Graduation Year

2007

Document Type

Thesis

Degree

M.S.

Degree Granting Department

Environmental Science and Policy

Major Professor

L. Donald Duke, Ph.D.

Keywords

Spiral wound membrane, Polyamide membrane, Membrane autopsy, Membrane cleaning, Membrane fouling

Abstract

Membrane biofouling is a common occurrence in water treatment plants that utilize reverse osmosis (RO). As bacteria and biofilm material build up on the membrane surface, it becomes more difficult for clean water to permeate through the membrane, and more pressure is required to produce the same amount of water. When pressures become critically high, membranes must be cleaned. This process is expensive in terms of chemical cost, labor, and downtime. Even after membranes have been cleaned, they can re-foul quickly if the cleaning did not effectively remove the biofilm. The water treatment plant in Dunedin, FL, which uses RO for treating groundwater, has experienced membrane biofouling since it began operation in 1992. Without the means to systematically evaluate a multitude of cleaning strategies on the bench scale, cleaning optimization must be conducted on the production skid level, which restricts the evaluation of alternative protocols.

This problem is typical for many RO plants. The objectives of this project are: (1) using a multi-level and systematic approach, develop cleaning strategies for biofouled membranes that will lead to improved cleaning and decreased operational costs; (2) develop other cleaning strategies that will add to the scientific knowledge base; (3) quantify the effects of improved protocols; and (4) determine the policy implications of developed protocols in terms of cost suitability to Dunedin and elsewhere in Florida. This project consists of three phases, with phases progressively more similar to the water production environment. In the first phase, a series of bench tests were performed in the laboratory. Fouled membrane swatches were soaked and agitated in different cleaning solutions for different lengths of time, at different temperatures and pH.

Protein and carbohydrate assays were then performed on both the cleaning solution and the membrane swatch to determine which conditions yield most complete removal of protein and carbohydrate from the membrane surface. Results indicate that carbohydrate removal does not appear to depend strongly on pH or temperature. Protein removal increases with increasing pH and is slightly greater at higher temperatures. The second phase of testing employed a 4"x6" stainless steel flat-sheet module in which cleanings were performed under different conditions to document corresponding changes in water flux and salt rejection. Operational parameters were based on pertinent literature and optimization results from Phase 1. Results indicate that water flux increases in response to cleaning at increasing pHs and increasing temperatures with best performances occuring after 30 minutes of cleaning. Salt rejection appears to decrease with pH.

The most effective cleaning protocols, determined through trials in Phases 1 and 2, were put to the test again in Phase 3 where cleanings were performed on a specially constructed single-element cleaning system (for 8.5" x 40" elements), designed to clean a membrane element in isolation. This phase also served as final verification of new cleaning protocols before implementation on the production scale. Results from this phase were inconclusive due to mechanical problems. A multi-level, systematic cleaning evaluation leads to better understanding of the dependence of biofilm material removal and membrane performance on critical factors such as temperature, pH, time of cleaning, and chemical dose, which results in improved cleaning protocols and ultimately cost savings to RO water utilities such as Dunedin.

Share

COinS