Sei sulla pagina 1di 18
126 xm PoULUTION CONTROL ENGINEERING 6.3 DIFFUSION MODELS In the air pollution literature this next class of models is usually called diffusion mod- els. Most engineers would call them dispersion models because engineers reserve the word diffusion for molecular diffusion, which is not the principal mechanism described by these models. However, the preceding name is so common that it will be used here. 6.3.1 The Gaussian Plume Idea Most diffusion models use the Gaussian plume idea, which also is a material balance model. In it, one considers a point source such as a factory smokestack (which is not really a point but a small area that can be satisfactorily approximated as a point) and attempts to compute the downwind concentration resulting from this point source. The schematic representation and nomenclature are shown in Fig. 6.3, where the origin of the coordinate system is placed at the base of the smokestack, with the x axis aligned in the downwind direction. The contaminated gas stream (normally called a plume) is shown rising from the smokestack and then leveling off to travel in the x direction and spreading in the y and z directions as it travels Such plumes normally rise a considerable distance above the smokestack be- cause they are emitted at temperatures higher than atmospheric and with a vertical velocity. For Gaussian plume calculations the plume is assumed to be emitted from ys z= Oat base of stack FIGURE 6.3 Coordinate system and nomenclature for the Gaussian plume idea. AIR POLLUTANT CONCENTRATION MODELS 127 a point with coordinates 0,0, H, where H is called the effective stack height, which is the sum of the physical stack height (h in Fig. 6.3) and the plwne rise (Ah in Fig. 6.3). Physical stack height for any existing plant can be determined with ordinary measuring instruments. Plume rise is discussed in Sec. 6.4. For the moment we will assume that we are dealing with a point source located at 0,0, H that steadily emits anonbuoyant pollutant at emission rate Q (normally in g/s). Let us assume the wind blows in the x direction with velocity and that this velocity is independent of time, location, or elevation. The problem is to compute the concentration due to this source at any point (x, y, z) for x > 0. If molecular diffusion alone were causing the plume to mix with the surround- ing air, the plume would spread slowly and appear (if the pollutant is visible) as a thin streak moving straight down the sky (see Problem 6.33). The actual cause of the spread of plumes is the large-scale turbulent mixing that exists in the atmosphere, which may be visualized by comparing a snapshot of a plume with a time exposure of the same plume (Fig. 6.4). At any instant the plume will appear to have a twisting, snake-like shape as it moves down the sky. The twisting behavior is caused by the turbulent motion of the atmosphere that is superimposed on the plume's large-scale linear motion caused by the horizontal wind. This turbulent motion is random in nature, so that a snapshot taken a few minutes after the first would show the twists and turns in different places, but the overall form would be similar. However, time averages out these short-term variations of the plume, and thus a time exposure appears quite uniform and symmetrical. For this reason, if we placed a pollutant- concentration meter at some fixed point in the plume, we would see the concentration oscillate in an irregular fashion about some average value. The Gaussian plume ap- proach tries to calculate only that average value without making any statement about instantaneous values. The results obtained by Gaussian plume calculations should be considered only as averages over periods of at least 10 minutes, and preferably one-half to one hour. Snapshot ‘Time exposure FIGURE 64 Comparison of snapshot and time exposure of a visible plume. AIR POLLUTION CONTROL ENGINEERING 128 6.3.2 To derive the Gaussian plume formula, we will first take the viewpoint of a person along with the air, the Lagrangian viewpoint. From this viewpoint, the ground 5 to be passing below, much as the ground appears to be passing below a person long upwind of the stack from which the pollutant is Gaussian Plume Derivation riding appeal in an airplane. We begin ridin, emitted, so we will say that the initial concentration of the pollutant is zero (b = 0) (If b > 0, we must add the value of b to the value calculated here to obtain the best ¢ of the atmospheric concentration.) As we pass directly over the stack we pass into a region of high concentration. This high concentration is localized in a thin thread of contaminated air that passes directly over the stack. After we have passed the stack we will see this thread of contaminated air expand by turbulent mixing. To find out how it expands by turbulent mixing, we will perform a material balance around some small cube of space near the center of the plume. The dimen- sions of this small cube are shown in Fig. 6.5. Let us consider a material that is neither created nor destroyed in the atmosphere, so that the two right-most terms of estim; Eq. (6.1) are zero. The remaining terms are Accumulation all flow all flow ) a of 6.12) Coe) keane aml The accumulation rate is the time derivative of the amount contained, which is the product of the concentration and the volume. But the volume of the cube is not changing with time, so a ae de —(cV) = V— = Ax Ay Az — (6.13) ar ar s no bulk flow (i.e., convection) into or out of the cube we are considering Accumulation rate There because the cube is moving with the local wind velocity. However, there are flows through all six faces of the cube due to turbulent mixing. We do not have a clear and complete physical or mathematical picture of the complex subject of turbulent mixing, but one may approximate it by saying that the flux of material being mixed FIGURE 65 Dimensions of the cube used for the material balance. across any surface is where ¢ = conce n = distar K = turbu Because the flux mo dimensions of mass This dimension is and we will see the conduction or mass the same; rather it s} and mass diffusion: turbulent dispersion concentration to loy transport or K-thee coefficient, K, iso} Our cube has (in Fig looksi looks in the plus» (6.14) twice, we e two faces can bed Net flow into the in the x direct where the first ter second represents Procedure we can 9c/dz and two im six faces by turbu the accumulation analogous ones ir