Design of Greenhouse Ventilation Systems with Computational Fluid Dynamics: Balancing Performance and Energy Sustainability

The global population growth and rising demand for food have emphasized the critical role of protected agriculture, particularly in controlled-environment greenhouses [1]. These structures provide a means to optimize crop yield and quality by offering protection from adverse weather conditions and pests, thus ensuring year-round production. However, the operational sustainability of modern greenhouses is significantly challenged by their high energy consumption [2], primarily driven by the need for precise climate control [3]. Conventional mechanical systems, such as fans and cooling pads, require substantial electricity and contribute to a large carbon footprint. In response, there has been a global push towards more sustainable and economically viable solutions [4]. Natural ventilation, which significantly reduces the reliance on mechanical systems, outperforms as a highly effective and passive method to regulate greenhouse climate by allowing passive heat and mass transfer [5]. This passive climate control mechanism in greenhouses not only lowers electricity costs but also minimizes carbon footprints, so as to align with global efforts towards energy-efficient and sustainable practices [6]. The efficiency of natural ventilation, however, is a complex interplay of various factors, including the geometric configuration of the greenhouse, its environmental conditions, and the design of ventilation openings and ducts. Optimizing this system is not merely a consideration of design but a core strategy for transforming energy-intensive greenhouse operations into a more sustainable form of agriculture.

The study of natural ventilation in greenhouses has traditionally relied on empirical and experimental methods. These conventional approaches, such as physical models in wind tunnels and full-scale experiments using tracer gases, are often costly and time-consuming [7], [8]. They also present important practical challenges, including the difficulty of precisely measuring airflow and temperature distribution in every region of a large and complex environment. These limitations often restricted the quantity of design variations that could be tested, hence a slow and resource-intensive process for a comprehensive optimization. Computational fluid dynamics (CFD) has emerged as an invaluable tool for analyzing and optimizing airflow patterns in complex environments like greenhouses [9], [10]. Unlike traditional experimental methods, which are often expensive and laborious, CFD simulations provide a detailed visualization of air movement, temperature distribution, and mass transfer. This allows researchers to virtually test various design modifications and predict their impact on the overall performance of the ventilation system [11]. CFD-guided optimization has been successfully applied to improve natural ventilation by optimizing vent sizes, locations, and types to enhance thermal comfort for crops and reduce energy consumption [12], [13].

It has been highlighted that when energy-consuming devices such as fans, heating, and cooling systems are employed, the configuration and geometry of air ducts play a crucial role in system performance. A key aspect of effective natural ventilation is the design and shape of air ducts, which can significantly influence airflow resistance and overall system efficiency [14]. Rectangular ducts are widely used due to their ease of fabrication and installation, but they create higher pressure drops and greater turbulence, which can impede airflow [15]. Oval ducts, on the other hand, offer a compromise and provide lower airflow resistance than rectangular ducts, while being more adaptable for installation in confined spaces [16]. Circular ducts are considered the most aerodynamically efficient, as their smooth and curved shape minimizes pressure loss and turbulence [17]. To accurately model the greenhouse environment, the dense plant canopy should be taken into account. The plant acts as a significant obstacle to airflow; it increases drag and redistributes the air within the greenhouse [18]. These effects could be precisely captured when researchers model the plant canopy as a porous medium by assigning specific hydraulic resistance properties to the crop volume [19]. This approach provides a realistic representation of how plants influence the microclimate, hence enabling more accurate predictions of temperature distribution and ventilation efficiency.

Despite extensive research on greenhouse ventilation, a significant gap remains in the systematic optimization of duct shape and quantity, particularly for natural ventilation systems. Most existing studies have focused on a single duct shape or a limited set of configurations, leaving substantial room for performance improvements unexplored. This study addressed that gap by conducting a comprehensive and CFD-guided optimization of duct geometry and quantity for enhanced efficiency of natural ventilation. Besides, this research investigated the impact of circular, oval, and rectangular duct shapes and quantities on the efficiency of natural ventilation and their associated energy saving in greenhouses. The primary objective is to identify the optimal duct configuration that maximizes airflow and thermal uniformity, while effectively reducing energy consumption. Performance was assessed through simulations that evaluated the effect of each design on crop thermal comfort and overall temperature distribution within the greenhouse. By systematically varying duct parameters, the most effective design was determined to cater for specific greenhouse geometries and climatic conditions. This comprehensive CFD-guided analysis is expected to highlight configurations that offer substantial cost saving and promote more sustainable greenhouse operations.

This section presents the physical model, computational setup, and methodology employed in CFD simulations. Additionally, the validation procedures are outlined to ensure the reproducibility and reliability of the simulation results.

Figure 1a provides a comprehensive overview of the geometric model. Figure 1a illustrates the computational domain, which features a single-span greenhouse measuring 20 m in length, 10 m in width, and 6 m in height. The internal volume includes a simulated plant canopy, modeled as a porous medium with dimensions of 18 m × 8 m × 2 m and positioned at ground level. The model incorporated a natural ventilation system consisting of four outlets, each with a cross-sectional area of 1 × 3 m², symmetrically located on the opposing sidewalls. Figure 1b presents a visual summary of six investigated configurations of distinct roof inlets. Two primary scenarios were considered: 1) a high-count inlet configuration with ten smaller ducts (R10, O10, and C10), each having a cross-sectional area of 3 m²; 2) A low-count configuration with three larger ducts (R3, O3, and C3), each with a cross-sectional area of 10 m². For both scenarios, three different duct shapes were analyzed: rectangular (R), oval (O), and circular (C). This systematic variation in shape and quantity allows a detailed comparative analysis of their influence on ventilation performance.

In CFD simulations, the air was modeled as an incompressible fluid, with a constant density assumption at standard atmospheric conditions.

The steady-state flow field was simulated using ANSYS Fluent 16.1, based on the governing equations of continuity, momentum, and energy. Airflow was modeled using the Reynolds-Averaged Navier-Stokes equations. The shear stress transport k–$\omega$ model was employed due to its proven accuracy in predicting separated and swirling flows, rendering it appropriate for the complex geometry and flow behavior characteristic of naturally ventilated greenhouses [20]. The governing equations for mass, momentum [21], and energy are as follows:

Here, $\rho$ denotes the fluid density, $u_\text{i}$ is the velocity component, P represents the static pressure, and $\mu$ is the dynamic viscosity of the fluid. The term $\tau_\text{ij}$ corresponds to the Reynolds stress tensor, while $g_\text{i}$ indicates the body force component. The symbol $h$ refers to the specific enthalpy, $k_\text{eff}$ is the effective thermal conductivity accounting for both molecular and turbulent contributions, and $S_\text{h}$ represents the volumetric source term associated with heat generation within the domain. Table 1 presents a list of constants incorporated for the simulation.

The simulations were conducted using the following boundary conditions. The ground temperature was set to a constant 304 K, while the temperatures of the sidewalls and roof were maintained at 303 K. Airflow entering through the roof inlets was defined using a logarithmic velocity profile, a standard approach for modeling atmospheric boundary layers. This velocity profile, shown in Figure 1a, varies as a function of height ($y$) from the ground and was derived from a predefined atmospheric profile consistent with values reported in a previous study [23]. The outlet boundaries were assigned a pressure-outlet condition with a gauge pressure of 0 Pa. A comprehensive mesh independence study was conducted to ensure the accuracy and reliability of the simulation results. The mesh was refined iteratively, with mass flow rate at the outlets selected as the key parameter for convergence analysis. Mesh independence was achieved when the difference in mass flow between two consecutively finer meshes was less than 1%, to ensure that the simulation results were not influenced by mesh resolution. The computational domain was discretized using a tetrahedral mesh with an average element size of 100 mm.

The meshed geometry of the greenhouse is illustrated in Figure 2. The main panel displays the overall tetrahedral mesh applied to the computational domain, while the inset provides a close-up view of a specific region, thus highlighting the mesh quality and resolution. This figure supports the robustness of the mesh strategy and the credibility of the numerical analysis.

This study provided an understanding of the impact of different duct shapes and quantities on key performance metrics relevant to greenhouse ventilation. The primary parameters analyzed included: velocity, temperature distribution, airflow streamlines, wall shear stress (WSS), ventilation rate, air changes per hour (ACH), and annual energy saving.

Temperature distribution, which directly affects crop health and productivity, is assessed to evaluate the uniformity of cooling or heating within the greenhouse. Airflow streamlines are used to visualize circulation patterns, revealing recirculation zones or dead spots that can diminish ventilation efficiency.

WSS serves as an indicator of energy losses and surface friction effects, while ventilation rate and ACH quantify air renewal, which is critical for maintaining optimal CO₂ levels and humidity control.

Ultimately, annual energy savings are estimated by comparing different duct configurations; this offers an economic assessment that links choices of ventilation design to long-term operational efficiency and sustainability.

The analysis focused on key performance indicators, including velocity fields, temperature distribution, airflow streamlines, WSS, ventilation rate, energy saving, and ACH. These parameters allow a systematic comparison of six cases, to highlight the influence of duct geometry and quantity on the overall ventilation efficiency and thermal comfort inside the greenhouse.

In Figure 3, the magnitude of airflow within the greenhouse volume is illustrated at a vertical plane cross-section of the greenhouse, which is perpendicular to the inlet plane. The panels were organized into two groups: The top row displays the ten-inlet configurations (R10, O10, and C10), while the bottom row shows the three-inlet configurations (R3, O3, and C3). In all panels, the highest velocities are concentrated near the inlet openings. As the air moves into the greenhouse, its velocity decreases remarkably and the air transitions to the cooler areas. Comparing the different shapes, the panels for the circular (C10 and C3) and oval (O10 and O3) ducts show a more gradual and widespread velocity reduction within the greenhouse. In contrast, the panels for the rectangular ducts (R10 and R3) appear to have a slightly more abrupt velocity drop, with a more concentrated high-velocity zone at the inlets.

Figure 4 demonstrates the temperature distribution across the middle surface of the canopy for the six different ventilation configurations. The temperature was distributed relatively uniformly across the canopy for all configurations. Nevertheless, a closer examination revealed that the circular duct configurations (C10 and C3) exhibited more extensive regions of blue coloring, which signified a lower overall temperature on the plant canopy surface, compared to the rectangular duct configurations (R10 and R3). The latter indicates a higher prevalence of warmer areas owing to different duct configurations.

Figure 5 presents a three-dimensional visualization of airflow streamlines for each of the six duct configurations. These streamlines trace the complete airflow path from the inlets to the outlets, thus offering insight into internal circulation behavior. The rectangular configurations (R10 and R3) exhibit more pronounced localized swirling and recirculation zones near the inlet regions. In contrast, the circular (C10 and C3) and oval (O10 and O3) configurations display smoother and more direct flow paths throughout the greenhouse interior. This visualization underscores the significant influence of duct shape on airflow behavior, with smoother profiles promoting more uniform and efficient ventilation.

Figure 6 illustrates the WSS contours with a color scale ranged from 0 to 283 × 10^-6 Pa on the exterior surfaces of the greenhouse for all six duct configurations. WSS serves as a key indicator of frictional energy losses and the aerodynamic drag experienced by airflow along the surfaces. In all configurations, the highest WSS values were observed near the sharp edges and inlet openings, where flow separation and shear were most pronounced. A comparative analysis revealed that the rectangular ducts (R10 and R3) exhibited both larger high-WSS areas and higher WSS magnitudes than their circular (C10 and C3) and oval (O10 and O3) counterparts. This suggested that rectangular ducts generated greater frictional drag, thus reducing overall flow efficiency compared to shapes that are more streamlined.

In Figure 7, the results demonstrate a trade-off between energy efficiency and ventilation capacity among the six duct designs. The left chart shows annual energy savings (in Joules) on the Y-axis, while the right chart shows the number of air changes per hour (ACH) on the Y-axis. The X-axis in both charts represents the different ceiling configurations. The results revealed that the R3 configuration saved the lowest amount of energy, whereas the C3 configuration conserved the highest amount. Additionally, the R3 and R10 configurations achieved the highest ACH, whereas the C10 configuration obtained the lowest ACH.

Table 2 provides a numerical summary of the simulation results, with a breakdown of the corresponding values for mass flow rate, temperature of plant area, and power saving for the six configurations (R10, O10, C10, R3, O3, and C3). The mass flow rates are ranged from 18.066 kg/s for C10 to 18.375 kg/s for R3 and R10. The temperature of plant area shows a range from 297.427 K for C3 to 298.692 K for R3. Power saving in watts, varies significantly, with R3 having the lowest power consumption at 95 466.357 and C3 having the highest at 122 736.9075.

This research offered a systematic and integrative evaluation of greenhouse ventilation by concurrently examining multiple interdependent parameters, including velocity distribution, thermal regulation, WSS, energy consumption, and air change rates. In contrast to prior studies that have typically considered these factors in isolation, the present work provided a holistic framework that captured the coupled influence of aerodynamic and thermal behaviors on ventilation efficiency. Furthermore, three distinct duct geometries, i.e., rectangular, circular, and oval, were investigated across different inlet configurations, thus representing the first comparative assessment of its kind. This multi-faceted approach highlighted the critical role of duct design in optimizing greenhouse microclimates while simultaneously advancing energy-efficient and sustainable operation strategies.

Velocity fields for the six duct configurations in Figure 3 highlighted how duct geometry and quantity influenced airflow penetration. The rectangular ducts of R10 and R3 exhibited localized high-velocity regions near the inlets, which rapidly decayed into stagnant zones. This behavior has been mentioned in previous studies, where sharp-edged openings generated stronger entrance jets but poorer overall mixing [24]. By contrast, the circular and oval ducts (C10, O10, C3, and O3) distributed velocity more uniformly across the greenhouse; this improved canopy ventilation and reduced dead zones. Such smoother velocity attenuation could promote better microclimate regulation and gas exchange at the plant level [25]. An investigation into multi-chapel greenhouses discovered that combining roof and side wall vents with different opening layouts, particularly when vent shapes were smooth rather than sharp-edged, resulted in significantly more uniform airflow distribution and reduced hot spots inside the canopy [26]. Environmental Monitoring and CFD Simulations in another study reported that roof vent placement and shape influenced not only the magnitude of airflow, but also the downstream mixing and microclimate uniformity under natural ventilation [27]. These studies supported our findings that circular and oval ducts outperformed rectangular designs in terms of velocity decay and dead-zone reduction. However, in the arched greenhouse case, some configurations with large rectangular vents under high external wind showed better air change rates, thus implying that rectangular vents may retain their merits in specific climatic or wind conditions [28].

The temperature contours, as shown in Figure 4, indicated relatively uniform conditions across all duct designs, but with distinct differences in canopy cooling efficiency. The circular duct configurations (C10 and C3) yielded noticeably cooler canopy temperatures, consistent with the enhanced airflow distribution observed earlier. Prior work emphasized that improved mixing and reduced stratification could lower canopy temperature, thereby enhancing photosynthetic efficiency and yield [29]. In contrast, rectangular ducts (R10 and R3) demonstrated warmer localized areas which could potentially increase thermal stress for plants. This outcome aligned with findings about the crucial impact of duct geometry on microclimate uniformity in greenhouse environments [5]. Modifying roof shape to improve air circulation led to significantly lower maximum canopy temperatures in sweet pepper greenhouses, in parallel to the advantage witnessed in the case of circular ducts in the current study [30]. Similarly, the semi-closed study of greenhouses demonstrated that combining cooling ducts with carefully placed vents reduced hot spots and improved average thermal comfort within the crop canopy [31]. In contrast, research suggested that when canopy height was tall ($\approx$ 1.8 m), even large side vent openings yielded less cooling benefit possibly due to reduced wind penetration deeper into the canopy; this explained the underperformance of rectangular ducts relative to rounded ones in certain scenarios [32]. On the other hand, a study discovered larger discrepancies of temperature in configurations with non-uniform vent layout or occluded flow paths, hence indicating that flow path obstruction by canopy or by geometry could degrade thermal uniformity [33]. These comparisons affirmed that duct shape (rounded versus sharp-edged), quantity of inlets, and canopy height are critical in modulating cooling efficiency. When designing for thermal comfort and energy saving, the findings in this study suggested that circular or oval ducts provided a superior trade-off, especially where tall crop canopy or obstructed airflow might reduce the effectiveness of ventilation in rectangular designs.

The rectangular ducts generated more pronounced swirling zones close to the inlet edges, indicative of flow separation due to sharp corners (see Figure 5). Such recirculation could trap heat and humidity, thus leading to non-uniform microclimates as observed in experimental studies of greenhouse ventilation [34]. Similarly, a study reported that sharp-edged ducts in greenhouses led to less uniform airflow and increased thermal stratification [35]. Conversely, circular and oval ducts allowed smoother airflow trajectories with fewer vortices, which enhanced the renewal of indoor air. This observation corresponded with CFD analyses, which showed that rounded duct inlets reduce turbulence intensity and promote energy-efficient ventilation [36], [37].

Figure 6 quantifies WSS patterns, reflecting energy dissipation through duct–air interactions. The rectangular ducts present higher and more extensive WSS regions around inlet edges, consistent with their sharper geometry. This indicated greater frictional drag and higher energy losses, which negatively affected overall system efficiency. By comparison, the circular and oval ducts exhibited significantly lower WSS, thus reinforcing their advantage in reducing frictional losses. This supported the notion that smoother duct geometries not only improve airflow distribution but also extend duct durability and reduce long-term maintenance costs.

As shown in Figure 7, while the R3 and R10 configurations achieve the highest ACH, their energy savings are the lowest, highlighting a trade-off between ventilation capacity and energy efficiency. Similar findings reported that higher ACH did not necessarily translate into optimal energy use [38]. Conversely, the circular duct configurations, particularly the C3, achieve the highest energy saving while maintaining satisfactory ACH. This balance between ventilation and cost aligned with sustainable greenhouse management strategies, which prioritize reduced energy consumption without compromising crop health [38]. Based on these findings, product shape is proved to be strongly conducive to both economic and environmental outcomes.

The circular ducts (C3 and C10) consistently produced lower canopy temperatures; with the C3 reaching the coolest value of 297.427 K. These configurations exhibited higher power saving, particularly the C3 at 122 736 W, which significantly outperformed the rectangular ducts (see Table 2). Although rectangular ducts (R3 and R10) allowed slightly higher mass flow rates, they corresponded to higher plant temperatures and lower energy efficiency. This pattern was consistent with earlier CFD-based evaluations of greenhouse ventilation, where duct geometry dictated the balance between airflow rate and energy performance [22]. Consequently, the data reinforced that rounded duct geometries were more advantageous for balancing climate control, crop comfort, and cost-effectiveness.

Despite the comparative analysis performed in this study, several limitations should be acknowledged. First, the simulations assumed steady-state boundary conditions and neglected transient environmental variations, such as diurnal changes in solar radiation, external wind fluctuations, and seasonal temperature shifts, which can significantly influence greenhouse microclimates. Second, the plant canopy was modeled as a homogeneous porous medium, which simplified the complex aerodynamic interactions between individual leaves and stems. This model of plant canopy may lead to discrepancies when predicting localized airflow patterns and heat transfer near plant surfaces. Additionally, the thermal properties of the greenhouse materials were considered constant, when possible variations were ignored due to aging or moisture content. Future work should aim to implement fully transient CFD and fluid-structure interaction (CFD-Fuel Stratified Injection) simulations with dynamic environmental inputs to capture real-world ventilation performance. Incorporating detailed plant morphology and leaf-level turbulence modeling could further improve the accuracy of microclimate predictions. Finally, experimental validation in operational greenhouses would strengthen the reliability of the numerical findings and allow refinement of model parameters for broader applications.

The present study thoroughly examined the role of duct geometry and quantity in enhancing the efficiency of greenhouse ventilation system. By analyzing airflow patterns, temperature distribution, WSS, and overall energy performance, the research highlighted how variations in duct design could substantially influence both the microclimatic conditions and the energy demands of greenhouses. The comparative evaluation of different duct geometries revealed notable differences in their aerodynamic and thermal performance, as summarized below:

Circular and oval ducts generally:

• Promoted smoother airflow streamlines;

• Reduced recirculation zones;

• Achieved more uniform temperature distributions across the canopy.

Rectangular ducts were associated with:

• Localized high-velocity regions;

• Elevated WSS values;

• Greater frictional losses, leading to reduced energy efficiency.

Among all configurations tested, the C3 (circular, 3 inlets) was the most energy-efficient and it offered the highest power saving while maintaining favorable thermal conditions. The R3 (rectangular, 3 inlets) demonstrated less effective temperature control and lower energy efficiency, despite a high air change rate. These results underscored the importance of geometric optimization in natural ventilation systems and the balance of airflow uniformity, energy consumption, and thermal comfort at the crop level. Circular ducts were particularly advantageous when long-term operational costs and plant health were prioritized. Overall, this study provided valuable insights for designing energy-efficient greenhouse ventilation system, a reference framework for future experimental validation and large-scale implementation.