ESTIMATION Of AQUIFER PARAMETERS FROM PUMPING TEST DATA USING GENETIC ALGORITHM

Abstract

Modeling of groundwater systems is usually plagued by the uncertainties associated with hydraulic parameters of aquifer systems. A certain level of certainty and understanding of the hydraulic parameters of an aquifer system is necessary for efficient groundwater resources planning and management. Parameter estimation is an inverse problem, which involves the prediction of a function within a domain, given an error criterion with respect to a set of observed data. The pumping (aquifer) test is the standard technique for estimating aquifer parameters viz., transmissivity (7), storage coefficient (S), hydraulic conductivity (K), and thickness of the aquifer (H) for which the graphical method is widely used. Various numerical methods have been developed to estimate aquifer parameters and eliminate the subjectivity of traditional graphical type curve methods. Because of its inherent characteristics, Genetic Algorithm optimization technique avoids the subjectivity, long computation time, and ill-posedness often associated with conventional optimization techniques. In the present study, the efficacy of the Genetic Algorithm (GA) optimization technique is assessed in estimating partially penetrated confined aquifer parameters from the pumping test data. Computer codes are developed to obtain optimal partially penetrated confined aquifer parameters (hydraulic conductivity K, Storage Coefficient S, and Aquifer Thickness H) using Genetic Algorithm optimization technique. Applicability, adequacy, and robustness of the developed code are tested using sets of theoretical data generated by a model and field test data. Estimation of hydraulic parameters of partially penetrated confined aquifer from the time drawdown pumping field test data using curve match procedure were made, and those parameter values are compared with the one estimated by Genetic Algorithm optimization Technique. The Genetic Algorithm technique yielded significantly low values of the sum of squares of errors (SSE) as compared to graphical method; thus indicating that the GA technique is an efficient and reliable method for estimating various aquifer parameters, especially in a situation where the graphical matching is poor. Furthermore, the performance evaluation of the developed GA-based computer code is made to show that the fitness value (SSE) of the iii best point in a population reduces with increasing generation number and population size. The analysis of the sensitivity of the parameters during the performance of GA is also made to indicate that a unique set of aquifer parameters was obtained for both datasets. The GA based computer program with interactive window developed in this study is user friendly and can serve as a teaching and research tool, which could also be useful for practicing hydrologists and hydro-geologists. iv