Doctor of Philosophy (PhD)
Civil and Environmental Engineering
Unstructured grid is a tessellation of geometric shapes in irregular patterns that provides flexibility in grid design for groundwater modeling. However, groundwater modeling is mostly developed with uniform grid tessellation and layer, which could simplify model structure or cause expensive computational costs in high-resolution simulations. Unstructured grid incorporates non-uniform horizontal and non-uniform vertical discretizations providing the capability to replicate complex hydrostratigraphy, capture geologic features that are crucial for groundwater flow simulation, and reduce computational costs while maintaining a high resolution for areas of interest. This study contains three parts to investigate unstructured-grid approach on constructing high-fidelity groundwater models, comparisons with analytical relief well evaluation, and optimization implementation on relief well operations. The first study develops a three-dimensional (3-D) groundwater model using MODFLOW-USG on an unstructured grid to evaluate relief well performance at the Profit Island vicinity levee and conduct comparative analysis with conventional seepage analysis (i.e., blanket theory). Consequently, this case study indicates that 3-D modeling is a more accountable and precise tool than blanket theory. The second study proposes relief well operations optimization as an alternative to increase levee and floodwall factor of safety against underseepage, which is achieved by a multi-objective mixed-integer non-linear programming implemented with a 3-D unstructured groundwater model. The approach is demonstrated at the Inner Harbor Navigation Canal, New Orleans, Louisiana. Relief well operations optimization provides decision-makers useful tradeoffs between averaged factor of safety deficit, total pumping rate, and the number of pumping wells for making pumping decisions. The third study advances unstructured grid to develop a high-fidelity 3-D groundwater model using a lithology model for the Louisiana Capital Area Groundwater Conservation District (District) on the Southern Hills Aquifer System. Groundwater flow simulation, water budget analysis, uncertainty analysis, and comparisons with satellite data provide comprehensive understanding of groundwater storage variation in the District and evaluate groundwater depletion induced by excessive prolonged groundwater withdrawals in the Baton Rouge area. This study successfully demonstrates the capability of unstructured-grid approach on groundwater modeling to simulate groundwater dynamics within complex subsurface hydrogeological structures, evaluate relief well performances, and optimization implementation.
Chen, Ye-Hong, "Investigation of Groundwater Depletion and Leveel Underseepage with Unstructured-Grid Modeling Approach" (2022). LSU Doctoral Dissertations. 5789.
Tsai, Frank T.-C.