A novel renewable polygeneration system for hospital buildings: Design, simulation and thermo-economic optimization

Applied Thermal Engineering 67 (2014) 43e60

Contents lists available at ScienceDirect

Applied Thermal Engineering journal homepage: www.elsevier.com/locate/apthermeng

A novel renewable polygeneration system for hospital buildings: Design, simulation and thermo-economic optimization Annamaria Buonomano a, Francesco Calise a, *, Gabriele Ferruzzi b, Laura Vanoli b a b

DII, Univ. of Naples Federico II, P.leTecchio 80, 80125 Naples, Italy DiT, Univ. of Naples Parthenope, Centro Direz. IS.C4, 80143 Naples, Italy

h i g h l i g h t s
A new solar trigeneration model is designed and dynamically simulated.
The system supplies electrical, cooling and heating energy to Hospital buildings.
The energy demand data are measured for one year.
The economic profitability of the system is satisfactory even without public funding.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 3 October 2013 Accepted 1 March 2014 Available online 15 March 2014

This paper presents an analysis of a possible energy retrofit of an existing University Hospital District, located in Naples (Italy), by using an innovative renewable polygeneration system. This system integrates both Concentrating PhotoVoltaic/Thermal collectors (CPVT) and Solar Heating and Cooling (SHC) technologies. The CPVT parabolic trough collectors are equipped with triple-junction PhotoVoltaic (PV) cells: this technology usually shows ultra-high energy conversion efficiencies. The main components of the system are: CPVT collectors, a single-stage LiBreH2O absorption chiller, storage tanks and balance of plant devices. The system is assumed to be installed at a University Hospital District located in Naples (Italy), equipped with a gas-turbine cogeneration system. The data regarding cooling, heating and electrical demands and productions are measured for a one-year operation. The CPVT produces electrical energy, which is consumed in part by the system parasitic loads, whereas the eventual surplus is fed in the electrical grid. Simultaneously, the CPVT provides heat, used for space heating, for domestic hot water and/or to drive the absorption chiller, which produces cooling energy. The system is designed and dynamically simulated in TRNSYS environment, including detailed and validated mathematical models for the simulation of all the components. The results are analysed from both energy and economic points of views, using parametric analyses and thermo-economic optimizations. The energy performance of the system is excellent since all electrical and thermal energies produced by the renewable system are consumed by the user. The economic results show that the system can be profitable (pay-back period around 12 years) even without any public funding. In case of feed-in tariffs, the system becomes extremely profitable from the economic point of view. The thermo-economic optimization, based on a mixed heuristic/deterministic algorithm, shows that the system profitability can be further improved, increasing solar field area and decreasing storage specific volumes for m2 of collectors installed. Ó 2014 Elsevier Ltd. All rights reserved.

Keywords: CPVT Solar cooling District heating Optimization TRNSYS

1. Introduction Over the last decades, due to the not sustainable trends in energy supply and demand, the energy sector has increasingly

* Corresponding author. E-mail addresses: [email protected], [email protected] (F. Calise). http://dx.doi.org/10.1016/j.applthermaleng.2014.03.008 1359-4311/Ó 2014 Elsevier Ltd. All rights reserved.

focused on energy savings through measures aiming at the reduction of energy demands and the increase of energy efficiency. Both measures play a critical role in addressing environmental and economic goals. Nowadays, in Europe nearly half of the energy consumption is needed in the heating sector and, at the same time, the energy demand for cooling and air-conditioning is rising rapidly [1]. Therefore, the promotion and use of renewable energy heating and cooling systems and equipment have become

44

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

Nomenclature A ACH AHE AHU BOP C CHCP CHP CHW COP ce cg c CPVT CSHCP CT CW DE DH DHC DHW E ETC J G GT hc h H HE HF HS HVAC HW IAM I LHV M _ m OC ORC P PE PEC PEM PTSC PV PVT Q RS S

area Absorption CHiller Auxiliary Heat Exchanger Air Handling Units Balance of Plant concentration ratio Combined Cooling Heating and Power Combined Heat and Power CHilled Water Coefficient of Performance electricity unitary cost [V/kW h] thermal energy unitary cost [V/kW h] specific heat [J/kg K] Concentrating PhotoVoltaic Thermal solar collectors Concentrating Solar Heating Cooling and Power Cooling Tower Cooling Water loop District Energy District Heating District Heating and Cooling Domestic Hot Water electric power [kW] Evacuated Tube solar Collectors cost function [V] incident radiation on the PV surface Gas Turbine convective heat transfer coefficient [W/m2 K] hours [h] enthalpy [kJ/kg] Heat Exchanger Hot Fluid Hydraulic Separator Heating Ventilation and Air Conditioning Hot Water Incident Angle Modifier global solar irradiance [kW/m2] natural gas Lower Heating Value [kJ/Sm3] mass of fluid [kg] mass flow rate [kg/s] Operating Cost [V] Organic Rankine Cycle power production [kWh] Primary Energy [kWh] Primary Energy Consumption [KWh] Polymer Electrolyte Membrane Parabolic Trough Solar Collectors PhotoVoltaic panels PhotoVoltaic Thermal solar collectors heat [kWh] Reference System tank layer surface area [m2]

necessary to fulfil the European targets in the renewable energy sector [2], as well as to significantly contribute to the reduction of the EU’s energy consumption and energy import dependence [3]. As a result, due to a cooperative effort among researchers and government agencies [4], several innovative efficient systems have been investigated, such as: ventilation systems [5], advanced solar cooling systems [6], ground-source heat pumps [7], cogeneration

SCF SHC SPB ST T TK UA V WHE

Solar Collector Fluid Solar Heating and Cooling Simple Pay Back Solar Trigeneration temperature[ C] tank Overall loss coefficient [kJ/(h K)] volume [m3] Waste Heat recovery boiler

Greek letters absorptance [e] control function [e] control function [e] thickness [m] conductivity [W/m K] ε emissivity [e] εHE effectiveness of the load heat exchanger [e] h efficiency [e] hel,t thermoelectric conversion efficiency [e] s StefaneBoltzmann constant [W/m2 K]

a b g d l

Subscripts a outside air dry bulb ap aperture area ACH absorption chiller b beam radiation back back CT cooling tower CPVT concentrating photovoltaic solar collectors conc concentrator cool cooling el electricity f fluid front front gross gross electrical power inv inverter HE Heat Exchanger in inlet mod module connections net net out outlet opt optical pump pump PV PV layer PVT PhotoVoltaic Thermal solar collectors req required sky sky equivalent temperature sub metallic substrate TK tank th thermal energy top top surface area tot total radiation

[8], renewable microgeneration [9], etc. In this framework, the optimal combination of emerging high-efficiency technologies, of renewable energy sources and of district heating and cooling, also coupled with measures to reduce the energy demand of buildings, has to be properly considered when planning, designing, building and renovating industrial or residential areas [4]. The present work focuses on the investigation of a novel solar polygeneration system

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

[10] serving an existing University Hospital district. Here, solar energy is used to simultaneously produce electricity and thermal energy for district heating and cooling and domestic hot water scopes. The technologies included in the investigated trigeneration system are: Solar Heating and Cooling (SHC), District Heating and Cooling (DHC) and Concentrating PhotoVoltaic Thermal collectors (CPVT). Such technologies (SHC, DHC and CPVT) have been separately investigated in literature. Concerning the SHC technology, in order to support the implementation of SHC technologies and to reduce their relative costs, several solar technology design tools have been developed [11]. In particular, advanced modelling and simulation tools of SHC plants have been recently developed with the aim at analysing and optimizing the system layout [12], the control strategy [13] and the components operation [14]. Many authors have been involved in investigations and optimizations of some innovative SHC system configurations including: heat pumps for summer and/or winter auxiliary energy [15], concentrating solar collectors and double effect absorption chillers [8], PVT collectors [16], linear parabolic CPVT collectors [17], dish parabolic CPVT collectors [18], pure plant oil reciprocating engines as auxiliary device [19], solar-biomass hybrid absorption cooling system [20], CHCP system with transcritical CO2 [21] and with an Organic Rankine Cycle (ORC) driven by solar energy [22], etc. Although such research effort, SHC is still at the margin of the market, mostly due to the very high installation cost. From this point of view, the combination of SHC technologies with flat or concentrating PVT systems represents a way to increase the utilization of solar energy, more efficiently during the whole year, and the system costeffectiveness [23]. As above mentioned, the presented paper is based on combined SHC and CPVT technologies. In particular, conventional concentrating solar thermal collectors (e.g. Parabolic Trough Collectors) adopted in SHC systems are here replaced by CPVT, which simultaneously produce electricity and heat [23]. Although one of the most promising applications of PVT collectors consists in their integration in building facades [24], this technology is also very attractive when coupled with some concentrating devices, such as parabolic trough [25] and dish [26]. Recent CPVT prototypes adopt novel PV materials [27], such as multi-junction solar cells, able to approach a nominal efficiency of 40% and to allow the system to operate up to 240  C at reasonable conversion efficiency (slightly lower than 20%) [28]. Literature review shows a small number experimental and theoretical works dealing with CPVT: a novel miniature prototype CPVT, operating at low temperatures, based on a dish concentrator and a silicon PV cell [29]; a high temperature CPVT adopted in solar cooling and water desalination applications [30]; a Parabolic Trough CPVT prototypes [31]. A low-concentrating CPVT system, using truncated parabolic concentrators reflecting sunlight onto the surface of PV cells, operating as the evaporator of a heat pump, was also analysed in Ref. [32]. Experimental results of that study showed that an average COP of 4.8 can be achieved [32]. In Ref. [33] the electrical and thermal efficiencies of Parabolic Trough CPVT collectors, as a function of the PV technology for different concentration ratios, were analysed. In particular, three technologies are investigated: super cell, GaAs cell and concentrating silicon cell arrays. The experimental results show that their respective average electrical efficiencies are 3.63%, 8.94%, and 3.67%, whereas the corresponding thermal efficiencies are 45.17%, 41.69%, and 34.53%. In the authors’ knowledge, although a large number of papers analysed separately CPVT and SHC technologies, literature review regarding CPVT devices coupled to SHC systems is very scarce, as reported in Ref. [34]. Similarly, only few papers analysed the possibility of integrating CPVTs in SHC systems. In particular, the

45

theoretical feasibility of integrating concentrating CPVTs (based on triple-junction cells) with SHC based on a single-effect LiBreH2O absorption chiller, was recently presented in Ref. [35]. Here, the system performance was calculated by using a simplified system layout (including an auxiliary heater and control equipment). The thermal efficiency of the PVT resulted to be stably slightly lower than 60%, while the increase in the device operating temperature produces a negligible reduction of PVT electrical efficiency, it ranged approximately between 19% and 23% as a function of the PVT operating temperature, varying between 65  C and 120  C, respectively. Authors also highlighted that the system can be economically profitable (under certain conditions) and improvable by optimizing the system layout (e.g.: including a storage tank) [35]. The authors of the presented study investigated the CPVT-SHC system economic profitability as function of the system future scale economies and possible public funding [18]. In these papers it was also pointed out that a remarkable surplus of heat is produced through CPVTs. In fact, in all the studies available in references it was concluded that the fluctuation and mismatch of solar radiation and user thermal/cooling demands determine a large amount of heat that often cannot be supplied to the user for space heating and cooling purposes. This surplus could be eventually used to match additional thermal energy demands (e.g. Domestic Hot Water, DHW) in buildings characterized by discontinuous (either on daily or on seasonal terms) activities. As a consequence, the main potential applications, on a smallscale, of such solar Polygeneration systems are: hospitals (which have large and constant cooling and thermal loads), hotels, residential districts, schools, commercial buildings and so forth [36]. From this point of view, several studies pointed out that very high energy and economic feasibility of polygeneration plants can be achieved in hospital buildings or districts. In hospitals, DHC security issues are addressed through CHP and CHCP technologies. From this point of view, a theoretical analysis focused on the feasibility of a hybrid plant applied to a real hospital located in Ferrara, Italy, is presented in Ref. [37]. Here, the environmental benefits achievable through a shift from the adopted conventional system, to various high efficiency hybrid systems directly fed by renewable energy is investigated. A case study of a CHP system for a hospital in Greece is presented in Ref. [38] Here, the authors performed a selection of the most suitable CHP unit, showing that the cogeneration system is most economically profitable when the main gas engine operates 8000 h/year. In this case, the total annual energy cost has been reduced by 32.4% and the Internal Rate Return (IRR) for 20 year lifetime of system is 19% [38]. Possible energy savings in hospitals may be also achieved implementing the pinch technology, as shown in Ref. [39], where the authors estimated a 38% reduction of the thermal power, using four heat exchangers [39]. In Ref. [40] it was presented an optimization model developed for analysing the effects of financial market conditions and energy prices on the optimal structure of a CHCP system, designed for providing energy services for a hospital located in Zaragoza (Spain). The annual energy services demands of this hospital were estimated and expressed on hourly basis of two representative days per month. A dynamic analysis, performed in order to investigate the feasibility of a trigeneration plant serving a hospital located in North Italy, was presented in Ref. [41]. In this paper, the CHCP plant was intended to integrate the existing natural gas fired-boiler central plant. Electric and heat loads for both domestic hot water and process steam were estimated hourly from the monitored consumption, while, space heating and cooling loads were computed, hourly, using the building energy software tool implemented in TRNSYS. Several thermo-economic analyses were also presented in literature. In particular, in Ref. [42], two approaches for the exergoeconomic analyses of a trigeneration system supplying a hospital with

46

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

heating, cooling and power, were presented. Here, the optimization model was based on aggregate energy-flows or energy flow-rates simulated with an approach developed and presented, by the same authors, in Ref. [43]. An economic analysis of a trigeneration system serving a Slovenian hospital was presented in Ref. [44]. This analysis, only focused on the economics of the system and on the optimization of the cooling production, returned that a cogeneration system based on a gas turbine with compression or absorption chillers is the most promising one. Finally, a mixed integer programming model for finding the optimal design and the optimal management strategy of a trigeneration system in a hospital complex were presented in Ref. [45]. Note the adopted model requires the energy load profiles of the user, which were derived with simplified assumptions based on the user typology. No literature works are focused on the innovative solar trigeneration system here presented. Such system is based on the integration of CPVT and SHC technologies, applied to an existing University Hospital District energy located in Naples (Italy). The original idea of the presented paper is matching the measured energy demand of electricity, space heating and cooling (by an absorption chiller), recorded during a 1-year experimental campaign, and domestic hot water of the existing District Energy (DE) through this novel Concentrating SHC and Power (CSHCP) system. In this paper, the solar field consists of novel CPVT prototypes (presented in Ref. [17]). The CPVT produces electrical energy, which is in part consumed by the system parasitic loads, whereas the eventual excess is fed in the electrical grid. Simultaneously, the CPVT provides heat that can be used by the SHC system for space heating, for domestic hot water and/or to drive the singleeffect LiBreH2O absorption chiller, producing cooling energy. In this innovative system layout, with respect to authors’ previous works (e.g. Ref. [26]), a different thermal storage system, consisting of two thermal storage tanks for heating and Domestic Hot Water (DHW) scopes, is also modelled. In particular, this new system layout includes novel control strategies for the DHW management. The system performance is designed and dynamically simulated in TRNSYS environment. Here, a new TRNSYS type for the simulation of DHW tank is developed. To this scope of a new algorithm has been developed in Fortran and linked to the TRNSYS environment. Furthermore, the simulations of the proposed trigeneration system are based on real measured data regarding the user thermal, cooling and electrical demands of the year 2012. To this scope, a 1year experimental campaign has been performed. The results are analysed from both energy and economic points of views, and a parametric analysis and a rigorous thermo-economic optimization of the system, based on a mixed heuristic/determinist algorithm, are also performed. In particular, the parametric analysis aims at evaluating the variations of system energetic and economic performance as a function of the design/operational parameters. This analysis is carried on starting from an initial configuration and varying one parameter per time. Then, a rigorous-thermo economic optimization is also performed. Differently from the parametric analysis, the optimization considers the variation of all the design/ operational parameters simultaneously and aims at determining the values of those parameters which minimize/maximize a selected objective function. In particular, as usual in thermoeconomic theory, in this paper an economic performance parameter is selected as objective function [46]. Therefore, the combination of design parameters found by the optimization process is the one maximizing system profitability. Finally, it is also worth noting that the methodology implemented in the present paper is very different from the common simplified feasibility studies (typically based on average values of the parameters). In fact, the calculations presented in this paper are based on a very complex dynamic simulation model which allows

one to evaluate the variations of each parameter every minute, also including a detailed thermo-economic optimization. Obviously, it must be considered that the presented methodology is very expensive both for the implementation of the model and for the calculations (for example the optimization requires about 1 week of continuous calculations). Therefore, the methodology adopted in this study makes sense only for large, expensive and innovative systems which could not be analysed through well-established approaches. 2. System layout The Solar Trigeneration (ST) system investigated in this paper combines SHC and CPVT technologies. This system is assumed to be used as a repowering of an existing and larger cogeneration system, which serves an existing University District Hospital located in Naples, South Italy. 2.1. District hospital The Federico II University Hospital is located on the hill area of Naples. The hospital, designed and started in the early sixties, was completed in 1972 and it consists of a district of 21 lots of different volumes buildings, as shown in Fig. 1. In each of the 21 lots, several buildings including departments, clinics, services and equipment are located. The hospital is also integrated with the Faculty of Medicine and Surgery of the University Federico II, whose classrooms are located in the same buildings. The district buildings are connected by service tunnels on two levels and by main and secondary roads. The complex has about 2800 beds and covers a total area of 443,000 m2. The buildings total surface area and volume are about 270,000 m2 and 1,130,000 m3. In Fig. 1 an orthophoto of the district area is shown. The high district energy consumptions are due to the specialized services provided by the hospital. In general, countries regulations outline how these requirements may be fulfilled, through proper design and operation of technical systems and the building itself. This is mainly accomplished by regulations for thermal insulation, lighting, indoor temperature levels and ventilation [47]. The high functionalities that the district hospital must guarantee implies the remarkable use of electricity and high space heating and cooling demands. These are determined also by the high infiltrations and air changes required by the strict indoor air quality levels (due to functions as surgery, intensive care units, white rooms, outpatient clinics, etc.). In addition, hospital administrative and executive offices as well as research departments and laboratories are often equipped with dedicated air-conditioning devices (e.g. splits, electric heaters, etc.). The proper evaluation of all the district buildings gains and occupancy, the determination of the thermal parameters of the buildings envelopes, as well as the identification of all energy end-uses within the hospital buildings have been carried out through an energy audit. This in situ investigation, developed by the University of Naples Federico II, aims at planning energy conservation measures to be implemented. Data about thermal transmittances of the main buildings components are shown in Table 1. In this table, for sake of brevity, due to the high number of investigated thermal zones included in the hospital buildings, only the range of variation of external envelope thermal properties, infiltration flow rates and internal gains due to lighting equipment, are reported. In the same table, occupancies, mechanical ventilation air changes, light and equipment loads of the district clinics, surgeries and schools are also reported. In this study, data about the annual demands for district electric-power and thermal energy (for heating and cooling scopes) have been measured during the whole year 2012 and monitored for intervals of one hour. The existing trigeneration plant provides all

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

47

Fig. 1. University Hospital “Federico II”.

energy vectors required by the buildings (power, hot water for domestic uses, hot fluid for the space heating, steam for specific uses, chilled fluid for space cooling). This plant consists of a cogenerative gas turbine for the electricity and thermal production. The gas turbine waste heat is recovered in order to feed energy for heating and domestic hot water scopes and, for cooling ones by feeding an absorption chiller device, as better reported in the next paragraph. Note that, electricity production, district electricity consumptions, district heat and cooling consumptions are all measured, while the consumption due to space heating scopes is calculated by the authors as a difference between the provided total monitored heat consumption and that one calculated for domestic hot water, as a function of the hospital needs. In Fig. 2 the dynamic profiles of monitored electricity (requested (Eel,req) and produced (Eel,CHP)), space heating (Qheat,req) and cooling (Qcool,req) loads and the calculated domestic hot water (QDHW,req) load are shown. It is worth noting that the district hospital presents very high energy consumptions. In particular, it is possible to observe the intensive use of energy for space cooling/heating and for DHW needs. As typical condition for hospitals, the DHW demand is almost constant during the year; the district space heating system results to be operating all over the year (also during the summer season, due to air handling unit post heating coils); the space cooling demand is requested only during summer months. As a consequence, the cooling operating period range from May 19th to November 15th. The electrical energy produced through the existing gas turbine (Eel,CHP) is lower than the relative request (Eel,req) only during some hours of the summer months (JuneeJuly). This occurrence can be ascribed to the increase of cooling energy demanded by dedicated electrical devices. For the investigated district hospital, the annual energy requirements and peak loads are reported in Table 2.

2.2. Existing energy equipment The hospital district is served through a CHCP system which provides electricity, space heating and cooling and domestic hot water. Power and heat produced by the CHCP is supplied to the users via a complex underground pipe network. This CHCP system, connected to the electricity and gas national grids for emergency back-up generation, consists of:
a CHP system, which includes a gas turbine, GT, coupled to a waste heat recovery boiler, WHE, and two auxiliary boilers;
5 absorption single-stage LiBreH2O chillers, ACH, coupled to 8 cooling towers, CT;

The CHP system serves a hospital, which must perform critical lifesaving functions (even when the supply of natural gas and electricity from the utility grid is interrupted). Thus, it is designed to maintain critical life-support systems. In fact, the existing CHP system operates independently of the grid during emergencies and it is capable of black start. The generation of a base electricity load is performed through the GT prime mover, fed by natural gas. From the prime mover the engine exhaust heat is recovered through the WHE, which recovers the waste heat in order to provide steam or hot water for both heating and DHW circuits. In order to bridge the peak heat demand, a number of two dedicated gas-fired auxiliary heaters, can be activated. Part of the produced thermal energy is also used to feed the absorption chillers, ACH, for the production of cold water for space cooling scopes. These ACHs are coupled with the cooling towers, each of them is equipped with four axial fans. Therefore, the energy produced in the main station is then transported to the so called district heating substations in a high pressure water filled extensive underground pipe network. Note that the complex is served through 21 sub-stations, one for each complex buildings. In each sub-station, by plate heat exchangers the required thermal energy is transferred to the end users. The

Table 1 Buildings thermal parameters and schedules. Average external envelope thermal properties

Shading systems Infiltration flow rate Temperature set-points Heating HVAC season Occupation schedule Offices and schools Patients rooms Occupancy Offices Schools Patients rooms People gain Light work Moderate/heavy work Other gain (pc, monitor.) Lights schedules Offices and schools Patients rooms Lights gain

U wall 1.20e2.10 W/m2 K U floor wall 2.07e2.38 W/m2 K U roof wall 0.91e3.28 W/m2 K U windows 4.61e5.80 W/m2 K Aluminium horizontal slats 0.50e3 ACH 20O24  C (winter) and 26e28  C (summer) October 1steApril 30th From 8:00 to 12:00/18:00 from Monday to Friday 24 h all week 0.11 m2/person 1.10 m2/person 0.10 m2/person 100 W/person 120 W/person 20 W/m2 From 8:00 to 21:00 from Monday to Friday 24 h all week 10e20 W/m2

48

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

Fig. 2. University Hospital annual energy demands.

production of cold water for the space cooling is partly decentralized and provided by room dedicated electric chillers. The district buildings are equipped with 2-pipes hydraulic (radiators and fan coils) and aeraulic (AHU units) networks. Note that in the district sub-stations, the temperature level of the thermal energy to be supplied depends on the specific end user target. The use of equipment and devices which require hot fluid for air-conditioning scopes also during the cooling period, combined with the demand for domestic hot water, increases the profitability of the installed CHP system. Since the thermal requests are quite constant during the day, as it typically occurs for hospitals, a high total efficiency is achieved by the CHCP system. In fact, due to the hospital even load profiles occurring almost all year round and to its nearly total utilization of waste heat, very high system efficiency are achieved. Data about the hospital CHCP and HVAC systems are reported in Table 3. 2.3. Solar Trigeneration (ST) system The Solar Trigeneration (ST) system is assumed to be used as a repowering of the existing CHCP system above described. A simplified ST system layout is schematically shown in Fig. 3; here, only the main components are displayed. Regarding the system operating fluids, six different loops are taken into account:
Solar Collector Fluid, SCF: pressurized water flowing from the source sides of the tanks to the solar field;
Hot Fluid, HF: pressurized water flowing from the load sides of the tanks to the devices using solar thermal energy;
Cooling Water, CW: water flowing in the condenser and absorber of the Absorption Chiller (ACH);
CHilled Water, CHW: water flowing in the evaporator of the ACH supplying space cooling devices;
Hot Water, HW: water supplying space heating devices;
Domestic Hot Water, DHW: water supplying domestic devices.


a solar collector field, CPVT, consisting of concentrating parabolic trough solar collectors. Each collector is equipped with a water-cooled receivers placed on the focus of the parabola and covered by a PV multi-junction layer; the device, also equipped with a one-axis tracking system, concentrates on its focus only the beam radiation. The CPVT field can operate up to 100  C and heats the SCF loop.
a LiBreH2O single-effect Absorption Chiller (ACH), heated up by the CPVT solar field. The generator of the ACH is fed via the HF loop, its condenser and absorber are cooled through the CW loop leaving the closed-circuit Cooling Tower (CT), while the evaporator of the ACH supplies Chilled Water (CHW) for space cooling demands;
a closed-circuit Cooling Tower (CT) which provides Cooled Water (CW) to both condenser and absorber of the ACH;

Table 3 Hospital existing CHCP system e design data. Gas turbine Gas turbine power Electrical efficiency Exhaust gas temperature Exhaust gas flow rate Heat recovery boiler Heating capacity Outlet steam temperature Steam pressure Steam flow rate Absorption chillers (ACH) Cooling capacity Rated COP Evaporator flow rate Condenser flow rate

Table 2 Hospital current annual energy consumptions and peak loads.

Space heating Space cooling Domestic hot water Electricity demand Electricity production

The components included in the system layout are diffusely presented in the previous works developed by some of the authors (e.g. Ref. [15]). The main system components are:

Energy [GWh]

Peak load [MW]

44.8 15.4 2.3 31.8 44.2

13.6 7.7 0.6 6.1 6.4

Chilled water temperatures Cooling water temperatures Cooling towers (CT) Nominal capacity Number of fans per CT Spray water flow rate per CT Inlet water temperature Outlet water temperature

5.67 31.5 510 21

MWe %  C kg/s

9 192 12 13

MW  C bar t/h

4  2.5 1  4.7 0.8 439 528 660 1121 12.5e5.5 30e36

MW

8  6.3 4 5840 39.5 29.5

e m3/h m3/h  

C C

MW e m3/h  C  C

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

49

Fig. 3. System layout.


a Thermal Storage system (TK1) that supplies heat for space heating and cooling purposes; it consists of a vertical hot storage tank;
a Thermal Storage system (TK2) that stores the surplus heat from TK1 to be supplied for domestic hot water purposes, it consists of a vertical hot storage tank with an internal heat exchanger;
a fixed-volume pump (P1) for the HF loop; a variable-speed pump (P2) for the SCF loop; a fixed-volume pump (P3) for the CW loop; a fixed-volume pump (P4) for the CHW loop; a fixedvolume pump (P5) for the HW loop;
an inertial chilled/hot water storage tank (TK3), adopted in order to reduce the number of start-up and shut-down events of the absorption chiller ACH;
a plate-fin heat exchanger (HE2) in the HF loop, that transfers heat from the HF to the hot water (CHW) to be supplied to the fan-coils during winter;
a plate-fin heat exchanger in the solar loop, which is used to produce domestic hot water (HE1) when the solar irradiation is higher than the ACH (or HE2) thermal demand;
an Hydraulic Separator (HS), that balances the flows between the primary and secondary hydraulic circuits;
some Balance of the Plant (BOP) equipment (the majority not displayed in Fig. 3 for sake of simplicity), such as pipes, mixers, diverters, valves, and controllers, required for the system operations. The presented ST system was dynamically simulated in TRNSYS environment [48]. The basic operating principle of the ST system can be here summarized as it follows. The running logics of pumps P1, D1 and P2 is briefly summarized in Fig. 4. The solar irradiation is converted into electrical energy and thermal energy through the CPVT field. The produced electrical energy is delivered to the public grid and/or consumed by the user, while the thermal energy heats the CPVT outlet water temperature up to a desired set point Tset,CPVT. The overall CPVT efficiency strongly depends on this temperature; thus, in order to reach its desired set point (which depends on the winter and summer operation modes) a variable speed pump, P2, is used. This pump, P2, is managed by a feedback controller e operating by a secant method e which varies continuously the SCF water flow rate and deactivates the fluid flowing in the CPVT in case

of low radiation (under 50 W/m2), which can potentially determine a cooling effect of the fluid. Generally, when the solar radiation is low and/or the user demand is high, CPVT outlet temperature may be also lower than its set-point. Conversely, in case of scarce thermal energy demand and/or high solar irradiation, the SCF temperature may overcome the set point. In this case, SCF fluid is cooled down 100  C by the HE1, which produces domestic hot water DHW. The activation of the heat exchanger HE1 also prevents CPVT outlet temperature from overcoming the fluid boiling temperature, around 120  C. Therefore, the fluid exiting from CPVT and/ or HE1 may supply either TK1, which stores the produced HF hot stream, or TK2. In particular, this flow is managed through the diverter and mixer, D1 and M1. The control system measures the temperature of top side of tank TK1. When such temperature is higher than Tset,CPVT  DTTK, the flow is switched to TK2 (as shown in Fig. 4). Then, temperature of TK1 decreases due to the heat supplied for space heating or cooling. When such temperature falls outside the fixed dead bands, the control system switches the flow to TK1. Then, this cycling goes on. On the load side of tank TK1, the fluid HF is pumped by using P1 to the primary side of the heat exchanger HE2, during winter, and to the generator of the ACH, during summer. In Fig. 4, Tmin is the minimum allowable temperature for the activation of P1. P1 is managed by a controller, which measures the temperature of the fluid exiting from TK1. When this temperature is lower than a minimum allowable value (Tmin, equal to 75  C), P1 is deactivated, so that solar energy can heat TK1 up to the required minimum values. Therefore, during summer, the temperature of the hot fluid used to drive the ACH is always higher than the minimum value of 75  C. Similarly, in winter, the temperature of the HF going to HE2 is always higher than 75  C. Condenser and absorber of the ACH are both coupled to the CT, which provides cooling water, CW. The produced hot (HW)/chilled (CHW) water passes through an inertial tank TK3 adopted in order to simulate the piping capacity. HW and CHW are supplied to heating and cooling devices, which provide the demanded space heating and cooling energy. In particular, pumps P5 (in winter) and P4 (in summer) supply HW and CHW to the end user. The energy produced through the ST system, in terms of electricity, space heating and cooling and domestic hot water, is so delivered to the user. As always occurs for renewable systems, the consumption of the energy produced by the ST is assumed to be

50

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

not exploded for sake of brevity, include: i) the CPVT model for the calculation of power, Pel,CPVT, and thermal, QCPVT, production by solar, ii) temperatures controllers for the system management, which are highly connected to the other sub-models, iii) the storage management systems and iv) the thermal and cooling energy production system. Obviously, these blocks are highly coupled to each other in order to allow the simulation to be executed. The ST system behaviour is clearly described in Section 4. As above mentioned, the investigated system originates from layouts developed in previous works, where built-in and user-developed components are described in detail, e.g. Ref. [34]. The majority of the models used for the components (e.g. pumps, mixers, diverters, valves, controllers, auxiliary heater, absorption chiller, cooling tower, plate-fin heat exchanger, building, etc.) were taken from the software library. For the CPVT model, a new TRNSYS type has been developed by the authors in Fortran and then linked to the software. The model of the CPVT was recently presented in literature and it will be here shortly resumed [17]. Moreover, in this paper additional models have been implemented in order to carry out the simulations. In particular, a novel model of a stratified storage tank with internal heat exchangers (used for DHW preparation) is here developed and added to the TRNSYS library. Then, specific models have been also developed for the calculations of primary energy savings and for the evaluation of the economic profitability of the system. Note that all adopted models of the components, previously developed by the authors or included in the software library, are validated against experimental data. Therefore, in this section only a brief description of some of the main new components used for the system simulation is provided. 3.1. Parabolic trough solar thermal collectors [17] The concentration ratio is defined as the ratio between the aperture area, Aap, of the concentrator and the area, APVT, of the two PV triangular sides (sketch reported in Ref. [17]) of the receiver:

CPVT ¼

Aap APVT

(1)

The optical efficiency (hopt) of the concentrator is assumed being constant [35]. Therefore, the radiation incident on the PV surface is:

GPVT ¼ APVT Ib CPVT hopt IAMth

(2)

priority with respect to the remaining available energy flows (gasturbine, gas boiler, etc). The design data of the ST system components are reported in Table 4.

As commonly done in concentrating systems, in the previous equation only the beam incident radiation (Ib) is considered. This radiation is corrected by considering both the optical efficiency of the receiver and the Incidence Angle Modifier (IAM) [49]; it takes into account that the incident radiation decrease when the angle of incidence increases. The IAM, related to the thermal production, is evaluated on the basis of data experimentally calculated by Bernardo et al. [50]. Simultaneously, additional thermal energy is absorbed by the top thermal absorber.

3. System model

Qtop ¼ Atop Itot atop

Fig. 4. Running logics of P1 and P2 as a function of the CPVT and TK1 temperatures.

As mentioned before, the analysis is performed by a complex scientific approach, including detailed modelling of all the components of the system. All these components and the system as a whole have been dynamically simulated in TRNSYS environment, which includes a large library of built-in components, often validated with experimental data [48]. A schematic block diagram, showing the top level of the TRNSYS model, is reported in Fig. 5. The simulation model is split in several sub-models or blocks, as summarised in Fig. 5. The main sub-models reported in the figure, and

(3)

In this case, the top surface area, Atop, can convert both beam and diffuse radiation, i.e. the total radiation (Itot), since the insolation incident on that surface is not concentrated. Assuming the top surface area as a grey surface and considering that the area of the top surface is much smaller than that of the sky, the radiative heat transfer between the top absorber and the sky can be calculated as it follows [49]:


 4 4 Qtopsky ¼ Atop εtop s Ttop  Tsky

(4)

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60


 4 4 QPVTconc ¼ APVT sεPVT TPVT  Tconc

Table 4 Main design parameters for the solar trigeneration system. Parameter

Description

Value

Unit

cf NCPVT ACPVT VTK1 VTK2 Tset,CPVT,summ

Water specific heat Number of CPVT collector CPVT aperture area TK1 volume/CPVT area TK2 volume CPVT outlet set-point temperature, summer CPVT outlet set-point temperature, summer TK2 activation deadband TK3 volume P4, P5 flow P4 flow/number of CPVT collectors ACH rated CHW temperature difference ACH rated cooling capacity, PACH ¼ 1.3QP4cfDTrated P2 flow/CPVT area Number of Tank nodes Tank height P1 flow P3 flow TK1 temperature minimum temperature for system activation, summer TK1 temperature minimum temperature for system activation, winter TK1 temperature for switching to TK2, summer TK1 temperature for switching to TK2, winter DHW nominal inlet temperature

4.187 400 12 20 40 90

kJ/kg K e m2 l/m2 M3  C

90



5 20 2.08 105 520 5 1215



Tset,CPVT,wint

DTTK VTK3 QP4 ¼ QP5 qP4 DTrated PACH qP2 nTK HTK QP1 QP3 Tmin,summ

Tmin,wint TTK,set, summer TTK,set,winter Tset,DHW

C

C m3 kg/h kg/h  C kW

100 kg/h m2 5 e 2 m 5 1.08 10 kg/h 4.85 105 kg/h  75 C

75



85 85 55



C

C C  C 

Here, the sky equivalent temperature (Tsky) is calculated using a routine included in the software. Ttop is the temperature of the top surface. Similarly, assuming the area of the concentrator much bigger than the one of the PVT receiver and assuming both PVT and concentrator as grey surfaces, the radiative heat transfer between the PVT and the concentrator is calculated as [49]:

51

(5)

where TPVT and Tconc are respectively the PVT and the concentrator surfaces temperatures. The convective heat transfer between the PVT and the air is calculated as it follows [49]:

Qconv;PVT ¼ APVT hc;PVT ðTPVT  Ta Þ

(6)

The forced convection between the top absorber and the air is [49]:

  Qconv;top ¼ Atop hc;top Ttop  Ta

(7)

The gross electrical power produced by the PV layer is:

PPVT;gross ¼ CPVT APVT Ib hopt hPV IAMel

(8)

Note that this energy is calculated assuming the concentrated beam radiation incident on the PV layer (corrected through the concentrator optical efficiency and by the IAM coefficient) corrected through the electrical efficiency of the PV, hPV. The electrical efficiency of the triple-junction PV (hPV) is experimentally related to the concentration ratio and to the temperature [35]:

hPV ¼ 0:298 þ 0:0142 lnðCPVT Þ þ ½  0:000715 þ 0:0000697 lnðCPVT ÞðTPVT  298Þ

(9)

Note that this equation returns ultra-high values of electrical efficiency, also approaching 40%, as usual in IIIeV PV cells. The IAMel is also evaluated on the basis of the experimental data provided by Bernardo et al. [50]. The net power produced by the system is reduced of the amount of electricity lost in the module connections and in the inverter, considering the corresponding efficiencies (hmod and hinv) [35].

PPVT;net ¼ PPVT;gross hmod hinv

(10)

Finally, the heat absorbed by the cooling fluid is:


 _ f Hf ;out  Hf;in Qf ¼ m

(11)

In the previous equation, the enthalpies of the inlet and outlet cooling fluids (Hin and Hout) are calculated by the thermo-physical property subroutine included in TRNSYS environment. Therefore, the overall energy balance on a control volume that includes the entire triangular receiver is:

APVT Ib CPVT hopt IAMth þ Atop Itot atop _ f ðHout  Hin Þ þ CPVT APVT Ib hopt hPV IAMel ¼ m
 4 4 þ APVT Ib CPVT hopt IAMth rPVT þ Atop εtop s Ttop  Tsky
 4 4 þ APVT hc; PVT ðTPVT  Ta Þ  Tconc þ APVT sεPVT TPVT   þ Atop hc; top Ttop  Ta

(12)

A second energy balance takes into account the control volume that includes the metallic substrate and the fluid channel, assumed as a heat exchanger. The energy balance for this heat exchanger is:

_ f ðhout  hin Þ ¼ εHE m _ f cf ðTsub  Tin Þ m

Fig. 5. Top level block diagram of the ST system model.

(13)

where Tsub is the temperature of the metallic substrate. For these given boundary conditions (inlet temperature and mass flow rate, beam and total radiations and relative angle of

52

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

incidence, environment and sky temperature, ambient pressure and wind velocity), the unknowns are five, namely: PVT temperature, substrate temperature, fluid outlet temperature, temperature of top receiver surface (facing the sky) and temperature of the concentrator. Therefore, three more equations, in addition to Eqs. (12) and (13), must be considered. The third of the required five equations is derived from an energy balance on a control volume including the PVT layer, and the metallic substrate.

T  Ttop T  Tsub _ f ðHout  Hin Þ þ Atop sub APVT PVT ¼ m rPVTsub rtop

(14)

A fourth energy balance can be considered with respect to the control volume that includes the top side of the substrate and the top surface of the triangular receiver:

Atop


 Tsub  Ttop 4 4 þ Atop Itop ¼ Atop Itop rtop þ Atop εtop s Ttop  Tsky rtop   þ Atop hc;top Ttop  Ta (15)

Finally, the last energy balance takes into account the control volume that includes only the parabolic concentrator.


 4 4 þ Itot Aconc aconc APVT sεPVT TPVT  Tconc
 4 4 þ Aconc hc;conc;front ðTconc  Ta Þ ¼ Aconc sεconc;back Tconc  Tsky þ Aconc hc;conc;back ðTconc  Ta Þ

Mn cf

DP X     dTn;2 _ f cf Tn1;2  Tn;2 þ Tn;2  Tnþ1;2 ¼ m dw dp ¼ 1

   Sn l  Tn1;2  Tn;2 þ Tnþ1;2  Tn;2 Sn G   X   gj UAjn Tn;j  Tn;2 þ UAn Ta  Tn;2 þ þ

j¼1

The first sum on the right hand side of Eq. (17) represents the heat transfer caused by the mass flows. Here, DP is the number of mass flow rates which can occur in one segment of the n-th store node. The second sum accounts the conductivity between the vertical boundary layers of the store. The heat loss toward the surroundings is accounted by the third term. Here, UAn represents the heat transfer capacity rate between the store n-th node and the outdoor environment. Thus, the fourth term on the right hand side of the same equation represents the heat transfer between the heat exchanger nodes, on the boundary vertical partitions, and the store ones. Here, gj ¼ 1 if the n-th store node is in contact with the horizontal boundary n-th node of the heat exchanger, otherwise gj ¼ 0. UAjn is the heat transfer capacity rate between the i-th node of the heat exchanger and the store. The energy balance for a n-th heat exchanger (lefti,1 and righti,3 vertical partitions) node is given by the following equation. Note that a left side heat exchanger is considered.

Mn cf;HE

    dTn;1 _ f;HE cf;HE Tn1;1  Tn;1 þ Tn;1  Tnþ1;1 ¼ m dw     þ UAn Ta  Tn;1 þ g1 UA1n Tn;2  Tn;1

(16)

3.2. Storage tanks (TK1 and TK2) Two types of storage tanks are adopted: i) fluid-filled sensible energy storage tanks and ii) storage tanks with internal heat exchangers and connections for charge and discharge. The first type of storage tank is adopted for solar energy storage and it is modelled by using the TRNSYS built-in Type 4d [48]. The second type of storage tank, adopted for domestic hot water scopes, is a stratified fluid storage tank with an internal heat exchanger and connections for direct charge and discharge, as briefly reported in the following. Some of the assumptions of this second storage tank model are: - the tank is subjected to thermal stratification; therefore, it is divided into N fully-mixed equal sub-volumes. As a result, the tank thermal losses toward the outdoor ambient can be individually specified [51]; - the effect of stratified charging or discharging is modelled by assuming variable inlet positions such that the entering fluid may be added to the tank node at a temperature as nearly equal to its own temperature as possible. Note that for the stratified charging the water inlet position is set above the outlet one, vice versa for the stratified discharging. The temperatures of the N store nodes are calculated on the basis of unsteady energy and mass balances and they are obtained by solving a set of differential equations similar to the next. Here, the change of internal energy with the time occurring in the n-th store node can be calculated as it follows:

(17)

(18) In Eq. (18) the vertical temperature boundary nodes, Tn1,1 or Tnþ1,1, which no heat exchangers are connected to, represents the inlet and outlet temperatures of the considered heat exchanger. 3.3. Economic model One of the main goals of this paper is the thermo-economic optimization of the analysed system aiming at promoting its commercialization. To this scope a detailed economic model was implemented: here, capital and operating costs and several indexes required for evaluating the economic profitability of the system are taken into account. Some equations for the calculation of the capital costs of the components were presented by the authors in previous studies (e.g. Ref. [16]). However, these equations cannot be used for the case investigated in this paper due to the different sizes of the components. Therefore, new cost functions, derived from manufacturers’ data, valid for the selected capacities (Table 4) of the components were developed. In order to estimate the investment cost (Ji) of the different components, the next equations were used:

JCPVT ¼ 600ACPVT

(19)

3 2 3  0:0393PACH þ 244:53PACH þ 95494 JACH ¼ 105 PACH

(20)

3 2 þ 0:0276PCT þ 11:709PCT þ 33; 554 JCT ¼ 5,106 PCT

(21)

JTK ¼ 494:9 þ 0:808VTK

(22)

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

53


 2 Jpump ¼ 1:08  0:00000002Qpump þ 0:0285Qpump þ 388:14 (23)  JHE ¼ 150

AHE 0:093

0:78 (24)

The yearly savings in terms of operating costs are strictly related to the amounts of energy (electrical, cooling and thermal) delivered to the user. As regards the electrical energy, the operating costs are calculated by taking into account also the electricity flows exchanged with the grid. In fact, the model considers the eventual purchase/selling from/to the grid at the current price or cost, respectively, on the basis of a time-dependent tariff system. In Italy, as in many other countries, the cost of electric energy is calculated according to time-dependent tariffs. In particular, three different timeframes are established, namely: F1, F2, F3, corresponding to peak, medium and off-peak periods. The cost of electricity to be purchased was assumed equal to 0.14 V/kW h, 0.11 V/kW h and 0.090 V/kW h, in F1, F2 and F3, respectively. The prices for electricity sale were equal to 0.090 V/kW h, 0.070 V/kW h and 0.045 V/ kW h, in F1, F2 and F3, respectively. Then, the model compares the present time step with the time-dependent cost map, in order to evaluate the corresponding time-frame (F1, F2 or F3). Finally, the model evaluates the present cost of electricity [16]. Conversely, the users of the considered District Hospital pay a constant price both for thermal energy and for cooling energy. Therefore, the savings due to the production of space heating and Domestic Hot Water are proportional to the corresponding amount of energy produced and delivered to the user. The same occurs for the cooling energy. The considered costs of thermal (cut) and cooling energy (cuc) are 0.075 V/kW h and 0.0433 V/kW h. Therefore, the savings in terms of operating costs are:

OC ¼

X

Eel;net;i ce;i þ Qheat cut þ QDHW cut þ Qcool cuc

(25)

i

The economic profitability is calculated by using the Simple Pay Back (SPB) index [16]. 4. Results and discussion As discussed in the previous sections, the aim of this paper is the design and the simulation of a novel Solar Trigeneration, ST, system to be installed at the District Hospital of Naples, Italy. The scope of this system is to improve the energy conversion efficiency of the Hospital power plant. A base case study was developed using the Meteonorm weather data of Naples, South of Italy. The main design/ operational parameters are summarized in Table 4. It is worth noting that, for sake of brevity, such table summarizes only the main parameters used in the simulations. In fact, the complete set of input parameters required to run the simulations is significantly larger, being very high the number of components included in the system. The capacities of the components included in the ST system are accurately selected in order to cover only a small part of the overall heating, cooling and electrical demands. This implies the full utilization of the thermal energy produced through the ST system and the consequent improvement of its economic profitability. In particular, the nominal electrical power of the ST system was selected approximately at 800 kW. As a consequence, the nominal thermal power is about 2.0 MW, which is sufficiently lower than the Hospital demand. It is worth noting that a solar system with a capacity designed on the nominal electrical power demanded by the Hospital (about 6.1 MW) would require an enormous available area for the installation of the solar collectors.

Fig. 6. Temperature: summer day.

Further details about the specific design parameters can be found in Refs. [17,52] for the CPVT collector and Ref. [16] for other components. Note that some of the design parameters shown in the previous table are closely related to each other. All components were designed parametrically and resulted automatically re-sized as a function of the solar field capacity. Thus, a variation of the number of solar collectors determines a proportional variation of several design parameters, such as ACH and HE capacities, TK volumes, etc. Similarly, a variation of qP4 determines a corresponding proportional variation in the capacity of pumps, ACH and HE, whereas the size of the CPVT field is not affected by such parameter. Note also that in this analysis the CPVT outlet temperature is set at 90  C during the summer and 80  C in the winter. In previous studies, lower winter operating temperatures (around 45e50  C) were selected in order to improve both CPVT thermal and electrical efficiencies. Such selection is due to the fact that the ST must supply both Hospital high temperature (radiators) and low temperature (fancoils and air handling units) HVAC devices. As a result, during winter, the CPVT outlet temperature was kept as high as possible. Note that, when the top temperature of TK1 is lower than 75  C, no heat is supplied for space heating and cooling by the ST. In fact, in winter this is the minimum temperature required by the radiators; conversely, in summer, 75  C is the minimum temperature of the hot water driving the absorption chiller. Obviously, when TK1 temperature is below 75  C, the user will be supplied by the existing gas-fired trigeneration system and the solar loop will increase TK1 temperature up to the desired set-point temperature. As typically occurs in hospitals, the DHW supply temperature was set at 55  C. When the temperature of TK2 is lower than this set point, no renewable DHW is delivered to the user. This control of the TK2 is managed through an on/off controller which operates using dead bands of 3  C and it executes the following actions: it closes the DHW supply valve when the TK2 temperature reaches 58  C and opens the valve when the TK2 temperature goes down to 52  C. As mentioned before, the simulation tool developed in this paper includes hundreds of components. Namely, the system layout includes about 200 TRNSYS types. In addition, the tool also includes detailed energy demand data for the District Hospital under investigation. As a consequence, each simulation returns a huge amount data, such as dynamic profiles of powers, temperatures and mass flow rates. Furthermore, the tool also allows one to integrate results on whatever time basis (days, weeks, months or year). This capability is very useful for a better interpretation of the results since the integration allows one to mitigate the oscillations achieved during the dynamic operation.

54

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

Obviously, it is not possible to report all the tremendous amount of data calculated in the paper. Therefore, for sake of brevity, results are summarized presenting temperature and powers profiles for a representative summer day (the winter one is omitted). A weekly integration is also presented in order to show the variation of the parameters during the year. Finally, yearly results are presented including both a sensitivity analysis and a thermo-economic optimization aiming at determining the values of the system design parameters, minimizing the pay back period. 4.1. Summer day The selected representative summer day is July 30th. In particular, Fig. 6 shows the temperature plots from 0.00 (5040 h) to 24.00 (5064 h). Here, the operation of the controllers managing the pump P2 and the mixers and diverters (M1 and D1) installed upstream to TK1 and TK2 is clearly shown. When there is no radiation incident on the CPVT collectors, the water temperatures of both tanks TK1 and TK2 decrease due to thermal losses. Approximately around 6.30 am, the radiation sensors activate the solar loop, as a consequence an increase of the temperature exiting from the CPVT (Tout,CPVT) occurs. All the heat produced by the CPVT field is delivered to TK1, until its top temperature is below the fixed switching set-point (85  C). Therefore, the temperature of TK1 increases and the temperature of TK2 keeps on decreasing. From 6:30 to 7:30, while TK1 is heated by the CPVT, some heat is demanded by the ACH, which produces chilled water at 6.5  C (Tin,user). Because TK1 is driving the ACH, around 8.00 am the top temperature of TK1 falls below the minimum allowable temperature (75  C). Such occurrence determines the ACH stand-by with the consequent solar heating of TK1. In the early morning, the balance between the heat delivered by the CPVT and the heat demanded by the ACH, implies a stable operation of TK1 around a top temperature of 78  C. Simultaneously, in order to keep the CPVT outlet temperature at 90  C the variable speed pump P2 continuously varies the flow entering in the CPVT. The fluctuations shown in such figure are typical of the operation of feedback controllers. Subsequently, around 10.00 am, the increase of the beam radiation incident on the CPVT implies a corresponding increase of the CPVT heat production and, consequently, of the heat delivered to TK1. Therefore, around 11:30, TK1 top temperature overcomes the switching set point (85  C) and the heat produced by the CPVT is delivered to TK2, determining an increase of its bottom and top temperatures (TTK2,top and TTK2,bot) as shown in the figure. While no heat is supplied by the CPVT to TK1, ACH keeps on demanding heat to TK1, thus the temperature of TK1 decreases and, consequently, mixers and diverters switch to TK1 again. This cycle is clearly shown in Fig. 6, where these fluctuating trends of TK1 and TK2 temperatures are displayed. This hysterical behaviour is determined by the control strategy implemented in this work. In fact, as above mentioned, P2 flow can supply alternatively TK1 or TK2. In particular, when TK1 top temperature is higher than 85  C, P2 flow is switched from TK1 to TK2. Then, when TK1 top temperature falls below 83  C (2  C dead-band), P2 flow is switched again to TK1. This strategy is implemented using TRNSYS Type2b, ON/OFF Differential Controller. As a consequence, TK1 top temperature (TK1,top) oscillates between 83  C and 85  C, determining also a corresponding oscillation of the temperature of the hot stream exiting from ACH (TACH,hot,out). Obviously, the oscillation of TK1 top temperature (TK1,top) show effects on the temperature profiles of all nodes of the stratified tank TK1 (e.g. TK1,bottom), as shown in Fig. 6. The dynamic behaviour of the system can be also and better interpreted observing the heat flows shown in Fig. 7. Here, the trend of the heat produced by the CPVT (QCPVT) is highly affected by the availability of the beam solar radiation. This figure also shows

the ACH deactivation around 8.00 am, above discussed, and the heat demanded to TK1 by the ACH by TK2 for DHW production (QTK1 and QTK2). Note that, in the first hours of the morning the amount of heat demanded to TK1 is higher than the one produced by the CPVT. In fact, during these hours the TK1 storage capacity is partly used to supply heat to the ACH. Conversely, the opposite occurs during the central hours of the day. Note also that, the heat is demanded to TK2 only after 12:45, when the top temperature of TK2 reaches the minimum allowable level (55  C) required by the user. In the same figure it is also shown that the cooling power produced by the ACH (Qcool) is significantly lower than the total cooling power required by the whole District Hospital (Qcool,req). Obviously, the availability of beam radiation also affects the trends of the electrical power produced by the CPVT (Pel,CPVT), which is one order of magnitude lower than the one required by the whole District Hospital (Pel,req), as shown in Fig. 8. 4.2. Weekly analysis The integration of the results on a weekly basis allows one to obtain a better interpretation of the results due to the mitigation of all the fluctuations shown in the previous figures. In Fig. 9, thermal and electrical energy flows related to the CPVT loop are shown. Note that the solar loop considered in this work is equipped with a concentrating solar collector (CPVT) whose performance is particularly sensitive to the availability of beam radiation. As a consequence, during the winter both CPVT thermal (QCPVT) and electrical (PCPVT) productions strongly decrease. Note also that both total (Itot) and beam (Ibeam) radiations shown in such figure are referred to the CPVT aperture area. The CPVT is NeS oriented using a single-axis tracking system. This arrangement maximizes the radiation incident during summer, reducing the one incident in winter. This circumstance further contributes to increase the spread between the winter and summer performance (both thermal and electrical) of the CPVT under investigation. Fig. 9 also shows the plot of the electrical energy demanded by the entire District Hospital, which is higher in summer. In fact, some end users of the hospital are not connected to the district cooling loop, determining an increase of the demanded electrical energy due to the operation of the electrical heat pumps. A similar increase is also observed in the winter coldest periods when the heat provided by the district heating system is not sufficient and the additional heat is obtained by using electrical space heating devices. The trend of the weekly thermal energy produced by the CPVT also affects the plots of the weekly thermal and heating, cooling and DHW energies produced by the solar system (Figs. 10, 11 and 12). In particular, Fig. 10 shows that the

Fig. 7. Summer day: heat flows.

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

55

Fig. 8. Summer day: electrical powers.

Fig. 10. Weekly analysis: thermal energy.

thermal energy produced by the proposed system (QHE2) is largely lower than the one demanded by the user (Qheat,req), as before reported. Unfortunately, HE2 thermal production dramatically decreases during the coldest weeks, where the heat demand increases. Conversely, Fig. 11 shows that the maximum cooling energy demand (Qcool,req) is simultaneous with the ACH cooling production (QACH), even if the capacity of the ACH is absolutely low for the magnitude of the cooling energy demanded by the District Hospital. In addition, Fig. 11 also shows that all the heat required to drive the ACH is supplied by the tank TK1. Fig. 12 shows that the DHW heating demand is dramatically larger than the DHW produced by the system. The maximum DHW production is typically achieved during the middle seasons where the thermal energy produced by the CPVT is high and the system cooling and heating demands are lower. Therefore, the tank TK2 is rarely heated by the CPVT and the heat exchanger HE1 operates only when the solar loop temperature overcomes 100  C, which is a circumstance rarely achievable only during the summer season. Therefore, the majority of the thermal energy produced by the CPVT is delivered to TK1 all year long. In fact, Fig. 13 shows the ratios between HE1, TK1 and TK2 thermal energies with respect to the one produced by the CPVT. This Figure shows that the contribution of HE1 is marginal, whereas the heat supplied to TK2 is significantly lower than the one used by TK1. This implies that the temperature of the tank TK1 is for long periods below the set-point fixed for switching to TK2. Finally, Fig. 14 shows the weekly trends of some of the most significant system performance parameters. Here, both CPVT thermal and

electrical efficiencies are calculated with respect the total radiation (hel,CPVT and ht,CPVT) and the beam radiation (hel,CPVT* and ht,CPVT*). The electrical and thermal efficiencies, calculated with respect to the beam radiation, oscillate around 20% and 50%, respectively. Slight reductions of the CPVT efficiencies are observed during winter as a consequence of the higher thermal losses toward the environment and lower available beam radiation. As expected, this reduction is much more significant for the thermal efficiency which is much more sensitive to the thermal losses to the environment. The COP of the ACH is also stable due to control management of the ACH inlet temperature.

The results of the annual simulations are summarized in Table 5, where both thermal and electrical energies are shown. According to the trends discussed above, the cooling energy is much higher than the thermal output, due to the poor thermal performance of the CPVT during the winter season. DHW production (QHE1 and QTK2) is definitely marginal with respect to the heat used for space heating and cooling purposes (QTK1). In fact, the TK1 top temperature is rarely higher than 85  C, because the continuous heat demand due to space heating and cooling. Such table also clearly shows that all the electrical and thermal energies produced by the system are largely lower than the ones demanded by the user. This circumstance allows one to achieve a full utilization of the solar energy. Finally, such table shows that the majority of the electricity

Fig. 9. Weekly analysis: thermal and electrical energy, solar loop.

Fig. 11. Weekly analysis: cooling energy.

4.3. Yearly analysis

56

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

Fig. 12. Weekly analysis: domestic hot water.

produced by the system occurs during the peak hours (F1), whereas its contribution during the off-peak hours (F3) is marginal since these hours often occur during the night. Conversely, the electricity demanded by the user in F3 is high too. The overall energetic and economic performance parameters are summarized in Table 6. Such results show the high value of the electrical efficiency, due to the use of IIIeV PV cells. Thermal efficiency, calculated with respect to the incident beam radiation, is very high, too. The capital cost of the system under investigation is estimated in about 3.6 MV. However, a saving of about 0.24 MV/ year is obtained, determining a SPB period of about 15 years. This result is achieved not considering any type of public funding. However, if a feed-in tariff of 0.30 V/kW h (similar to the one adopted in Italy for PV systems), the SPB period becomes 6.1 years. Considering also a possible feed-in tariff of 0.10 V per kW h of thermal energy produced by the CPVT, the SPB becomes 4.1 years. 4.4. Sensitivity analysis and thermo-economic optimization A comprehensive sensitivity analysis was developed, in order to analyse the dependence of the performance on the most significant design variables and/or boundary conditions, when all the remaining parameters remain fixed. Among these variables, for sake of brevity, the sensitivity analysis will be shown only with respect to the following parameters: CPVT area, TK volumes (VTK1), flow rate of P4 (qP4). The most important parameter for the design of the polygeneration system under investigation is the number of CPVT

Fig. 13. Weekly analysis: CPVT thermal energy utilization.

Fig. 14. Weekly analysis: performance parameters.

collectors and, thus, the total CPVT field area. According to the design procedure developed in this paper, a variation of the CPVT area determines a corresponding variation of several additional design parameters. In particular, increasing the CPVT area, TK1 volume also increases, being fixed the ratio between TK1 volume and CPVT area (VTK1). As a consequence, due to the constant ratio between TK1 and TK2 volumes (equal to 10), TK2 volume also increases. Then again, the maximum flow rate of pump P2 also increases being fixed the parameter qP2. Similarly, an increase of CPVT area also determines an increase of P4 flow rate, being the parameter qP4 fixed. The capacity of the ACH also increases since it is strictly related to the nominal flow rate of pump P4. It can be concluded that, according to the procedure implemented in this work, a variation of CPVT area determines a re-sizing of the majority of the components included in the system. The results of this sensitivity analysis are provided in Fig. 15. Here, it is clearly shown that an increase of CPVT area determines a negligible decrease of its electrical and thermal efficiencies due to the slightly higher operating temperatures achieved in case of larger CPVT solar field. The graph also shows that the amount of CPVT heat delivered to the tank TK1 is almost constant (only a slight reduction is detected). In fact, as discussed before, an increase of CPVT area determines a

Table 5 Annual results: energy flows. Parameter

MWh/year

Itot Ibeam PCPVT Paux QCPVT QC,ACH Qh,ACH QHE2 QTK1 QTK_DHW QHE1 Qheat,req Qcool,req QDHW,req Pel,net Pel,req Pel,req,F1 Pel,req,F2 Pel,req,F3 Pel,net,F1 Pel,net,F2 Pel,net,F3

9.22Eþ03 5.67Eþ03 1.15Eþ03 1.60Eþ02 2.91Eþ03 1.50Eþ03 1.83Eþ03 8.12Eþ02 2.65Eþ03 2.13Eþ02 1.99Eþ01 7.98Eþ02 1.49Eþ03 1.49Eþ03 1.14Eþ03 3.17Eþ04 1.22Eþ04 7.51Eþ03 1.20Eþ04 7.80Eþ02 1.98Eþ02 1.60Eþ02

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60 Table 6 Annual results: economic and energetic parameters. Parameter

Value

Unit

DC

2.53Eþ02 3.60Eþ03 14.25 0.205 0.126 0.531 0.327 0.82

kV/year kV years / / / / /

Jtot SPB

hel,CPVT hel,CPVT* ht,CPVT ht,CPVT* COPACH

proportional increase of ACH and HE2 capacities, resulting in a slight variation of ratio between TK1 and CPVT thermal energies. In fact, Fig. 15 also shows that all the thermal, cooling and electrical energies linearly increase for higher CPVT area. This result is caused by the above discussed sizing procedure. The economic results show that the SPB periods rapidly decreases for higher CPVT area, showing that the larger the solar field the better the economic performance. This result is due to the fact that the capital cost assumed for the CPVT is sufficiently balanced by the savings due to its thermal and electrical productions. This is also due to the fact that, in the considered range of variation, all the heat produced by the CPVT is used by the thermal/cooling loads. However, for extremely large CPVT solar fields, some amount of heat should be rejected being higher than the one demanded by the users. This would result in a non-monotonic trend of the SPB periods, not shown in Fig. 15 due to the restricted range of variation of CPVT area. Fig. 16 shows the sensitivity analysis as a function of VTK1, i.e. the ratio between TK1 volume and CPVT area. As a consequence, an

Fig. 15. Sensitivity analysis: CPVT area.

57

increase of this parameter determines a corresponding increase of both TK1 and TK2 (being the ratio between TK1 and TK2 volumes fixed). Results show that both electrical and thermal efficiencies of the CPVT are scarcely sensitive to the increase of TK volumes. In addition, an increase of TK volume determines a slight increase of the ratio between the heat delivered to TK1 and the heat produced by the CPVT, which increases approaching an asymptotic value around 0.72. This is basically due to the summer operation, as shown in Fig. 16, in which the cooling energy produced by the system increases for higher TK1 volumes. Conversely, the winter thermal energy produced is maximum around 40 L/m2. As a consequence, DHW production decreases for larger TK1 volumes. This result is in accordance with previous results found in literature. In fact, it is well known that larger TK1 volumes improve the system storage capacity, increasing the amount of solar energy delivered to the user. On the other hand, the larger the volume of TK1, the higher the thermal losses toward the environment. From the economic point of view, results show that the best configuration (lowest SPB) is achieved for the smaller TK volumes. Note that this is due to the fact that the capacity of the solar system under investigation is significantly lower than user demand. Therefore, the need of a storage system is marginal since the user will always consume all the thermal energy produced by the solar system. Obviously, when the ratio between CPVT and ACH/HE2 capacities varies, these results may significantly change. This analysis is shown in Fig. 17. Note that, for the considered design procedure, a variation of qP4 does not affect the capacities of the components included in solar loop (CPVT, P2, TK1 and TK2) but determines a proportional increase of the capacities of ACH and HE2, also causing a possible increase of the space heating and cooling production. In fact, an increase of qP4 determines an increase of the heat demanded to TK1 by ACH and/or HE2, as clearly shown in Fig. 17. As a consequence, TK1 top temperature rarely approaches the switching set point temperature, since a large amount of heat is continuously demanded by the ACH or by HE2. Consequently, the ratio between the heat delivered to TK1 and the one produced by the CPVT increases also causing a corresponding reduction of the heat produced by the TK2. In other words, an increase of such parameters moves part of the energy produced by the CPVT from TK2 to TK1, and therefore from DHW to space heating and cooling. However, this is accomplished using larger pumps, ACH and HE2, determining a higher capital cost of the system. Nevertheless, using larger HE2 and ACH capacities does not determine significant savings in terms of operating costs, since the additional amount of heat and cooling energy produced by such components is obtained losing a proportional amount of heat produced for DHW. The overall results is that the SPB increases for larger values of qP4. Finally, a thermo-economic optimization is also performed. Differently from the sensitivity analysis discussed above, the optimization consider the simultaneous variation of all the design parameters aiming at determining the minimum of a selected objective function. This procedure is implemented using complex mathematical algorithms. In particular, the optimization was carried out using the TRNOPT plug-in included in the TRNSYS package. This software links the TRNSYS dynamic simulation to the optimization algorithm. Such algorithms were developed by Lawrence Berkeley National Laboratory and included in the GENOPT package. In our case, the optimization was performed using a Generalized Search Method (G S), which allowed us to calculate the optimal value avoiding the calculation of partial derivatives. In particular, the modified version of the HookeeJeeves algorithm [53] was used. The latter allows for robust and efficient optimization, taking into account the approximation of the objective function due to the TRNSYS solving technique and avoids the risk of achieving local

58

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

about 1.0 GWh of electrical energy, 1.5 GWh of cooling energy and 1.2 GWh of thermal energy (space heating and domestic hot water). These energetic savings correspond to a saving of about 253 kV/ year in terms of operating costs. The system is designed in order to maximize the utilization of the energy produced. In fact, the dynamic plots showed that the capacity of the system, in terms of electricity, cool, heat and DHW, is very often lower than the demand of the Hospital. This criterion allows one to maximize system profitability. Results also showed that the performance of the system is dramatically affected by the weather conditions, being poor in winter and excellent in summer. This remarkable difference is basically due to the dramatic decrease of availability of beam radiation in winter. The dynamic simulations also proved the robustness of the control strategies implemented in this work. In fact, thermal energy produced by the solar loop, is mostly delivered to TK1 for space heating and cooling purposes; only a small amount of solar thermal heat is used for DHW by the tank TK2. This selection allows one to maximize the electrical and thermal efficiencies, calculated with respect to the beam radiation, oscillate around 20% and 50%. The overall capital cost of the system is about 3.6 MV, resulting in a SPB of about 14 years. The thermo-economic optimization showed that the economic performance of the system may be improved increasing solar field capacity and reducing tank volumes. In this case, SPB may reduce down to about 11 years. A substantial reduction of the SPB (about 4 years) is obtained in case of feed-in tariff. Therefore, authors concluded that the system under investigation is excellent from the energetic point of view since the majority of the energy produced is used by the user. This circumstance is

Fig. 16. Sensitivity analysis: TK1 volume.

minimum points. Such methodology is very efficient, since the optimization usually is completed in a relatively low number of simulations, compatible with reasonable computational times. In order to reduce this latter quantity, the number of variables considered in the optimization was relatively small. In particular, the optimization operated on three variables (CPVT area, TK1 specific volume, qP4), using the same ranges of variations used in the parametric analyses. The optimization target is the minimum of the SPB. The optimization process is shown in Fig. 18, where the values of the objective function (SPB) and the values of the design variables is shown for each iteration. The convergence is achieved after about 100 simulations, which means about 1 week of continuous calculation. This process shows the same results obtained by the sensitivity analysis. The minimum value of the SPB is achieved for the highest CPVT area, the lowest tank volume and the lowest value of qP4. As discussed before, in such configuration, the economic profitability of the system is maximized. 5. Conclusions In this paper the investigation of a solar trigeneration system is presented. The dynamic simulation of the system is performed by means of a zero-dimensional transient simulation model developed in TRNSYS. The SHC system under investigation is based on a field of CPVT solar collectors coupled with a single-stage LiBreH2O absorption chiller, supporting the cogeneration plant installed in a University Hospitals District located in Naples (Italy). Results of the simulation are excellent from the energetic point of view since the system under investigation allows one to save

Fig. 17. Sensitivity analysis; qP4.

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

Fig. 18. Optimization: objective function and design variables.

basically due to the peculiarity of the user (District Hospital) which demands constants and large amounts of thermal, cooling and electrical energy all year long. The economic performance is similar to the one achieved by other renewable technologies: in case of feed-in tariff SPB periods are very good, becoming excessively long when such incentives are not available. References [1] IEA, I.E. Agency, World Energy Outlook 2013, r.d.l.F. International Energy Agency, 75739 Paris Cedex 15, France, 2013. [2] C. Brandoni, M. Renzi, F. Caresana, F. Polonara, Simulation of hybrid renewable microgeneration systems for variable electricity prices, Appl. Therm. Eng. (2013) in press. [3] IEA, Solar Cooling Position Paper, Task 38 Solar Air-Conditioning and Refrigeration. Available at: http://archive.iea-shc.org/task38/, (accessed Jan 2014). [4] E. Parliament, EC. Directive 2009/28/EC of the European Parliament and the Council on the promotion of the use of energy from renewable sources. [5] R.K. Calay, W.C. Wang, A hybrid energy efficient building ventilation system, Appl. Therm. Eng. 57 (2013) 7e13. [6] D. Chemisana, J. López-Villada, A. Coronas, J.I. Rosell, C. Lodi, Building integration of concentrating systems for solar cooling applications, Appl. Therm. Eng. 50 (2013) 1472e1479. [7] Y. Guo, G. Zhang, J. Zhou, J. Wu, W. Shen, A techno-economic comparison of a direct expansion ground-source and a secondary loop ground-coupled heat pump system for cooling in a residential building, Appl. Therm. Eng. 35 (2012) 29e39. [8] A. Jamali, P. Ahmadi, M.N. Mohd Jaafar, Optimization of a novel carbon dioxide cogeneration system using artificial neural network and multi-objective genetic algorithm, Appl. Therm. Eng. 64 (2014) 293e306. [9] E. Entchev, L. Yang, M. Ghorab, E.J. Lee, Performance analysis of a hybrid renewable microgeneration system in load sharing applications, Appl. Therm. Eng. (2013) in press. [10] A. Coronas, S.S. Murthy, J. Carles Bruno, Editorial for the Special Issue of Applied Thermal Engineering on Polygeneration, vol. 50, 2013, pp. 1397e 1398. [11] C. Monné, S. Alonso, F. Palacín, L. Serra, Monitoring and simulation of an existing solar powered absorption cooling system in Zaragoza (Spain), Appl. Therm. Eng. 31 (2011) 28e35. [12] F. Calise, M.D. D’Accadia, L. Vanoli, Thermoeconomic optimization of solar heating and cooling systems, Energy Convers. Manag. 52 (2011) 1562e1573. [13] B.J. Huang, C.W. Yen, J.H. Wu, J.H. Liu, H.Y. Hsu, V.O. Petrenko, J.M. Chang, C.W. Lu, Optimal control and performance test of solar-assisted cooling system, Appl. Therm. Eng. 30 (2010) 2243e2252. [14] H.Z. Hassan, A.A. Mohamad, H.A. Al-Ansary, Development of a continuously operating solar-driven adsorption cooling system: thermodynamic analysis and parametric study, Appl. Therm. Eng. 48 (2012) 332e341. [15] F. Calise, Thermoeconomic analysis and optimization of high efficiency solar heating and cooling systems for different Italian school buildings and climates, Energy Build. 42 (2010) 992e1003. [16] F. Calise, M. Dentice d’Accadia, L. Vanoli, Design and dynamic simulation of a novel solar trigeneration system based on hybrid photovoltaic/thermal collectors (PVT), Energy Convers. Manag. 60 (2012) 214e225. [17] F. Calise, L. Vanoli, Parabolic trough photovoltaic/thermal collectors: design and simulation model, Energies 10 (2012) 22. [18] A. Buonomano, F. Calise, A. Palombo, Solar heating and cooling systems by CPVT and ET solar collectors: a novel transient simulation model, Appl. Energy 103 (2013) 588e606.

59

[19] F. Calise, A. Palombo, L. Vanoli, Design and dynamic simulation of a novel polygeneration system fed by vegetable oil and by solar energy, Energy Convers. Manag. 60 (2012) 204e213. [20] B. Prasartkaew, S. Kumar, Experimental study on the performance of a solarbiomass hybrid air-conditioning system, Renew. Energy 57 (2013) 7. [21] J. Wang, P. Zhao, X. Niu, Y. Dai, Parametric analysis of a new combined cooling, heating and power system with transcritical CO2 driven by solar energy, Appl. Energy 94 (2012) 58e64. [22] F.A. Al-Sulaiman, F. Hamdullahpur, I. Dincer, Performance assessment of a novel system using parabolic trough solar collectors for combined cooling, heating, and power production, Renew. Energy 48 (2012) 161e172. [23] P.G. Charalambous, G.G. Maidment, S.A. Kalogirou, K. Yiakoumetti, Photovoltaic thermal (PV/T) collectors: a review, Appl. Therm. Eng. 27 (2007) 275e286. [24] T.T. Chow, J.W. Hand, P.A. Strachan, Building-integrated photovoltaic and thermal applications in a subtropical hotel building, Appl. Therm. Eng. 23 (2003) 2035e2049. [25] I.K. Karathanassis, E. Papanicolaou, V. Belessiotis, G.C. Bergeles, Multi-objective design optimization of a micro heat sink for concentrating photovoltaic/ thermal (CPVT) systems using a genetic algorithm, Appl. Therm. Eng. 59 (2013) 733e744. [26] A. Buonomano, F. Calise, M. Dentice d’Accadia, L. Vanoli, A novel solar trigeneration system based on concentrating photovoltaic/thermal collectors. Part 1: design and simulation model, Energy 61 (2013) 59e71. [27] F. Cotana, F. Rossi, A. Nicolini, Evaluation and optimization of an innovative low-cost photovoltaic solar concentrator, Int. J. Photoenergy 2011 (2011). [28] K. Nishioka, T. Takamoto, T. Agui, M. Kaneiwa, Y. Uraoka, T. Fuyuki, Annual output estimation of concentrator photovoltaic systems using high-efficiency InGaP/InGaAs/Ge triple-junction solar cells based on experimental solar cell’s characteristics and field-test meteorological data, Sol. Energy Mater. Sol. Cells 90 (2006) 57e67. [29] A. Kribus, D. Kaftori, G. Mittelman, A. Hirshfeld, Y. Flitsanov, A. Dayan, A miniature concentrating photovoltaic and thermal system, Energy Convers. Manag. 47 (2006) 3582e3590. [30] G.K. Mittelman, Abraham Mouchtar, Ornit Dayan, Water desalination with concentrating photovoltaic/thermal (CPVT) systems, Sol. Energy 83 (2007) 1322e1334. [31] J.S. Coventry, Performance of a concentrating photovoltaic/thermal solar collector, Sol. Energy 78 (2005) 211e222. [32] G. Xu, X. Zhang, S. Deng, Experimental study on the operating characteristics of a novel low-concentrating solar photovoltaic/thermal integrated heat pump water heating system, Appl. Therm. Eng. 31 (2011) 3689e3695. [33] M. Li, X. Ji, G.L. Li, Z.M. Yang, S.X. Wei, L.L. Wang, Performance investigation and optimization of the trough concentrating photovoltaic/thermal system, Sol. Energy 85 (2011). [34] F. Calise, A. Palombo, L. Vanoli, Maximization of primary energy savings of solar heating and cooling systems by transient simulations and computer design of experiments, Appl. Energy 87 (2010) 524e540. [35] G. Mittelman, A. Kribus, A. Dayan, Solar cooling with concentrating photovoltaic/thermal (CPVT) systems, Energy Convers. Manag. 48 (2007) 2481e 2490. [36] G. Chicco, P. Mancarella, Distributed multi-generation: a comprehensive view, Renew. Sustain. Energy Rev. 13 (2009) 535e551. [37] G. Bizzarri, G.L. Morini, New technologies for an effective energy retrofit of hospitals, Appl. Therm. Eng. 26 (2006) 161e169. [38] G.K. Alexis, P. Liakos, A case study of a cogeneration system for a hospital in Greece. Economic and environmental impacts, Appl. Therm. Eng. 54 (2013) 488e496. [39] A. Herrera, J. Islas, A. Arriola, Pinch technology application in a hospital, Appl. Therm. Eng. 23 (2003) 127e139. [40] M.A. Lozano, J.C. Ramos, M. Carvalho, L.M. Serra, Structure optimization of energy supply systems in tertiary sector buildings, Energy Build. 41 (2009) 1063e1075. [41] G. Pagliarini, C. Corradi, S. Rainieri, Hospital CHCP system optimization assisted by TRNSYS building energy simulation tool, Appl. Therm. Eng. 44 (2012) 150e158. [42] E. Cardona, A. Piacentino, Optimal design of CHCP plants in the civil sector by thermoeconomics, Appl. Energy 84 (2007) 729e748. [43] A. Piacentino, F. Cardona, EABOT e energetic analysis as a basis for robust optimization of trigeneration systems by linear programming, Energy Convers. Manag. 49 (2008) 3006e3016. [44] D. Ziher, A. Poredos, Economics of a trigeneration system in a hospital, Appl. Therm. Eng. 26 (2006) 680e687. [45] P. Arcuri, G. Florio, P. Fragiacomo, A mixed integer programming model for optimal design of trigeneration in a hospital complex, Energy 32 (2007) 1430e1447. [46] G. Tsatsaronis, Thermoeconomic analysis and optimization of energy systems, Prog. Energy Combust. Sci. 19 (1993) 227e257. [47] N. Sendhil Kumar, K. Matty, E. Rita, W. Simon, A.l. Ortrun, C. Alex, W. Roland, G. Tim, M. Tapas Kumar, Experimental validation of a heat transfer model for concentrating photovoltaic system, Appl. Therm. Eng. 33e34 (2012) 175e182. [48] S.A. Klein, A Transient System Simulation Program, Solar Energy Laboratory, TRNSYS. University of Wisconsin, Madison, 2006. [49] F.P. Incropera, D.P. DeWitt, Fundamentals of Heat and Mass Transfer, fifth ed., John Wiley & Sons, Jefferson City, United States of America, 2001.

60

A. Buonomano et al. / Applied Thermal Engineering 67 (2014) 43e60

[50] L.R. Bernardo, B. Perers, H. Håkansson, B. Karlsson, Performance evaluation of low concentrating photovoltaic/thermal systems: a case study from Sweden, Sol. Energy 85 (2011) 11. [51] S. Furbo, S. Svendsen, Description and Operation of the Reference Heater, TIL, Technical University Denmark, January 1987.

[52] F. Calise, A. Palombo, L. Vanoli, A finite-volume model of a parabolic trough photovoltaic/thermal collector: energetic and exergetic analyses, Energy 46 (2012) 283e294. [53] R. Hooke, T.A. Jeeves, “Direct search” solution of numerical and statistical problems, J. Assoc. Comput. Mach. 8 (1961) 17.