Abstract:
Use of artificial neural networks (ANNs) is becoming increasingly common in the analysis of groundwater hydrology and water resources problems. In this research, an ANN was developed and used to estimate aquifer parameter values, namely transmissivity and storage coefficient, from pumping test data for a large diameter well. The ANN was trained to map time-drawdown and well diameter data (input vector) to its corresponding transmissivity and storage coefficient values (output vector). Based upon a pre-specified range of aquifer parameters, the input vectors were generated from the analytical solution of Papadopulos and Copper for large diameter well in a homogeneous, isotropic, non-leaky confined aquifer. The ANN was trained with a fixed number of drawdown data points corresponding to a varying pre-specified range of aquifer parameters and time-series values. Once the network is trained to an acceptable level of accuracy, it produces an output of aquifer parameter values for any input vector. The results obtained with the ANN are in good agreement with published values. A significant advantage of the ANN approach is that it overcomes the problem of determining the storage coefficient, which when determined by traditional type curve matching method is of questionable reliability.
