HR: 09:30h
AN: H11B-07    [PDF]
TI: Optimization of Engineering Design of Subsurface Environmental Remediation Systems: Development and Testing of Community Benchmark Problems
AU: * Mayer, A S
EM: asmayer@mtu.edu
AF: Department of Geological and Mining Engineering and Sciences Michigan Technological University, 1400 Townsend Drive, Houghton, MI 49931 United States
AU: Miller, C T
EM: casey_miller@unc.edu
AF: Center for the Advanced Study of the Environment Department of Environmental Sciences and Engineering University of North Carolina, CB #7400 105 Rosenau Hall, Chapel Hill, NC 27514 United States
AB: It is well established that the design of economically efficient subsurface remediation systems can benefit from the joint use of formal optimization and simulation models. It is also well known that obtaining the optimal solution for such designs is usually difficult and computationally expensive, due to the characteristic nonlinear, nonconvex nature of the solution spaces. We believe that more rapid progress on optimal design methods might result from both improved methods of evaluation and comparison of existing methods on realistic problems and from the investigation of novel methods not yet studied in subsurface remediation field. This work responds to these needs. We have designed a set of systematic test problems to be attacked by the engineering and mathematics community, as a means for benchmarking and comparing optimization approaches. The test problems pose many of the difficulties anticipated in solving real-world problems such as (a) mixed continuous and integer, nonlinear objective functions, (b) the combination of boundary conditions and system parameters gives rise to complex relationships between the objective function, the decision variables, the constraints, and the state variables, (c) evaluation of the objective function is based on solving model equations that are difficult to solve accurately and quickly; and (d) the number and range of decision variables is potentially enormous. The physical problems include water supply design problems in freshwater and freshwater-saltwater systems, a contaminant plume capture zone design problem, and a contaminant plume pump-and-treat design problem. Problem domains are specified in terms of hydraulic conductivity distributions- from homogeneous domains to spatially-correlated random fields- and in terms of confined vs. unconfined conditions. Each problem is specified completely in mathematical and numerical terms, but sufficient flexibility is allowed to provide for a wide range of problem solution approaches. For example, the user is free to select a flow and transport simulator and to specify spatial and temporal discretizations. This paper will describe the test problems and initial results.
DE: 1831 Groundwater quality
DE: 1832 Groundwater transport
DE: 1884 Water supply
SC: Hydrology [H]
MN: 2003 Fall Meeting