Evaporative Cooling Options for Building Air-Conditioning: A Comprehensive Study for Climatic Conditions of Multan (Pakistan)

This study provides comprehensive details of evaporative cooling options for building air-conditioning (AC) in Multan (Pakistan). Standalone evaporative cooling and standalone vapor compression AC (VCAC) systems are commonly used in Pakistan. Therefore, seven AC system configurations comprising of direct evaporative cooling (DEC), indirect evaporative cooling (IEC), VCAC, and their possible combinations, are explored for the climatic conditions of Multan. The study aims to explore the optimum AC system configuration for the building AC from the viewpoints of cooling capacity, system performance, energy consumption, and CO2 emissions. A simulation model was designed in DesignBuilder and simulated using EnergyPlus in order to optimize the applicability of the proposed systems. The standalone VCAC and hybrid IEC-VCAC & IEC-DEC-VCAC system configurations could achieve the desired human thermal comfort. The standalone DEC resulted in a maximum COP of 4.5, whereas, it was 2.1 in case of the hybrid IEC-DEC-VCAC system. The hybrid IEC-DEC-VCAC system achieved maximum temperature gradient (21 ◦C) and relatively less CO2 emissions as compared to standalone VCAC. In addition, it provided maximum cooling capacity (184 kW for work input of 100 kW), which is 85% higher than the standalone DEC system. Furthermore, it achieved neutral to slightly cool human thermal comfort i.e., 0 to −1 predicted mean vote and 30% of predicted percentage dissatisfied. Thus, the study concludes the hybrid IEC-DEC-VCAC as an optimum configuration for building AC in Multan.


Introduction
Air-conditioning (AC) loads of buildings are increasing rapidly due to the increase in population, urbanization, and standards of living [1]. Although, many environment friendly AC systems are used nowadays that are free from HCFC/HFC refrigerants. However, there are many environmentally harmful systems that are still under practice to control/regulate temperature and humidity. Thus, increasing AC use is somehow responsible for global warming and CO 2 emissions around the globe [2]. Alternatively, heat-driven adsorption/ absorption-based environment friendly AC systems are also in practical use [3,4]. In 2017, worldwide global electricity consumption in the residential sector was around 496,583 and 398,729 ktoe in commercial and public sector buildings, respectively [5]. Heating, ventilation and air-conditioning (HVAC) of buildings consume a huge amount of total building electricity, which is also responsible for the increase in CO 2 emissions [6,7]. Therefore, environment friendly HVAC systems are principally needed worldwide. Evaporative cooling systems Buildings in Pakistan are generally built with brick wall with low to moderate insulation without any AC load calculation. The VCAC systems are usually preferred in commercial buildings, however, it is not financially feasible at least for low-income residents [21,49]. Therefore, multistage and/or hybrid types of evaporative coolers could be an eye-catching solution for providing desired conditions throughout year. Consequently, seven kinds of evaporative cooling configurations (DEC, IEC, VCAC, IEC-DEC, DEC-VCAC, IEC-VCAC, and IEC-DEC-VCAC) are explored in this study for climatic conditions of Multan (Pakistan). The effectiveness of evaporative cooling systems is determined from the lab-scale experimental apparatus, which are used as design values for DesignBuilder simulation. Building AC load of Agricultural Engineering Department of Bahauddin Zakariya University and associated human thermal comfort were estimated in EnergyPlus. Cooling capacity, coefficient of performance (COP) and reduction in CO 2 emissions by the systems were investigated and compared accordingly.
Energies 2020, 13, x FOR PEER REVIEW 3 of 23 without any AC load calculation. The VCAC systems are usually preferred in commercial buildings, however, it is not financially feasible at least for low-income residents [21,49]. Therefore, multistage and/or hybrid types of evaporative coolers could be an eye-catching solution for providing desired conditions throughout year. Consequently, seven kinds of evaporative cooling configurations (DEC, IEC, VCAC, IEC-DEC, DEC-VCAC, IEC-VCAC, and IEC-DEC-VCAC) are explored in this study for climatic conditions of Multan (Pakistan). The effectiveness of evaporative cooling systems is determined from the lab-scale experimental apparatus, which are used as design values for DesignBuilder simulation. Building AC load of Agricultural Engineering Department of Bahauddin Zakariya University and associated human thermal comfort were estimated in EnergyPlus. Cooling capacity, coefficient of performance (COP) and reduction in CO2 emissions by the systems were investigated and compared accordingly.

Proposed Evaporative Cooling Systems' Configuration
As the study area lies in a warm desert (hot and dry) climate (BWh) class of Köppen climate classification system [47], therefore, the EC systems are usually operated in the region except for monsoon (rainy) season. However, the potentials of multistage EC and hybrid systems are not explored as far as the local market of Pakistan is concerned. Therefore, this study investigates seven combinations of EC and VCAC systems, which are named as: (a) DEC, (b) IEC, (c) VCAC, (d) IEC-DEC, (e) DEC-VCAC, (f) IEC-VCAC, and (g) IEC-DEC-VCAC. Schematic illustrations and the corresponding psychrometric behavior are provided in Figure 2a to 2g, respectively. The key information for each proposed system is provided as follows:

Proposed Evaporative Cooling Systems' Configuration
As the study area lies in a warm desert (hot and dry) climate (BWh) class of Köppen climate classification system [47], therefore, the EC systems are usually operated in the region except for monsoon (rainy) season. However, the potentials of multistage EC and hybrid systems are not explored as far as the local market of Pakistan is concerned. Therefore, this study investigates seven combinations of EC and VCAC systems, which are named as: (a) DEC, (b) IEC, (c) VCAC, (d) IEC-DEC, (e) DEC-VCAC, (f) IEC-VCAC, and (g) IEC-DEC-VCAC. Schematic illustrations and the corresponding psychrometric behavior are provided in Figure 2a-g, respectively. The key information for each proposed system is provided as follows: • DEC system (Figure 2a): It cools process air isenthalpically (process 1-2). The DEC system can cool the process air up to wet-bulb temperature of the ambient air.

•
IEC system (Figure 2b): The working air (process 1-3) is cooled isenthalpically like DEC, and the cooling affect is transferred to product air by heat transfer between wet and dry channels. The fact sensibly cools the product air (process 1-2) up to wet-bulb temperature.

•
VCAC system (Figure 2c): It is the conventional compressor-based vapor compression air conditioning unit that cools the process air below the dew point. The required conditions are obtained by mixing/reheating the product air (process 1-2, dotted line). • DEC system ( Figure 2a): It cools process air isenthalpically (process 1-2). The DEC system can cool the process air up to wet-bulb temperature of the ambient air.

•
IEC system (Figure 2b): The working air (process 1-3) is cooled isenthalpically like DEC, and the cooling affect is transferred to product air by heat transfer between wet and dry channels. The fact sensibly cools the product air (process 1-2) up to wet-bulb temperature.

•
VCAC system (Figure 2c): It is the conventional compressor-based vapor compression air conditioning unit that cools the process air below the dew point. The required conditions are obtained by mixing/reheating the product air (process 1-2, dotted line). • IEC-DEC ( Figure 2d): It is a double stage EC unit that cools the air by indirect (process 1-2) followed by direct (process 2-3) EC concepts. ): It is also a hybrid system that is composed of IEC (process 1-2), IEC (process 2-3) and VCAC (process 3-4) units.

Building Energy Simulation
The present study focuses on air-conditioning requirements for the building of the Agricultural Engineering Department of Bahauddin Zakariya University, Multan, Pakistan. The primary layout of the building is shown in Figure 3, which occupies a floor of 3251.7 m 2 area and volume of 15,669.0

Building Layout
The present study focuses on air-conditioning requirements for the building of the Agricultural Engineering Department of Bahauddin Zakariya University, Multan, Pakistan. The primary layout of the building is shown in Figure 3, which occupies a floor of 3251.7 m 2 area and volume of 15,669.0 m 3 . The details of inputs parameters used for simulation are provided in Table 1 and consist of building geometrical information (height, wall/ roof construction material, glazing/window construction material, and overall heat transfer coefficient), thermal gains, energy efficiency setpoints, and zone-wise AC operation schedule.

of 21
Energies 2020, 13, x FOR PEER REVIEW 5 of 23 building geometrical information (height, wall/ roof construction material, glazing/window construction material, and overall heat transfer coefficient), thermal gains, energy efficiency setpoints, and zone-wise AC operation schedule.     [19,. Therefore, preliminary layout, thermal properties, construction materials, Energies 2020, 13, 3061 6 of 21 overall thermal coefficient, building symmetry, direction, and all other physical properties for the studied building were designed in DesignBuilder and simulated using EnergyPlus. The proposed seven combinations for AC units ( Figure 2) were designed in DesignBuilder using HVAC module in generic loop air handling unit (AHU). Hourly weather data of Multan (Pakistan) are used for the simulation, which was obtained from TRNSYS repository, provided by World Meteorological Organization (WMO) [77], cross checked using station weather data from Meteonorm. A complete fresh air circulation in AHU was selected. The modules of Direct Research Special and Indirect Research Special were selected in DesignBuilder for DEC and IEC analyses with design WBE of 0.90 and 0.605, respectively (Table 1). A direct expansion (DX) cooling coil was used in the case of the VCAC system. Thereby, combinations of single stage, multistage, and hybrid AC systems were successively arranged inside the AHU.
The corridors of the studied building are excluded from simulation as they are exposed to atmosphere. The washrooms area is also excluded from simulation as local design. The operation schedule for seven AC system configuration was selected from 8 am to 4 pm, which are typical office hours in the building.

Energy Estimations
Wet-bulb effectiveness (WBE) can be calculated by the ratio of temperature differences as given by [2]: where, ε wb is WBE (-) and T is temperature ( • C). The subscripts OA and SA represent outside air and supply air, respectively. The subscripts DBT and WBT represent dry bulb and wet-bulb temperatures, respectively. Cooling capacity (Q) was determined using Equation (2) as given by [78]: where, Q denotes the cooling capacity (kW), m denotes air mass flow rate of~14 kg s −1 , C p represents specific heat capacity of air (1.005 kJ kg −1 K −1 ), and ∆T is the difference between outside and supply air temperatures ( • C). Coefficient of performance (COP) was defined as the ratio of the cooling effect produced to the work input (prime movers, fans), represented by Equation (3): Equation (2) was also used for VCAC with an assumption of no latent heat transfer. However, considering latent heat transfer, Equations (4)- (7) can be used as given by [53]: where, ρ air represents density of air (kg m −3 ), u denotes the user-set volumetric flow of air (m 3 s −1 ), h out is total enthalpy of the air at supply side (J/kg), h in is total enthalpy of the air at inlet side (J/kg), X min is minimum of inlet and outlet air humidity ratios (kg/kg), Q net is the total cooling capacity (W), Q sensible is the total sensible energy transfer (W), and Q latent is the total latent energy transfer (W).

Cooling Loads
Cooling loads faced by the AC system were classified into transmission, infiltration, internal, and equipment loads. Transmission loads are mainly generated from heat transfer from walls, roof, basement, and windows. Infiltration loads include air change through the opening of doors, clearance beneath the doors, cracks, and crawl spaces. Internal loads include heat generated from the human body, lighting, appliances, and solar heat gain from windows through conduction. Heat generated from the human body depends on the level of activity of the body in terms of metabolic rate. Typically, a seated office worker generates 65 W m −2 [50] heat varied by body surface area as given by Equation (8) [79]: Thermal comfort of a human body mainly depends on operative temperature (approximately equal to the average of dry bulb temperature of air and surrounding surfaces' temperature), relative humidity and velocity of air. As per well-known heat and mass transfer guidelines by Cengal Y. [80], lower and upper acceptable limits of operative temperature for typical Pakistani shalwar-kameez clothing [50,51] are 18 • C and 23 • C (for winter: 1.145 clo) and 21 • C and 25 • C (for summer: 0.775 clo), respectively, at 50% relative humidity and air flow of 0.15 m s −1 [80]. The heat gains through human body (Q hb ), exterior surfaces (Q ext ), sol-air temperature (T s-a ), walls (Q wall ), fan motors (Q m ), floor (Q sp ), windows (Q glazing ), air change per hour (ACH), sensible infiltration (Q s,i ), and latent infiltration (Q l,i ) were calculated by Equations (9)-(18), respectively [80]. For details, please refer to the cited literature [78].
where, Q hb denotes total heat generated from human body (W); Q sens represents sensible heat gain from the skin and lungs; Q lat represents latent heat from the body (sweating, moist breath); h o represents heat transfer coefficient (W m −2 • C); A s represents surface area of the wall (m 2 ); T a stands for ambient air temperature ( • C); T s represents surface temperature ( • C); α solar represents solar absorptivity constant [81]; q solar represents incident solar radiation (W m −2 ); E represents surface emissivity [50]; ϕ represents solar incident angle relative to the latitude; T sur represents mean temperature of surrounding surfaces ( • C); T s-a denotes sol-air temperature ( • C); U wall represents wall's overall heat transfer coefficient (W m −1 • C −1 ); T in and T out represent inside and outside air temperatures, respectively; Q m represents heat gain from motor (W); P r represents power of the motor (W); L f represents load factor; U f represents utilization factor and η m denotes efficiency of the motor; Q sp represents total heat loss from the floor (W); U floor represents overall heat transfer coefficient for the floor (W m −1 • C −1 ); P floor Energies 2020, 13, 3061 8 of 21 represents floor perimeter (m); Q glazing represents total heat transfer through the windows; Q c , Q e , and Q f represent heat transfer through center, edge and frame of the windows, respectively; U g represents overall heat transfer coefficient of the window; A g denotes surface area of the window; ACH represents air change/hour; m in represents air infiltration inflow rate (m 3 h −1 ); V represents building volume (m 3 ); Q s,i and Q l,i represent sensible and latent heat transfer due to infiltration, respectively; h fg represents latent heat of evaporation (kJ kg −1 ); and X in & X out represent inside and outside air humidity ratios (g/g), respectively.

Thermal Comfort Indices
Thermal comfort can be quantified using predicted mean vote (PMV) and predicted percentage of dissatisfaction (PPD) as per Fanger's model and ISO standard 7730 [82]. Thermal comfort is a structure of various factors including temperature, relative humidity, velocity of air, rate of human body metabolism (directly linked to performance activity), and clothing insulation. It is important to note that PMV is only an average of thermal comfort sensation for a group of persons, so, individual thermal comfort sensation can vary from this average value. To counter this effect, PPD (as a function of ±2 PMV) is introduced. The PMV and PPD was calculated using Equations (19)-(22) [82] and Equation (23) [82]: where, Met represents metabolic rate of the body per unit area (W m −2 ), Q work represents heat loss from working human body (W m −2 ), P a vapor pressure of ambient air, T air represents temperature of the conditioned space ( • C), r cb represents ratio of clothed to naked body, T cloth represents temperature of the clothes ( • C), MRT represents mean radiant temperature ( • C), h conv represents convective heat transfer coefficient of the conditioned space (W m −2 • C −1 ), R c represents thermal resistance of clothes (m 2 • C W −1 ), u air represents velocity of air flow (m s −1 ), and Cloth insul represents insulation of clothes (clo). Figure 4 shows the numerical values of WBE achieved by DEC and IEC systems for climatic conditions of Multan (Pakistan) [83]. The effectiveness results presented in Figure 4 cover the typical summer months (i.e., May to August) under Multan conditions. These results are based on the lab-scale DEC/IEC experimental apparatus from the authors' previous work [83].  (Table 1). Consequently, a simulation model was designed in DesignBuilder and simulated using EnergyPlus to optimize the applicability of the proposed seven systems (Figure 2) for the studied building in Multan (Pakistan). The details of study area, AC load, and simulation are provided in Section 3. Accordingly, the results are presented in Figure 5a-b for a monthly basis, which shows the ambient and product air conditions for all of the seven AC configurations for temperature, relative humidity, humidity ratio, and enthalpy, respectively. numerical values of 0.90 and 0.605 are used as design values in the simulation models for DEC and IEC systems, respectively (Table 1). Consequently, a simulation model was designed in DesignBuilder and simulated using EnergyPlus to optimize the applicability of the proposed seven systems (Figure 2) for the studied building in Multan (Pakistan). The details of study area, AC load, and simulation are provided in Section 3. Accordingly, the results are presented in Figure 5a-b for a monthly basis, which shows the ambient and product air conditions for all of the seven AC configurations for temperature, relative humidity, humidity ratio, and enthalpy, respectively.

Results and Discussion
The IEC-DEC-VCAC system provided a temperature gradient of 13 °C (maximum) in May, whereas the IEC system was only able to achieve 4 °C (Figure 5a). It is important to mention that the temperature gradient is defined as the difference between ambient and product air temperatures. The studied hybrid system options (i.e., DEC-VCAC, IEC-VCAC and IEC-DEC-VCAC) performed relatively better as compared to all standalone system options. The standalone VCAC system resulted in a temperature gradient of 7 °C in May, which also produces almost uniform performance irrespective of ambient conditions round the clock throughout the month. Maximum temperature gradient for all configurations was found mostly in May and June months, whereas it was negative in winter months. The DEC system provided a relative humidity difference of 36% (maximum) in May (Figure 5b). The DEC system provided a humidity ratio of 0.000023 kg kg -1 in July (Figure 5c). The DEC system provided the highest enthalpy of 35.13 kJ kg -1 in June (Figure 5d).   In support to the above discussion, an insight of hourly performance variation is presented in Figure 6 for all configurations accordingly. The presented results are depicted for 19 June conditions as a hypothetical representative of a summer day. The findings of these hourly results are not much different, as reported above. Among the studied systems, the IEC-DEC-VCAC system resulted in a maximum temperature gradient i.e., 21 °C (at 3 pm) followed by IEC-VCAC i.e., 20.7 °C (at 3 pm), whereas the IEC system provided the lowest temperature gradient of just 9.8 °C (at 3 pm). From midnight to three past blazing noon, standalone EC systems resulted in a temperature gradient which depended on ambient air. It solidifies the statement that performance of standalone EC systems is dependent on ambient air conditions. In contrast, standalone VCAC and hybrid EC systems were only impacted by varying load. The hybrid EC systems easily meet the required conditions at higher power consumption. The IEC-DEC-VCAC system provided a temperature gradient of 13 • C (maximum) in May, whereas the IEC system was only able to achieve 4 • C (Figure 5a). It is important to mention that the temperature gradient is defined as the difference between ambient and product air temperatures. The studied hybrid system options (i.e., DEC-VCAC, IEC-VCAC and IEC-DEC-VCAC) performed relatively better as compared to all standalone system options. The standalone VCAC system resulted in a temperature gradient of 7 • C in May, which also produces almost uniform performance irrespective of ambient conditions round the clock throughout the month. Maximum temperature gradient for all configurations was found mostly in May and June months, whereas it was negative in winter months. The DEC system provided a relative humidity difference of 36% (maximum) in May (Figure 5b). The DEC system provided a humidity ratio of 0.000023 kg kg −1 in July (Figure 5c). The DEC system provided the highest enthalpy of 35.13 kJ kg −1 in June (Figure 5d).
In support to the above discussion, an insight of hourly performance variation is presented in Figure 6 for all configurations accordingly. The presented results are depicted for 19 June conditions as a hypothetical representative of a summer day. The findings of these hourly results are not much different, as reported above. Among the studied systems, the IEC-DEC-VCAC system resulted in a maximum temperature gradient i.e., 21 • C (at 3 pm) followed by IEC-VCAC i.e., 20.7 • C (at 3 pm), whereas the IEC system provided the lowest temperature gradient of just 9.8 • C (at 3 pm). From midnight to three past blazing noon, standalone EC systems resulted in a temperature gradient which depended on ambient air. It solidifies the statement that performance of standalone EC systems is dependent on ambient air conditions. In contrast, standalone VCAC and hybrid EC systems were only impacted by varying load. The hybrid EC systems easily meet the required conditions at higher power consumption.  One-day-averaged product air conditions of the seven AC systems are presented in Figure 7. It is worth mentioning that the expanded psychrometric chart with radiation and convection dimensions used in Figure 7 is taken in raw form from a state-of-the-art study authored by Teitelbaum et al. [84]. The IEC-DEC-VCAC resulted in a neutral to slightly cool thermal comfort sensation as per ISO 7730 [82] requirements (with matching mean radiant temperature MRT) for a person doing light office work, whereas the IEC-VCAC system resulted in neutral to slightly warm thermal comfort sensation due to varying MRT. The standalone VCAC resulted in a few shades of discomfort. The DEC showed maximum deviation from its mean radiant temperature. One-day-averaged product air conditions of the seven AC systems are presented in Figure 7. It is worth mentioning that the expanded psychrometric chart with radiation and convection dimensions used in Figure 7 is taken in raw form from a state-of-the-art study authored by Teitelbaum et al. [84]. The IEC-DEC-VCAC resulted in a neutral to slightly cool thermal comfort sensation as per ISO 7730 [82] requirements (with matching mean radiant temperature MRT) for a person doing light office work, whereas the IEC-VCAC system resulted in neutral to slightly warm thermal comfort sensation due to varying MRT. The standalone VCAC resulted in a few shades of discomfort. The DEC showed maximum deviation from its mean radiant temperature. From the PMV and PPD point of view, hourly annual thermal comfort sensation resulted from the seven AC configurations and is presented in Figure 8. All the proposed AC configurations achieved thermal comfort at some point or another throughout the whole year. However, on 19 June, all the hybrid EC systems (i.e., DEC-VCAC, IEC-VCAC and IEC-DEC-VCAC) achieved neutral to slightly cool thermal sensation with PPD variation of 30% to 40% (Figure 8). From the PMV and PPD point of view, hourly annual thermal comfort sensation resulted from the seven AC configurations and is presented in Figure 8. All the proposed AC configurations achieved thermal comfort at some point or another throughout the whole year. However, on 19 June, all the hybrid EC systems (i.e., DEC-VCAC, IEC-VCAC and IEC-DEC-VCAC) achieved neutral to slightly cool thermal sensation with PPD variation of 30% to 40% (Figure 8). From the PMV and PPD point of view, hourly annual thermal comfort sensation resulted from the seven AC configurations and is presented in Figure 8. All the proposed AC configurations achieved thermal comfort at some point or another throughout the whole year. However, on 19 June, all the hybrid EC systems (i.e., DEC-VCAC, IEC-VCAC and IEC-DEC-VCAC) achieved neutral to slightly cool thermal sensation with PPD variation of 30% to 40% (Figure 8). Performance of the seven AC systems based on the temperature gradient specifically in summer months (May, June, July, August) is presented in Figure 9. The IEC-DEC-VCAC system provided a temperature gradient of 10.9 • C (maximum) followed by 10.8 • C, 9.9 • C, 6.5 • C, 5.8 • C, 5.1 • C, and 3.2 • C (minimum) provided by IEC-VCAC, DEC-VCAC, IEC-DEC, VCAC, DEC, and IEC, respectively.
Performance of the seven AC systems based on the temperature gradient specifically in summer months (May, June, July, August) is presented in Figure 9. The IEC-DEC-VCAC system provided a temperature gradient of 10.9 °C (maximum) followed by 10.8 °C, 9.9 °C, 6.5 °C, 5.8 °C, 5.1 °C, and 3.2 °C (minimum) provided by IEC-VCAC, DEC-VCAC, IEC-DEC, VCAC, DEC, and IEC, respectively.
Performance of the seven AC configurations based on 1-day-averaged WBE for the climatic conditions of Multan (Pakistan) is presented in Figure 10. The IEC-DEC-VCAC system provided maximum WBE of 1.5 followed by IEC-VCAC i.e., 1.45, whereas, the standalone IEC system provided WBE of 0.5 (minimum). All the hybrid EC systems provided WBE of more than 1 due to pretreatment in EC stage and further compression at VCAC stage.  Performance of the seven AC configurations based on 1-day-averaged WBE for the climatic conditions of Multan (Pakistan) is presented in Figure 10. The IEC-DEC-VCAC system provided maximum WBE of 1.5 followed by IEC-VCAC i.e., 1.45, whereas, the standalone IEC system provided WBE of 0.5 (minimum). All the hybrid EC systems provided WBE of more than 1 due to pretreatment in EC stage and further compression at VCAC stage. Variation in cooling capacity, work input and COP of the seven AC systems for summer months is presented in Figure 11a-c, respectively. The IEC-DEC-VCAC system provided the highest cooling capacity of 184 kW, followed by IEC-VCAC with 183.5 kW, whereas, the standalone IEC system provided the lowest cooling capacity of 63 kW (Figure 11a). The standalone EC systems provided the lowest work input of 20 to 30 kW compared to the VCAC system with 108 kW. It is important to note that IEC-DEC-VCAC provided relatively lower work input among the VCAC-based systems (Figure Variation in cooling capacity, work input and COP of the seven AC systems for summer months is presented in Figure 11a-c, respectively. The IEC-DEC-VCAC system provided the highest cooling capacity of 184 kW, followed by IEC-VCAC with 183.5 kW, whereas, the standalone IEC system provided the lowest cooling capacity of 63 kW (Figure 11a). The standalone EC systems provided the lowest work input of 20 to 30 kW compared to the VCAC system with 108 kW. It is important to note that IEC-DEC-VCAC provided relatively lower work input among the VCAC-based systems (Figure 11b). The standalone DEC system provided the highest COP of 4.5 followed by the IEC-DEC system with 4.2 due to relatively lower work input. It is also worth mentioning that the IEC-DEC-VCAC system provided a relatively higher COP of 2 among the VCAC-based systems (Figure 11c). VCAC-based AC systems resulted in low COP compared to the DEC and IEC-DEC systems due to higher work input at direct expansion stage (cooling coil). In addition, the details of annual electricity consumption (kWh) of the studied AC system configurations are provided in Figure S1 (Supplementary Materials).  Additionally, annual sensible and latent heat gains from different sources (i.e., surface conduction, windows, people, infiltration, and equipment and lights) into the building are presented in Figure 13. Surface conduction contributed the highest heat gain of 79.1% of the total heat gains. Heat gain from surface conduction is highest due to poor insulation of walls and roof during building construction in Pakistan. Based on the above discussions, it was concluded that the EC hybrid systems, especially the IEC-DEC-VCAC system provides a superior performance compared to other AC systems. Although COP of IEC-DEC-VCAC is lesser compared to other AC systems due to higher  Figure 12 depicts the annual water consumption by the seven AC systems. The IEC-DEC system resulted in the highest water consumption of 1256 m 3 year −1 due to multistage EC systems followed by 1245, 1045, 1035, 704, and 697 m 3 year −1 water consumption by the IEC-DEC-VCAC, DEC, DEC-VCAC, IEC, and IEC-VCAC systems, respectively. The standalone VCAC system consumed no water for cooling due to usage of air-cooled coil. All the hybrid systems provided relatively lower water consumption compared to their respective standalone systems.  Additionally, annual sensible and latent heat gains from different sources (i.e., surface conduction, windows, people, infiltration, and equipment and lights) into the building are presented in Figure 13. Surface conduction contributed the highest heat gain of 79.1% of the total heat gains. Heat gain from surface conduction is highest due to poor insulation of walls and roof during building construction in Pakistan. Based on the above discussions, it was concluded that the EC hybrid systems, especially the IEC-DEC-VCAC system provides a superior performance compared to other AC systems. Although COP of IEC-DEC-VCAC is lesser compared to other AC systems due to higher Additionally, annual sensible and latent heat gains from different sources (i.e., surface conduction, windows, people, infiltration, and equipment and lights) into the building are presented in Figure 13. Surface conduction contributed the highest heat gain of 79.1% of the total heat gains. Heat gain from surface conduction is highest due to poor insulation of walls and roof during building construction in Pakistan. Based on the above discussions, it was concluded that the EC hybrid systems, especially the IEC-DEC-VCAC system provides a superior performance compared to other AC systems. Although COP of IEC-DEC-VCAC is lesser compared to other AC systems due to higher work input, this system provided the highest cooling capacity and WBE, and it can achieve the required conditions subject to a higher power consumption rate.
Energies 2020, 13, x FOR PEER REVIEW 16 of 23 work input, this system provided the highest cooling capacity and WBE, and it can achieve the required conditions subject to a higher power consumption rate. Annual greenhouse gas (CO2) emissions from the seven AC configurations is presented in Figure  14. CO2 emissions were obtained by multiplying the annual electricity consumption with the CO2 emission equivalent of 0.505 kgCO2 kWh -1 [85]. The DEC system emitted a minimum CO2 i.e., 332,8 kgCO2 year -1 due to the absence of a compressor followed by IEC, with 129,74 kgCO2 year -1 . Among the VCAC-based systems, standalone VCAC system emitted the highest CO2 of 274,883 kgCO2 year -1 . Moreover, the IEC-DEC-VCAC system emitted relatively the lowest CO2 of 241,134 kgCO2 year -1 among the VCAC hybrid systems.  Annual greenhouse gas (CO 2 ) emissions from the seven AC configurations is presented in Figure 14. CO 2 emissions were obtained by multiplying the annual electricity consumption with the CO 2 emission equivalent of 0.505 kgCO 2 kWh −1 [85]. The DEC system emitted a minimum CO 2 i.e., 332.8 kgCO 2 year −1 due to the absence of a compressor followed by IEC, with 129.74 kgCO 2 year −1 . Among the VCAC-based systems, standalone VCAC system emitted the highest CO 2 of 274,883 kgCO 2 year −1 . Moreover, the IEC-DEC-VCAC system emitted relatively the lowest CO 2 of 241,134 kgCO 2 year −1 among the VCAC hybrid systems.
Energies 2020, 13, x FOR PEER REVIEW 16 of 23 work input, this system provided the highest cooling capacity and WBE, and it can achieve the required conditions subject to a higher power consumption rate. Annual greenhouse gas (CO2) emissions from the seven AC configurations is presented in Figure  14. CO2 emissions were obtained by multiplying the annual electricity consumption with the CO2 emission equivalent of 0.505 kgCO2 kWh -1 [85]. The DEC system emitted a minimum CO2 i.e., 332,8 kgCO2 year -1 due to the absence of a compressor followed by IEC, with 129,74 kgCO2 year -1 . Among the VCAC-based systems, standalone VCAC system emitted the highest CO2 of 274,883 kgCO2 year -1 . Moreover, the IEC-DEC-VCAC system emitted relatively the lowest CO2 of 241,134 kgCO2 year -1 among the VCAC hybrid systems.

Conclusions
The present study aimed to explore the optimum AC system configuration comprised of direct evaporative cooling (DEC), indirect evaporative cooling (IEC), and vapor compression AC (VCAC) systems for the climatic conditions of Multan (Pakistan). The study area lies in a warm desert (hot and dry) climate (BWh) class of Köppen climate classification. In this regard, building AC load and human thermal comfort requirements for the Agricultural Engineering Department building of Bahauddin Zakariya University, Multan, Pakistan, have been investigated. Standalone evaporative cooling and standalone vapor compression AC (VCAC) systems are commonly used in Pakistan. Therefore, the study proposed seven AC system configurations named as: DEC, IEC, VCAC, IEC-DEC, DEC-VCAC, IEC-VCAC, and IEC-DEC-VCAC. The systems were designed in DesignBuilder and simulated using EnergyPlus.
The IEC-DEC-VCAC system provided a temperature gradient of 21 • C (maximum) subject to a higher power consumption, whereas, standalone EC systems failed to compete. All three hybrid systems achieved neutral to slightly cool PMV thermal comfort with 30 to 40% PPD. The IEC-DEC-VCAC system provided the highest WBE and cooling capacity of 1.4 and 184 kW, respectively. On the other hand, the IEC system provided the lowest WBE and cooling capacity of 0.50 and 63 kW, respectively. The DEC system yielded the highest COP of 4.5, whereas the IEC-DEC-VCAC system provided a low COP of~2.0 due to higher work input. The IEC-DEC-VCAC emitted an average annual CO 2 of 241,134 kg of CO 2 per year. Among the seven AC configurations, the standalone VCAC system resulted in the highest annual CO 2 emissions i.e., 274,883 kg of CO 2 per year. The standalone VCAC and hybrid IEC-VCAC & IEC-DEC-VCAC system configurations could achieve the desired human thermal comfort for the study area. However, the study concludes the hybrid IEC-DEC-VCAC as an optimum configuration (among the studied options) for building AC in Multan due to its better performance in terms of temperature gradient, thermal comfort, and CO 2 emissions.

Conflicts of Interest:
The authors declare no conflict of interest.

∆T
Difference between outside and supply air temperatures ( Inside air humidity ratio (g g −1 ) X min Minimum of inlet and outlet air humidity ratios (kg/kg) X out Outside air humidity ratio (g g −1 ) α solar Solar absorptivity constant ε wb Wet-bulb effectiveness (-) η m Efficiency of motor ρ air Density of air (kg m −3 ) ϕ Solar incident angle relative to the latitude