Optimum Design and Operation of Diatomite Filtration Plants

Abstract

The primary objective of this study was to develop a digital‐computer program for use in designing diatomite filtration plants that would produce filtered water of requisite quality at least cost. To achieve the objective, it was necessary to be able to predict the variation of filter cake resistance for various conditions, because cake resistance is one of the primary factors influencing costs. Diatomite filtration equations that were theoretically developed from the generally accepted filtration rate equation were used to describe the head loss‐time relationships during formation of diatomite filter cakes. A computer program was developed for determining the optimum operating conditions for a specific type of plant that filters a particular water and uses a specified grade of filter aid. This program was named the Program for Optimization of Plant Operation (POPO). The program determines the optimum combination of filtration rate, body feed concentration, and terminal head loss by simply computing costs of filtration for any desired number of combinations and choosing the ten most economical.

Some Anomalies in Precoat Filter Operation and Their Effect on Filtration Performance Prediction

“…Theoretical equations 1 for filter design resulted from this research, as well as techniques for optimizing the design and operation of precoat filters. 2 The accep tance of precoat filtration by many communities has been attributed, in general, to a better understanding of filtration theory. For example, a report published in 1965 indicated that approx imately 81 precoat filtration plants had been completed for the filtration of pota ble water, and more than 60 percent of new swimming pools made use of these filters.…”

Design requirements of precoat filters for water filtration

“…In addition, this procedure requires many time consuming calculations. ' There fore, Dillingham (25,27) The large range of values of the B index indicated by these results demonstrates that the only way to determine the value of the g index and to describe its variation with body feed and suspended solids concentrations is by collecting filtration data with the water to be filtered at the pro posed plant site. Even data collected at a nearby plant which filters water from a similar source may be of little value.…”

“…8 for flat septa or Equation 6 for cylindrical septa. Dilling ham(25,27) neglected initial dilution when calculating t by approximating X by t since when X is used there are not explicit solutions for t of either Equation 8 or Equation 6.Arora(3) has stated that this can result in serious error.Initial dilution can be considered and explicit solutions of Equations 8 and 6 obtained by replacing X with t -t^ since X = t -t^ when t is large (Equation 30). The errors in doing this to calculate the filter run length are negligible since X approaches t -t^ very rapidly.…”