more than 90 percent of the sulfur. Cost estimates at present indicate that for a 500-mw plant burning coal at 10 percent ash and 3 percent sulfur, about 36 cents per ton of coal burned would be required.

C. RECENT DEVELOPMENTS

1. A modification of the alkalized alumina process uses sodium aluminate in lieu of alkalized alumina. This change permits use of smaller equipment and saves materials handling costs because the aluminate is more efficient in removing sulfur dioxide from stack gases. Sodium aluminate is commercially available. According to current studies, sodium aluminate absorbs 0.34 gram of sulfur dioxide per gram of absorbent, compared to 0.25 gram of absorbed sulfur dioxide per gram of alkalized alumina. The amount of removal is 85 percent of the stoichimetric amount for the aluminate and 65 percent of the stoichimetric amount for alkalized alumina.

It should be possible to regenerate and recycle the reacted sodium aluminate and recover sulfur in the regeneration step. Fly ash from coal-fired plants does not hinder absorption efficiencies of sodium aluminate even when present in amounts equal to the amounts of the absorbent.

2. Economic data on the molten carbonate process indicates that sulfur recovery credits would probably offset operating costs. Both capital and operating costs for the molten carbonate process are lower than corresponding costs for the alkalized alumina, catalytic oxidation, or Reinluft processes.

Bench-scale tests of the molten carbonate process show greater than 99 percent removal of sulfur oxide from gas streams containing between 0.3 and 3.0 percent sulfur dioxide. The hydrogen sulfide product can be converted into elemental sulfur or sulfuric acid depending on nearby marketing conditions. A small pilot plant to test the process under continuous operating conditions is in the design stage.

1. INTRODUCTION TO METEOROLOGI

CAL DISCUSSION

The atmosphere is the medium that transports and disperses air pollution. If there is a source such as a power plant emitting sulfur dioxide, meteorology determines concentrations at exposed receptors.

This discussion will be limited to the meteorological aspects of air pollution from coal-burning power plants without any control of SO, emissions. Primarily, a nominal 1,000-mw plant size will be considered. It will be assumed that if a plant site is unsuitable for a 1,000-mw plant, it is also unsuitable for plants of larger size. The siting of nuclear plants will not be considered as these are adequately treated in Chapter III." The discussion will also be limited to the problems of gaseous pollutants from the power plants as it will be assumed available technology will be used for the effective removal of particulates.

2. SOURCE CONDITIONS

A nominal 1,000-mw power plant will emit approximately 500 tons of SO, per day, assuming that it will burn about 9,000 tons of coal per day containing 3 percent sulfur, and 90 percent of the sulfur is converted to SO2, which is emitted from the stack. The most important factors in determining emission rate are the amount of fuel burned and its sulfur content. Ranges of sulfur contents of coal used in the United States have been given as 0.5 to 4.5 per

(109)

cent by weight for bituminous and 0.4 to 0.8 percent for anthracite (1).

Emission factors determining stack plume elevation above grade are physical stack height, stack shape (or diameter of the stack opening), and velocity and temperature of the effluent at the top to the stack.

Trends in steam station design indicate that the most likely physical stack height for a nominal 1,000-mw power plant is 800 feet above grade (2), although there exist 1,000-mw power plant units with 500 foot stacks and much smaller units with stacks higher than 800 feet. In hilly terrain it is conceivable that stacks as much as 1,200 or 1,500 feet in height may ultimately be required (3). In other countries large power plants have been designed with multiflue stacks (4,5). However, in the United States new, or proposed, stacks are of the simplest design, with the least possible restrictions to air flow through the flue and out of an uncomplicated circular opening at the top. Stack diameters would be expected to be about 30 feet or more in order to conduct about 3 million cubic feet per minute of stack effluent at velocities of approximately 75 feet per second, full load. Higher velocities, say 90 to 120 feet per second, will normally increase the fan and motor size and cost, and chimney maintenance costs, and may inhibit plume rise due to buoyancy (6).

Effluent temperatures decrease as power plant unit size increases. A nominal 1,000-mw power plant would be expected to have an effluent temperature of about 250° F. (7). ·

3. METEOROLOGICAL FACTORS

The meteorological factors affecting plume rise are wind speed, wind shear, air temperature, air pressure, and stability. In various equations (8, 9, 10) for estimating the height of rise of a plume from a stack the plume rise is inversely proportional to some power of the wind speed. That is, the higher the wind speed, the less the plume rise. Plume rise equations become useless as the wind speed approaches zero. Figure 1 shows plume rise for several TVA steam plants with respect to wind speed (11). The plume rise equations, which assume uniform atmospheric conditions, are unreliable with light winds when stable layers at altitudes above the top of the stack inhibit plume rise, or when there is appreciable wind shear. Air temperature and air pressure determine the air density. The effect of differences in air pressure is relatively insignificant, but the effect of differences in temperature on plume buoyancy is considerable. It follows then, that significantly higher stack plumes would occur in colder air. When the temperature decreases with altitude (the lapse rate) by 5.4° F. per 1,000 feet, or nearly so, the atmosphere is said to have "neutral" stability. If the temperature decrease is greater, the conditions are unstable. A volume of air which is given a vertical motion under unstable conditions will be accelerated upward or downward. On the other hand, if the temperature decrease is less than for the neutral condition; the conditions are stable, and a volume of air which is displaced will tend to return to its original level.

The available plume rise equations take into account the expected damping effect of stable conditions, but do not agree with respect to the effect of instability on the height of plume rise. The well-known equation of Holland (12) suggests that a value between 1.1 and 1.2 times the height of the plume rise for the neutral condition should be used for the unstable conditions, and between 0.8 and 0.9, for stable conditions; however, recently proposed equations (13) suggest that there is no difference in the plume rises for neutral and unstable cases. The assumption is that the turbulence is more vigorous in unstable than in neutral conditions and that plume rise may even decrease lightly with increasing

110

instability. The need for additional study of the effect of stability is indicated by the Tilbury studies (4), which have shown that plume rise is not significantly affected by lapse rate: the potential decrease in altitude during inversion conditions, due to the temperature gradient, tends to be compensated by the lack of turbulence and slower spread of the plume.

Plume rise is also affected by aerodynamic factors, the merging of plume from multiple stacks, convection from the heat of the plant (in the case of large power plants) and indirectly by the physical height of the stack itself. When a stack plume is emitted in a disturbed air flow caused by wind blowing over structure or irregular terrain, the standard plume rise may not apply. Wind tunnel studies (14-17) are useful when information cannot be otherwise deduced.

Additional buoyancy results when two or more plumes merge from stacks that are close together, and the resulting larger plume will usually have a greater rise. Extension of any formula to multiple stacks is speculative but some helpful data are available from TVA (6, 11) and other sources; however, much of the data are in terms of resulting ground-level concentrations rather than plume behavior itself.

Thomas, Carpenter, and Gartrell (6) report that when thermal convections are present, the heat emission from a large power plant or similar energy source may initiate thermals and, in fact, establish a semi-independent local atmospheric circulation. In this circumstance, the plume may rise several thousand feet and often produces cumulus clouds.

Stone and Clark (4) state, "an important effect-previously suspected and now substantially confirmed-is that plume rise is itself dependent on the height of the stack. This arises from the observation that turbulence decreases with height above ground level (particularly at heights over 300 feet) which leads to slower mixing and increased rise of the plume."

In simplifying the treatment of dispersion, it is often realistic to assume that atmospheric dispersion begins at a height above the actual stack top. This height (h) is called "the effective stack height" and it is the sum of the actual

stack height (h,) and the rise of the plume after emission (Ah), as shown in figure 2, due to inertial and buoyancy force. Most of the time effective stack heights for a nominal 1,000 mw power plant will be higher than 1,000 feet above grade, and effective heights of 2,000 or 3,000 feet would not be unusual. Generally, atmospheric conditions that would affect the plume are significantly different from the meteorological conditions measured near the ground by a Weather Bureau station. With level terrain, maximum ground concentrations decrease as the effective stack height increases, and the point at which the maximum occurs moves farther from the base of the stack.

The basic factors considered when estimating downwind (18, 19) concentrations from a stack are the emission rate, effective stack height, wind speed, and atmospheric stability (which determines the rate of diffusion). As discussed earlier, a higher wind speed decreases the effective stack height while at the same time it also increases dispersion as shown in figure 3. Therefore, there is a critical wind speed at which maximum ground concentrations occur; this wind speed varies with physical stack height and plume rise parameters.

When making diffusion estimates, it is usually assumed that the wind speed is constant with height. However, roughness elements on the ground engender mechanical turbulence, which affects wind speed profile as shown in figure 4.

Thermal turbulence also affects wind speed. Usually in the daytime, upward motions transfer the momentum "deficiency" due to any friction losses near the earth's surface through a relatively deep layer, consequently the wind speeds at the effective stack height are normally greater at night than during the day as shown in figure 5.

Atmospheric stability categories may be determined from the temperature lapse rate, which may be measured by balloon-borne or tower mounted instruments, or determined from the effects of turbulence on any of several types of instruments exposed to the wind, such as a wind vane (20).

In the absence of measurements, stability conditions may be estimated. Pasquill (18, 21) has provided a key to stability categories that

111

is based on surface wind speed, the amount of incoming solar radiation, and cloud conditions. Turner (22) has fitted these categories to the ESSA-Weather Bureau airport station surface observations on punch cards available from the National Weather Records Center, Asheville, N.C.

Instability enhances dispersion, particularly in the vertical direction; whereas the effect of a stable vertical temperature stratification is to decrease the horizontal eddy diffusion and practically eliminate the vertical eddy diffusion.

Figure 6 shows a stack emitting a plume within a temperature inversion layer based at the surface. The temperature increases with height in the layer and the plume exhibits practically no vertical diffusion. Such a condition, called "fanning" is most likely to occur in rural areas during a night when the sky is clear and winds are light. The effluent trail may be narrow, widening gradually on a straight line from the stack, or it may resemble a meandering river.

Depending on the duration of the stable period and the wind speed at the effective stack level, the effluent may travel for many miles with very little dilution. However, during the following morning after the ground has been heated by the sun, air near the ground will be warmed and become turbulent so that parts of the plume will be carried rapidly to the ground as shown in figure 7. This condition, called "fumigation" results in relatively high ground concentrations, lasting for periods ranging from a few minutes to an hour or more. Actual measurements of concentrations from large power stations under fumigation conditions will be discussed later.

Early field observations of fumigation from stacks of moderate size, approximately 200 to 500 feet high, were made by Hewson and Gill (23), Church (24), meteorologists at the AEC National Reactor Testing Station (25), Lowry (26), and others. Methods of calculating inversion breakup fumigation concentrations have been given by Hewson (27), Turner (8), and the ASME (10).

A type of fumigation can also occur in a limited mixing layer with light winds when the effluent is released and contained within a small