Year

2012

Season

Spring

Paper Type

Master's Thesis

College

College of Computing, Engineering & Construction

Degree Name

Master of Science in Mechanical Engineering (MSME)

Department

Engineering

First Advisor

Dr. James Fletcher

Second Advisor

Dr. Richard V. Conte

Third Advisor

Dr. Adel El-Safty

Department Chair

Dr. Murat M. Tiryakioglu

College Dean

Dr. Mark A. Tumeo

Abstract

While extensive mathematical and numerical work has been done in terms of modeling the mainstream flow in a tube with porous walls, very little has been done experimentally to confirm these various solutions, and what has been done has focused on large sintered metal tubes used in nuclear power applications. Furthermore, these solutions are quite mathematically complex and arduous to implement. In this work, the mainstream flow through a porous polymer membrane tube is examined and a method for calculating the through-membrane flow rate and axial pressure drop is presented. Two membrane tubes are tested experimentally, and a simple set of modeling equations that are physically intuitive are presented which fit the data. A characterization test is described which can be used to determine the permeability coefficient, kD, for a membrane sample, which can in turn be used to calculate the through-membrane flow rate and axial pressure drop. The models are then evaluated by performing flow-through experiments and measuring the pressures and flows within the membrane. For both membranes tested, the permeability coefficient is determined to be kD = 5.9394 × 10−14 m2 . For the tube diameters (2 mm and 8 mm) and flow rates (100-500 sccm) tested, it is shown that for dimensionless tube lengths bL = L/(dReD) ≥ 0.3, a model that assumes fully developed flow through the entire tube accurately describes the through-membrane flow rate data. The fully-developed model consistently under-predicts the experimental data for axial pressure drop, therefore it is assumed that the discrepancy is due to an additional pressure loss from the developing region. This loss is determined empirically using the data. The model’s validity is examined and compared to that of other authors for the range of flow rates tested.

Share

COinS