Modeling of Fugitive Dust Emission for ... - ACS Publications

However, a construction sand and gravel processing plant is often a major source of air pollution, due to its associated fugitive dust emission. To pr...
0 downloads 0 Views 91KB Size
Environ. Sci. Technol. 2001, 35, 2073-2077

Modeling of Fugitive Dust Emission for Construction Sand and Gravel Processing Plant C H I N G - H W A L E E , * ,† L I - W E N T A N G , † A N D C. T. CHANG‡ Department of Environmental Engineering, Da-Yeh University, 112, Shan-Jiau Road, Da-Tsuen, Chang-Hwa, Taiwan, R.O.C., and Department of Environmental Engineering, National I-Lan, Institute of A and T, I-Lan, Taiwan, R.O.C.

Due to rapid economic development in Taiwan, a large quantity of construction sand and gravel is needed to support domestic civil construction projects. However, a construction sand and gravel processing plant is often a major source of air pollution, due to its associated fugitive dust emission. To predict the amount of fugitive dust emitted from this kind of processing plant, a semiempirical model was developed in this study. This model was developed on the basis of the actual dust emission data (i.e., total suspended particulate, TSP) and four on-site operating parameters (i.e., wind speed (u), soil moisture (M), soil silt content (s), and number (N) of trucks) measured at a construction sand and gravel processing plant. On the basis of the on-site measured data and an SAS nonlinear regression program, the expression of this model is E ) 0.011‚u2.653‚M-1.875‚s0.060‚N0.896, where E is the amount (kg/ ton) of dust emitted during the production of each ton of gravel and sand. This model can serve as a facile tool for predicting the fugitive dust emission from a construction sand and gravel processing plant.

Introduction In Taiwan, a typical process for a construction sand and gravel processing plant (CSGPP) involves gravel transportation, crushing, screening, washing, storage, and hauling. During the operation of the plant, a high concentration of dust which seriously damages the surrounding air quality is often produced; hence it is worthwhile to understand the air quality and the dust emission characteristics during the operation of a CSGPP. A local CSGPP was selected to be studied for this purpose. On the basis of on-site observation at this CSGPP, the parameters of soil moisture, soil particle size, wind speed, and traffic flow were found to be the main factors affecting the dust emission during operations. Another important reason for choosing the aforementioned parameters is that they can be easily measured on-site in any CSGPP. Thus, the on-site data of wind speed, soil moisture, soil silt content (-75 µm), number of trucks, and total suspended particulate (TSP) are measured in this local plant. These measured data are intended to be used to determine the relationship between the dust emission quantity and the four aforementioned parameters. In this study a quantitative * Corresponding author phone: 886-4-853-2835; fax: 886-4-8531157; e-mail: [email protected]. † Department of Environmental Engineering, Da-Yeh University. ‡ National I-Lan, Institute of A and T. 10.1021/es001237y CCC: $20.00 Published on Web 04/14/2001

 2001 American Chemical Society

FIGURE 1. Operation flowsheet for selected construction sand and gravel processing plant. model was developed in this study to describe this relationship.

Background of Selected CSGPP A typical construction sand and gravel processing plant located in Central Taiwan was selected for investigation in this study. This local CSGPP processes mainly gravel and sand excavated from a nearby river. The excavated gravel is transported by truck to this plant. The incoming trucks unload the river gravel onto a bar screen which can eliminate giant stones. The undersized material is treated by a trommel screen to dispose of the larger stones. The remaining material is then subjected to a series of screening, crushing, conveying, and washing devices to obtain several grades of clean products. The final products produced by this plant are principally different sizes of gravel or sand. There are five piles of different sized products which can be observed at this plant. These products can easily be sold for civil engineering applications in Taiwan. Figure 1 displays the operation flowsheet for this plant, and Figure 2 shows the plant layout drawing. On the basis of an on-site observation, the main factors affecting the dust emission behavior of this plant are summarized below. (1) Truck dumping: When the incoming trucks unload and dump river gravel onto the bar screen, a noticeable fine dust is emitted. (2) Mechanical treatment: The dumped river gravel and sand is subjected to a series of mechanical treatments such as screening, crushing, and conveying. These mechanical treatment processes often produce a large quantity of fine dust. (3) Product loading: The final clean products are temporarily stored in different piles on the ground and then loaded onto the trucks by a front-end loader. The movement of this loader and the accompanying process often create dust emission. (4) Wind erosion: Most of the VOL. 35, NO. 10, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2073

FIGURE 2. Layout of selected CSGPP and location of sampling points. ground area of this plant is unpaved. When the unpaved ground and sand piles are subjected to natural wind, fine dust is emitted into the air. (5) Vehicle movement: Both loaded and empty trucks moving around on the unpaved ground soil also cause a fine dust emission. (6) Moisture of sand and soil: The high moisture content of sand piles and unpaved ground soil obviously minimize dust emission, an effect which can be clearly noticed during a rainy day. (7) Particle size of sand and soil: The high concentration of fine-sized particles contained in the sand pile and ground soil also result in a high concentration of fine dust emission.

Development of Mathematical Model The main purpose of this study is to develop a mathematical model to predict the amount of fine dust which can be emitted from a CGSPP under different operating conditions. The first step is to identify the parameters which must be included in the model. The considered parameters must be able to represent the dust emission behavior of the plant. Several predictive emission factor equations can be found in the U.S. EPA’s AP-42 (1) and other literatures (2-13). No exact emission factor equation can be extracted from this literature (1-13) to predict the dust emission behavior for Taiwan’s CSGPP. However, in the AP-42, the form of the predictive emission factor equations for an “aggregate handling and storage pile” can be used as a basis for the development of the CSGPP’s dust emission model. The form of this equation is given below (1)

u ( 2.2) E ) k(0.0016) (M2 )

1.3

1.4

(1)

where E is the emission factor (kg/ton), k is the particle size multiplier (dimensionless, for TSP, k ) 0.74), M is the material moisture content (%), and u is the mean wind speed (m/s). Equation 1 is used to predict the quantity of particulate emissions generated by dropping the material onto a receiving surface during the aggregate storage activities. For 2074

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 10, 2001

emissions from equipment traffic (trucks, front-end loaders, dozers, etc.) traveling between or on piles, the AP-42 recommends the following equation for vehicle traffic on an unpaved surface (1)

E ) k(1.7)

W (12s )(48S )(2.7 ) (w4 ) (365365- p) 0.7

0.5

(2)

where E is the emission factor, kg/VKT, k is the particle size multiplier (dimensionless, for TSP, k ) 0.8), s is the silt content (-75 µm) of road surface material (%), S is the mean vehicle speed (km/h), W is the mean vehicle weight (ton), w is the mean number of wheels, and p is the number of days with at least 0.254 mm of precipitation per year. On the basis of an on-site observation of the dust emission behavior in this CSGPP and the form of eqs 1 and 2, the parameters of soil moisture, soil particle size, wind speed, and traffic flow were found to be the main factors affecting the dust emission during the operation hours of the plant. Thus, these four parameters were selected to represent the dust emission behavior of the plant in this study. The dust emission factor (E) can be calculated by measuring the onsite total suspended particulate (TSP) data. The TSP as measured by a standard high-volume air sampler has a relatively coarse size range. An effective cutting point of 30 µm aerodynamic size is frequently assigned to the standard high-volume sampler (1). Another important reason for choosing the aforementioned parameters is that they can be easily measured on-site in any CSGPP. Thus, the on-site data of wind speed (u), soil moisture (M), soil silt content (s) (-75 µm), number of trucks (N), and total suspended particulate (TSP) are measured in this local plant. On the basis of the above discussion, the final form of the mathematical model for predicting the amount of fine dust which can be emitted from the plant is developed below:

E ) a‚ub‚Mc‚sd‚N f

(3)

Here, E is the emission factor (i.e., the quantity of dust emitted from each ton of gravel and sand produced, kg/ton) and can be calculated from the on-site TSP, wind speed, and

production rate data. The parameters of a, b, c, d, and f in eq 3 are dimensionless constants. These constants can be obtained by using the nonlinear regression method to fit the actual measured data of wind speed (u), soil moisture (M), soil silt content (s) (-75 µm), and number of trucks (N) with the calculated value of E. The calculation of the aforementioned parameters will be discussed later in this study.

Experiments TSP. On the basis of the wind direction exhibited at this local CSGPP, four TSP sampling points (i.e., A, B, C, and D) were chosen to set up the TSP equipment to measure the on-site data. The location of these sampling points are shown in Figure 2. In this study, the TSP data were measured by using the U.S.A. Graseby/GMW high-volume sampler, and all were measured during daytime operation hours at the local plant. The duration of each TSP measuring period was exactly 1 h. A total of 27 TSP measuring periods was conducted in this study. Each measuring period had four sets of TSP (i.e., sampling point A, B, C, and D) on-site data. Moisture. During each TSP measuring period, the ground soil near the TSP sampling points and the materials from the five product piles (see Figure 2) were sampled to measure their moisture content. A total of nine samples were taken for moisture measurement during each measuring period. The average moisture content of these nine samples was taken to represent the moisture content of the material in the measuring period. The moisture content (M) of each sample was determined by the following equation:

M)

W1 - W 2 × 100% W1

where W1 is the original material weight and W2 is the material weight after drying at 100 °C for 2 h. Silt Content. The samples obtained for moisture measurement were also used for silt content determination in this study; thus, nine samples were taken for silt measurement during each TSP measuring period. Each obtained sample was dried at 100 °C for 2 h. Then, each dried sample was treated with a dry-screening device (Retsch Co., Model: AS 200 digit) with a 200 mesh (i.e., 75 µm) screen for 30 min to determine the amount of material having a size less than 75 µm. The shaking speed of this screening device was set at 2880 vibrations/minute for the screening tests. After drying and screening, the weight of the material was measured to calculate the silt content. The average silt content of these nine samples was taken to represent the silt content of the material in this measuring period. The silt content (s) of each sample is determined by the following equation:

s)

W200 × 100% W

where W is the material weight after drying at 100 °C and W200 is the weight of material passing 200 mesh (75 µm) screen after drying at 100 °C. Wind Speed. A wind-speed measurement instrument (Davis Instrument Co.) was set up at the downwind sampling point to measure the on-site wind-speed, wind-direction, temperature, humidity, and pressure data during each TSP measuring period. Such data can be downloaded by pclink3 software at intervals of one minute. The average wind-speed data from these downloads were taken to represent the wind speed in each measuring period. Number of Vehicles. On the basis of on-site observation, the main vehicles causing dust emissions appear to be trucks unloading feed materials and trucks loading end-products. Both kinds of trucks have similar hauling capacities, and their wheel numbers range from 10 to 14. The vehicle speed

TABLE 1. On-Site Measured TSP Data TSP measuring period no.

A (µg/m3)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

489.64 396.44 379.31 271.08 196.68 197.55 97.10 82.06 59.31 202.78 173.48 151.82 152.05 108.85 155.59 237.60 200.37 260.51 258.12 347.01 338.12 394.45 457.12 273.33 371.21 161.72 233.42

sample point B (µg/m3) C (µg/m3) 970.94 558.76 579.99 367.75 175.88 257.08 119.98 95.59 74.57 154.37 137.51 139.04 122.38 79.10 142.87 127.74 330.20 450.00 414.56 572.07 308.74 294.47 372.82 242.67 347.34 134.59 191.88

505.58 369.91 478.13 677.42 1206.19 458.21 230.20 265.26 250.54 534.18 483.45 467.10 388.61 324.15 133.74 189.65 541.44 489.12 919.93 1532.83 1048.38 2528.58 1770.46 579.39 357.92 471.28 594.78

D (µg/m3) 1297.19 870.09 1864.18 538.12 508.63 380.81 125.24 88.27 100.02 317.18 174.70 135.43 186.16 120.96 180.18 208.24 274.82 440.41 828.04 1152.48 677.36 1614.58 2123.44 476.77 413.36 159.84 298.34

of both types of trucks is about the same, since they always move in a routine route inside this CSGPP. Thus, the factor of vehicle traffic speed and number of wheels has been omitted, and only the number of trucks was considered in this study. During each TSP measuring period, the number of trucks unloading and loading at this plant was recorded.

Results and Discussion A total of 27 TSP measuring periods were conducted in this study. The duration of each period was exactly 1 h. Within each measuring period, there were 4 TSP measured data points, 9 moisture-content analysis data points, 9 silt-content analysis data points, 60 wind-speed measured data points, and 1 truck-number recorded datum. These four TSP data were measured at the sampling points A, B, C, and D, respectively, which are tabulated in Table 1. On the basis of an on-site observation and measurement of wind direction, the average TSP data measured at sampling points A and B can be considered as the upwind TSP concentration for this CSGPP, whereas the average TSP data measured at sampling points C and D were considered as the downwind TSP concentration. The net TSP concentration of this local plant during its operation hours is equal to the downwind concentration minus the upwind concentration. The width of this CSGPP, which is perpendicular to the wind direction, is equal to approximately 70 m. The maximum height of the storage piles inside the plant is about 5 m. Assume that there is no significant dust emission above a height of 5 m in the plant and that within these 5 m the TSP concentration is uniformly distributed without any change. Thus, the effective air volume (i.e., containing the emitted fine dust) which passes through the plant every hour is equal to

u (wind speed, m/s) × 60 s × 60 min × (70 m × 5 m) The hourly production of sand and gravel at this CSGPP is nearly 35 tons. On the basis of the aforementioned information, emission factor E can be calculated by the following VOL. 35, NO. 10, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2075

FIGURE 3. A comparison between actual calculated and predicted emission factors.

TABLE 2. On-Site Measured M, u, s, and N Data and Calculated Values of E

TABLE 3. SAS Nonlinear Regression Program

TSP measuring period no.

calcd E (kg/ton)

moisture, M (%)

wind speed, u (m/s)

silt content, s (%)

no. of vehicles, N

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

0.0059746 0.0031271 0.0191685 0.0092389 0.0149796 0.0100326 0.0041591 0.0042739 0.0020281 0.0177031 0.0058742 0.0029172 0.0016633 0.0044901 0.0000918 0.0003867 0.0030855 0.0063865 0.0439362 0.0610411 0.0295183 0.1616580 0.0683876 0.0050559 0.0014237 0.0096423 0.0074104

2.67 1.96 1.82 1.95 3.01 4.46 5.16 6.06 8.47 8.48 6.19 6.33 8.14 4.27 4.17 3.93 3.66 4.19 6.12 3.27 3.29 2.96 2.43 2.85 3.66 4.37 3.62

0.97 0.61 0.77 0.89 0.62 1.45 1.67 1.35 0.52 1.99 0.94 0.52 0.28 0.97 0.33 0.66 0.60 1.62 2.27 1.92 1.52 2.60 1.24 0.52 1.50 1.60 0.88

0.52 0.43 0.31 0.45 0.82 1.17 1.35 0.99 0.88 0.42 0.20 0.32 0.38 0.44 1.06 0.45 0.56 0.56 0.43 0.88 1.42 1.02 1.19 1.54 0.98 0.82 1.50

6.00 9.00 21.00 11.00 18.00 13.00 9.00 15.00 14.50 5.00 12.50 11.00 9.00 9.00 11.00 5.50 9.50 5.00 8.50 10.50 9.50 12.00 17.00 18.00 14.00 10.00 11.00

equation

tsp × 70 × 5 × u × 60 × 60 E) 35 × 109

(4)

where E is the emission factor, (i.e., the quantity of dust emitted during the production of each ton of gravel and sand, kg/ton) and tsp is the net TSP concentration in each TSP measuring period.

[

(point D TSP + point C TSP) 2 (point A TSP + point B TSP) (µg/m3) 2

]

Here u is the average wind speed in each TSP measuring period (m/s). The calculated E value as well as the measured mean wind speed (u), mean moisture content (M), mean silt content 2076

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 10, 2001

(s), and number (N) of trucks for each TSP measuring period are presented in Table 2. The relationship between the calculated E value and the parameters of u, M, s, and N can be obtained by using a nonlinear regression technique to fit the actual measured data (see Table 2) to eq 3. In this study a nonlinear regression program in the SAS package, version 6.12 (i.e., a statistical package provided by Da-Yeh University), was developed to obtain the dimensionless constants a, b, c, d, f shown in eq 3 and listed in Table 3. On the basis of this program and the input data given in Table 2, the bestfitting result shows that the values of constants a, b, c, d, and f are 0.011, 2.653, -1.875, 0.060, and 0.896, respectively. Thus, the final form of this developed model for calculating the quantity of dust which can be emitted from a CGSPP under different operating conditions (i.e., u, M, s, and N) is given below:

E ) 0.011‚u2.653‚M-1.875‚s0.060‚N0.896

(5)

The fitting result of eq 5 gives a mean square deviation (MSD) of 0.000096, where

MSD )

1 n-1

Acknowledgments

n



concluded that eq 5 can reasonably represent the actual dust emission behavior of the plant.

(Ep - Ec)i2

i)1

where n is the total number of TSP measuring periods, Ep is the E value predicted by eq 5, and Ec is the E value calculated by eq 4. On the basis of eq 5 and the actual measured data of u, M, s, and N shown in Table 2, the predicted emission factor (Ep) for each TSP measuring period is calculated and plotted in Figure 3. For comparison, the actual calculated emission factors (Ec) versus the number of measuring periods are also plotted in Figure 3. It can been seen that the general trend of actual calculations of E can be well predicted by eq 5. However, for some data points, there is a noticeable error between the Ec and Ep. The accuracy of this developed model (i.e., eq 5) can be improved by obtaining more on-site measured data of u, M, s, and N to fit eq 3. The positive values (i.e., b ) 2.653, d ) 0.060, and f ) 0.896) shown in eq 5 mean that an increase in wind speed, silt content, and number of trucks will increase the quantity of dust emitted from the CSGPP, whereas the negative value of c (-1.8785) in eq 5 reveals that a high soil-moisture content results in a low concentration of dust emission. The description stated above coincides with the on-site observation of the dust emission behavior of the plant. Thus, it can be

The financial support of the Department of Environmental Protection of Nantou County, the Republic of China, is gratefully acknowledged. The authors also wish to express appreciation to Dr. Cheryl Rutledge for her editorial assistance.

Literature Cited (1) U.S. EPA. Compilation of Air Pollution Emission Factors (AP42); U.S. EPA: Research Triangle Park, NC, 1996. (2) Kleinman, M. T.; Pasternack, B. S.; Eisenbud, M.; Kneip, T. J. Environ. Sci. Technol. 1980, 14, 62-66. (3) Nicholson, K. W. Atmos. Environ. 1993, 27A, 181-188. (4) Clausnitzer, H. J. Environ. Qual. 1996, 25, 4, 877-884. (5) Kulshrestha, U. C. Atmos. Environ. 1996, 30, 24, 4149-4154. (6) Midwest Research Institute. Characteristics of Mud/Dirt Carryout onto Paved Roads from Construction and Demolition Activities; U.S. EPA-600/R-95-171; EPA: Research Triangle Park, NC, 1996. (7) Liu, F.; Hong, Y. J. Aerosol Sci. 1996, 27, 1, S93-S94. (8) Venkatram, A., et al. Atmos. Environ. 1999, 33, 1093-1102. (9) Xuan, J. Atmos. Environ. 1999, 33, 1767-1776. (10) Xuan, J., et al. Atmos. Environ. 2000, 34, 4565-4570. (11) Beer, T., et al. Math. Comput. Modelling 1995, 21, 9, 131-135. (12) Callot, Y., et al. Geodinamica Acta 2000, 13, 245-270. (13) Venkatram, A.Atmos. Environ. 2000, 34, 1-11.

Received for review May 9, 2000. Revised manuscript received January 29, 2001. Accepted February 6, 2001. ES001237Y

VOL. 35, NO. 10, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2077