Viz Lab Summer Grant 2004

2008 Reports 2007 Reports 2006 Reports 2005 Reports 2004 Reports

Visualizing the Flow of Contaminants Through Soil

Dr. Steven Sternberg Chemical Engineering

SUMMARY

Building on a series of projects I was involved with over the past school year, I finished the following films:

This project is a continuation of my work involving the measurement of a contaminant moving through heterogeneous porous media. I have built and operated a lab scale sandbox in my lab, in which I measure concentration of a contaminant throughout the actual space and throughout time. The analysis from each experiment is time consuming (4 to 6 hours) and, as part of my visualization projects goals, I wanted to find ways to reduce time spent on the initial data analysis. The method I investigated was to learn to use the Fluent software available for faculty use in the Viz lab. This software allows me to pre-calculate pressure fields, local and global velocities, and to estimate expected travel times. This knowledge is essential in calculating the transport properties in my experimental systems and can be used to greatly reduce the effort in data analysis. I have not yet quite obtained all this information, but I have made an excellent start and hope to complete the work this summer.

I have also used the viz lab resources to explore new ways to visualize the data sets I have collected (and am currently collecting) in two ways:

1) develop ‘movies’ to show the time evolution of my experimental system for posting on my web page

2) Improve the data analysis through automation so that greater levels of analysis can be completed in less time.

DATA SETS

The data sets I create are from experiments designed to investigate the flow of a ground water contaminant through heterogeneous soils. The particular phenomenon I am interested in is the additional mixing caused by the flow as it encounters distinct changes in soil permeability. These changes are called heterogeneities, and are created in the experiments by forming an inclusion of different permeability into a larger region. Data is collected with a 16 x 16 grid of ion specific electrodes embedded into the soil. Each electrode is measured every ¾ second. Total experimental time may be 5000 to 15000 seconds. This creates data files of plain text ranging in size from 3 to 10 MB. Greater detail of the experiments and data formats are available upon request. A simple, and small file size visualization showing this phenomenon will be an excellent addition to future presentations and for my web page research interests.

A second aspect of my project is to develop a new data analysis approach exploiting the statistical nature of this phenomenon, which suggests that the process may be amenable to stochastic modeling based on the first four statistical moments of the distribution. This program could be developed in FORTRAN to a) preprocess the experimental data b) analyze it using standard statistical routines and c) post analysis including the visualization as described above.

CONCLUSION

This work is ongoing. I have learned how to use the Fluent software, and have built models in it of each of my experimental systems. I have made preliminary calculations of my actual experimental systems, but have found that they need to be trained to my historical data sets. The actual values of system parameters input to fluent do not adequately match my own measurements. This upcoming summer, I plan to create the data mappings to allow accurate predictions, and to find ways to have Fluent create data files of point velocities, pressures, and expected travel times to each position in the experiments.