0 Introduction

Accompanied with the rapid economic growth and substantial increase in population, a wide variety of buildings have emerged. These buildings inevitably caused increasing building energy consumption which has come into focus all over the world^[1-2]. In recent years, many energy evaluation models were used to evaluate the energy performance and save the energy of buildings. Han et al.^[3-5]proposed the novel radial basis function (RBF) based on affinity propagation (AP). Geng et al.^[6-7] proposed a convolutional neural network (CNN) integrating the cross-feature (CF) (CF-CNN) method for energy optimization and analyzed the structural information of the syntactic dependency for the relation extraction task. Specifically, the concept of nearly zero energy buildings is proposed with a target of reducing the building energy demands to a minimum degree and greatly improving the energy use efficiency^[8-9]. The main technical measures for realizing the nearly zero energy buildings are usually done by means of passive building design through adapting to the local climatic characteristics and site conditions. This article focuses on high energy consumption in buildings and contributes to near zero energy consumption and intends to achieve a breakthrough in passive building design.

Among the building energy consumption components, the air conditioning energy consumption turns out to make up the most important part^[10-11]. Especially in tropical and subtropical regions, the hot and humid climate conditions would intensify the use frequency of air conditioning systems and greatly increase the air conditioning energy consumption^[12-13]. This phenomenon is further outstanding for public high-large space buildings, in which over 50% of total energy consumption came from air conditioning^[14].

Considering the large proportion of air conditioning system occupying in the whole building energy consumption under hot and humid climatic conditions, an effective passive design to reduce the air conditioning energy consumption for high and large space buildings is highly necessary. Taking a deep inspection of the parts constituting the air conditioning energy consumption, two aspects are specially focused on. One is the cold source which is used to produce necessary cold air to eliminate indoor cooling load. The other one is the power equipment which is used to transport the cold air. Therefore, reasonably designing and optimization for air conditioning's cold source and power equipment play the key role in radically reducing air-conditioning energy consumption level for hot-humid areas.

Natural ventilation methods have attracted world-wide attention due to their lower energy consumption and less pollution^[15-16]. For high-large space buildings, the unique structural features with open large space actually provide great convenience for natural ventilation. However, since the outdoor air temperature in hot-humid regions usually present higher values during the long-term summertime period and cannot be directly used as reasonable cold source for air conditioning, the simple natural ventilation usually produces insufficient ventilation and cannot satisfy the cooling demands of high and large space buildings. Therefore, how to make use of the open large space to realize efficient natural ventilation and provide suitable indoor thermal environments through appropriate passive design is a major issue for maximally reducing the air-conditioning energy consumption in high-large space buildings.

For the long-term high air temperatures in hot-humid regions, effective precooling of the outdoor air would be a better choice for providing cold source. Research results have shown that the soil temperature below underground to a certain depth could reach 9℃ lower than the air temperature^[17]. In this way, the soil below underground could serve as natural and stable cold source for cooling the outdoor air through effective earth to air heat exchange during the long-term summertime in hot-humid regions. The earth to air heat exchange system (EAHES) has been applied as an energy-saving technology to passing buildings^[18]. Several researches have discussed the influences of operating parameters (e.g., pipe material, pipe diameter, air velocity) on the thermal performance of the EAHES^[19-22]. As can be seen from these studies, the EAHES technology could serve as stable cold/heat sources and contribute to energy-saving air conditioning design, thus greatly reducing building energy consumption.

To enhance the natural ventilation effects and increase the transmission power for transporting cold air, the solar chimney (SC) effects, which adopt thermal pressure to convert solar thermal radiation energy into air kinetic energy, have become the focus of attention in realizing enhanced natural ventilation. A series of articles have been conducted through field experiments^[23] and numerical simulations^[24] to analyze the various SC effects under different working conditions. Several factors influencing the system wind-thermal environmental performance of solar ventilation are discussed including the air gap depth, opening size, outdoor air temperature, solar radiation^[25-30]. These researches demonstrate that setting roof-top solar ventilation tower for high-large space buildings would efficiently enhance natural ventilation through the thermal pressure effects. This approach of applying roof-top solar chimney would act as effective power equipment transporting cold air and enhance natural ventilation effects in high and large space buildings.

Although the EAHES technology and SC approach have been widely accepted as passive measures in air conditioning application field, the current researches usually focus on the performance analysis of one single aspect. Specifically, the high and large space buildings have unique inner space and special structures. The comprehensive wind-thermal environmental characteristic of system applying both EAHES as cold source and SC approach as air transmission power have not been well expressed. The quantitative analysis and comparison for such a passive natural ventilation system under different design parameters of EAHES and SC methods need detailed discussion. Most of the existing researches only discussed the performance characteristic of the EAHES and SC, and did not discuss the thermal comfort of the system. The thermal comfort is an indispensable and important factor to evaluate a passive technology of buildings in architectural design.

Considering the actual application values and drawbacks of the existing researches, this paper takes an actual typical high-large theater located in tropical climatic regions as the study object. A passive composite enhanced natural ventilation system by applying both the EAHES technology and SC approach is designed for the target high and large space building. Wind-thermal environmental characteristic of system under different system design conditions by both considering the underground tunnel and solar chimney are discussed by applying the computational fluid dynamics (CFD) numerical simulation technique. The ventilation quantity (V_q), the distributions of air velocity (V_a) field and air temperature (T_a) field, human vertical temperature gradient difference (T_gd) are considered for displaying the performance of the passive composite enhanced natural ventilation system. This study would contribute to enhancing natural ventilation effects and help realize the nearly zero energy building concept for high and large space buildings in hot-humid tropical climate zones.

1 Designed Passive Composite Enhanced Natural Ventilation System 1.1 Designing Composite Enhanced Natural Ventilation System

Hainan Province has the biggest tropical region in China, widely known by its typical year-round high temperatures. Meteorological data show that the annual average temperature is 22.5-25.6 ℃, the annual sunshine hours are 1780-2600 h, the total solar radiation is 4500-5800 MJ/m², and the annual precipitation is 1500-2500 mm. This study chose an actual high and large space building, a theatre in Hainan, as the study object. The simplified flat layout map of the theatre is shown in Fig. 1(a). The building dimensions are 36.5 m long, 26 m high and 26.4 m wide. The audience area consists of 19-storey terraced seating, and each storey is filled up with 33 seats. This study assumes that the theater is running in full load condition, namely 627 people sitting in the theater.

References show that the soil temperature at 6 m below underground in Hainan could stably stay at about 17 ℃^[31]. In this way, the underground soil could provide sufficient cooling capacity for the outdoor air with high temperatures. Therefore, this study applies underground tunnels to realize earth-air heat exchange process and convey cold air into the theatre. Then considering the strong solar radiation amount in Hainan area, this study aims to apply the solar chimney to realize natural ventilation under thermal pressure. The solar chimney is set on the roof-top of the theatre with three-dimensional dimensions of 6 m long, 1 m wide, 20 m high. Considering that the theater usually operates both during the daytime and nighttime, the solar chimney effects could help enhance the natural ventilation during the daytime with abundant solar radiation. Then for the rainy days or nighttime lacking sufficient solar radiation, an axial flow fan is equipped in the solar chimney, as shown in Fig. 1(b). In this way, the ventilation system could ensure continuous sufficient ventilation quantity under various weather conditions. Fig. 1(b) displays the schematic diagram for the designed whole ventilation system of the theatre.

The theoretical air conditioning process and energy conservation effects of this designed system is briefly summarized as follows:

1) The designed system has two advantages compared with the traditional refrigerating methods. First, the underground tunnel-based EAHES takes the lower soil temperature as cold source. Second, the seating locations in the theatre gradually rise from front to back and therefore a large cavity under the audience seats is formed, which can be taken as plenum chamber. The cooled air first enters the plenum chamber, and then feeds into the theatre inner space from under the audience seats. The supply air process realizes the displacement ventilation process and contributes to the uniformity of both air flow field and temperature field. In addition, the cooled air could be directly applied to the audience area and discharged from the top of theatre without considering the cooling load above the audience. Compared with the dilution ventilation method, this displacement ventilation method from under the audience seats needs less refrigerating capacity and presents better energy conservation features.

2) In the meantime, the solar chimney continuously heats the air inside it and enlarges the air temperature difference between inside and outside the chimney, and generates significant air density difference which thus induces the air inside the solar chimney flowing outside. With the decrease of the air quantity inside solar chimney, the air pressure difference between the solar chimney and theatre inner space then induces the air inside theatre to circulate naturally, which reflects the exhaust capacity of the SC effects.

Taken as a whole, the underground tunnel-based EAHES not only serves as stable natural cold source, but also provides the theatre's displacement ventilation, while the roof-top solar chimney provides transportation power for natural ventilation by making good use of the strong solar radiation amount in Hainan to realize SC effects. Integrating the underground tunnel-based EAHES and the roof-top solar chimney effects, a passive composite enhanced natural ventilation system for high and large space buildings in tropical regions is proposed as an energy-saving air conditioning method.

Since the construction project has not been completed yet, this paper adopts the CFD to evaluate and analyze the system wind-thermal environmental performance. This paper considers seven simulation research groups respectively concerning different design conditions. The research process is shown in Fig. 2 and the detailed description will be shown in Section 1.2 and Section 1.3.

1.2 System Design Factors

Different system structural forms would cause varied influences on system wind-thermal environmental performance. Therefore, several basic system design factors are considered. First, the underground tunnel-based EAHES consists of two parts. One is the underground tunnel which is used for cooling outdoor air, the other is the plenum chamber under the audience seats which conveys the air dynamic pressure to static pressure and makes the supply air distributing more uniformly. Considering the characteristics of underground tunnel-based EAHES, three aspects of factors are considered, namely the underground tunnel length, the underground tunnel number, the plenum chamber's air supply location. Among them, the underground tunnel length and underground tunnel number would influence the cooling effects of EAHES, while the plenum chamber's air supply location would mainly influence the air supply effects.

Second, the solar chimney as the power equipment plays a key role in the system. Considering the structural composition of solar chimney, four factors are considered, namely the solar radiation intensity, the solar chimney height and width, the absorber plate angle. Fig. 3 shows the schematic diagram of solar chimney structure model. As shown in Fig. 3, the solar chimney height, length and width can be displayed by AE, AB and AD. Then the DHGC represents the absorber plate which can rotate at a certain rotation angle θ around rotation axis(DC). The solar chimney inlet and outlet are located in ABCD and EFGH respectively.

Each design factor corresponds to one research group, as shown in Table 1.

Among these research groups, Group One mainly considers the influences of different underground tunnel lengths on cooling and ventilation effects of the composite enhanced natural ventilation system by setting single underground tunnel, side air supply. Group Two changed single underground tunnel to double underground tunnel compared with Group One, and would discuss the system cooling and ventilation differences between working conditions No. 1-4 and working conditions No. 5-8. Then Group Three changed side air supply location to bottom air supply location compared with Group Two, and would express the system cooing and ventilation differences between working conditions No. 5-8 and working conditions No. 9-12. Group Four considers working conditions No. 13-15 and mainly discusses the solar radiation effects on the system wind-thermal environmental performance. Group Five to Group Seven mainly considers the influences of solar chimney on cooling and ventilation effects of the composite enhanced natural ventilation system. Group Five considers working conditions No. 16-18 and mainly discusses the solar chimney height effects on the system wind-thermal environmental performance. Group Six considers working conditions No. 19-21 and mainly discusses the solar chimney width effects on the system wind-thermal environmental performance. Group Seven considers working conditions No. 22-24 and mainly discusses the Absorber plate angle effects on the system wind-thermal environmental performance. For working conditions No.13-24, working conditions No.7 is the control group.

Through comparison analysis with multiple research groups, the designed system wind-thermal environmental performance would be discussed under different design parameters, and the SC-EAHES integrated composite enhanced natural ventilation system with better wind-thermal environmental characteristic would be expressed.

1.3 System Wind-Thermal Environmental Performance Evaluation Index

In order to express the air cooling effects and system ventilation effects of the composite enhanced natural ventilation system under different design condition, the wind-thermal environmental performance aspects are considered.

First, the ventilation quantity (V_q) is a major aspect for displaying the ventilation effects of the composite enhanced natural ventilation system. Second, the distributions of air velocity (V_a) field and air temperature (T_a) field for the auditorium area are important representations for expressing the cooling and ventilation effects of the designed system. In addition, this displacement ventilation system would usually cause bigger human vertical temperature gradient difference (T_gd) which would obviously affect human thermal comfort. According to the related design standard for heating, ventilation and air conditioning, the first-level thermal comfort corresponds to the PPD (Predicted Percentage of Dissatisfied) less than 10%, and research results show that the PPD under 10% usually corresponds to the limited human vertical temperature gradient difference under 4 ℃/m^[32-33]. Therefore, the vertical temperature difference between head and foot in the auditorium area is also a necessary index from the view of human thermal comfort.

Therefore, three aspects of indexes are applied for evaluating the comprehensive performance of this designed passive composite enhanced natural ventilation system, as listed in Table 2.

2 Numerical Simulation Methods

A series of system design factors affecting system wind-thermal environmental performance in research groups is discussed. Therefore, numerical simulation methods are used to analyze and express the comprehensive system wind-thermal environmental performance under different working conditions. Considering the computational accuracy and model complexity, CFD simulation technology is applied. Aiming at the simulation requirements, the simulation theoretical basis, model hypothesis, boundary conditions, solution method, mesh generation and grid independence test, and reliability verification are respectively illustrated as below.

2.1 Model Hypothesis

In order to simplify calculation during the simulation process, several simplification and hypothesis are made: 1) The calculation condition is steady-state, including steady-state outdoor air temperature, solar radiation intensity, tunnel temperature and human heat dissipation; 2) The simulation medium (air) is considered as incompressible Newton fluid; 3) The air flow state is turbulent; 4) The air satisfies Boussinesq hypothesis; 5) Wall heat storage is not considered. For the mathematical models, this paper adopted the realizable K-ε(K is Turbulent kinetic energy, ε is turbulent energy dissipation)turbulence model applied for this simulation process which was discussed by Liu et al.^[34]that the realizable K-ε turbulence model was in a good agreement with the measured data for natural ventilation in high and large space buildings. The DO radiation model in Fluent is chosen for this study because the DO radiation model could better deal with the local heat source. This study assumes that all surface areas of the solar chimney are imposed with solar radiation. For all surface areas of the solar chimney, the solar radiation intensity is set as a constant value of 600 W/m² according to the annual solar radiation amount in Hainan during summertime. Then according to the average daily temperature during the daytime in Hainan, the calculated outdoor air temperature is set as 32.2 ℃ (305.35 K) and the relative air thermal physical parameters are determined, as shown in Table 3.

According to the simulation conditions listed in Table 3, ANSYS Fluent (v16.1), which is widely used for solving complex flow problems, is adopted to simulate the airflow and temperature distributions for the study object due to its strong abilities for processing structured and unstructured mesh^[35-37]. Therefore, the heat transfer and flow problems for the whole composite enhanced natural ventilation system would be expressed by ANSYS Fluent (v16.1). The output results would evaluate the system wind-thermal environmental performance under different working conditions. This research firstly simplified the physical model for this designed system. All the surfaces are considered as zero-thickness, and the audience are filled up with people sitting still in the seats. Each person is expressed with an equivalent cuboid of 0.4 m×0.4 m×0.9 m. Fig. 4 shows a 3-D sample simulation case with double underground tunnel and side air supply, including 3-D computational domain and three orthographic views from different angles.

2.2 Boundary Conditions

For the system wall surfaces, calculation results show that the total cooling load of wall surfaces is 20.69 kW and that of the roof is 40.37 kW. This study did not consider the cooling load difference caused by orientations, and therefore the wall's heat flow density was set 7.2 W/m² with the total wall heating area of 2871 m² while the roof's heat flow density was sustained as 47.4 W/m² with the total roof heating area of 851 m². Considering that the soil temperature at 6 m below underground in Hainan could stay at 290.15 K, the underground tunnel wall temperature was assumed to stay constant at a fixed temperature. Then the solar chimney was set in the form of heat flux density which relates to solar radiation intensity.

Since the system air supply and system air outlet is directly connected with the outside environment, the air supply and outlet applied the pressure boundary conditions (pressure inlet for underground tunnel air supply and pressure outlet for solar chimney air outlet). The turbulence intensity is calculated as follows^[38]:

$ I=u^{\prime} / \bar{u}=0.16\left(\operatorname{Re}_{D_H}\right)^{-1 / 8} $

(1)

where I is turbulence intensity, u′ is turbulence pulsation velocity, m/s; $ \bar{u}$ is turbulence average velocity, m/s; Re_DH is Reynolds number corresponding to hydraulic diameter.

The hydraulic diameter can be calculated by system air supply size and the system inlet air temperature was consistent with the outdoor air temperature. For the internal heat source of this system, this study considers the heat dissipation of human body as the main internal heat source. The personnel heating load was set 56 W/person and the human heat flux density was set 31.82 W/m² with the human body surface area of 1.76 m².

2.3 Solution Method

In this study, the simulation process is considered as steady-state. After comprehensively taking the computational expense into account, the SIMPLE(Semi-Implicit Method for Pressure Linked Equations) algorithm was applied to calculate related physical variables for this simulation problem. The strength of the SIMPLE method is that together with implicit time treatment of the flow variables, a steady state solution can be efficiently obtained or rather large time steps for unsteady flow computations are used. To specify the desired value of the convergence of residual, simulations of different geometries were done, and the convergence history was checked for each one-order decrease of residuals from 1×10^-3 to 1×10^-7. Results found that 1×10^-6 residual is suitable for energy equation and radiation equation while 1×10^-3 is adequate for the other equations to indicate the solution convergence. Then after multiple simulation experiments, the relaxation factors in SIMPLE algorithm were determined (Table 4), which could ensure correct calculation results and make the equations converge quickly.

To ensure the simulation accuracy, this study employed the second-order upwind discrete scheme with second-order precision for discretizing the momentum equation, energy equation, K equation, ε equation and radiation equation. For the pressure equation, the Body Force Weighted discrete scheme was adopted due to the buoyancy-driven flow feature.

2.4 Mesh Generation and Grid Independence Test

To create a non-uniform structured mesh generation for the computational domain, ICEM software was employed due to its powerful grid generation ability and higher grid quality^[39-40]. The mesh type is Tetra/Mixed. Since the system working area, underground tunnel air supply and solar chimney air outlet are main areas that require further attention, the corresponding mesh density for these areas is increased.

In order to avoid the influence of mesh number on simulation results, five mesh elements are considered, which are 2959186, 3584694, 4820342, 5574533 and 6351568. According to the meteorological data information in Hainan, the solar radiation amount is set 600 W/m². Under the solar radiation amount of 600 W/m², the system ventilation quantity is calculated with different mesh elements. Results show that the system ventilation declines with the increase of mesh number. When the mesh number reached 4820342, the ventilation quantity almost did not obviously decline. Therefore, the mesh element with 4820342 is utilized as it shows the best number of elements with respect to both accuracy and computation time.

2.5 Reliability Verification

Several researchers have investigated the solar chimney ventilation effects with actual field experimental measurements. This study mainly focuses on the simulation reliability and accuracy of the designed natural ventilation system. Therefore, a comparison verification is conducted according to a field case experiment. Li et al.^[41] used the devices (Fig. 7) to investigate the ventilation performance of solar energy chimney.

As shown in Fig. 5, the chimney walls of the device are high-density slabs with small thermal conductivity coefficient. The right wall is the movable wall ranging from 400-1200 mm, and the left and the rear walls are Trombe walls, the heat gain of which is expressed by calorific values of the electrothermal film attached evenly on the chimney walls. During the experiments, different solar radiation intensities (200 W/m²-400 W/m²) are simulated by modulating transformer voltage to alter calorific values of electrothermal film. Hot bulb anemoscopes are used to test the chimney internal air velocity. Then the solar chimney ventilation quantity can be calculated with the formula:

$ M=\bar{v} \times A \times \rho_{\text {air }} $

(2)

where M is Natural ventilation airflow rates, kg/s; $\bar{v} $ is Mean velocity, m/s; A is Cross section area of chimney, m²; ρ_air is air density, kg/m³.

In order to verify the simulation reliability of this study, the influence of different solar radiation intensities on the ventilation performance of solar chimney is simulated by ANSYS Fluent (v16.1) with the same size of chimney channel (700 mm) as Ref. [41]. Fig. 6. show that relative error rates between the field experiments in Ref. [41] and the CFD simulations are 13.12%, 11.72% and 12.86% with the respective solar radiation heat flux of 200 W/m², 300 W/m² and 400 W/m².

All of relative errors are less than 15% and the relative error variation trends stays the same. Therefore, the simulation methods in this study are effective and reliable overall.

3 Results and Discussion 3.1 System Ventilation Quantity under Different System Design Conditions

In the relevant working conditions for underground tunnels, the simulation results show that in Groups One to Three, V_q will decrease with the increase of the length of the underground tunnel. Among them, Group 2 has the largest decrease, from 16.65 to 12 kg/s. Group One dropped the least from 16.75 to 13.75 kg/s. In general, V_q of the double underground tunnel is similar to that of the single underground tunnel, while V_q under the bottom air supply method is larger than that of the side air supply. The main reason is that the air supply direction at the bottom is the same as that at the top. For side air supply, there is a 90° difference between the air supply direction and the air outlet direction at the top. Compared with bottom air supply method, the side air supply method could provide an additional bending, thus increasing the system total resistance and declining the system ventilation amount.

In the relevant working conditions for solar chimneys, the simulation results show that in Groups Four to Seven, V_q will increase through increasing the solar intensity, enlarging the height and width of solar chimney and also the inclination of the solar panel. Among them, Group Five displays the largest increase, increasing from 9.73 to 16.21 kg/s. Group Seven has the least increase varying from 13.64 to 14.16 kg/s with a limited increase of 0.52 kg/s. Group Six and Group Four were followed by an increase of 5.12 kg/s and 2.72 kg/s, respectively.

3.2 Wind-Thermal Environmental Characteristic with Different Underground Tunnel Design Conditions

Both considering the inlet's air supply effect and for the convenience of maintenance personnel, the underground tunnel size was set as 2 m high and 6 m wide. After preliminary simulation experiments, results show that when the underground length is shorter than 80 m, the auditorium back-row seats present higher air temperature values which weakens research significance, while excessively increasing the underground tunnel length would then increase the system investment cost. Therefore, simulation experiments Group One designed four underground tunnel lengths varying from 80 to 140 m (Table 2). For Group One, there is single underground tunnel with bottom air supply, solar radiation intensity is 600 W/m², the solar chimney is 20 m high, 6 m long and 1 m wide with 0.6° absorber plate. After ANSYS Fluent's simulations for Group One, the system wind-thermal environmental performance indexes are analyzed.

In Group One, the distributions of V_a field for the auditorium area are expressed as below. Fig. 7(a) shows the local velocity distributions for the auditorium area of working conditions from No. 1 to No. 4 at cross section z=0 m. Results show that the cold air which has been cooled by underground tunnel-based EAHES enters the plenum chamber through the side air supply. V_a gradually decreases to below 0.2 m/s. Among these working conditions with different underground tunnel lengths, working conditions No. 1 and No. 2 present larger velocity variation gradient while working conditions No. 3 and No. 4 express a relatively uniform velocity variation.

Then T_a distributions for the auditorium area of working conditions from No. 1 to No. 4 at cross section z=0 m are presented in Fig. 7(b). Results show that with the increase of underground tunnel lengths, T_a in plenum chamber gradually declines. The maximum temperature values of the auditorium area are 303.33, 302.92, 302.5 and 297.92 K respectively for working conditions No. 1 to No. 4. Among them, working condition No. 1 and No. 2 present higher temperature values for the auditorium back-row seats. Then after statistical calculation, the maximum head-foot temperature differences are obtained respectively 2.5, 3.4, 3.31 and 3.4 ℃ for these four working conditions No. 1 to No. 4, which shows that the head- foot temperature difference increases first, then stabilizes with the increase of underground tunnel lengths.

Considering the system wind-thermal environmental performance in Group One, with the increase of underground tunnel length, the air temperature in plenum chamber gradually decreases. Specifically, working conditions No. 1 and No. 2 generate larger air velocity variation gradient while working conditions No. 3 and No. 4 express a relatively uniform velocity variation and better cooling effects for the auditorium area. The maximum head-foot temperature difference first increases and then stabilizes, which generally accord with the first-level thermal comfort.

Since the results in Group One with single underground tunnel display several disadvantages, such as the higher local auditorium temperatures in back-row seats and larger velocity gradient in plenum chamber, this study considers adding more underground tunnels and adopts double underground tunnel system to analyze the system wind-thermal environmental performance. In this way, the simulation experiments were conducted according to Group Two which was set with double underground tunnels while the other design parameters are kept the same with Group One.

In Group Two, the distributions of V_a field for the auditorium area are expressed in Fig. 8. Since the cross-section z=8 m shows the greatest V_a disturbance, cross section z=8 m was selected to analyze the double underground-tunnel system wind-thermal environmental performance. As shown in Fig. 8(a), V_a for the middle-row seats is higher than V_a for the front-row seats and the back-row seats in the auditorium area. As the underground tunnel length increases, V_a in the outlet under the middle audience seats decays to 0.22 m/s. The working conditions from No. 5 to No. 8 present lower V_a compared with Group One. V_a entering the plenum chamber in Group Two is about half those in Group One, which avoids excessive V_a. Among the cases in Group Two, working condition No. 8 with 120 m long underground tunnel shows better uniformity compared with others.

T_a distributions for the auditorium area in Group Two at cross section z=8 m are shown in Fig. 8(b). With the increase of underground tunnel lengths, T_a in plenum chamber is respectively 298.75, 297.92, 296.67, 295.42 K, which are about 2 K lower than those in Group One. The maximum head-foot temperature difference for working conditions No. 5 to No. 8 are 2.5, 2.5, 3.3, 4.17 ℃. Among them, working conditions No. 5 to No. 7 show smaller head-foot temperature difference compared with Group One while working condition No. 8 has larger head-foot temperature difference which may exceed first-level thermal comfort. However, the 140 m long underground tunnel presents excessive head-foot temperature difference which would influence human thermal comfort in auditorium area.

In the underground tunnel-based EAHES, the bottom air supply location of plenum chamber is also a widely used method, as shown in Fig. 9. Therefore, this section mainly analyzes the influence of different plenum chamber's air supply locations on system wind-thermal environmental performance. According to Group Three in Table 1, simulations with bottom air supply locations are conducted and the system wind-thermal environmental performances for Group Three are obtained. Since Group Two adopted the side air supply location while Group Three applied the bottom air supply location, the other simulation conditions stay the same with each other. Therefore, results in Group Two are considered as control group in this section.

On Group Three, V_a distributions for the auditorium area in Group Three are shown in Fig. 10(a). Cross section z =8 m of working conditions from No. 9 to No. 12 was selected to express the bottom air supply location's system wind-thermal environmental performance. As shown in Fig. 17, working conditions No. 9 to No. 12 all present larger V_a around air supply inlets and caused great velocity nonuniformity in the plenum chamber. Specifically, working conditions No. 9 and No. 10 show larger V_a for the back-row seats in the auditorium area, while with the increase of tunnel lengths, the air supply nonuniformity was relieved. Although the ventilation volume of the bottom air supply method is larger than that of the side air supply method, its air supply has great heterogeneity. This is because the air supply outlet of the bottom air supply method is arranged below the rear part of the theater, and its air supply direction is consistent with that of the upper air outlet. However, compared with the larger theater, the air supply outlet arranged behind the theater is relatively small, which cannot meet the uniformity of the whole large space of the theater. Therefore, compared with Group Two, the bottom air supply system in Group Three shows obvious V_a nonuniformity and would influence the system air supply effects.

T_a distributions for the auditorium area in Group Three at cross section z =8 m are shown in Fig. 10(b). With the increase of underground tunnel lengths, T_a in plenum chamber is respectively 299.17, 298.33, 297.08, 295.83 K, which demonstrates about 0.4 K higher than those in Group Two. In addition, the whole auditorium under working conditions No. 9 and No. 10 present larger temperature difference between front rows and back rows, in which the air temperatures of back rows all show higher values. Then the maximum head-foot temperature difference for working conditions No. 9 to No. 12 are 2.5, 2.92, 3.33 and 4.17 ℃ respectively, the values of which are generally similar with those in Group Two. Among them, working condition No. 12 also displays larger head-foot temperature difference exceeding first-level thermal comfort. Taken on the whole, considering the cooling effects and better air supply uniformity, Group Two demonstrates better performance and shows its priority to Group Three.

As discussed above, the composite enhanced natural ventilation system shows different system wind-thermal environmental performances due to their different design parameters, namely the 80/100/120/140 m long underground tunnel length, single/double underground tunnel, side/bottom plenum chamber's air supply location. The contrast analysis between Group One, Two and Three shows that Group One and Group Three both present larger V_q than Group Two. But considering the cooling effects and air supply uniformity, Group Two shows its great advantages expressed by lower air supply temperature and smaller V_a gradient compared with Group One and Group Three. Then considering the three group experiments, it can be concluded that compared with the working conditions with 80 m and 100 m long underground tunnel lengths, the working conditions with 120 m and 140 m underground tunnel lengths could generate lower air supply temperatures and relatively uniform velocity variations. Therefore, the increase of underground tunnel length would help enhance the system cooling effects and provide better air supply distributions. However, the maximum head-foot temperature difference of Group Two and Three shows that the 140 m long underground tunnel length cannot satisfy the first-level thermal comfort, which has exceeded the limited gradient value of 4 ℃. Therefore, by comprehensively considering the system wind-thermal environmental performance index and investment cost, working condition No. 7 in Group Two expresses its better comprehensive performance compared with other working conditions in Group One to Three and its T_gd also accords with the first-level thermal comfort degree.

3.3 Transmission Characteristic Analysis under Different Solar Chimney Design Conditions

Solar radiation is the power source of the composite enhanced natural ventilation system. The solar radiation intensity varies with seasons, times of a day and weather conditions. Therefore, this section mainly analyzes the influence of different solar radiation intensity on system wind-thermal environmental performance. For the comparison analysis, working condition No.7 was set as reference control condition, and working conditions No.13 to No.15 consider different solar radiation intensities varying from 200 to 800 W/m².

In Group Four, V_a distributions for the auditorium area at cross section z =8 m are presented in Fig. 20. As shown in Fig. 11(a), when the solar radiation intensity is 200 W/m², V_a gradually decays from 0.44 to 0 m/s in plenum chamber. Among the conditions shown in Fig. 20, working condition No. 14 presents lower V_a for the seats in the auditorium area and also shows good uniformity. However, as the solar radiation intensity increases, the increase of V_q led to the increase of V_a and a slower decline occurs in x direction.

T_a distributions for the auditorium area at cross section z =8 m are expressed in Fig. 11(b). T_a in plenum chamber are 295.42, 296.25, 296.67 and 297.08 K, respectively. When the solar radiation intensity is 200 W/m², T_a distribution in plenum chamber is obviously stratified. In plenum chamber, T_a of the upper part is higher than that of the bottom. But as the solar radiation intensity gradually increases to 400 W/m² and above, stratification of V_a is almost negligible. Then after statistical calculation, the maximum head-foot temperature differences 4.59 ℃, 3.75 ℃, 3.34 ℃ and 3.08 ℃ are obtained respectively for working conditions No. 13, No. 14, No. 7 and No. 15, which shows that as the solar radiation increases, the head-foot temperature difference decreases, which improves the thermal comfort of audiences. As the solar radiation increases, the quantity of air taking more heat increases and the supply air temperature is higher. The effects of both aspects decrease the head-foot temperature difference.

As shown in Figs. 7-8 and Fig. 10, the solar radiation intensity contributes to larger difference in transmission power and further leads to the difference in cooling effects. But the outdoor background solar radiation cannot be artificially controlled. To optimize the system ventilation effects, the structural parameters of solar chimney are considered from three aspects of solar chimney height, solar chimney width and angle of absorber plate. Three research groups from Group Five to Group Seven are designed for the system wind-thermal environmental performance analysis under different solar chimney parameters.

For Group Five to Group Seven, the underground tunnel length is set as 120 m with double tunnels and side air supply location. Each group discusses one single parameter variation of the solar chimney, and working condition No.7 was set as reference control condition. Since the change of solar chimney parameters would mainly influence the air distributions and characteristics inside the solar chimney, the local air velocity and air temperature distributions inside the solar chimney are discussed.

In Group Five, V_a distributions and T_a distributions inside the solar chimney are presented in Fig. 12. V_a of the upper part is slower than that of the bottom and T_a of the upper part is higher than that of the bottom. Resistance of the boundary layer is large, so V_a of the boundary layer is smaller than that of the middle. T_a of the boundary layer is stable and higher than that of the middle. As the solar chimney height increases from 10 m to 20 m, V_a inside does not increase much. But as the solar chimney height increases from 20 m to 25 m, V_a of the bottom increases to 2.5 m/s and above. As the solar chimney height increases, T_a inside increases and the temperature difference between inlet air and outlet air also intensifies, then the thermal pressure is enlarged to induce ventilation.

In Group Six, V_a distributions and T_a distribution for the solar chimney are presented in Fig. 13. Compared with Group Five, the average T_a of Group Six decreases from 316.45 to 302.94 K and the maximum temperature difference between inlet and outlet decreases as the solar chimney width increases. Therefore, the thermal pressure is reduced and V_a declines from 2.69 m/s to 1.66 m/s with the increase of the solar chimney width. Although the area of the solar chimney increases and the solar chimney absorbs more solar radiation, T_a in the middle maintains a low temperature, which is about 303.15 K.

Integrating the results in Group Five and Six, it is acknowledged that whether the solar chimney height or width increases with the other variables unchanged, the system ventilation effects are improved. Several existing articles have shown that the inner air velocity could be strengthened by increasing the ratio of height to width of solar chimney^[42-43], which may not totally accord with the simulation results in this study. The reasons may be due to the high-large space of the target buildings. The whole ventilation is caused not only by the SC effects but also the underground tunnel cooling effects. Therefore, the increase of total inner space of target building helps strengthen the whole system ventilation.

For Group Seven, pre-simulations for more conditions with absorber plate angles varying from 0° to 30° of the solar chimney are conducted. Results show that as the plate angles increase from 0.2° to 0.8°, the system wind-thermal environmental performance presents obvious variations while the other plate angles did not show effective influences on the system wind-thermal environmental performance. Considering the cost of the solar chimney, the plate angles varying from 0.2° to 0.8° are chosen as working conditions, as shown in Group Seven. Working condition No.7 is still set reference condition to conduct comparison analysis. V_a distributions and T_a distributions for the solar chimney are presented in Fig. 14. As the width of solar chimney increases from 0.2° to 0.8°, T_a and the temperature difference between inlet and outlet do not intensify much. As the absorber plate angles increases, V_a has no regular change trend. Because there is a backflow that hinders the airflow, this phenomenon may not totally accord with the simulated results in Refs. [44-46] which emphasized the effective effects of increasing absorber plate angles on ventilation, which may be account of the differences in the size of the study objects.

4 Conclusions

According to the structure features of high and large space buildings, the solar chimney (SC) approach can be applied to serve as power equipment transporting cold air in air-conditioning systems, and the earth to air heat exchange system (EAHES) could be then used as cold source to realize the enhancement of the natural ventilation effects. Taking a theatre in Hainan Province of China as study object, a newly composite enhanced natural ventilation system was proposed. In this system, the underground tunnel-based EAHES serves as the natural cold source and provides the theatre's displacement ventilation. The roof-top solar chimney then provides transportation power for natural ventilation by making good use of the stronger solar radiation in Hainan to realize SC effects.

Research on seven simulation groups concerning the underground tunnel length, underground tunnel number, air supply location, solar radiation intensity, solar chimney height and width, and absorber plate angle of the proposed system respectively are conducted by using ANSYS Fluent (v16.1). For the underground tunnel under the same solar chimney design conditions, results show that the system ventilation quantity (V_q) declines with the increase of underground tunnel length. Groups One and Three both present larger V_q than Group Two, but Group Two shows its great advantages in the cooling effects and air supply uniformity.

Then for the solar chimney under the same underground tunnel design conditions, results show that with the increase of solar radiation intensity, V_q grows while the head-foot temperature difference (T_gd) decreases. Since the solar radiation would greatly influence the system transmission power, Groups Five to Seven concerning the solar chimney structure are discussed. Results show that the solar chimney factors affect V_a distributions and T_a distributions inside the solar chimney, and then cause the difference in V_q. Increasing the height, width, and absorber plate angle of solar chimney would all contribute to larger V_q, but the increasing rate of V_q increases for Group Five while decreases for Group Six and Group Seven. Although the transmission power of solar chimney could be increased by enhancing the solar chimney sizes and absorber plate angle, the construction cost should be also considered.

From an overall perspective, taking the comprehensive system wind-thermal environmental performances of V_q, V_a and T_a field uniformity, human vertical thermal comfort index T_gd and the construction cost into account, the system under working condition No. 7 with 120 m long, side air supply, double underground tunnel and 20 m high, 1 m wide, 0.6° absorber plate angle solar chimney shows its priority in better comprehensive performance. This paper originates from the concept of nearly zero energy building design, and mainly contributes to the energy saving of air conditioning in large space buildings in hot-humid tropical climate zones. The discussion for multiple research groups could provide theoretical basis and reference for the thermal comfortable air conditioning system design of high and large space buildings. In the near future, more detailed theoretical analysis of each factor would be discussed by proposing theoretical hypotheses and verifying them with appropriate field experiments.