an ability totake thermoeconomics and LCAe and their tradeoff relationshipseinto account in product/process design decision making. Thiscombined approach was integrated into an EMIS, identifying whereenvironmental loads were generated and tracked their distributionto the final products of trigeneration systems. In an attempt toaddress the ongoing debate, an innovative environmental alloca-tion method was proposed.

The allocation proposal for trigeneration systems considered thatenvironmental loads of the cogeneration module were distributedamong the consumers of the final products, resulting in overallreduced emissions derived from the combined production. Suchreductions were evaluated in proportion to the emissions associatedwith obtaining heat and electricity separately via conventionalsystems. The assistant developed also provided Umberto with thecapability of calculating economic costs. A potent tool was obtainedfor energy, economic, and environmental analysis, to study thedistribution of resources throughout productive systems as well asthe formation process of costs and environmental loads.

Effective environmental related strategies connect the reductionof emissions with a system’s operational strategy (consumption ofresources). Therefore the usage of EMIS with double input fromthermoeconomics and LCA tools could be promoted to (1) analyzethe distribution of material and energy resources throughouta productive system; (2) serve the numerical registration andinterpretation of environmental effects; (3) calculate unit economiccosts and study the cost formation process; (4) identify the mostenvironmentally beneficial among competing technologies; and(5) allow an emission-efficient economy to develop.

By incorporating environmental information on the usage andconsumption of resources into Umberto software, the approach ofmaterial flow networks gave insight on the environmental loadsassociated with each flow of the system. Thus, the consumers ofa productive system can be made aware of the environmental loadsassociatedwith the consumption of each product (either internal orfinal). This information can be very useful for the introduction ofstrategies oriented to changes and improvements in the design andoperation of productive systems as well as in consumption patternsand resource conservation, contributing to the development ofa more sustainable economy.

Acknowledgments

This work was developed within the framework of researchproject ENE2007-67122, funded in part by the Spanish Government(Energy program) and the European Union (FEDER program).

M. Carvalho et al. / Environmental Modelling & Software 30 (2012) 71e8080

Special thanks are extended to the Umberto Competence Centre atHTW Berlin for the support given to Monica Carvalho during herresearch stay.

References

Abusoglu, A., Kanoglu, M., 2009. Allocation of emissions for power and steamproduction based on energy and exergy in diesel engine powered cogeneration.Energy & Fuels 23, 1526e1533.

Awasthi, A., Chauhan, S.S., 2011. Using AHP and Dempster-Shafer theory forevaluating sustainable transport solutions. Environmental Modelling & Soft-ware 26 (6), 787e796.

Carvalho, M., et al., 2010. Distribution of resources and allocation of environmentalloads applied to trigeneration systems. In: Proceedings of the 23nd Interna-tional Conference on Efficiency, Cost, Optimization, Simulation and Environ-mental Impact of Energy Systems, Lausanne, Switzerland.

Carvalho, M., 2011. Thermoeconomic and environmental analyses for the synthesisof polygeneration systems in the residential-commercial sector�. Ph.D. thesis,University of Zaragoza, Spain. <zaguan.unizar.es/record/5744/files/TESIS-2011-025.pdf> (accessed 01.11.11).

Carvalho, M., Serra, L.M., Lozano, M.A., 2011a. Geographic evaluation of trigenera-tion systems in the tertiary sector. Effect of climatic and electricity supplyconditions. Energy 36 (4), 1931e1939.

Carvalho, M., Serra, L.M., Lozano, M.A., 2011b. Optimal synthesis of trigenerationsystems subject to environmental constraints. Energy 36 (6), 3779e3790.

Chicco, G., Mancarella, P., 2008. Assessment of the greenhouse gas emissions fromcogeneration and trigeneration systems. Part I: Models and indicators. Energy33 (3), 410e417.

CHOSE, 2001. Energy savings by combined heat cooling and power plants (CHCP) inthe Hotel Sector. Final Report, SAVE Project. <http://www.inescc.pt/urepe/chose> (accessed 01.11.11).

Colonna, P., Gabrielli, S., 2003. Industrial trigeneration using ammoniaewaterabsorption refrigeration systems (AAR). Applied Thermal Engineering 23 (4),381e396.

COM 2002/91/EC, Directive on the energy performance of buildings. Official Diary ofthe EC, 4.1.2003, L1, 65e71.

Díez, E., McIntosh, B.S., 2009. A review of the factors which influence the use andusefulness of information systems. Environmental Modelling & Software 24 (5),588e602.

Ecoinvent v2.0 Database, 2007. Swiss Centre for Life Cycle Inventories, Dübendorf,Switzerland.

El-Sayed, Y.M., Evans, R.B., 1970. Thermoeconomics and the design of heat systems.Transactions of ASME Journal of Engineering for Power 92, 27e34.

El-Sayed, Y.M., Gaggioli, R.A., 1989. A critical review of second law costing methods:Parts I and II. ASME Journal of Energy Resource Technology 111, 1e15.

Gaggioli, R., 1983. Second law analysis for process and energy engineering. In:Efficiency and Costing, ACS Symposium Series, vol. 235, pp. 3e50.

Guinée, J.B. (Ed.), 2002. Handbook on Life Cycle Assessment: Operational Guide tothe ISO Standards. Kluwer Academic Publishers, Dordrecht; Boston.

IDEMAT Database, Section for Environmental Product Development, Faculty ofIndustrial Design, 2001. Delft University of Technology, Delft, The Netherlands.<http://www.idemat.nl> (accessed 01.11.11).

Lingo, 2008. The Modeling Language and Optimizer. <http://www.lindo.com>

Chicago, USA (accessed 01.11.11).Lozano, M.A., Valero, A., 1993. Theory of the exergetic cost. Energy 18 (9), 939e960.Lozano, M., Carvalho, M., Ramos, J.C., Serra, L.M., 2009a. Thermoeconomic analysis

of simple trigeneration systems. International Journal of Thermodynamics 12(3), 147e153.

Lozano, M.A., Ramos, J.C., Carvalho, M., Serra, L.M., 2009b. Structure optimization ofenergy supply systems in tertiary sector. Energy and Buildings 41 (10),1063e1075.

Lozano, M.A., et al., 2009c. Energy cost analysis of simple trigeneration systemsunder variable operation conditions. In: Proceedings of the 22nd InternationalConference on Efficiency, Cost, Optimization, Simulation and EnvironmentalImpact of Energy Systems Foz do Iguaçu, Brazil.

Maglorie, P., et al., 2002. Economie d’énergie en trigénération. International Journalof Thermal Sciences 41, 1151e1159.

Marimón, M.A., Arias, J., Lundqvist, P., Bruno, J.C., Coronas, A., 2011. Integration oftrigeneration in an indirect cascade refrigeration system in supermarkets.Energy and Buildings 43 (6), 1427e1434.

Möller, A., 2010. Material and energy flow-based cost accounting. ChemicalEngineering & Technology 33 (4), 567e572.

Phylipsen, G., et al., 1998. Handbook on International Comparisons of EnergyEfficiency in the Manufacturing Industry. Utrecht University Press, theNetherlands.

Pokorný, J., 2006. Database architectures: current trends and their relationships toenvironmental data management. Environmental Modelling & Software 21 (11),1579e1586.

Porter, M., van der Linde, C., 1995. Green and competitive: ending the stalemate.Harvard Business Review 73, 120e134.

Rosen, M., 2008. Allocating carbon dioxide emissions from cogeneration systems:descriptions of selected output-based methods. Journal of Cleaner Production16 (2), 171e177.

SAIC e Scientific Applications International Corporation, 2006. Life Cycle Assess-ment: Principles and Practice. National Risk Management Research Laboratory,U.S. Environmental Protection Agency, Cincinnati, Ohio, US.

Serra, L.M., et al., 2009. Polygeneration and efficient use of natural resources.Energy 34 (5), 575e586.

Simapro, 2008. Life Cycle Assessment Software. PRé Consultants, Amersfoort, TheNetherlands. <http://www.pre.nl/simapro> (accessed 01.11.11).

Summerheat Project, 2008. Cooling in Summer with Heat from CHP Technologies.Euro Heat & Power English Edition, vol. 5. <http://www.eu-summerheat.net>.http://www.eu-summerheat.net/, III/2008, pp. 22e25.

TRIGEMED, 2003. Promotion of Tri-Generation Technologies in the TertiarySector in Mediterranean Countries. Survey Report, SAVE Project. <http://www.managenergy.net/resources/368>. http://www.managenergy.net/resources/368.

Turner, G.M., Hoffman, R., McInnis, B.C., Poldy, F., Foran, B., 2011. A tool forstrategic biophysical assessment of a national economy e the Australianstocks and flows framework. Environmental Modelling & Software 26 (9),1134e1149.

Umberto, version 5.0, ifu Hamburg GmbH and ifeu GmbH, Hamburg. <http://www.umberto.de/en>, 2006 Germany (accessed 01.11.11).

Viaggi, D., Bartolini, F., Raggi, M., 2009. Combining linear programming and prin-cipaleagent models: an example from environmental regulation in agriculture.Environmental Modelling & Software 24 (6), 703e710.

Viere, T., et al., 2010. J. Verbund-simulation e strategic planning and optimization ofintegrated production networks. Chemical Engineering & Technology 33 (4),582e588.

Wang, J.-J., Jing, Y.-Y., Zhang, C.-F., Shi, G.-H., Zhang, X.-T., 2008. A fuzzymulti-criteria decision-making model for trigeneration system. EnergyPolicy 36 (10), 3823e3832.

Wohlgemuth, V., et al., 2006. Combining discrete event simulation and materialflow analysis in a component-based approach to industrial environmentalprotection. Environmental Modelling & Software 21 (11), 1607e1617.

Xu, Z., Coors, V., 2011. Combining system dynamics model, GIS and 3D visualizationin sustainability assessment of urban residential development. Building andEnvironment 47, 272e287.