Water distributions systems have generally been operated to provide a steady supply of water at a sufficient pressure. However, stringent water quality regulations have forced water utilities to maintain water quality throughout the distribution system. As a result, chloramine disinfection has gained significant popularity because it provides a longer lasting residual and forms substantially fewer regulated disinfection by-products (DBPs). Unfortunately, chloraminated distribution systems are more complex, relative to chlorinated systems, because they are impacted by a variety of parameters. Thus, the ability to represent these complex dynamics with a distribution system water quality model could assist in making operational and management decisions to maintain adequate water quality.
The objectives of this study were to perform a field-scale assessment associated with chloramine dynamics, DBP formation, and nitrification, and assess the ability of existing bench-scale models to represent complex water quality dynamics utilizing EPANET-MSX — a multi-species distribution system network water quality solver. The results of the field-scale study and modeling assessment were used to identify potential “knowledge gaps” that exist in current advanced water quality models and provide motivation for future improvements associated with distribution system water quality modeling.
A two week field-scale study was performed to evaluate chloramine dynamics, DBP formation, and nitrification in a distribution system. The samples and analysis included measurements for: total and combined chlorine, ammonia, nitrite, nitrate, pH, alkalinity, and organic carbon (chloramine dynamics); sulfate (as a pseudo-tracer); N-nitrosodimethylamine (DBP formation); and the absence/presence of ammonia-oxidizing bacteria (AOB) and anaerobic ammonia-oxidizing bacteria (ANAMMOX) (nitrification). A distribution system network model was used to simulate existing water quality models with EPANET-MSX. The observed and model predicted results were utilized to assess model performance.
The chloramine dynamics and NDMA model provided mixed results for representing the observed water quality data. During the first week of the study, the model represented the observed results reasonably well. However, during week two, more significant deviations were observed (particularly with monochloramine and ammonia) that may not be explained by the lack of modeling associated with pipe wall interactions or nitrification. The model provided reasonable NDMA predictions, but the assessment was limited due to the small sample size and analytical uncertainty associated with very low NDMA concentrations.
The absence/presence analysis of nitrifying bacteria resulted in three AOB-positive and six ANAMMOX-positive samples (of 52 total samples), which provided the first evidence of ANAMMOX occurring within a potable water system. This discovery introduces another possible mechanism that can contribute to nitrification in chloraminated distribution systems.
These results indicate that EPANET-MSX is capable of representing complex water quality dynamics and network hydraulics throughout the distribution system, but further improvements can be made to improve the underlying model accuracy. Inconsistencies between modeled and observed results illustrate the potential “knowledge gaps” and may be attributed to the following factors that were not considered: temporal variability in the influent water quality concentrations, interaction between biological and chemical species (nitrification), reactions between the bulk fluid and pipe wall, and accurate representation of spatial distribution and stochastic consumer demands.