Viz Lab Summer Grant 2006

2007 Reports 2006 Reports 2005 Reports 2004 Reports

Visualization and Animation of Contaminant Movement in Porous Media

Steve Sternberg Chemical Engineering

Overview

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, 2-D sandbox in my lab in which I measure concentration of a contaminant throughout space and time. Data collection is mostly automatic, and creates measures of potential (as mV) at 256 positions every 0.75 seconds for up to 10,000 seconds. Initial data analysis is time consuming because the electrode at each position must be calibrated independently (calibration data is obtained during data collection). Data from the initial analysis consists of a large, complex matrix of the time and spatial dependence of several properties. Summary numbers or tables do not adequately provide a non-expert an easy way to understand the results. Finally the overall goal of my research is to improve the predictive power of mathematical models.

The goals of my summer project were:

1. Find a way to reduce time spent on the initial data analysis.
2. Develop new, visual ways to present the results.
3. Explore a new way to compare my experimental results with predictive theory.

Results

1. Find a way to reduce time spent on the initial data analysis

Discussions with other V-Labbers convinced me to explore new graphical user interfaces. Especially interesting was for me to examine just were and how often the mouse was used. I was spending a huge amount of time clicking a mouse on tiny input boxes.

Figure 1: Old Interface

I replaced the tiny click boxes with easy to see and use slider controls. I also realized that at the end of each electrode analysis, I clicked three times. I replaced that with a single function click. The savings added up to six fold decrease in time required per run.

Figure 2: New Interface

The new interface (shown in Figure 2) groups all the click-spots and sliders the left side middle. This project alone made my summer time spent in the V-lab worthwhile.

2. Develop new, visual ways to present the results

This part of the work—finding ways to show my results visually—was the most fun part of the project. The next several images show examples of what I created. Each shows a 2-D plan of the sand box with flow inlet and outlet in one of the corners. The slightly shaded quarter region highlights the location of a distinct heterogeneity in the flow field. The actual color mapping shows the position of a flowing fluid front at a certain time. These static pictures demonstrate the progression of a contaminant plume as it moves into and through the sandbox.

These pictures are just a subset of the images I created from my data sets. I found this work very interesting and fun. It genuinely displays the results in an easy to discuss format.

While I was working on this, a discussion ensued at the V-lab into animating the data sets. It was actually quite straight forward and used the images I created to make the above static images. One such animation is shown below.

Explore a new way to compare my experimental results with predictive theory

The other goal—to explore ways to compare the experimental work with predictive theory was performed with the modeling software FLUENT. The software is used for computational fluid dynamics. I was able to build models of my system, calculate local and global velocity fields, and generate pressure profiles. I was not successful in creating contaminant transport models, nor in using FLUENT to compare and contrast results. I learned how to use the modeling software, and was pointed in the direction of a package more suitable for my work—Groundwater Vistas. This is a windows based front end for the USGS MODFLOW programs. I am planning to explore this third goal in detail using the new software package.