- Email: [email protected]
A life cycle multi-objective economic and environmental assessment of distributed generation in buildings
Energy Conversion and Management 97 (2015) 420–427
Contents lists available at ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
A life cycle multi-objective economic and environmental assessment of distributed generation in buildings Amir Safaei a,b,⇑, Fausto Freire b, Carlos Henggeler Antunes a,c a
INESC Coimbra, R. Antero de Quental 199, 3000–033 Coimbra, Portugal ADAI-LAETA, Department of Mechanical Engineering, University of Coimbra, Rua Luis Reis Santos, 3030-788 Coimbra, Portugal c Department of Electrical and Computer Engineering, University of Coimbra, 3030–290 Coimbra, Portugal b
a r t i c l e
i n f o
Article history: Received 6 November 2014 Accepted 12 March 2015 Available online 8 April 2015 Keywords: Lifecycle assessment (LCA) Multi objective linear programming (MOLP) Distributed Generation (DG) Cogeneration Solar Pareto frontiers Buildings
a b s t r a c t Distributed generation, namely cogeneration and solar technologies, is expected to play an important role in the future energy supply mix in buildings. This calls for a methodological framework to assess the economic and environmental performance of the building sector when such technologies are employed. A life-cycle model has been developed, combining distributed generation and conventional sources to calculate the cost and environmental impacts of meeting the building energy demand over a defined planning period. Three type of cogeneration technologies, solar photovoltaic and thermal, as well as conventional boilers along with the Portuguese electricity generation mix comprise the energy systems modeled. Pareto optimal frontiers are derived, showing the trade-offs between different types of impacts (non-renewable cumulative energy demand, greenhouse gas emissions, acidification, eutrophication) and cost to meet the energy demand of a commercial building. Our analysis shows that according to the objective to employ distributed generation (reducing cost or environmental impacts), a specific design and operational strategy for the energy systems shall be adopted. The strategies to minimize each type of impact and the associated cost trade-offs by exploring the solutions located on the Pareto optimal frontiers are discussed. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Energy use in the building sector accounts for more than 40 percent of the EU energy consumption [25]. The generation of energy at the point of consumption, employing Distributed Generation (DG) technologies, is pointed out as a key option for promoting energy efficiency and use of renewable sources in building sector [48]. Combined Heat and Power (CHP) technologies and solar technologies, namely Photovoltaic (PV) and Solar Thermal (ST), represent promising onsite generation alternatives and several strategies, including the introduction of financial incentives, have been implemented to further increase the share of DG in the building sector energy mix [46]. A comprehensive view on distributed multi generation by Chicco and Mancarella [11] also underlines this point. The introduction of different types of DG calls for a framework to assess the building economic and environmental performance when such technologies are employed. Some studies have assessed ⇑ Corresponding author at: ADAI-LAETA, Department of Mechanical Engineering, University of Coimbra, Rua Luis Reis Santos, 3030-788 Coimbra, Portugal. E-mail address: [email protected] (A. Safaei). http://dx.doi.org/10.1016/j.enconman.2015.03.048 0196-8904/Ó 2015 Elsevier Ltd. All rights reserved.
DG technologies considering efficiency and economic perspectives, by using techniques other than optimization. Xuan et al. [54] examined the application of a gas-fuelled reciprocating Combined Heat and Power (CHP) in a commercial building in China. Wu and Rosen [53] employed an energy equilibrium model to compare conventional and NG cogeneration-based district energy systems for heating, cooling and electrical services. Marantan et al. [33] demonstrated the potential of tri-generation NG CHP application in commercial buildings. The importance of cogeneration systems for sustainable energy use was underlined in Çakir et al. [8]. Gunes [22] examined the application of fuel cell-based Total Energy System (TES) for residential buildings. Dentice d’Accadia et al. [15] dealt with the application of a small scale fuel cell cogeneration (electrical power <15 kW) to light commercial application users. Another study on fuel cells [17] offered a methodology for assessing the performance of two types of fuel cells in terms of primary energy demand and CO2 emissions. Bhattacharyya and Quoc Thang [5] found that economic feasibility of medium and large scale cogeneration systems is vulnerable to changes in buy-back rate of electricity and investment costs. Mone et al. [36] investigated the underlying factors in economic feasibility of CHP systems using commercially available gas
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427
turbines. Ullah et al. [51] evaluated the application of a single tubular Solid Oxide Fuel Cell (SOFC) with an electrolyte of YttriaStabilized Zirconia ceramic powder. Ranjbar et al. [41] reported the energy and exergy assessments of a trigeneration system based on a SOFC. Raj et al. [40] investigated the effect of temperature, stoichiometry and the degree of humidification on the performance of a planar solid oxide fuel cell. Brandoni et al. [7] assessed the impact of micro-generation technologies on the mid- and longterm sustainability of urban areas. A dynamic energy analysis of a residential building-integrated cogeneration system under different boundary conditions was presented in Rosato et al. [45]. Regardless of the purpose to install DG (e.g. reducing cost or environmental impacts), factors such as characteristics of different types of DG, dynamic energy costs, different types of building energy demand and their variation, varying solar resources, and national policy frameworks to promote each type of DG should be taken into account to decide on the design and operating strategy of DG. Optimization models applied to DG systems typically consider only cogeneration systems, focusing on economic and technical aspects. For instance, Alanne et al. [3] discussed the techno-economic optimization of Stirling engine micro-cogeneration systems in residential buildings. Casisi et al. [10] studied the effect of different economic support policies on the optimal synthesis and operation of a renewable-based distributed energy supply system for an industrial area. Calise et al. [9] dealt with the design and simulation of a prototype of a small-scale solar CHP system based on evacuated flat-plate solar collectors and organic Rankine cycle. Abdelhady et al. [1] discussed the design of a small scale stand-alone solar thermal co-generation plant for an isolated region in Egypt. Akikur et al. [2], analyzed the performance of a cogeneration SOFC system combined with solar energy. Kabalci et al. [28] proposed a smart monitoring system for renewable based distributed energy systems. Bortolini et al. [6] evaluated the technical and economic design of a combined PV and battery energy storage system. Arcuri et al. [4] proposed a combination of heat pumps, absorption chillers and cogeneration systems, discussing the optimal operation strategy of the energy systems maximizing annual short- and long-term economic returns. Cho et al. [12] developed a model to minimize the total cost of energy usage for a building based on energy efficiency constraints for each component. Dorer and Weber [18] assessed the emission performance of residential micro-cogeneration systems with dynamic whole-building simulation programs. Hawkes and Leach [23] developed an equivalent annual cost minimization model to determine the driving factors behind the investment in fuel cell CHP technologies. A sensitivity analysis showed that the results were sensitive to capital cost, energy import/export prices, plant life_ time, and the temporal precision selected for the study [24]. Inan et al. [26] examined the effect of exchange rate on the cogeneration systems fixed and variable costs in Turkey. The performance assessment of various building cogeneration through energy and exergy efficiencies was discussed in Kanoglu and Dincer [29]. Kong et al. [30] developed an energy optimization model for gas turbine cogeneration systems. Mavrotas et al. [34] discussed an optimization framework for energy supply systems in commercial buildings by considering the demand uncertainty. Monteiro et al. [37] developed a model for planning micro-CHP plants in agreement with the Portuguese energy legal framework. Pan et al. [39] presented a real-time optimum operation strategy to improve the efficiency of the cogeneration systems, integrating turbine generator and cooling tower. Yokoyama et al. [55] discussed the optimal design of a gas engine cogeneration system for electricity and hot water supply, using a branch and bound method. In this regard, relatively few recent studies have looked at the combination of renewable and CHP. Such studies include Akikur et al. [2], who examined the performance analysis of a cogeneration
421
system using solar energy and SOFC technology. Ismail et al. [27] discussed the sizing optimization of a system consisting of PV panels, CHP and storage, using genetic algorithms. Rezvan et al. [44] developed a robust optimization model to determine the optimum capacity of DG for buildings in the case of demand uncertainty. Moreover, despite environmental impacts associated with the building sector is a key issue, only some studies incorporate environmental aspects into the design and operation of DG. Ren & Gao [42] developed a single-objective model for the integrated plan and evaluation of DG systems. The model was extended to a multi-objective model in Ren et al. [43], minimizing energy costs and operating CO2 emissions. Lu et al. [31] presented a multiobjective optimization based for energy management of a district site in China. A systematic optimization procedure to select and size a cogeneration plant fuelled by natural gas, evacuated tube solar collectors, and gasified biomass is discussed in Rubio-Maya et al. [46]. Torchio [50] compared a district heating CHP and a distributed generation CHP in terms of energy and economic criteria, as well as CO2 and NOx emissions. Wang et al. [52] proposed a multi-objective optimization for combined cooling, heating and power system driven by solar energy. Such studies mainly assess operating CO2 (or NOx or SO2), and do not adopt Life-Cycle (LC) approaches, whereas any comparison among energy supply options must employ a LC approach [32]. An exception is Osman et al. [38] who looked at both costs andtwo types of environmental impacts (greenhouse gases, tropospheric ozone depletion), taking into account only CHP technologies. The incorporation of environmental aspects into the analysis calls for multi-objective models and techniques in which decisions should be made exploiting the trade-offs between the conflicting axes of evaluation of the merits of distinct solutions, which are made operational by multiple objective functions to be optimized. A model to minimize the Life-Cycle Costs (LCC) of meeting the energy demand (power, heating, cooling) of a commercial building by integrating DG and conventional energy sources was presented in Safaei et al. [47]. Here, the model is extended into a multiobjective linear programming (MOLP) model by introducing four additional environmental impacts as objective functions along with cost. To calculate the environmental impacts of different energy solutions, a Life-Cycle Assessment (LCA) was conducted considering the impacts related to construction and operation of energy systems, as well as the upstream processes related to their fuel input, i.e. NG for CHPs. MOLP techniques are employed to unveil and exploit the trade-off between the competing objectives, i.e. minimizing cost versus minimizing each type of impact category. Section 2 provides a brief overview of the model and its inputs. In Section 3 the Pareto optimal frontiers obtained for cost vis-à-vis environmental impacts are presented. This is followed by the discussions of the results in Section 4. Finally, in Section 5 the main conclusions are drawn. 2. MOLP model The mathematical model was presented in Safaei et al. [47] for cost optimal design and operation of DG in Portuguese commercial buildings. The model incorporates different types of CHP technologies (Micro-Turbines – MT, Internal Combustion Engines – ICE, Solid Oxide Fuel Cells – SOFC), separate production of electricity - Portuguese electricity generation mix in 2011 [20] and heat (onsite boilers), renewable sources (ST and PV), and auxiliary cooling systems (Absorption Chiller – AC, Compression Chiller – CC). The specifications of the energy systems considered and the main assumptions are described below: Solar PV systems: 4 kWp (KW peak) mono-crystalline PV system with lifetime of 30 years and three 2500 W inverters.
422
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427
ST systems: a solar thermal system consisting of flat-plate absorbers with black chrome coating on copper, a 2000 liters steel heat storage, two 40 W pumps, and the necessary piping for the operation of the system. The transportation of the solar system parts and construction works for mounting of the panels are also accounted for. MT cogeneration system: a 60 kWe (kW electricity) MT system with an operating life of six years with four maintenance sessions for this period. The turbine runs on natural gas, has lubricant consumption of 4.4 liters per year, with electrical and thermal efficiencies of 26% and 52% at 100%-load, respectively. SOFC cogeneration system: a tubular cell design 125 kWe natural gas SOFC engine with operating life of 8 years. During this period, the engine has one major overhaul session to change the fuel stack and 7 maintenance sessions each with 0.15 m3 tap water consumption. The rated electrical efficiency and the power to heat ratio are 45% and 1.20 at 100%-load, respectively. ICE cogeneration system: a 172 kWe natural gas engine with a rated electrical efficiency of 33% and power to heat ratio of 0.61 at 100% load. The lifetime of the system is assumed to be 6 years. AC: a 104 kW AC with a coefficient of performance (COP) of 0.7 and lifetime of 20 years. CC: a 15 kW heat pump with a COP of 2.2 and lifetime of 20 years. The emission of refrigerant R134a during operation is taken into account. Conventional boiler: a natural gas boiler with a rated efficiency of 80% and lifetime of 20 years. National grid: Portuguese electricity generation mix in 2011. This base model was applied to a hotel complex located in the city of Coimbra, Portugal. The data for hourly electricity consumption, as well as hourly cooling and heating demands (calculated using building simulation) were used to generate the loadduration-curves of the building. The year was divided into three seasons – Hot (H), Mild (M) and Cold (C) -, each being represented by three load duration curves (for each type of cooling, heating and power demand) with the following block-loads: peak1 (P1), peak2 (P2), high1 (H1), high2 (H2), medium1 (M1), medium2 (M2) and low (L). The block-loads are defined in a way to represent the dynamic cost of electricity. The MOLP model formulation in this article explicitly considers five objective functions to be minimized: LC Equivalent Annual Costs (LCEAC), and LC Environmental impacts (LCEI): nonrenewable Cumulative Energy Demand (CED), Greenhouse Gas (GHG) emissions, acidification, and eutrophication. LCEAC is the cost per year of meeting the building energy demand, comprising the fixed and variable costs. LCEAC is frequently used as a decision making indicator when evaluating investment projects of unequal life spans [19]. Different categories of constraints are included in the model, namely: meeting electricity, heat, and cooling energy demands; capacity constraints to certify that the energy output of the generation units (CHPs, PV, ST, boiler, CC, AC) are within their operation ranges; national policy constraints to guarantee the fulfillment of the EU [16]/8/EC on the promotion of cogeneration systems; and consistency constraints to guarantee the internal coherence of the model. The decision variables determine the number of panels installed for solar systems, and the number of units installed and their output for the other generation units, at each year throughout the planning period. Input parameters to the model include the magnitude of different types of energy demand of the building (power, heating, cooling) at each block-load throughout the planning period, fixed and operating costs of the units, their
environmental impacts and the economic indicators, e.g. interest rate. The full description of the base model, and case study characteristics are discussed in Safaei et al. [47]. The model enables a thorough assessment of the economic implications and environmental impacts associated with meeting the building energy demand. The environmental impacts of the building employing conventional sources of energy (onsite boiler, grid) are estimated, and the design and operating strategy of energy systems combining DG to reduce each type of impact is explored. LCA was used to calculate the LCEI arising from one unit energy output of energy systems based on the CML method [21], as summarized in Tables 1 and 2. The MOLP model enables to derive Pareto optimal frontiers, which display the trade-offs between cost-efficient solutions and the solutions with lower non-renewable CED, GHG, acidification and eutrophication impact values. Each solution consists of a set of energy systems and their respective operation planning to meet the level of cost and environmental impacts determined by the DM. The assumptions to calculate the Pareto optimal frontiers are as follows: Unlike the base model in [47], the capacity of a conventional boiler is considered as a variable, so the optimal solution (according to the defined objective function) gives out the optimal capacity of the boiler and its output throughout the planning period. The model considers exporting electrical power to the grid. The emissions resulting from the exported power are credited by their deduction from the total emissions, since it was assumed that they avoid the generation of electricity (according to the production mix). This has been accounted by the model using the ‘‘avoided-burden approach’’ [13]. To derive the Pareto optimal frontier, an algorithm based on the methodology proposed in Sylva & Crema [49] was implemented in GAMS [35]. 3. Results 3.1. Pareto frontier for cost and CED The Pareto frontier obtained for non-renewable CED and cost objective functions is shown in Fig. 1. The values correspond to the estimated LCEAC and non-renewable CED of meeting the building energy demand for a year. For solution a in Fig. 1, no DG is permitted to be installed, being the set of energy systems composed of boiler (257 kW), CC (120 kW), and the grid (Portuguese mix 2011). Solution b corresponds to the minimum non-renewable CED, being the set of energy systems composed of ICE (172 kW), PV system (80 kW, the maximum capacity due to area restrictions), boiler (25 kW size), AC (208 kW), and the grid (Portuguese mix 2011). The set of energy systems for the most cost-effective solution (solution x at the most extreme left part of the Pareto frontier) is composed of an ICE (172 kW), a boiler with minimal size (5 kW), 1 AC (104 kW), 30 kW of CC, and the grid. Compared to the case employing only conventional systems, i.e. solution a in Fig. 1, employment of ICE at its cost-optimal operating strategy brings about both financial and energy savings. Thus solution a is dominated by the solution x located on the Pareto frontier. Safaei et al. [47] discussed that under certain fuel cost conditions, ICE and ST represent cost-effective DG solutions for the Portuguese commercial building sector. In order to reduce nonrenewable CED, the output of grid and boiler slightly increases, replacing ICE to partly address the building energy demand. To further reduce non-renewable CED, PV systems can be added to the energy mix. The number of PV installations gradually increases along the Pareto frontier as the DM desires for less non-renewable
423
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427 Table 1 Results of LC impact assessment of cogeneration systems—input to the MOLP model.
CED GHG Acidification Eutrophication
MJ/kW he g CO2 eq/kW he g SO2 eq/kW he g PO34 eq/kW he
ICE 100% load
ICE 75% load
ICE 50% load
MT 100% load
MT 75% load
MT 50% load
MT 25% load
SOFC 104% load
SOFC 100% load
SOFC 93% load
SOFC 85% load
SOFC 78% load
SOFC 68% load
SOFC 62% load
13.89 771
15.24 846
17.40 964
17.76 990
19.24 1054
23.07 1402
35.45 1961
10.69 625
10.46 621
9.82 611
9.62 608
9.44 605
9.44 605
9.23 602
0.468
0.494
0.530
0.533
0.558
0.623
0.833
0.280
0.276
0.265
0.262
0.259
0.259
0.256
0.105
0.113
0.123
0.111
0.118
0.137
0.199
0.092
0.090
0.087
0.086
0.085
0.085
0.084
Table 2 Results of LC impact assessment of solar and conventional systems—input to the MOLP model. Grida
PV kW he CED GHG Acidification Eutrophication a
MJ g CO2 eq g SO2 eq g PO34 eq
1.80 90 0.472 0.21
ST
Boiler
kW hth 4.70 350 0.78 0.54
0.39 21 0.158 0.10
5.75 320 0.168 0.04
Modeled according to Portuguese electricity generation mix in 2011 [20].
CED; the solution with minimum CED (solution b) employs an 80 kW PV system. Figs. 2 and 3 show the optimal configuration of the energy systems to minimize non-renewable CED. The X-axis shows the demand block-loads (peak, high, medium, low) throughout a year of the planning period for three hot (H), mild (M) and cold (C) seasons. ICE operates to meet the building base-load and remaining demand is met via grid and boiler. The cooling demand is completely met by 2 ACs that run on the waste heat from ICE and boiler. 3.2. Pareto frontier for cost and GHG
CED (GJ)
Fig. 4 shows the Pareto frontier obtained for GHG and cost objective functions. The format of the Pareto frontier, and the types of energy systems along solutions located on the Pareto frontier are similar to the Pareto obtained for cost and non-renewable CED (Fig. 1). The employment of ICE can therefore provide both economic and GHG savings for the building under study. The set of energy systems pertaining to the solutions a (conventional) and x (least cost solution) were described in Section 3.1. Solution c minimizes GHG emissions, being the set of energy systems composed of an ICE (172 kW), PV systems (80 kW), a boiler (170 kW size), AC (208 kW), and the grid (Portuguese mix 2011).
20000 18000 16000 14000 12000 10000 8000 6000 4000 2000 0 395
α
ω
β
400
405
410
415
420
LCEAC (k€) Fig. 1. Cost vs. CED Pareto frontier.
425
430
435
The set of energy systems and their operation planning for the solution with minimum cost, solution x, was explained in Section 3.1. In order to reduce GHG, the output of ICE increases to replace the electricity imported from the grid. As Fig. 4 displays, the slope of the frontier declines toward solutions with higher costs. This signifies that for such solutions it bears more cost to reduce GHG by one unit, since ICE is not operating according to its cost optimal design, being rather operating to decrease GHG emissions even in the block-loads in which its operation is not economical. Higher production by ICE directly (via meeting the electricity demand) and indirectly (via reducing the electrical energy needed for cooling) decreases the electricity import from grid. With the limit over the thermal output of ICE,1 it is emission saving using its thermal energy for cooling purposes (by driving the AC) rather than for direct heating purposes. So the ICE operates almost in full-load during hot and medium seasons (to feed the AC), and stops in some block-loads in cold season (Peak 2, High 2), during which a 170 kW boiler is employed to meet the thermal demand. Except this, the operation strategy of the energy systems was found to be comparably similar to Figs. 2 and 3. The optimal design of the energy systems to minimize GHG (solution c in Fig. 4) is composed of an ICE (172 kW), 80 kW PV systems, boiler (170 kW) and 2 ACs (208 kW). Overall, the ICE operates to meet the base-load where the extra electricity and heat needed are met via grid and boiler, respectively. PV is able to satisfy less than 1% needs of the building (due to rooftop constraints); in relative terms this only reduces the building GHG emissions by 3%, but in absolute terms 80 kWp (kW peak) PV provides an annual savings of 44 tonne CO2 eq. 3.3. Pareto frontier for cost and Acidification In this section, solutions positioned on the Pareto frontier obtained to show the trade-off between the building cost and its acidification impact are explored (Fig. 5). As Tables 1 and 2 show, the Portuguese generation mix in 2011 [20] has significantly higher acidification impact (per kWh electricity) than DG. Thus, in order to reduce the building acidification impact, the use of electricity from grid should be moderated and ultimately avoided. The employment of cogeneration technologies, specifically SOFC, provides significant savings compared to conventional design of energy systems and results in a total negative balance of acidification emissions. The Pareto frontier is composed of five curves. The solutions positioned on section A of the Pareto front employ ICE, boiler, AC, and PV systems. The solutions on section B contain the same type of energy systems, plus SOFC systems (Fig. 5). Each jump in the value of cost objective function represents the employment of an extra SOFC unit. 1 In order to benefit from the national remuneration regime, as explained in Safaei et al. [44], the overall efficiency of the cogeneration system should satisfy the Primary Energy Savings (PES) threshold defined in Decree-law 2004/8/EC (2004) and its Portuguese successor [14]. Safaei et al. [44] discussed that satisfying this threshold dictates that at least 77% of the annual thermal output of the ICE systems should be consumed onsite.
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427
500 450 400 350 300 250 200 150 100 50 0
PV grid ICE
low
medium2
high2
medium1
high1
peak2
low
peak1
medium2
high2
medium1
high1
peak2
low
peak1
medium2
high2
medium1
high1
peak1
export peak2
kW
424
H H H H H H H M M M M M M M C C C C C C C Fig. 2. Electrical power—optimal operation planning of energy systems to minimize non-renewable CED (solution b in Fig. 1).
300 250
kW
200 150 Boiler
100
ICE
50
Thermal demand peak1 peak2 high1 high2 medium1 medium2 low peak1 peak2 high1 high2 medium1 medium2 low peak1 peak2 high1 high2 medium1 medium2 low
0
H H H H H H H M M M M M M M C C C C C C C Fig. 3. Heating power—optimal operation planning of energy systems to minimize non-renewable CED (solution b in Fig. 1).
1600 α
GHG (tonne CO2 eq)
1400 1200 1000 800
ω γ
600 400 200 0 390
400
410
420
430
440
450
LCEAC (k€)
reduce acidification, due to their high power-to-heat ratio, which allows them to supply most of the building electrical power needs while not over-supplying the thermal demand. Consecutive drops in the acidification impact in section B of Fig. 5 are due to the use of an extra unit of SOFC. Ultimately 4 SOFCs are employed along with an ICE, 80 kW PV, 107 kW boiler, 1 AC and 45 kW CC to meet the building demand. This setting corresponds to the configuration of energy systems to minimize acidification impact (solution d) in Fig. 5. SOFCs operate to meet the base-loads, while ICE is employed to satisfy peak demand (Figs. 6 and 7). As discussed, acidification emissions savings can be acquired through exporting extra power to grid; therefore throughout the year, 150 kW electrical power, the maximum power allowed due to policy restrictions, is exported to grid. With five CHP systems
Fig. 4. Cost vs. GHG Pareto frontier.
2500
Acidificaon (kg SO2 eq)
α
The solution using only conventional sources, solution a, is also displayed in Fig. 5. The employment of the ICE (in solution x) decreases annual acidification by more than 40% compared to conventional sources. All the solutions positioned on section A of Fig. 5 employ an ICE; however, the contribution of the ICE to meet the building demand increases toward costlier solutions. The capacity of PV installations also increases, until they occupy the total rooftop space. Acidification can be reduced annually by roughly 350 kg SO2 eq. as an effect of employing 80 kW PV systems. All solutions positioned on the right hand side of section A of Fig. 5 employ 80 kW PV systems. The drop in the value of acidification impact, from section A to section B, is the result of employment of SOFC in the subsequent solutions. SOFC systems are the most efficient DG technology to
2000
B
A
1500 ω ICE, PV ICE, 1 SOFC, PV
1000
ICE, 2 SOFC, PV ICE, 3 SOFC, PV
500
ICE, 4 SOFC, PV
0 δ
-500 390
440
490
540
590
640
690
740
LCEAC (k€) Fig. 5. Cost vs. acidification pareto frontier.
790
840
425
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427
700 600 500 400 pv
300
ICE
200
SOFC
100
export high2
medium2
low
high1
medium1
peak2
low
medium2
high2
medium1
high1
peak2
low
peak1
medium2
high2
medium1
C
high1
C
peak2
H H H H H H H M M M M M M M C
peak1
peak1
0
C
C
C
C
Fig. 6. Electrical power—optimal operation planning of energy systems to minimize acidification impact (solution d in Fig. 5).
450 400 350 300 250 200 150 100 50 0
Boiler ICE SOFC peak1 peak2 high1 high2 medium1 medium2 low peak1 peak2 high1 high2 medium1 medium2 low peak1 peak2 high1 high2 medium1 medium2 low
Thermal demand
H H H H H H H M M M M M M M C C C C C C C
Eutrophicaon (kg PO4 eq)
Fig. 7. Heating power—optimal operation planning of energy systems to minimize acidification impact (solution d in Fig. 5).
2000 1500 1000 ω
α ICE,PV 1 SOFC, ICE,PV
500
2 SOFC, ICE,PV 3 SOFC, ICE,PV
0
4 SOFC, ICE,PV
-500 -1000 390
η
440
490
540
590
640
690
740
790
increasing by one unit for each curve toward less eutrophication impacts. On top of SOFC and ICE, 80 kW PV systems are also employed, which results in an annual saving of up to 410 kg PO34 eq. The optimal planning patterns of the energy systems in each block-load to minimize eutrophication impact, corresponding to solution g in Fig. 8, are similar to Figs. 6 and 7. The set of energy systems is composed of ICE (172 kW), SOFC (500 kW), PV systems (80 kW), boiler (107 kW), AC (104 kW), CC (45 kW), and the grid (Portuguese mix 2011).
840
LCEAC (k€) Fig. 8. Cost vs. Eutrophication Pareto frontier.
operating, a relatively high annual energy cost is experienced. The cooling base demand is met by a 104 kW AC, along with 45 kW CC that are employed to meet the rest of demand. 3.4. Pareto frontier for cost and Eutrophication Fig. 8 shows the Pareto frontier derived for cost and eutrophication impact (kg PO34 eq.). The format of Pareto frontier, and the types of energy systems along solutions located on the curve are parallel to the Pareto frontier obtained for cost vs. acidification (Fig. 5). In order to reduce the eutrophication impact, the imported (exported) electrical energy gradually decreases (increases). The most viable technology to increase the export to grid (without violating the policy framework) is SOFC due to high power-to-heat ratio, so its installation capacity increases according to the level of eutrophication reduction desired by the DM. The Pareto frontier is composed of five curves, the number of SOFC installation
4. Discussions The format of curves and the types of energy systems along the solutions located on the Pareto frontiers obtained for cost vis-à-vis CED (Fig. 1) and GHG (Fig. 4) were found to be similar. The same association was observed between Pareto frontiers obtained for cost vis-à-vis acidification (Fig. 5) and eutrophication impacts (Fig. 8). This has clear implications for decision making within the building sector, demonstrating that the same combination of DG can be employed to reduce both non-renewable CED and GHG, or acidification and eutrophication. Figs. 1 and 4 showed that ICE is a viable cogeneration technology in order to minimize the cost, GHG emissions or non-renewable CED arising from meeting the building energy demand. The operating strategy of ICE should be optimized according to the required value of each objective function. The Pareto frontiers obtained for cost vis-à-vis eutrophication and acidification impacts indicate that in order to minimize those impacts, energy use from grid should be curbed and ultimately avoided. Employment of any type of DG can mitigate the impacts, but SOFC is the best DG solution to be used for this purpose while not oversupplying the building thermal demand.
426
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427
Our analysis shows that PV has higher potential compared to ST to reduce the impacts of the building when combined with cogeneration systems. Employing CHP might not leave space for the employment of ST systems, since the building thermal demand could be entirely met by CHP. Half of the electrical energy produced onsite, however, could be exported to grid (as long as not violating the policy framework) and this provides more space for the employment of PV systems. That is, although PV has higher installation cost, it also has higher potential to reduce the environmental impacts of the building compared to ST, due to the dynamics of national policies and conditions.
5. Conclusions Considering the high impacts of building sector, it is well established that any rigorous assessment of building energy systems should also incorporate environmental aspects. By combining optimization techniques and LCA methodology, this work presents a systematic framework for incorporating environmental aspects into evaluation and selection of energy systems for the building sector. An LCA was implemented to assess the LCEI related to construction and operation of alternative types of energy systems, as well as the upstream emissions related to their fuel input, such as natural gas. That is, the framework adopts a life-cycle approach that is fundamental for a thorough assessment of energy systems. The result of LCA was used as an input into the multi-objective model to optimize the design and operational strategy of energy systems to mitigate each type of building environmental impact and costs. The model adequately enables assessing the trade-offs underlying economic implications and environmental impacts associated with meeting the building energy demand. The proposed methodological framework allows selecting the design and optimizing the operation strategy of energy systems according to the DM’s objectives, i.e. reducing costs or environmental impacts. The application of this framework to the Portuguese commercial building sector led to the conclusion that an operational strategy of energy systems to reduce non-renewable CED also reduces GHG emissions, and vice versa. The same association was observed between acidification and eutrophication impacts. Furthermore, the importance of incorporating the national policies to promote DG into the analysis was underlined. Given that the inputs to the model (environmental impacts of the energy systems and costs) are adjusted to the study, the proposed framework could be applied to other types of buildings, such as residential or industrial, and other locations. For this purpose, an LCA should be adequately framed according to the location to regard geographical properties (for solar systems) and incorporate the correct fuel input (NG) upstream emissions. Future developments include performing sensitivity analysis over several input parameters to the model, e.g. efficiency parameters and emission factors of DG. Other extensions include developing a robust optimization framework to provide a cost- and/or environmentally-robust solutions in the face of uncertainty.
Acknowledgements The authors acknowledge the Portuguese Science and Technology Foundation (FCT) projects PTDC/SEN-TRA/117251/ 2010 and PEst-OE/EEI/UI0308/2014. This work was developed under the Energy for Sustainability Initiative of the University of Coimbra and supported by the R&D Project EMSURE (Energy and Mobility for Sustainable Regions, CENTRO 07 0224 FEDER 002004).
Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.enconman.2015. 03.048.
References [1] Abdelhady S, Borello D, Tortora E. Design of a small scale stand-alone solar thermal co-generation plant for an isolated region in Egypt. Energy Convers Manage 2014;88:872–82. [2] Akikur RK, Saidur R, Ping HW, Ullah KR. Performance analysis of a cogeneration system using solar energy and SOFC technology. Energy Convers Manage 2014;79:415–30. [3] Alanne K, Söderholm N, Sirén K, Beausoleil-Morrison I. Techno-economic assessment and optimization of Stirling engine micro-cogeneration systems in residential buildings. Energy Convers Manage 2010;51(12):2635–46. [4] Arcuri P, Florio G, Fragiacomo P. A mixed integer programming model for optimal design of trigeneration in a hospital complex. Energy 2007;32(8):1430–47. [5] Bhattacharyya SC, Quoc Thang DN. Economic buy-back rates for electricity from cogeneration: Case of sugar industry in Vietnam. Energy 2004;29(7):1039–51. [6] Bortolini M, Gamberi M, Graziani A. Technical and economic design of photovoltaic and battery energy storage system. Energy Convers Manage 2014;86:81–92. [7] Brandoni C, Arteconi A, Ciriachi G, Polonara F. Assessing the impact of microgeneration technologies on local sustainability. Energy Convers Manage 2014;87:1281–90. [8] Çakir U, Çomakli K, Yüksel F. The role of cogeneration systems in sustainability of energy. Energy Convers Manage 2012;63:196–202. [9] Calise F, d’Accadia MD, Vicidomini M, Scarpellino M. Design and simulation of a prototype of a small-scale solar CHP system based on evacuated flat-plate solar collectors and Organic Rankine Cycle. Energy Convers Manage 2015;90:347–63. [10] Casisi M, De Nardi A, Pinamonti P, Reini M. Effect of different economic support policies on the optimal synthesis and operation of a distributed energy supply system with renewable energy sources for an industrial area. Energy Convers Manage 2015;95:131–9. [11] Chicco G, Mancarella P. Distributed multi-generation: a comprehensive view. Renew Sustain Energy Rev 2009;13(3):535–51. [12] Cho H, Luck R, Eksioglu SD, Chamra LM. Cost-optimized real-time operation of CHP systems. Energy Build 2009;41(4):445–51. [13] Clift R, Doig A, Finnveden G. The application of life cycle assessment to integrated solid waste management: Part I – Methodology. Transactions of the Institute of Chemicals Engineers, Part B: Process Saf Environ 2000;78:279–89. [14] Decree-Law n° 23/2010 of 25 of March (in Portuguese). [15] Dentice d’Accadia M, Sasso M, Sibilio S, Vanoli L. Micro-combined heat and power in residential and light commercial applications. Appl Therm Eng 2003;23(10):1247–59. [16] Directive 2004/8/EC (2004) Directive 2004/8/EC of the European parliament and the council on the promotion of cogeneration based on a useful heat demand in the internal energy market.. [17] Dorer V, Weber R, Weber A. Performance assessment of fuel cell microcogeneration systems for residential buildings. Energy Build 2005;37(11):1132–46. [18] Dorer V, Weber A. Energy and CO2 emissions performance assessment of residential micro-cogeneration systems with dynamic whole-building simulation programs. Energy Convers Manage 2009;50(3):648–57. [19] Fuller SK, Petersen SR. Life-Cycle Costing Manual for the Federal Energy Management Program. 1995 Edition. NIST handbook; 1996. p. 135. [20] Garcia R, Marques P, Freire F. Life-cycle assessment of electricity in Portugal. Appl Energy 2014;134:563–72. [21] Guinée JB, Gorrée M, Heijungs R, Huppes G, Kleijn R, Koning A de, et al. Handbook on life cycle assessment. Operational guide to the ISO standards. I: LCA in perspective. IIa: Guide. IIb: Operational annex. III: Scientific background. Kluwer Academic Publishers, ISBN 1-4020-0228-9, Dordrecht, 2002, p. 692. [22] Gunes MK. Investigation of a fuel cell based total energy system for residential applications. MSc thesis. Virginia Polytechnic Institute and State University, Virginia; 2001. [23] Hawkes A, Leach M. Solid oxide fuel cell systems for residential microcombined heat and power in the UK: key economic drivers. J Power Sources 2005;149:72–83. [24] Hawkes A, Leach M. Impacts of temporal precision in optimization modelling of micro-combined heat and power. Energy 2005;30(10):1759–79. [25] IEA. Transition to Sustainable Buildings; Strategies and Opportunities to 2050; 2013. Available at: . _ _ A, Izgi E, Ay S. Modelling of the change in national exchange rate model [26] Inan depending on the economic parameters of a natural gas cogeneration system: Turkey case. Energy Convers Manage 2009;50(4):1049–55.
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427 [27] Ismail MS, Moghavvemi M, Mahlia TMI. Genetic algorithm based optimization on modeling and design of hybrid renewable energy systems. Energy Convers Manage 2014;85:120–30. [28] Kabalci E. A smart monitoring infrastructure design for distributed renewable energy systems. Energy Convers Manage 2015;90:336–46. [29] Kanoglu M, Dincer I. Performance assessment of cogeneration plants. Energy Convers Manage 2009;50(1):76–81. [30] Kong XQ, Wang RZ, Li Y, Huang XH. Optimal operation of a micro-combined cooling, heating and power system driven by a gas engine. Energy Convers Manage 2009;50(3):530–8. [31] Lu H, Yu Z, Alanne K, Xu X, Fan L, Yu H, et al. Parametric analysis of energy quality management for district in China using multi-objective optimization approach. Energy Convers Manage 2014;87:636–46. [32] Lund C, Biswas W. A review of application of lifecycle analysis to renewable energy systems. B Sci Technol Soc 2008;28(3):209–2429. [33] Marantan A, Popovic P, Radermacher R. The potential of CHP technology in commercial buildings; characterizing the CHP demonstration building. ASHRAE Symposium on CHP Technologies for the New Century, ASHRAE Transactions, vol. 108, part 1; 2002. [34] Mavrotas G, Diakoulaki D, Florios K, Georgiou P. A mathematical programming framework for energy planning in services’ sector buildings under uncertainty in load demand: the case of a hospital in Athens. Energy Policy 2008;36(7):2415–29. [35] McCarl BA et al. McCarl Expanded GAMS User Guide. GAMS Release 24.2.1. Washington, DC, USA:GAMS Development Corporation; 2013.
427
[42] Ren H, Gao W. A MILP model for integrated plan and evaluation of distributed energy systems. Appl Energy 2010;87(3):1001–14. [43] Ren H, Zhou W, Nakagami KI, Gao W, Wu Q. Multi-objective optimization for the operation of distributed energy systems considering economic and environmental aspects. Appl Energy 2010;87(12):3642–51. [44] Rezvan AT, Gharneh NS, Gharehpetian GB. Robust optimization of distributed generation investment in buildings. Energy 2012;48(1):455–63. [45] Rosato A, Sibilio S, Scorpio M. Dynamic performance assessment of a residential building-integrated cogeneration system under different boundary conditions. Part I: Energy analysis. Energy Convers Manage 2014;79:731–48. [46] Rubio-Maya C, Uche-Marcuello J, Martínez-Gracia A, Bayod-Rújula AA. Design optimization of a polygeneration plant fuelled by natural gas and renewable energy sources. Appl Energy 2011;88(2):449–57. [47] Safaei A, Freire F, Antunes CH. A model for optimal energy planning of a commercial building integrating solar and cogeneration systems. Energy 2013;61:211–23. [48] Shrestha RM, Marpaung COP. Supply- and demand-side effects of power sector planning with demand-side management options and SO2 emission constraints. Energy Policy 2005;33(6):815–25. [49] Sylva J, Crema A. A method for finding the set of non-dominated vectors for multiple objective integer linear programs. Eur J Oper Res 2004;158(1):46–55. [50] Torchio MF. Comparison of district heating CHP and distributed generation CHP with energy, environmental and economic criteria for Northern Italy. Energy Convers Manage 2015;92:114–28. [51] Ullah KR, Akikur RK, Ping HW, Saidur R, Hajimolana SA, Hussain MA. An experimental investigation on a single tubular SOFC for renewable energy based cogeneration system. Energy Convers Manage 2015;94:139–49. [52] Wang M, Wang J, Zhao P, Dai Y. Multi-objective optimization of a combined cooling, heating and power system driven by solar energy. Energy Convers Manage 2015;89:289–97. [53] Wu YJ, Rosen MA. Assessing and optimizing the economic and environmental impacts of cogeneration/district energy systems using an energy equilibrium model. Appl Energy 1999;62(3):141–54. [54] Xuan J et al. Evaluation of the introduction of a combined heat and power system in a commercial building in Shanghai. J Asian Arch Build Eng 2006;5(2):385–90. [55] Yokoyama R, Shinano Y, Taniguchi S, Ohkura M, Wakui T. Optimization of energy supply systems by MILP branch and bound method in consideration of hierarchical relationship between design and operation. Energy Convers Manage 2015;92:92–104.
Contents lists available at ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
A life cycle multi-objective economic and environmental assessment of distributed generation in buildings Amir Safaei a,b,⇑, Fausto Freire b, Carlos Henggeler Antunes a,c a
INESC Coimbra, R. Antero de Quental 199, 3000–033 Coimbra, Portugal ADAI-LAETA, Department of Mechanical Engineering, University of Coimbra, Rua Luis Reis Santos, 3030-788 Coimbra, Portugal c Department of Electrical and Computer Engineering, University of Coimbra, 3030–290 Coimbra, Portugal b
a r t i c l e
i n f o
Article history: Received 6 November 2014 Accepted 12 March 2015 Available online 8 April 2015 Keywords: Lifecycle assessment (LCA) Multi objective linear programming (MOLP) Distributed Generation (DG) Cogeneration Solar Pareto frontiers Buildings
a b s t r a c t Distributed generation, namely cogeneration and solar technologies, is expected to play an important role in the future energy supply mix in buildings. This calls for a methodological framework to assess the economic and environmental performance of the building sector when such technologies are employed. A life-cycle model has been developed, combining distributed generation and conventional sources to calculate the cost and environmental impacts of meeting the building energy demand over a defined planning period. Three type of cogeneration technologies, solar photovoltaic and thermal, as well as conventional boilers along with the Portuguese electricity generation mix comprise the energy systems modeled. Pareto optimal frontiers are derived, showing the trade-offs between different types of impacts (non-renewable cumulative energy demand, greenhouse gas emissions, acidification, eutrophication) and cost to meet the energy demand of a commercial building. Our analysis shows that according to the objective to employ distributed generation (reducing cost or environmental impacts), a specific design and operational strategy for the energy systems shall be adopted. The strategies to minimize each type of impact and the associated cost trade-offs by exploring the solutions located on the Pareto optimal frontiers are discussed. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Energy use in the building sector accounts for more than 40 percent of the EU energy consumption [25]. The generation of energy at the point of consumption, employing Distributed Generation (DG) technologies, is pointed out as a key option for promoting energy efficiency and use of renewable sources in building sector [48]. Combined Heat and Power (CHP) technologies and solar technologies, namely Photovoltaic (PV) and Solar Thermal (ST), represent promising onsite generation alternatives and several strategies, including the introduction of financial incentives, have been implemented to further increase the share of DG in the building sector energy mix [46]. A comprehensive view on distributed multi generation by Chicco and Mancarella [11] also underlines this point. The introduction of different types of DG calls for a framework to assess the building economic and environmental performance when such technologies are employed. Some studies have assessed ⇑ Corresponding author at: ADAI-LAETA, Department of Mechanical Engineering, University of Coimbra, Rua Luis Reis Santos, 3030-788 Coimbra, Portugal. E-mail address: [email protected] (A. Safaei). http://dx.doi.org/10.1016/j.enconman.2015.03.048 0196-8904/Ó 2015 Elsevier Ltd. All rights reserved.
DG technologies considering efficiency and economic perspectives, by using techniques other than optimization. Xuan et al. [54] examined the application of a gas-fuelled reciprocating Combined Heat and Power (CHP) in a commercial building in China. Wu and Rosen [53] employed an energy equilibrium model to compare conventional and NG cogeneration-based district energy systems for heating, cooling and electrical services. Marantan et al. [33] demonstrated the potential of tri-generation NG CHP application in commercial buildings. The importance of cogeneration systems for sustainable energy use was underlined in Çakir et al. [8]. Gunes [22] examined the application of fuel cell-based Total Energy System (TES) for residential buildings. Dentice d’Accadia et al. [15] dealt with the application of a small scale fuel cell cogeneration (electrical power <15 kW) to light commercial application users. Another study on fuel cells [17] offered a methodology for assessing the performance of two types of fuel cells in terms of primary energy demand and CO2 emissions. Bhattacharyya and Quoc Thang [5] found that economic feasibility of medium and large scale cogeneration systems is vulnerable to changes in buy-back rate of electricity and investment costs. Mone et al. [36] investigated the underlying factors in economic feasibility of CHP systems using commercially available gas
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427
turbines. Ullah et al. [51] evaluated the application of a single tubular Solid Oxide Fuel Cell (SOFC) with an electrolyte of YttriaStabilized Zirconia ceramic powder. Ranjbar et al. [41] reported the energy and exergy assessments of a trigeneration system based on a SOFC. Raj et al. [40] investigated the effect of temperature, stoichiometry and the degree of humidification on the performance of a planar solid oxide fuel cell. Brandoni et al. [7] assessed the impact of micro-generation technologies on the mid- and longterm sustainability of urban areas. A dynamic energy analysis of a residential building-integrated cogeneration system under different boundary conditions was presented in Rosato et al. [45]. Regardless of the purpose to install DG (e.g. reducing cost or environmental impacts), factors such as characteristics of different types of DG, dynamic energy costs, different types of building energy demand and their variation, varying solar resources, and national policy frameworks to promote each type of DG should be taken into account to decide on the design and operating strategy of DG. Optimization models applied to DG systems typically consider only cogeneration systems, focusing on economic and technical aspects. For instance, Alanne et al. [3] discussed the techno-economic optimization of Stirling engine micro-cogeneration systems in residential buildings. Casisi et al. [10] studied the effect of different economic support policies on the optimal synthesis and operation of a renewable-based distributed energy supply system for an industrial area. Calise et al. [9] dealt with the design and simulation of a prototype of a small-scale solar CHP system based on evacuated flat-plate solar collectors and organic Rankine cycle. Abdelhady et al. [1] discussed the design of a small scale stand-alone solar thermal co-generation plant for an isolated region in Egypt. Akikur et al. [2], analyzed the performance of a cogeneration SOFC system combined with solar energy. Kabalci et al. [28] proposed a smart monitoring system for renewable based distributed energy systems. Bortolini et al. [6] evaluated the technical and economic design of a combined PV and battery energy storage system. Arcuri et al. [4] proposed a combination of heat pumps, absorption chillers and cogeneration systems, discussing the optimal operation strategy of the energy systems maximizing annual short- and long-term economic returns. Cho et al. [12] developed a model to minimize the total cost of energy usage for a building based on energy efficiency constraints for each component. Dorer and Weber [18] assessed the emission performance of residential micro-cogeneration systems with dynamic whole-building simulation programs. Hawkes and Leach [23] developed an equivalent annual cost minimization model to determine the driving factors behind the investment in fuel cell CHP technologies. A sensitivity analysis showed that the results were sensitive to capital cost, energy import/export prices, plant life_ time, and the temporal precision selected for the study [24]. Inan et al. [26] examined the effect of exchange rate on the cogeneration systems fixed and variable costs in Turkey. The performance assessment of various building cogeneration through energy and exergy efficiencies was discussed in Kanoglu and Dincer [29]. Kong et al. [30] developed an energy optimization model for gas turbine cogeneration systems. Mavrotas et al. [34] discussed an optimization framework for energy supply systems in commercial buildings by considering the demand uncertainty. Monteiro et al. [37] developed a model for planning micro-CHP plants in agreement with the Portuguese energy legal framework. Pan et al. [39] presented a real-time optimum operation strategy to improve the efficiency of the cogeneration systems, integrating turbine generator and cooling tower. Yokoyama et al. [55] discussed the optimal design of a gas engine cogeneration system for electricity and hot water supply, using a branch and bound method. In this regard, relatively few recent studies have looked at the combination of renewable and CHP. Such studies include Akikur et al. [2], who examined the performance analysis of a cogeneration
421
system using solar energy and SOFC technology. Ismail et al. [27] discussed the sizing optimization of a system consisting of PV panels, CHP and storage, using genetic algorithms. Rezvan et al. [44] developed a robust optimization model to determine the optimum capacity of DG for buildings in the case of demand uncertainty. Moreover, despite environmental impacts associated with the building sector is a key issue, only some studies incorporate environmental aspects into the design and operation of DG. Ren & Gao [42] developed a single-objective model for the integrated plan and evaluation of DG systems. The model was extended to a multi-objective model in Ren et al. [43], minimizing energy costs and operating CO2 emissions. Lu et al. [31] presented a multiobjective optimization based for energy management of a district site in China. A systematic optimization procedure to select and size a cogeneration plant fuelled by natural gas, evacuated tube solar collectors, and gasified biomass is discussed in Rubio-Maya et al. [46]. Torchio [50] compared a district heating CHP and a distributed generation CHP in terms of energy and economic criteria, as well as CO2 and NOx emissions. Wang et al. [52] proposed a multi-objective optimization for combined cooling, heating and power system driven by solar energy. Such studies mainly assess operating CO2 (or NOx or SO2), and do not adopt Life-Cycle (LC) approaches, whereas any comparison among energy supply options must employ a LC approach [32]. An exception is Osman et al. [38] who looked at both costs andtwo types of environmental impacts (greenhouse gases, tropospheric ozone depletion), taking into account only CHP technologies. The incorporation of environmental aspects into the analysis calls for multi-objective models and techniques in which decisions should be made exploiting the trade-offs between the conflicting axes of evaluation of the merits of distinct solutions, which are made operational by multiple objective functions to be optimized. A model to minimize the Life-Cycle Costs (LCC) of meeting the energy demand (power, heating, cooling) of a commercial building by integrating DG and conventional energy sources was presented in Safaei et al. [47]. Here, the model is extended into a multiobjective linear programming (MOLP) model by introducing four additional environmental impacts as objective functions along with cost. To calculate the environmental impacts of different energy solutions, a Life-Cycle Assessment (LCA) was conducted considering the impacts related to construction and operation of energy systems, as well as the upstream processes related to their fuel input, i.e. NG for CHPs. MOLP techniques are employed to unveil and exploit the trade-off between the competing objectives, i.e. minimizing cost versus minimizing each type of impact category. Section 2 provides a brief overview of the model and its inputs. In Section 3 the Pareto optimal frontiers obtained for cost vis-à-vis environmental impacts are presented. This is followed by the discussions of the results in Section 4. Finally, in Section 5 the main conclusions are drawn. 2. MOLP model The mathematical model was presented in Safaei et al. [47] for cost optimal design and operation of DG in Portuguese commercial buildings. The model incorporates different types of CHP technologies (Micro-Turbines – MT, Internal Combustion Engines – ICE, Solid Oxide Fuel Cells – SOFC), separate production of electricity - Portuguese electricity generation mix in 2011 [20] and heat (onsite boilers), renewable sources (ST and PV), and auxiliary cooling systems (Absorption Chiller – AC, Compression Chiller – CC). The specifications of the energy systems considered and the main assumptions are described below: Solar PV systems: 4 kWp (KW peak) mono-crystalline PV system with lifetime of 30 years and three 2500 W inverters.
422
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427
ST systems: a solar thermal system consisting of flat-plate absorbers with black chrome coating on copper, a 2000 liters steel heat storage, two 40 W pumps, and the necessary piping for the operation of the system. The transportation of the solar system parts and construction works for mounting of the panels are also accounted for. MT cogeneration system: a 60 kWe (kW electricity) MT system with an operating life of six years with four maintenance sessions for this period. The turbine runs on natural gas, has lubricant consumption of 4.4 liters per year, with electrical and thermal efficiencies of 26% and 52% at 100%-load, respectively. SOFC cogeneration system: a tubular cell design 125 kWe natural gas SOFC engine with operating life of 8 years. During this period, the engine has one major overhaul session to change the fuel stack and 7 maintenance sessions each with 0.15 m3 tap water consumption. The rated electrical efficiency and the power to heat ratio are 45% and 1.20 at 100%-load, respectively. ICE cogeneration system: a 172 kWe natural gas engine with a rated electrical efficiency of 33% and power to heat ratio of 0.61 at 100% load. The lifetime of the system is assumed to be 6 years. AC: a 104 kW AC with a coefficient of performance (COP) of 0.7 and lifetime of 20 years. CC: a 15 kW heat pump with a COP of 2.2 and lifetime of 20 years. The emission of refrigerant R134a during operation is taken into account. Conventional boiler: a natural gas boiler with a rated efficiency of 80% and lifetime of 20 years. National grid: Portuguese electricity generation mix in 2011. This base model was applied to a hotel complex located in the city of Coimbra, Portugal. The data for hourly electricity consumption, as well as hourly cooling and heating demands (calculated using building simulation) were used to generate the loadduration-curves of the building. The year was divided into three seasons – Hot (H), Mild (M) and Cold (C) -, each being represented by three load duration curves (for each type of cooling, heating and power demand) with the following block-loads: peak1 (P1), peak2 (P2), high1 (H1), high2 (H2), medium1 (M1), medium2 (M2) and low (L). The block-loads are defined in a way to represent the dynamic cost of electricity. The MOLP model formulation in this article explicitly considers five objective functions to be minimized: LC Equivalent Annual Costs (LCEAC), and LC Environmental impacts (LCEI): nonrenewable Cumulative Energy Demand (CED), Greenhouse Gas (GHG) emissions, acidification, and eutrophication. LCEAC is the cost per year of meeting the building energy demand, comprising the fixed and variable costs. LCEAC is frequently used as a decision making indicator when evaluating investment projects of unequal life spans [19]. Different categories of constraints are included in the model, namely: meeting electricity, heat, and cooling energy demands; capacity constraints to certify that the energy output of the generation units (CHPs, PV, ST, boiler, CC, AC) are within their operation ranges; national policy constraints to guarantee the fulfillment of the EU [16]/8/EC on the promotion of cogeneration systems; and consistency constraints to guarantee the internal coherence of the model. The decision variables determine the number of panels installed for solar systems, and the number of units installed and their output for the other generation units, at each year throughout the planning period. Input parameters to the model include the magnitude of different types of energy demand of the building (power, heating, cooling) at each block-load throughout the planning period, fixed and operating costs of the units, their
environmental impacts and the economic indicators, e.g. interest rate. The full description of the base model, and case study characteristics are discussed in Safaei et al. [47]. The model enables a thorough assessment of the economic implications and environmental impacts associated with meeting the building energy demand. The environmental impacts of the building employing conventional sources of energy (onsite boiler, grid) are estimated, and the design and operating strategy of energy systems combining DG to reduce each type of impact is explored. LCA was used to calculate the LCEI arising from one unit energy output of energy systems based on the CML method [21], as summarized in Tables 1 and 2. The MOLP model enables to derive Pareto optimal frontiers, which display the trade-offs between cost-efficient solutions and the solutions with lower non-renewable CED, GHG, acidification and eutrophication impact values. Each solution consists of a set of energy systems and their respective operation planning to meet the level of cost and environmental impacts determined by the DM. The assumptions to calculate the Pareto optimal frontiers are as follows: Unlike the base model in [47], the capacity of a conventional boiler is considered as a variable, so the optimal solution (according to the defined objective function) gives out the optimal capacity of the boiler and its output throughout the planning period. The model considers exporting electrical power to the grid. The emissions resulting from the exported power are credited by their deduction from the total emissions, since it was assumed that they avoid the generation of electricity (according to the production mix). This has been accounted by the model using the ‘‘avoided-burden approach’’ [13]. To derive the Pareto optimal frontier, an algorithm based on the methodology proposed in Sylva & Crema [49] was implemented in GAMS [35]. 3. Results 3.1. Pareto frontier for cost and CED The Pareto frontier obtained for non-renewable CED and cost objective functions is shown in Fig. 1. The values correspond to the estimated LCEAC and non-renewable CED of meeting the building energy demand for a year. For solution a in Fig. 1, no DG is permitted to be installed, being the set of energy systems composed of boiler (257 kW), CC (120 kW), and the grid (Portuguese mix 2011). Solution b corresponds to the minimum non-renewable CED, being the set of energy systems composed of ICE (172 kW), PV system (80 kW, the maximum capacity due to area restrictions), boiler (25 kW size), AC (208 kW), and the grid (Portuguese mix 2011). The set of energy systems for the most cost-effective solution (solution x at the most extreme left part of the Pareto frontier) is composed of an ICE (172 kW), a boiler with minimal size (5 kW), 1 AC (104 kW), 30 kW of CC, and the grid. Compared to the case employing only conventional systems, i.e. solution a in Fig. 1, employment of ICE at its cost-optimal operating strategy brings about both financial and energy savings. Thus solution a is dominated by the solution x located on the Pareto frontier. Safaei et al. [47] discussed that under certain fuel cost conditions, ICE and ST represent cost-effective DG solutions for the Portuguese commercial building sector. In order to reduce nonrenewable CED, the output of grid and boiler slightly increases, replacing ICE to partly address the building energy demand. To further reduce non-renewable CED, PV systems can be added to the energy mix. The number of PV installations gradually increases along the Pareto frontier as the DM desires for less non-renewable
423
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427 Table 1 Results of LC impact assessment of cogeneration systems—input to the MOLP model.
CED GHG Acidification Eutrophication
MJ/kW he g CO2 eq/kW he g SO2 eq/kW he g PO34 eq/kW he
ICE 100% load
ICE 75% load
ICE 50% load
MT 100% load
MT 75% load
MT 50% load
MT 25% load
SOFC 104% load
SOFC 100% load
SOFC 93% load
SOFC 85% load
SOFC 78% load
SOFC 68% load
SOFC 62% load
13.89 771
15.24 846
17.40 964
17.76 990
19.24 1054
23.07 1402
35.45 1961
10.69 625
10.46 621
9.82 611
9.62 608
9.44 605
9.44 605
9.23 602
0.468
0.494
0.530
0.533
0.558
0.623
0.833
0.280
0.276
0.265
0.262
0.259
0.259
0.256
0.105
0.113
0.123
0.111
0.118
0.137
0.199
0.092
0.090
0.087
0.086
0.085
0.085
0.084
Table 2 Results of LC impact assessment of solar and conventional systems—input to the MOLP model. Grida
PV kW he CED GHG Acidification Eutrophication a
MJ g CO2 eq g SO2 eq g PO34 eq
1.80 90 0.472 0.21
ST
Boiler
kW hth 4.70 350 0.78 0.54
0.39 21 0.158 0.10
5.75 320 0.168 0.04
Modeled according to Portuguese electricity generation mix in 2011 [20].
CED; the solution with minimum CED (solution b) employs an 80 kW PV system. Figs. 2 and 3 show the optimal configuration of the energy systems to minimize non-renewable CED. The X-axis shows the demand block-loads (peak, high, medium, low) throughout a year of the planning period for three hot (H), mild (M) and cold (C) seasons. ICE operates to meet the building base-load and remaining demand is met via grid and boiler. The cooling demand is completely met by 2 ACs that run on the waste heat from ICE and boiler. 3.2. Pareto frontier for cost and GHG
CED (GJ)
Fig. 4 shows the Pareto frontier obtained for GHG and cost objective functions. The format of the Pareto frontier, and the types of energy systems along solutions located on the Pareto frontier are similar to the Pareto obtained for cost and non-renewable CED (Fig. 1). The employment of ICE can therefore provide both economic and GHG savings for the building under study. The set of energy systems pertaining to the solutions a (conventional) and x (least cost solution) were described in Section 3.1. Solution c minimizes GHG emissions, being the set of energy systems composed of an ICE (172 kW), PV systems (80 kW), a boiler (170 kW size), AC (208 kW), and the grid (Portuguese mix 2011).
20000 18000 16000 14000 12000 10000 8000 6000 4000 2000 0 395
α
ω
β
400
405
410
415
420
LCEAC (k€) Fig. 1. Cost vs. CED Pareto frontier.
425
430
435
The set of energy systems and their operation planning for the solution with minimum cost, solution x, was explained in Section 3.1. In order to reduce GHG, the output of ICE increases to replace the electricity imported from the grid. As Fig. 4 displays, the slope of the frontier declines toward solutions with higher costs. This signifies that for such solutions it bears more cost to reduce GHG by one unit, since ICE is not operating according to its cost optimal design, being rather operating to decrease GHG emissions even in the block-loads in which its operation is not economical. Higher production by ICE directly (via meeting the electricity demand) and indirectly (via reducing the electrical energy needed for cooling) decreases the electricity import from grid. With the limit over the thermal output of ICE,1 it is emission saving using its thermal energy for cooling purposes (by driving the AC) rather than for direct heating purposes. So the ICE operates almost in full-load during hot and medium seasons (to feed the AC), and stops in some block-loads in cold season (Peak 2, High 2), during which a 170 kW boiler is employed to meet the thermal demand. Except this, the operation strategy of the energy systems was found to be comparably similar to Figs. 2 and 3. The optimal design of the energy systems to minimize GHG (solution c in Fig. 4) is composed of an ICE (172 kW), 80 kW PV systems, boiler (170 kW) and 2 ACs (208 kW). Overall, the ICE operates to meet the base-load where the extra electricity and heat needed are met via grid and boiler, respectively. PV is able to satisfy less than 1% needs of the building (due to rooftop constraints); in relative terms this only reduces the building GHG emissions by 3%, but in absolute terms 80 kWp (kW peak) PV provides an annual savings of 44 tonne CO2 eq. 3.3. Pareto frontier for cost and Acidification In this section, solutions positioned on the Pareto frontier obtained to show the trade-off between the building cost and its acidification impact are explored (Fig. 5). As Tables 1 and 2 show, the Portuguese generation mix in 2011 [20] has significantly higher acidification impact (per kWh electricity) than DG. Thus, in order to reduce the building acidification impact, the use of electricity from grid should be moderated and ultimately avoided. The employment of cogeneration technologies, specifically SOFC, provides significant savings compared to conventional design of energy systems and results in a total negative balance of acidification emissions. The Pareto frontier is composed of five curves. The solutions positioned on section A of the Pareto front employ ICE, boiler, AC, and PV systems. The solutions on section B contain the same type of energy systems, plus SOFC systems (Fig. 5). Each jump in the value of cost objective function represents the employment of an extra SOFC unit. 1 In order to benefit from the national remuneration regime, as explained in Safaei et al. [44], the overall efficiency of the cogeneration system should satisfy the Primary Energy Savings (PES) threshold defined in Decree-law 2004/8/EC (2004) and its Portuguese successor [14]. Safaei et al. [44] discussed that satisfying this threshold dictates that at least 77% of the annual thermal output of the ICE systems should be consumed onsite.
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427
500 450 400 350 300 250 200 150 100 50 0
PV grid ICE
low
medium2
high2
medium1
high1
peak2
low
peak1
medium2
high2
medium1
high1
peak2
low
peak1
medium2
high2
medium1
high1
peak1
export peak2
kW
424
H H H H H H H M M M M M M M C C C C C C C Fig. 2. Electrical power—optimal operation planning of energy systems to minimize non-renewable CED (solution b in Fig. 1).
300 250
kW
200 150 Boiler
100
ICE
50
Thermal demand peak1 peak2 high1 high2 medium1 medium2 low peak1 peak2 high1 high2 medium1 medium2 low peak1 peak2 high1 high2 medium1 medium2 low
0
H H H H H H H M M M M M M M C C C C C C C Fig. 3. Heating power—optimal operation planning of energy systems to minimize non-renewable CED (solution b in Fig. 1).
1600 α
GHG (tonne CO2 eq)
1400 1200 1000 800
ω γ
600 400 200 0 390
400
410
420
430
440
450
LCEAC (k€)
reduce acidification, due to their high power-to-heat ratio, which allows them to supply most of the building electrical power needs while not over-supplying the thermal demand. Consecutive drops in the acidification impact in section B of Fig. 5 are due to the use of an extra unit of SOFC. Ultimately 4 SOFCs are employed along with an ICE, 80 kW PV, 107 kW boiler, 1 AC and 45 kW CC to meet the building demand. This setting corresponds to the configuration of energy systems to minimize acidification impact (solution d) in Fig. 5. SOFCs operate to meet the base-loads, while ICE is employed to satisfy peak demand (Figs. 6 and 7). As discussed, acidification emissions savings can be acquired through exporting extra power to grid; therefore throughout the year, 150 kW electrical power, the maximum power allowed due to policy restrictions, is exported to grid. With five CHP systems
Fig. 4. Cost vs. GHG Pareto frontier.
2500
Acidificaon (kg SO2 eq)
α
The solution using only conventional sources, solution a, is also displayed in Fig. 5. The employment of the ICE (in solution x) decreases annual acidification by more than 40% compared to conventional sources. All the solutions positioned on section A of Fig. 5 employ an ICE; however, the contribution of the ICE to meet the building demand increases toward costlier solutions. The capacity of PV installations also increases, until they occupy the total rooftop space. Acidification can be reduced annually by roughly 350 kg SO2 eq. as an effect of employing 80 kW PV systems. All solutions positioned on the right hand side of section A of Fig. 5 employ 80 kW PV systems. The drop in the value of acidification impact, from section A to section B, is the result of employment of SOFC in the subsequent solutions. SOFC systems are the most efficient DG technology to
2000
B
A
1500 ω ICE, PV ICE, 1 SOFC, PV
1000
ICE, 2 SOFC, PV ICE, 3 SOFC, PV
500
ICE, 4 SOFC, PV
0 δ
-500 390
440
490
540
590
640
690
740
LCEAC (k€) Fig. 5. Cost vs. acidification pareto frontier.
790
840
425
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427
700 600 500 400 pv
300
ICE
200
SOFC
100
export high2
medium2
low
high1
medium1
peak2
low
medium2
high2
medium1
high1
peak2
low
peak1
medium2
high2
medium1
C
high1
C
peak2
H H H H H H H M M M M M M M C
peak1
peak1
0
C
C
C
C
Fig. 6. Electrical power—optimal operation planning of energy systems to minimize acidification impact (solution d in Fig. 5).
450 400 350 300 250 200 150 100 50 0
Boiler ICE SOFC peak1 peak2 high1 high2 medium1 medium2 low peak1 peak2 high1 high2 medium1 medium2 low peak1 peak2 high1 high2 medium1 medium2 low
Thermal demand
H H H H H H H M M M M M M M C C C C C C C
Eutrophicaon (kg PO4 eq)
Fig. 7. Heating power—optimal operation planning of energy systems to minimize acidification impact (solution d in Fig. 5).
2000 1500 1000 ω
α ICE,PV 1 SOFC, ICE,PV
500
2 SOFC, ICE,PV 3 SOFC, ICE,PV
0
4 SOFC, ICE,PV
-500 -1000 390
η
440
490
540
590
640
690
740
790
increasing by one unit for each curve toward less eutrophication impacts. On top of SOFC and ICE, 80 kW PV systems are also employed, which results in an annual saving of up to 410 kg PO34 eq. The optimal planning patterns of the energy systems in each block-load to minimize eutrophication impact, corresponding to solution g in Fig. 8, are similar to Figs. 6 and 7. The set of energy systems is composed of ICE (172 kW), SOFC (500 kW), PV systems (80 kW), boiler (107 kW), AC (104 kW), CC (45 kW), and the grid (Portuguese mix 2011).
840
LCEAC (k€) Fig. 8. Cost vs. Eutrophication Pareto frontier.
operating, a relatively high annual energy cost is experienced. The cooling base demand is met by a 104 kW AC, along with 45 kW CC that are employed to meet the rest of demand. 3.4. Pareto frontier for cost and Eutrophication Fig. 8 shows the Pareto frontier derived for cost and eutrophication impact (kg PO34 eq.). The format of Pareto frontier, and the types of energy systems along solutions located on the curve are parallel to the Pareto frontier obtained for cost vs. acidification (Fig. 5). In order to reduce the eutrophication impact, the imported (exported) electrical energy gradually decreases (increases). The most viable technology to increase the export to grid (without violating the policy framework) is SOFC due to high power-to-heat ratio, so its installation capacity increases according to the level of eutrophication reduction desired by the DM. The Pareto frontier is composed of five curves, the number of SOFC installation
4. Discussions The format of curves and the types of energy systems along the solutions located on the Pareto frontiers obtained for cost vis-à-vis CED (Fig. 1) and GHG (Fig. 4) were found to be similar. The same association was observed between Pareto frontiers obtained for cost vis-à-vis acidification (Fig. 5) and eutrophication impacts (Fig. 8). This has clear implications for decision making within the building sector, demonstrating that the same combination of DG can be employed to reduce both non-renewable CED and GHG, or acidification and eutrophication. Figs. 1 and 4 showed that ICE is a viable cogeneration technology in order to minimize the cost, GHG emissions or non-renewable CED arising from meeting the building energy demand. The operating strategy of ICE should be optimized according to the required value of each objective function. The Pareto frontiers obtained for cost vis-à-vis eutrophication and acidification impacts indicate that in order to minimize those impacts, energy use from grid should be curbed and ultimately avoided. Employment of any type of DG can mitigate the impacts, but SOFC is the best DG solution to be used for this purpose while not oversupplying the building thermal demand.
426
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427
Our analysis shows that PV has higher potential compared to ST to reduce the impacts of the building when combined with cogeneration systems. Employing CHP might not leave space for the employment of ST systems, since the building thermal demand could be entirely met by CHP. Half of the electrical energy produced onsite, however, could be exported to grid (as long as not violating the policy framework) and this provides more space for the employment of PV systems. That is, although PV has higher installation cost, it also has higher potential to reduce the environmental impacts of the building compared to ST, due to the dynamics of national policies and conditions.
5. Conclusions Considering the high impacts of building sector, it is well established that any rigorous assessment of building energy systems should also incorporate environmental aspects. By combining optimization techniques and LCA methodology, this work presents a systematic framework for incorporating environmental aspects into evaluation and selection of energy systems for the building sector. An LCA was implemented to assess the LCEI related to construction and operation of alternative types of energy systems, as well as the upstream emissions related to their fuel input, such as natural gas. That is, the framework adopts a life-cycle approach that is fundamental for a thorough assessment of energy systems. The result of LCA was used as an input into the multi-objective model to optimize the design and operational strategy of energy systems to mitigate each type of building environmental impact and costs. The model adequately enables assessing the trade-offs underlying economic implications and environmental impacts associated with meeting the building energy demand. The proposed methodological framework allows selecting the design and optimizing the operation strategy of energy systems according to the DM’s objectives, i.e. reducing costs or environmental impacts. The application of this framework to the Portuguese commercial building sector led to the conclusion that an operational strategy of energy systems to reduce non-renewable CED also reduces GHG emissions, and vice versa. The same association was observed between acidification and eutrophication impacts. Furthermore, the importance of incorporating the national policies to promote DG into the analysis was underlined. Given that the inputs to the model (environmental impacts of the energy systems and costs) are adjusted to the study, the proposed framework could be applied to other types of buildings, such as residential or industrial, and other locations. For this purpose, an LCA should be adequately framed according to the location to regard geographical properties (for solar systems) and incorporate the correct fuel input (NG) upstream emissions. Future developments include performing sensitivity analysis over several input parameters to the model, e.g. efficiency parameters and emission factors of DG. Other extensions include developing a robust optimization framework to provide a cost- and/or environmentally-robust solutions in the face of uncertainty.
Acknowledgements The authors acknowledge the Portuguese Science and Technology Foundation (FCT) projects PTDC/SEN-TRA/117251/ 2010 and PEst-OE/EEI/UI0308/2014. This work was developed under the Energy for Sustainability Initiative of the University of Coimbra and supported by the R&D Project EMSURE (Energy and Mobility for Sustainable Regions, CENTRO 07 0224 FEDER 002004).
Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.enconman.2015. 03.048.
References [1] Abdelhady S, Borello D, Tortora E. Design of a small scale stand-alone solar thermal co-generation plant for an isolated region in Egypt. Energy Convers Manage 2014;88:872–82. [2] Akikur RK, Saidur R, Ping HW, Ullah KR. Performance analysis of a cogeneration system using solar energy and SOFC technology. Energy Convers Manage 2014;79:415–30. [3] Alanne K, Söderholm N, Sirén K, Beausoleil-Morrison I. Techno-economic assessment and optimization of Stirling engine micro-cogeneration systems in residential buildings. Energy Convers Manage 2010;51(12):2635–46. [4] Arcuri P, Florio G, Fragiacomo P. A mixed integer programming model for optimal design of trigeneration in a hospital complex. Energy 2007;32(8):1430–47. [5] Bhattacharyya SC, Quoc Thang DN. Economic buy-back rates for electricity from cogeneration: Case of sugar industry in Vietnam. Energy 2004;29(7):1039–51. [6] Bortolini M, Gamberi M, Graziani A. Technical and economic design of photovoltaic and battery energy storage system. Energy Convers Manage 2014;86:81–92. [7] Brandoni C, Arteconi A, Ciriachi G, Polonara F. Assessing the impact of microgeneration technologies on local sustainability. Energy Convers Manage 2014;87:1281–90. [8] Çakir U, Çomakli K, Yüksel F. The role of cogeneration systems in sustainability of energy. Energy Convers Manage 2012;63:196–202. [9] Calise F, d’Accadia MD, Vicidomini M, Scarpellino M. Design and simulation of a prototype of a small-scale solar CHP system based on evacuated flat-plate solar collectors and Organic Rankine Cycle. Energy Convers Manage 2015;90:347–63. [10] Casisi M, De Nardi A, Pinamonti P, Reini M. Effect of different economic support policies on the optimal synthesis and operation of a distributed energy supply system with renewable energy sources for an industrial area. Energy Convers Manage 2015;95:131–9. [11] Chicco G, Mancarella P. Distributed multi-generation: a comprehensive view. Renew Sustain Energy Rev 2009;13(3):535–51. [12] Cho H, Luck R, Eksioglu SD, Chamra LM. Cost-optimized real-time operation of CHP systems. Energy Build 2009;41(4):445–51. [13] Clift R, Doig A, Finnveden G. The application of life cycle assessment to integrated solid waste management: Part I – Methodology. Transactions of the Institute of Chemicals Engineers, Part B: Process Saf Environ 2000;78:279–89. [14] Decree-Law n° 23/2010 of 25 of March (in Portuguese). [15] Dentice d’Accadia M, Sasso M, Sibilio S, Vanoli L. Micro-combined heat and power in residential and light commercial applications. Appl Therm Eng 2003;23(10):1247–59. [16] Directive 2004/8/EC (2004) Directive 2004/8/EC of the European parliament and the council on the promotion of cogeneration based on a useful heat demand in the internal energy market.
A. Safaei et al. / Energy Conversion and Management 97 (2015) 420–427 [27] Ismail MS, Moghavvemi M, Mahlia TMI. Genetic algorithm based optimization on modeling and design of hybrid renewable energy systems. Energy Convers Manage 2014;85:120–30. [28] Kabalci E. A smart monitoring infrastructure design for distributed renewable energy systems. Energy Convers Manage 2015;90:336–46. [29] Kanoglu M, Dincer I. Performance assessment of cogeneration plants. Energy Convers Manage 2009;50(1):76–81. [30] Kong XQ, Wang RZ, Li Y, Huang XH. Optimal operation of a micro-combined cooling, heating and power system driven by a gas engine. Energy Convers Manage 2009;50(3):530–8. [31] Lu H, Yu Z, Alanne K, Xu X, Fan L, Yu H, et al. Parametric analysis of energy quality management for district in China using multi-objective optimization approach. Energy Convers Manage 2014;87:636–46. [32] Lund C, Biswas W. A review of application of lifecycle analysis to renewable energy systems. B Sci Technol Soc 2008;28(3):209–2429. [33] Marantan A, Popovic P, Radermacher R. The potential of CHP technology in commercial buildings; characterizing the CHP demonstration building. ASHRAE Symposium on CHP Technologies for the New Century, ASHRAE Transactions, vol. 108, part 1; 2002. [34] Mavrotas G, Diakoulaki D, Florios K, Georgiou P. A mathematical programming framework for energy planning in services’ sector buildings under uncertainty in load demand: the case of a hospital in Athens. Energy Policy 2008;36(7):2415–29. [35] McCarl BA et al. McCarl Expanded GAMS User Guide. GAMS Release 24.2.1. Washington, DC, USA:GAMS Development Corporation; 2013.
427
[42] Ren H, Gao W. A MILP model for integrated plan and evaluation of distributed energy systems. Appl Energy 2010;87(3):1001–14. [43] Ren H, Zhou W, Nakagami KI, Gao W, Wu Q. Multi-objective optimization for the operation of distributed energy systems considering economic and environmental aspects. Appl Energy 2010;87(12):3642–51. [44] Rezvan AT, Gharneh NS, Gharehpetian GB. Robust optimization of distributed generation investment in buildings. Energy 2012;48(1):455–63. [45] Rosato A, Sibilio S, Scorpio M. Dynamic performance assessment of a residential building-integrated cogeneration system under different boundary conditions. Part I: Energy analysis. Energy Convers Manage 2014;79:731–48. [46] Rubio-Maya C, Uche-Marcuello J, Martínez-Gracia A, Bayod-Rújula AA. Design optimization of a polygeneration plant fuelled by natural gas and renewable energy sources. Appl Energy 2011;88(2):449–57. [47] Safaei A, Freire F, Antunes CH. A model for optimal energy planning of a commercial building integrating solar and cogeneration systems. Energy 2013;61:211–23. [48] Shrestha RM, Marpaung COP. Supply- and demand-side effects of power sector planning with demand-side management options and SO2 emission constraints. Energy Policy 2005;33(6):815–25. [49] Sylva J, Crema A. A method for finding the set of non-dominated vectors for multiple objective integer linear programs. Eur J Oper Res 2004;158(1):46–55. [50] Torchio MF. Comparison of district heating CHP and distributed generation CHP with energy, environmental and economic criteria for Northern Italy. Energy Convers Manage 2015;92:114–28. [51] Ullah KR, Akikur RK, Ping HW, Saidur R, Hajimolana SA, Hussain MA. An experimental investigation on a single tubular SOFC for renewable energy based cogeneration system. Energy Convers Manage 2015;94:139–49. [52] Wang M, Wang J, Zhao P, Dai Y. Multi-objective optimization of a combined cooling, heating and power system driven by solar energy. Energy Convers Manage 2015;89:289–97. [53] Wu YJ, Rosen MA. Assessing and optimizing the economic and environmental impacts of cogeneration/district energy systems using an energy equilibrium model. Appl Energy 1999;62(3):141–54. [54] Xuan J et al. Evaluation of the introduction of a combined heat and power system in a commercial building in Shanghai. J Asian Arch Build Eng 2006;5(2):385–90. [55] Yokoyama R, Shinano Y, Taniguchi S, Ohkura M, Wakui T. Optimization of energy supply systems by MILP branch and bound method in consideration of hierarchical relationship between design and operation. Energy Convers Manage 2015;92:92–104.