Estimating Groundwater Flow Parameters Using Response Surface Methodology.

摘要

This thesis examined the use of response surface methodology (RSM) as a parameter estimation technique in the field of groundwater flow modeling. Using RSM, an attempt was made to calibrate three hydraulic parameters (porosity, transverse permeability, and rate of recharge) of an existing two-dimensional, steady-state flow model. The model simulated groundwater flow for a portion of landfill 10 located on Wright-Patterson Air Force Base, Ohio. The model had previously been calibrated by graphical matching observed water-levels to predicted water-levels. Using the parameter values from the earlier calibration effort as a starting point, a central composite design was developed and the simulation executed at each design point. A residual sum of squares function was used as the calibration criteria and an empirical model of the error surface was developed. Of the three hydraulic parameters, only transverse permeability had a significant effect on the response. The regression model also indicated the response had a high degree of variability. A graph of the regression equation revealed no local optima within the design region indicating the initial parameter estimates may not have been warranted. Parameter estimation, Model calibration, Response surface methodology, Groundwater flow models.