Numerical Evaluation of a Negative Pressure Ventilation System for Ammonia Emission from a Solid-Covered Manure Storage Tank

Abstract

Ammonia (NH₃) emissions from temporary manure storage tanks represent a significant environmental concern in livestock production systems. While combining solid covers with negative pressure ventilation is a promising strategy to mitigate these emissions, there is currently a lack of systematic research on its design optimization and performance. This study employs Computational Fluid Dynamics (CFD) to evaluate the effectiveness of a solid-covered manure storage tank combined with negative pressure ventilation for controlling NH₃ emissions. A validated CFD model was developed to simulate airflow and ammonia transport under open-field and covered conditions. The influences of tank headspace depth, slot type (top and side), and slot location on outlet ammonia concentration were investigated. Results show that headspace depth is one of the important parameters affecting ammonia transport, with deeper headspaces consistently reducing outlet NH₃ concentrations. Compared with no-slot scenarios, top slots could increase ammonia emissions by inducing impinging-jet effects, whereas side slots exhibited depth-dependent impacts, reducing emissions at 1.0 and 1.6 m depths but increasing them at 0.4 m depth. All the differences in ammonia emission across the simulations can be attributed to the difference in the near-wall velocity. The findings provide useful guidance for the design and optimization of ammonia mitigation strategies in manure storage systems.

1. Introduction

Livestock production is a major source of environmental pollutants, with ammonia (NH₃) being one of the most critical gases emitted from manure management systems. Ammonia emissions from livestock operations substantially contribute to atmospheric nitrogen deposition [1], soil acidification [2], secondary particulate matter formation [3], and nutrient imbalance in terrestrial and aquatic ecosystems [4]. Additionally, high NH₃ concentrations within animal houses and surrounding regions adversely affect animal health [5], worker safety [6], and overall farm sustainability [7,8]. Consequently, mitigating NH₃ emissions from manure storage and treatment processes has become a central concern in sustainable livestock production.

Among the stages of manure management, i.e., collection, storage, treatment, and land application, manure storage is often the dominant contributor to NH₃ emissions. During storage, urea in urine is rapidly hydrolyzed to ammonium (NH₄⁺) by urease enzymes, a fraction of which volatilizes as NH₃ depending on temperature, pH, and air exchange rate [9,10]. In open or partially covered manure storage systems, this process occurs continuously and is frequently intensified by environmental factors such as high temperature, wind speed, and agitation [11,12,13,14]. Temporary manure storage tanks, commonly used prior to land application or treatment, are typically open to the atmosphere and therefore become a significant and difficult-to-control source of on-farm NH₃ emissions [15].

To mitigate NH₃ volatilization, various strategies have been investigated, including chemical additives, floating covers, and air treatment systems [16,17,18]. Chemical approaches, such as slurry acidification, effectively suppress NH₃ volatilization by lowering pH but may be costly, corrosive, or impractical for temporary storage [19]. Physical covers, such as straw, plastic sheets, or floating membranes, reduce emission by limiting gas diffusion to the atmosphere; however, they may induce heat accumulation, anaerobic conditions, and safety concerns associated with gas buildup [20].

Recently, covered manure storage systems integrated with ventilation have been introduced in practice. In such a system, the headspace is sealed to prevent uncontrolled gas release, while low-rate or micro-negative-pressure ventilation continuously extracts air for downstream treatment. The extracted air can subsequently be directed to biofilters, acid scrubbers, or other NH₃ removal technologies. Although such systems offer a promising balance between emission control and ventilation requirements, their design and optimization remain challenging due to the complex interactions among airflow patterns, gas concentration gradients, and environmental conditions inside the covered storage tank.

In the investigation of livestock-related ammonia emissions, both field experiments and numerical simulations have been widely employed. Field measurements are generally costly and time-consuming and are poorly suited for systematic parametric analyses, whereas Computational Fluid Dynamics (CFD) has been extensively applied to investigate ammonia emission and transport processes under various situations. Previous studies have combined CFD simulations with wind tunnel or experimental measurements to examine airflow characteristics, concentration boundary layers, and ammonia mass transfer mechanisms, demonstrating that CFD can reliably predict ammonia emission behavior when appropriate turbulence and mass transfer models are adopted [21,22,23]. CFD has also been employed to analyze scale effects and airflow patterns in manure tanks. The authors pointed out that conventional similarity criteria, specifically the Reynolds number, failed to adequately represent ammonia transport mechanisms, thereby underscoring the necessity of geometric similarity with the full-scale prototype [24]. In addition, CFD studies in livestock buildings have highlighted the critical role of ventilation configuration and airflow organization in controlling ammonia concentration and dispersion [25,26]. Regarding the application of CFD in the investigation of ammonia emissions from temporary manure storage tanks, the available literature remains very limited. Zhao et al. [27] employed CFD to assess ventilation strategies for personnel entry into manure tanks; however, the positive-pressure ventilation adopted in that study discharged contaminated air directly into the atmosphere, making it unsuitable for ammonia emission control or air quality management. To date, CFD studies systematically analyzing NH₃ emission behavior in solid-covered manure storage tanks equipped with negative-pressure ventilation systems have not yet been reported.

Therefore, the present study aims to develop a CFD-based model for analyzing and optimizing NH₃ emission control strategies in a solid-covered temporary manure storage tank. The specific objectives are: (1) to develop and validate a CFD model capable of simulating NH₃ diffusion and transport from a manure storage tank; (2) to evaluate the effect of a solid cover design integrated with a micro-negative pressure ventilation system on air extraction and NH₃ removal; and (3) to explore the influence of different slot configurations on the internal flow field characteristics of the manure tank and ammonia transmission. Specifically, the area-averaged NH₃ concentration at the extraction outlet was adopted as the key performance indicator to quantitatively evaluate the system’s emission control capability under different headspace depths and unintended leakage scenarios (represented by slot configurations).

2. Materials and Methods

Given the complexity of ammonia dispersion in manure storage, which is influenced by multiple interacting factors, including air temperature, air velocity, manure temperature, total ammonia nitrogen in the manure, etc. [9,14], the present study focused primarily on evaluating the effect of different ventilation modes on ammonia dispersion. To simplify the analysis and maintain consistency in the interpretation results, the following assumptions were adopted: (i) the slurry surface was treated as a constant ammonia source with a uniform and time-invariant mass fraction, which has been verified and has been used in multiple ammonia emission-related studies [22,23,24]; (ii) the study does not address the effects of temperature on ammonia emissions; and (iii) the simulation was treated as steady-state.

2.1. CFD Model

2.1.1. Computational Domain

This research considered four simulation scenarios, including one open-field scenario and three solid-covered with negative-pressure ventilation manure storage tank scenarios (no slot, side slot, and top slot configurations). The manure storage tank (excluding manure) was dimensioned as 10 m × 5 m × 2 m (length × width × depth), with its size determined based on field measurements from a commercial pig farm in Hangzhou, China. Three headspace depths, defined as the remaining space above the slurry surface, were considered to correspond to different manure storage volumes: 1.6 m (shallow storage with less manure), 1.0 m (medium storage), and 0.4 m (deep storage with more manure). For the open-field scenario, the manure tank was placed within a computational domain of 100 m × 50 m × 25 m (length × width × height) (Figure 1).

For the solid-covered manure storage tank, the solid cover was enclosed with two circular openings, as illustrated in Figure 2a. One served as the inlet for the ventilation system, and the other one served as the outlet, which was technically connected with the exhausting fan for air extraction. In practical solid-covered manure storage systems, leakage slots may occur due to imperfect sealing at panel joints or structural connections. To account for this, two representative slot configurations were introduced as simplified substitutes: a top slot located at the tank roof, and a side slot located at the junction between the side wall and the top cover, each with dimensions of 100 cm × 1 cm (Figure 2b–d).

To sum up, three solid-covered scenarios were investigated: (i) an idealized fully sealed manure storage tank without any slots (NS), (ii) a tank with only a side slot, and (iii) a tank with only a top slot. For the slot-equipped scenarios, three slot locations were further examined: near inlet (NI), middle (M), and near outlet (NO) (Figure 2e).

2.1.2. Boundary Conditions

In the uncovered (open-field) scenario, the inlet was specified as a velocity inlet with three prescribed velocities (1, 3, and 5 m/s), while the outlet was set as a pressure outlet with a gauge pressure of 0 Pa. The top, bottom, and side boundaries of the open domain, as well as the tank walls, were set as “no-slip wall” without NH₃ mass fraction. The slurry surface was modeled as a no-slip wall with a constant mass fraction of NH₃ of 0.000453 and a constant temperature of 15 °C. For the solid-covered manure storage tank scenarios, the boundary conditions are summarized in

Table 1. Boundary conditions for a solid-covered manure storage tank.

Item	Boundary	Property	Value
Inlet	Pressure inlet	Gauge pressure	0 Pa
Outlet	Velocity inlet	Velocity magnitude	−4 m/s
		Air temperature	22 °C
		NH3 mass fraction	0
Slot (if have)	Pressure inlet	Gauge pressure	0 Pa
Slurry surface	No slip wall	No heat flux	-
	Ammonia release surface	NH3 mass fraction	0.000453
Other walls	No slip wall	No heat flux	-

. The inlet and leakage slots were set as a “pressure inlet” with a gauge pressure of 0 Pa. The outlet was modeled as a “velocity inlet” boundary condition with a prescribed velocity of −4 m/s. The negative sign denoted that the airflow was directed out of the computational domain, representing the exhaust. This setup was adopted because of the requirement of the applied simulation software ANSYS Fluent 2024. Additionally, the reason for setting −4 m/s was to ensure the ventilation system operated under a slight negative pressure (approximately −10 Pa) at the inlet. The setup was validated by preliminary simulations.

2.1.3. Ammonia Mass Fraction Determination

The ammonia mass fraction can be calculated by the following equations [9]:

$$\mathrm{\omega }={\displaystyle \frac{c_{nh3,g}}{{\rho }_m}}\approx {\displaystyle \frac{c_{nh3,g}}{{\rho }_a}}={\displaystyle \frac{\mathrm{T}\mathrm{A}\mathrm{N}·\mathrm{f}}{{\mathrm{K}}_{\mathrm{h}}·{\mathsf{\rho }}_{\mathrm{a}}·{10}^3}}$$

(1)

where c_nh3,g is mass concentration of ammonia gas, kg/m³, ${\mathsf{\rho }}_{\mathrm{m}}$ and ${\mathsf{\rho }}_{\mathrm{a}}$ are the densities of the air–ammonia mixture and pure air, respectively, kg/m³, given the extremely low mass fraction of ammonia in the mixture, ${\mathsf{\rho }}_{\mathrm{m}}$ was assumed equal to ${\mathsf{\rho }}_a$ in this study; TAN is the total ammonia nitrogen, mg/L, $\mathrm{f}$ is the fraction of un-ionized ammonia, which can be determined by Equations (2) and (3), and ${\mathrm{K}}_{\mathrm{h}}$ is the Henry’s constant, the ratio of ammonia concentration in liquid to ammonia concentration in air at the air–liquid interface, which can be determined by Equation (4), proposed by Rong et al. [22].

$$\mathrm{f}={\displaystyle \frac{{10}^{\mathrm{p}\mathrm{H}}}{{10}^{\mathrm{p}\mathrm{H}}+\frac{1}{{\mathrm{K}}_{\mathrm{d}}}}}$$

(2)

$${\log }_{10}{\mathrm{K}}_{\mathrm{d}}=-0.09018-{\displaystyle \frac{2729.92}{{\mathrm{T}}_{\mathrm{m}}}}$$

(3)

$${\log }_{10}{\mathrm{K}}_{\mathrm{h}}=-3.337+{\displaystyle \frac{1977.309}{{\mathrm{T}}_{\mathrm{m}}}}$$

(4)

where K_d is the dissociation constant, T_m is the manure temperature, K.

The ammonia mass fraction of 0.000453 used here was selected from the study by Rong et al. [22], which can also be determined based on Equations (1)–(4) under a TAN concentration of 6800 mg/L, slurry pH of 8.5, air density of 1.15 kg/m³, and slurry temperature of 25 °C.

2.2. Governing Equations

All simulations were conducted with double precision under steady-state conditions. The standard gravitational acceleration was enabled. The steady-state governing equations were based on the Reynolds-Averaged Navier–Stokes (RANS) approach, as shown in Equation (5):

$$\mathrm{d}\mathrm{i}\mathrm{\nu }\left(\mathsf{\rho }\mathrm{\Phi }\mathrm{u}\right)=\mathrm{d}\mathrm{i}\mathrm{\nu }\left({\mathrm{\Gamma }}_{\mathrm{\Phi }}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{\Phi }\right)+{\mathrm{S}}_{\mathrm{\Phi }}$$

(5)

where $\mathsf{\rho }$ is the density, kg m⁻³; Φ represents the common variables of interest, i.e., velocity, m s⁻¹, temperature, K, species, turbulent kinetic energy, m² s⁻², and its dissipation rate, m² s⁻³; u is the velocity, m s⁻¹; ${\mathrm{\Gamma }}_{\mathrm{\Phi }}$ is the transport coefficient dependent on Φ; and S_Φ is the source term dependent on Φ.

The SST k-omega model has been widely regarded as a particularly well-suited model for simulating ammonia emissions from a surface [23,24] and was used in the following simulations. Ammonia diffusion under steady state in the air was governed by the species transport equation:

$$\nabla ·\left(\mathsf{\rho }\overrightarrow{\mathrm{v}}{\mathrm{Y}}_{\mathrm{i}}\right)=-\nabla ·\overrightarrow{{\mathrm{J}}_{\mathrm{i}}}+{\mathrm{S}}_{\mathrm{i}}$$

(6)

where $\mathsf{\rho }$ is the density, kg m⁻³; $\overrightarrow{\mathrm{v}}$ represents the velocity of the diffusing species; Y_i is the mass fraction of the species i, dimensionless; J_i is the diffusion flux of the species, kg m⁻² s⁻¹; S_i is the user-defined sources.

2.3. Simulation Scheme and Convergence Criteria

Velocity-pressure coupling was implemented using the coupled algorithm. Second-order upwind schemes were selected for calculating pressure, momentum, turbulent kinetic energy, turbulent dissipation rate, and energy. Convergence was considered achieved when the following criteria were satisfied: (i) absolute residuals for continuity and x-, y-, and z- velocity components were below 1 × 10⁻³; (ii) the energy residual was below 1 × 10⁻⁶; and (iii) the relative variation in average outlet ammonia concentration over any consecutive 100-iteration intervals within the latest 1000 iterations was less than 0.01%. All simulations were conducted on laboratory workstations.

2.4. Mesh Distribution and Grid Independence Test

The structural meshing method was applied to all cases. A mesh independence test was conducted using a representative configuration consisting of a manure storage tank with a headspace depth of 1.0 m and a top slot located at the middle position. Three mesh densities were evaluated: coarse (~3 million cells), medium (~6 million cells), and fine (~9 million cells). To meet the requirements of the enhanced wall treatment, the y⁺ values were maintained at approximately 1 for all cases.

To evaluate grid convergence, three key parameters were monitored: the outlet ammonia concentration (C_out), the pressure at the top slot (P_t), and the average velocity at the top slot (v_t). Notably, C_out was determined by calculating the area-weighted average of the ammonia concentration across the outlet surface, which was consistently adopted for all subsequent analyses.

Table 2. Mesh independence test results.

Mesh Level	Cout 1 (mol/m3)	Pt 2 (Pa)	vt 3 (m/s)
Coarse	0.0088744	−8.649	3.822
Medium	0.0088739	−8.614	3.814
Fine	0.0088688	−8.618	3.815

¹: the ammonia concentration at the outlet, ²: the pressure at the top slot, ³: the velocity at the top slot.

summarizes the mesh independence results. The relative differences in the monitored parameters across the three mesh densities were approximately 0.4%. The differences between the medium and fine meshes for all three parameters were less than 0.1%. Therefore, the medium mesh was adopted for all subsequent simulations, providing an optimal balance between numerical accuracy and computational efficiency.

2.5. CFD Model Validation

Model validation is essential to confirm the credibility of the CFD model for simulating ammonia emission, diffusion, and transport. The validation process in this study was conducted using experimental data from wind tunnel tests reported in the study [23], focusing on two aspects: turbulence and NH₃ mass transfer. Figure 3 shows the geometric model. The wind tunnel had dimensions of 3.67 m (length) × 0.35 m (width) × 0.35 m (height), with an ammonia release surface of 0.6 m × 0.35 m (Figure 3a) Velocity and NH₃ concentration were measured at discrete heights above the NH₃ solution surface (defined as Z = 0, the liquid surface) at the wind tunnel’s center (Figure 3b,c). The measurement heights included 0.003, 0.005, 0.01, 0.015, 0.03, 0.045, 0.06, 0.1, and 0.16 m, covering the near-wall boundary layer to the core flow region. Specifically, the measurement point at 0.003 m was used for velocity measurement only (Figure 3c). Boundary conditions in the numerical simulations were consistent with those in this study, as summarized in

Table 3. Boundary conditions used in the validation simulations.

Item	Boundary	Property	Value
Inlet	Velocity inlet	Velocity magnitude	0.1, 0.2, 0.3 and 0.4 m/s
		Air temperature	22 °C
		NH3 mass fraction	0
Outlet	Pressure outlet	Gauge pressure	0 Pa
Slurry surface	No slip wall	No heat flux	-
	Ammonia release surface	NH3 mass fraction	0.0007175
Top wall	No slip wall	No heat flux	-
Side wall
Bottom wall

.

The normalized value was used to validate the NH₃ transportation model in this study. Normalized ammonia concentration, C_N, can be expressed as:

$${\mathrm{C}}_{\mathrm{N}}={\displaystyle \frac{{\mathrm{C}}_{\mathrm{p}}-{\mathrm{C}}_{\mathrm{i}}}{{\mathrm{C}}_{\mathrm{o}}-{\mathrm{C}}_{\mathrm{i}}}}$$

(7)

where C_o is NH₃ concentration at the outlet, which was the area-weighted average of NH₃ concentration at the outlet in this study, kg/m³; C_i is NH₃ concentration at the inlet, which was zero because the inlet mass fraction of NH₃ was zero in this study, kg/m³; and C_p is the local NH₃ mass concentration, kg/m³.

Mass concentration was used in this section to ensure consistency with the validation experiment reported in Ref. [23], whereas the subsequent results were primarily reported in terms of molar concentration. The conversion between mass and mole concentrations was performed using the following relation:

$${\mathrm{C}}_{\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}}={\mathrm{C}}_{\mathrm{m}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r}}\times {\mathrm{M}}_{{\mathrm{N}\mathrm{H}}_3}$$

(8)

where ${\mathrm{C}}_{\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}}$ is the mass concentration, kg/m³; ${\mathrm{C}}_{\mathrm{m}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r}}$ is the molar concentration, mol/m³; ${\mathrm{M}}_{{\mathrm{N}\mathrm{H}}_3}$ is the molar mass of ammonia (0.017031 kg/mol).

3. Result and Discussion

3.1. Numerical Model Validation

Figure 4 compares the experimental and numerical results. The close agreement between measured and simulated profiles of air velocity and ammonia concentration indicated the validity of the proposed CFD model. Specifically, the coefficient of determination (R²) value of measured and simulated velocities was 0.9883, while that for normalized ammonia concentration (C_N) reached 0.9975, indicating a strong linear correlation between the numerical results and experimental data. Minor discrepancies in velocity were observed, likely due to instrumental errors and the inherent challenges of near-wall measurements, where small deviations in measurement height can cause significant changes in recorded concentrations [23]. Additionally, the CFD simulation itself may also introduce minor errors due to the simplification. Given that these uncertainties were negligible, the proposed CFD model was deemed sufficiently accurate for simulating NH₃ dispersion in this study and was therefore used in all subsequent simulations.

3.2. Ammonia Emission from the Manure Tank Without a Solid Cover

Figure 5 presents the velocity and ammonia concentration contours of the manure tank under open-field conditions. At a wind speed of 1 m/s, the airstream started to enter the manure tank zone near the middle of the opening, forming a shear layer with the internal airflow there. Higher wind speed caused the air to enter the manure tank earlier and generated a thicker shear layer. Such a phenomenon was consistent with the airflow pattern through parallel slatted floors reported by Qin et al. [24]. Regarding ammonia concentrations, increasing the wind speed from 1 to 5 m/s raised the ammonia emission rate from 181.4 to 397.1 g/h (at 3 m/s) and finally to 604.4 g/h. Additionally, the ammonia concentration near the manure tank on the downwind side can reach 1.2 × 10⁻³ mol/m³, exceeding the human olfactory threshold [28], indicating the necessity of effective mitigation measures for the open temporary manure storage tank.

3.3. Solid-Covered Manure Storages

Figure 6a shows the ammonia concentrations at the outlet of the solid-covered manure storage tank, and Figure 6b illustrates the velocity contour of the horizontal cross-section 0.01 m above the slurry surface under three different depths. As the headspace depth increased from 0.4 to 1.6 m, the outlet ammonia concentration decreased from 9.13 × 10⁻³ to 8.57 × 10⁻³ mol/m³, indicating that cover depth has a clear influence on ammonia emissions. This trend could be attributed to the differences in the near-wall air velocity. As can be observed in Figure 6b, the air velocity at 0.01 m above the slurry surface was higher for 0.4 m headspace than for 1.6 m. Additionally, the smaller sectional area of the 0.4 m headspace case likely contributed to the higher air velocity. Near-wall velocity decreased with increasing tank depth, reducing the ammonia mass transfer coefficient [29].

In addition to the reduced mass transfer coefficient, the increased headspace depth significantly expanded the total air volume within the tank. This larger volume promotes greater dilution and mixing of the ammonia as it is dispersed from the source to the outlet, further contributing to the observed decrease in outlet concentration. This explained why greater tank headspace depths resulted in lower outlet ammonia concentrations.

Regarding the internal airflow pattern, as shown in Figure 6c, at a headspace depth of 0.4 m, the flow was relatively uniform and dispersed from the inlet to the outlet. As the depth increased, the streamlines became more intertwined, forming “vortex-like” flow patterns.

3.4. Effect of Top Slot

Figure 7a shows the outlet ammonia concentration for the solid-covered manure storage tank with a top slot at the middle position under different headspace depths, while Figure 7b presents the air velocity contour at a plane 0.01 m above the slurry surface. As the headspace depth increased from 0.4 m to 1.6 m (specifically 0.4, 1.0, and 1.6 m), the outlet ammonia concentration decreased from 9.91 × 10⁻³ to 8.87 × 10⁻³ and to 8.74 × 10⁻³ mol/m³, respectively. This demonstrates a consistent decreasing trend similar to the no-slot scenario. Compared with the no-slot scenario, all top-slot scenarios exhibited higher outlet ammonia concentrations across all depths. This increase can be attributed to the impinging jet effect of the top slot. The vertically downward airflow from the top slot merged with the main airflow, resulting in a higher near-wall air velocity and thereby enhancing the convective mass transfer of ammonia (Figure 7b). Similar impinging-jet-induced enhancement of near-wall transport has been widely reported in ventilation and heat/mass transfer studies [30,31,32]. Specifically, at a headspace depth of 0.4 m, the top slot contributed to the largest increase in outlet ammonia concentration, from 9.13 × 10⁻³ to 9.91 × 10⁻³ mol/m³, due to maximal enhancement of near-wall velocity at the slurry surface. At greater depths, the downward airflow from the top slot experienced greater velocity attenuation before reaching the slurry surface, resulting in a smaller effect on ammonia emission.

Figure 7c shows the outlet ammonia concentration for three top-slot locations at a headspace depth of 1.0 m. Simulations were performed for all three headspace depths (0.4, 1.0, and 1.6 m); however, as the results exhibited consistent trends across these depths, the intermediate depth of 1.0 m was selected as a representative case for detailed analysis to maintain conciseness. As shown in Figure 7c, increasing the distance between the top slot and the inlet (from positions NI to NO) led to a gradual rise in the outlet ammonia concentration, from 8.38 × 10⁻³ to 8.87 × 10⁻³, and finally to 8.91 × 10⁻³ mol/m³. A possible explanation for this trend was that ammonia accumulated downstream of the primary airflow and beneath the outlet due to convective transport along the main flow path. When the top slot was positioned closer to the outlet, the incoming airflow more effectively disturbed and entrained this locally accumulated ammonia, thereby enhancing its transport toward the outlet.

3.5. Effect of Side Slot

Figure 8a shows the ammonia concentration at the outlet of the solid-covered manure storage tank with a top slot at the middle position under different headspace depths (0.4, 1.0, and 1.6 m). Figure 8b shows the contour of air velocity on a horizontal plane located 0.01 m above the slurry surface. As the headspace depth increased from 0.4 to 1.0 and then to 1.6 m, the outlet ammonia concentration decreased from 9.41 × 10⁻³ to 8.56 × 10⁻³ and finally to 8.32 × 10⁻³ mol/m³, which was consistent with the trend observed in the no-slot scenario. Notably, the introduction of a side slot had a depth-dependent phenomenon; that is, at a headspace depth of 0.4 m, it led to an increase in outlet ammonia concentration compared with the no-slot scenario, whereas at depths of 1.0 and 1.6 m, it resulted in a reduction. This phenomenon could be explained by the fact that, under a constant outlet flow rate, adding a side slot increased the total effective inlet area, reducing both the average inlet velocity and the near-wall airflow velocity above the slurry surface, thereby suppressing ammonia mass transfer. The results for the 1.0 m and 1.6 m cases align well with this explanation. In contrast, at the shallow depth of 0.4 m, the airflow entering through the side slot directly disturbed the near-wall airstream, locally increasing the near-wall velocity (as evidenced in Figure 8b) and enhancing ammonia emission.

Figure 8c illustrates the outlet ammonia concentrations under three side-slot positions at a headspace depth of 1.0 m. The outlet ammonia concentration was 8.23 × 10⁻³ mol/m³ for the near inlet (NI) slot, increased to 8.56 × 10⁻³ mol/m³ for the middle (M) case, and then slightly decreased to 8.44 × 10⁻³ mol/m³ for the near outlet (NO) slot, indicating a non-monotonic dependence of ammonia emission on side-slot location. Although this trend is evident, the underlying flow and mass transfer mechanisms responsible for the observed variation are not fully resolved in the present study and require further investigation.

3.6. Limitations and Perspectives

This study assumes a constant ammonia mass fraction at the slurry surface. In reality, this mass fraction could vary significantly with several factors, such as total ammoniacal nitrogen (TAN), pH, and manure temperature, given the objective of this study, none of which were accounted for in the present study. Additionally, all the simulations were conducted under steady-state conditions; while this treatment was widely applied, admittedly, it possibly neglected many time-sensitive effects. Furthermore, the non-monotonic phenomenon observed in the cases relating to side-slot locations suggested that NH₃ transport was governed by complex interactions between airflow dilution and vortex dynamics that have not yet been fully resolved. Therefore, future work could address these limitations by incorporating unsteady-state simulations, more precise models (e.g., discrete phase modeling), and a dynamic mass fraction of ammonia that accounts for key affecting parameters. Such enhancements would improve the adaptability of the findings to a broader range of real-world environmental conditions.

4. Conclusions

This study focused on ammonia emission characteristics of a temporal manure storage tank equipped with a solid cover and negative pressure ventilation. The following conclusions can be drawn:

(1)

Under open-field conditions, as the wind speed increased from 1 to 3 and then to 5 m/s, the ammonia release rate rose from 181.4 to 397.1 and finally to 604.4 g/h. Furthermore, the downwind ammonia concentration exceeded the human olfactory threshold.

(2)

Within the assumptions of this CFD model, headspace depth appears to be one of the most important parameters affecting ammonia transport, with deeper headspaces consistently resulting in lower outlet ammonia concentrations due to reduced airflow velocity near the slurry surface.

(3)

Top-slot configurations generally increased outlet ammonia concentrations, particularly when positioned closer to the outlet, owing to enhanced near-surface convective mass transfer induced by downward impinging airflow.

(4)

Side-slot configurations exhibited a depth-dependent effect, increasing ammonia emissions at a headspace depth of 0.4 m but reducing emissions at depths of 1.0 m and 1.6 m.