Parameter Estimation and Contaminant Source Characterization in Horizontal Tracer Experiment in Two-Layered Porous Media

概要

論文の詳細を見る
Laboratory tracer experiments are conducted to investigate solute transport in 100-cm long, horizontally placed flow field under steady saturated water flow conditions. Uniformly packed homogeneous and two-layered flow fields having cross-sectional areas of 50×6cm^2 are of concern. NaCl breakthrough curves are measured with salt sensors inserted with regular intervals along the flow direction. Longitudinal and transverse dispersivities, retardation factor, leak duration at the source and source magnitude are estimated using genetic algorithm (GA) linked with an analytical solution of the advection-dispersion equation. The results show that the values of longitudinal dispersivity range from 0.0631 to 0.316cm while transverse dispersivity is estimated in the range from 0.0221 to 0.190cm in homogeneous or two-layered porous media. In addition, breakthrough curves corresponding to the identified set of values are recovered with good accuracy regardless of noise level, while retardation factor and source magnitude are less sensitive parameters to noisy data among others, indicating that application of GA to parameter estimation problem with noisy data is expected to recover the concentration profiles. Moreover, estimates in layered media depend on transport pathway of contaminant, suggesting the importance of observation location in parameter estimation.