Automatic model calibration applying global optimization techniques

This paper summarizes a study aimed at the application of global optimization techniques for the purpose of quantitative characterization of the Wolfville formation located in Nova Scotia, Canada. Aquifer parameters (transmissivity, storativity, areal recharge and boundary flux) are calibrated in order to yield the best possible match with the available field observations. The calibration is accomplished using a global approach to the inverse procedure in “black box” systems optimization which makes possible the simultaneous fitting of several tens of parameters. This study indicates that, even for a limited number of optimized parameters, a global search procedure should be considered. Numerical results are presented and discussed to show the validity of the approach.