Date of Award

Winter 2002

Document Type


Degree Name

Doctor of Psychology (PsyD)


Civil & Environmental Engineering

Committee Director

Gary C. Schafran

Committee Member

Avodeji Demuren

Committee Member

Willaim A. Drewry

Committee Member

Jaewan Yoon


Treatment plants utilize chlorine and chloramine both as a primary disinfectant for water purification and a secondary disinfectant for distribution system protection. The treatment goal is that the disinfectant residual be present throughout the distribution system until it reaches the customer's tap. Maintaining a residual throughout the system is critical for providing pathogen-free water and protecting human health. Therefore, water utility authorities are very interested in how disinfectants such as chlorine and chloramine decay as they travel through a water distribution system.

Current mathematical models typically divide the decay into two distinct phases. There is (1) decay occurring in the bulk phase of water and (2) decay attributed to a demand exerted by the pipe wall. Transport between these two phases has been described with a variety of mass transfer processes, which utilize dimensionless flow parameters such as the Reynolds, Schmidt and Sherwood number.

For many unlined cast iron water distribution grids, field data exhibited higher disinfectant decay rates than could be explained with conventional modeling. Using the Norfolk Naval Base as a test site, field data was collected from low flow areas of the water distribution system from 1999 through 2002. A new decay model was developed to account for biofilm characteristics as they vary in relation to various pipe wall materials. These materials were polyvinyl chloride, transite, cement-lined ductile iron, and unlined cast iron.

During the research, it was found that the heterogeneous surface profile of cast iron pipe significantly altered the hydraulic flow profile. As a result, turbulent flow can be assumed for all flow regimes in the mass transfer expressions. However, modifying the hydraulic radius and diameter to account for the tubercled wall surface did not improve the model results. These modifications overpredicted the mass transfer in most scenarios.

This research effort also identified three other decay sinks to be incorporated in the model. These sinks were diffusion, iron release from cast iron pipe walls, and microbial detachment events from pipe wall biofilms. Diffusion was found to be a significant sink under zero flow conditions for pipe diameters eight inches and smaller. Adjustments in the diffusion term to account for tubercled surface profiles also appear to be justified. Iron release episodes were very sporadic but were found to occur under zero flow and low flow conditions. Microbial wall sheds were found to be the most dominant decay sink for cast iron pipe in zero flow conditions. Microbial detachment events were indirectly measured by collecting and analyzing heterotrophic plate count samples using R2A agar.