By early 1968, maximum ground-level SO concentrations for two-unit operation were established and extrapolated to the three-unit level. Results indicated that at full-load three-unit operation, ground-level SO2 concentrations would exceed the TVA threshold of an 0.8-ppm 30-min average SO2 concentration during certain meteorological conditions.

2) Critical Plume Dispersion Model The meteorological and atmospheric plume dispersion conditions associated with the maximum ground-level SO2 concentrations were also identified. Without exception, the highest concentrations occurred with the limited layer mixing, or trapping, model.25 The criticality of this model, as applied to large power plant operations, was first realized by TVA in 1968 by the analysis of air monitoring network data from the TVA Bull Run Steam Plant (950-mw unit with 244-m stack). The data showed that the maximum

ground-level SO, concentrations were significantly higher than those estimated for coning dispersion, which was then considered by TVA as the principal plume dispersion model for its power plants. Prior to limited layer mixing dispersion, the stack effluent usually rises well above a surface-based inversion and is embedded in stable air aloft. Daytime thermally induced mixing subsequently builds upward through the plume and uniformly stirs the effluent to the ground. The Bull Run data further revealed that there was a principal meteorological synoptic pattern identified with limited layer mixing-type dispersion-anticyclonic circulation within the lower few thousand meters with pronounced stability and weak-to-moderate mean wind speed.

Limited layer mixing dispersion is most critical at large plants with tall stacks such as the Paradise plant for the following reasons:

1. Ground-level SO, concentrations can reach significantly higher levels than during coning dispersion.

2

2. High stacks offer material advantages for most meteorological conditions including those causing coning and inversion breakup dispersion; however, they are a relatively minor determinant of ground-level SO, concentrations during limited layer mixing dispersion. Under this type of dispersion, maximum plume height, which is directly related to ground-level concentration, is primarily determined by the depth and degree of stability of the stable air aloft rather than stack height. However, higher stacks can reduce the frequency of high groundlevel concentrations from limited layer mixing dispersion if they are tall enough to force a significant portion of the plume into that stable air that does not undergo temperature transition.

3. Maximum ground-level SO2 concentrations from limited layer mixing dispersion at Paradise may persist from two to five hr, beginning between 9 and 11 local time, and these concentrations occur relatively close to the plant, approximately three miles. The plume centerline ground-level concentration, x, identified with limited layer mixing dispersion, is expressed as:

where:

10°Q
X=
√20uH,

(ppm)

(1)

Q=SO2 emission rate (m3/sec)

y=standard deviation of limited layer mixing plume crosswind normal distribution (m)

u=mean horizontal wind speed between stack top and plume top (m/sec) H, height of mixing layer (m)

Equation (1) is used to calculate an allowable SO2 emission rate at Paradise to ensure that an 0.8-ppm 30-min average ground-level SO2 concentration is not exceeded.

3) Meteorological and Plume Dispersion Criteria Nine meteorological and plume dispersion criteria were established to determine the critical atmospheric conditions when plant operations could cause a 30-min average ground-level SO, concentration of 0.8 ppm to be exceeded. Parameter values for the criteria were developed during the 1969 preoperational field studies from meteorological measurements in the field when ground-level SO2 concentrations, projected to threeunit plant full-load operation, exceeded the 30-min average concentration of 0.8 ppm. To require plant generation load reduction, all criteria requirements must be met. The nine criteria, with respective parameter values, are shown in Table I.

The SO, emission limitation program at the Paradise Steam Plant is in effect every day throughout the year. On a day of implementation, the atmospheric dispersion conditions for the period 9 to 2 local time (LT) are checked by onsite meteorological measurements taken at 7-7:30 LT. Temperature profile data are collected by a recording instrument mounted in a fixed-wing airplane as it ascends to about 2,000 meters over the plant area. Wind profile measurements are also taken concurrently near the plant site from surface to about 2,000 meters, using the standard pilot balloon method.

These data, along with sky condition and maximum surface temperature, are processed by the power plant computer for scheduling the SO2 emission limitation, beginning about 9 LT and ending no later than 2 LT. The computer determines whether all parameters of the nine meteorological and plume dispersion criteria fall within the critical range of values. If so, the magnitude of SO2 emission limitation, or plant generation load reduction, is determined from the computer program-using the limited layer mixing plume dispersion model, Equation (1).

During each SO, emission limitation implementation period, the on-site meteorologist maintains close surveillance over the meteorological and plume dispersion conditions and notifies the plant staff should conditions become favorable to resume normal plant operations. At about 11 LT, the meteorologist checks conditions with wind speed profile measurements and cloud cover observations. If the wind speed or percent cloud cover fall out of the critical range, the SO emission limitation requirement is canceled.

PROGRAM RESTRICTIONS

There are two power system operational conditions in which the Paradise Steam Plant would not reduce generation load below a minimum level during a designated SO2 emission limitation period. These restrictive conditions became effective when further limitation in load would cause power system instability and risk a failure in the supply of firm power to customers.

PROGRAM EFFECTIVENESS

Since the beginning of the SO2 emission limitation program in September 1969 through November 30, 1972 (39 months), there have been 106 days when the meteorological conditions would have required some SO2 emission limitation if the plant had been operating at or near full capacity. However, due to unit outage or below full-load operation, there were only 41 days when generating load had to be reduced. The magnitude of the load reductions ranged from 26 to 960 mw with an average of 454 megawatts. The average duration of load reductions, generally between 9 and 2 LT, was 3.6 hr with a minimum of 0.4 hr and a maximum of 5.8 hr (Table II).

TABLE II.-SUMMARY OF SO2 EMISSION LIMITATION CONDITIONS, SEPTEMBER 1969 THROUGH
NOVEMBER 1972, PARADISE STEAM PLANT

An examination of the SO2 data from the Paradise Steam Plant air monitoring network before and after the implementation of the SO2 emission limitation program demonstrates that the program has been very effective. Frequency distributions of measured three-hr and 24-hr average SO, concentrations before and after the implementation of the SO2 emission limitation program are given for fourteen SO, monitoring instruments in Table III. Both periods include approximately 21 months of equivalent record. The "after" period extends only through June 25, 1971, since some of the SO, monitoring instruments were removed from the network after that time. The most significant feature of Table III is the comparative frequencies, before and after, of the SO; concentration classes which represent levels exceeding the national ambient air quality standards for SO. With respect to the three-hr average standard, 0.5 ppm, Table III shows that before the SO2 emission limitation program was initiated, there were ten incidents when measured SO2 concentrations exceeded the 0.5 ppm level during the 21-month period. After the SO, emission limitation program was initiated, the frequency of 0.5 ppm three-hr average SO, concentrations dropped to only two during the subsequent 21-month period. Likewise, for the 24-hr average SO2 standard of 0.14 ppm, the before and after frequency of occurrences exceeding the standard were eight and zero, respectively. The two "after" SO concentrations exceeding the 0.5 ppm three-hr average level occurred at two separate monitoring sites, whereas the ten "before" similar concentrations occurred at the rate of more than one per

TABLE III.-FREQUENCY DISTRIBUTION OF MEASURED SO, CONCENTRATIONS BEFORE AND AFTER IMPLEMENTATION OF SO2 EMISSION LIMITATION PROGRAM, PARADISE STEAM PLANT

Therefore, Kentucky's ambient SO2 standards were exceeded before but not after the initiation of this program. These data are even more impressive considering the fact that during the 30-month period preceding startup of Unit 3 and the initiation of the program, the average SO2 emission was 8.65 kg/sec, whereas during the first 39 months following the implementation of the program, the SO2 emission rate was 13.0 kg/sec-an increase of 50 percent. Therefore, this indicates that noncontinuous SO2 emission limitation programs can be used at large coal-fired power plants as an effective method for meeting required ambient SO2 standards.

PROGRAM IMPROVEMENTS

With present program procedures, plant generating load reductions are initiated about 9 LT regardless of the meteorological and plume configuration conditions at that time. A new procedure is being planned so that load reductions can be initiated with more precise consideration given to the time when the thermally induced mixing reaches the plume. Basically, this time will be estimated by predicting the rate of change of the surface temperature by use of an emperically derived diurnal temperature curve and the adiabatic projection of the temperature onto the morning aircraft temperature profile.

A second program improvement involves the augmentation of the existing air monitoring network to provide improved surveillance of maximum ground-level SO2 concentrations, particularly during periods of SO2 emission limitation. Should the data from this network indicate that the applicable ambient standards are being exceeded, immediate steps would be taken to adjust the threshold level in the program to further reduce SO2 emissions to ensure that the air quality standards would not be exceeded. Plans are also under way to provide the on-site meteorologist with mobile SO, sampling facilities which can be used for maintaining further surveillance over ground-level SO2 concentrations in the power plant area during periods of SO2 emission limitation.

OTHER TVA 802 EMISSION LIMITATION PROGRAMS

A brief description of the SO2 emission limitation programs for the Widows Creek and Cumberland Steam Plants, as well as a résumé of study plans for six other plants, follows:

WIDOWS CREEK STEAM PLANT PROGRAM

As discussed previously, the current SO2 emission limitation program at the Widows Creek Steam Plant was initiated in September 1970. The program was further augmented in August 1972, when a full-time onsite meteorologist was