Investigation of the fundamental assumptions relating compression‐permeability data with filtration

Abstract

AbstractA recently reported investigation indicates several inaccuracies in the methodology of compression‐permeability (C‐P) testing which suggest that previously reported agreement between C‐P and filtration data may be fortuitous. Until now, there has been no separate and direct confirmation of each of the two assumptions necessary to obtain a unique correspondence between C‐P and filtration data.The first assumption that the specific filtration resistance is a function solely of cumulative‐drag‐stress is generally accepted. Direct proof requires that the parabolic filtration equation, which is derived primarily on this assumption, describes both incompressible and compressible cake behavior. Most materials produce compressible cakes and “a priori” screening to find an incompressible cake requires identification of a material (Geon) that produces a cake with a linear axial pressure distribution. Results show that the parabolic filtration equation fits both types of cake behavior but an equation based on constant filtration resistance describes only incompressible cake behavior. To engineering accuracy and for dilute slurries, the assumption is verified.The second assumption that the cumulative‐drag‐stress equals the cake pressure drop is a macroscopic force balance and experimental verification requires a filter chamber designed to measure both of these quantities. A theoretical development, based on integral averaging, and experimental results both indicate that the ratio of cumulative‐drag‐stress to cake pressure drop correlates extremely well with cake porosity.A unique one‐to‐one correspondence between C‐P and filtration data is not possible without “a priori” knowledge of filter cake porosity. Previously reported agreement between C‐P and filtration data can probably be attributed to the L/D dependence of C‐P specific filtration resistances. Consequently, C‐P data can be used as a research tool to simulate filtration data but predicted filtration times based solely on C‐P data can be in considerable error.