Content area
Groundwater, the major source of human drinking water, is susceptible to contamination from industrial and agricultural activities. This research develops a web-based simulation system of remote high performance computing model for contaminant transport and retention in soils. A three-dimensional advection-dispersion-reaction MRTM model, based on previous experimental and theoretical studies, is proposed to analyze the transport and retention of chemical contaminants in groundwater flowing through soils. Since three-dimensional experiments are difficult to implement and verify, this simulation system provides scientists an alternative to trace the contaminant movement in soils outside laboratories.
The alternating direction implicit (ADI) algorithm is used in this study to reduce the computational complexity. Although the ADI method is very efficient to solve the governing advection-dispersion-adsorption equations in the three-dimensional MRTM model, achieving higher order accuracy with different boundary conditions remains a difficult research topic. This research develops a new numerical scheme to achieve second-order accuracy with the Neumanntype boundary conditions. Furthermore, parallel computing is used to achieve high performance using powerful multiprocessor computers.
A web-based simulation system provides users a friendly interface for remote access to the system through Internet browsers, so as to utilize remote computing resources transparently and efficiently. In the client-side computing one-dimensional MRTM simulation system, the legacy code written in FORTRAN and C are wrapped and reused with Java code, which provides the web-based graphic user interface (GUI). The server-side computing three-dimensional MRTM simulation system integrates the remote high performance computing resources, database management systems, online visualization functionality, and web-based user-friendly GUIs. Given access to the Internet, users can execute and manage remote high performance computing jobs anywhere anytime, even through a web browser from a laptop personal computer.
In summary, this research has the following four contributions: (1) Extending the one-dimensional MRTM transport and retention model to three-dimensional applications. (2) Using the alternating direction implicit (ADI) numerical algorithm to reduce the computational complexity and achieving second-order accuracy with the Neumann-type boundary conditions. (3) Using parallel computing to achieve high performance on multiprocessor parallel computers. (4) Integrating the web-based system with user-friendly interfaces, which provide client-side and server-side computing services and web-based visualization functionality.