Analysis of the Effect of Exhaust Configuration and Shape Parameters of Ventilation Windows on Microclimate in Round Arch Solar Greenhouse

Abstract

The round-arch solar greenhouse (RASG) is widely used in the alpine and high latitude areas of China for its excellent performance. Common high temperature and high humidity environments have adverse effects on plants. It is extremely important to explore a reasonable and efficient ventilation system. A three-dimensional numerical simulation model of greenhouse ventilation considering crop canopy airflow disturbance was established. A robust statistical analysis to determine the validity of the model was calculated to thoroughly validate its overall performance. Microclimate distribution characteristics of nine kinds of exhaust configuration in greenhouse in summer were analyzed comparatively. It was determined that the highest ventilation efficiency could be achieved by adopting the combined configuration of rolling film at the south corner of the greenhouse and pivoting the window at the north side of the roof. In winter, the opening angle of ventilation window at the north side of the roof was less than 40° to ensure the rapid cooling of the interior of the greenhouse without the crops being affected by the cold environment. Through optimization analysis, the ventilation configuration with a deviation angle of 25° and a width of 900 mm is more reasonable (10 m span). The research results provide theoretical guidance for the design of the ventilation structure in RASG and further improve the sustainable development of the facility’s plant production.

1. Introduction

Chinese solar greenhouses are widely used in northern China as energy efficient facilities (as shown in Figure 1). Currently China’s energy-efficient solar greenhouses have been able to produce winter tomatoes, cucumbers, and other thermophilic vegetables under the condition of no warming measures, relying only on solar radiation [1,2,3]. The high production efficiency and annual intensive production of solar greenhouses have improved the land utilization rate, greatly increasing the production per unit area, effectively alleviating the problem of difficult supply of vegetables in the off-season in China, especially in the north, and greatly increasing the economic income of farmers [4,5,6]. Traditional solar greenhouse structures are mostly brick-walled or soil-walled thermal storage structures. The interior of the assembled round-arch solar greenhouse (RASG) is equipped with a better water circulation and heat storage and release system than brick walls. The unique semi-circular assembled skeleton structure and the heat collection device give RASG the advantages of better lighting conditions, low construction costs, and short cycle times [7,8,9].

Greenhouse agriculture aims to ensure the normal growth of plants at different stages. The regulation of greenhouse microclimate is important for efficient production, and ventilation is the key measure for greenhouse environmental regulation [10,11,12,13]. Ventilation not only safeguards the respiration of crops in the greenhouse but also regulates the air temperature and humidity in the greenhouse as well as other environmental factors that are conducive to crop growth. Summer heat is frequent, and ventilation of greenhouses is necessary for greenhouse crop production. The high temperature in the greenhouse at midday in winter is also detrimental to the crop and proper ventilation can play a role in cooling and dehumidifying the greenhouse, thereby reducing the occurrence of crop mycosis. At present, the main measure of microclimate control in solar greenhouses in production is natural ventilation. Opening the vents of the greenhouse provides convective heat exchange through the action of buoyancy and wind pressure [14,15,16]. Natural ventilation is easy to operate, consumes less energy, and can meet the growing needs of crops in the greenhouse. Reasonable ventilation measures can not only regulate the temperature and humidity distribution in the greenhouse but also maximize the use of solar energy. However, traditional regulation relies heavily on the experience of agricultural producers, which often results in wasted resources.

Extensive research has been conducted by research scholars on the effect of natural ventilation in greenhouses, and certain results have been achieved. In the past, the tracer gas method and energy balance method were mainly used to study the microclimate characteristics of greenhouse ventilation [17,18,19]. Scholars have conducted a large number of experiments on the different effects of ventilation methods using the above methods. They analyzed the feedback effect of crop growth pattern and crop canopy structure on air flow distribution in a greenhouse [20,21,22]. However, the tracer gas method is influenced by changes in experimental conditions and the environment. The sensitivity of gas flow fluctuations can reduce the stability of experimental results. The approach of energy balance is based on treating the air inside the greenhouse as a whole and assuming that the air is mixed evenly. It also cannot reflect the spatial distribution of temperature and humidity.

With the development of science and technology, the computational fluid dynamics (CFD) method has been found to be an effective and stable technical tool in the study of greenhouse ventilation and is becoming more and more widely used around the world [23,24,25,26,27]. The CFD method takes the three conservation equations of fluid mass, momentum, and energy as the theoretical basis for numerical calculations. Combined with the fluid turbulence model, it can simulate and predict the airflow pattern and the spatial distribution of temperature and humidity, CO₂ concentration, and other factors in the greenhouse in two or three dimensions. Many scholars have analyzed the effects of natural ventilation on air flow and heat distribution in passive greenhouses located in tropical and subtropical countries by numerical simulation [28,29,30,31]. Park and Lee et al. [32,33] used CFD method to simulate the natural ventilation of the greenhouse in the coastal area of South Korea, and evaluated the ventilation efficiency and ventilation volume. Santolini et al. [34] used the CFD method to study the wind-driven ventilation in Italian greenhouses, focusing on the influence of the net on the airflow distribution in the greenhouse. Akrami et al. [35] used CFD to analyze the ventilation requirements of different vent opening scenarios in Egyptian greenhouses, showing the importance of the entrance location to the proposed sustainable greenhouse system. Saberian et al. [36] established a computational fluid dynamics model of a full-scale gable greenhouse in Iran. The variation of solar heat load and the influence of the regional dominant wind profile on the heat transfer coefficient of the wall surface are both predicted. There are also a large number of research reports on greenhouse ventilation in different regions of China [37,38,39,40]. Zhang et al. [41] studied the relationship between different vent openings and air humidity changes in Chinese solar greenhouses through CFD simulation. The simulation results show that with the increase in greenhouse ventilation openings, natural ventilation increases, resulting in a decrease in indoor air humidity and temperature. He et al. [42] successfully designed a three-dimensional CFD model based on an 11-span plastic greenhouse. The model was verified by the measured air temperature and relative humidity, and the influence of vent structure and opening size on greenhouse climate was studied.

RASGs are given priority for construction due to their good performance characteristics. The effectiveness of ventilation in RASG applied in the high latitudes of northern China has been investigated. However, the climatic characteristics of summer and winter, which differ greatly, have not been taken into account when ventilation configurations are investigated. The optimal ventilation strategy for RASGs in northern China is not clear. The influence of exhaust configuration and shape parameters of the greenhouse ventilation window on the indoor airflow pattern and distribution is not clear. In the actual construction of the greenhouse construction process, workers rely mostly on experience to design and build greenhouse ventilation window configuration and shape size. The optimum size and location parameters for ventilation configuration are not yet clear and there is a lack of uniform construction standards. The purpose of this study is to investigate the construction standards and structural parameters (i.e., location and dimensions) of a ventilated configuration for RASG suitable for northern China, systematically revealing the effect of different combinations of ventilation configuration on temperature and humidity as well as airflow patterns in the greenhouse. It provides theoretical guidance for further improving the sustainability of facility plant production.

2. Materials and Methods

2.1. Technical Route of the Research Program

In this study, a combination of experimental and simulation techniques is used to investigate the optimization strategy of the ventilation configuration of a round-arch solar greenhouse. The technical route of the research program is shown in Figure 2. The three dimensions of the experimental control greenhouse are measured to facilitate accurate physical CFD model construction. The microenvironment inside and outside the greenhouse is continuously monitored by indoor environmental sensors and outdoor weather stations. On the one hand, it is used for the boundary condition input of the simulation model to improve the model accuracy, and, on the other hand, it is used for the comparison and verification of the model calculation results and the actual measurement results. The validated model simulates different ventilation conditions to optimize the ventilation configuration. Finally, based on the simulation and optimization results, the actual construction of a round-arch solar greenhouse with an optimal ventilation configuration is carried out to compare the ventilation effect.

2.2. Experimental Site and Measuring Procedures

The experimental procedures are carried out in a round-arch solar greenhouse belonging to the Laboratory of Horticultural Facility Design and Environmental Control at Shenyang Agricultural University, China (41.8° N, 123.4° E). The greenhouse azimuth is 7° west-south. The lighting roof is facing the south, and the west angle enables the greenhouse to receive more solar radiation. The experimental greenhouse adopts steel structure material to bear the load, and the skeleton spacing is uniform to ensure uniform stress. The greenhouse has a semi-circular cross-sectional shape, which provides a large space and good light. The length of the greenhouse is 60 m, the span is 10 m, and the ridge height is 5 m. The southern roof of the greenhouse is a double-layer structure. The vents are arranged on the ground side of the south roof. The vents are 0.7 m high from the ground and the two adjacent vents are at the same distance. The ventilation window baffle is driven by the rotating shaft to open and close. The top has a flip-up window ventilation configuration with an angle of 10° to the center line of the greenhouse. During the day, only a layer of polyethylene film on the south roof skeleton ensures the reception of solar radiation. At night, as the outdoor temperature drops, there is an urgent need to cover with another layer of insulation to reduce the heat loss through long-wave radiation. Since the traditional brick wall structure is eliminated, the solar water circulation heat storage-release system is adopted to realize the heat storage and the heat preservation of the greenhouse. In the greenhouse, 60 rows of tomato plants were planted at an interval of 1 m along the north-south direction. The length of a single row is 8 m, with 18–22 plants and an average height of 1.2 m. The greenhouse microclimate measurement experiments are carried out under typical sunny weather conditions in summer and winter, respectively. Figure 3 shows the display of the experimental greenhouse.

The environmental parameters of the solar greenhouse include: indoor and outdoor air temperature and humidity, the temperature and humidity around the leaves, the external wind speed and direction, the breeze speed between the canopy, the saturated vapor pressure of air and canopy, and the indoor and outdoor solar radiation intensity. An outdoor weather station is installed at a straight distance of 5.5 m from the greenhouse in the southeast direction. A quarter of the greenhouse length is used as the basis for the selection of equidistant environmental monitoring sections, and data sensors are arranged in all sections. There are 51 ELITECH temperature and humidity automatic recorders (RC-4HC type, temperature range: −40 °C to 85 °C, temperature precision: ±0.5 °C, temperature resolution: 0.1 °C; humidity range: 0% to 100% RH, humidity precision: ±3% RH, humidity resolution: ±3% RH) made in China that are placed at different heights from the ground. In each environmental monitoring section, the solar radiation intensity is measured by a DELTA OHM solar radiation automatic recorder (Pyranometer-HD2302.0, range: 0.01 W/m² to 2 × 10⁵ W/m², resolution: 0.01 W/m²) made in Italy. The BENETECH anemometer (GM8901 type, range: 0.30–45 m/s, precision: 0.01 m/s, accuracy: ±3%) made in China is used to measure the actual wind speed at a distance of 0.5 m from the bottom and top vents in the three environmental monitoring sections. All instrument probes are shielded from solar radiation to avoid the measurement deviation caused by the direct radiation of solar radiation.

In the crop section, microclimate monitoring points are arranged at equal intervals along the north-south direction, 0.7 m above the ground within the crop canopy. The temperature and humidity of the plant leaves are measured with the temperature and humidity automatic recorders. The atmospheric pressure around the plants is measured with the BENETECH digital barometer (GM510 type, range: ±10.00 kpa, accuracy: ±0.3%) made in China. The breeze speed around the plants is measured with the BENETECH anemometer (GM8901). The leaf area index (LAI) is calculated by measuring the characteristic size of the leaves with a ruler.

Figure 4 shows the arrangement of several cross-sections of temperature and humidity environment monitoring points in the greenhouse. The measured data are collected in constant time steps of 30 s and the data curves are output via a computer connection. The collected indoor and outdoor environmental data is compiled into the model calculations using User-Defined Functions (UDF) as boundary conditions. In addition, the predicted root-mean-square division (PRMSD) of the numerical model is calculated from Equation (1) to thoroughly validate its overall performance [43].

$$PRMSD=\frac{1}{{\overline{D}}_m}\sqrt{\frac{{\displaystyle {\sum }_1^N{\left(D_m-D_s\right)}^2}}{N}}$$

(1)

Here $N$ is the number of environmental monitoring points, $D_m$ is the measured parameter, and $D_s$ is the simulated value. The length and width of the plants are measured using vernier calipers, the number of plants planted in the greenhouse are counted, the planting density is calculated, the leaf area of the planted crops is calculated, and the average of 10 calculations is taken as the result.

Figure 4. Experiment scene: (a) different cross-sectional locations of environmental measurement points to monitor microclimate changes; (b) experiment snapshots inside greenhouse.

Figure 4. Experiment scene: (a) different cross-sectional locations of environmental measurement points to monitor microclimate changes; (b) experiment snapshots inside greenhouse.

2.3. Modeling and Numerical Simulation

The special environmental phenomena of greenhouses require specific physical models for their simulation. The physical quantities that are originally continuous in the time and space domains are replaced by a set of variable values at a finite number of discrete points. Partial differential equations about the relationship between the field variables at these discrete points is established by certain principles, and finally the system of algebraic equations is solved by association to obtain an approximation of the desired variables [44].

2.3.1. Fundamental Equations and Mathematical Model

In this paper, the three-dimensional simulation of greenhouse natural ventilation is introduced in detail by using ANSYS 19.2 commercial CFD software. In addition, the numerical analysis of Geometry and mesh design uses Solidworks and Fluent Mesh software tools. The parameterized simulation has been performed in parallel on a workstation (Intel I9-10940X 3.3 GHz processors; 128 GB Fully DDR4 buffered memory). When the external climatic conditions change, the microclimate of the interior changes accordingly. CFD is a numerical method for solving the governing equations of fluid flow. By using the finite volume method, the system is simulated in discrete space and time and the unsteady three-dimensional turbulence conservation equations for the velocity and temperature fields are solved on an unstructured grid. The governing equations for fluid flow and heat transfer are based on three conservation laws [5]. Equations (2)–(6) represent the conservation equations for mass, momentum, and energy, respectively.

The Reynolds time-averaged equation method is introduced to time-average the transient Navier-Stokes equation, and the standard k-ɛ turbulence model with high simulation accuracy and good convergence is used to supplement the equations reflecting the turbulence characteristics to simulate and calculate the internal flow and temperature fields of the greenhouse [11].

$$\frac{\partial \rho }{\partial t}+\frac{\partial (\rho u_i)}{\partial x_i}=S_m$$

(2)

where $S_m$ is the mass added to the continuous phase.

$$\frac{\partial \left(\rho u\right)}{\partial t}+div\left(\rho uu\right)=div(\mu gradu)-\frac{\partial p}{\partial x}+S_u$$

(3)

$$\frac{\partial \left(\rho v\right)}{\partial t}+div\left(\rho vu\right)=div(\mu gradv)-\frac{\partial p}{\partial x}+S_v$$

(4)

$$\frac{\partial \left(\rho w\right)}{\partial t}+div\left(\rho wu\right)=div(\mu gradw)-\frac{\partial p}{\partial x}+S_w$$

(5)

$grad()=\frac{\partial ()}{\partial x}+\frac{\partial ()}{\partial y}+\frac{\partial ()}{\partial z}$; $S_u$, $S_v$ and $S_w$ are momentum generalized source terms.

$$\frac{\partial }{\partial t}\left(\rho E\right)+\frac{\partial }{\partial x_i}\left(u_i\left(\rho E+P\right)\right)=\frac{\partial }{\partial x_i}\left(k_{eff}\frac{\partial T}{\partial x_i}-{\displaystyle \sum _j{h}_j{J}_j+u_j{({\tau }_{ij})}_{eff}}\right)+S_h$$

(6)

Here, $k_{eff}$ is the effective transmission coefficient and h represents the apparent enthalpy of an ideal gas.

$$\frac{\partial }{\partial t}\left(\rho k\right)+\frac{\partial }{\partial x_i}\left(\rho k{u}_i\right)=\frac{\partial }{\partial x_j}\left[\left(\mu +\frac{{\mu }_t}{{\sigma }_k}\right)\frac{\partial k}{\partial x_j}\right]+G_k-\rho \epsilon$$

(7)

$$\frac{\partial }{\partial t}\left(\rho \epsilon \right)+\frac{\partial }{\partial x_i}\left(\rho \epsilon u_i\right)=\frac{\partial }{\partial x_j}\left[\left(\mu +\frac{{\mu }_t}{{\sigma }_{\epsilon }}\right)\frac{\partial \epsilon }{\partial x_j}\right]+C_{1\epsilon }\frac{\epsilon }{k}{G}_k-C_{2\epsilon }\rho \frac{{\epsilon }^2}{k}$$

(8)

Here, $k$ is the turbulence kinetic energy, m² s⁻²; $\epsilon$ is the dissipation rate; $u_i$ is the speed vector; ${\mu }_t$ is the eddy viscosity, Pa s; ${\sigma }_k$, and ${\sigma }_{\epsilon }$ is the Turbulent Prandtl numbers; $C_{1\epsilon }$ and $C_{2\epsilon }$ are the model courants; and $G_k$ is the generation of $k$ due to mean velocity gradients.

Since the humidity field of the solar greenhouse needs to be considered, the component transport equation needs to be added to the calculation equation to consider the gas in the greenhouse as a mixture of pure air and water vapor (wet air). The moisture content is derived by calculating the mass fraction of water vapor, which in turn gives the relative humidity of the air [45]. The component transmission equation is shown below:

$$\frac{\partial \left(\rho Y_i\right)}{\partial t}+\nabla \left(\rho v{Y}_i\right)=-\nabla J_i+R_i+S_i$$

(9)

Among them, R_i is the production of chemical reaction rate; S_i is defined as the source term and the additional rate of discrete phase; and G_i is the diffusion flux of the i th substance resulting from the concentration gradient.

To solve the humidity of wet air, we need to work out the enthalpy of wet air, which satisfies the energy conservation equation:

$$\frac{\partial \left(\rho h\right)}{\partial t}+\frac{\partial \left(\rho uh\right)}{\partial x}+\frac{\partial \left(\rho vh\right)}{\partial y}+\frac{\partial \left(\rho wh\right)}{\partial z}=\frac{\partial }{\partial x}\left({\Gamma }_h\frac{\partial h}{\partial x}\right)+\frac{\partial }{\partial y}\left({\Gamma }_h\frac{\partial h}{\partial y}\right)+\frac{\partial }{\partial z}\left({\Gamma }_h\frac{\partial h}{\partial z}\right)+S_h$$

(10)

Water vapor concentration a in wet air satisfies the mass conservation equation:

$$\frac{\partial \left(\rho Y_w\right)}{\partial t}+\frac{\partial \left(\rho u{Y}_w\right)}{\partial x}+\frac{\partial \left(\rho v{Y}_w\right)}{\partial y}+\frac{\partial \left(\rho w{Y}_w\right)}{\partial z}=\frac{\partial }{\partial x}\left({\Gamma }_y\frac{\partial Y_w}{\partial x}\right)+\frac{\partial }{\partial y}\left({\Gamma }_y\frac{\partial Y_w}{\partial y}\right)+\frac{\partial }{\partial z}\left({\Gamma }_y\frac{\partial Y_w}{\partial z}\right)+S_c$$

(11)

Here ${\Gamma }_h$ is the effective diffusion coefficient of enthalpy; ${\Gamma }_y$ and S_c are the diffusivity and source term of water vapor, respectively [44,45,46].

The relationship between the mass concentration of water vapor and moisture content is as follows:

$$Y_w=\frac{d_w}{1+d_w}$$

(12)

Here $d_w$ is the moisture content of wet air, %.

Solar radiation is an important factor affecting the distribution of temperature and humidity fields in the greenhouse. The radiative heat transfer process between indoor and outdoor should be considered in the simulation process, and the solar ray tracing method is chosen to load the solar model. The discrete ordinates radiation model (DO) has a wide range of application and considers the effect of scattering. The gray band model is used to calculate the gray body radiation and non-gray body calculation, and is one of the lesser models that can be used for translucent media [11]. Any optical thickness scenario can be calculated using the DO model. The transfer solution for radiation is performed in Equation (13).

$$\nabla \left(I_{\lambda }\left(\overrightarrow{r}, \overrightarrow{s}\right)\overrightarrow{s}\right)+\left({\alpha }_{\lambda }+{\sigma }_s\right)I_{\lambda }\left(\overrightarrow{r}, \overrightarrow{s}\right)={\alpha }_{\lambda }{n}^2\frac{\sigma T^4}{\pi }+\frac{{\sigma }_s}{4\pi }{\displaystyle \underset{0}{\overset{4\pi }{\int }}{I}_{\lambda }\left(\overrightarrow{r}, {\overrightarrow{s}}^{\prime }\right)}\Phi \left(\overrightarrow{s}\cdot {\overrightarrow{s}}^{\prime }\right)d{\Omega }^{\prime }$$

(13)

Here, $I_{\lambda }$ represents the monochromatic luminance, W/(m² sr¹); $\overrightarrow{r}$ represents the solar position vector; $\overrightarrow{s}$ represents the solar direction vector; ${\alpha }_{\lambda }$ represents the spectral absorption coefficient; ${\sigma }_s$ represents the scattering coefficient; $\sigma$ represents the Stefan-Boltzmann constant, $\sigma$ = 5.669 × 10⁻⁸ W/m² K⁴; $n$ represents the refractive index; $\Phi$ represents the diffusion phase function; and ${\Omega }^{\prime }$ represents the solid angle.

The plants in the greenhouse have some influence on the indoor air flow, such as the respiration and transpiration of the crops, so the influence of the crop cannot be ignored in the numerical simulation [46,47,48]. In order to describe tomato plants in porous media in the greenhouse, the source term is added to the momentum equation according to the Darcy-Forchheimer law. The porous medium is calculated as Equation (14):

$$S_u=\left(-\frac{\mu }{\alpha }u\right)+\left(-C_u\cdot \frac{1}{2}\rho \left|u\right|u\right)=\left(-D_u\mu u\right)+\left(-C_u\cdot \frac{1}{2}\rho \left|u\right|u\right)$$

(14)

Here, $S_{\mathrm{u}}$ is the source term of the momentum equation; $\mu$ is the dynamic viscosity, m²/s; $\alpha$ is the permeability of the porous medium, m²; $D_{\mathrm{u}}$ is the viscous drag coefficient; and $C_{\mathrm{u}}$ is the inertia drag coefficient.

There is a temperature difference between the temperature of the canopy area of the planted crop in the solar greenhouse and the temperature of the air inside the greenhouse, and the magnitude of the apparent heat exchange between the planted crop and the environment inside the greenhouse depends mainly on the aerodynamic properties of the planted crop canopy. From the conservation of energy in the canopy of growing crops, it is known that the plants in the solar greenhouse will store some of the heat from solar radiation, which is converted into latent heat and released into the air inside the greenhouse. The sensible heat exchange between the crop canopy and the room air––the latent heat of transpiration of the crop––is calculated using the following equation. Finally, it is added to the energy equation in the form of a source term. The source of internal heat is calculated as in Equation (15) [49,50]:

$$S_{\varphi }=2LAI\rho c_p\frac{T_c-T_i}{r_a}+LAI\rho \lambda \frac{H_c-H_a}{r_a+r_s}$$

(15)

Here $S_{\varphi }$ is the source term of the energy equation; $LAI$ is the leaf area index, m²/m²; $T_c$, and $T_i$ are the temperatures of the crop and the indoor air, K; $r_a$ denotes the aerodynamic resistance of tomato leaves, s/m; $\lambda$ is the latent heat of evaporation of water, J/kg; $H_c$, and $H_a$ are the relative humidity of the crop and the indoor air; and $r_s$ is the average impedance of the crop stomata, s/m.

The aerodynamic impedance of the crop boundary layer and the average impedance of the crop porosity are calculated in Equations (16) and (17) [51]:

$$r_a=\left\{\begin{array}{ll}840{(\frac{d}{\left|T_c-T_i\right|})}^{0.25}& u\le 0.1 \mathrm{m}/\mathrm{s}\\ 220(\frac{d^{0.2}}{u^{0.8}})& u\ge 0.1 \mathrm{m}/\mathrm{s}\end{array}\right.$$

(16)

$$r_s=200\left(1+\frac{1}{\exp \left(0.05\left(R_{gi}-50\right)\right)}\right)\cdot \left(1+0.11\exp \left(0.34\frac{D_i}{100}-10\right)\right)$$

(17)

Here, $R_{gi}$ is the internal solar radiation; $D_i$ is the saturated water vapor pressure difference; and $d$ is the characteristic blade length per unit length. All parameters are added to the control equation in the form of source terms.

In general, the ratio of the total area of plant leaves to the land area is defined as the leaf area index (LAI). The calculation equation is shown in Equation (18) [52]:

$$LAI=\frac{{\displaystyle \sum _1^n{S}_{leaf}}}{S_{soil}}$$

(18)

where n is the number of leaves; $S_{leaf}$ is the area of each leaf; and $S_{soil}$ is the area covered by soil. In this study, we measured and obtained the average values for different types of leaves to obtain the value of the greenhouse LAI for this experiment.

2.3.2. Boundary Conditions and Geometric Modeling

The computational domain and mesh partitioning cases used in the model are shown in Figure 5a. The model is used to simulate the natural ventilation in the form of combined ventilation at the top and bottom of the arched circular heliostat. The airflow effect under the action of external wind pressure and internal buoyancy is considered in the simulation. The wall boundary conditions are used to define the soil boundary and the surface of the enclosure structure of the arched greenhouse. The top and sides of the external wind flow field of the calculation domain are set up in the form of symmetrical walls. The ground outside the greenhouse is set to a cyclical temperature surface that varies with the external ambient temperature. The boundary below the greenhouse soil is set to 0 °C wall. The windward side of the wind flow field is set as the velocity inlet and the opposing leeward side is set as the pressure outlet boundary condition. In addition, Equation (19) represents the external near-surface airflow profile, simulating different wind velocity gradients at vertical ground level around the greenhouse. In order to avoid the interference of the initial airflow effect on the airflow pattern in the greenhouse, the distance between the surrounding external flow field side interface and the greenhouse should not be less than 10 H. The leeward boundary should be kept at least 15 H from the greenhouse, with H defined as the height of the ridge height of RASG [11,34]:

$$\frac{U(y)}{U(y_r)}={\left(\frac{y}{y_r}\right)}^{\zeta }$$

(19)

Here $U(y)$ is the wind velocity at y height, $U(y_r)$ is the reference wind velocity at the reference height ($y_r=10\mathrm{m}$), and $\zeta$ is the coefficient of roughness.

The same meshing method is used to discuss the ventilation configuration separately (see Figure 5b). The top ventilation configuration is divided into pivoting the window at top (PT), and a rolling film type (RF) and pivoting the window at the north side of the roof (PN). The bottom ventilation configuration is divided into pivoting the window outward (po), pivoting the window inward (pi), and rolling film (rf). Different ventilation configurations should be comprehensively considered and combined to conduct ventilation simulation exploration in the following cases: PT (po), PT (pi), PT (rf), RF (po), RF (pi), RF (rf), PN (po), PN (pi), and PN (rf).

The financial cost, density, specific heat, and thermal conductivity of these materials were obtained from the literature to obtain the specific values of these parameters [1,5,38]. The thermal performance parameters of the materials for each component in the numerical calculation are shown in

Table 1. The thermal performance parameters of the materials in the numerical calculation.

Material Name	Thickness (mm)	Density (kg m−3)	Specific heat Capacity (J kg−1 K−1)	Thermal Conductivity (W m−1 K−1)	Radiation Absorption Rate (%)	Radiation Transmittance (%)
Air	-	1.225	1006.43	0.0242	-	-
Polystyrene board (north roof)	110	30	1368	0.042	-	-
Wood	370	500	2520	0.29	0.7	-
Soil	-	1800	828	1.16	0.86	-
Polystyrene film	0.1	970	750	0.34	0.1	0.8
Tomato canopy	-	300	1000	0.173	0.9	-

.

Based on the experimental test, it is found that high temperature and high water vapor concentration are very easy to occur when the greenhouse ventilation management is improper. Therefore, it is necessary to discuss which ventilation configuration can reduce the temperature and humidity to the suitable temperature for plant growth in the short term. The indoor temperature in summer and winter is as high as 50 °C and 32 °C, respectively, and humidity even reaches more than 85% and continues to rise. Detailed thermal microclimate and boundary conditions are summarized in

Table 2. Initial boundary conditions in the three-dimensional model.

Classifications	Boundary Conditions	Parameters	Boundary Conditions	Parameters
Environmental parameter (Summer)	Outside temperature	26.2 °C	Inside air temperature	50 °C
	Outside humidity	51.5%	Canopy temperature	47.3 °C
	Prevailing wind velocity	3.4 m s−1	Inside air humidity	75.6%
	Inside soil surface temperature	55.7 °C	Canopy humidity	85.2%
Environmental parameter (Winter)	Outside temperature	−7.6 °C	Inside air temperature	32 °C
	Outside humidity	35.8%	Canopy temperature	30.7 °C
	Prevailing wind velocity	3.2 m s−1	Inside air humidity	82.4%
	Inside soil surface temperature	26.6 °C	Canopy humidity	86.3%
Plant canopy (Summer)	Porosity	0.85	Internal loss factor (C1)	0.2
	Pressure drop coefficient [42,43]	0.32	leaf area index (LAI)	2.1
	d	8 mm	Saturated water vapor pressure difference	550 Pa
	Latent heat of evaporation	2.43 J·kg−1	ra	792 s·m−1
			rs	200 s·m−1
Plant canopy (Winter)	Porosity	0.78	Internal loss factor (C1)	0.2
	Pressure drop coefficient [42,43]	0.32	leaf area index (LAI)	2.4
	d	12 mm	Saturated water vapor pressure difference	535 Pa
	Latent heat of evaporation	2.38 J·kg−1	ra	864 s·m−1
			rs	200 s·m−1

as the initial values of the next transient simulation. The boundary conditions under different microclimates in summer and winter are listed to better understand the model establishment and analysis process. The steady-state calculation is first performed to observe the overall convergence, and then the transient time format is used for calculation.

The solver algorithm in this study is a pressure-velocity coupled method, and the SIMPLE algorithm is used to make the calculation with strong stability. In the spatial discretization settings, the cell-based least squares method is chosen for the gradient term. The second-order format is chosen for the pressure term and the second-order windward format is chosen for the momentum term. All other items are in first order windward format. The residual convergence criteria for solving each equation are shown in

Table 3. Residual convergence criteria for each equation.

Continuity Equation	X-Momentum Equation	Y-Momentum Equation	Z-Momentum Equation	Energy Equation	DO-Intensity
1 × 10−3	1 × 10−3	1 × 10−3	1 × 10−3	1 × 10−6	1 × 10−6

[5,11]. In order to determine the appropriate time step value, we set up five different time steps to simulate the same scene. As the calculation time increases, the deviation of temperature and humidity also increases. It leads to unfavorable factors in the stability of calculation results. In order to improve the computational efficiency without distortion of the calculation results, the time step used in this study is 60 s.

2.3.3. Grid Irrelevance Verification

The accuracy of the model mesh directly affects the accuracy of the calculation results. The same boundary conditions are set for the three mesh densities to simulate the sudden high temperature in summer and winter. The management strategy of the experimental greenhouse and the ventilation mode in summer is that the bottom vent rotates 60° and the top vent rotates 60°. The ventilation mode in winter is closed by the bottom vent, and the top vent is rotated by 30°. The greenhouse model used to analyze the grid density adopts the same ventilation strategy as the experimental greenhouse. Three simulation schemes with different mesh numbers are used for discretizing the cell mesh.

Figure 6 shows the temperature difference between the measured value and the simulated value of the monitoring points. Compared with the measured values, it is found that the simulation results of rough grid have high distortion, and it is difficult to accurately describe the indoor temperature fluctuations. The temperature distribution of the middle grid and the fine grid is more similar, and the coincidence degree with the experimental measurement is higher. Further, the correlation (error) test between the simulated and measured values of the average temperature at different times under different grids is carried out, and the reliability of the grid simulation is evaluated by PRMSD parameters. The results of the discrepancy analysis of the mesh are shown in

Table 4. Comparison results of sensitivity analysis.

Variables	Rough Grid	Middle Grid	Fine Grid
Number of elements in the greenhouse	721,673	1,613,721	2,355,716
Number of elements in the plants	228,246	395,738	625,712
Minimum orthogonal quality	0.21	0.35	0.3
Maximum skewness	0.79	0.68	0.76
Discrete time (min)	3.52	4.13	8.1
PRMSD of air temperature in summer (%)	6.4	4.2	3.4
PRMSD of air temperature in winter (%)	6.6	4.7	3.9

. The middle grid has the lowest maximum distortion. The fine grid increased the number of grids by 46% and the discretization time increased significantly. The computational load on the computer is greatly increased, resulting in longer computation times. The PRMSD value of the air temperature using the rough grid is the highest, and the PRMSD values using the middle grid and the fine grid are not much different. Although the PRMSD value of the fine mesh is the lowest, considering the computational load of computer and the time cost of calculation, the middle grid is selected to minimize the computational complexity while ensuring the simulation accuracy.

3. Results and Discussion

3.1. Validation of the Simulation Model

Figure 7a shows the wind rose diagrams of Shenyang area in summer and winter, respectively. As the Shenyang area belongs to the northern temperate zone influenced by the monsoon semi-humid continental climate, the airflow characteristics in summer and winter are opposed. The frequency of airflow directed to the southwest reaches a maximum of 23.7% in summer, while the frequency of southerly and southeasterly winds remains at about 10%. Due to the topography of northern China, northeasterly winds are prevalent in winter with a maximum frequency of 17.6%. As shown in the Figure 7d, the wind direction angle of the south wind direction is defined as 0°, and the angle is increased clockwise with the height as the axis to represent different wind directions. The PRMSD of three positions in summer are 6.69%, 6.35%, and 17.36%, respectively. The PRMSD in winter are 8.11% and 13.63%, respectively. The results show that there is a certain deviation between the simulated value and the measured value of the temperature at the measurement point, but both are within a reasonable error range, and the overall agreement is good.

In the analysis process, 12 measurement points at different positions are selected. They include crop canopy interior and higher air microclimate characteristics. In order to fully verify the accuracy of the model, the environmental changes within 1000 s of ventilation are compared. A total of 10 time scenarios are set after opening the vent. As the first scenario, S01-Second100 represents the state of the vent opening for 100 s. Figure 7c summarizes the distribution of temperature and the relative humidity datasets obtained by experimental tests as well as numerical simulations. The overall statistical data are normally distributed.

Table 5. Test statistic results for the model.

F Test to Compare Two Variances	H0: σ(Dm)2 = σ(Ds)2	H1: σ(Dm)2 ≠ σ(Ds)2
	Temperature	Relative Humidity
F	0.059	0.006
p-value	0.808	0.938
95% confidence interval	{−0.239, 0.267}	{−0.254, 0.252}
Two Sample t-test	H0: µDm = µDs	H1: µDm ≠ µDs
	Temperature	Relative Humidity
T	0.111	−0.010
p-value	0.912	0.992
95% confidence interval	{−0.587, 0.657}	{−2.232, 2.230}
Mean (µ)	µDm = 38.24 °C	µDm = 72.43%
	µDs = 38.17 °C	µDs = 72.36%

The null hypothesis (H₀) is accepted.

shows the results of statistical analysis to determine the validity or negativity of the CFD model. The normality of the measurement and simulation data sets is verified, and a hypothesis test is performed to determine whether the variance of the dataset is homogeneous. The value of the comparison statistic F is calculated at the significance level of 0.05, and the values of F are 0.059 and 0.006. The p values of temperature and relative humidity are 0.808 and 0.938, respectively. The p-value is much higher than the significance level, which is why the null hypothesis raised cannot be rejected. Therefore, it can be considered that the variances of the data are equal (σ_(Dm)² = σ_(Ds)²).

Once the equality of variance is determined, we begin to propose and calculate the hypothesis comparison of the mean difference [53,54]. For these cases, the null hypothesis is established as H0: µ_Dm = µ_Ds. The p-value obtained is 0.912 and 0.992 for temperature and relative humidity, values that are greater than the significance level of 0.05, so the null hypothesis is accepted. This test also includes the confidence interval for the difference in means, for the case of temperature is {−0.587, 0.657} and for humidity is {−2.232, 2.230}; as both intervals include zero, it can be interpreted that, such as the hypothesis test, the difference between the means of the observed and simulated data is not significantly different from zero. Therefore, considering that the mean value of the datasets is statistically similar, the simulation model can be accepted.

Through statistical analysis, the hypothesis test of the difference of the homogeneous variance measure is carried out to test the validity or the rejection of the numerical simulation model. This analysis is complemented by a comparison of the simulated data with the data measured through goodness-of-fit parameters such as mean absolute error ($MAE$) Equation (20), root-mean-square error ($RMSE$) Equation (21), and, finally, through the coefficient of determination ($R^2$) Equation (22) [54,55]:

$$MAE=\frac{1}{N}{\displaystyle \sum _{i=1}^N\left|D{m}_i-D{s}_i\right|}$$

(20)

$$RMSE=\sqrt{{\displaystyle \sum _{i=1}^N\frac{{\left|D{m}_i-D{s}_i\right|}^2}{N}}}$$

(21)

$$R^2=1-\frac{{{\displaystyle \sum _{i=1}^N\left|D{m}_i-D{s}_i\right|}}^2}{{{\displaystyle \sum _{i=1}^N\left|D{m}_i-\overline{Dm}\right|}}^2}$$

(22)

where $N$ is the number of samples, $D{m}_i$ are the measured values of temperature and relative humidity, $D{s}_i$ are the simulated temperature and relative humidity values at a point $i$, and Tm is the average of the measured values.

The quantitative results obtained for the goodness-of-fit parameters between the measured and simulated data (

Table 6. Comparison of measured and simulated temperature and humidity.

	Temperature (°C)			Relative Humidity (%)
	MAE *	RMSE *	R2	MAE *	RMSE *	R2
S01-Second100	0.27	0.33	0.85	0.23	0.28	0.93
S02-Second200	0.24	0.33	0.82	0.32	0.40	0.88
S03-Second300	0.38	0.48	0.86	0.41	0.48	0.78
S04-Second400	0.47	0.56	0.73	0.37	0.49	0.72
S05-Second500	0.24	0.30	0.72	0.23	0.32	0.75
S06-Second600	0.28	0.33	0.82	0.21	0.28	0.90
S07-Second700	0.26	0.32	0.85	0.30	0.36	0.82
S08-Second800	0.30	0.37	0.87	0.18	0.26	0.93
S09-Second900	0.33	0.39	0.89	0.30	0.39	0.88
S10-Second1000	0.18	0.22	0.94	0.27	0.36	0.93

MAE *: Mean absolute error; RMSE *: Root-mean-square error.

) shows that, for temperature, the $MAE$ values behave within the range of 0.18 °C and 0.47 °C and for $RMSE$ between 0.22 °C and 0.56 °C. For relative humidity, the $MAE$ values behave within the range of 0.18% and 0.41%, while for $RMSE$ they are within 0.26% and 0.49%.

The 1:1 scatter plot and the linear regression curve for a 95% confidence interval are constructed for temperature in each simulated scenario (Figure 8). In general terms, there is good agreement between the measured and simulated data, which is verified with coefficients of a determination ($R^2$) ranging from 0.72 to 0.94 (Table 6).

This is also conducted for the relative humidity (Figure 9), for which values of $R^2$ between 0.72 and 0.93 are obtained (Table 6).

The same statistical analysis method is used to verify the accuracy of the model in winter. The statistical results show that the error of each parameter is within a reasonable range.

3.2. Analysis of Airflow Pattern and Thermal Behavior in Summer

As can be seen in Figure 10, the ventilation configuration at the bottom of the pivoting the window outward is not conducive to the reduction of indoor temperature. The high temperature gas retention is obvious and the temperature drop rate is low. The greenhouse with rolling film ventilation at the bottom has a better cooling effect. The greenhouse with pivoting the window inward at the bottom has a better cooling effect on the upper part of the air, but the temperature of the crop canopy is higher because of the air diversion effect of the ventilation window panel. The ventilation configuration of pivoting the window at the north roof promotes the drop of the air temperature between the canopy of crops, so that the crops can quickly enjoy the appropriate growth temperature conditions. The ventilation configuration at the bottom of pivoting the window inward causes higher wind speed in the south of the greenhouse, but the effect of the deflector causes the wind speed between the canopy to be smaller. The distribution of humidity and temperature is similar, and the dehumidification effect of RF (rf) and PN (rf) is more satisfactory. The microclimate distribution of the north–south section shows that more vortex airflow is generated inside the greenhouses of RF (pi) and PN (pi), mainly concentrated in the upper part of the canopy, which hinders the airflow circulation between the canopy.

The adoption of pivoting the window outward in the bottom ventilation configuration hinders the southwesterly airflow from the outside to enter the greenhouse, and the average air speed in the greenhouse is maintained at a low level of 0.13 m/s–0.20 m/s. Reasonable ventilation configuration increases the average air speed from 0.15 m/s to 0.32 m/s. The dehumidification effect of the top ventilation configuration of pivoting the window at top (PT) and the bottom structure of pivoting the window outward (po) is deficient because the concentration of water vapor in the air is seriously affected by the airflow. Ventilation flow directly determines whether effective dehumidification is possible. The high humidity area in the greenhouse with a poor dehumidification effect is mainly concentrated at the top of the greenhouse and near the north wall. The indoor humidity distribution of RF (rf) and PN (rf) is more reasonable.

The ventilation configuration of pivoting the window outward (po) and pivoting the window inward (pi) leads to higher pressure and airflow velocity near the ventilation window at the bottom. Poor air mobility in the crop canopy near the north wall results in lower wind speeds nearby. The bottom rolling film (rf) ventilation configuration leads to more uniform air flow and avoids adverse reactions caused by sudden local strong air flow. The pivoting window inward (pi) at the bottom guides airflow through the canopy and forms a sawtooth dehumidification effect. There are still small areas of high humidity in the greenhouse. When the bottom ventilation window is rolling film (rf), the humidity in 60%-80% of the canopy section is basically maintained at the same level as the outside world.

3.3. Ventilation Performance Estimation and Optimal Structure Selection

The appropriate ventilation performance evaluation index can more intuitively reflect the ventilation effect of the ventilation configuration. In order to further investigate the influence of the ventilation configuration on the ventilation effect, the ventilation flow rate has been defined as an important characteristic evaluation parameter that gives a visual indication of the ventilation performance. The ventilation flow rate (VFR) is defined as the amount of the greenhouse air exchange per second (Equation (23)), which reflects the local ventilation rate in the RASG [38].

$$VFR=\frac{{\overline{m}}_A{A}_V}{\rho V}$$

(23)

Here, $VFR$ is the ventilation flow rate of the greenhouse; ${\overline{m}}_A$ is the average mass flow of the roof vent, kg/s; $\rho$ is the density of air, kg/m³; $V$ is the greenhouse volume, m³; and $A_V$ is the area of the roof vent, m².

In order to obtain a quantitative statistical analysis of the ventilation effect, Figure 11 compares the ventilation flow rate inside the greenhouse in the first 600 s of different cases. By comparing the top and bottom ventilation configurations, it can be concluded that the air circulation rate at the bottom of rolling film (rf) reaches a very high level. The VFR of PN (rf) and RF (rf) reaches 0.58 s⁻¹ and 0.51 s⁻¹, respectively, far exceeding the VFR of other cases. The VFR of PN (rf) is more than three times that of the lowest PT (po). Under the same bottom ventilation configuration, the VFR of PT accounts for 55.4% to 63.8% of the air circulation rate of RF. When the bottom ventilation configuration is pi or rf, the greenhouse with the top ventilation configuration of PN has the highest air circulation rate. The top and bottom ventilation configurations are prioritized according to the air circulation rate. Top ventilation configuration: PN > RF > PT; bottom ventilation configuration: rf > pi > po.

In order to determine the optimal ventilation configuration, Figure 12 comprehensively compares the air and crop temperatures, VFR, and air humidity at a time step of 600 s in nine cases. PN (rf) has the lowest air temperature, indicating that it has the highest cooling range. The temperature in the greenhouse crop area of PN (rf) is 34.21 °C, 3.18 °C lower than that of PT (po) at 37.39 °C. An excellent cooling effect directly determines the rationality of the ventilation configuration. Crops exposed to high humidity for a long time are prone to diseases. PN (rf) effectively reduces the greenhouse humidity to 56.73%, which is 13.69% lower than a PN (po) of 70.42%. It is obvious that the VFR of PT (po) is the lowest at 600 s, indicating that the air circulation permeability is poor, and the air exchange rate between external fresh air and indoor air is low. The VFR of PN (rf) and RF (rf) are 0.58 s⁻¹ and 0.51 s⁻¹, significantly higher than other cases. The negative sign in the figure only indicates the direction of the bar chart and has no practical significance. Considering the comparison of the above environmental indicators, PN (rf) is regarded as the best ventilation configuration scheme because all the indicators of PN (rf) are optimal.

3.4. Structural Optimization of the Proposed Scheme Based on Summer and Winter

To further optimize the shape parameters of the ventilation window of RASG, it is necessary to determine the location and width of pivoting the window at the north roof to reduce the time and cost of greenhouse construction. It will provide theoretical guidance for the actual construction of the greenhouse ventilation configuration. As shown in Figure 13, the rotation fulcrum of the ventilation window at the top is defined as point D₁, the deviation angle of point D₁ from the center line of the greenhouse is α, the width of the ventilation window is H, and the span of the greenhouse is 2R. The axis direction is defined as the positive X-axis direction along the north of the greenhouse, and the positive Y-axis direction along the vertical height. The coordinate position D₁ is (R·(1 + sinα), R·cosα).

Figure 14a vividly shows the variation of indoor air temperature with an angle and width at 600 s. With the increase in angle, the air temperature decreases first and then increases. This indicates that the angle changes affect the air temperature fluctuations, especially when the angle of ventilation window is 25°, when the ideal effect of the ventilation window with a different width is achieved. On the premise of ensuring a good cooling effect, it is recommended that the width of the ventilation window be at least 800 mm. Figure 14b shows the distribution of VFR with a width and angle in the first 600 s. The VFR increases significantly when the width of the ventilation window is below 700 mm. The VFR remains above 0.58 s⁻¹ when the width of the ventilation window exceeds 700 mm, and the increase rate is not obvious. When the width of the ventilation window is less than 300 mm and the angle is less than 20°, the VFR does not exceed 0.37 s⁻¹. Ventilation window angles of more than 25° and a width of not less than 600 mm meet the requirements of higher ventilation flow. When the angle of the ventilation window is less than 25° and the width of the ventilation window is 600–800 mm, the width increases with the decrease in the angle. Figure 14c shows the variation of indoor air humidity. When the ventilation width is less than 500 mm, the humidity is maintained above 59% and the dehumidification effect is poor. When the width exceeds 500 mm, the contour of humidity takes on a trough shape as the angle increases. It can also be considered that with the increase in the angle, the width to reach the ideal humidity decreases first and then increases. According to the definition effect of dehumidification being 57%, the angle is less than 20°, and the width is 600–800 mm. The angle is 20–30° and the width is 600 mm to meet the dehumidification requirements. If the angle is more than 30°, the minimum width should be 800 mm.

In order to reasonably determine the construction standard of pivoting window at the north roof, it is necessary to quantitatively analyze the effect of both the width and angle of the ventilation window on the ventilation effect. The air temperature, air humidity, and ventilation flow rate obtained from cases with the same width and different angles are averaged to avoid the influence of the angle on the width. As shown in Figure 15a, the mean values of environmental indicators under different widths are compared at a time step of 600 s. The ventilation flow rate increases significantly when the width is less than 700 mm. Air temperature and air humidity decrease significantly with increasing width. Under the effect of wind pressure and pressure inside and outside the greenhouse, larger vents cause more heat and water vapor content inside the room to be carried out. The width of the ventilation window reaches the buffer range of 700 mm–900 mm. When the width exceeds 900 mm, all indicators change slowly. When the angle is 25°, the increase in ventilation flow is significantly reduced, and the increase is within 100 m³. More specifically, air temperature and humidity take on a trough shape, reaching its lowest value at an angle of 25°. Generally, the span of the greenhouse varies depending on the needs of the user and the cost of construction. General ventilation window size standards are as follows:

D₁ (R·(1+sin25°), R·cos25°)

D₁ is the rotation fulcrum of the ventilation window at the top and R is half of the span of RASG.

Figure 15. Comparison of greenhouse average temperature, humidity, and ventilation flow with different ventilation configuration parameters: (a) width; (b) angle.

Figure 15. Comparison of greenhouse average temperature, humidity, and ventilation flow with different ventilation configuration parameters: (a) width; (b) angle.

The northeasterly wind prevails in winter in northern China. As the north slope ventilation configuration is located on the north windward side of the greenhouse, the cold outdoor air enters the room through the vent for heat exchange. Larger vent opening angles and larger vent window widths can produce a stronger circulation of cold air indoors, further leading to frost damage to the crop. Figure 16 shows the indoor airflow distribution and the temperature distribution of different cross sections when the ventilation window is opened at an angle of 30°. Outside cold air enters the interior of the greenhouse through the north slope vents. The airflow in the eastern part of the greenhouse is significantly influenced by the outside wind direction. The cold air from outside blows vertically through the vents to the crop and then moves along the soil surface towards the south bottom corner of the greenhouse. Similarly, there is a higher airflow circulation in the greenhouse area west of the middle section of the greenhouse. The vents with red arrows in Figure 16b indicate outside airflow into the greenhouse, while the other lighter colored arrows in the vents indicate outflow. Figure 16c shows that the entry of airflow from both sides drive the hot indoor air towards the middle of the greenhouse, resulting in lower temperatures in the inlet section and higher temperatures in the middle section of the greenhouse, particularly concentrated in the crop canopy area. The local maximum temperature difference reaches 22 °C. Temperature distribution from the eastern section shows that the air temperature entering the greenhouse drops below the appropriate growth temperature for the crop, which can cause some degree of damage to plants. The angle of the ventilation window opening and the width of the ventilation window affect the direction and temperature of the airflow reaching the top of the crop canopy [56].

In order to further analyze the influence of the ventilation window opening angle on indoor airflow, Figure 17 shows the microclimate distribution under different ventilation window opening angles when the width of the fixed ventilation window is 700 mm and the time step is 300 s. The temperature inside the greenhouse is maintained at a high level for a ventilation window opening angle less than 30°, and the temperature drop in the crop canopy is not significant. When the vent window is opened at an angle of more than 40°, the cooling rate of indoor air is significantly higher and the overall average temperature of the greenhouse area north of the vent window is lower. The air temperature directly below the ventilation window drops below 5 °C when the ventilation window is opened at an angle of 50°, which causes an extremely negative impact on the growth of the crop. The cloud map on the north-south section shows that the microclimate activity in the eastern region is more intense and the airflow disturbance is obvious. The cold air entering from the vent gradually spreads to the canopy area. When the opening size of the vent reaches 50°, the airflow velocity directly below the vent reaches more than 1.5 m/s. The cold airflow with a higher speed will invade the interior of the canopy, too, and cause damage to the plants due to freezing. Therefore, considering the effects of temperature and airflow, it is recommended that the opening angle of the ventilation windows on the north roof in winter is lower than 40°.

To further analyze the effect of ventilation window width on the indoor microclimate, Figure 18 shows contour plots of the width on the average indoor air temperature and wind speed at 700 s of ventilation for different ventilation window opening angles. The effect of the ventilation window width on temperature is small when the ventilation window opening angle is 10°, and the increase oin average wind speed is not significant when the ventilation window width exceeds 800 mm. When the ventilation window is opened at an angle of 20°, the cooling range and ventilation rate of the ventilation window width in the range of 800 mm to 900 mm are higher. The ventilation effect decreases when the width of the vent window exceeds 900 mm. When the ventilation window opening angle is between 30° and 40°, increasing the ventilation window width also increases the ventilation effect. However, the difference between 900 mm and 1000 mm is smaller, especially for the increase in the average wind speed. Considering that the longer the width of the ventilation window, the higher the cost, it is recommended that the width of the ventilation window should be 900 mm.

3.5. Configuration Upgrade and Performance Experimental Comparison

Based on the optimal ventilation configuration parameters of RASG obtained from the simulation, a solar greenhouse with an optimized new ventilation configuration and ventilation strategy can be constructed with the basic structural dimensions of the experimental greenhouse as a reference. A schematic diagram of two RASGs with different ventilation configurations is shown in Figure 19. PT (pi) is the experimental control greenhouse. PN (rf) is a newly constructed RASG with a bottom rolling film ventilation configuration and a top pivoting window at the north roof. Both greenhouses are planted with 0.5 m high tomato plant, and both use the same configuration of a water circulation heat storage system for winter night heating production. Environmental data in both greenhouses are monitored by temperature and humidity sensors. The bottom and top vents are kept open during the summer, and the top vents of both greenhouses are opened at the same angle of 25°. In order to ensure the same cross-sectional area when the bottom vents of both greenhouses are opened, the bottom pull film ventilation configuration at PT (pi) is rolled at a height of 0.5 m and the bottom pivoting window at PN (rf) is turned at an angle of 50°. During the experiment, the vent is closed to make the average air temperature reach 50 °C, and then the vent is continuously opened for 600 s to observe the internal microclimate changes. The average outdoor wind speed reaches 3.3 m/s.

The experimental data of PN (rf) were measured in summer by using the same experimental instrument and measuring point arrangement scheme. Figure 20 shows the variation of air temperature and wind speed with time in two greenhouses. After the opening of the vent, the air temperature of the two greenhouses decreases significantly, and the cooling rate of PN (rf) is higher than that of PT (pi). At 600 s, the internal air temperature difference between the two greenhouses is 2.85 °C. The average wind speeds of PT (pi) and PN (rf) are 0.21 m/s and 0.37 m/s, respectively. The air flow of PN (rf) is better, which is beneficial to the air exchange of the plant canopy. The ventilation driven by wind pressure is mainly caused by the wind speed of the inlet of the greenhouse. Ventilation driven by thermal pressure is mainly caused by the density difference at different levels, which is associated with the temperature variation at different height locations [6,18]. The outside wind speed changes in real time during the greenhouse ventilation. Wind pressure driving plays a leading role at high wind speeds, and thermal pressure is more important at low wind speeds. The bottom vent of PN (rf) is more permeable, which can create higher inlet velocity than the rotating window of PT (pi). In addition, previous contributions show that they have only achieved a certain foundation for the top ventilation structure [11]. This study proves that the combined ventilation form of top and bottom creates a new and more efficient ventilation configuration, which further enriches the optimization ideas and strategies of greenhouse ventilation. The results further show that the ventilation configuration optimization method in this study has high feasibility and practicality.

4. Conclusions

In this study, the ventilation configuration type and size parameters of RASG in a high latitude cold region were optimized by experimental and simulation methods. On the basis of good experimental verification, the RASG ventilation optimization model was constructed, and the influence law and mechanism of different ventilation configuration parameters on greenhouse microclimate were systematically expounded. The difference of ventilation configuration was analyzed from the dynamic mode of temperature, humidity, and wind speed. The reasonable ventilation combination structure was determined, and the calculation formula of structural parameters in the actual greenhouse construction process was proposed, which provided a theoretical basis for greenhouse construction. The results from this study have indicated that:

(1)

A three-dimensional numerical simulation model of CRASG ventilation considering crop canopy airflow disturbance is established.

(2)

The bottom ventilation configuration of the rolling film increases the air flow into the interior of the greenhouse and results in more uniform airflow through the crop canopy. The adoption of the north roof ventilation configuration enhances the airflow through the interior of the greenhouse.

(3)

Based on the hot summer climate conditions, the optimal ventilation configuration obtained is PN (rf). In winter, the opening angle of the north roof ventilation window is less than 40° to ensure the rapid cooling of the interior of the greenhouse without the crops being affected by the cold environment. After optimization, the deviation angle between the rotation point of the ventilation window and the center line of the greenhouse is 25° and the width is 900 mm.

(4)

A 10-m span RASG is established to verify the ventilation effect of PN (rf). The experimental results further verify the reliability of the numerical simulations and prove that the optimal ventilation configuration construction parameters of RASG obtained in this study are reasonable and effective.

The results of the study can provide theoretical guidance for the construction of RASG ventilation configuration to further improve the efficiency of crop environmental regulation and reduce energy consumption. The further realization of crop quality improvement and high yield is of far-reaching significance to the development of the agricultural greenhouse industry.