Optimization of Greenhouse Structure Parameters Based on Temperature and Velocity Distribution Characteristics by CFD—A Case Study in South China

Abstract

Greenhouses are applied to mitigate the deleterious effects of inclement weather, which facilitates the optimal growth and development of the crops. South China has a climate characterized by high temperature and high humidity, and the temperature and relative humidity inside a Venlo greenhouse are higher than those in the atmosphere. In this paper, the numerical model of the flow distribution of a Venlo greenhouse in South China was established using the CFD method, which mainly applied the DO model, the k-e turbulence model, and the porous medium model. The porous resistance characteristics of tomatoes were obtained through experimental research. The inertial resistances of tomato plants in the x, y, and z directions were 80,000,000, 18,000,000, and 120,000,000, respectively; the viscous resistances of tomato plants in the x, y, and z directions were 0.43, 0.60, and 0.63, respectively. The porosity of tomato plants was 0.996. The average difference between the temperature of the established numerical model and the experimental temperature was less than 0.11 °C, and the average relative error was 2.72%. This research also studied the effects of five management and structure parameters on the velocity and temperature distribution in a greenhouse. The optimal inlet velocity is 1.32 m/s, with the COF of velocity and temperature being 9.23% and 1.18%, respectively. The optimal skylight opening is 1.76 m, with the COF of velocity and temperature being 10.68% and 0.88%, respectively. The optimal side window opening is 0.67 m, with the COF of velocity and temperature being 9.25% and 2.10%, respectively. The optimal side window height is 1.18 m, with the COF of velocity and temperature being 9.50% and 1.33%, respectively. The optimal planting interval is 1.40 m, with the COF of velocity and temperature being 15.29% and 0.20%, respectively. The results provide a reference for the design and management of Venlo greenhouses in South China.

1. Introduction

Greenhouses are applied to mitigate the deleterious effects of inclement weather, which facilitates the optimal growth and development of the crops. Compared to traditional solar greenhouses, it has the advantages of high strength, light weight, good seismic performance, fast construction speed, small construction quantities, land saving, and a high degree of factory production. It is easy to achieve standardized design and management, and is conducive to the rapid development of standardized greenhouses [1]. However, a Venlo greenhouse is prone to high temperatures in South China due to its subtropical climate [2].

There are many cooling methods commonly used for the purpose of solving representative high temperature issues in greenhouses in China, including spray cooling [3], natural ventilation [4], and fan-pad system [5]. Arbel et al. showed that the cooling performance of spray cooling reduced the temperature by 8.5 °C to 12.0 °C [6]. McCartney et al. have shown that natural ventilation cooling can reduce temperatures by 1.3 °C to 3.6 °C [7], and it is not suitable to be applied in South China due to its weak cooling amplitude. Ayad Saberian et al. evaluated the temperature reduction of a fan-pad system in a subtropical desert greenhouse by 15–25 °C compared to natural ventilation methods by a computational fluid dynamics (CFD) approach [8]. Fan-pad systems are mostly used for ventilation and cooling in greenhouses, and there are many studies on individual parameters, such as fan velocity [9], side windows [10], and skylights [11], which have made important contributions to the optimization of the indoor temperature distribution. Young-Bok Kim et al. studied the effects of the open level of the side window to control the temperature and relative humidity in a fog cooling greenhouse. The target temperature and relative humidity were set at 28 °C and 75%, respectively. The three modes of the side window open level were 0%, 50%, and 100%. The results showed that the average dry bulb temperatures of the inside air were 28.2, 27.2, and 26.3 °C, respectively [12]. Huang et al. established a CFD model based on the three-dimensional turbulence model equation of Reynolds number and the Darcy–Forkheimer law. Models were created for both no-crop and crop conditions to simulate different wind speeds and temperatures. Corresponding simulation diagrams were generated and analyzed. The results show that changing the wind velocity to 3 m/s or the fan position to 1 m is better than the situation where the wind speed is 2 m/s and the fan position is 1.3 m [13]. Khaoua et al. adopted the two-dimensional computational fluid dynamics (CFD) method to numerically analyze the influence of wind speed and the structure of the roof ventilation opening on the airflow and temperature distribution in the separated greenhouse. The numerical model successfully verified the ventilation rate data collected in a 2600 m² four-span greenhouse, which was divided into two compartments by plastic partitions. The results show that, under the same conditions, the wind speed variation range of the crop cover layer at different exhaust hole and compartment positions is 0.1–0.5 m s⁻¹, and the temperature difference variation range is 2–6 °C [14].

The test methods are limited by the disadvantages of limited sampling points, measurement errors, and uncontrollable environments [15]. Improvements in computing facilities, together with theoretical and experimental studies, increased the understanding of the biophysical process in a greenhouse system [16]. At present, computational fluid dynamics (CFD) has been widely applied for greenhouse climate simulation [17], which can be combined with actual environmental parameters to set initial conditions to simulate the characteristics of greenhouse microclimate distribution under different conditions [18,19].

Ventilation plays an indispensable role in the maintenance of optimal conditions for plant growth and development [20,21]. Tsafaras et al. [22] studied the greenhouse equipped with a pad and fan system, and found the modified greenhouse design resulted in about 14% higher fresh weight production and more than 40% water saving was achieved on evaporative cooling. Similarly, Lee and Short [23] used CFD to evaluate natural ventilation rates and airflow distributions in a multi-span greenhouse, while Santolini et al. [24] studied the effect of shading screens on airflow patterns within the greenhouse through CFD simulations. Wang [25] et al. found that the cooling effect was obvious in plastic greenhouses with side windows and top ventilation. He [26] et al. found that the indoor temperature was most uniform, and the ventilation rate was highest under the configuration of top and side windows. Wang [27] et al. suggested that skylights and side windows can enhance the ventilation and cooling effect. However, they lacked research on factors such as crop planting density and side window height.

This paper study the temperature spatial distribution prediction in South China greenhouse by CFD. It also optimized the management and structure parameters by the validated CFD model, including fan velocity, side window height, side window opening, skylight opening, and planting density.

2. Materials and Methods

2.1. Experimental Materials

The Venlo greenhouse is located at the Test Base of Guangdong Academy of Agricultural Sciences in Baiyun district, Guangzhou (113°26′ E, 23°23′ N, and 72 m altitude), and oriented in a north-south direction. The greenhouse has dimensions of 30 × 8 × 6.2 m. The greenhouse contained 3 rows and 19 columns of tomato plants. Tomato plants were planted with the hydroponic nutrient solution, with two adjacent rows spaced 3.2 m apart. The greenhouse has side windows on the left and right sides, with the lower edge of the window being 0.5 m above the ground, and the side window opening height being 1.6 m. The leeward side of the roof was provided with a long strip of glazed skylights, which were opened using an articulated push-open structure, and all skylights were closed during the test period. The temperature sensors were equipped in the greenhouse at the locations shown in Figure 1 with a red color. Temperature sensors (RC-4HC, −20 °C~+60 °C, ±0.1 °C measured value, Jiangsu Jingchuang Electronics Co., Ltd., Xuzhou, China) were used to monitor the temperature in the greenhouse, and the recording time interval of the sensor was set to 15 min.

The measurements were taken in the afternoon of 17 June 2023. The average temperature, relative humidity, and solar radiation of air outside the glasshouse at 2 p.m. were 34.65 °C, 63.4% and 48,976 W/m², respectively, which were monitored by a meteorological station (M6, Shandong Renke Measurement and Control Technology Co., Ltd., Jinan, China).

2.2. Mathematical Model

In this study, SolidWorks 2021 was used for 3D modelling of a single glass greenhouse with a length of 30 m, a width of 4 m, and a height of 6.2 m. ICEM 2022 R1 and Fluent 2022 R1 were used for meshing and simulation of the test greenhouse.

Since the air flow inside the greenhouse was slow, which ranged from 0.24 to 1.12 m/s, and the temperature change was small, which was in line with the Boussinesq assumption. The airflow inside the greenhouse was regarded as a turbulent flow with normal temperature, low velocity, and incompressible. The model was solved by the finite volume method (FVM) and the SIMPLEC algorithm to discretize the established Reynolds time-averaged N-S equations. The basic governing equations of the greenhouse could be described in the following forms [28]:

(1)

Mass conservation equations:

$${\displaystyle \frac{\partial u}{\partial x}}+{\displaystyle \frac{\partial v}{\partial y}}+{\displaystyle \frac{\partial w}{\partial z}}=0$$

(1)

where u, v, and w are the velocity components in the x, y, and z direction, m/s.

(2)

Conservation of momentum equation:

$${\displaystyle \frac{\partial \left(\rho uu\right)}{\partial x}}+{\displaystyle \frac{\partial \left(\rho uv\right)}{\partial y}}+{\displaystyle \frac{\partial \left(\rho uw\right)}{\partial z}}={\mu }_{eff}\left({\displaystyle \frac{{\partial }^{2}u}{\partial x^2}}+{\displaystyle \frac{{\partial }^{2}u}{\partial y^2}}+{\displaystyle \frac{{\partial }^{2}u}{\partial z^2}}\right)-{\displaystyle \frac{\partial p}{\partial x}}$$

(2)

$${\displaystyle \frac{\partial \left(\rho \nu u\right)}{\partial x}}+{\displaystyle \frac{\partial \left(\rho \nu \nu \right)}{\partial y}}+{\displaystyle \frac{\partial \left(\rho \nu w\right)}{\partial z}}={\mu }_{eff}\left({\displaystyle \frac{{\partial }^2\nu }{\partial x^2}}+{\displaystyle \frac{{\partial }^2\nu }{\partial y^2}}+{\displaystyle \frac{{\partial }^2\nu }{\partial z^2}}\right)-{\displaystyle \frac{\partial p}{\partial y}}$$

(3)

$${\displaystyle \frac{\partial \left(\rho wu\right)}{\partial x}}+{\displaystyle \frac{\partial \left(\rho wv\right)}{\partial y}}+{\displaystyle \frac{\partial \left(\rho ww\right)}{\partial z}}={\mu }_{eff}\left({\displaystyle \frac{{\partial }^{2}w}{\partial x^2}}+{\displaystyle \frac{{\partial }^{2}w}{\partial y^2}}+{\displaystyle \frac{{\partial }^{2}w}{\partial z^2}}\right)-{\displaystyle \frac{\partial p}{\partial z}}-\rho g\beta \left(T-T_{ref}\right)$$

(4)

where μ_eff is the effective viscosity; p is the pressure on the fluid microcell, Pa; T is the air temperature, K; T_ref is the air reference temperature, K; ρ is the density of the air, kg/m³; β is the coefficient of thermal expansion, β = T⁻¹, 1/K; q is the heat source, W; C_p is the specific heat capacity, J/(kg·K).

(3)

Conservation of energy equation:

$${\displaystyle \frac{\partial \left(\rho uT\right)}{\partial x}}+{\displaystyle \frac{\partial \left(\rho vT\right)}{\partial y}}+{\displaystyle \frac{\partial \left(\rho wT\right)}{\partial z}}={\displaystyle \frac{q}{C_p}}+{\displaystyle \frac{{\lambda }_{eff}}{C_p}}\left({\displaystyle \frac{{\partial }^{2}T}{\partial x^2}}+{\displaystyle \frac{{\partial }^{2}T}{\partial y^2}}+{\displaystyle \frac{{\partial }^{2}T}{\partial z^2}}\right)$$

(5)

where λ_eff is the effective thermal conductivity, W/(m·K).

(4)

Radiation model:

The solar radiation model is loaded in the CFD model by the solar ray tracing method because the experiment is conducted in the middle of summer, and the solar radiation has a great influence on the flow field characteristics in the greenhouse. In addition, a discrete-coordinate radiation model (DO model) is chosen to consider the effect of thermal radiation. The DO model treats the radiation equation (RTE) propagating along the s_r direction as a field equation, and the specific equation can be expressed in the following form:

$${\displaystyle \frac{\mathrm{d}I\left(\overrightarrow{r},\overrightarrow{s}\right)}{\overrightarrow{s}}}+\left(\alpha +{\sigma }_s\right)I\left(\overrightarrow{r},\overrightarrow{s}\right)=\alpha n^2{\displaystyle \frac{\sigma T^4}{\pi }}+{\displaystyle \frac{{\sigma }_{\mathrm{s}}}{4\pi }}{\int }_0^{4\pi }I\left(\overrightarrow{r},\overrightarrow{s}\right)\varphi \left(\overrightarrow{s},{\overrightarrow{s}}^{\prime }\right)\mathrm{d}{\mathsf{\Omega }}^{\prime }$$

(6)

where $\overrightarrow{r}$ is position vector; $\overrightarrow{s}$ is direction vector; ${\overrightarrow{s}}^{\prime }$ is scattering direction vector; s is stroke length, m; I is the radiation intensity, W/m², which depends on the position and direction; α and n are the absorption coefficient and refractive index, and $\sigma$ is Stefan-Boltzmann constant, W/(m²·K⁴); ${\sigma }_{\mathrm{s}}$ is scattering coefficient; $T$ is the average temperature, °C; $\varphi$ is the Phase function; ${\mathsf{\Omega }}^{\prime }$ is a solid angle.

(5)

Porous model

There are many gaps, irregular spaces, and complex surfaces inside real tomato crops. It is inappropriate to restore a real geometric model, which clearly requires a significant amount of mesh and time for computation. Therefore, the real model is often simplified as porous media. It is necessary to improve the accuracy of the data by taking the average reading within one minute. The collected data are fitted into Equation (11)

$$\mathsf{\Delta }\mathrm{p}=\mathsf{\alpha }{\mathrm{v}}^2+\mathrm{b}\mathrm{v}$$

(7)

Many drag coefficients were tested in a wind tunnel experiment, and their values are usually 0.31 or 0.32. The study obtained the drag coefficients of cherry tomatoes by means of a ventilated drag test rig, where the porosity of the tomato crop was measured by the drainage method. The inertial drag factor ${\mathrm{C}}_2$ and the viscous drag coefficient ${\displaystyle \frac{1}{\mathsf{\alpha }}}$ were calculated based on the pressure drop characteristic relationship equation, applying the principle of equality of coefficients and comparing it with the fitted linear equations, so as to calculate the inertial drag coefficient $\mathsf{\alpha }$ and the viscous drag coefficient $\mathrm{b}$.

$$\mathsf{\Delta }\mathrm{p}={\displaystyle \frac{1}{2}}{\mathrm{C}}_2\mathsf{\rho }{\mathrm{v}}^2+{\displaystyle \frac{\mathsf{\mu }}{\mathsf{\alpha }}}\mathrm{v}$$

(8)

where Δp is the pressure drop between the inlet and outlet, ρ is the air density, and μ is the air viscosity.

A porous medium is a solid skeleton with interconnected pores in which single-phase or multiphase fluids can flow. As a measure of the structural properties of porous media, porosity is commonly used. In this experiment, the volume of the tomato plant was measured using the drainage method.

$$\epsilon ={\displaystyle \frac{v_{drainage}}{v_{bucket}}}$$

(9)

where $\epsilon$ is the porosity of the tomato crop, $v_{drainage}$ is the volume of water drained from the container into the tomato crop, and $v_{bucket}$ is the volume of the container filled with water.

2.3. Measurement of the Porous Parameters

Fresh tomato plants were obtained from a greenhouse located at the Tianhe Experimental Base in Guangzhou, China. The experimental platform is designed to gain the inertia resistance and viscous resistance coefficients in the x, y, and z directions, respectively. It is made up of a suction fan (CZR-900, air volume: 1320 m³/h, Shanghai Yongshang Electrical Appliance Co., Ltd., Shanghai, China), differential pressure gauge (testo 510, range: ±13.79 kPa, accuracy: ±0.01 kPa, Testo Instruments International Trade (Shanghai) Co., Ltd., Shanghai, China), anemometer (testo 410-1, range: 0.4~20 m/s, accuracy: ±0.1 m/s, Testo Instruments International Trade (Shanghai) Co., Ltd., Shanghai, China), current regulator, air tunnel, tomato plants and measuring holes, whose position is shown in Figure 2. Tomato plants were divided into several 1 m long plants, which were placed into a platform in x, y, and z directions, respectively. And the testing process should be carried out as quickly as possible to prevent blade detachment or dehydration from data errors. Adjust the current regulator to allow the negative pressure fan to reach the measured wind velocity, while recording the average value of the differential pressure gauge within one minute to ensure accuracy. The steps are as follows:

First, place the tomato plants as shown in the figure. Adjust the fan velocity through the current regulator and measure the air velocity at the fan outlet position. Then, place two tubes with differential pressure gauges at the inlet of the fan and the inlet of the air duct, where H0 remains constant, another pipe was used to measure H1, H2 and H3 in sequence, and obtain the average value by three times measurement. Finally, the ventilation resistance coefficient of tomatoes was obtained through the method of function fitting and equal coefficient correspondence.

Figure 3 shows the fitted equation of experimental data. As Figure 3 shows, the minimum R-squared is 0.96, which indicates the fitting equation is accurate. As a consequence, the viscous resistance and the inertial resistance of tomato plants are shown in

Table 1. The resistance parameters of tomato plants.

Items		Values
Inertial resistance C2 (m−1)	X-axis	80,000,000
	Y-axis	18,000,000
	Z-axis	120,000,000
Viscous resistance ${\displaystyle \frac{1}{\alpha }}$ (m−2)	X-axis	0.43
	Y-axis	0.60
	Z-axis	0.63

.

2.4. Numerical Method

2.4.1. Mesh Independence Research

The computational domain is an area of 8 m × 5.6 m × 30 m, which was developed by the SolidWorks 2023 program and was meshed by the ICEM 2022 R1 program. Wet-curtain and fan blade were simplified as a face to reduce the quantity of mesh. The air inlet was set at the wet curtain where air passed through, and the air outlet was set at the fan, side windows, or skylights. The computational domain was discretized by a tetra/mixed mesh type. Velocity, pressure, and temperature were calculated by relative discretization equations after integration in each part. The mesh density was thickened in significant parts such as the fan outlet, wet-curtain inlet, planting porous, side windows, and skylights for better gradients in the simulation of such parts, whose details are shown in Figure 4.

Mesh independence plays a vital role in assessing mesh quality, which will affect the results of the simulation. The mesh size is closely related to the accumulation of error [29,30]. Large mesh quantity increases computational time and cumulative error, while small mesh quantity increases the iterative error, which is difficult to converge. Therefore, it is notable to ensure a proper mesh size, which facilitates enhancing computation efficiency as well as calculation accuracy. The main purpose of this part is to ascertain that mesh quantity has little effect on mesh quality and simulation results by monitoring the air velocity with different mesh quantities under the same simulation boundary condition. As illustrated in Figure 4b, the velocity values of four points on the cross-section at a distance of 0.5 m from the air inlet are employed as evaluation criteria. Upon increasing the amount of mesh to 1.147 million, the monitored velocity on P1~P4 exhibits a tendency to converge and ceases to undergo significant changes.

2.4.2. Boundary Condition

In previous CFD studies on greenhouse applications, the k-e turbulent model has unique advantages in solving high Reynolds coefficient problems. A standard k-e model is an effective tool for simulations, as it can reduce the calculation time. The inlet turbulence intensity (I) was calculated by using Equation (10).

$$Re=\rho vd/\mu$$

(10)

$$I=0.16(Re{)}^{-{\displaystyle \frac{1}{8}}}$$

(11)

where I is turbulence intensity, and the value is calculated as 3.42%; $Re$ is Reynolds number, and the value is calculated as 224,481.72; d is the characteristic diameter, m.

In order to simplify the model and improve the simulation efficiency, the following assumptions are made for the model. The house gas is a Newtonian fluid; The air in the greenhouse is incompressible during the flow and conforms to the Boussinesq assumption. In this study, steady-state simulation was used because the outside temperature of the glass shed was stable during the experimental test, and the negative pressure fan operated continuously.

The air inlet of the wet curtain was set as a velocity inlet with a velocity of 1.12 m/s and a temperature of 26.85 °C. The outlet boundary condition was set as a pressure outlet. When there were side windows and a skylight in the model, a pressure outlet was applied for their boundary condition. The solid wall was a wall without slip, and the thermal boundary condition was set to temperature. Considering that the heat production of the tomato plants ultimately acts on the maintenance of body temperature, the tomato plant was set as a thermostat.

A solar tracking model was added to simulate the changes in solar radiation angle during solar motion, and the average radiation during the experiment was 48,976 W/m². Compared to solar radiation, the respiration heat generated by plants is very small, so its impact on heat transfer is ignored. The specific thermodynamic parameters of the greenhouse are shown in

Table 2. Thermal dynamic parameters of greenhouse materials.

Physical Property/unit	Air [31]	Glass [32]	Soil [33]	Tomato Plants [34]
Density/kg·m−3	1.184	2530	1620	990
Cp/J·kg·K−1	1006.58	840	1480	3680
Thermal Conductivity/W·m·K−1	0.02604	1.20	1.3	0.476
Absorption Coefficient	0.20	1.20	0.8	0.014
Diffusion Coefficient	0	0.1	0.2	0.80
Scattering Coefficient	1.00	1.00	1.0	0.80
Refractive Index	0.86	0.85	0.9	0.95

.

2.4.3. Solution of Simulation

The simulation is initialized by a steady solver to avoid calculation errors, which often happen at the start of an unsteady calculation. In addition, a pressure-based type solver was employed to simulate temperature changes. The gravitational acceleration was set to 9.81 m/s².

All of the simulations were performed with the help of an i9-12900KF Sixteen-Core-Processor with 64 GB RAM (Dell (China) Co., Ltd., Xiamen, China). The CFD code mainly uses the finite volume method by Fluent 2022 R1. Furthermore, a first-order implicit transient formulation was implemented. In the pressure–velocity coupling, a SIMPLE algorithm was employed for the solution of the Navier–Stokes equations, while a least squares cell-based method was utilized for the computation of the gradient. The standard discretization schemes were employed for the computation of pressure. First-order upwind discretization schemes were employed for the calculation of momentum, turbulent kinetic energy, and turbulent dissipation rate. Respectively, the convergence standard of momentum, continuity, energy were their value meet 10⁻³, 10⁻³, 10⁻⁶, and 10⁻⁵.

Table 3. Design of simulation groups.

Simulation	Velocity (m/s)	Skylight Opening (m)	Side Window Opening (m)	Side WindowHeight (m)	Planting Interval (m)
1	0.72	0	0	1.18	1.40
2	0.92	0	0	1.18	1.40
3	1.12	0	0	1.18	1.40
4	1.32	0	0	1.18	1.40
5	1.52	0	0	1.18	1.40
6	1.32	0	0	1.18	1.40
7	1.32	0.44	0	1.18	1.40
8	1.32	0.88	0	1.18	1.40
9	1.32	1.32	0	1.18	1.40
10	1.32	1.76	0	1.18	1.40
11	1.32	0.88	0	0.78	1.40
12	1.32	0.88	0	0.98	1.40
13	1.32	0.88	0	1.18	1.40
14	1.32	0.88	0	1.38	1.40
15	1.32	0.88	0	1.58	1.40
16	1.32	0.88	0	0.98	1.40
17	1.32	0.88	0.33	0.98	1.40
18	1.32	0.88	0.67	0.98	1.40
19	1.32	0.88	1.00	0.98	1.40
20	1.32	0.88	1.33	0.98	1.40
21	1.32	0.88	1.33	0.98	0.46
22	1.32	0.88	1.33	0.98	0.67
23	1.32	0.88	1.33	0.98	0.96
24	1.32	0.88	1.33	0.98	1.40
25	1.32	0.88	1.33	0.98	2.13

lists all the imulation groups studied in this article under different effect factors.

2.5. Evaluation Method

The air condition in the greenhouse is well described by uniformity of the temperature and velocity, which is adopted by Zhang et al. [35], Yu et al. [36], and Hemming et al. [37]. In this paper, the evaluation criteria of temperature and velocity distribution under different conditions are the coefficient of variation.

$${CV}_T={\displaystyle \frac{\sqrt{{\sum _{i=1}^N\left(T_i-T_a\right)}^2}}{T_a}}$$

(12)

$${CV}_v={\displaystyle \frac{\sqrt{{\sum _{i=1}^N\left(v_i-v_a\right)}^2}}{v_a}}$$

(13)

where ${CV}_T$ and ${CV}_v$ are the coefficients of variation of temperature or velocity, $T_i$ and $v_i$ are the measured values of temperature or velocity, and $T_a$ and $v_a$ are the average temperature and velocity, respectively. $N$ is the number of measurement locations.

3. Result

3.1. Model Validation

The comparison between the measured and simulated temperature at 27 points is shown in Figure 5. It is obvious that the maximum discrepancy between the measured and simulated temperatures at each measurement point is 1.4 °C, while the mean temperature discrepancy at each measurement point is 0.5 °C, and the average relative error is 2.72%, which indicates that the model has high accuracy. The discrepancy between the average temperature of the model and the actual temperature can be attributed to the fluctuations in actual radiation during the test period. The radiation value of the model, however, represents the average radiation during the test period, and the discrepancy between the actual radiation and the modelled radiation was not considered, resulting in the elevated average temperature of the model.

3.2. Airflow Distribution

The ventilation processes in the greenhouse strongly influence the indoor climate organization, which is closely linked to the interaction between airflow uniformity and plant growth. The distribution of air velocity within the greenhouse sections is shown in Figure 6. Influenced by the pressure drop inside and outside the greenhouse, the inflow air velocity is significantly higher than the internal air velocity, indicating that the negative pressure ventilation effect is significant and has a strong traction effect on the wind direction of the airflow inside the greenhouse.

The airflow distribution in three planes of the longitudinal section (X) of the greenhouse is given in Figure 6a, with X = 0.7, X = 5.1, and X = 7.3 m, respectively. The highest and lowest velocities of the plant porous domain are 0.8 m/s and 0.002 m/s, respectively, while they are 1.1 m/s and 0 m/s in the whole domain. The airflow distribution in three planes of the longitudinal section (Z) of the greenhouse is given in Figure 6b, with Z = 0.7, Z = 7.5 m, Z = 14.25 m, Z = 21 m, and Z = 29.25 m, respectively. In the triangular area above the wet curtain and air outlet, the airflow velocity is slower than the average value. In the crop area, the wind velocity between crops is slightly higher than the velocity inside the crops because the crop obstructs the airflow, with a difference of approximately 0.2 m/s.

3.3. Temperature Distribution

Figure 7a shows the contour of the temperature distribution of the greenhouse along the x-axis. The temperature of each section gradually increases with height. The temperature in the plant area ranges from 34 °C to 40 °C, and the highest temperature is seen above the fan. The temperature was observed to be higher in the region above the wet curtain and fan, reaching approximately 40–50 °C. The thermal environment within the greenhouse represents a significant factor influencing the optimal growth of indoor crops. The temperature distributions across the greenhouse Z-axis section at specific points, including 0.75 m, 7.5 m, 14.25 m, 21 m, and 29.25 m, are illustrated in Figure 7b. In the triangular area above the wet curtain and air outlet, the airflow velocity is slow, resulting in the air absorbing too much solar radiation and creating a high-temperature region. It can be observed that the temperatures at each point within the longitudinal section increased with height in the horizontal position. Furthermore, a notable symmetrical distribution was evident on both sides of the greenhouse, indicating a more consistent temperature distribution on both sides of the greenhouse. The temperature exhibits greater fluctuation within the two longitudinal sections of the greenhouse, specifically at Z = 0.75 m and Z = 29.25 m. Additionally, the temperature at the bottom is observed to be lower. However, with the increase in height, the temperature gradient increases, reaching approximately 49.5 °C at the top.

3.4. Effect of the Inlet Velocity

The side window opening, side window height, skylight opening, and planting interval were set as 100%, 1,18 m, 1.76 m, and 1.4 m, respectively, while the velocity of inlet ranges from 0.72 m/s to 1.32 m/s to study the effects of flow distribution in the greenhouse (Figure 8). As shown in Figure 8a, the air inside the greenhouse cools by about 0.13 °C, and the mean velocity increases by about 0.1 m/s for every 0.2 m/s increase in air velocity at the air inlet. It was found that the COF of velocity was gradually reduced (Figure 8b), and the coefficient of temperature had a sudden change when the inlet velocity increased. From the perspective of the COF combining velocity and temperature, when the inlet velocity was 1.32 m/s, the velocity and temperature distribution in the greenhouse was more uniform, and the corresponding COF are 9.23% and 1.18%, respectively.

3.5. Effect of the Skylight Opening

The inlet velocity, side window opening, side window height, and planting interval were set as 1.32 m/s, 100%, 1,18 m, and 1.4 m, respectively, while the velocity of skylight opening changed range from 0 m to 1.76 to study the effects of skylight opening in the greenhouse. Figure 9a shows the variation trend of average velocity and temperature as the skylight increased gradually. The results showed that the average velocity first increased and then decreased, while the average temperature decreased as the skylight opening increased. As shown in Figure 9b, when the skylight opening increased gradually, the COF of velocity and temperature showed a trend of decreasing first and then increasing. The minimum coefficient of variation of velocity and temperature in the greenhouse was observed at a skylight opening of 1.76 m, while the COF was 10.68% and 0.88%, respectively.

3.6. Effect of Side Window Opening

The inlet velocity, skylight opening, side window height, and planting interval were set as 1.32 m/s, 1.76 m, 1.18 m, and 1.4 m, respectively, while the side window opening was varied from 0% to 100% to study the effects of the side window opening on velocity and temperature distribution. As shown in Figure 10a, the average velocity shows a trend of increase, while the average temperature shows a trend of decrease with the increase of the side window opening. As the side window opening increased from 0 m to 1.33 m, temperature decreased from 33 °C to 30 °C, and velocity increased from 0.59 m/s to 0.97 m/s. When the side window opening increased gradually, the COF of velocity and temperature showed a trend of decreasing first and then increasing (Figure 10b). The minimum coefficient of variation of velocity and temperature in the greenhouse was observed at the side window opening of 0.67 m, while the COF was 9.25% and 2.10%, respectively.

3.7. Effect of Side Window Height

The inlet velocity, skylight opening, side window opening, and planting interval were set as 1.32 m/s, 1.76 m, 100%, and 1.18 m, respectively, while the height of the side window was set from 0.72 m to 1.58 m to study the effects of the side window height. Figure 11a shows the effect of side window height on the average velocity and temperature inside the greenhouse. Obviously, as the height of the side window increases, there are certain fluctuations in temperature and velocity. Figure 11b shows the effects of the side window height on both the coefficient of variation of velocity and temperature distribution. The minimum COF of velocity and temperature were both showed in height of 1.18 m, but it showed high temperature at the same time. Based on the above results, it is considered that its optimal parameter is 1.18 m, and the COF was 9.50% and 1.33%, respectively.

3.8. Effect of Planting Interval

The inlet velocity, skylight opening, side window opening, and side window height were set as 1.32 m/s, 1.76 m, 100%, and 1.18 m, respectively, while the planting interval changed range from 0.46 m to 2.13 m to study the effects of the planting interval. As shown in Figure 12a, the average velocity shows a downward trend with the increase of planting density, while the average temperature shows a trend of first decreasing and then increasing. Figure 12b shows the effects of planting interval on velocity uniformity and temperature uniformity in greenhouses. As the planting interval increased, the coefficient of variation of temperature and velocity decreased. However, the coefficient of variation of temperature increased suddenly when the plant interval reached 2.13 m. Considering the average temperature, average velocity, and COF comprehensively, the optimal planting interval in the greenhouse is 1.4 m, and the COF was 15.29% and 0.20%, respectively.

4. Discussion

4.1. Numerical Model

Cooling is usually used to overcome the excessive heat effect produced by greenhouse crops, and the combination of a well-designed greenhouse and an appropriate cooling strategy is an effective way to solve the problem of high temperatures in greenhouses. This paper establishes a numerical model of the temperature distribution in the greenhouse and optimizes the greenhouse structure.

Ventilation is important for crop production and quality in greenhouse cultivation, and side window openings, side window heights, and skylight window openings are the basis of greenhouse design with a suitable planting strategy. As shown in Figure 6, the velocity distribution of X and Z axes reveals that the velocity flow rate over the wet curtain and crop area is slow, with a maximum wind velocity of only 0.2 m/s in the crop area. This is attributed to the obstruction of air by the canopy of the crop in question. By using a fan-pad method to cool down, Xu et al. found that the temperature fluctuated within the range of 35 °C to 41 °C, which is similar to these results [38]. In the crop-growing area, temperature variations were minimal, remaining at approximately 35 °C due to equilibrium in heat exchange between the crop, air, and solar radiation.

The average relative error between the temperature test value and the simulation value is 2.72%, which was more precise than the previous research [21]. The reason might be that the parameters of tomato plants used in this study are closer to the actual values, causing the higher accuracy of the numerical model.

4.2. Inlet Velocity

It is not difficult to find that the air slowly passes through the plant area, which is replaced by a porous medium. There is no doubt that the lowest velocity occurred in the interior of the porous medium because of its obstruction, resulting in air circulation, too. Meir Teitel et al. simulated the same porous medium in the plant canopy domain using the Horkheimer equation [39]. The highest velocity was seen at the outlet of the fan, where air rapidly ran out of the greenhouse. The reason for this phenomenon is that the energy of the inlet air velocity is limited, and air spreads out into the greenhouse unevenly, resulting in this phenomenon where the wind velocity value is lower than the average level in the triangular area above the wet curtain and the triangular area above the fan.

The effect of mechanical ventilation on the uniformity of airflow organization in greenhouses is measured by the coefficient of variation, as shown in Figure 8. We adjust the wind velocity of the wet curtain flowing into the greenhouse by changing the power of the negative pressure fan. The inlet velocity gradually increased from 0.78 m/s to 1.32 m/s, and it was observed that the velocity variation curve gradually decreased, which is similar to the research of Guo et al. [40]. Guo et al. found that with the increase of inlet velocity, the coefficient of variation of temperature and velocity inside the greenhouse decreased. However, when the temperature coefficient of variation curve decreases to a certain extent, it suddenly increases, which is somewhat different from the research of Guo et al. This may be due to excessive inlet velocity, causing the temperature of the fan and wet curtain to be too low, resulting in local temperature unevenness. Will changing the magnitude of wind velocity, temperature, and wind velocity uniformity continue to change, and further gradient and experimental research are needed.

4.3. Skylight Opening

The skylight provides an opportunity for indoor hot air to exchange with external cold air, promoting the uniformity of indoor air. The installation of skylights promotes heat exchange between the air inside and outside the greenhouse, and the appropriate skylight opening facilitates uniform air velocity and temperature distribution. To investigate the effect of skylight opening on the indoor velocity uniformity and temperature uniformity, the size of skylight opening was varied to assess the velocity uniformity and temperature uniformity inside the greenhouse. As shown in Figure 9a, with the skylight opening increasing gradually, the coefficient of variation of velocity was found to fluctuate sinusoidally. When the skylight opening increased, the average indoor temperature gradually decreased, while the average velocity first increased and then decreased. The results of Wang’s research indicate that when the opening angle of the side window is 45°, the condition where the skylight is opened at 60° is more conducive to the ventilation and cooling of the greenhouse environment than the conditions where the opening angle of the skylight is 45°or 75°. Under the combined ventilation method of using side windows and skylights with reasonable opening degrees, the overall average temperature of the greenhouse decreased from 38.4 °C to 36.9 °C, and the average temperature in the crop area of the greenhouse decreased from 38.1 °C to 37.2 °C. The average wind velocity in the greenhouse has increased from 0.45 m/s to 0.89 m/s [41].

4.4. Side Window Opening

The side windows are usually opened together with the skylight to achieve a higher cooling effect. On the basis of optimizing the skylight opening to 100%, this study further concluded that the uniformity of greenhouse gas organization is best when the side window opening is 0.67 m. It indicates that adjusting the opening of the side windows helps to reduce indoor temperature and increase indoor air flow velocity. Many researchers have confirmed this conclusion, too. A similar phenomenon was also shown in the research of Sun et al. [42], which studied the minimum temperature uniformity coefficients, reached 0.36 while its maximum is calculated to be 0.24 in this study. Kim et al. reported that in a naturally ventilated greenhouse, when the height of the side window opening increased from 0.6 m to 1.2 m, the indoor and outdoor temperature difference decreased from 14.0 °C to 7.1 °C [43]. Meir Teitel et al. found that the steady-state values of temperature decrease when the side window opening increases from 0.4 m to 1.4 m [44]. Moreover, the pattern of side window opening and the coefficient of variation are not very clear. It can only be concluded that when the side window opening is 0.67 m, the coefficient of variation of temperature and velocity is the lowest, and the air is more uniform compared to the other four situations. Referring to the coefficient of variation curve of the side window opening in Figure 10, it is recommended to adjust the height of the side window to 0.67 m to achieve better cooling, ventilation, and uniformity.

4.5. Side Window Height

This paragraph considers the height of the bottom of the side window from the ground as a factor affecting the greenhouse climate environment. The previous section has already optimized the inlet velocity, skylight opening, and side window opening, but the study of side window height is also particularly important. The change in the height of the side window from the ground will result in a change in the airflow towards the plant area. Being too low or too high can lead to uneven wind velocity and temperature in the plant area, and in severe cases, it can damage plant growth. Many scholars have studied the height of side windows in the environmental control field, but they have overlooked the height of the side windows from the ground [44,45]. This article achieves the effect of side window height on greenhouse velocity and temperature uniformity by keeping the window area constant and controlling the height of the window above the ground. Of course, the mutual influence between side window height and side window opening has not been studied yet; further research will be conducted in the future.

4.6. Planting Interval

The indoor air was most uniform when the planting interval was 1.40 m, which may be related to the respiration and transpiration of tomato plants, and the plant and the indoor air were fully exchanged, resulting in the indoor air temperature becoming uniform. As the plant interval continued to increase, the unevenness coefficient of indoor air temperature did not continue to decrease, indicating that the heat exchange between the plant and the air reached saturation, and continued to increase; the plant interval was not able to reduce the uniformity of air. It appears that the best temperature uniformity in the glasshouse, planting interval was positively correlated with the mean velocity, and planting interval was generally negatively correlated with the mean temperature. When plant interval is 2.4 m, both the soil and the indoor air accumulates a large heat load due to solar radiation and could not absorb the sensible heat of the air through the plants, resulting in the highest average indoor temperatures at this time, and the plants also had the effect of improving the uniformity of the indoor air velocity. The larger the distance between crops, the higher the indoor air velocity. Considering the greenhouse area and field management, 1.40 m is advised for farmers to plant in line with the actual situation and practical management. Considering the planting interval, temperature, and velocity uniformity, the planting interval in greenhouses should be selected as 1.40 m. Furthermore, studying planting density is a complex issue. It is necessary to determine the relationship between crop height [46], row spacing, and quantity through more experiments.

5. Conclusions

In this paper, the numerical model of the flow field distribution of the Venlo greenhouse in South China was established by using the CFD method, and its structural parameters were optimized. The following main conclusions were drawn:

(1)

The porous resistance characteristics of tomatoes were obtained through experimental research. The inertial resistances of tomato plants in the x, y, and z directions were 80,000,000, 18,000,000, and 120,000,000, respectively; the viscous resistances of tomato plants in the x, y, and z directions were 0.43, 0.60, and 0.63, respectively. The porosity was 0.996, and the R-squared of the curve fitting reached 0.96.

(2)

The average difference between the temperature of the established numerical model and the experimental temperature is less than 0.11 °C, and the average relative error is 2.72%, indicating a relatively high experimental accuracy.

(3)

Based on the established numerical model, the effects of wind velocity, skylight opening, side window opening, side window height, and planting interval on the velocity of greenhouse temperature were studied, and the optimal parameter combination was obtained. The optimal wind velocity, skylight opening, side window opening, side window height, and planting interval of the Venlo greenhouse are 1.32 m/s, 1.76 m, 0.67 m, 1.18 m, and 1.40 m, respectively.