Document Type


Publication Date

January 1984


The direct filtration water treatment scheme does not include sedimentation and in some cases flocculation. Compared to conventional treatment, direct filtration has lowered capital costs, reduced space requirements, decreased sludge quantities, and reduced coagulant dosages. One objective of this research was the statistical comparison of the direct filtration, Utah Valley Water Purification Plant (Orem, Utah) and the conventional Little Cottonwood Water Treatment Plant (Salt Lake City, Utah). These treatment plants are the two most compatible treatment plants having the highest correlation of source water in the local area. The Little Cottonwood plant receives approximately 65 percent of its sources water from Deer Creek Reservoir whereas the Utah Valley plant receives all of its source water from Deer Creek Reservoir. Data from August 1, 1980, through August 31, 1983, were obtained from the daily water quality and plant operation logs of the two treatment plants. Utilizing the computer, the data were clocked into season means and compared staistically in several fashions. The water quality parameter of most importance in the comparison is the finished water turbidities. The most benficial results were obtained from a two-way analysis of variance using an F-ratio as the reference for signifiicance. The F-ratio for the finished water turbidity at the degree of significance, alpha = 0.01, proved not significant. The overall statistical analysis exhibits that the Utah Valley plant produces not only acceptable finished water turbinities well below EPA's maximum contaminant level of 1 TU, but one that is also comparable in quality to that of the conventional processes of the Little Cottonwood Treatment Plant. Another objective of this research was the operation of two pilot-scale direct filtration systems at the Utah Valley treatment plant. The pilot plant treated the same source water and used the same dual-filter media as the Utah Valley treatment plant. The pilot plant flow processes consisted of a rapid mix basin, a flocculation basin, and the filter column. Filters were evaluated by the filter performance index (F.P.I.), which is based on the quantity of turbidity removed, the volume of water produced during a filter run, and effluent quality. The highest F.P.I. values were achieves during the filter runs using alum as the primary coagulant and a cationic polymer as a coagulant aid. These filter runs produced a product water with finished water rubidities considerably below the EPA maximum contaminant level of 1 TU. The most successful filter runs were characterized with filter loading rates ranging from 3 to 5 gpm/ft^2, and alum and polymer dosages ranging from 2.1 to 8.1 mg/l and 0.70 to 2.2 mg/l, respectively. These filter runs treated raw water with average turbidities from 13.6 to 22.8 NTU.