PARAMETER ASSESSMENT IN FLOW THROUGH POROUS MEDIA

Abstract

Effective planning and management of aquifers requires modeling of conceptualized system. Assessment of reliable parameters is vital for meaningful system simulation. Optimisation - simulation models are under continuous investigations to auto - calibrate such models resulting in assessment of hydraulic conductivity, specific yield, dispersivity, recharge estimations and recognition of an acceptable modeled structure. This study assesses parameters in confined and unconfined aquifers by genetic algorithm (GA) and simulated annealing (SA). These heuristic methods are found ideally suited for combinatorial optimization problems involving non-convex objective functions. An inverse parameter identification model based on coupled flow-solute transport simulations is developed. Twenty seven aquifer parameters for the nine zones of the confined aquifer are estimated by the coupled numerical models. Normally distributed noise was added to the data to examine their efficacy in the field problem. Both models based on SA and GA responded to the noisy data well and were also found to be independent to the initial guess of the parameters. Necessary modifications in the coupled parameter inversion algorithms were made to apply it to unconfmed aquifer of Mahi Right Bank Canal (MRBC) Command area of Kheda District, Gujarat, India. Estimated zonal hydraulic conductivity, longitudinal and transverse dispersivity values were compared with zonal values of the flow region for the calibrated model.