Analyzing the effects of passive design strategies on building ventilation performance and thermal comfort using simulation-based approach

. One of the sustainable approaches to improving indoor thermal comfort and building ventilation efficiency, particularly in developing and populated countries, is to implement the optimum passive design solutions. However, the interaction between ventilation performance, indoor thermal comfort, and passive design features of buildings is scarcely studied in the domain of airflow modeling. Therefore, this study aims at assessing the ventilation performance and indoor thermal comfort with regards to passive design strategies, specifically, opening characteristics such as the Opening to Wall ratio (OWR) and Window to Wall ratio (WWR) in a Computational Fluid Dynamics (CFD) environment. The CFD simulations run is found pivotal for determining the spatial variation of Predicted Mean Vote (PMV) and Predicted Percentage Dissatisfied (PPD) to gauge thermal comfort and the natural ventilation performance in terms of air change rate per hour (ACH). The proposed methodology is implemented in classroom spaces of an academic building located in a warm and humid region. In the study, a field investigation is carried out to collect precise temperature and air velocity data needed to validate the resulting CFD model. The findings show that increasing OWR and WWR percentages can enhance ventilation efficiency and indoor thermal comfort.


Introduction
Humans spend 80% to 90% of their life inside buildings, and it is very important to have an indoor built environment comfortable for living [1]. However, in recent years, indoor environmental quality (IEQ), which includes thermal, acoustic, visual, and indoor air quality (IAQ) of urban areas has been declining, due to the factors such as inadequate building ventilation, poor ambient air quality, excessive presence of indoor dust particles, and the usage of coal, wood, and other solid fuels for cooking [2]. This often demands the building to turn towards mechanical ventilation techniques at the expense of spending more energy on the building's operation and management. For a developing and populous nation such as India, striking a balance between improving indoor thermal comfort and building energy usage, therefore, remains a challenge.
In India, buildings account for around 45% of the nation's total energy consumption, and to offer indoor thermal and visual comfort to building occupants, lighting and ventilation controls account for over 73% of the energy utilised in Indian buildings [3]. Therefore, innovative, affordable, and sustainable technology must be implemented to construct buildings that are more functional and comfortable. Therefore, implementing and predicting the effectiveness of natural ventilation in buildings in relation to the outdoor environment is a critical parameter to be considered. Moreover, it needs * Corresponding author: albert@iitb.ac.in to necessitate the evaluation of existing techniques for analysing the ventilation performance of buildings that can deliver an acceptable ventilation rate to dilute indoor air contaminants and a thermally comfortable environment for building occupants.
Furthermore, ASHRAE guidelines state that indoor thermal comfort and ventilation rate for acceptable IAQ is also linked to the design characteristics of the building. This is the reason there has been an increase in interest among academics and professionals in assessing occupant health and indoor thermal comfort in relation to building design strategies. [4]. The essential element of architectural or passive design strategies is the building's access to natural lighting and convective airwaves. Building orientation and shape, opening-towall ratio (OWR), window-to-wall ratio (WWR), insulation of the roof and walls, glazing units and their characteristics, overhang specifications, installation of solar-shading devices, wall thickness, colour and texture of the walls, and construction materials are just a few examples of passive design elements [5]. The variation in pressure across the inlet and outlet openings (windows and doors), however, serves as a measure of natural air movement in building spaces. Therefore, OWR and WWR are among the important passive design elements that need to be looked at while analysing ventilation efficiency and indoor thermal comfort in buildings. To examine the effectiveness of ventilation in buildings, empirical models, small-scale and large-scale experimental models, multi-zone network models, and Computational Fluid Dynamics (CFD) models are some of the techniques that are widely accepted [6,7,8]. CFD has become a crucial technique in recent years due to its ability to effectively solve and capture transport and turbulence equations, which is a challenging, complex, and multi-variable problem to solve for airflow models in a ventilation system [9]. In most cases, CFD is taken into consideration while designing new buildings and evaluating the performance of existing buildings in aspects such as ventilation, pollutant transport, and fire safety.
The direct application of CFD for design and parametric studies of wind-driven natural ventilation of buildings has been covered in several studies. In order to predict the airflow rate in buildings with different design factors, such as shape, wind direction, and various opening configurations, Asfour and Gadi (2007) conducted CFD simulation runs [10]. Similarly, using CFD simulations, Rizk et al. (2018) demonstrated how the estimated indoor airflow within a conventional building model is significantly impacted by the outside environment [11]. Meanwhile, Himaya et al. (2015) highlighted that the difference in enthalpy, or the amount of heat equivalent in thermodynamic measures, between indoor and outdoor environments affects the cooling impact brought on by natural ventilation [12].
Furthermore, the relationship between natural ventilation strategies, indoor thermal comfort, and IAQ has also been the focus of recent ventilation-based building research. A study examined the indoor comfort conditions in a typical classroom in a school and investigated the effect of natural ventilation on both indoor thermal comfort and IAQ [13]. This study conducted a field survey and, with the help of correlation analysis, determined that implementing natural ventilation strategies at each break time positively affects the reduction of CO2 concentration without compromising indoor thermal comfort.
Similarly, in addition to determining the natural ventilation potential for a windcatcher, air velocity, and temperature variables were taken into consideration to assess indoor thermal comfort conditions in a CFD setting by Goudarzi et al. (2021) [14]. Likewise, Gupta et al. (2021) also used the CFD simulation package "Cradle scSTREAM" to examine the thermal comfort of a naturally ventilated hostel building in the composite climate of Jaipur, India. The primary result of the study is a thermal comfort range based on the ASHRAE-55 standard [15].
Overall, the literature indicates that studies have been carried out in the field of evaluation of current approaches for analysing the ventilation performance of buildings that can provide an acceptable ventilation rate and a thermally comfortable environment for building occupants. Even though the investigation of natural ventilation performance using a potent tool for resolving airflow problems, such as CFD, is well-advanced, the relationship of it with indoor thermal comfort is still determined by air temperature and velocity values and ranges. Indoor thermal comfort, however, is not only related to air temperature and velocity but also clothing insulation, metabolic rate, radiant temperature, and relative humidity, according to international comfort standards as those set by ASHRAE and the International Standard Organization (ISO) [4]. The two most highly reliable thermal comfort indices are predicted mean vote (PMV) and predicted percentage of dissatisfaction (PPD). However, Fanger suggests that the expectancy factor needs to be included for naturally ventilated buildings in warm regions for estimating PMV-PPD [16].
Additionally, indoor thermal comfort will vary point by point depending on changes to passive design features, so a single number or its ranges for the entire indoor space will not be sufficient. Therefore, investigating the spatial variation of indoor thermal comfort indices and their relationship with natural ventilation potential and passive design characteristics will add value to this domain, particularly for a room configuration where occupants are dispersed throughout the building spaces, such as classrooms.
To address the aforementioned gaps, the primary objectives of this study are: 1. Assessment of the ventilation performance and indoor thermal comfort concerning passive design strategies, specifically, opening characteristics such as the Opening to Wall ratio (OWR) and Window to Wall ratio (WWR) in a Computational Fluid Dynamics (CFD) environment for a classroom setting.

To determine the spatial variation of Predicted
Mean Vote (PMV) and Predicted Percentage Dissatisfied (PPD) to evaluate indoor thermal comfort and the natural ventilation performance in terms of air change rate per hour (ACH) using CFD simulation runs. Figure 1 below depicts the methodology adopted for this study, which can be fragmented into five main steps. Data from the building's case study room is gathered in the first step which is necessary for modeling the building space. In the second stage, the reference test case study room is modeled and simulated in the CFD environment in ANSYS Fluent 2020 R2 [17]. The CFD analysis is detailed in subsequent sections. Next, the simulated results are then gathered for the following parameters: temperature, air velocity, air change rate (ACH), and thermal comfort indices (PMV-PPD). The temperature and air velocity statistics are also validated against the case study room's field data. The same case study room is used as the basis for parametric simulations for passive design elements that are carried out in the following step. Finally, the dataset is assembled in the last stage, as shown in Figure 1, to comprehend the interactions between ventilation performance and indoor thermal comfort with regard to passive design components for buildings. The mathematical background for the airflow analysis utilised to calculate the ventilation performance and indoor thermal comfort in the CFD environment for the study is provided in the following section.

Airflow analysis in Computational Fluid Dynamics (CFD)
The problem of 3D indoor airflow is governed by the conservation of mass, momentum, and energy equations. Mass conservation in three-dimensional airflow is given by [18]: where P is pressure and ߤ is viscosity. Next, the energy conservation equation for the airflow, given in averaged, condensed form is as follows: where T is temperature, hs is sensible enthalpy and λ is thermal conductivity.
Further, it is necessary to include Reynolds decomposition in the continuity equation because of the turbulent nature of the airflow: where ‫ݑ‬ ത is the average velocity for a steady flow, and u í is the fluctuation velocity The governing equations for turbulent flow are as follows: where τ ij denotes viscous stress tensor.
The above equations (7), (8), (9), and (10) are combined to form Reynolds Average Navier Strokes Equation (RANS) as given below: where ρu ത i u í is Reynolds stress tensor, F ഥ is body force and s ij is the mean rate of the strain tensor.
Further, the standard k-ɛ model is used for closing the set of governing equations. The k-ɛ model is a turbulence model in which transport equations are solved for obtaining the values of the turbulent kinetic energy k and its dissipation rate ε [19]. μ t =c μ ρ k ε (12) where ߤ ௧ is turbulent viscosity. The aforementioned airflow modeling equations are solved in ANSYS Fluent 2020 R2 in a steady-state condition for the study.

Case study: Formulation of Base case
The study considers a reference classroom in an academic building to demonstrate the proposed methodology's applicability in a tropical climate such as India. The building is located in Mumbai, Maharashtra, India as represented in Figure 2. The location of the building has a latitude of 19.07°N and an east longitude of 72.87°E. Mumbai's climate is characterised as warm and humid, with an average annual mean temperature of 27.2 °C and average annual precipitation of 242.2 cm. In the months from September to November, Mumbai experiences lesser humidity, lower diurnal temperatures, and moderate wind speeds, these conditions are ideal for using natural ventilation in buildings. The study is conducted in mid-November, when the relative humidity is 62%, the average temperature is 26°C, and the mean wind speed is estimated to be 1 m/s (ISHRAE data Mumbai).

Fig. 2. Location of the reference building
The reference classroom is located on the second floor of the two-storied academic building. Table 1 lists the geometry specification information that is gathered from the field survey. There are 40 benches occupying a total area of 76.2 m 2 in the classroom. Figure 3 shows the geometry-specific model of the classroom created in Autodesk AutoCAD 2023 [20].

Geometry and computational domain
In Ansys 2020 R2, the AutoCAD model is imported into SpaceClaim. The model's geometry is cleaned up in SpaceClaim in order to be processed further. The geometry's interferences are also cleaned up and removed. The dimensions of the classroom are 10.08 m (L) × 7.56 m (W) × 2.5 m (H), and the entire computational domain taken into account in the study is Length × Width × Height = 12 H × 4 H × 5H, per the standard recommended for the numerical analysis in CFD environment [14].  . 3. AutoCAD model of test study room (a and b). All dimensions in meters.

Mesh generation
The next step in airflow modeling is meshing, which involves dividing up the geometry's entire volume into discrete, recognizable volumes so that the Navier-Stokes and turbulence equations can be solved and the truncation error in numerical simulation may be ignored. The unstructured meshing with tetrahedral elements is chosen in the study to perform meshing, which can be seen in Figure 4. The overall mesh element counts ranged from 73,845 and every mesh metric was confirmed to be within the allowable range as depicted in Table 2.
Further, the grid independence test of the mesh is also examined as shown in Figure 5. The average temperature is the chosen parameter to perform the grid independence test against the element size. The optimum fine element size, which is found to be 0.02 m based on the test findings, is established.

Boundary conditions and solver settings
In Fluent, boundary conditions are set for the base case after mesh generation. In this case, the wind-driven flow E3S Web of Conferences 396, 02023 (2023) https://doi.org/10.1051/e3sconf/202339602023 IAQVEC2023 dominates and the buoyancy impact is discarded. The fluid material is set as "Air Ideal Gas" and the Reynolds Average Navier Strokes Equation (RANS) and standard k-ɛ is used to simulate turbulence of the indoor air. A fixed temperature of 25°C has been set for the walls, ceiling, and floor. The walls are set to be in the no-slip boundary condition. The wall material is concrete with the following characteristics: Density = 2400 kg/m 3 ; Specific gravity = 750 J/kg-K, and Thermal conductivity = 2.5 w/m-K. For the Mumbai region for the days of November, hourly-based ISHRAE data are used to determine the air temperature and wind speed. The simulation considers outlets as pressure outlets and inlets as velocity inlets. Humans with a value of 70W/person for the classroom are employed as heat sources in the simulations.

Validation of simulation model
The airflow velocity and temperature reported by CFD in this study for the base classroom are validated by comparing them with the experimental field results collected by an anemometer.

Fig. 5. Grid independence analysis test
Using an HTC-AVM-08 anemometer with a reading accuracy of 5.0%, the field measurements were carried out in the classroom during two consecutive days on November 11 and 12, 2022, from 9 AM to 6 PM with varied occupancies. Two mechanical fans were put on while the windows and doors remained open, respectively during these hours. In the classroom, the anemometer was positioned at X = 0.32 m from the back, Z = 0.6 m above the floor, and Y = 3.78 m from the inlet of the reference classroom. In order to compare the simulated result with the field measurement, the same point is taken into consideration. Figure 6 below illustrates the variations in airflow temperature and velocity between the field measurement and simulated outcomes. The reference case boundary conditions are based on hourly data from the ISHRAE Mumbai climate file rather than real-time weather station data, and it also considers symmetric wall assumptions. The differences between the results of the turbulence model and field experiments are explained by these two significant factors. For air velocity and temperature, the Root Mean Squared Error (RMSE) between CFD results and field results are found to be "0.12" and "0.69," respectively, which is acceptable for numerical validation [21]. Figure 7 represents the velocity and temperature profiles of the base classroom at different floor levels.

Thermal comfort and ventilation performance assessment
Following the validation of the base case classroom, the study conducted parametric simulations of OWR and WWR. It is performed in a CFD environment to examine ventilation performance and indoor thermal comfort for a naturally ventilated classroom setting when doors and windows are kept fully open to external conditions. Given that the air velocity and temperature were seen to be constant throughout the month, 10 AM is chosen as the time for the parametric simulation.
Indoor thermal comfort is determined in terms of the corrective PMV-PPD considering an expectancy factor of 0.5 at the floor height of 0.6 m following ASHRAE codal regulation. The spatial variation of indoor thermal comfort is plotted in ANSYS Fluent using a UDF (User Defined Function) file based on the PMV and PPD equation provided by the ASHRAE code. The UDF is essentially a C programme or a C function that ANSYS Fluent can dynamically load to augment its default functionality. These indices are the function of airspeed, temperature, clothing insulation, metabolic rate, radiant temperature, and relative humidity. The constant value of relative humidity = 62%; clothing insulation = 0.55; radiant temperature = air temperature, and metabolic rate = 1.2 are assumed for the parametric simulations. Further, the findings of the simulation are used to estimate the PMV and PPD while considering the varying values of air velocity and temperature within a classroom. The acceptable level of indoor thermal comfort, as defined by ASHRAE, is provided by PMV values in the 0.5 range and that of PPD within 10% [4]. The efficiency of natural ventilation is measured in air changes per hour (ACH). One ACH denotes the amount of outdoor air that has been introduced into the space being vented in one hour. It is calculated by dividing the volumetric air flow rate by the volume of the space.
For performing the parametric simulations of OWR and WWR, the study takes the following scenarios into account, as given in Table 3 (a and b). OWR is not advised below 5% for warm and hot locations. Additionally, openings for neighboring classes are not permitted in academic buildings' adjacent walls. Therefore, the analysis takes into account three OWR cases: the base case, the lowest OWR recommended, and the highest OWR possible in the indoor functional space. For these simulations, the standard case of three windows and two doors is maintained. Further, the study considers three different WWR scenarios: the lowest WWR recommended (three windows), the lowest WWR recommended with two windows, and the highest WWR that can be achieved in a classroom. For parameterized cases of WWR and OWR, the spatial variation of PMV and PPD can be seen in Figure  8 for a height of 0.6 m above the floor. The areaaveraged velocity, temperature, PMV, PPD, and ACH values calculated in a simulated environment are shown in Table 4 as a summary table for the respective models.

Discussion
Due to the design and season of the year considered for the base case classroom, all parametric instances are simulated with the inlet opening (windows) being presumed to be in a windward direction using crossventilation techniques. Table 4 indicates that PPD% is declining while OWR% is increasing and ACH is also advancing concurrently.
Although less air volume is delivered by Model 1 into the classroom and has a lower percentage of OWR than Model 2, it performs better than Model 2 overall. In Model 1, the air is successfully circulated in the classroom and has reached the outlet with a reasonable velocity because the inlet's reduced aperture size has improved its velocity and since the inlet's and outlet's aperture sizes are nearly identical.  Further, Model 5 is the worst-performing scenario with two windows when it comes to parametric simulations for WWR. More stagnant air can be detected in this situation close to the wall's non-window side. Overall, Model 3 and Model 6 have delivered the best ventilation performance and indoor thermal comfort, respectively. OWR and WWR are therefore important passive design elements that should be evaluated for analysing ventilation effectiveness and indoor thermal comfort. Additionally, to improve indoor airflow and velocity within indoor spaces with crossventilated techniques in wind-driven ventilation, apertures' positions and relative dimensions of inlet and outlet should be given priority over increasing their sizes.

Conclusion
The study investigated the ventilation performance and indoor thermal comfort with regard to passive design strategies, considering, opening characteristics such as OWR and WWR in a CFD environment. Using field data gathered for air velocity and temperature, the study's first section validated a base case simulation of a E3S Web of Conferences 396, 02023 (2023) https://doi.org/10.1051/e3sconf/202339602023 IAQVEC2023 classroom on the second floor of an academic building in a tropical climate. The effectiveness of ventilation and indoor thermal comfort are evaluated in the next section using parametric models for OWR and WWR in a natural ventilation setting. The findings illustrate that by implementing higher OWR and WWR percentages in building envelopes, ventilation effectiveness, and indoor thermal comfort can be improved. Nevertheless, the relative sizes of the inlet and outlet apertures and their positions can alter the indoor airflow and velocity, which may increase or reduce the effectiveness of ventilation and the level of indoor thermal comfort.
Future research should focus on developing such sensors and transducers that can accurately evaluate natural ventilation performance and indoor thermal comfort at different locations of the indoor space. This would help determine the optimum passive design parameters for sustainable habitats as a result of incorporating these variables into the indoor environment design.