KWRRI Personnel

Chandramouli Viswanathan
Scientist III

354B Oliver Raymond Bldg.
University of Kentucky
Lexington, KY 40506

EDUCATION
Ph.D (Water Resources Engineering) Indian Institute of Technology, Madras, India, 1997
M.E. (Water Resources Engineering and Management), Annamali University, India, 1990
B.E. (Civil Engineering), Annamali University, India, 1985

PROFESSIONAL EMPLOYMENT
2006 - current: Scientist III, Kentucky Water Resources Research Institute, University of Kentucky & Adjunct Faculty, Department of Civil Engineering, University of Kentucky
2005 to 2006: Visiting Faculty, Department of Civil Engineering, University of Kentucky
2004: Associate Professor in Civil Engineering, Indian Institute of Technology, Guwahati, Assam, India
2003 to 2004: Visiting Faculty, Department of Civil Engineering, University of Kentucky
1999 to 2003: Assistant Professor in Civil Engineering, Indian Institute of Technology, Guwahati, Assam, India
1997-1999: Lecturer in Civil Engineering, National Institute of Technology, Tiruchirappali, Tamil Nadu, India
1986-1997: Assistant Engineer, Water Resources Organization, Government of Tamil Nadu, India

RESEARCH INTEREST AND EXPERTISE

• Water Resources Systems Analysis: Optimization - Reservoir operation and river system water resources planning
• Artificial Intelligence in Water Resources Engineering: Neural Networks, Fuzzy logic, Genetic algorithm and Expert Systems
• Artificial Intelligence in Environmental Engineering
• Remote Sensing and GIS applications in Water Resources Engineering
• Surface water and ground water quality modeling, TMDL studies
• Groundwater - Field Pumping test, conjunctive use of ground and surface water studies, Ground water modeling

Recent Research Activities in UK

Study 1: Backfilling, predicting and classifying microbial databases
ANN models were developed for predicting and classifying of fecal coliform, A-typical and E-coli bacteria using several input variables. Influence of each input in the ANN model for the considered output was demonstrated using Relative Strength Effect (RSE) concept. Input data selection using RSE and their advantages were demonstrated. A strategy to identify anomalous datasets using cross validation procedure was also evolved and reported in this study.

Study 2 : RSE based training termination to facilitate better training the small microbial datasets
In many environmental databases, limited availability of data is a major constraint. A novel training termination algorithm was developed using RSE for ANN back propagation training. In this approach, artificially a dummy input variable with random values was included in training and the RSE of the dummy variable was monitored. When the RSE value of the dummy variable reaches a minimum value, the training termination was performed. This approach facilitates training with more data since no testing dataset is required for this approach (which is required in conventional ANN modeling).

Study 3 : Predicting Enteric Virus presence in surface water
ANN models based on conventional and RSE based training were developed to predict Enteric virus presence. These models gave improved performance when compared to multiple logistic regression models. Input variables selected to predict these variables were chosen carefully to represent the age and load factors with flow change indices. These models showed very good predicting abilities.

Study 4 : Total Maximum Daily Load (TMDL) modeling for South Elkhorn, Town Branch, Wolf Run and Cane Run watersheds using EPA-BASINS and HSPF models
Developed fecal TMDLs for South Elkhorn, Town Branch, Wolf Run and Cane Run watersheds in Kentucky State. This research work involves geo-processing and watershed modeling. ARC-View, BASINS software were used for initial geo-processing of the spatial data. Hydrological Simulation Program Fortran (HSPF) modeling was performed for these watersheds. After hydrology calibration, water quality calibration for fecal coliforms was done using field observations. Bacterial load was estimated using EPA Bacterial Indicator Tool. Several simulations for different scenarios were taken up for finalizing the TMDLs.
(Ref: Report submitted to Kentucky Division of Water)

Study 5: Regional Ground water Model for Paducah Gaseous Diffusion Plant (PGDP) – Evaluation and Re-calibration
The regional ground water model for the Paducah Gaseous Diffusion Plant (PGDP) was developed in stages by US Department of Energy (DOE). Investigations identified contaminated ground water around the plant. Plumes of Trichloroethene and radioactive Technitium 99 were spread beyond the DOE property boundary. The finite difference model constructed using MODFLOW for ground water flow. The Solute Transport modeling is done using MODFLOWT model using the GW-Vistas interface. Several sensitivity analyses for hydraulic conductivity, leakage, recharge, constant head boundary conditions (for Ohio river), stream boundary conditions (for Little Bayou and Big Bayou creeks), attenuation rate of Trichloroethene (TCI), etc were handled and their influence over the ground water plume movement in different stress periods were documented. Different remedial options are being explored using the regional ground water model for further investigations. As per recent suggestions of Congress, for buying the properties around PGDP, (which are above the contaminated ground water plume) analyses were taken up by performing several simulations with different site conditions. Evaluation of regional ground model developed using FEMWATER is in progress.

PUBLICATIONS/PRESENTATIONS
V.Chandramouli, G.Brion, T.Neelakantan, S.Lingireddy (2006) Backfilling missing microbial concentrations in a riverine database using artificial neural networks submitted to Water Research, Elsevier. This paper is accepted for publication (paper No. WR6041).

V.Chandramouli, S.Lingireddy, G.Brion (2006), A Robust Training Terminating criterion for neural network modeling of small datasets, submitted to Journal of Computing in Civil Engineering, ASCE (American Society of Civil Engineers) This paper is accepted for publication (paper No. CP/2004/022533)

V.Chandramouli, Paresh Deka, (2005) Neural network based decision support system for optimal reservoir operation, Journal of Water Resources Management, KluwerAcademic Publishers, Springer, 19, 447-464.

Brion, G, V.Chandramouli, T.R.Neelakantan, Lingireddy, S, Girones, R, Lees, D, Allard, A and Vantarakis, A, (2005), Artificial Neural Network prediction of viruses in Shellfish, Applied and Environmental Microbiology, American Society of Microbiology, Sep, 5244-5253.

Paresh Deka, V.Chandramouli, (2005) Fuzzy Neural Network model for river flow prediction, Journal of Hydrologic Engineering ASCE (American Society of Civil Engineers), 10(4), 302-314.

Ramesh, S.V.T, V.Chandramouli, (2005) Improved weighing methods, deterministic and stochastic data driven models for estimation of missing precipitation records, Journal of Hydrology, Elsevier, 312, 191-206.

V.Chandramouli, Baleshwar Singh, Prajnan Goswami (2004). Study on Bankline migration of Brahmaputra river in Dhubri region, Assam using remote sensing imageries, Indian Journal of Hydrology, Indian Association of Hydrologists, Roorkee, India, 8(3),22-25.

Paresh Deka, V.Chandramouli, (2003) Fuzzy-neural network model for deriving river stage-discharge relationship, Hydrological Sciences Journal, Oxford, 48(2), 197-209.

V.Chandramouli, K.A.Kuppusamy, K.Manikandan, (2002). Study on water sharing in a multi reservoir system using dynamic programming - neural network model, International Journal of Water Resources Development, Vol.18 (3), 425-438.

Chandramouli, V., and Raman, H., (2001) Multi Reservoir modeling with dynamic programming and neural network, Jl of Water Resources Planning & Management, ASCE (American Society of Civil Engineers), 127(2), 89-98.

H. Raman, V.Chandramouli, (1996). Deriving a general Operating Policy for reservoir using Neural Network, Journal of Water Resources Planning and Management, ASCE (American Society of Civil Engineers), 122(5), 342-347.

Ramesh, S. V. T, V.Chandramouli, Ormsbee, L., Effect of DEM (Digital Elevation Models) Resolution on the Hydrological and Water Quality Modeling, American Society of Civil Engineers ASCE World Environmental and Water Resources Congress, May 21-25, Omaha, Nebraska, USA.

V.Chandramouli, Ramesh, S. V. T, Ormsbee, L., Surface Water Assessment and Hydrologic Modeling under Karst Aquifer conditions, American Geophysical Union Conference, Nov-Dec, 2005, AGU Conference, San Francisco, USA.

V.Chandramouli, Manish Kumar, Chandan Mahanta. Optimization of reservoir operation using dynamic programming and fuzzy neural network, presented in the International conference on Hydrology and Watershed Management, held during Dec, 2002 at Jawaharlal Nehru Technological University, Hyderabad, India.

C.Sivapragasam, Liong, S.Y, V.Chandramouli, Towards efficient Multipurpose reservoir operation: A new approach, presented in the International conference IAHR, held during Aug, 2002 at National University, Singapore. Vol.II, pp 569-574.

P. Deka, V.Chandramouli, River flow prediction using fuzzy neural network modeling, presented in the International conference IAHR, held during Aug, 2002 at National University, Singapore. Vol.II, pp.711-714.

V.Chandramouli, Y.Stanley, GIS Based Simulation Model with the aid of Artificial Neural Network for Optimal Water Resource Management of a Multi Reservoir System, presented in the International conference on Remote sensing and GIS/GPS ICORG held during Feb, 2001 at JNTU, Hydrabad, India. Vol.I, Pp 447-452.

H.Raman and V.Chandramouli, 1996, Optimal operation of multi-reservoir system using Dynamic Programming and neural network, paper presented at the International Conference on Applications of Artificial Intelligence in Engineering XI, Clear Waters, Florida, USA in September, 1996.

COMMUNICATIONS
Telephone: 859-257-8005
Fax: 859-257-4404
E-mail Chandramouli Viswanathan