Identification of hydraulic conductivities tensor for heterogeous confined aquifers using swarm intelligence

Abstract

The identification of spatially distributed hydraulic conductivities belongs to the inverse problems which are complex, highly nonlinear and cannot be effectively solved using traditional optimization techniques. The use of stochastic optimization techniques is, therefore,more interesting. As such, the present study proposes a linked optimization-simulation approach for estimating hydraulic conductivities of confined heterogeneous aquifers. This linkage is based on combining a recent swarm optimization technique called crow search algorithm (CSA) with a finite element model. Here, the role of CSA is to estimate the optimal set of hydraulic conductivities that minimize an objective function that measures the discrepancy between the measured hydraulic heads and those computed by the finite element model. The effectiveness of our approach was tested on three synthetic aquifer problems under both known and unknown boundary conditions. The obtained results were found satisfactory even under noisy hydraulic heads measurement