HR: 1330h
AN: H12E-1023    [PDF]
TI: Identification of Complex Reactive Transport Processes in Groundwater Modeling
AU: * Thomas, B
EM: bthomas@ucla.edu
AF: UCLA, Civil & Environmental Engineering 5731 Boelter Hall/Box 951593, Los Angeles, CA 90035-1593
AU: Yeh, W W
EM: williamy@seas.ucla.edu
AF: UCLA, Civil & Environmental Engineering 5731 Boelter Hall/Box 951593, Los Angeles, CA 90035-1593
AU: Sun, N
EM: nezheng@seas.ucla.edu
AF: UCLA, Civil & Environmental Engineering 5731 Boelter Hall/Box 951593, Los Angeles, CA 90035-1593
AB: The study of innovative technologies for groundwater and soil remediation requires modeling complicated physical, chemical and biological processes in the subsurface. In recent years, various numerical models have been developed for solving these coupled advection-dispersion-reaction (ADR) problems. To make such a model a useful tool for assessment and design, however, both a modelAŸA›A›ƒ_sAªA›ƒ_zA›s structure and its parameters must be calibrated by observed data, and the reliability of the modeling parameters must be analyzed. The key to this process lies in identifying various reactive terms in the governing equations, such as adsorption and desorption, chemical decay or reaction, degradation by microorganisms, and adhesion to colloidal surfaces. A review of the literature shows that current research has several limitations, namely, that (1) the reactive procedure is often simplified as linear and equilibrium, (2) the functional form of the reactive term is often assumed to be known and that only a few constant parameters need to be identified, (3) the inverse problem is simply seen as a curve-fitting problem and solved by the traditional least squares method, and (4) most experiments and observations are in the lab scale. Many researchers have shown that the reactive processes in porous media are often nonlinear and kinetic and that their functional forms are often unknown. Moreover, the functional form of a reactive term obtained in the lab scale may not be appropriate for the field scale because of the physical and chemical heterogeneities of natural formations and the lack of observation data. Therefore, it is critical to develop a new methodology for identifying the various reactive terms in groundwater modeling. To this end, our research explores determining the structure of the reactive functionalities themselves based on available measurements. No a priori assumptions for the functional forms, such as linear versus nonlinear, are required. The procedure entails optimally constructing a series of polynomial basis functions to approximate the effects of the observed reaction behavior. This is accomplished through a coupling of genetic algorithms (GA) to sequential quadratic programming (SQP), which itself links to the ADR simulation model. A numerical example of the procedure is presented which demonstrates the GA/SQP progression along with the evolution of its concomitant parameter reliability.
DE: 1806 Chemistry of fresh water
DE: 1829 Groundwater hydrology
DE: 1831 Groundwater quality
DE: 1832 Groundwater transport
DE: 3260 Inverse theory
SC: Hydrology [H]
MN: 2003 Fall Meeting