AWWA ACE59865

Statistical methods of random selection were applied to the problem of water quality sampling in distribution systems. A sampling methodology was developed to address sampling goals using a statistical framework, thereby increasing the reliability of sample data by including estimation of standard errors and confidence levels. The technique of stratification was used to categorize locations and sampling times within the distribution system and several different schemes for allocation of the samples among the strata were investigated. The sampling methodology was applied to a case-study distribution system, the City of Durham, North Carolina, which serves approximately 150,000 customers. The goal of the case study application was to assess the effect of sampling location, time of sample collection, and overall sample size on the estimation of the proportion of samples with below-target chlorine residual concentration, P. To investigate the effectiveness of different sampling designs, expensive and time-consuming sampling programs are often undertaken. However, for this study the analysis was performed using numerical experiments based on a synthetic data set derived from a free chlorine model of the distribution system (EPANET2). Includes 2 references, tables, figure.