Design requirements for condensation-free operation of high-temperature cooling systems in mediterranean climate

Radiant cooling systems are a subject of increasing scientific interest due to their efficiency and ability to use high-temperature cooling sources. In hot and humid conditions, they have generally been studied in combination with dehumidification systems. For retrofit projects, a control system that would eliminate the need for dehu-midification would be beneficial. In the present study, a passive geothermal-based radiant high-temperature cooling system is studied in a Mediterranean climate. The system is operated with supply water temperature control using dew point temperature as a controlling variable. The system ’ s performance is compared with that of an all-air cooling system. The systems are evaluated using IDA-ICE building energy simulations, validated with on-site measurement data. The results show that the radiant cooling system produces the same level of thermal comfort with 40% lower energy use and 85% lower exergy consumption than the all-air system. The risk of condensation limits the cooling capacity of the radiant cooling system. Consequently, insufficient cooling capacity causes thermal discomfort for the occupants due to the operative temperature exceeding 26 ◦ C.


Introduction
As climate change is causing an increase in global temperatures, excessive temperatures that require cooling become increasingly common.Cooling demand is growing even in places where space cooling has traditionally been uncommon.Knutsson et al. [1] concluded that the contribution of anthropogenic climate change can be observed with highest confidence in the increasing occurrence of extreme temperature events such as heat waves.This will increase cooling demand globally, especially in buildings.For example, Isaac and van Vuuren [2] predicted that global electricity usage for air conditioning will increase by 7% annually between 2020 and 2030.In light of this development, the European Commission has identified the decarbonisation of buildings as a key goal in their 2016 strategy on heating and cooling [3].
Crawley [4] concluded that changes in building energy use will be the greatest in temperate and mid-latitude climates, where heating demand will decrease and cooling demand will increase.Similarly, Cellura et al. [5] predicted that cooling demand in the Mediterranean region would increase by 75%-104% between 2017 and 2035.This propagates the need for space cooling; Pierangioli et al. [6] found that in the Mediterranean climate cooling demand will come to dominate the thermal energy demand of buildings and that passive measures, such as improved insulation, will have to be supplemented with increased use of HVAC (heating, ventilation and air conditioning) systems.The long life cycle of buildings is (20-100 years [7]) creates an urgent need for retrofit solutions that are suitable for implementation in existing buildings that previously did not have cooling installations.
High-temperature cooling (HTC) systems are a subject of growing scientific and engineering interest because of their efficiency.HTC systems use a higher coolant temperature (17 • C or above) than conventional cooling systems (~7 • C).This means the temperature lift required from the chiller is reduced from approximately 30 • C to about 20 • C [8]. Seshadri et al. [8] demonstrated the potential of HTC systems in terms of smaller exergy losses and an improved chiller coefficient of performance (COP), compared with conventional high-temperature lift cooling systems.They found that decreasing the chiller's temperature lift from 29 • C to 21 • C resulted in a 29% improvement in sensible cooling efficiency.
Hydronic radiant cooling systems such as radiant cooling panels, which can be installed on the inner surfaces of existing structures, are well suited for retrofit projects.Their installation is non-invasive, and distribution pipes for the cooling water require much less space than the channels required for conventional convection-based systems.From an engineering perspective, such systems have commonly been prefabricated, which makes their installations in retrofitting projects more favourable and economical.Hao et al. [9] found that a radiant cooling ceiling combined with desiccant dehumidification can save 68.5% of chiller energy compared with a conventional all-air system.In terms of thermal comfort, Karmann et al. [10] found that in studies comparing radiant cooling systems with all-air systems either the radiant system was preferred or no clear preference one way or the other could be established.In other words, thermal comfort under radiant cooling systems is as good as or better than it is under all-air systems.An additional advantage of radiant cooling systems is their ability to utilise high-temperature energy sources such as solar and geothermal better than convection-based systems [11].This makes these systems better suited for ground-coupled geothermal passive cooling.In a ground-coupled geothermal passive cooling system, the heat extracted from the conditioned space is released directly into the ground without an active chiller.
The city of Barcelona (northeastern Spain) falls into the category of climates with an increased cooling demand.In summer, the outdoor air is hot and humid.The average temperature for July is 23.9 • C and the average relative humidity 67% [12].This poses a challenge for a radiant cooling system, since it has two conflicting targets.On the one hand, the cooling capacity of the system needs to be high enough to meet the cooling demand in the conditioned space.On the other hand, condensation on the cooling panel surfaces has to be prevented.Tang et al. [13] note that condensation risk is especially high in places where outdoor air can infiltrate the building directly.This means that buildings with high rates of air infiltration are especially challenging to serve because of condensation risk.
Some studies have investigated the solutions to the aforementioned problem of conflicting targets, but mostly on the cooling emission devices themselves.Rhee et al. [14] list a number of strategies to mitigate the condensation problem with radiant cooling systems.These strategies mainly rely on air dehumidification to ensure sufficient cooling capacity without causing condensation [15][16][17][18].However, the dehumidification of supply air will generally increase the system's cost, design effort required and operating complexity.For example, Wang et al. [18] developed a household size dehumidification system that would still require 40-48 • C heat source for desiccant regeneration.A solution that would achieve acceptable indoor air quality without the added complexity and cost of a dehumidification system would be favourable.Lim et al. [19] suggested controlling the supply water temperature according to the dew point temperature in the room that has the highest condensation risk.Their study focused on floor cooling, which leaves a knowledge gap on the applicability of this method to cooling with ceiling panels.Additionally, Rhee et al. [14] noted that further studies regarding the direct use of renewable energy as a cooling source are necessary to improve the system performance and to extend the application of HTC systems in buildings.
The general aim of the present study is to evaluate the system performance of a HTC system connected to a geothermal reservoir in the Mediterranean climate.The goal is to map the HTC system's cooling capacity considering condensation risk.The study also focuses on the evaluation of perceived thermal comfort in an office space served by the investigated system.The system's performance is also compared with the performance of an all-air cooling system.

Studied case
This study compares the performance of a geothermal HTC system with a conventional cooling system, namely an all-air system, in a retrofit case study.The study focuses on an indoor climate performance simulation of an office building that is a part of a school building complex located in Sant Cugat near Barcelona in Spain.Energy and exergy performances of the systems are also investigated.Simulations were conducted in IDA-ICE 4.8 (Indoor Climate and Energy) performance simulation program.The simulations are conducted following the standard EN15251:2007 [20].The accuracy of the IDA-ICE program was assessed through the IEA Solar Heating and Cooling Program, Task 22, Subtask C [21].IDA-ICE was found to produce accurate results, which showed good agreement with the reference programs.
The investigated administrative building is a part of a three-building complex together with a primary school and sports pavilion.The building complex was constructed in 1975.The building complex together with the layout of the studied building is shown in Fig. 1.The studied building is marked with a red circle in Fig. 1.The indoor air temperature as well as relative humidity and CO 2 -level in the building are measured using a Netatmo climate station [22] that is located in the north-eastern corner of the building, indicated with a black arrow in Fig. 1.The relevant parameters of the building are presented in Table 1.The values for building area and geometry, number of occupants and U-values of the building components were collected from the building owner.No measured data were available for the internal heat gains caused by lighting and office equipment, domestic hot water (DHW) use and air infiltration rate, so the values for these parameters were adopted from open literature [23][24][25][26] considering the studied building type and location.IDA-ICE's typical value of 0.5 W/m 2 floor area was used for the internal heat gain due to DHW.The occupation schedule is estimated according to the recommendation by Ref. [24], for an office building.The schedule is presented in Fig. 2. As shown in Fig. 2, the building is occupied during office hours (7-18) with varying intensity.During the month of August, the building is unoccupied due to school holidays.This is reflected in simulations to accurately model the use of the building.
Climate data inventory ASHRAE IWEC2 weather file for Barcelona Airport [27] was used for the simulations.The annual temperature distribution for the studied location is presented in Fig. 3. Fig. 3 shows the frequency of hourly temperatures in the climate data inventory.The hours with temperatures exceeding 24 • C are represented with a white column; 24 • C is the base temperature for calculating cooling degree days given by Eurostat [28].The total number of cooling degree days in the simulation data set is 164, less than the average of 2009-2018 years in Spain (246) but almost double the average of the EU-28 (88) over the same period [28].The design dry-bulb temperature for cooling is 29.6 • C [27].

Current HVAC-system
As shown in Table 1, the building has a total served area of 288 m 2 .The building complex underwent an envelope retrofitting process in 2018, which has increased external wall insulation.The U-value of the external walls was reduced to 0.27 W/m 2 K after the retrofit.Windows are equipped with manually operated shutters.The building is occupied during office hours (07-18), and the number of occupants is eight.DHW demand is estimated to be 4 l/person/day based on the demand of a typical office building according to Takata [25].The internal heat loads from lighting and office equipment are estimated to be 1.6 and W/m 2 floor area and are based on data provided by Gayral [23] and Ahmed et al. [24] respectively.
Currently the studied building is naturally ventilated without any cooling services.The ventilation air is supplied via infiltration through the envelope and openings, such as window openings.The amount of air infiltration through the building envelope was estimated to be 2.7 m 3 /h/ m 2 external wall with a pressure difference of 4 Pa, based on the results reported by Litvak et al. [26] in similar buildings in the Mediterranean climate.

Studied cooling system
As part of a joint European project (Geofit), the central heating network of the building complex presented in Fig. 1 will be retrofitted with a ground-source heat pump system (GSHP).The ground heat exchanger will be sized to cover the base heating load of the entire network, 41 kW.Therefore, the ground heat exchanger consisting of boreholes will have high installed thermal energy storage capacity.However, due to local regulations in Spain, cooling can only be installed in the administrative building.Therefore, the thermal energy storage capacity of the borehole system will be oversized for the modest cooling demand of the administrative building.Accordingly, the administrative building was chosen as a suitable demonstration site for an innovative cooling solution.Its relatively low cooling demand combined with an HTC system presents an opportunity to test geothermal passive cooling, where the ground heat exchanger is used as a heat sink without the heat pump.The building also has a stable occupation schedule, which is beneficial for radiant cooling systems that are generally ill-equipped to cope with fast, unexpected load changes during a day.The system is operated with a supply water temperature control method as outlined by Fig. 1.Building complex and layout of the studied building in Sant Cugat near Barcelona.Lim et al. [19].
The cooling panels are designed to cover the maximum ceiling area with five given panel sizes that vary from 950 mm × 277 mm-2550 mm × 277 mm.Selections of cooling panel technical parameters are based on the manufacturer's real performance data of the most updated cooling panel products available on the market, presented in Fig. 5 [29].
The cooling capacity of a hydronic system with a constant surface area is dependent on two variables that can be manipulated: mass flow and water supply temperature.Lim et al. [19] found that the most favourable performance was achieved by controlling the supply water temperature.Similar results were found by Arghand et al. [30], who tested a comparative passive cooling setup in a climate chamber with low cooling loads.In the present study the base supply water temperature for the cooling panels is set to 18 • C. The temperature was determined based on a thermal response test conducted on the site, which suggested this temperature would be achievable by passive cooling.With this supply temperature, the panels' cooling capacity is 40 W/m 2 panel area based on manufacturer data.To reduce the condensation risk, the panels are equipped with a recirculation shunt.The panel supply water temperature is maintained at least 1 • C above the dew point temperature in the room with the highest condensation risk, following a suggestion by Hao et al. [9].During unoccupied time the system is shut down and the indoor temperature is allowed to drift.The HTC system's indoor temperature set point is 25 • C with a 1 • C deadband, according to the request of the occupants.For a ceiling panel radiant cooling system, the temperature set point is one of the most sensitive variables in terms of both thermal comfort and energy efficiency [31].Therefore, choosing a different temperature set point would yield different results.
According to Atienza Marquez et al. [32], a radiant cooling system should always be coupled with a ventilation system to take care of ventilation requirements and latent heat loads.Therefore, to complement the passive geothermal HTC system, a ventilation system retrofit is also designed.The air flow supplied by the air handling unit (AHU) is simulated following the standards of ASHRAE 62.1-2016 [33].The specifications for single office rooms in this standard correspond to supply air flow of 0.4 l/s/m 2 floor area .The ventilation system is designed as demand-controlled ventilation.The system is equipped with heat recovery whose temperature efficiency was estimated to be 80% [34].Like the HTC system, the ventilation system is only operated during occupation to conserve energy.

All-air cooling system
To assess the performance of the studied HTC system, it is compared with an all-air cooling system.Here, the all-air system serves as a comparable alternative cooling system.This system is especially prevalent in the Mediterranean area.In an all-air system, chilled air is supplied through the ventilation system to provide the desired indoor climate.For the studied case, the all-air system is dimensioned in accordance with the Category II of the EN15251:2007 standard [20], to ensure adequate cooling capacity.For single offices the standard recommends a supply air flow of 4.2 l/s/m 2 floor area .The system control aims to keep the return air temperature at 24 • C during occupied time.To achieve this, the supply air temperature is varied between 15 • C and ambient outdoor air temperature.Supply water temperature to the AHU's cooling coil is set to 5 • C. The system is equipped with heat recovery that is equivalent to the HTC system.For unoccupied time the ventilation system is run on half speed, which allows the room temperatures to fluctuate in order to reduce energy use.The schematic designs of the baseline, all-air and HTC systems are presented in Fig. 6 (a), (b) and (c) respectively.

Model validation
The created building model was validated with indoor and outdoor temperature data measured onsite.Both sets of data were collected during June and July of 2019.The building model was simulated using measured outdoor temperature data from a weather station in the building site instead of the weather file used in performance evaluation simulations.The simulated temperature inside the studied zone in the north-eastern corner of the building ('Zone 1') was then compared with measured indoor temperature data from the same room.The considered zone is marked with a black arrow in Fig. 1.The indoor temperature was measured with 15-min intervals using a Netatmo climate station [22].The validation results are presented in Fig. 5. Fig. 7 shows the daily averages of measured and simulated air temperatures inside the studied zone.The median absolute difference (MAD = median(|T sim -T m |), where T m = median(T m )) between the hourly measured and simulated temperatures was 2.1%, while the largest absolute difference between the methods was 1.4 • C. The measurements were conducted during the building's normal use, which inevitably causes some deviations due to unpredicted variation in both internal and external loads, such as unexpected window openings and changes in cloud coverage.Overall, the low MAD and small absolute difference between measured and simulated temperatures indicate that the building model is valid.

Comparing the proposed system with an all-air system
Fig. 8 presents the operative temperature in Zone 1 during average cooling load.The figure shows, that during an average cooling load the HTC system and all-air system perform equally well.With the all-air system, the highest operative temperature during this week is 26.3 • C, compared to 26.4 • C for the HTC system.Both systems exceed comfort temperature for a duration of 7 h (13% of occupied time).Meanwhile in the baseline scenario (without any cooling services) the operative temperature stays continuously above the comfort limit of 26 • C, peaking at 29.1 • C. The system exceeds the comfort temperature for a duration of 55 h (100% of occupied time).This means that during average load, the HTC system and the all-air system perform equally well and clearly outperform the baseline scenario.
Fig. 9 shows the duration curves of operative temperature in the hottest zone for each scenario during the cooling season.Only occupied time is considered for the duration curves.In total the cooling season includes 671 h of occupied time.The hottest zones in the different scenarios are the zone in the north-western corner for baseline and allair systems and Zone 1 for the HTC system.Fig. 9 shows that the all- air system and the HTC system perform equally well in terms of keeping the operative temperature on a comfortable level.With the all-air system the operative temperature in the zone with highest cooling demand exceeds the comfort limit of 26 • C (represented in the figure with a black dotted line) for roughly 186 h during the cooling season, or about 28% of the total occupied time.With the HTC the respective numbers are 156 h corresponding to 23% of total occupied time.Both systems clearly outperform the baseline system, which exceeds the comfort limit for 560 h or 83% of the total occupied time.Similarly, Fig. 10 shows the duration curves for CO 2 -level in the zone with the highest cooling demand for each studied scenario, which are the second room from the west on the north side for the baseline scenario and the second room from west on the south side for both the all-air and HTC systems.Fig. 10shows that both all-air and radiant systems keep the CO 2 -level under the recommended limit of 1100 ppm during the entire cooling period.In the case of the baseline system the limit is exceeded for 139 h during the cooling season, or about 21% of the total occupied time.
In addition to the contributions to indoor climate, we also investigated the thermal energy and exergy performance of the studied systems.Only the thermal energy use of the systems is considered, the energy use of auxiliary components such as pumps and fans is outside the scope of this study.The results are presented in Table 2, which shows the cooling energy use during the considered cooling season and peak cooling power for each studied scenario.As can be seen, the peak cooling power need is similar for all-air and radiant (HTC) systems.However, the HTC system uses approximately 40% less thermal energy for cooling than the all-air system.Furthermore, the HTC system's cooling capacity is achieved with a much higher cooling temperature, 18 • C as opposed to 5 • C with the all-air system.This is beneficial for the planned use of passive cooling, as it eliminates the need for a chiller and will therefore be the more favourable solution from an environmental point of view.The higher cooling temperature of the HTC system is also beneficial for the system's exergy performance.
Fig. 11 presents the exergy performance of the HTC system and allair system.The exergy flow in the studied systems' cooling circuits Ėx cooling was calculated using Equation (1) [35]: where Qcooling is the (cooling) heat flow, T s is the supply water temperature to the cooling circuit, T r is the return water temperature from the cooling circuit and T 0 is the outdoor temperature.Studying a similar system, Sangi and Müller [35] showed that the majority of exergy destruction occurs in the cooling circuit.Therefore, this comparison using Equation ( 1) -which ignores the exergy destruction in other components of the systems, such as pumps, AHU and control units-was deemed justifiable.Fig. 11 illustrates the difference between the HTC and all-air systems' exergy performances.The low temperature difference between the ambient and the cooling supply water leads to relatively low exergy destruction in the HTC system compared with the much higher exergy destruction in the all-air system.This is due to the larger temperature difference between the ambient and the cooling supply water.The accumulated exergy destruction during the cooling season is approximately 90 Wh/m 2 floor area for the HTC system and 570 Wh/m 2 floor area for the all-air system.This shows that the all-air system consumes 6.5 times more exergy than the evaluated radiant HTC system.This implies that the HTC system has much wider opportunities to be combined with lowgrade renewable energy sources than the conventional all-air system.

Thermal performance of the studied HTC system
Fig. 12 and Fig. 13 present the simulation results of the studied Zone 1 during an average and the hottest working weeks during cooling season (July 1 5 and July 22 26 respectively).This zone was chosen as representative because it was found to present the highest risk for condensation in the simulations.In both figures, the working hours are shown with a white background, while unoccupied hours are depicted with a grey background.During unoccupied time, the cooling system is switched off and the temperature is allowed to drift.As Fig. 12 shows, during an average load the HTC system is capable of handling the cooling load.The operative temperature is mainly kept below the comfort limit of 26 • C, as the dew point temperature of the supply air mostly remains below 20 • C. The highest and most enduring overshoot is 0.4 • C and lasts for 5 h.The maximum predicted mean vote (PMV, calculated according to ISO 7730 [36]) that results from this indoor temperature overshoot is 0.64 (≈neutral).Furthermore, Fig. 13 shows that during the peak cooling load the cooling capacity of the HTC system is limited by supply air dew point temperature.Operative temperature in the hottest zone exceeds the comfort limit of 26 • C by up to 1.9 • C every afternoon, because the recirculation shunt forces the supply water temperature up to mitigate condensation risk.The PMV-value caused by  this indoor temperature overshoot is 1.0 (=slightly warm).The total occupied overshoot time during the week is 52 h 15 min, and the highest operative temperature is 27.9 • C.
To study how the system could be further improved to keep operative temperature below the comfort limit of 26 • C, we examined the hottest day during the studied period: July.Fig. 14 (a) presents the operative temperature inside the studied zone, supply water temperature and outdoor temperature using the designed condensation control, and shows that the system's cooling capacity is insufficient to keep the operative temperature under the desired comfort limit of 26 • C.This is due to the dew point temperature control limiting the supply water temperature to 22 • C at the lowest.The operative temperature peaks at 27.9 • C. The maximum cooling power of the system during peak load is

Discussion
The results of this study show that the studied hydronic radiant HTC system is capable of meeting the cooling demand of the evaluated building for 77% of the cooling season.Furthermore, the provided thermal comfort level is comparable with a conventional all-air cooling system, which is a common cooling system in Mediterranean countries.The HTC system performs significantly better than the all-air system in terms of energy and exergy efficiencies, due to the smaller temperature difference between the coolant and the ambient temperature.These results are congruent with results by Oxizidis and Papadopoulos [37] who conducted their study in similar climate in Greece.
To further improve the performance of the HTC system during peak load, a dehumidification system would be necessary to ensure condensation-free operation at a higher cooling capacity.However, the fact that the evaluated building in its current state is naturally ventilated would likely necessitate significantly more retrofit measures to improve the airtightness of the envelope.Currently, unwanted air infiltration is a significant source of moisture addition to the indoor air.Therefore, ensuring that the HTC system's supply water temperature will not drop below the dew point temperature in Zone 1 is likely to increase the retrofit cost considerably.Considering the modest amount of time that the system's cooling capacity is inadequate, this cost increase could be deemed unreasonable.
Experienced thermal comfort is subject to a number of factors including air temperature, mean radiant temperature, relative humidity, cold draught, metabolic rate, and clothing insulation.In the present study, operative temperature and CO 2 -level were given special consideration as indicators of the quality of the indoor comfort and environment.Review studies such as Karmann et al. [10] and Rhee et al. [14] Fig. 13.Operative temperature, supply air dew point temperature and supply water temperature under peak load in the studied zone.have shown that HTC systems are generally found to provide a comfortable indoor environment, but also that other factors affecting thermal comfort, such as cold draught and metabolic rate, should be considered when designing an HTC system.
In the conducted simulations the surface temperature of the radiant panels is assumed to be uniform.In reality, Ning et al. [38] found that the cooling panel's surface temperature directly below the cooling water pipe can be 4 • C lower than the surface temperature on the rest of the panel.The effect of the surface temperature distribution on the panels' cooling capacity and possible condensation hotspots should be carefully assessed.This study focused on the evaluation of the thermal energy performance and thermal comfort of the studied scenarios.Life cycle cost assessments and primary energy uses of the compared scenarios remain a subject for further study.

Conclusions
This study examined a hydronic radiant high-temperature cooling (HTC) system supplied by passive geothermal in the Mediterranean climate.Based on the results of the study, the following conclusions are drawn: • The evaluated HTC system is able to maintain the operative temperature below 26 • C for 77% of the cooling season in Barcelona without surface condensation.However, during cooling peak load (23.8 W/m 2 floor area ) the cooling capacity of the system is insufficient because of condensation risk, which leads to a maximum operative temperature of 27.9 • C. The predicted mean vote (PMV) value in this situation is 1.0, which is classified as a slightly warm feeling.
• To remove thermal discomfort the supply water temperature would need to be lowered to 17.1 • C, which would increase the risk of surface condensation on the radiant panels.The largest difference between the supply water temperature and dew point temperature is approximately 5 • C, creating a high risk of condensation on the panel surface.• Thermal comfort provided by the studied HTC system is as good as with an all-air system.However, the HTC system achieves this thermal comfort level using 40% less cooling energy than the all-air system.• The HTC system's ability to operate at a supply temperature higher than 18 • C facilitates the use of passive cooling and enables the evaluated HTC system to operate with 85% lower exergy consumption than the all-air system, which potentially makes the system more environmentally friendly.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 4 (
a) shows variation of the daily mean temperatures during the cooling period, that is, June-September in Barcelona.The two evaluated periods are indicated with grey shading.Period 1 (July 1-5) represents the average cooling load situation and Period 2 (July 22-26) represents Nomenclature COP Coefficient of performance DHW Domestic hot water Ėx cooling Exergy flow in cooling circuit [W] GSHP Ground source heat pump HTC High-temperature cooling HVAC Heating, ventilation and air conditioning IDA-ICE Indoor climate and energy simulation software IWEC2 International weather files for energy calculations 2et al.the cooling peak load.The average outdoor temperature during the entire cooling season is 22.2 • C.During Period 1 it is 20.9 • C and during Period 2 it is 26.0 • C. According to Fig.4(a), the highest outdoor temperatures and therefore the maximum cooling load occur during late July and early August, whereas in early June the cooling demand is low.In addition, the cooling performance of the evaluated systems was also assessed on the hottest day (July 24) in the studied period.The temperature variation during this day is presented in Fig.4(b).The average temperature during this day is 26.7 • C and the peak hourly temperature is 31.4• C.

Fig. 3 .
Fig. 3. Hourly mean temperature distribution in the climate file for Barcelona.

Fig. 4 .
Fig. 4. (a) Variation of daily mean temperature and outdoor humidity during the cooling season in Barcelona with indicated (grey area) periods used for system evaluation in this study (b) Variation of hourly temperatures during the hottest day in Barcelona.

Fig. 6 .Fig. 7 .
Fig.6.Schematic system designs of (a) building in its current state (b) all-air cooling (c) passive radiant cooling system with mechanical ventilation.

Fig. 9 .
Fig. 9. Duration of thermal discomfort with the compared systems.

Fig. 11 .
Fig. 11.Daily mean of the exergy flow per square meter cooled floor area in the compared systems.

Fig. 12 .
Fig. 12. Operative temperature, supply air dew point temperature and supply water temperature under average load in the studied zone.

Fig. 14 .
Fig. 14.Operating conditions during the hottest day in the dataset (a) with condensation control and (b) without condensation control.

Table 1
Parameters of the studied building.