Performance Assessment of Pre-Filtration Strainer of an Ultrafiltration Membrane System by Particle Size Analysis

Abstract

One of the industrial applications of ultrafiltration membrane system is water purification and wastewater treatment. Membranes act as physical barriers by eliminating particles such as pollen, yeast, bacteria, colloids, viruses, and macromolecules from feed water. The effectiveness of the membrane to separate particles is determined by its molecular weight cut-off and feed water characteristics. Typically, pre-filtration strainers are installed upstream of an ultrafiltration membrane system to separate large particles from the flow stream. The criteria for selection of the strainer pore size is unclear and is often determined by the feed water average particle size distribution. This paper is motivated by the hydraulic loading failure of a 125 μm strainer by average feed water particle size of 1.6 μm when the volumetric flow is at or greater than 40% of the rated design flow capacity. The objective of this paper are to: a) determine if the feed particle size distribution is a sufficient parameter for selection of pre-filtration strainer, b) evaluate the effect of feed flow velocity on strainer performance, and c) enhance strainer performance using vortex generator. In this experimental study, a Single Particle Optical Sensing, Accusizer, was used to analyze particle size distribution of five water samples collected at strainer feed, strainer filtrate, and strainer backwash. All samples were analyzed using a lower detection limit of 0.5 μm. In order to capture more counts of the larger particles present in the sample, a second analysis was done for each sample at a higher detection limit, 5.09 μm for feed sample, and 2.15 μm for the rest of the samples. Particle size data based on individual detection limits were statistically combined to generate comprehensive blended results of total number and total volume. The volume was determined based on assumption that each particle is spherically shaped. The Particle Size Distribution Measurement Accuracy is ±0.035 μm. Results showed that the feed particle size diameter and volume was insufficient to determine strainer size. Particle size distribution is needed at the feed, filtrate, and backwash to evaluate the strainer particle separation efficiency. It was observed that the total particle count in the filtrate (4.4 × 106) was an order of magnitude higher than the feed (3.2 × 105). Specifically, the total count for particles with diameter less than 7.22 μm were higher in the filtrate while larger particle size ≥ 7.22 μm were more in the feed stream. It appears that the large particles in the feed breaks down into smaller particles at the strainer interface and the small particles (≤ 7.22μm) passed through the pore into the filtrate. The particle breakdown, detachment of particles in the strainer pore into the filtrate, and particle to particle interactions are enhanced by increase in flow velocity hence increasing the hydrodynamic shear that acts on attached particles. A vortex generator inserted in to the strainer reduced pore clogging and pressure drop.