| The primary objective of this research has been to develop non-linear optimization models for determining the optimal design and operation of the soil aquifer treatment (SAT) systems. The numerical models are capable of determining the wet-dry schedules for single and multiple cycle operation periods. Tools for designing, analyzing, and operating SAT systems have been developed in this research.;Phases II and III of this research are the development of a mathematical programming procedure, to solve the soil aquifer treatment (SAT) system operation, formulated as a discrete time optimal control problem. The procedure interfaces a non-linear optimization model with a simulation code. The generalized reduced gradient (GRG2) method was used as the non-linear optimization code while the hydraulic simulation was performed using an unsaturated flow model, kinematic wave and another one-dimensional unsaturated flow model, HYDRUS, which incorporates the clogging effects, to provide a more sophisticated level of solution algorithm.;Various levels of modeling efforts are analyzed to derive the validity of the numerical computer models to solve the SAT system operation problem. The application of the models developed in this research, generates optimal operation schedules for the SAT systems. This new methodology is an engineering tool for water managers and decision makers for evaluating various modeling approaches for the design, operation, calibration and maintenance of the soil aquifer treatment (SAT) systems.;In phase I, more than one physically based infiltration models have been evaluated to describe the infiltration process used in the groundwater recharge. The soil moisture redistribution process has also been described in two ways. The optimization problem has a non-linear objective function, to maximize infiltration, which is subject to a set constraints. The model described above has non-linearity in both objective function and constraints, which requires a non-linear programming algorithm for solution. In this phase, a non-linear programming procedure was developed to solve the above model. The non-linear programming model was also expanded to calibrate actual field data obtained from a pilot soil aquifer treatment (SAT) system, in Tucson, Arizona. Results from the non-linear programming model generate optimal operation schedules for the SAT system. |
