Characteristics of Unorganized Hydrogen Sulfide Dispersion for Industrial Building Layout Optimization

Abstract

Hydrogen sulfide (H₂S) is the main toxic pollutant emitted to the atmosphere from auto-coating wastewater. Its unorganized dispersion poses a health challenge for workers. Defining safe working distance, which transfers the H₂S occupational exposure limit into industrial construction design regulation, would be a useful approach for reducing H₂S exposure risk. Therefore, in this study, an H₂S dispersion prediction, within 25 m, was performed by a computational fluid dynamics (CFD) method to explore the influence of temperature and wind speed on H₂S dispersion. With the temperature changes from 288 K to 303 K, the H₂S concentration at different observing points decreased. With wind speed changes from 2 m/s to 20 m/s, the plume layer structure was studied in the whole process. According to the H₂S distribution characteristics, when the sedimentation tank treatment capacity is less than or equal to 10 m³/h, the safe working distance of H₂S unorganized dispersion is 10 m. Hence, when there are workplaces within 10 m of the tank, closed measures should be taken for the sedimentation tank, or the manufacturer layout should be optimized to protect the environment and human health.

1. Introduction

Hydrogen sulfide (H₂S), a toxic gas with a rotten egg smell, widely exists in industrial production processes [1]. In fact, H₂S is commonly found in industrial sites such as geothermal power plants, livestock farms, paper mills, pharmaceutical plants, sewage treatment plants, landfills, and automobile factories [2,3,4], being a potential health hazard for workers. Therefore, analyzing different types of H₂S dispersion in different industrial plants is significant.

The H₂S potential health risks are divided into acute toxicity (high H₂S concentration) and chronic damage (low H₂S concentration) [5]. Mature operation specifications and safety measures have been conducted for acute poisoning prevention, including carbon-based material adsorption, metal catalytic oxidation, electrochemical treatment, and biological treatment [6,7]. Although the majority of H₂S would be systematically treated [8], there is still a considerable amount of H₂S that would be freely emitted to the atmosphere, due to concentration differences, pressure differences, or other reasons. As a result, many industrial sites are chronically filled with low-concentration H₂S. It was illustrated that chronic exposure to low levels of H₂S is associated with increased mortality and morbidity for respiratory diseases, disorders of the peripheral nervous system, heart failure, and diseases of the veins [9,10]. Therefore, workers who are exposed to low levels of H₂S for a long time are at great risk of H₂S chronic damage to their health.

Occupational exposure standards for low-concentration H₂S have been set. According to WHO, having about 7.5 mg/m³ H₂S in the atmosphere, people start to have discomfort symptoms, such as eye irritation and bad breathing [11]. The Chinese occupational exposure limits for hazardous agents in the workplaces stipulates that the maximum allowable concentration of H₂S is 10 mg/m³ [12]. Although the limits are specific, there are still big difficulties in implementing the policy, due to the lack of effective control measures of low-concentration H₂S in workplaces [13]. In this case, defining a safe working distance, which transfers the H₂S occupational exposure limit into industrial construction design regulation, would be a useful approach for reducing the potential H₂S exposure risk. Modeling the dispersion of low-concentration H₂S in different industrial application scenarios becomes an economical and efficient way to visualize H₂S dispersion. The simulation models would be conducive to avoiding potential health risk areas with high H₂S concentration, optimizing plant layout and ventilation design, and reducing the health damage of H₂S to humans.

Current studies on outdoor low-concentration H₂S dispersion simulation mainly treated a factory as a point source of H₂S, using the air pollution model aermod or calpuff, combined with a GIS (Geographic Information System) system [14,15,16], to investigate the effect of H₂S on the air quality of the surrounding area and the health effects on nearby residents [17], rather than workers, who stay much closer to the source. The dispersion distribution of H₂S was simulated at the kilometer scale [18]. However, the accuracy of the two models is not suitable for simulating low-concentration H₂S on the meter scale. For distances from 50 m to 250 m, the CFD method [19] was used to analyze high-concentration H₂S leakage events, which indicated that the CFD method could be suitable for short-distance and low-concentration simulation.

At present, no simulation study about short-distance unorganized H₂S dispersion has been reported to define the safe working distance. Low-concentration H₂S is still a hidden danger for workers. Therefore, it is necessary to study the unorganized H₂S dispersion for occupational exposure in a short distance, smaller than 25 m, from the source (inside manufacture) by CFD method. Based on the gas chromatography–mass spectrometry (GC–MS) method, using a multi-component gas analyzer, the main odorous component of automobile coating wastewater was analyzed to be H₂S. Therefore, an outdoor automobile coating sewage sedimentation tank is taken as the simulation scenario in this study to establish a dispersion model of low-concentration H₂S. The model explores the influencing factors and dispersion characteristics of low-concentration H₂S in this scene.

With the research gaps mentioned, the aim of this study is to conduct dispersion model of unorganized H₂S and define a proper safe working distance. The implementation steps are: first, select the proper models for surface source no-barrier H₂S dispersion that closely match the experimental measurements; then, analyze the wind speed and temperature influences on short-distance outdoor H₂S dispersion; finally, acquire the safe distances of different wind speeds and temperature. This study will clearly illustrate the risks of unorganized H₂S no-barrier dispersion from sewage sedimentation tank to human and help to create a safe working distance for H₂S exposure management.

2. Materials and Methods

2.1. Description of the Study Area

We studied an auto parts paint washing sewage sedimentation tank in 20 m × 10 m size, in Jinzhou city, Hubei Province (Figure 1a) and built a 3D sewage tank model (Figure 1b). In production processes, H₂S from auto parts paint dissolved in washing wastewater, conveyed to the sewage pool for centralized treatment. The sewage pool could be regarded as a surface emission source of odorous pollutants in a wide-open outdoor space. The odor of H₂S would affect the working environment. Note that the wastewater sedimentation tank, processing 10 m³/h, discharged regularly in work time. There was no barrier tall building near the sewage pool, except a small office and a sidewalk 4 m and 11 m away the sewage pool, respectively.

2.2. Experimental Measurement

Sampling dates were collected in winter 2020 and summer 2021, respectively. Three observing points (Figure 1d) were set at 0 m, 4 m, and 11 m, called P1, P2, and P3, to monitor H₂S concentration by the H₂S detector (SGA-608, SINGOAN). Active sampling by vacuum pump can quickly carry out real-time online detection of multi-component factors in the environment. We only needed one key to start the machine, and could take the initiative to sample and display the current measured value. The H₂S detected time of each observing point lasted 5 min, and the concentration average values are displayed in Figure 2. Meanwhile, the environmental parameters, including temperature and wind speed, were measured in continuous 1 min in both measurements. Additionally, the Cartesian coordinates of the three monitoring points were (10 m, 0 m, 1.5 m), (14 m, 0 m, 1.5 m), and (21 m, 0 m, 1.5 m), respectively. The measurement details and results were listed in

Table 1. Collection parameters of each measurement in experimental scene.

Measurement No.	Collection Date	Sampling Locations	Temperature	Wind Speed
1	21 September 2020	P1, P2, P3	302.5–303.3 K	3.2–4.6 m/s
2	8 March 2021	P1, P2, P3	287.8–288.4 K	15.8–16.2 m/s

.

2.3. Parameters Calculation

On the basis of double membrane theory, the process of H₂S release can be divided into three parts, which are liquid-phase, gas–liquid interphase, and gas-phase mass transfer. The sewage pool was assumed to have an interphase between liquid and gas phases. The three phases and the two films between them formed a five-layer structure to complete the three procedures mentioned above. All the phases and films were considered to be in a dynamic equilibrium. There were three assumptions in this theory: firstly, the interphase was assumed to have no mass transfer resistance; secondly, the sewage was regarded as H₂S saturated solution; thirdly, H₂S concentration in gas phase was zero. Besides, it is known that interphase mass transfer should obey Henry’s law. According to the setting above, the following equations were listed:

N = N_L = N_G

(1)

N = D (c_L – c_G)

(2)

c_i = Hp_i

(3)

N = D (c_L – Hp_G)

(4)

where N is the total mass transfer rate, N_L is the mass transfer rate in the liquid phase, N_G is the mass transfer rate in the gas phase, D is the mass transfer coefficient, c_L is the species concentration in liquid phase, c_G is the species concentration in gas phase, c_i is the concentration of species i in liquid phase when it balances with the material partial pressure p_i in the gas phase, H is the Henry’s law constant.

The formular of mass transfer coefficient is:

$$D=\frac{435.7T^{3/2}\sqrt{\frac{1}{M_A}+\frac{1}{M_B}}\times {10}^{-4}}{P{\left[{(\sum A)}^{\frac{1}{3}}+{(\sum B)}^{\frac{1}{3}}\right]}^2}$$

(5)

where T is the environment temperature, M is the relative molecular mass, $\sum A$ and $\sum B$ are the molar volume per gram, P is the environment pressure.

The main purpose of this study is to investigate the H₂S flow characters in the air. A computational volume of 40 m × 30 m × 3 m in x, y, z directions was established, with the volume center at origin point (x = y = z = 0 m) (Figure 1d). The computational domain included the clean air velocity inlet (blue arrow), the H₂S mass inlet (yellow arrow), and four mixture flow outlets (left, right, opposite, and up). The H₂S mass inlet was on the bottom surface and H₂S initially diffused upward. The model structure was shown in Figure 1c.

For a flowing H₂S computational cell, the governing equations are mass and momentum conservation. The mass conservation is reflected by the species transport equations, based on the assumptions above:

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

(6)

where in a turbulent flow

$$J_i=-(\rho D_{i,m}+\frac{{\mu }_t}{S{c}_t})\nabla Y_i$$

(7)

where t is time, ρ is fluid density, Y_i is the mass fraction of species i, u is the mean instantaneous velocity, J_i is the diffusion flux of the species i, D_i,m is the diffusion coefficient of species i in the mixture, Sc_t is the turbulence Schmidt number generally being equal to 0.7, and μ_t is the turbulence viscosity. The odor dispersion is dependent on the species gradient, the diffusion coefficients, and the turbulence viscosity [20]. Using Reynolds-averaged Navier–Stokes (RANS) method would be a good solution for the equations system with the two governing equations. CFD software offers several turbulent models. It has three different types, including realizable, RNG (Re-normalization group), and standard. These models perform differently in different cases. So, it is recommended that different studies choose their suitable models according to their own circumstances [21,22]. Using simulation results (linear fitting curves) of wind speed at 4 m/s and 16 m/s to match the onsite measured data is a reasonable method for estimating the suitable turbulence model.

2.4. Simulation Scheme and Boundary Conditions

For outdoor dispersion, the environment conditions would be unstable. Wind speed might be the dominant factor affecting outdoor H₂S no-barrier dispersion in a short distance [23]. Ana et al. [24] noticed that, at the same monitoring point, the summer H₂S concentration was much higher than winter. It was assumed that wind speed affected H₂S unorganized dispersion by changing turbulent state and mass transfer. Additionally, temperature mainly changed H₂S diffusion by affecting mass transfer through temperature difference. Herein, this study would verify these two factors on H₂S dispersion, respectively.

Control variable method was adopted to design the simulation plan. To analyze the effect of temperature, we controlled the wind speed at 4 m/s and changed the environment temperature from 288 K to 303 K. Observing the H₂S concentration change could help us understand the influence of temperature on H₂S dispersion. Similarly, for wind speed study, we kept environment temperature at 288 K and changed the wind speed from 2 m/s to 20 m/s to illustrate the influence on H₂S distribution. So, here is a list of all the environmental conditions we need to simulate (

Table 2. H₂S emission rate at different temperatures and wind speeds.

Wind Speed (m/s)	Temperature (K)	H2S Mass Flow Rate (kg/s)
1.0	288	1.50
1.5	288	1.51
2.0	288	1.62
4.0	288	2.18
	293	2.23
	298	2.29
	303	2.35
6.0	288	2.62
8.0	288	3.07
10.0	288	3.49
12.0	288	3.88
14.0	288	4.25
16.0	288	4.61
18.0	288	4.95
20.0	288	5.29

). The H₂S emission rates for model simulation at different temperatures and wind speeds are also shown below.

Except taking wind speed and temperature as variables, turbulent intensity and turbulent viscosity ratios also need to be calculated as inputs in k-ε model. The first step is to calculate the Reynolds number and the turbulent intensity. The second step is to acquire turbulent viscosity ratio from turbulent kinetic energy k and its dissipation rate ε. The turbulent viscosity ratio increases with the Reynolds number. The boundary conditions change with the wind speed. To build the H₂S dispersion model, the assumptions taken are as follows. First, the flow is considered to be three-dimensional turbulence. Second, the wind speed value is considered to not vary with height, and the wind is parallel to the ground. Third, gauge pressures of all pressure outlets are set as 0. Additionally, the CFD simulation solver in this study was PISO (Pressure-Implicit with Splitting of Operators), which is a pressure-based solver.

3. Results and Discussion

3.1. Subsection H₂S Concentration Results and Model Validation

The results of H₂S concentration measured on site are shown in Figure 2. Because of the sedimentation tank without any waste gas treatment equipment, the high concentration value was not out of expectation. At 1.5 m height, the H₂S distribution of two measurements were similar, and the H₂S concentration decreased with the distance increase away from the H₂S source. However, the environment conditions of the two measurements were different, and the effect of temperature and wind speed on the H₂S dispersion needed deeper analysis.

Before simulating, a suitable model needed to be ensured. About k-ε model selection, Muhammad et al. [25] found that, in the gas dispersion process of hydrogen, the turbulence model influence was very small for high Reynold number regions, such as the air injection source, but in the region far away from the injection source, where turbulence was weaker, the influence was larger. It meant that the choice of different k-ε models would have a great impact on the simulation results for turbulence with a small Reynolds number. They reported that the RNG and the standard models provided a better estimate of the region where turbulence is fully developed. To validate this opinion and find out the appropriate model for H₂S dispersion, realizable, RNG, and standard turbulence models were selected to simulate H₂S dispersion at 4 m/s and 16 m/s, because the simulation data could be fitted in six lines (Figure 3) to compare with the onsite data in Figure 2. There was a low accuracy for the standard model because the gap between the curves and the measurement points were large, since, for 4 m/s, the analogue results were generally lower than the field measurements, but for 16 m/s, they were higher than the field measurements. On the contrary, the performances of realizable were better in a high wind speed condition, but its results error of 4 m/s was larger than standard. The realizable model was still not accurate enough. The concentration curves of the RNG model under 4 m/s and 16 m/s had better down-trend synchronicity and were close to the measurement points. The result errors of RNG were less than 15%.

From the comparison above, it can be seen that there were significant differences between the three models, with RNG performing the best. Meanwhile, the applicability of the RNG model for the outdoor dispersion simulation of gaseous pollutants has been fully verified in the studies about dispersion around buildings [26,27,28]. Herein, the RNG model was chosen to simulate under various temperature and wind speeds.

3.2. H₂S Dispersion Simulation with Different Temperature

From the mass transfer coefficient formula (5), it was known that H₂S emission increases with temperature rise. Although, for the sewage sedimentation tank, the increase in H₂S production due to temperature rise is small [29]. Besides, Archana and Melanie [30] considered that the increase of H₂S production had little effect on the H₂S dispersion movement. The studies above illustrated that the temperature change would affect the H₂S concentration in two aspects: the production and the dispersion. As temperature rises, the production and the dispersion of H₂S would be enhanced simultaneously. Therefore, in order to further eliminate the interference, the H₂S mass flow rate was fixed at 2.18 kg/s (environment condition: 4 m/s and 288 K) to analyze the effect on dispersion caused by the enhancement of the molecular thermal.

The H₂S concentration decreased with the temperature rise in Figure 4, which meant the enhancement effect of high temperature on dispersion is greater than that on production. There were more H₂S coming out than coming in at the observing points. This was the reason for the H₂S concentration decrease. However, the differences were small, in a range from 0.16 to 0.21 ppm. The ratio of mean deviation to mean value of H₂S concentration reflected the temperature influence on H₂S dispersion (

Table 3. Mean value, mean deviation, and ratio of mean deviation to mean value of H₂S concentration with different temperatures at P1, P2, and P3.

H2S Concentration (ppm)	P1	P2	P3
Mean value	7.88	3.48	1.45
Mean deviation	0.055	0.063	0.051
Ratio	0.7%	1.8%	3.5%

). Among the three monitoring points, the ratio was no more than 3.5%, which explained that the annual temperature change had little effect on H₂S dispersion. So, the big differences during the two measurements at the three points were mainly due to wind speed.

3.3. H₂S Dispersion Simulation with Different Wind Speeds

Having wind speeds from 1 m/s to 20 m/s as the variable, the simulated H₂S plumes were shown in Figure 5 and Figure 6. The legend shows the H₂S concentration fraction of the initial concentration Q, which was the initial H₂S concentration above the sewage pool closing to the water surface.

The change of H₂S concentration fraction revealed the dispersion tendency of H₂S. Generally, with wind speed increasing, H₂S concentration declined with height increase in y direction, influenced range enlarged with the increase of distance in x direction, and the H₂S plume eventually stabilized at a limit height about 3 m during the wind speed changes. The whole plume shape was consistent with the report of Abdullah and Weiming [31] about leaked gas having the same, or similar, density as ambient air. For more details, the research range was divided into two parts. Region 1 was the range above the sewage pool (x coordinate from −10 m to 10 m), and region 2 was the rest (x coordinate from 10 m to 25 m).

Region 1 mainly observed the wind speed influence on H₂S emission. It was evaluated by the H₂S high concentration range (H₂S concentration above 0.5Q) change. Before 2 m/s, the turbulence caused by wind was not strong enough to produce a big mass transfer coefficient for strong mass transfer. So, the H₂S high concentration range was small. At the outlet of the tank, the H₂S concentration reached 0.9Q. That is because the velocity ratio between odorous compounds velocity at the source outlet and wind speed was small [32]. The H₂S compounds tended to disperse vertically upward. This characteristic was also noticed in Figure 7, where, before 2 m/s, the H₂S concentrations at P1, P2, and P3 were close. However, from 4 m/s, the stronger turbulence improved the mass transfer, and the vertical dispersion was improved. Comparing the H₂S plumes of different wind speeds in Figure 5 and Figure 6, the H₂S emission was strengthened with wind speed, increasing from 2 m/s to 10 m/s. High wind speeds let the ambient air have more power to carry the H₂S climbing up, presenting a larger influence height.

Surprisingly, the growth trend of the high-concentration H₂S range in region 1 changed in high wind speed condition. When the wind speeds were faster than 10 m/s, the high-concentration H₂S influence range shrank. This phenomenon was studied in depth, and the reason was found to be the change of flow field (Figure 8).

In the three cases (4 m/s, 10 m/s, and 18 m/s) in Figure 8, the slope of streamline above region 1 was positive, which meant air flowed upwards. The slope was minimum at 4 m/s and maximum at 10 m/s. The slope of 18 m/s was in the middle, which was slightly smaller than 10 m/s (Figure 9). The characteristics of air flow were exactly the reason why the influence range of high-concentration H₂S first increased and then decreased.

Region 2 mainly observed the influence of wind speed on the H₂S dispersion. When the wind speed was under 2 m/s, due to the small amount of H₂S emission, the whole region 2 was H₂S low-concentration range (H₂S concentration under 0.5Q). H₂S tended to disperse perpendicularly to the jet direction because the horizontal velocity was also small. Additionally, the edge of the plume was unsmooth. With wind speed augmenting to 4 m/s, the H₂S concentration changed quickly. This concentration rapidly changed, due to wind speed rise, which was also illustrated in Shen’s research [33]. It was displayed by the concentration gradient in Figure 7. Anyway, the plume edge began to be smooth. It meant that more H₂S dispersed down-wind, instead of flying up, which increased the H₂S fraction in region 2. With the further increase of wind speed, the height of the influence range gradually rose again and stabilized at about 3 m under 8 m/s. After 10 m/s, the H₂S influence range was basically unchanged. This phenomenon was also found in the outdoor H₂S dispersion study of a landfill [34]. The whole H₂S influence range fluctuation in Figure 6 and the streamlines state in Figure 8 explain that, as the distance between the observation point and the leak source increased, the influence of air flow on the distribution of H₂S became more obvious, which was consistent with the report of Majid et al [35].

Meanwhile, the H₂S plume structure was also changed. With the increase of wind speed from 4 m/s to 12 m/s, the thickness of each concentration layer increased. However, this trend reversed from 14 m/s. As the wind speed continued to increase, the layers whose fractions were higher than 0.5Q became thinner, and other layers whose fractions were smaller than 0.5Q became thicker. The H₂S distribution changed because of the flow field change, which was the same as region 1 (Figure 9). In the three cases, focusing on the slopes of streamlines in region 2, it was noticeable that the changes in characteristics were similar to the slopes in region 1. Using k to represent the slope, we could obtain that k₄ < k₁₈ < k₁₀. The change of streamline slope conducted the H₂S layer thickness change.

3.4. Safe Distance for Outdoor Sedimentation Tank

In certain industries, such as landfills, sewage treatment plants, and animal husbandry, workers face a higher risk of H₂S exposure. Therefore, H₂S occupational exposure is highly regarded, and relevant laws and regulations have been formulated in various countries [36,37].

As mentioned above, China specifies a H₂S maximum allowable concentration (MAC), which is 10 mg/m³. Uniformly converting the units of H₂S MAC to ppm for data analysis, 10 mg/m³ corresponded to 6.9 ppm. The initial measured max concentration Q was 13.34 ppm. So, the MAC of H₂S equaled 0.54Q. Hence, an area having H₂S concentration fraction higher than 0.54Q could be regarded as the unsafe region.

The area bounded by a 0.54Q curve and the x-axis was an unsafe region in Figure 10. The unsafe region of 2 m/s was about 11 m long and 0.5 m high, which was the distance away from the office monitoring point P2. As for lower wind speed, such as 1 m/s and 1.5 m/s, whose unsafe region were supposed to be smaller than 2 m/s, being similar with the H₂S plume change in region 1, the unsafe region first developed and then shrank with wind speed increase. According to the onsite data, mostly the wind speed in Jingzhou city was under 8 m/s, whose corresponding safe region was 17.2 m long and 1.5 m height. This is the most basic distance for workers to be safe, but the general safe region was supposed to be larger than 17.2 m. Under 12 m/s, the unsafe region reached its maximum range, which was 19.5 m long and 1.7 m height. With wind speed higher than 12 m/s, the unsafe region began to shrink.

Anyway, analyzing the partial enlarged drawing in Figure 10, it was found that the slope of H₂S concentration under different wind speeds had the same characteristics as the streamlines of different wind speeds. The unsafe region border slope before 10 m/s was always greater than the slope after 10 m/s, which meant H₂S dispersed further down-wind after 10 m/s. It was concluded that, under the worst environment conditions of 12 m/s, the safe distance was over about 20 m long and 1.8 m height. Therefore, the office at point 2 was suitable for workers only when the wind speed was less than 2 m/s. Additionally, the sidewalk at point 3 could be used as a daily road. In summary, for this model scenario, a work office was better be set at least 10 m away from the wastewater pool (down-wind direction). The adjusted sewage pool model is shown below in Figure 11.

4. Conclusions

For surface source unorganized H₂S dispersion modeling, RNG is the most suitable k-ε turbulence model that most closely matches the experimental measurements. Using the RNG model to explore the influences of temperature and wind speed on H₂S dispersion, the results show that temperature change indeed leads to different H₂S production and different H₂S concentration at the monitoring points, but the effect is much smaller than wind speed for H₂S dispersion. The different wind speed changes the flow field. Taking the wind speed of 12 m/s as the dividing line, H₂S dispersion shows two states: enhanced (less than 12 m/s) and unchanged (more than 12 m/s).

In this case, the CFD simulation provides a visualization for H₂S dispersion at short distance outdoors. Between 288 K to 303 K, the H₂S concentration outside the sewage sedimentation tank does not reach the occupational exposure limit when the wind speed is less than 2 m/s. However, when the wind speed increases from 4 m/s to 12 m/s, the safety distance increases from 7 m to the maximum of 10 m. Therefore, for the sewage sedimentation tank with sewage treatment capacity less than or equal to 10 m³/h, it is recommended that workshops or offices be built 10 m away from the pollution source. This guidance would have value for practical industrial environment layouts.