For one week, technically one week plus a little extra, in January 2008 it was my turn to run an experiment at the lab. While there is no diary to accompany this latest experiment, there are some highlights and teasers presented here. Since the results need to be reviewed and scrutinized there are no axis labels or interpretations provided. Just look at these for their aesthetic value (that's asking a lot) and watch your DavidPace.com RSS feed for new entries that provide more information. The ultimate goal is to get a thesis written, in which case all will be revealed.

What Happened in January 2008?

Here are some statistics from the experiment:

• Total Number of Plasma Discharges (Shots) Measured = 254,988
• Approximate Total Plasma Time ≈ 60 minutes
• Number of Separate Experiments Performed = 35
• Total Size of Compressed Data Files = 29.61 GB

The enormous data set possible from collecting so many shots is one advantage of working in a basic plasma laboratory. For comparison, consider how long it might take a fusion laboratory to acquire this many plasma shots. Let's say that DIII-D (I like this picture) can get 40 research-grade shots in one day. At such a rate it would take them just under 18 years to get 254,988 shots. That would only be the case if they ran the machine every day. In reality, DIII-D is unlikely to get that many shots throughout its entire life†.

Searching through this mountain of data can be a daunting task. With that in mind, here are a few pointers on how to find noteworthy behaviors in any kind of data.

If Something is Sticking Out, It Might be Important

This first figure features a few distinct peaks. The two largest ones stick out and catch the eye. Only the largest one is important, which I hope to prove to you very soon.

Peaks in a line plot are often important, unless labeled otherwise.

If There is a Hole Somewhere, Something Important is Probably Happening There

The next image shows a depression that is well localized. There is a lot of interesting behavior going on inside there and a lot of the flat region can be ignored. In a rushed review it would be a good idea to concentrate only on the region of the depression. In a thorough review, and all thesis projects are thorough, everything must be taken into account. So, while it is apparent where all the action is, the entire area needs to be examined.

When it comes to grad school, you'll find depression(s) everywhere.

If a Signal Changes Quickly, Then Something Important Probably Happened

The next image shows a transition from one type of oscillatory behavior to another, less regular oscillatory behavior. Sudden changes in physical systems are gold mines for research. It turns out that the change shown here is a very important part of what I am studying. Any place this behavior is observed will be noted and focused on for a broad spectrum of different analyses.

The little-wiggles to big-wiggles transition is ubiquitous in physics.

Contour Plots Always Provide Good Summaries

The contour plot below succinctly summarizes many different processes that occurred in that plasma. From some of my posters you can see that contours provide a large amount of information in a single display. Two distinct regions are labeled below, but that is an over-simplification of what the image represents. The “Neat Behavior” is fairly well understood and actually expected. The “Thesis Behavior” region, though looking much simpler and possibly more boring, is full of many more unresolved issues. This complete data set will hopefully provide the means by which some of these outstanding issues might be solved.

It is funny that the left side, which seems more complicated, is actually better understood than the right side.

What Next?

Until next time... I'll be sitting at my computer trying to make sense of this and all the other data we have collected over the past couple years.