OPTIMIZATION FOR COGENERATION SYSTEMS IN BUILDINGS BASED ON LIFE CYCLE ASSESSMENT

OPTIMIZATION FOR COGENERATION SYSTEMS IN BUILDINGS BASED ON LIFE CYCLE ASSESSMENT SUBMITTED: Marc 2005 REVISED: January 2006 PUBLISHED: May 2006 at ttp://itcon.org/2006/20/ EDITOR: Dana J. Vanier Ayat E. Osman, PD. Candidate University of Pittsburg, Department of Civil and Environmental Engineering email: aeo2@pitt.edu Robert Ries, Assistant Professor University of Pittsburg, Department of Civil and Environmental Engineering email: robries@pitt.edu SUMMARY: Tis paper presents a model tat is developed to optimize te selection and operation of energy systems in commercial buildings based on teir environmental performance. Te model can be used for decision support regarding infrastructure in bot design and operation of building energy systems. Te approac is composed of energy simulation to generate building s energy demand, life cycle assessment (LCA) to model different energy systems, and optimization model to optimize te selection and operation of tese energy systems. Te energy systems tat are discussed in tis paper are cogeneration systems, average electric grid, gas boilers, and absorption and electric cillers. Te performance criteria presented in tis paper are primary energy consumption (PEC) and troposperic ozone precursor potential (TOPP). KEYWORDS: cogeneration, life cycle assessment, optimization. 1. INTRODUCTION Commercial buildings in te US are one of te largest sectors tat purcase electricity. On average, te commercial and residential sectors pay significantly more for electricity tan te industrial sector, mainly because some of te industrial sectors generate teir own electricity (A. D. Little, 2000). Te significant environmental impacts resulting from te operation of buildings ave been reported by te U.S. Department of Energy: building operations represent approximately 40% of te annual total energy consumption in te United States wit approximately equal proportion of te annual related carbon dioxide production (37%) (U.S. DOE, 2003). Approximately 2236.2 million Metric tones (MMT) of carbon dioxide emissions from residential (1,212.0 MMT) and commercial (1,024.2 MMT) is due to U.S. energy-related carbon dioxide emissions in 2004 (EIA, 2004). Carbon dioxide is one of te most important greenouse gases (GHG): -- gases tat trap te solar eat close to te Eart s surface and absorb infrared radiation at particular wavelengt. Te Intergovernmental Panel on Climate Cange (IPCC) developed te global warming potential (GWP) concept to measure te impacts of different GHG on global warming by normalizing tese gases relative to carbon dioxide based on teir radiative forcing, wic is given te equivalence of 1 unit (IPPC, 2001). Oter GHG tat contribute to GWP are: (a) nitrogen oxides, totalling 1130 tousand metric tons (TMT) carbon dioxide equivalent (815 TMT originate from residential energy related activities and 315 TMT carbon dioxide equivalent from commercial sources); and (b) metane, totalling 7087 TMT carbon dioxide equivalent (6,968 TMT carbon dioxide equivalent from residential sources and 119 TMT carbon dioxide equivalent from commercial sources) (EIA, 2004). Terefore by summing tese numbers, te total GWP resulting from residential and commercial energy related activities considering carbon dioxide, nitrogen oxides and metane emissions is 2244 MMT carbon dioxide equivalent. In commercial buildings, natural gas is recognized as te principal fuel for space and water eating, wit electricity generated off-site used for cooling loads, ligting needs, office and oter equipment. Using natural gas-fired cogeneration systems will allow commercial buildings not only to generate teir own electricity but also to use energy efficiently because tese tecnologies can utilize te oterwise wasted termal energy for a variety of purposes, suc as space and water eating as well as cooling wit absorption cillers. In addition to energy efficiency tat results in lower primary energy resource consumption, natural gas-fired cogeneration ITcon Vol. 11 (2006), Osman and Ries, pg. 269

systems emit fewer pollutants tan conventional coal- and oil-fired systems. Tis is because natural gas as a lower sulfur, nitrogen, and carbon content tan coal resulting in lower emissions. Management of cogeneration systems is usually performed to minimize costs. Minimal consideration is given to optimizing te management of tese systems wit regards to environmental criteria. A study on energy management strategies for existing cogeneration systems sowed tat tese systems were not profitable due to a lack of a strong energy management strategy and te systems operation was constrained by te very complex electricity utility cost rate strategy (Benelmir and Fedit, 1998). On te oter and, it is necessary to manage energy systems based on environmental criteria to design sustainable tecnologies tat minimize te global impacts resulting from systems operation. Te complexity in operating cogeneration systems, especially wen considering integrating tem wit conventional systems suc as te electric grid and gas boilers, arises from te fact tat tere are many options to size and operate tese systems, depending on teir respective caracteristics, as well as te energy use profile of a building. Cogeneration systems can be sized to meet te annual peak termal or electric demand and can be operated to follow te termal load of te building, te electric load of te building, or periodically adjusted to follow eiter te electric and termal load depending on te objective of te user. Several of te studies relating to te application of combined eat and power (CHP) tecnologies were performed to investigate operational strategies mainly from te efficiency and economic perspectives (Arivalagan et al, 1995; Few et al, 1997; Fawkes et al, 1998; Jones, 1999; Brandon and Snoek, 2000; Gunes, 2001; Marantan et al, 2002; Jalalzade et al, 2002; Yodovard et al, 2001; Ellis and Gunes, 2002). Operations Researc (OR) tecniques were used to optimize te operation of utility plants to minimize operating costs or to maximize revenue (O Brien and Bansal, 2000). Te focuses of te studies on cogeneration systems ave been limited to gas and steam turbines (Marecal and Kalitventzeff, 1998; Venkates and Cankong, 1995); owever, one study investigated te optimization of solid oxide fuel cell and gas turbine combined cycle wit regards to cost and emission rates (Burer et al, 2003). Witin tose studies te environmental impacts tat were addressed were limited to te assessment of te carbon dioxide emissions resulting from te operation of tese systems (Cung et al, 1997; Wu and Rosen, 1999; and Burer et al, 2003). Te life cycle assessment (LCA) framework provides a tool to understand and analyze te performance of tese energy systems by considering te different products life stages and teir impact on te environment (ANSI/ISO, 1997). Te environmental impact can be global suc as greenouse gases, regional, suc as acid rain, or local, suc as smog formation. In addition, te efficiencies of tese energy systems directly impact te consumption of primary energy resources. Terefore, in ligt of te necessity for te sustainable use of energy resources and cleaner environment, te understanding of te environmental impact associated wit te production of energy and te application of tese energy systems in buildings is important in building design. Few LCA studies in te literature address te application of cogeneration systems in buildings. Te LCA studies include te assessment of te environmental impacts resulting from various electric generation systems (Micaelis, 1998; Gagnon et al, 2002), natural gas combined cycle (NGCC) systems (Spat and Mann, 2000; Lombardi, 2003), and solid oxide fuel cell cogeneration system (Pent, 2003). Te model presented in tis paper intends to optimize te selection and operation of energy systems based on teir potential life cycle environmental impacts by integrating LCA and OR tecniques. By using a ypotetical case study of a commercial office building, te analysis of te results sowed tat te model is appropriate for infrastructure decision support in bot te design and operation of building energy systems. For design, te model can be used to determine energy system strategies. For design and evaluation of te operation pase, te model can be used to ascertain te most efficient operating strategy to be implemented based upon predicted energy use from ourly building energy simulation data. Tis improves system performance wen compared to a fixed operating strategy suc as wen following te termal or electric load of a building. Te model could also be useful for making operational decisions wen predictions of sort-term expected loads are fairly well known. Te model can also be used to assess alternative building energy use reduction efforts regarding electrical and termal energy and to determine wic alternative would ave te best result for te effort invested. Lastly, te model can be used for single building energy systems or tose tat are designed to serve te energy requirements of multiple buildings in an area, suc as campus or municipal energy systems. ITcon Vol. 11 (2006), Osman and Ries, pg. 270

2. APPROACH A life cycle optimization model is developed in order to consider environmental performance in te design of building operations. Te approac consists of te following stages: Energy simulation, Life cycle assessment and Optimization FIG. 1: Illustration of life cycle energy optimization model. Te first stage, energy simulation, is used to define te building s caracteristics and determine te building energy use profile. Te second stage, life cycle assessment, is used to develop energy systems models tat are used in meeting te building s energy demand. Te emissions results obtained from te LCA model are used as coefficients of te decision variables in te optimization model. Te tird stage, optimization, is used to determine te optimum energy systems and operational strategies used to meet building s energy demand. Depending on te user s objective, te optimization model could be used to acieve a single objective or a combination of objectives, suc as minimizing primary energy consumption and emissions. Fig. 1 sows an illustration of te study approac. 2.1 Energy Simulation Energy simulation is used to obtain te ourly eating, cooling, and electrical loads of a building. Energy simulation allows te user to define a number of building caracteristics, suc as: ITcon Vol. 11 (2006), Osman and Ries, pg. 271

building geometry and construction materials, location wit specific weater caracteristics, building size, building use, specific occupancy caracteristics, equipment use scedules and ligting use scedules. Life cycle assessment framework Goal and scope definition Inventory analysis Impact Analysis Interpretation Direct applications Product development and improvement Strategic planning Public policy making Marketing Oter FIG. 2: Illustration of life cycle assessment framework (ANSI/ISO, 1997). Energy simulation software is ten used to generate te building ourly energy use tat matces te building s caracteristics as defined by te user. In addition, energy-efficient strategies can be defined, suc as day-ligting wit associated dimming of artificial ligts, using energy-efficient ligts, improving insulation trougout, improving windows, reducing infiltration, incorporating passive solar eating, sading windows, adding termal mass, installing iger efficiency HVAC, relocating ducts to inside te termal envelope, enancing HVAC controls, and using an economizer cycle (SBIC, 1996). Energy use results can also be obtained in annually and montly basis. Tere are a number of energy simulation software packages available tat can be used to generate te electrical and termal demand profiles at te required time step. Energy-10 (SBIC, 1996) as been used in te current model implementation. Te ourly eating, cooling, and electrical loads of te building become parameters in te optimization model. 2.2 Life Cycle Assessment Te LCA model is developed following te International Organization for Standardization (ISO) framework (ANSI/ISO, 1997). LCA studies te environmental aspects and potential impacts trougout a product s life (i.e., cradle-to-grave) from raw material acquisition troug production, use and disposal. According to te International Standards, te LCA pases include definition of goal and scope, inventory analysis, impact assessment, and interpretation, as sown in Fig. 2. 2.2.1 Goal of te study: Te goal of tis study is to create LCA models for four types of building energy systems: grid-based energy systems, cooling systems, eating systems and cogeneration systems. Tese models assess te life cycle environmental impact of te production of energy for buildings. Te outputs of tese models are also used in te optimisation model. ITcon Vol. 11 (2006), Osman and Ries, pg. 272

2.2.2 Scope of te study: Te scope of te study covers te following product systems: Grid-based energy systems (two systems are modelled to supply electricity for a building): 1. US average electric grid: te electric generation mix in te US supplied by te grid is modelled as follows: 53% coal, 17% natural gas, 17% nuclear, 9% ydro, 2% oil, 2% waste, 0.4% geotermal and 0.15% wind (IEA, 1998). An average grid loss of 6.5% is considered in te process; te total efficiency of te electric generation process is 32% (EIA, 2002). 2. Natural gas combined cycle (NGCC): A 500-MW NGCC power plant wit 49% electrical efficiency is used in te scenarios to represent te best available central generation tecnology. Specifications and assumptions are acquired from a life cycle assessment study of a natural gas combined cycle power generation system (Spat and Mann, 2000). Te plant configuration consists of two gas turbines, a tree pressure eat recovery steam generator, and a condensing reeat steam turbine. Natural gas is fed into a gas turbine wic drives te generator. Waste eat from te turbine is captured by te eat recovery steam generator wic provides steam for te steam turbine wic in turn also drives a generator (Spat and Mann, 2000). In suc a system, usually two tirds of te electric power is provided by te gas turbine and one tird by te steam turbine (Hay, 2000). Emissions from te NGCC process are obtained from EPA's AP-42 emission factors for gas turbines (EPA, 1995). Cooling systems (two systems are modelled to supply cooling for a building): 1. Absorption ciller (AC): A 1.5-MW two-stage absorption ciller uses water as te refrigerant and litium bromide as te absorbent. Te AC is driven by eat. Operation data from a commercially available absorption ciller, (York, 1997), is used to create te process model. Auxiliary energy required for te operation of te AC is supplied by U.S. average electric grid. Te coefficient of performance (COP) of te AC is 1.05. 2. Electric ciller (EC): A 196-kW EC is modelled. Electric cillers provide cilled water for all air conditioning applications tat use central station air andling or terminal units and are driven by electricity. Operation data from a commercially available absorption ciller (York, 1999), is used to create te process model. Te COP of te EC is 4.6. Heating systems: A 1-MW natural gas-fired boiler wit efficiency of 88.7% is modeled to supply te required eating demand in a building. Te emissions from te boiler are acquired from EPA s AP-42 for gas boiler (EPA, 1995). Cogeneration systems: tree cogeneration systems are modelled to supply te electrical and/or termal requirements of a building): 1. Microturbine (MT): Te MT modelled is a 60-kWe natural gas-fired MT unit wit 80% overall efficiency of fuel input, 28% of wic is electrical and 52% is termal. Process efficiencies and emissions at part load operations are also used in developing te microturbine model adapted from a Capstone microturbine (GHG, 2003). Microturbines are just emerging as a future distributed resource tat will be ideally sized to meet te electric load profiles of many commercial and institutional end-users. Tey are mostly run on natural gas and exaust eat can be recovered for ot water or steam loads. 2. Internal combustion engine (ICE): Commercially available reciprocating engines for power generation range from 0.5-kW to 6.5-MW. Different size ranges are modelled including process efficiencies and emissions at part load operations adapted from Caterpillar gas engines (Caterpillar Inc., 1999). Natural gas reciprocating engines are a popular coice for commercial combined eat and power applications due to teir good part-load operation and availability of size ranges tat matc te load of many commercial and institutional end-users. System overall efficiency is about 88% of fuel input, 33% of wic is electrical and 55% is termal. Steam or ot water can be generated from recovered eat tat is typically used for space eating, reeat, domestic ot water and absorption cooling. 3. Solid oxide fuel cell (SOFC): Anoter emerging tecnology for combined eat and power application is te SOFC. A 110-kWe SOFC wit 80% overall efficiency of fuel input, 47% of wic is electrical and 26-33% is termal, is used in te model. Te difference in termal ITcon Vol. 11 (2006), Osman and Ries, pg. 273

efficiency between te 26% and 33% SOFC is tat te 26%-termal SOFC uses part of te generated eat to provide eat for te fuel reformation process. Te SOFC process model is adapted from Siemens Westingouse SOFC (Bessette et al, 2001). Natural gas is reformed to ydrogen gas witout loss. Wit exaust temperature of up to 600 F, steam or ot water can be generated from recovered eat tat is typically used for space eating, reeat, domestic ot water and absorption cooling. 2.2.3 Functional units and system boundaries Te functional unit: Te functional unit is a measure of performance of te functional outputs of te product system. Te primary purpose of a functional unit is to provide a reference to wic te inputs and outputs are related (ANSI/ISO, 1997). Te functional unit used in tis study is 1-kW of energy consumption. Te product system boundaries: Te system boundaries determine wic unit processes sall be included witin te LCA (ANSI/ISO, 1997). System boundaries considered in tis study are: elementary flow at system boundaries and te defined tecnology specifications. Fig. 3 sows a scematic of a typical product system used in developing te LCA model. 2.2.4 Life cycle inventory analysis Te inventory analysis involves data collection and calculation procedures to quantify relevant inputs and outputs of te product system. Tese data also constitute te input to te life cycle impact assessment (ANSI/ISO, 1997). Te major product systems constituting te production and use of energy generation investigated in tis study include unit processes tat are linked to one anoter by product flows across te systems boundaries, suc as energy flow, intermediate product flows witin te systems boundaries, suc as auxiliary energy flow, and elementary flows to te environment, entering a unit process, suc as natural gas or leaving a unit process, suc as air emissions and water effluents. LCA software, Global Emission Model for Integrated Systems (GEMIS) (Fritsce and Scmidt, 2003), is used to design te LCA models and generate te emissions resulting from te production and use of energy in te case building under study. In addition, GEMIS database is used in modelling te upstream processes, suc as fuel and material exploration, production, and distribution. In developing te LCA model, processes are defined wic converts, transports, or produces a product, for example a resource, suc as natural gas is converted to electricity by linking all te processes involved in te extraction, transportation, and conversion to electricity. 2.2.5 Life cycle impact assessment Te impact assessment step of te LCA evaluates te significances of potential environmental impacts using te results of te life cycle inventory analysis. Te environmental impact indicators cosen to quantify te potential contribution of te products inventory flow are: Primary Energy Consumption, Global Warming Potential (GWP), Troposperic Ozone Precursor Potential (TOPP) and Acidification Potential (AP). Primary energy consumption is a quantitative measure of te total amount of primary energy resources needed to deliver energy. Resources are products tat can be converted to energy carriers e.g. oil and coal from wic fuels can be derived, wind, ydro-power etc. Tis impact addresses only te depletion effect of resource extraction, i.e. te upper end of te process cains, and not impacts resulting from extraction processes, suc as emissions. Te impact of primary energy use determines te availability of natural resources, wic translates to issues suc as efficiency, conservation, sustainable energy use, etc. ITcon Vol. 11 (2006), Osman and Ries, pg. 274

FIG. 3: Energy flow in a typical product system GWP is te mass-based equivalent of te radiative forcing of green ouse gases (GHG), based on te specific forcing of CO 2. It is expressed in CO 2 equivalents. Because GHG, suc as metane and carbon dioxide, ave different atmosperic residence times, te GWP is determined as an integral over a period of time; usually, GWP data refer to a time orizon of 100 years. Altoug trends in levels of te GHG are well known, teir effects on global temperature and climate are muc less certain. Most computer models predict global warming of 1.5-5 C; suc warming would ave profound effects on rainfall, plant growt, and sea levels wic migt rise as muc as 0.5-1.5 meters (Manaan, 1994). TOPP is te mass-based equivalent of te ozone formation rate from precursors, measured in ozone precursor equivalents. Te TOPP represents te potential formation of near-ground (troposperic) ozone wic can cause summer potocemical smog. Altoug not a great treat to te global atmospere, smog does pose significant azards to living tings and materials in local urban areas. Ozone, wic serves an essential protective function in te stratospere, is te major cause in te troposperic smog. Surface ozone levels are used as a measure of smog. Ozone poto toxicity raises a particular concern in respect to trees and crops. In addition, ozone is responsible for most of te uman respiratory system distress and eye irritation resulting from exposure to smog, for instance, breating is impaired at ozone levels at about 0.1 ppm (Manaan, 1994). Acidification potential (AP) is te result of aggregating acid air emissions, expressed in SO 2 equivalents. Te SO 2 equivalents express te acidification potential (AP) and are calculated from te molecular weigts and te proton binding potential of te respective emissions (by definition AP is equal to one for SO 2 ). Acid rain spreads out over several undred to several tousand kilometres; tis classifies it as a regional air pollution problem compared to a local air pollution problem, smog, or a global one, suc as greenouse gases. Emissions from industrial operations and fossil fuel combustion are te major sources of acid-forming gases. Some of te impacts of acid rain are: direct pytotoxicity to plants from excessive acid concentrations, destruction of sensitive forests, respiratory effects on umans and animals, acidification of lake water wit toxic effects to lakeflora and fauna, and corrosion to exposed structures, electrical relays, equipment, and ornamental materials especially tose made of limestone (Manaan, 1994). ITcon Vol. 11 (2006), Osman and Ries, pg. 275

Te following environmental impact indicators, GWP, TOPP, and AP equivalents, are calculated as follows: Indicator equivalenc e [ e Indicator ] = i i were, e i = mass of emission (i) in kg, and Indicator i = environmental impact indicator of emission (i), in [kg/kg] Table 1 sows te emission equivalents tat express GWP, TOPP, and AP. Life cycle (LC) emission factors are representative values tat attempt to relate te quantity of a pollutant released to te atmospere wit an activity associated wit te release of tat pollutant. In tis study, te emission factors are expressed as te weigt of te pollutant in terms of te environmental impact indicators divided by a unit (kw) consumed. Tese emission factors are ten used as coefficients of te decision variables in te objective function of te optimization model. 2.3 Optimization 2.3.1 Model Description Te optimization model developed in tis study allows for integrating cogeneration systems wit a grid-based electric generation systems, as well as, eating and cooling systems. Te optimization model results in a mixed integer linear programming (MILP) problem. Wen solved, te values of te decision variables represent te optimum operational strategy according to a linear objective function subject to te specified constraints. Variables are composed of continuous and binary variables. Continuous decision variables are used in te formulation of equipment performance caracteristics, energy balance, and supply-demand relationsips. Binary variables (0-1 variables) are used to determine if a particular cogeneration unit is used at a certain time or not. Te binary variables also ensure tat te selected cogeneration unit will only operate at a particular part load level at a certain time. Linear equations are used to formulate te constraints describing te correlations between te capacities and efficiencies of energy systems corresponding to teir rated and part load operation. Te objective function of te optimization problem is formulated by using continuous decision variables for energy supply and te emission factors as coefficients of te variables for te energy systems considered. Table 2 sows te parameters and decision variables used in te formulation of te optimization model. 2.3.2 Objective Function Te objective function is formulated to minimize te total LC emissions expressed in kg (e.g. kg of CO 2 equivalents wen minimizing LC GWP). Te LC emissions include te emissions from te life cycle of te processes from resource extraction, production, and operation. (1) [ ] n P _ GRID EF _ GRID + + [ ] COGEN EF _ COGEN H _ B EF _ BOILER 24 Minimize (2) = 1 u= 1 p P up up TABLE 1:Emission equivalents for CO 2 (IPCC, 1996-a), TOPP (EEA, 2000), and SO 2 (EEA, 2000). Emission Equivalents CO 2 CH 4 N 2 O NOx NMVOC CO SO 2 HCL CO 2 equivalents 1 21 316 - - - - - TOPP equivalents - 0.014-1.22 1.0 0.11 - - SO 2 equivalents - - - 0.696 - - 1.0 0.878 ITcon Vol. 11 (2006), Osman and Ries, pg. 276

TABLE 2: Parameters and decision variables for optimization model. Parameters Description Index number for operating ours, = 1, 2,..., 24. p u Index number for part load operating level of a particular cogeneration unit. For MT units p = 25, 50, 75, 100 and for te ICE units p= 50, 75, 100. Index number for cogeneration units, u = 1, 2,..., n, were n is te total number of cogeneration units. C Cooling required for space cooling at our. COGEN_MaxCap up Te maximum electric generation capacity from a cogeneration unit u operating at part load level p at our. COGEN_MinCap up Te minimum electric generation capacity from a cogeneration unit u operating at part load level p at our. EF_BOILER LC emission factor resulting from generating 1-kW of termal energy from a gas boiler. EF_COGEN up Life Cycle (LC) emission factor resulting from a cogeneration unit u operating at part load level p to generate 1- kw of electric energy. EF_GRID LC emission factor resulting from obtaining 1-kW of electricity from te grid. H Heating required for space and water eating at our. P Power required for miscellaneous electric demand, oter tan cooling, at our. PH_RATIO up Power to eat efficiency ratio of a cogeneration unit u operating at part load level p. Decision Variables C_AC Cooling obtained from te absorption ciller at our. C_EC Cooling obtained from te electric ciller at our. COGEN up Binary variable for cogeneration unit u operating at part load level p at our, 1 if unit is in use { 0 oterwise H_AC Heating required to drive te absorption ciller to supply te cooling demand at our. H_B Heating obtained from te gas boiler at our. H_COGEN up Heat obtained from cogeneration unit u operating at part load level p at our. H_EXCESS Excess eat remaining after all te eating requirements are met at our. P_COGEN up Power obtained from cogeneration unit u operating at part load level p at our. P_EC Power required to drive te electric ciller to supply te cooling demand at our. P_EXCESS Excess power remaining after all te power requirements are met at our. P_GRID Power obtained from te grid at our. 2.3.3 Constraints: Energy balance and supply-demand Te building s power demand consisting of power required for miscellaneous office equipment and ligts, and power required for cooling if an electric ciller is used, must be satisfied eac our. Eac our, power can be supplied by te grid and/or specific cogeneration unit(s) operating at a particular load. n P _ GRID + P _ COGEN P + P _ EC (3) u = 1 p P up Te building s termal demand, consisting of termal energy required for space and water eating, and termal energy required for cooling if an absorption ciller is used, must be satisfied eac our. At a particular our, termal energy can be supplied by a gas boiler and/or specific cogeneration unit(s) operating at a particular load. n H _ B + H _ COGEN H + H _ AC (4) u = 1 p P up ITcon Vol. 11 (2006), Osman and Ries, pg. 277

Te building s cooling demand must be satisfied eac our. At a particular our cooling can be supplied by an absorption ciller and/or an electric ciller. C _ AC + C _ EC = C (5) Te ourly excess electrical energy is defined as: n P _ EXCESS = P _ GRID + P _ COGEN [ P + P _ EC ] (6) u= 1 p P up Te ourly excess termal energy is defined as: n H _ EXCESS = H _ B + H _ COGEN u= 1 p P up [ H + H _ AC ] Performance caracteristics of energy systems Eac our, te electric energy generated from a cogeneration unit operating at a particular part load level is equal to te product of te termal energy required by te unit and te power to eat ratio of tat unit. P _ COGEN = H _ COGEN PH _ RATIO up up up (8) Eac our, te electric energy obtained from a cogeneration unit operating at a particular part load level is greater tan or equal to te minimum capacity of tat unit and less tan or equal to te maximum capacity of tat unit. P _ COGEN COGEN _ MinCap COGEN up up up (9) P _ COGEN COGEN _ MaxCap COGEN up up up Eac our, te cooling energy obtained from an absorption ciller is equal to te product of te termal energy required and te coefficient of performance (COP) of te absorption ciller. C _ AC = H _ AC AC (10) COP Eac our, te cooling energy obtained from an electric ciller is equal to te product of te electric energy required and te COP of te electric ciller. C _ EC = P _ EC EC (11) COP 3. ASSUMPTIONS AND LIMITATIONS Main assumptions made in te study are: Termal and electric conversion efficiencies of te cogeneration systems are acievable. Cogeneration systems are capable of following a specific termal or electrical load of te building. Te termal and electric energy produced from a cogeneration process is of utilizable quality. No eat or electric loses from a cogeneration process are considered oter tan tose captured by te conversion efficiencies. No credit is taken for any electrical energy generated above te demand. Some of te limitations of tis study are: Te current study is a ypotetical case wic migt not apply to real world scenario. Environmental impact indicators used in tis study are not representative of compreensive environmental impact analysis but represent a class of widely used environmental parameters, wic could be used for comparative analysis wit previous and future studies. A compreensive environmental impact analysis, including economic impacts, would be more valuable if te study was done on an actual setting; owever, because te current study is done on a ypotetical building, te results could provide a general understanding of te performance of energy systems in buildings and ways to minimize te environmental impacts of teir use. Some of te principal environmental impact indicators not addressed in tis study are economic impacts, uman toxicity, ecological toxicity, particulates formation and indoor air quality. (7) ITcon Vol. 11 (2006), Osman and Ries, pg. 278

Economic implications are not considered in te analysis. Cost analysis is not performed at tis stage of te study, wic is a key determinant in real world application of cogeneration systems. 4. CASE STUDY 4.1 Problem Formulation Te case building is a 100,000 square foot commercial office building. Te climate cosen as average cooling degree days (CDD) less tan 2000 and eating degree days (HDD) less tan 5500. Design caracteristics for te case building are based on U.S. average construction data obtained from te literature (Sezgan et al, 1995). Some of te main building caracteristics are: Wall materials: 8-inc masonry, rigid insulation, and gypsum board, Roof construction: flat, build-up roofing, rigid insulation, and gypsum board ceiling and Windows: double glazed wit aluminium frames In tis paper, energy use for te case building is presented for a typical day in August. Te termal load of te building consists mainly of domestic water eating, te electric load consists of electricity required for miscellaneous electric equipment and ligting. Cooling demand can be added to te termal load if an absorption ciller is used or can be added to te electric load if an electric ciller is used. Termal and electrical energy storage systems are not considered. Te optimization model considers ten 60-kWe MT units, average electric grid power, a gas boiler, and absorption and electric cillers. Eac MT unit is modeled to operate at four part load levels: 25%, 50%, 75%, and 100% load. Anoter problem is formulated to present a base case scenario representing conventional practice, were te only source for electricity is te electric grid, te eating source is te gas boiler, and cooling can be supplied by te AC and/or EC. Te performance caracteristics of te MT unit are given in Table 3. Te efficiency of te average electric grid is 32% and te coefficient of performance (COP) of te AC is 1.05, and te COP of te EC is 4.6. Te emission factors of te energy systems obtained from te LCA model are given in Table 4. TABLE 3: Efficiencies of MT. Electrical Efficiency % Termal Efficiency % Overall Efficiency % Part Load 100% 75% 50% 25% 100% 75% 50% 25% 100% 75% 50% 25% 60-kW MT 28 24.2 20.0 13.1 52 56.4 56.7 58.0 78.4 80.7 76.7 71.1 TABLE 4: Emission factors of energy systems. System Part Load Electric Grid Gas Boiler MT MT MT MT 100% 75% 50% 25% PEC [kw/kw of energy use] 3.09 1.18 3.99 4.32 5.22 7.97 TOPP [kg TOPP Equiv./kW] 0.0035 0.00021 0.00083 0.00081 0.0064 0.0038 GWP [kg CO 2 Equiv./kW] 0.787 0.254 0.749 0.795 1.067 1.479 In tis example two optimization problems are presented. Te first optimization problem is formulated wit te objective to acieve minimum LC primary energy consumption expressed in kw. Te second optimization problem is formulated wit te objective to acieve minimum LC TOPP expressed in kg of ozone precursor potential. Te tird optimization problem is formulated wit te objective to acieve minimum LC GWP expressed in kg of carbon dioxide equivalent. ITcon Vol. 11 (2006), Osman and Ries, pg. 279

4.2 Discussion and Results 4.2.1 First Optimization Problem: minimize total LC primary energy consumption. Te total minimum LC primary energy consumption (PEC) in a typical day in August is found to be 15579 kw. Te optimum operational strategies are found to be: MT units operating at full load are used to supply power primarily and partially power is obtained from te grid during some ours. Refer to Fig. 4. During all ours of te day, cogenerated eat from te MT units is used to meet te eating requirements of te building and te rest of te eat is used for cooling wit AC. Refer to Fig. 5. Cooling is met primarily by AC driven by eat from MT and partially by EC during peak ours, refers to Fig. 6. FIG. 4: Power generation during a typical day in August for minimum PEC. 800 800 700 700 600 600 kw of Heating 500 400 300 kw of Cooling 500 400 300 200 200 100 100 0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324 0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324 H_MT 100% Heating Demand Termal for Cooling Demand Total Termal Demand C_AC C_EC Cooling Demand FIG. 5: Termal energy generation during a typical day in August for minimum PEC. FIG. 6: Cooling energy generation during a typical day in August for minimum PEC. In te base case scenario, te total minimum LC primary energy consumption in a typical day in August is found to be 15759 kw. Electricity is supplied by te electric grid, eating by te gas boiler, and cooling by te electric ciller. Wen comparing te results from te MT optimization problem to te results from te base case scenario, tere is no significant decrease in primary energy consumption, owever, te operation strategies were considerably different sowing te sensitivity of te model towards optimizing energy use by minimizing te LC primary energy consumption. 4.2.2 Second Optimization Problem: minimize total LC TOPP Te total minimum LC TOPP in a typical day in August is found to be tree kg of ozone precursor potential. Te optimum operational strategies are found to be: MT units operating at 75% and 100% load are used to supply power during all ours of te day. Refer to Fig. 7. Te cogenerated eat from te MT units is used to ITcon Vol. 11 (2006), Osman and Ries, pg. 280

meet te eating requirements of te building and te rest of te eat is used for cooling wit AC. Refer to Fig. 8. Cooling is met primarily by AC driven by eat from MT and partially by EC during peak ours, refer to Fig. 9. No power is obtained from te grid at any time. FIG. 7: Power generation during a typical day in August for minimum TOPP. kw of Heating 800 700 600 500 400 300 200 100 0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324 H_BOILER H_MT 75% H_MT 100% Heating Demand Termal for Cooling Demand Total Termal Demand kw of Cooling 800 700 600 500 400 300 200 100 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 C_AC C_EC Cooling Demand FIG. 8: Termal energy generation during a typical day in August for minimum TOPP. FIG. 9: Cooling energy generation during a typical day in August for minimum TOPP. In te base case scenario, te total minimum LC TOPP is 13 kg of ozone precursor potential. Electricity is supplied by te electric grid, eating by te gas boiler, and cooling by AC driven by eat from te gas boiler. Wen comparing te results from te MT optimization problem to te results from te base case scenario, it is found tat 77% reduction in ozone precursor potential can be acieved by using MT cogeneration system instead of conventional systems. 4.2.3 Tird Optimization Problem: minimize total LC GWP Te total minimum LC GWP in a typical day in August is found to be 3043 kg of carbon dioxide equivalent. Te optimum operational strategies are found to be: MT units operating at 100% load are used to supply power during most operating ours of te day (Hour 7-20) and partial power is obtained from te grid during te rest of te day. Te cogenerated eat from te MT units is used to meet te eating requirements of te building and te rest of te eat is used for cooling wit AC. Cooling is met primarily by AC driven by eat from MT and partially by EC during peak ours. ITcon Vol. 11 (2006), Osman and Ries, pg. 281

In te base case scenario, te total minimum LC GWP is 3998 kg of carbon dioxide equivalent. Electricity is supplied by te electric grid, eating by te gas boiler, and cooling by EC. Wen comparing te results from te MT optimization problem to te results from te base case scenario, it is found tat 24% reduction in GWP can be acieved by using MT cogeneration system instead of conventional systems. In addition, results are found to vary according to te user s objectives. Hence, wen designing energy systems, a olistic approac sould be taken to investigate different parameters in order to optimize system selection wit te minimum environmental impacts. In tis example, tradeoffs are seen between using microturbine cogeneration system as opposed to conventional systems. Tis approac can elp te decision maker in designing energy systems in commercial buildings tat would reduce environmental impacts suc as GWP and TOPP rater tat only economical implications. Results sow tat tis model is applicable in te selection and operation of cogeneration systems wile integrating tem wit grid-based electricity, gas boiler, and absorption and electric cillers. Also, te MILP is useful for optimization wit respect to environmental criteria rater tan merely economical objectives, wic is important in designing sustainable energy systems for buildings. Te model is currently being augmented to consider not only environmental but also capital, maintenance, and operational costs of tese energy systems wic will impact te results, especially te number of units selected for operation. 5. CONCLUSION In tis paper an LCA optimization model is presented wic uses environmental criteria for te selection of energy systems and optimization of te operational strategies tat integrate cogeneration systems wit utility energy systems in commercial building applications. Suc an approac could be used for sustainable planning wen designing for optimum energy management in buildings by considering lower primary energy consumption, and emissions wile maximizing processes efficiencies. Te LCA optimization model will be useful for: Selecting energy systems for building applications, Designing operational strategies wile considering system s caracteristics, suc as efficiencies, capacities, and emissions, as well as variable loads of buildings, Analyzing te effects of various termal and electrical energy use in buildings on te performance of energy systems, Control of energy systems in operation and Predicting te environmental life cycle impact resulting from te life cycle of a building s energy systems. Given resource constraints and pollution generation from building energy systems, a model tat can be used in decision making regarding te optimization of impacts from building energy use can be a valuable tool for infrastructure management. 6. REFERENCES A.D. Little. (2000). Opportunities for Micropower and Fuel Cell/Gas Turbine Hybrid Systems in Industrial Applications. Rep. No. 38410, Vol.1: Main Text. Cambridge, MA. ANSI/ISO. (1997). Environment Management-Life Cycle Assessment-Principles and Framework, 14040-1997, NSF International, Ann Arbor, MI. Arivalagan A., Ragavendra B.G., and Rao A.R.K. (1995). Integrated energy optimization model for a cogeneration based energy supply system in te process industry, Electrical Power & Energy Systems, Vol. 17, No. 4, 227-233. Benelmir R. and Feidt M. (1998). Energy Cogeneration Systems and Energy Management Strategy. Energy Conversion and Management, Vol. 39, No. 16-18, 1791-1802. Bessette N.F., Borglum B.P., Scicl H., and Douglas D.S. (2001). Siemens SOFC Tecnology on te way to economic competitiveness. Siemens Power Journal. <ttp://www.siemenswestingouse.com> (July 25, 2002). Brandon P. and Snoek C. (2000). Microturbine Cogeneration. ASHRAE Transactions, Vol. 106, No. 1, 669-74. ITcon Vol. 11 (2006), Osman and Ries, pg. 282

Burer M., Tanaka K., Favrat D., and Yamada K. (2003). Multi-criteria optimization of a district cogeneration plant integrating a solid oxide fuel cell-gas turbine combined cycle, eat pumps and cillers. Energy, Vol. 28, 497-518. Caterpillar Inc. (1999). Gas engine tecnical data: G3304 NA. <ttp://www.caterpillar.com> (September 25, 2004). Cung W., Wu J., and Fuller J.D. (1997). Dynamic energy and environment equilibrium model for te assessment of CO 2 emission control in Canada and te USA. Energy Economics, Vol. 19,103-124. Ellis M.W. and Gunes M. (2002). Status of Fuel Cell Systems for Combined Heat and Power Applications in Buildings. ASHRAE Transactions, Vol. 108, No. 1, 1032-1044. EIA (2002). Montly Energy Review April 2002. Energy Information Administration <ttp://www.eia.doe.gov> (May, 2002). EIA (2004). Emissions of Greenouse Gases in te United States 2004. December 2005. Energy Information Administration Office of Integrated Analysis and Forecasting, U.S. Department of Energy, Wasington, DC. <ttp://www.eia.doe.gov> (January, 2006). EPA (1995). Compilation of air pollutant emission factors. AP-42: Stationary Point and Area Sources, 5t ed., Vol. 1, Researc Triangle Park, N.C., Environmental Protection Agency <ttp://www.epa.gov/ttn/ceif/ap42> (May 15, 2002). EEA (2000). Environmental Signals 2000. Regular Indicator Rep., European Environmental Agency, Copenagen, Denmark. Fawkes S.D., Jacques J.K., Linnell C., and Watkins A.A. (1998). Practical Modeling of Micro Combined Heat and Power Applications for Local Autority District Heating Scemes. International Journal of Energy Researc, Vol.12, 611-629. Few P.C., Smit M.A., and Twidell J.W. (1997). Modelling of a Combined Heat and Power (CHP) Plant Incorporating a Heat Pump for Domestic Use. Energy, Vol. 22, No. 7, 651-659. Fristce U.R. and Scmidt K. (2003). Gemis Manual, Institute of Applied Ecology, Darmstadt, Germany.<ttp://www.oeko.de/service/gemis/en/g4_manual1.pdf> (January 15, 2004). Gagnon L., Belanger C., and Uciyama Y. (2002) Life Cycle Assessment of Electricity Generation Options: Te Status of Researc in Year 2001. Energy Policy, Vol. 30, 1267-78. GHG (2003). Environmental Tecnology Verification Report: Combined Heat and Power at a Commercial Supermarket-Capstone 60 kw Microturbine CHP System. Catswort, CA., Greenouse Gas Tecnology Center Researc Institute, <ttp://www.epa.gov/etv> (June 10, 2004). Gunes M. (2001). Investigation of a Fuel Cell Based Total Energy System for Residential Applications. MS Tesis, Dept. of Mecanical Engineering, Virginia Polytecnic Institute and State University, Blacksburg, VA. Hay N.E. (1988). Guide to Natural Gas Cogeneration. USA: Te Fairmont Press. IEA (1998). World energy outlook: Electricity information, IEA Books, International Energy Agency, Paris. IPCC (1996). Climate cange 1995. Te science of climate cange, J. T. Hougton, L. G. Meiro Filo, B. A. Callander, N. Harris, A. Kattenberg and K. Maskell, eds., Intergovernmental Panel on Climate Cange, Cambridge Univ. Press, Cambridge, U.K. IPPC (2001). Intergovernmental Panel on Climate Cange, Climate Cange 2001: Te Scientific Basis. Cambridge University Press, Cambridge, UK. (www.ipcc.c) Jalalzade-Azar A., Slayzak S., and Ryan J. (2002). Evaluation of Cooling, Heating, and Power (CHP) for Office Buildings. ASHRAE Transactions, Vol. 108, No. 1, 1045-1051. Jones S. (1999). Micro-cogeneration Optimal Design for Service Hot Water Termal Loads. MS Tesis, Dept. of Mecanical Engineering. Georgia Institute of Tecnology, Atlanta. ITcon Vol. 11 (2006), Osman and Ries, pg. 283

Lombardi L. (2003). Life cycle assessment comparison of tecnical solutions of CO 2 emissions reductions in power generation. Energy Conversion and Management, Vol. 44, No. 1, 93-108. Manaan S.E. (1994). Environmental Cemistry, CRC Press LLC, USA. Marantan A., Popovic P., and Radermacer, R. (2002). Te Potential of CHP Tecnology in Commercial Buildings-Caracterizing te CHP Demonstration Building. ASHRAE Transactions, Vol. 108, No. 1, 1025-1031. Marecal F. and Kalitventzeff B. (1998). Process integration: Selection of te optimal utility system. Computers Cem. Engng., Vol. 22, Suppl., S149-S156. Micaelis P. (1998). Life Cycle Assessment of Energy Systems. <ttp://www.rcep.org.uk/studies/energy/98-6067/micaelis.tm> (May 3, 2002). O Brien J. M. and Bansal P. K. (2000). Modelling of cogeneration systems Part 1: istorical perspective, Proc. Of te Institution of Mecanical Engineers, Vol. 214, Part A, 115-124. Osman A. and Ries R. (2004). Life Cycle Impact Analysis of Energy Systems in Buildings. Journal of Infrastructure Systems, 10 (3), 87-97. Pent M. (2003). Life Cycle Analysis of Fuel Systems Components. Handbook of fuel cells-fundamentals, tecnology and applications, W. Vielstic, A. Lamm, and H. A. Gasteiger, eds., Wiley, Cicester, U.K., 1293-1317. Sezgan O., Franconi E.M., Koomey J.G., Greenberg S.E., Afzal A., and Sown L. (1995). Tecnology data caracterizing space conditioning in commercial building: Application to end-use forecasting wit Commend 4.0. Energy Analysis Program, Ernest Orlando Lawerance Berkeley National Laboratory, University of California, Berkeley, Calif. Spat P. and Mann M. (2000). Life Cycle Assessment of a Natural Gas Combined-Cycle Power Generation System. U.S. Dept. of Energy, Office of Scientific and Tecnical Information, Oak Ridge, Tenn. SBIC (1996). Designing low energy buildings: Passive solar strategies and Energy-10 software. Sustainable Buildings Industry Council, Wasington, D.C. U.S. DOE (2003). Annual Energy Review 2003, Energy Information Administration, DOE/EIA-0384, Department of Energy, Wasington, D. C. Venkates B. and Cankong V. (1995). Decision Models for Management of Cogeneration Plants, IEEE Transactions on Power Systems, Vol. 10, No. 3, 1250-1256. Wu J.Y. and Rosen A.M. (1999). Assessing and optimizing te economic and environmental impacts of cogeneration/district energy systems using an energy equilibrium model. Applied Energy, Vol. 62, 141-154. Yodovard P., Kedari J., and Hirunlab J. (2001). Te Potential of Waste Heat Termoelectric Power Generation from Diesel Cycle and Gas Turbine Cogeneration Plants. Energy Sources, Vol. 23, 213-224. York (1997). Millennium YPC Two Stage Direct-Fired Absorption Ciller-Heaters. York International Corporation, York, PA, USA. <ttp://www.york.com> (Marc, 2002). York. (1999). Liquid Cillers: Water Cooled Ciller and Remote Condenser Models. York International Corporation, York, PA, USA. <ttp://www.york.com> (Marc, 2002). ITcon Vol. 11 (2006), Osman and Ries, pg. 284