Dominic L. Boccelli

A reactive species model for chlorine decay and THM formation under rechlorination conditions

Chlorine is typically used within drinking water distribution systems to maintain a disinfectant residual and minimize biological regrowth. Typical distribution system models describe the loss of disinfectant due to reactions within the water matrix as first order with respect to chlorine concentration, with the reactants in excess. Recent work, however, has investigated relatively simple dynamic models that include a second, hypothetical reactive species. This work extends these latter models to account for discontinuities associated with rechlorination events, such as those caused by booster chlorination and by mixing at distribution system junction nodes. Mathematical arguments show that the reactive species model will always represent chlorine decay better than, or as well as, a first-order model, under single dose or rechlorination conditions; this result is confirmed by experiments on five different natural waters, and is further shown that the reactive species model can be significantly better under some rechlorination conditions. Trihalomethane (THM) formation was also monitored, and results show that a linear relationship between total THM (TTHM) formation and chlorine demand is appropriate under both single dose and rechlorination conditions. This linear relationship was estimated using the modeled chlorine demand from a calibrated reactive species model, and using the measured chlorine demand, both of which adequately represented the TTHM formation.

Efficient removal of endosulfan from aqueous solution by UV-C/peroxides: A comparative study

This study explored the efficiency of UV-C-based advanced oxidation processes (AOPs), i.e., UV/S2O8(2-), UV/HSO5(-), and UV/H2O2 for the degradation of endosulfan, an organochlorine insecticide and an emerging water pollutant. A significant removal, 91%, 86%, and 64%, of endosulfan, at an initial concentration of 2.45 μM and UV fluence of 480 mJ/cm(2), was achieved by UV/S2O8(2-), UV/HSO5(-), and UV/H2O2 processes, respectively, at a [peroxide]0/[endosulfan]0 molar ratio of 20. The efficiency of these processes was, however, inhibited in the presence of radical scavengers, such as alcohols (e.g., tertiary butyl alcohol and isopropyl alcohol) and natural organic matter (NOM). The inhibition was also influenced by common inorganic anions in the order of nitrite > bicarbonate > chloride > nitrate ≈ sulfate. The observed pseudo-first-order rate constant decreased while the degradation rate increased with increasing initial concentration of the target contaminant. The degradation mechanism of endosulfan by the AOPs was evaluated revealing the main by-product as endosulfan ether. Results of this study suggest that UV-C-based AOPs are potential methods for the removal of pesticides, such as endosulfan and its by-products, from contaminated water.

Real Time Water Demand Estimation in Water Distribution System

Accurate modeling of chemical transport in water distribution systems depends on accurate knowledge of temporally and spatially variable water demands. Typical network models would include water demands that are allocated from billing or census data, and thus may not be appropriate for specific operational analysis, such as emergency events arising from intentional or accidental contamination. During such an event, water consumption patterns may be significantly different from those assumed when developing the hydraulic model, and may change significantly over short time periods due to the unusual circumstances of the event. To allow accurate hydraulic and water quality prediction in real-time, the water demands should be updated continuously to reflect current conditions. The development of such a real-time water demand calibration method poses many complex issues such as identifiability and uncertainty of the water demand estimates. Given the sparsity of data that are likely to be available in real time, prior statistical information about water demands must be incorporated in the calibration procedure. In this paper, a method and algorithms are proposed for a real time water demand calibration process. A predictor-corrector methodology is proposed to predict statistical hydraulic behavior based on prior estimation of water demands, and then correct this prediction using new, real-time measurements. The problem is solved using the extended Kalman filter, which is a linear algorithm that calculates the estimate of water demands and their uncertainty. As part of the Kalman filter calculation, we calculate direct sensitivities of system hydraulic responses with respect to water demands. Results of numerical experiments illustrate the impacts of statistical demand variability, hydraulic measurement accuracy and sampling design on demand estimation. This paper was presented at the 8th Annual Water Distribution Systems Analysis Symposium which was held with the generous support of Awwa Research Foundation (AwwaRF).

Bayesian Approach for Real-Time Probabilistic Contamination Source Identification

AbstractDrinking water distribution system models have been increasingly utilized in the development and implementation of contaminant warning systems. This study proposes a Bayesian approach for probabilistic contamination source identification using a beta-binomial conjugate pair framework to identify contaminant source locations and times and compares the performance of this algorithm to previous work based on a Bayes’ rule approach. The proposed algorithm is capable of directly assigning a probability to a potential source location and updating the probability through the use of a backtracking algorithm and Bayesian statistics. The evaluation of the performance associated with the two algorithms was conducted by a simple comparison, as well as a simulation study in terms of a conservative chemical intrusion event through both a small skeletonized network and a large all-pipe distribution system network. Results from the simple comparison showed that the beta-binomial approach was more responsive to ch...

Development of a neural-based forecasting tool to classify recreational water quality using fecal indicator organisms

Users of recreational waters may be exposed to elevated pathogen levels through various point/non-point sources. Typical daily notifications rely on microbial analysis of indicator organisms (e.g., Escherichia coli) that require 18, or more, hours to provide an adequate response. Modeling approaches, such as multivariate linear regression (MLR) and artificial neural networks (ANN), have been utilized to provide quick predictions of microbial concentrations for classification purposes, but generally suffer from high false negative rates. This study introduces the use of learning vector quantization (LVQ)--a direct classification approach--for comparison with MLR and ANN approaches and integrates input selection for model development with respect to primary and secondary water quality standards within the Charles River Basin (Massachusetts, USA) using meteorologic, hydrologic, and microbial explanatory variables. Integrating input selection into model development showed that discharge variables were the most important explanatory variables while antecedent rainfall and time since previous events were also important. With respect to classification, all three models adequately represented the non-violated samples (>90%). The MLR approach had the highest false negative rates associated with classifying violated samples (41-62% vs 13-43% (ANN) and <16% (LVQ)) when using five or more explanatory variables. The ANN performance was more similar to LVQ when a larger number of explanatory variables were utilized, but the ANN performance degraded toward MLR performance as explanatory variables were removed. Overall, the use of LVQ as a direct classifier provided the best overall classification ability with respect to violated/non-violated samples for both standards.

Dynamic Water-Quality Simulation for Contaminant Intrusion Events in Distribution Systems

AbstractThe design of contamination warning systems and the performance of forensic tools are dependent on the performance of the event detection algorithms (EDA). However, most current EDA evaluation approaches do not account for the actual changes of common water-quality parameters in response to a contaminant. Thus, the objective of the current study was to develop water-quality models to represent the dynamics of chlorine, hydrogen ion concentration (pH), and conductivity in response to two contaminants [potassium cyanide (KCN) and nicotine] using experimental data. For chlorine-contaminant dynamics, a two-species second-order model was used to represent the reactions between chlorine and the background dissolved organic carbon as well as the contaminant. To simulate the change in pH, an equilibrium model was used to account for various water-quality species and was coupled with the dynamic chlorine model. As for electrical conductivity (EC), a step response, which is a linear relationship to the amou...

Model-Based Event Detection for Contaminant Warning Systems

AbstractSecurity issues have become increasingly important within distribution systems, which have led to the development of event detection algorithms (EDAs) to provide timely detection of intrusion events. The current study develops a localized model-based event detection algorithm that utilizes nonspecific water quality sensors to identify water quality anomalies. The proposed EDA focuses on evaluating a series of multivariate error signals between the observed signals and the model estimated signals based on a moving time-window of error statistics. The likelihood of the multivariate error signals is estimated using the product of univariate kernel density estimation (KDE), which is a type of nonparametric representation of the error distribution. A comprehensive analysis was performed using synthetic events to explore the combination of the moving window-pairs and bandwidth with respect to three injection strengths and two injection durations. In addition to the synthetic events, the EDA was also eva...

Sequencing Human Mitochondrial Hypervariable Region II as a Molecular Fingerprint for Environmental Waters

To protect environmental water from human fecal contamination, authorities must be able to unambiguously identify the source of the contamination. Current identification methods focus on tracking fecal bacteria associated with the human gut, but many of these bacterial indicators also thrive in the environment and in other mammalian hosts. Mitochondrial DNA could solve this problem by serving as a human-specific marker for fecal contamination. Here we show that the human mitochondrial hypervariable region II can function as a molecular fingerprint for human contamination in an urban watershed impacted by combined sewer overflows. We present high-throughput sequencing analysis of hypervariable region II for spatial resolution of the contaminated sites and assessment of the population diversity of the impacting regions. We propose that human mitochondrial DNA from public waste streams may serve as a tool for identifying waste sources definitively, analyzing population diversity, and conducting other anthropological investigations.

Comparing Single- and Multi-Species Water Quality Modeling Approaches for Assessing Contamination Exposure in Drinking Water Distribution Systems

New software such as EPANET-MSX enables water quality models that account for multiple reactive species in the distribution system. This allows for a more complete analysis of network water quality, including processes such as adsorption and biological inactivation by a disinfectant. For each of these reaction processes, three models are presented: single-specie conservative, single-specie reactions modeled by wall demand or bulk decay, and multi-species. The implications of model selection on the resulting exposure to contaminants that undergo these reaction processes are investigated for a hypothetical intentional contamination event and a simple single pipe system.

Preliminary Spatial-Temporal Statistical Analysis of Hourly Water Demand at Household Level

A methodology for modeling time series of hourly urban water use is presented based on separating the data into three components: trend, seasonality, and autocorrelation. Each component is represented by a model whose parameters are estimated. The series is transformed by removing each of the three components and the last transformation produces only a random error series. In the process of identifying the most suitable model for the autocorrelation component of the series, a large number of alternative autoregressive moving average (ARMA) models are assessed in terms of statistics that measure accuracy, parsimony, and randomness of the residuals. Hourly spatially aggregated water use in Cincinnati, Ohio, during the month of October, 2008 is modeled as an example. The model explains approximately 50% of the variance of this series, divided as trend (8.28%), seasonality (11.51%), and autocorrelation (29.94%). The methodology will be applied in a future study of a large data set of individual service connection hourly demand series, spatially aggregated under different schemes.