Improving the Performance of Houston's Southwest Wastewater Treatment Plant, U.S.A.

Abstract

In 1981, the City of Houston's Southwest Wastewater Treatment Plant (144,000 m3/d) was experiencing occasional afternoon bulking with loss of solids to the effluent. The plant employs sludge reaeration with about even allocation of tank volume to mixed liquor and return sludge aeration. A dynamic model of solids transfer from influent to mixed liquor to settling tank to return sludge aeration and wasting was developed for use on a desk-top computer. The model predicted that a counter-intuitive reduction in the return sludge rate would redistribute the solids to less in the mixed liquor and more in the sludge reaeration tanks, but would also generate a sludge blanket during the transition period. However, if the change were made when there would be several low-flow days, then the blanket would stay within the settling tanks. The Thanksgiving Holidays of 1981 were chosen for the change in the return sludge rate. Low flow occurred as expected and there was no loss of solids. The effluent suspended solids improved as expected, but the SVI also improved from 172 ml/g to 72 ml/g. The reduction in the return sludge flow rate allowed extra hydraulic capacity for storm flows. The improved control of solids was demonstrated by simulating a storm flow of 4.0 m/h overflow rate by taking settling tanks out of service with resulting effluent TSS of less than 20 mg/L.

Sludge Blanket Response to Storm Surge in an Activated‐Sludge Clarifier

“…The model of Garrett et a!. (1984) was based on overflow and underflow clarifier rates and solids settling velocity.…”

“…This concentration is much greater than would be predicted by the exponential expression. This was the primary reason for the selection of the power expression, although the comparison was not presented in the paper of Garrett et al (1984). Rachwal et al (1982) and Smollen and Ekama (1984) concluded that the exponential expression gives a better correlation between consideration of these opinions, Garrett et aI.…”

“…In 1982, tests of peak~ow to the final clarifiers were conducted at an operating 4730-m 3 /1i (30-mgd) activated-sludge plant (Southwest WWTP) in Hou&'tbn and reported by Garrett et al (1984). Tests included overflow rates as high as 4.06 m!h (2400 gpd/sq ft).…”

“…The settling velocity of an undiluted sample was correlated with the diluted 30-minute settled sludge volume in an unstirred I-L graduated cylinder. For return sludge, the OSV30 was determined as DSVI X C, where C is the solids concentration in percent.At the time of the report by Garrett et al (1984), principal models used to relate the settling velocity to solids concentration were the power model used by Yoshioka et al ( 1957) and Dick and Young (1972) and the exponential model used by Vesilind (1968). These models, respectively, are as follows:ABSTRACT: Aspects of operation and design for stonn flows at a 190-m 3 /h (1.2-mgd) dry weather design flow activated-sludge plant were tested.…”

“…At the time of the report by Garrett et al (1984), principal models used to relate the settling velocity to solids concentration were the power model used by Yoshioka et al ( 1957) and Dick and Young (1972) and the exponential model used by Vesilind (1968). These models, respectively, are as follows:…”