Even NASA Screws Up

The Windtunnel was damaged. I suppose that's slightly deceptive, its not nearly as dramatic as previous tests have been (there have been instances where the entire test apparatus has been torn off the test stand, but I can't show you the video because of security regulations). The damage that occurred recently is merely a tail strut failure on the Large Rotor Test Apparatus (LRTA). I can't show the pictures of the damage but I think this picture will help.

So the strut that broke was the one in the back of this image. It should also be noted that this is the 80x120 foot windtunnel not the 40x80 foot windtunnel where the LRTA presently is installed. Since the tail strut was broken, all the tests have been stopped. This is a bad thing as there is a time limit on how long we can use the wind tunnel before the navy decides to kick us out so they can test their own project. For those of you impressed by big numbers the LRTA houses two 3000 horsepower electric motors (2,237 kW for those of you who prefer the inferior metric system)

So my work for the past week has been focused on analyzing the data gathered from the hover runs (the air flow in the wind tunnel is off but they spin the rotor up to simulate hovering conditions). My job has been to sift through the data collected from each run and determine the quality of the data gathered. For each run there are about 20 million data points gathered, so its by no means a simple task to organize and graph all of that data. That's where MatLab comes in to the picture. Last year a couple of interns wrote a program that would quickly plot all of the data gathered for each run. The program is fittingly named QuickPlot. It does a great job of plotting the data, sometimes. Its not the most reliable program and it is very peculiar about how you use it. So my end goal is to really develop the program to make it more adaptable and make an automated channel quality checker. My present approach has my program computing the Fast Fourier Transform (FFT) of the data and then determine the magnitude of various frequencies in the FFT. This works for detecting some types of channel failure but its by no means perfect, oh well good thing I have more time to work on my program while I'm here.