详情

原文传递 DATA COLLECTION TO SUPPORT A SIMPLIFIED BACTERIAL REGROWTH MODEL FOR DISTRIBUTION SYSTEMS

题名：	DATA COLLECTION TO SUPPORT A SIMPLIFIED BACTERIAL REGROWTH MODEL FOR DISTRIBUTION SYSTEMS
作者：	Francis A. DiGiano Weidong Zhang Donald E. Francisco Melissa Wood
关键词：	distribution；collection；mplif；growth；systems；simp；data；model；centration；calibration
摘要：	The main objectives of this research were to: (1) collect field measurements of bacterial regrowth over a 12-month study (September 1998-August 1999) in the Durham and Raleigh distribution systems to add to the data base from our previous 1 7-month study of the same systems and (2) initiate development of a deterministic model of bacterial regrowth that will eventually be linked with components that account for uncertainty in parameter estimates to yield a more realistic model. Heterotrophic plate count (HPC) by the R2A agar method, total coliforms, biodegradable dissolved organic carbon (BDOC), chlorine residual (in Durham), combined chlorine (in Raleigh), disinfectant demand, disinfectant decay rates, total organic carbon (TOC) and temperature were measured monthly in the finished water and 10 stations in each distribution system. A tracer study was conducted in Durham to measure mean water residence time. The study design was unique because it indided a negative step input of fluoride at one water treatment plant (WTP) in Durham (Williams WTP) and a simultaneous switch in coagulant at the other (Brown WTP). This permitted calculation of the percent contribution of water from each WTP at each sampling station and a weighted average of the mean residence time of water originating from each WTP. Mean water residence time was defined from interpretation of the concentration-time data of chemical tracers at each sampling station. We believe that mean residence time can be correlated with water age as calculated in hydraulic network models. Based on the HPC data in the Durham and Raleigh distribution systems, the effectiveness of chlorine and combined chlorine in controlling regrowth was judged about equal and confirmed our previous findings. High HPC was found at stations with low chlorine concentrations in Durham and low combined chlorine concentrations in Raleigh. Low residual concentrations were generally associated with long mean residence time as expected. Although the BDOC was less than 0.5 mg/L, it was sufficient to produce regrowth at stations with low disinfectant residual. The bulk decay rates for free and combined chlorine showed that a simple, first-order rate model is simplistic; concentration of oxidizable organic matter should be included. A mechanistic model of regrowth was developed for a long section of pipe of a fixed diameter under steady-state flow conditions. The model included bacterial growth and attachment to the pipe wall, subsequent bacterial detachment, and chlorine inactivation of bacteria in bulk solution. The model showed that regrowth is very sensitive to the free-chlorine dosage in the finished water. Decreasing the chlorine dosage from 0.7 to 0.5 mg/L increased the HPC by several orders of magnitude. Once a biofilm was established on the pipe, neither increasing the chlorine dosage nor decreasing the BDOC in the finished water reduced regrowth substantially over 100 hours of simulation. The tracer data collected in both Durham and Raleigh will provide an excellent basis for calibration of existing network models such as EPANET. About 27 sets of monthly water quality data are available for testing the accuracy of model predictions for chlorine or chloramine residuals in such models. The development of new models that include bacterial regrowth may be advanced as well by use of our HPC data. The data sets would provide for testing of uncertainty in parameter estimates because in such models a distribution of parameter values must be included and this requires calibration with many sets of data.
报告类型：	科技报告