A novel optimisation framework for interpretation of unconfined aquifer pumping test data

Abstract

The complex well function formulations developed for unconfined aquifer systems make the determination of aquifer parameters difficult and inefficient using classical methods. In addition, the dimensional dependency of aquifer parameters, as well as the non-linear and non-convex nature of inverse groundwater problems, can make the stand-alone use of the metaheuristic algorithms inefficient in terms of computation time and effort, producing non-unique solutions. Therefore, a novel optimisation framework was established to interpret pumping test data collected from an unconfined aquifer. The proposed approach works with four inputs that are based on the hybrid use of two non-dimensional physical and newly introduced two non-physical parameters. The method has the benefit of the simplicity of traditional methods and the accuracy of the differential evolution algorithm (DEA). The capability of the proposed scheme was broadly examined using several pumping test scenarios, including hypothetical and real field test datasets. Sensitivity analysis was also performed to understand the uncertainty associated with the estimated flow parameters. The results show that the proposed scheme, powered by the DEA, is able to achieve outstanding estimation performance compared with conventional methods and other nature-inspired algorithms.