1. Introduction
For the past few years, with social and economic development, the people’s health-centered national fitness strategy has also been progressing [
1]. The policy drive, holding sports events such as the Winter Olympics, and improving material living standards have increased the demand for physical fitness sports and resulted in an explosive growth of sports buildings [
2]. In light of today’s major building energy consumption problems [
3], research on green energy conservation in sports buildings is significant to save energy, protect the environment, and create a comfortable and healthy sports and fitness environment.
Natatoriums, as a class of sports buildings, are characterized by large permeable spaces, high comfort requirements, high temperature, high humidity, and high energy consumption [
4]. The indoor water temperature of swimming pools should be controlled within a comfortable range, and the constant heat dissipation of the pools and the large internal energy consumption are especially obvious in winter. Swimming pools require constant temperature and humidity throughout the year, requiring year-round mechanical exhaust and air conditioning to regulate the temperature. The water in swimming pools is now commonly purified by circulation. The constant temperature in the building results in higher evaporation, higher indoor water vapor content in all seasons, and higher relative humidity [
5,
6]. These characteristics of swimming pools require much ventilation; therefore, the rational use of natural ventilation is the primary way to reduce the energy consumption of swimming pools [
7].
The traditional idea of energy saving for enhancing indoor ventilation is mainly an environmental conditioning system with natural ventilation as a supplement and mechanical ventilation as the main component. Air conditioning systems regulate the room’s high temperature and humidity [
8]. Thermal and wind pressure ventilation are the two most common primary forms of natural ventilation in buildings. However, these two primary forms of ventilation are influenced by outdoor wind speed, building shape, and surrounding environmental weather and need to be more easily controlled [
9]. In recent years, the popular design of a double-layer maintenance system can organize the airflow inside the building and achieve good wind circulation by using the cavity between the double-layer skin, which can better achieve summer ventilation, heat insulation, and winter insulation. However, this structure is not widely used because of its limitations, such as expensive costs and occupying a large indoor floor area [
10,
11,
12]. Compared with complete natural ventilation, mechanically assisted natural ventilation reorganizes airflow or even “forces” it locally to redirect it, making this ventilation achieve better results than natural ventilation. However, this mechanical device, locally used as auxiliary power in the building, requires additional energy to drive it [
13,
14,
15]. According to a survey, the energy efficiency of air conditioning heat pumps improved by 2.2 from 1985 to 2005, while it improved by only 0.8 from 2005 to 2015 [
16]. In other words, the energy efficiency ratio of the technical means of mechanically assisted natural ventilation is close to the theoretical limit.
Natatoriums are often in a high-temperature and high-humidity internal thermal and humid environment due to the particular characteristics of the space. The indoor thermal and humid environment seriously affects people’s comfort without the use of mechanical equipment such as air conditioners, and the human body often feels hot or cold [
17,
18]. Using the natural environment to regulate the internal environment of swimming pools is a new idea for energy saving in swimming pool buildings [
19]. Therefore, this paper proposes a more practical construction technique for using natural ventilation: the variable roof technique. This construction, which can be controlled artificially, converts the internal space of swimming pools into the outdoor environment during the high-temperature season, converts the indoor space into the outdoor space, and quickly regulates the heat and humidity of the environment [
20]. The technology of variable roofs can achieve the purposes of low carbon and energy saving but also enhance the flexibility and interest of the internal space of the swimming pool and position the swimming pool from the perspective of the double meaning of space and green energy saving.
Retractable roof technology (
Figure 1) has climate-responsive properties. Firstly, it can make full use of climate conditions to regulate the indoor environment by regulating the opening and closing of the envelope to achieve connected isolation from the outdoor climate [
21]. Secondly, it can solve the problems of fresh air, lighting, moisture drainage, mold prevention, and temperature regulation in swimming pools all at once. It can separate from the outdoor environment when the outdoor climate is unfavorable and regulates the internal environment traditionally [
22,
23].
Retractable roof technology also has the characteristic of spatial variability. It can use active technical means to control the building space, breaking the deadlock of previous research focusing on passive energy-saving technologies (
Table 1). It also links energy-saving means with the building space to form a variable space, opening up new directions and theories for the design ideas and energy-saving research of swimming pools [
24,
25].
This paper takes the Jiading Natatorium of Tongji University (
Figure 2) as an example. It uses Autodesk Ecotect Analysis 2011 to simulate and analyze the variable ventilation of its open and closed roof technology to verify the advantages of this new environmental conversion technology.
2. Methodology
Computational fluid dynamics (CFD) involves various disciplines, including fluid dynamics, computational methods, and computer graphics. CFD simulation entails creating computer models based on architectural designs and using Autodesk Ecotect Analysis 2011 to simulate the wind environment around and within completed and operational buildings. By integrating relevant knowledge of architectural technology with simulation results, architects can accurately predict and vividly describe the architectural wind environment of design proposals through scientific analysis. This, in turn, enables architects to compare multiple design options and refine architectural design proposals more effectively. In recent years, both domestically and internationally, there has been significant research and practice in this field, with CFD serving as a promising method and tool in the architectural design domain [
26,
27].
CFD numerical simulation involves solving the fundamental equations governing fluid motion using numerical methods to simulate fluid behavior. This enables the derivation of discrete physical quantities in the field, such as wind speed and direction, at different spatial locations when testing and analyzing the wind field inside or around buildings. This approach is applicable to studying idealized standalone buildings, two-dimensional street canyons, or three-dimensional building clusters. Furthermore, it can simulate natural urban environments, facilitating the evaluation of building ventilation performance, risk assessment, and building design optimization to enhance airflow dynamics [
28].
Direct numerical simulation (DNS), large eddy simulation (LES), and Reynolds-averaged Navier–Stokes (RANS) methods are commonly used numerical simulation techniques for turbulent flows, each with essential differences in how they handle turbulent flow. DNS involves solving the Navier–Stokes equations directly to simulate flow, considering turbulent structures of all scales, including the smallest eddies. LES, however, simulates turbulent flow by spatially filtering the flow field, retaining large-scale turbulent structures while filtering out small-scale eddies. However, for high-speed turbulent flows, such as those encountered in hypersonic conditions, the grid resolution must be very high to capture the required scales accurately. DNS and LES methods require high computational resources and are difficult to apply to complex models or simulations involving multiple buildings. In contrast, RANS primarily involves time-averaging the flow field to simulate turbulent flow, neglecting instantaneous variations, and using turbulence models to handle turbulent terms in the equations, thereby closing the equations. RANS does not require the flow field’s detailed spatial or temporal resolution, making it less computationally demanding. It is typically suitable for large-scale flow problems and is currently the most widely used turbulent flow simulation method [
29].
The SST k-ω and RNG k-ε models have superior capabilities in capturing flow details, making them suitable for complex flow simulations or scenarios requiring high-resolution flow details. The realizable k-ε and standard k-ω models also exhibit good performance in capturing flow details, making them suitable for simulations with high-resolution flow requirements. In contrast, the standard k-ε model used in this study, while relatively weaker at capturing some flow details compared to the previous four models, still provides satisfactory simulation results for overall flow trends. It can generally be used in typical scenarios. This model consumes fewer hardware resources during computation, exhibits low fluctuations and high precision in numerical calculations, possesses reasonable convergence rates and relatively low memory requirements, and yields effective solutions for flow problems around complex geometries. It is also commonly used in simulations related to building structures [
30].
The equations and methods employed in CFD simulations are critical components of the study. To ensure our approach’s scientific rigor and reliability, we have adopted mainstream CFD simulation techniques and utilized formulas and models widely applied in architectural simulation.
This study utilized the standard k-ε model, a commonly used turbulence model in CFD simulations. It is based on two transport equations used to describe turbulent kinetic energy and turbulent dissipation rate. The standard k-ε model is widely applied in architectural CFD simulations due to its relatively good computational stability and accuracy.
Furthermore, various architectural simulation studies have validated and extensively used the standard k-ε model. These studies encompass different types of buildings, complex wind environments, and various ventilation systems. This widespread validation and application make the standard k-ε model a mainstream choice for architectural CFD simulations.
The article utilized Ecotect Winair as the CFD numerical simulation tool. It employed the standard k-ε model to simulate the wind environment around the Wuhu National Fitness Center and various types of roofed sports buildings. The standard k-ε model is based on two transport equations, one for describing the transport and diffusion of turbulent kinetic energy k, and the other for describing the turbulent dissipation rate ε [
31]. The fundamental equations of the standard k-ε model are as follows:
The transport equation for turbulent kinetic energy k is given by:
where
represents the turbulent kinetic energy generated by the mean velocity gradient;
is the turbulent viscosity determined by turbulent kinetic energy and dissipation rate; and
is the turbulent Prandtl number for k.
The transport equation for turbulent dissipation rate ε is given by:
where
and
are model constants and
is the turbulent Prandtl number for
.
The formula for turbulent viscosity
is given by:
is an empirical constant in the standard k-ε model.
3. Research Model and Setting Information
3.1. External Environment
Shanghai is situated on the eastern coast of China, at the mouth of the Huangpu River. Its geographic coordinates range approximately from 30°40′ to 31°53′ north latitude and from 120°52′ to 122°12′ east longitude [
32]. As shown in
Table 2, Shanghai experiences a subtropical humid monsoon climate with distinct seasons. Summers are hot and humid, while winters are cold and damp. The total annual sunshine duration is approximately 1945.3 h, with a sunshine percentage of 45%. July is the hottest month, with an average monthly temperature exceeding 28 °C. The summer season extends for about three months, with extreme temperatures reaching up to 39.6 °C. The average relative humidity during the hottest month is 76.6%. Shanghai’s summers are characterized by sultry heat, with the plum rains leading to high humidity levels and reduced solar radiation [
33].
The primary function of the open–close roof is to enhance ventilation and heat dissipation. The most unfavorable parameters are selected, namely the highest temperature in summer and solar radiation. The outdoor wind speed is set to the annual average of 1.5 m/s, with the predominant wind direction in summer.
3.2. Model Simplification and Setting
As this study primarily focuses on the impact of retractable roofs on the indoor environment of the swimming pool, the architectural model has been simplified accordingly. Only the swimming pool hall with its retractable roof is retained, while other sections of the sports center are omitted. The aesthetically pleasing curved surfaces of the swimming pool hall are represented by a plane inclined based on the roof’s slope (
Figure 3,
Table 3) [
34,
35].
The main variables that could affect the effectiveness of opening and closing the swimming pool include solar radiation, external wind conditions, outdoor temperature, and the ratio of opening area to roof area [
36]. Among these variables, only the ratio of opening area is directly related to the building itself. Therefore, the simulation only considers the ratio of opening area and applies a gradient of 12.5% to generate five operational scenarios. According to
Table 3, the opening area ratios are 0%, 12.5%, 25%, 37.5%, and 50% (representing the actual maximum opening area).
In this computational fluid dynamics (CFD) simulation, the computational domain is divided into discrete, unstructured grids. Each grid cell represents a control volume, where equations are discretized. By solving the discretized equations, the velocity distribution of the fluid can be obtained at each grid cell.
In this study, we employed a simplified model to simulate the ventilation environment of a stadium. The focus was on investigating the impact of a retractable roof on indoor ventilation. Therefore, the model primarily encompasses the main hall of the stadium and the retractable roof section while disregarding other non-critical parts. The plan dimensions of the building model were set to 50 m by 70 m, with an overall height of 20 m, an eave height of 13 m, and a total roof height of 7 m.
In the model setup, key parameters included solar radiation, external wind speed, outdoor temperature, and the opening area ratio to the roof area. Within this setup, the ratio of the opening area to the roof area is a directly relevant parameter. Therefore, the simulation process primarily considered opening area ratios of 12.5%, 25%, 37.5%, and 50% as simulation parameters for different operational scenarios.
The details of the model setup were referenced from the case studies presented in
Table 1.
Table 1 outlines various types of stadiums with retractable roofs constructed domestically, including their substructure systems, roof styles, opening mechanisms, and roof materials. These pieces of information provided crucial references for the model setup.
By comparing these case studies, we ensured that the model setup aligns with the structural characteristics of stadiums in reality. For instance, the choice of substructure system influences the stability of the building, the roof style determines the opening method, and the opening mechanism and roof material directly affect the ventilation performance. These insights guided the selection of parameters in the model setup, ensuring that the simulation results were relevant to real-world engineering.
In summary, providing a detailed description of the model setup and associating it with the case studies in
Table 1 helps ensure the scientific rigor of the simulation process and the reliability of the results.
3.3. Grid Independence Verification
This study conducted grid independence verification to ensure the accuracy and stability of the CFD simulation results. This verification aims to confirm the stability of the simulation results under different grid densities, thereby ensuring that the chosen grid density is sufficiently fine.
In this study, we utilized an unstructured mesh generation method. The initial mesh density was determined based on the complexity of the building and critical areas to ensure sufficient accuracy in the simulation. Three different mesh density schemes were created to test grid independence: low, medium, and high density. The mesh details for each scheme are as follows:
Low density: partitioned based on more prominent mesh elements to simplify the simulation.
Medium density: finer than the low density to ensure sufficient mesh details in critical areas.
High density: the finest partition used to validate the stability of the simulation results.
We selected critical parameters in the simulation, such as wind speed and pressure distribution inside the building, for grid independence testing. Simulations were conducted for these parameters under the three mesh densities, and the results were compared to ensure stability.
The test results indicate that the simulation results stabilize as the mesh partitioning progresses from low to high density, with the variation in critical parameters gradually decreasing. The parameter variation between medium and high densities is minimal, indicating good stability of the simulation results.
Based on these test results, we have selected the medium-density mesh scheme. This choice ensures both the accuracy of the simulation and avoids excessive consumption of computational resources. This scheme achieved stable results in the simulation, demonstrating the effectiveness of grid independence verification.
4. Simulation Analysis and Results
The wind speed cloud diagrams of four operating conditions were compared, as shown in
Table 4, showing that different ratios of the opening area could have a significant effect on the indoor wind speed field. When the ratio of the opening area was 12.5%, the wind speed would be from 0.11 to 3.01 m/s. Although the maximum wind speed was high, the overall indoor airflow field would be uneven, and the area with low wind speed would be significant, forming a dead corner for the airflow. As the ratio of the opening area was increased to 25%, the maximum wind speed at the data point would be slightly decreased, and the wind speed would be about 0.01 to 2.76 m/s. However, the wind speed field was uniform, and the low wind speed area would be significantly reduced. Under the two operating conditions of 37.5% and 50%, the maximum wind speed was further reduced, the wind speed field was more uniform, and there was no apparent low wind speed area in the room [
37]. The influence of the ratio of the opening area on the indoor wind speed field can also be seen from the airflow streamline diagrams of the four working conditions. As the ratio of the open area continues to increase, the area affected by airflow changes from more concentrated to more diffused. The open and closed roof technology will enhance the natural ventilation in the room [
38,
39].
Observe
Table 4 and
Table 5. We analyzed the wind speed, pressure, and streamline under different window opening ratios. Firstly, we observed wind speed distribution at different window opening ratios. At a 12.5% opening ratio, the indoor wind speed was high but uneven, with significantly low wind speed areas. As the opening ratio increased to 25%, the wind speed slightly decreased, but the distribution became more uniform, and the low wind speed areas were reduced. Further increases to 37.5% and 50% opening ratios resulted in continued decreases in wind speed, leading to a more uniform wind speed distribution without significant low wind speed areas [
40]. Consequently, the 37.5% opening ratio was deemed the ideal choice. Secondly, we examined the effect of window opening ratios on thermal pressure ventilation. Increasing opening ratios enhanced indoor thermal pressure ventilation, resulting in better natural ventilation and more uniform airflow. However, excessively high opening ratios could lead to uncomfortably high indoor wind speeds [
41,
42]. Finally, we synthesized the considerations of wind speed uniformity and thermal pressure ventilation effects. Despite the broader and stronger airflow at a 50% opening ratio, it lacked sufficient moderation. In contrast, the 37.5% opening ratio provided adequate ventilation without generating excessive wind pressure, ensuring both uniform and stable indoor ventilation.
In
Figure 4, the “distance” in the horizontal coordinate refers to the distance to the facade of the building. Negative distance values indicate that the air has passed through the building, implying significant changes in airflow after passing through. Positive distance values indicate that the air has not yet passed through the building, suggesting that airflow may be less affected by the building at this point, resulting in smaller changes.
As shown in
Figure 4, the following information can be obtained:
For all opening area ratios, wind speed shows a trend of initially increasing and then decreasing with distance. This suggests that the maximum wind speed does not occur at the point closest to the opening but rather forms at a certain distance away [
43].
At a 12.5% opening area ratio, the wind speed rises most rapidly in the graph, reaching the highest maximum wind speed. This may be because smaller openings concentrate airflow, resulting in higher wind speeds in specific areas.
From 12.5% to 50% opening area ratio, the maximum wind speed gradually decreases. Meanwhile, the wind speed distribution becomes gentler with the increasing opening area, suggesting a more even distribution of airflow indoors.
At both ends of the distance in the graph, all wind speed values are close, suggesting that at locations further from the opening, wind speed is less influenced by the opening area [
44].
5. Discussion
As retractable roofs become more prevalent, an increasing number of sports-related buildings are embracing this flexible and adaptable roof design. This section begins by categorizing and classifying common roof forms and opening mechanisms in sports arenas. Subsequently, computer software simulations are employed to analyze each type, examining the influence of variable roof designs and opening rates on the architectural wind environment.
5.1. The Common Forms of Retractable Roofs
Common types of retractable roofs include sliding roofs, single-sided sliding roofs, folding roofs, and rotating roofs. Among them, sliding roofs can take three forms: flat, sloped, and domed, depending on the roof shape. As shown in
Figure 5 and
Table 6, sliding roofs can adjust opening and closing without significantly altering the roof shape, meeting architectural lighting and ventilation requirements. Single-sided sliding roofs and folding roofs are designed with more specific considerations for architectural lighting or ventilation needs compared to their dual-sided counterparts. Folding roofs, with the advantage of “multi-unit and miniaturization,” reduce concerns about structural stability during design. Rotating roofs effectively block wind force, mitigating the risk of excessive wind speeds at the opening and ensuring indoor comfort to some extent. When designing stadiums, the appropriate retractable roof forms can be selected based on specific requirements [
45,
46].
To facilitate simulation analysis, the building is uniformly set to the standard plan dimensions of small- to medium-sized sports facilities: 50 m × 70 m × 20 m. The building’s uniform height is set to 20 m, with the eaves height adjusted downwards to 13 m, and the total roof height set at 7 m (
Figure 6). For different types of retractable roofs, the opening range is set to one-third of the total roof area. For single-sided sliding roofs, the opening is one-sixth of the total roof area (equivalent to one-third of one side of the sloped roof).
5.2. Simulations of Wind Environments for Different Forms of Retractable Roofs
5.2.1. Simulation of Wind Environments for Different Roof Types with Constant Opening Ratios
As shown in
Table 7 and
Figure 7, we set the opening ratio of all six roof types to 37.5%. Different roof forms have distinct effects on the distribution of indoor wind speeds at the same opening ratio. Single-sided sliding slope roofs and sliding gable roofs exhibit a distribution pattern where wind speed increases from one side to a central peak and then decreases on the other side, reflecting the concentration effect that roofs may induce. Sliding dome roofs, on the other hand, display broader areas of high wind speeds, suggesting that dome structures may lead to a more uniform airflow distribution. In contrast, the wind speed distribution on flat roofs of sliding roofs appears more even without prominent peaks, indicating that flat roof structures may help disperse wind speeds. Wind speeds on the flat roofs of folding roofs peak in the central area but exhibit overall gradual changes, while those on the flat roofs of folding telescopic roofs show relatively lower wind speeds in the central area with peaks on both sides, possibly due to their unique structural design [
47,
48]. In summary, the shape of the roof and the opening ratio have a significant impact on wind speed distribution, which is a critical consideration for ventilation design.
5.2.2. Wind Environment Simulation Analysis for Six Types of Roofs at Various Opening Ratios
As shown in
Table 8, these line graphs depict the wind speed distribution for six different roof opening configurations. Each configuration is represented across four different opening area ratios (12.5%, 25%, 37.5%, and 50%).
From these graphs, it is evident that both single-sided sliding slope roofs and sliding gable roofs show a decreasing trend in peak wind speeds as the opening area increases. Compared to sloped roofs, sliding dome roofs exhibit higher maximum wind speeds at smaller opening area ratios, possibly due to the guiding effect of the dome structure on airflow. The flat roof structure of sliding roofs demonstrates relatively gradual changes in wind speed at moderate opening area ratios (25% and 37.5%), suggesting a more stable airflow distribution. For folding roofs with flat tops and folding telescopic roof structures with flat tops, the variation in maximum wind speed is relatively small across different opening area ratios, indicating consistent wind speeds maintained by folding roof structures at various opening area ratios. The folding telescopic structure appears to offer a more uniform wind speed distribution, especially at larger opening area ratios [
49].
Moreover, for all types of roofs, the peak wind speed decreases as the opening area increases, and the wind speed distribution becomes more uniform, which aligns with the earlier analysis. The location of the wind speed peak is influenced by the type and size of the opening, but the trend changes remain relatively consistent for each roof type.
5.3. Live Experiment and Validation
To ensure the accuracy of simulation results, we designed a series of experiments to validate the simulated data and compare it with actual conditions. The objective of the experiments is to verify the predicted indoor ventilation effectiveness of different window opening ratios through actual measurement data.
Experiment Design: The experiments were conducted at the Jiading Sports Center in Shanghai. We measured vital parameters such as indoor wind speed and temperature at different window opening ratios. These parameters are related to variables in the simulation, and experimental data can be used to assess the accuracy of the simulation results.
The experiments focus on evaluating the impact of different window opening ratios on indoor ventilation. We selected four different opening ratios, 12.5%, 25%, 37.5%, and 50%, and conducted multiple measurements at each ratio to ensure the reliability of the experimental data.
Experimental Method: Multiple sensors were installed at different locations within the sports center to measure wind speed and direction. The experiments covered a range of window opening ratios from 12.5% to 50%, with multiple measurements taken at each ratio to ensure the reliability and stability of the data.
The experimental results indicate that indoor wind speed and the actual experimental results are consistent with the simulated results, demonstrating the accuracy of the simulation. Wind speed varies with the increase in window opening ratio. At a 12.5% window opening ratio, the indoor wind speed is relatively high, with a maximum speed of 3.2 m/s, but the airflow field is less uniform, with many low-speed areas. As the window opening ratio increases to 25%, the wind speed becomes more uniform, with the maximum speed decreasing to 2.8 m/s. The wind speed field is most uniform at a window opening ratio of 37.5%, with a maximum speed of approximately 2.6 m/s and significantly fewer low-speed areas. Under a 50% window opening ratio, indoor ventilation is significantly enhanced, but the indoor temperature slightly decreases due to excessive ventilation.
Experimental Analysis: The experimental results suggest that a window opening ratio of 37.5% performs best regarding indoor ventilation effectiveness and wind speed uniformity. At this ratio, indoor ventilation is sufficient while avoiding temperature decreases and discomfort caused by excessive ventilation. Although a 50% window opening ratio provides higher ventilation effectiveness, it may lead to a decrease in indoor temperature, affecting comfort.
Experimental Conclusion: Based on the experimental results, the optimal window opening ratio is 37.5%, which ensures good ventilation effectiveness while maintaining indoor temperature stability. The experimental results are consistent with the simulated results, validating the accuracy of the simulation. Additionally, the experiment demonstrates that a reasonable window opening ratio significantly impacts the comfort and energy efficiency of the sports center.