Srinivasa Lingireddy

A neural network approach to identifying non-point sources of microbial contamination

Commonly measured fecal bacteria concentrations in water and rainfall data were utilized as inputs for training a neural network model to distinguish between urban and agricultural fecal contamination present in inputs to a drinking water reservoir. Seven sites were selected that represented differing degrees of fecal contamination arising from agricultural, urban, or a blend of both land use activities. The absence of human sewage at the inlet sites to the reservoir was determined by analysis for coprostannol and serotyping of male-specific coliphage. Analyses for fecal coliform (FC), fecal streptococci (FS), total coliform (TC) and coliphage were conducted over 2 years from weekly samples collected from these sites during dry and rainy times during warm seasons. The average concentrations of microorganisms measured were highly variable and analysis of FC/FS ratios was not able to differentiate between urban or agriculturally impacted sites. A neural network model was written that used bacterial and weather data to differentiate between three site classifications: urban, agricultural and a mixture of these. The validity of the source identification, neural network model was verified through case study.

A neural-network-based classification scheme for sorting sources and ages of fecal contamination in water

Artificial neural networks (ANNs) were successfully applied to data observations from a small watershed consisting of commonly measured indicator bacteria, weather conditions, and turbidity to distinguish between human sewage and animal-impacted runoff, fresh runoff from aged, and agricultural land-use-associated fresh runoff from that of suburban land-use-associated-fresh runoff. The ANNs were applied in a cascading, or hierarchical scheme. ANN performance was measured in two ways: (1) training and (2) testing. An ANN was able to sort sewage from runoff with < 1% error. Turbidity was found to be relatively unimportant for sorting sewage from runoff, while gross measurements of gram-negative and gram-positive bacteria were required. Predictions clustered tightly around the known values. ANN classification of aged suburban runoff from fresh, and agricultural runoff from suburban was accomplished with > 90% accuracy.

Hydraulic network calibration using genetic optimization

The validity of a hydraulic network model depends not only on the accuracy of its physical and geometric data but also on the accuracy of certain parametric data such as pipe roughness coefficients and nodal demands. Difficulties associated with economical and reliable measurements for these parameters often dictate estimation of these parameters through model calibration. This paper describes an optimization approach to calibrate a network model for pipe roughness coefficients, and spatial as well as temporal demand adjustment factors. The proposed model obtains an optimal solution by minimizing a nonlinear objective function subject to a set of linear and nonlinear constraints using a powerful search technique based on a genetic algorithm. Application of the optimal calibration model to water distribution systems using synthetic calibration data demonstrates capabilities of the proposed algorithm to generate good solutions in an efficient and robust manner.

AQUIFER PARAMETER ESTIMATION USING GENETIC ALGORITHMS AND NEURAL NETWORKS

Abstract Computational tools such as genetic algorithms and neural networks are becoming increasingly popular in scientific applications involving mathematical modeling. These tools emulate natural biological processes in an attempt to build more robust and efficient mathematical models. The present study explores the applicability of genetic algorithms and neural networks for aquifer parameter estimation, in an optimization framework. Although optimization models based on genetic algorithms are more robust than conventional nonlinear programming techniques, they often necessitate many computationally expensive function evaluations. On the other hand, genetic algorithms can also tolerate approximate function evaluations. The present study employs artificial neural networks that provide quick but reasonably accurate function evaluation, in conjunction with genetic algorithms. Such an optimization framework makes the resulting calibration model highly robust and efficient. Applicability of the proposed mode...

Artificial neural network prediction of viruses in shellfish.

ABSTRACT A database was probed with artificial neural network (ANN) and multivariate logistic regression (MLR) models to investigate the efficacy of predicting PCR-identified human adenovirus (ADV), Norwalk-like virus (NLV), and enterovirus (EV) presence or absence in shellfish harvested from diverse countries in Europe (Spain, Sweden, Greece, and the United Kingdom). The relative importance of numerical and heuristic input variables to the ANN model for each country and for the combined data was analyzed with a newly defined relative strength effect, which illuminated the importance of bacteriophages as potential viral indicators. The results of this analysis showed that ANN models predicted all types of viral presence and absence in shellfish with better precision than MLR models for a multicountry database. For overall presence/absence classification accuracy, ANN modeling had a performance rate of 95.9%, 98.9%, and 95.7% versus 60.5%, 75.0%, and 64.6% for the MLR for ADV, NLV, and EV, respectively. The selectivity (prediction of viral negatives) was greater than the sensitivity (prediction of viral positives) for both models and with all virus types, with the ANN model performing with greater sensitivity than the MLR. ANN models were able to illuminate site-specific relationships between microbial indicators chosen as model inputs and human virus presence. A validation study on ADV demonstrated that the MLR and ANN models differed in sensitivity and selectivity, with the ANN model correctly identifying ADV presence with greater precision.

Neural Network–Derived Heuristic Framework for Sizing Surge Vessels

AbstractSurge vessels provide efficient protection against low and high transient pressures in water distribution systems. However, they can be quite expensive, and any reduction in surge vessel size can significantly reduce surge protection costs. Graphical and other heuristic methods reported in literature are limited to sizing surge vessels for simple rising mains. Attempts to use more structured optimization techniques have been largely unsuccessful because of their impractical computational requirements. This article proposes a robust framework for developing surge protection design tools and demonstrates the usefulness of the framework through an example surge vessel sizing tool. The essence of the proposed framework is in the identification of key transient response parameters that influence surge vessel characteristics from seemingly unmanageable transient response data. This parameterization helps the sizing tool to exploit the similarity between transient responses of small pipe networks and sub...

A Neural Network Based Classification Scheme for Sorting Sources and Ages of Fecal Contamination in Water

Artificial neural network (ANN) modeling that used a set of simple bacterial measurements and informational inputs was successfully applied to data observations from a small watershed for the purposes of distinguishng between human sewage and animal-impacted runoff, fresh runoff from aged, and agricultural land use associated fresh runoff from that of suburban land-use associated fresh runoff. The ANN approach was able to classify sewage from heavily contaminated runoff with greater than 99% accuracy. Turbidity was found to be relatively unimportant as an input variable for sorting sewage from runoff, while gross measurements of gram-negative and gram-positive bacteria were required. ANN classification of aged suburban runoff from fresh, and agricultural runoff from suburban was accomplished with greater than 90% accuracy.

Identification of Pollution Sources Via Neural Networks

Neural networks are already used in watershed management to forecast, predict, and control water supplies, but to date have not been widely used to predict, control, or improve water quality. Yet, neural network based models can be applied to identify the presence of potentially hazardous, pathogen laden, agricultural runoff in receiving waters. This is an entirely novel approach for analysis of water quality indicators that, when expanded beyond source identification, may lead to development of new watershed monitoring and management paradigms.

Sensor Placement Guidance for Small Utilities

Contamination warning systems (CWS) are strategies to lessen the effects of contamination in water distribution systems by delivering early indication of events. A critical component of CWS, online quality monitoring, involves a network of sensors that assess water quality and alert an operator of contamination. Utilities developing these monitoring systems are faced with the decision of what locations are optimal for deployment of sensors. The TEVA-SPOT software was developed to analyze the vulnerability of systems and aid utilities in designing sensor networks. However, many small utilities do not have the technical or financial resources needed to effectively use TEVA-SPOT. As a result, a sensor placement algorithm was developed and implemented in a commercial network distribution model (i.e. KYPIPE) as a simple tool to aid small utilities in sensor placement. The developed tool was validated using 12 small distribution system models and multiple contamination scenarios for the placement of one and two sensors. Results were compared with data from identical simulations runs in TEVA-SPOT to verify the effectiveness of the proposed algorithm. Because the proposed algorithm uses a simple complete enumeration strategy for sensor placement, the algorithm was able to select the same or superior nodes to those selected by TEVA-SPOT.