Estimating hydraulic parameters for the complex hydrodynamic models

Abstract

Hydrodynamic models have usually been calibrated manually, by trial-and-error, with different values of hydraulic parameters and hydraulic characteristics of boundary conditions. This method of calibration and estimation of hydraulic parameters requires an extensive knowledge and experience of experts, but whether the resulting solution includes an optimal set of parameters still remains an open question. An optimization method founded upon the Gauss-Marquardt-Levenberg algorithm, along with PEST software based on that algorithm, introduces automation of model calibration with regularization, which substantially reduces the effect of expert judgment on the result. The method also introduces the so-called 'pilot points', which transcend the concept of homogeneous zones with the values of hydraulic parameters of a hydrogeological system or zones with the specified boundary conditions. Mineral ore deposits are the most complex in geological and hydrogeological terms, so the lignite coal mine 'Tamnava - West Field' was selected as a good area to test the approach. Applying the proposed method, the spatial distributions of the horizontal (Kx=Ky) and vertical (Kz) components of hydraulic conductivity are determined in the study area.

Keywords

• calibration
• optimization
• pest
• hydraulic conductivity