Effect of ventilation patterns on indoor thermal comfort and air-conditioning cooling and heating load using simulation

Over the past three years, regulations have been implemented to combine natural ventilation (NV) and air conditioning to mitigate the risk of disease transmission, particularly in response to the COVID-19 outbreak. As we know, simultaneous use of NV and air conditioning can make it challenging to achieve indoor thermal comfort. This paper aims to analyze the effect of NV on the air conditioning`s cooling and heating load in a classroom through simulation. A simulation model was developed using EnergyPlus software with an OpenStudio interface software. Simulation results demonstrate that continuous use of NV alongside an air conditioner increases the cooling load from 1.06 to 1.75 times during summer and a 1.54 to 9.49 times heating load increase during winter. On the other hand, intermittent NV every hour results in a cooling load increase from 1.05 to 1.46 times in summer and a heating load increase from 1.13 to 4.63 times in winter. Moreover, employing NV based on the outside air temperature can reduce the cooling load at the air conditioner with set-point 26℃—28℃ from 0.94 to 0.88 times. The outcomes of this study are expected to serve as a reference for determining strategies that effectively combine NV and air conditioning to meet various needs without causing a significant increase in energy consumption. Additionally, the results are expected to be useful for reducing AC energy consumption in extremely hot and cold weather with some strategies of NV application.


Introduction
Global warming has been an important issue for many parties to focus on reducing high energy consumption and costs as much as possible with many strategies.Heating, ventilation, and air-conditioning (HVAC) systems contribute of about 40-60% of the total energy consumption of the building sector (Zhang, et al. 2021).One of the passive design strategies to decrease HVAC energy use is using as much natural ventilation (NV) as possible when the outdoor temperature is not extremely high or low.Besides the passive design strategy, ventilation is essential for good indoor air quality maintenance for human health.Enhancement of NV was considered a key measure for a safe and healthy indoor space, and NV was prioritized over mechanical ventilation, and special measures should be taken with mechanical ventilation (filters, higher rates, etc.) (Monge-Barrio 2022).The study also shows that ventilated windows can be used to significantly reduce the energy demand for cooling and/or heating and improve the indoor comfort performance for occupants depending on the season (Imessad et al. 2014;Liu et al. 2017).However, using only NV in extreme hot and cold weather without air conditioners is known to be impractical, as achieving indoor thermal comfort would be difficult.Based on previous studies, some NV techniques can improve indoor thermal comforts, such as Trombe walls, double skin façade, solar chimneys, solar walls, wind towers, wind catchers, and wing walls, which can adjust the indoor temperature, drop a couple of degree Celsius is summer, and increase a couple of degree Celsius in winter, and also reduce cooling and heating energy (Zhang, et al. 2021).
Besides energy consumption and indoor thermal comfort issues, HVAC and NV have become important components to discuss as the COVID-19 pandemic began to spread at the beginning of 2020 worldwide, which has brought important issues for providing the best environmental conditions with good indoor air quality in the rooms to minimize virus transmission.Studies showed that the coronavirus disease can be transmitted via airborne (Morawska, et al. 2020;Guo et al. 2020;Li, et al. 2021;Jiang, et al. 2021).REHVA (Federation of European Heating, Ventilation, and Air Conditioning Associations) and ASHRAE (American Society of Heating, Ventilation, and Air Conditioning Engineers), recommended preventive measures for reducing the airborne disease transmission risk (ASHRAE 2020;REHVA 2020;Fantozzi et al. 2022).The reduction of infection probability was evaluated in different environments as a function of the increase in ventilation rate ensured by HVAC systems (Fantozzi et al. 2022).In naturally ventilated buildings, the use of CO 2 concentration can be used as an indicator of the infection probability.Although this method may present some limitations, to ascertain whether ventilation is sufficient and to ensure proper ventilation, based on the Ministry of Education, Sports, Science, and Technology, Japan (MEXT) it is possible to measure CO 2 concentration level as a guideline for ventilation with the standard for school environment hygiene is 1500 ppm, and Japan government's new coronavirus infectious disease countermeasures subcommittee says that 1000 ppm or less is desirable for restaurants that assume eating and drinking without masks (METI 2020;MEXT 2020).Some studies also recommended keeping indoor CO 2 concentration below 1000 ppm to reduce airborne transmission (Di, et al. 2021;MHLW 2022;Du et al. 2019).The NV appliance is considered to be sufficient to reduce the CO 2 concentration level.However, the air conditioner is favorable to use to achieve indoor thermal comfort in the hot and cold seasons.The combination of air conditioners and NV has been introduced in the pandemic era to reduce airborne disease transmission risk while keeping room in the thermal comfort range.However, there is evidence to suggest that there is excessive energy loss through uncontrolled or unnecessary air infiltration (Orme 2001).Study shows that air conditioner energy use in junior high schools increased 2 times in summer and 1.9 times in winter during the pandemic compared to air conditioner energy use before the pandemic (Sekartaji et al. 2023).It occurred because to reach indoor thermal comfort in a room with an air conditioner on and also with NV, the air conditioner temperature during the pandemic was set lower in summer and higher in the winter than before the pandemic (Sekartaji et al. 2023).
By the above issues related to the air conditioner energy use and NV, this research aims to assess NV's impact on air conditioner cooling and heating load using simulation when NV is used together with an air conditioner.Difference ventilation patterns will be simulated in this research to assess the impact of each ventilation pattern on the cooling and heating load in the room.This research result is not limited to the COVID-19 issue, but it applies to other airborne diseases, so it can be a reference for NV operation patterns when the air conditioner is on in the future.NV is a passive design strategy to decrease air conditioning energy use.However, in the extremely hot and cold season, air conditioners must still be operated to achieve thermal comfort in the room.Indoor thermal comfort will be difficult to be achieved if NV is used simultaneously in a full air conditioning operating room.For the last three years, the regulation of combining NV and air conditioning has been set to prevent disease transmission risk due to the COVID-19 outbreak in 2020.
There are two major objectives in comparing each mixed-used ventilation pattern to fully air-conditioned without NV intervention.First is when NV is necessary for reducing airborne virus transmission risk.Second is when NV is used as a strategy for energysaving issues, which has to be addressed, and reducing airborne virus transmission risk is not the issue.This study aims to assess the impact of NV on indoor thermal comfort and the cooling and heating load of air conditioning systems using simulation.Simulation allows users and designers to understand the interrelation between design and parameters, and the result is more energy conscious with a better comfort level attained throughout (Clarke 2001).A simulation model was developed using EnergyPlus software with an OpenStudio interface to analyze the impact of NV on the cooling and heating load of a classroom.The model was calibrated using measured data from an existing building located in Oita City.
The results of this research are also hoped to be useful for reducing AC energy consumption in extremely hot and cold weather with some strategies of NV application in full AC-operated rooms.Table 1 shows the abbreviation list.

Methodology
This research has two parts of the methodology.First is the field measurement of CO 2 level and calculation of the air change per hour in the elementary classroom during the COVID-19 pandemic when NV is used together with an air conditioner.In this part, the CO 2 concentration level will be analyzed and air change per hour in several measurement times.The result of the air change per hour will become the reference for the second part of the methodology.Figure 3 shows the classroom floor plan and the measurement item position.
The second part is the simulation of the cooling and heating load of the different ventilation patterns in the junior high school classrooms using the building energy simulation program EnergyPlus with energy graphical interface software, OpenStudio, and 3D Google Sketchup for modeling.EnergyPlus, a building energy modeling (BEM) software, is used to evaluate the energy consumption of buildings, which is relatively applicable with an excellent correlation between actual data compared to other energy simulation software, and have various weather data set (Tran et al. 2021;Bui et al. 2020;Fumo 2014).The OpenStudio application is a graphic energy modeling tool connected to 3D Google Sketchup for modeling, and EnergyPlus is an energy simulation tool.Schedules, loads construction and materials, HVAC setting, and selection can be edited in OpenStudio, and it has a high level of results visualization (Jarić et al. 2016).
The research framework of the simulation can be seen in Fig. 1.The actual monitoring data will be compared to the simulation data for validation of the simulation.In many studies, energy monitoring data and BEM coordination were demonstrated to be complementary (Tran et al. 2021).Actual monitoring data are obtained from Private Finance Initiative (PFI) monitoring data collected by the Oita Municipal office,  which collected air conditioner energy use data in several junior high schools in Oita City, Japan, from 2019 to 2021.The school selected for this validation is Joto Junior High School which uses Electric Heat Pump for the air conditioner system.
In the validity of simulation data, AC energy use from PFI data monitoring is compared with AC energy use of simulation results with daily average AC operating times and air room temperature in each classroom.The AC energy-use is electric power data used for AC (kWh).On the other hand, the output of the simulation result for new ventilation patterns are cooling load (kWh) and heating load (kWh).New parameters for simulation are divided into 4 independent variables, which are shown in Fig. 2. Dependent variables in this simulation are the cooling load for the summer season and the heating load for the winter season.The final results evaluate the impact of different parameters, such as ventilation patterns, design strategies, AC set-point, and air change per hour (for ventilation patterns 2 to 6), on the AC cooling and heating load.
The ventilation patterns (VP) parameter approach is divided into 3 parts, first is fully naturally ventilated without an air conditioner (VP1), the second is the air-conditioned room without NV(VP2), and the last is mixed mode ventilation, which means a combination between an air conditioner and NV (VP3 to VP6).While VP3, VP5, and VP6 have NV continuous operation, VP4 has an intermittent operation, which is closed and open per hour.All the cases from VP2 to VP6 have air conditioners in continuous operation from 08:00 to 16:00.NV at other times is set to be closed.Considering the Oita weather data is not available in EnergyPlus weather data, the weather data used for this simulation is Kagoshima weather data, which is considered to have the closest climate data to Oita City.The summer simulation period is from June to September 2019, while the winter simulation period is from January to March 2019 and December 2019.

CO 2 concentration level measuring data result
The classroom plan and CO 2 measurement item position are shown in Fig. 3.The balcony is the open area on the south side, while the corridor is a closed area on the Fig. 2 Independent and dependent variable scheme north side, which is the main access to the classroom.First, using the measured CO 2 concentration, the ventilation volume is estimated, and the ventilation situation is grasped and evaluated.The Formula used to estimate the ventilation volume is derived from Formula (1) (Ueno et al. 2016).Equation ( 1) is the indoor pollutant concentration t hours after the start of pollutant generation when the indoor pollutant concentration is completely uniform diffusion.Measurement tool was determined in point ① because it won't get in the way of student learning activities.
where, C: Indoor contamination concentration (kg/m 3 ) C 0 : Indoor concentration before pollution (kg/m 3 ) Assuming that the concentration in the room before the occurrence of contamination is equal to the concentration in the outside air, the concentration of contamination C in the room increases with time, becomes a constant value when t = ∞, and is expressed by the following equation (Ueno et al. 2016).
(1) Solving this for the ventilation volume yields the following equation, which is used to estimate the ventilation volume.Formula (1) and ( 2), indoor contamination concentrations (kg/m 3 ), and the number of pollutants (kg/h) have been shown to be mass units.However, formulas (3) and ( 4), indoor contamination concentrations (m 3 /m 3 ), and number of pollutants (m 3 /h) show the volume or capacity unit.The change of the unit is because the measurement tool used in this experiment uses ppm (m 3 /m 3 ) as the CO 2 unit.
where, m: Number of pollutants generated per person (m 3 /h/person) n: Number of people in the classroom (people) C: Indoor contamination concentration (m 3 /m 3 ) C 0 : Indoor concentration before pollution (m 3 /m 3 ) However, before the pollution occurred, the indoor concentration (= outside air CO 2 concentration) was 410 ppm, and the number of people in the classroom was 36.The period used for calculation is shown in Table 2 and Fig. 4. The period used in the In addition, regarding the amount m of pollutants, the source of carbon dioxide in the classroom is the people in the classroom.Therefore, the amount of carbon dioxide exhaled is estimated by referring to Formula 4.
Table 2 shows the calculation results.However, E is the ventilation frequency.During the measurement period, the ventilation volume on September 8, when the carbon dioxide concentration was the highest, was 561.3 m 3 /h, and the ventilation rate per hour was 3.12 times/h, higher than the residential standard of 0.5 times/h.However, the ventilation volume per person is 15.6 m 3 /h/person.This value is lower than the generally required ventilation volume of 20-30 m 3 /h/person, and it cannot be said that the ventilation situation is good.On the other hand, the ventilation volume on September 9, when the ventilation volume was the highest, was 1454 m 3 /h, and the ventilation volume per person was 40.4 m 3 /h/person, indicating good ventilation condition.On the other two days, the ventilation volume was 20 m 3 /h/person or more, and there was no problem.As mentioned above, it is presumed that the wind inflow was obstructed on September 8, and it was confirmed that good ventilation conditions could not be maintained on such a day.

PFI data monitoring and validation
Figure 5 shows the comparison between PFI data monitoring and validation from the summer of 2019 to 2021.Meanwhile, Fig. 6 shows the comparison between PFI data monitoring and validation in winter 2018-2019 to 2020-2021.It can be seen that in the summer of 2020 and winter from December 2020, there is AC energy use increase.Besides the longer AC operation times, it is caused by the lower AC setting temperature in summer and higher AC setting temperature in winter.Based on the author`s previous study, it occurred due to window opening regulation recommended by Japan government regarding airborne virus transmission risk prevention during the COVID-19 pandemic, which required temperature control to achieve indoor thermal comfort in classrooms (Sekartaji et al. 2023).The lower AC setting temperature in summer and higher AC setting temperature in winter control room air temperature, which is highly affected by hot outside air temperature in summer and cold outside air temperature in winter.
The naturally ventilated period is difficult to be simulated since the ventilation rate, air velocity, and air change per hour data are absent.For the validity of the AC energy-use data, NV is set in OpenStudio with an average of 6 air changes per hour from June 2020 in summer and from March 2020 in winter (during the COVID-19 pandemic).Adjustment of air change per hour is also carried out to match with actual monitoring data.

Modeling
The simulation modeling is divided into two models, one is for single corridors, and the other is for double corridors.Simulation modeling A (Fig. 7) has two thermal zones, thermal zone 1 is the classroom, and thermal zone 2 is the north corridor.Simulation Modeling B (Fig. 8) has three thermal zones, thermal zone 1 is the classroom, thermal zone 2 is the north corridor, and thermal zone 3 is the south corridor.The PFI data monitoring for validating is the data from the school, which has similar modeling to modeling A, which only has 1 corridor on the north side.
One of the parameters or independent variables in this simulation is design strategies.Table 3 shows the design strategy list and the description of detailed materials of each design strategy.There are four design strategies (DS), DSa, DSb, DSc, and DSd.DSa and DSb have one corridor, while DSc and DSd have double corridors.The difference between DSa and DSb or DSc and DSd is the wall insulation and double glazing for outdoor windows.Table 4 shows the characteristics of each zone.
The HVAC system in this simulation used the air loop package rooftop heat pump (Fig. 9), with Coil Cooling DX Single Speed 1 Rated coefficient of performance (COP) 4, Coil Heating DX Single Speed 1 COP 4, and fan with 0.9 efficiencies.

Ventilation pattern (VP) 1
Ventilation pattern (VP) 1 is an entirely natural ventilated room simulation.In this simulation, air change per hour is 6 ach.Summer design day is considered a summer peak day of August 28th.While winter design day is considered a winter peak day, which is December 21st.For the analysis of indoor thermal comfort, the air temperature comfort range in the school classroom is 18℃ to 28℃ based on the Ministry of Education, Culture, Sport, Science and Technology Japan (MEXT).Table 5 shows the Case names of VP1.
Figure 10 shows the result of the summer design day simulation.It shows that Case 1b has the highest air temperature during the day.While Case 1c has the lowest air temperature during the day.It indicates that a double corridor has a great impact on reducing air temperature on a hot summer day.However, conversely, wall insulation and pair glazing windows make air temperature higher than a room without wall insulation and pair glazing windows.Presumably, trapped hot air in the gap of the glazing and the insulation itself during the night caused higher temperatures during the day.The indoor air temperature comfort range will be analyzed.The air temperature range in summer is divided into three ranges, below 24 °C, between 24 to 28 °C, and above 28 °C.The indoor air temperature comfort range was obtained from acceptable Predicted Mean Vote (PMV) by International Organization for Standardization (ISO) 7730:2005 as ranging for existing buildings between -0.7 and + 0.7 (ISO, Ergonomics of the thermal environment-Analytical determination and interpretation of thermal comfort using calculation of the PMV and PPD indices and local thermal comfort criteria 2005).Calculation data assumed to determine range standard with acceptable PMV by ISO are shown in Table 6.

Table 4 Thermal zone characteristics
PMV calculation result of an air temperature of 24.5 °C for a lower limit is -0.68, and 28 °C for an upper limit is + 0.75 (Tanaka et al. 2006;Fukuyo and "PMV and PPD Calculation", 2021).Since, in OpenStudio software, the range of 24.5 °C is difficult to be calculated, 24 °C is considered as a lower limit of indoor air temperature comfort range in this study.
Based on the result of summer indoor air temperature range percentage (Fig. 11), DSc and Dsd have the highest percentage of air temperature comfort range, while DSa has the smallest percentage of air temperature comfort range.Although DSc and DSd have the same percentage of air temperature comfort range, the air temperature above 2 °C in DSc is higher than in DSd.It indicates the DSd has a great design for the summer season to increase the air temperature comfort range and decrease the hot temperature in the room.It is caused by the corridor in the South, which blocks the solar radiation from the South in the summer season.Although Fig. 11 shows that the number times of indoor air temperature between 24 to 28 °C is higher than indoor air temperature above 28 °C, it should be underlined that the total number of times is within 24 h, not only during the day or when the classroom is occupied but unoccupied times also calculated.It becomes the limitation of this study since the simulation could not calculate zone temperature only when the room is occupied.
Figure 12 shows the air temperature result on the winter design day.It shows that Case 1b has the highest air temperature, followed by Case 1d, Case 1a, and Case 1c.In contrast to the summer simulation result, the winter double corridor has a bad impact on receiving heat to keep the room warm.Solar radiation from the South is applicable to receive heat optimally in the winter.Thus, the corridor on the South blocked the solar radiation from the South.Based on the winter indoor air temperature comfort range (Fig. 13), DSb has the highest percentage of air temperature comfort range, while DSc has the smallest percentage of air temperature comfort range.

Ventilation pattern (VP) 2
Ventilation pattern (VP) 2 is the simulation of a fully air-conditioned room without NV.There are 20 cases in summer and 20 cases in winter based on the design strategies and AC set-point.
Figure 14 shows the cooling load in each case.It can be seen that wall insulation and pair glazing windows just have a small effect on decreasing the cooling load in summer.This small effect also only be seen in AC set-point 24 °C and 25 °C, while in AC set-point 26 °C to 28 °C.On the contrary, wall insulation and pair glazing windows have been affected by increasing cooling load in the summer and effective thermal insulation plays an important role in reducing cooling and heating load so that the wall insulation or pair glazing type selection could be adjusted and studied more.
Based on the VP2 simulation result in the winter (Fig. 15), DSb in AC set-point 20℃ has the smallest heating load.Unlike summer, wall insulation and pair glazing window significantly impact decreasing heating load in winter.The rising heating load spike can be seen in DSc when the corridor on the South side blocks a significant amount of solar radiation in winter, which is wanted by occupants to reduce the heating energy load.As is known, indoor thermal may be impacted by solar radiation transmitted through windows.A south-facing room in the winter season, a southern-facing window introduces sunlight to form a sun patch inside the room (Sun et al. 2011), and the impact of the solar radiation mainly concentrates on the near-window zone, and its scope will vary with the latitude of the location (Song et al. 2022).

Mixed mode ventilation patterns
Figure 16 shows a cooling load comparison between mixed-mode ventilation patterns and air-conditioned rooms without NV, VP2.It can be validated that the higher the air change per hour, the higher the cooling load increase, and the higher the AC set-point, the lower the cooling load increase.However, in VP5, where NV is used when the outside air temperature is between 18 °C to 28 °C, the cooling load is decreased in AC setpoint 27 °C and 28 °C.Based on these comparisons, the design strategies also influence the cooling load increase.DSa has the smallest cooling load increase, followed by DSb, DSc, and DSd, with the highest cooling load increase.
Figure 17 shows a winter heating load comparison between mixed mode, and VP2.Compared to the summer season, mixed mode and VP2 in the winter, the heating load is much higher than the cooling load increase.VP3 (Fig. 17a), VP4 (Fig. 17b), and VP6 (Fig. 17d) show that DSb, has the highest heating load increase compared to VP2, while DSc has the lowest heating load increase compared to VP2, followed by DSa in VP3 and VP6.On the other hand, in the use of intermittent NV, VP4, DSd has a lower heating load increase than other design strategies.Figure 16d shows that the heating load in DSb Fig. 16 Cooling load comparison between mixed mode and VP2 in summer has an extreme increase, from 5.33 to 9.46 times in 8 air changes per hour.It means that pair glass and wall insulation which greatly impacts the energy saving in DSb has become impractical when NV is used together with an air-conditioner.In the VP5 case, when NV is used based on the outside air temperature range (18 °C to 28 °C), the heating load is not changed from VP2 because in the winter, outside air temperature is commonly under 18℃, which causes NV is not used in this case.

Regression analysis
Table 7 shows the numerical value of each independent variable.These numerical values are input for model prediction.Value 1 is considered the lowest value of the cooling and heating load results, and value 5 is considered to have the highest cooling and heating load.

Summer regression analysis
Tables 8 and 9 show the analysis of the variance of cooling load in summer.The predicted Formula for regression analysis of summer simulation is shown in Formula (5).Table 10 shows the correlation of each independent variable in summer.ASP has the highest correlation value, while VP has the smallest correlation value.It indicated the AC set-point has a strong influence on energy efficiency.However, the difference in the ventilation pattern of NV has no significant correlation with energy efficiency.
Predicted formula for regression analysis of summer simulation: (5) 51.15532 + 6.058303 VP + 41.63068 DS + 186.157ASP + 65.92199 ach  Figure 18 shows the simulation and prediction value correlation of cooling load in summer.The simple regression equation was y = 0.9556x + 39.41 and R 2 = 0.9556.Since the R 2 value is 0.9556, it can be said that the simulation cooling load result is highly correlated to the prediction value of the cooling load.

Winter regression analysis
Tables 11 and 12 show the analysis of the winter heating load variance.The predicted Formula for regression analysis of summer simulation is shown in Formula (6).Table 13 shows the correlation of each independent variable in winter.VP has the highest correlation value, while DS has the smallest one.It indicated that different ventilation patterns co influence NV energy loss and energy efficiency for.However, the different design strategy has no significant correlation for energy loss or energy efficiency for using NV.
Predicted Formula for regression analysis of winter simulation: Figure 19 shows the simulation and prediction value correlation of cooling load in summer.The simple regression equation was y = 0.7932x + 864.71 and R 2 = 0.7898.Since the R 2 value is 0.7898, it can be said that the simulation heating load result is correlated to the prediction value of the cooling load.

Discussion
Based on this result, the study proposes energy efficiency solutions by comparing the impact of different parameters on AC cooling and heating load.In the present simulation, we analyzed three different ventilation patterns used in the classroom as a target     simulation.First, a fully natural ventilated classroom without an air conditioner is simulated to assess the indoor thermal comfort as a dependent variable with 90% occupancy in the room.It shows that indoor thermal comfort could not be reached.Second, the fully air-conditioned classroom without NV intervention and the mixed-used ventilated classroom have been simulated to assess the cooling and heating load.Design strategies, air change per hour setting, AC set-point, and ventilation patterns using a time-based window opening schedule and outside air temperature are the basic parameters used in this simulation.We hypothesize that these parameters strongly correlate to the cooling and heating load.However, the simulation results show that some parameters do not correlate strongly to the cooling and heating load.Design strategies, one of the parameters for cooling and heating load, correlate more in the summer season than in the winter.Two different layout plans as a design strategies approach, reveal different results for energy-saving strategies.Double corridor type, in which one of the corridors is placed on the south side, gives a benefit in summer season in terms of energy-saving due to blocking solar radiation from the south side.Besides the indoor thermal comfort benefit in the summer, the corridor or space on the south side could be architecturally functional as an addition study area or meeting area, not necessarily functioning as a corridor, while the corridor on the north side could be normally used as the corridor.On the other hand, this type of corridor has declined indoor thermal comfort in the winter, so it needs more energy to reach indoor thermal comfort in the winter.Moreover, in a fully air-conditioned room without NV intervention, the wall insulation and pair glazing design strategies in the summer slightly deteriorate indoor thermal comfort when the AC set-point is higher than 27 °C so that the cooling load little higher than in the classroom without the wall insulation and pair glazing in this AC set-point.Meanwhile, in an AC set-point under 27 °C, the cooling load of a room with wall insulation and pair glazing is slightly lower than the room without it.In mixed mode ventilation patterns, in all AC set-point cases, wall insulation and pair glazing slightly deteriorate indoor thermal comfort in the classroom.This case is different in heating load result in the winter season, which shows that this wall insulation and pair glazing strongly influence the energysaving strategy due to the lower heating load caused by this design strategy.As already stated in the introduction, there are two major objectives in the comparison between mixed-mode ventilation and fully air-conditioned cooling and heating load.VP3, VP4, and VP6 are mixed-mode ventilation patterns for reducing airborne disease transmission risk, while VP5 is for energy-saving related issues.The cooling and heating load results can be interpreted depending on the issue to addressed.Following the hypothesis, the cooling and heating load in VP6 is the highest of all other mixed mode ventilation patterns.VP6 is the mixed-mode ventilation pattern in which NV is used throughout the day regardless of the outside air temperature.The increase of heating load in VP6, compared to VP2, is extremely high, near to 10 times VP2 in the highest AC set-point (24 °C), 8 ach, and DSb design strategy.While the increase of cooling load in VP6 compared to VP2 is not too much different from the increase of cooling load in VP3 compared to VP2.VP3 and VP5 are the mixed mode ventilation patterns based on outside air temperature.Based on the present simulation results, VP5 is the lowest increase in cooling load and no change in heating load compared to VP2 from all other ventilation patterns.It could be a good energy-saving strategy in the summer season, in AC set-point above 26 °C.However, this ventilation pattern could not be considered a good strategy for reducing airborne disease transmission risk in the peak of summer and in winter.The peak of summer could be underlined since, in June, the outside air temperature tends to be not higher than 28 °C, so this VP5 ventilation pattern could be applied for energy-saving and health-related issues.Based on the simulation result of VP4, this ventilation pattern is the best strategy to reduce cooling load to reduce airborne disease transmission risk issues.Based on the author's previous study found that CO 2 concentration would exceed 1000 ppm after 30 min when windows are closed without NV in an occupied classroom with a discussion-type class (Sekartaji et al. 2022).While in VP4, the window opening pattern is per hour, it considers attaining in reducing airborne diseases transmission risk with the higher air change per hour.Based on the CO 2 concentration level measurement result (Table 2), air change per hour 8 is the most appropriate for reducing CO 2 concentrations to 656.6 ppm.Air change per hour under 3 is considered undesirable due to the CO 2 concentrations level being near 1000 ppm.

Conclusion
In summary, this study can be concluded as follows: 1. Indoor thermal comfort in the naturally ventilated occupied classroom without an air conditioner during the day (08:00-16:00) could not be reached regardless of design strategies.2. AC set-point has the highest correlation value for the cooling load in summer, while the ventilation pattern has the highest correlation value for the heating load in winter.3.In an effort to reduce airborne disease transmission risk issues, cooling and heating load in ventilation pattern based on window opening schedule, one hour close, and one hour open, VP4, is lower than the ventilation pattern based on outside air temperature, VP3, and ventilation pattern regardless outside air temperature, VP6. 4. Ventilation pattern window opening when the outside air temperature is between 18 to 28 °C (VP5) slightly can reduce the cooling load in the summer when the AC setpoint is above 26 °C for an energy-saving strategy.5.No change is found in heating load between a fully air-conditioned classroom without NV intervention (VP2) and a classroom with a ventilation pattern window opening when the outside air temperature is between range 18 to 28 °C (VP5) in winter.6.A design strategy with double corridors on the north and south sides can reduce the cooling load in an air-conditioned classroom without NV intervention and mixedmode ventilation in summer.On the other hand, this design strategy deteriorates indoor thermal comfort inwinter regardless of the ventilation pattern.

Fig. 3
Fig. 3 Classroom floor plan and CO 2 measurement item position (①) Fig. 5 Data monitoring and simulation comparison in summer

Fig. 6
Fig. 6 Data monitoring and simulation comparison in winter

Table 1
Abbreviation list

Table 2
Ventilation volume

Table 3
Design Strategy list and materials

Table 5
Case name of VP1

Table 6
PMV calculation parameters and the assumption values

Table 7
Independent variables and the value

Table 8
Analysis of variance 1 of cooling load in summer

Table 9
Analysis of variance 2 of cooling load in summer

Table 10
Correlation of each parameter for cooling load in summer

Table 11
Analysis of variance 1 of heating load in winter

Table 12
Analysis of variance 2 of heating load in winter

Table 13
Correlation of each parameter for heating load in winter