Summer 2009 - Magnetic Flux Leakage Analysis
This research was conducted during a summer internship at Rice University in the RiSYS Engineering Lab. I worked with a graduate student, Andrew Lynch, under Professor Fathi Ghorbel. During this internship, I focused on pipe inspection robotics using internal and external methods. These methods primarily involve magnetic flux leakage (MFL) to detect changes in ferromagnetic structures. I developed a novel method for analyzing MFL data to detect defects in pipes.
Documentation
The Defect Locator and Axial Averaging Algorithm (DLAA) uses the maxima and minima of the radial magnetic flux data to located defects and find the average of corresponding axial data. The basic process of the algorithm begins with a moving average filter to minimize radial noise. Then, the maxima and minima of the radial magnetic flux data are located. These maxima and minima denote defects on the pipe. Next, the mean of the axial magnetic flux data is determined for each defect. Finally, a wall thinning curve is produced by relating the means for each defect and the corresponding wall thicknesses.
See my paper for more details on the algorithm.
See my paper for more details on the algorithm.
