The Study Of Pipeline Magnetic Flux Leakage Signal Processing Based On Three-Axis Sensors

The safe operation of steel pipe is related to the quality of steel pipe, the pipeinspection is a necessary means to confirm the quality of steel pipe and reduceaccidents.There are many conventional methods for steel pipe defects detection, such aseddy current testing, magnetic particle testing, magnetic flux leakage testing, ultrasonictesting.Magnetic flux leakage testing method has been widely applied with the advantagesof fast detection speed, high detection rate, high sensitivity and simple operation.With thedevelopment of modern industry, people put forward higher requirements for steel pipeinspection system.Steel pipe detection system not only can detect defects in steel pipe, alsoshould have high accuracy and precision,also hope to be able to identify more types ofdefects.The traditional one-dimensional sensor is insufficient in this respect.Therefore, thisdissertation conduct a study on the improvement of conventional magnetic flux leakagetesting methods of steel pipe defects.The main work and contributions are as follows:First of all, this dissertation put the three axis sensor technology into the magnetic fluxleakage detection system, and study the location of sensor in a probe.By using a fixed threeaxis sensor probe array, steel pipe get through the sensor array to scan the steel pipe, so thatthe probe can collect signals of magnetic flux leakage.This dissertation also puts forwardthe basic composition of three axis sensor magnetic flux leakage detection system.Then, this dissertation conduct the study on two signal processing tasks of three axissensor magnetic flux leakage signals: denoising and singularity detection.Wavelettransform has a excellent perform on time-frequency analysis, this dissertation use thewavelet threshold denoising theory to denoise and modulus maxima theory to detect thesingularity point of three axis sensor magnetic flux leakage signals.Getting the results ofmagnetic flux leakage signal denoising and singular point detection of several defects withspecial parameters through the relevant software simulink technology, proves the feasibilityand accuracy of this method, also proves the superiorities of three axis sensor has thanconventional one dimensional sensor: data is more comprehensive and the accuracy ishigher.Finally, this dissertation conduct the research on algorithm flow and hardware basis ofsignal processing subsystem, and put forward the implementation diagram of signal processing subsystem.