An Investigation of using statistically designed experiments and a computer model

Abstract

This paper outlines the results of an experimental program aimed at established the best formation parameters for dynamic Zr (IV) — PAA membranes formed on porous ceramic supports. The membrane formation parameters investigated are the pressure, the circulation velocity, the Zr (IV) concentration, and the PAA concentration. The formation pressure was allowed to vary between 200 psig (1.4 MP a) to 800 psig (5.4 MP a), the circulation velocity from 6 fps (1.8 m/s) to 24 fps (7.3 m/s), the Zr (IV) concentration from 10 -4M to 9 × 10 -4M, and the PAA concentration from 50 ppm to 250 ppm. Twenty statistically designed experiments were conducted to evaluate the effect of these four parameters on the flux and rejection of the resulting membrane at the end of the formation procedure. The results were used in a multiple linear regression computer program named BMD-PIR which develops the measured values with an accuracy of +_ 2 percent. Using these models, it is possible to form a membrane with a given value of flux or rejection within their ranges. For the range of variables used in this study, the maximum membrane rejection was 91% corresponding to a flux of 10 gpd/ft 2(1.7 cm/hr) for 0.05 M NaCl and 800 psi (5.4 MP a). High rejection was obtained by increasing the membrane formation pressure and by increasing the PAA concentration. High flux can be achieved by increasing the formation pressure, by increasing the Zr (IV) concentration up to 9.0 × 10 -4M -, and by increasing the circulation velocity.