The Research On Cracks' MFL Analysis And Signal Identification

Magnetic flux leakage (MFL) testing has been applied extensively in recent years, particularly the research and the expanding application to the corrosion inspection; magnetic flux leakage (MFL) testing has already been a comparatively mature inspective technique. But during the storage tank and pipeline service, there may also form a lot of more dangerous crack defect except corrosion defect. But resultantly a lot of factors to influence the examination result, this increases a large number of difficulties for MFL examination and evaluation to crack defect.This text with Magnetic flux leakage of the crack defect for research object, use academic analysis, finite element analysis, experiment research method to carry on the analysis for Magnetic flux leakage testing of the crack defect, and apply neural network to to carry on intelligent identify for the signal of Magnetic flux leakage testing of the crack defect. This text keeps the crack defect of tank floor plate simplification and turns into a few kinds comparatively suppleness analysis model in figure, namely the V shape section crack defect, rectangle section crack defect and combinational form section crack defect. Theoretically apply equivalent strip dipole to analyze magnetic flux leakage field for the crack defect, and with Maxwell equations for the theoretic foundation, and apply ANSYS the finite element software to make the simulation analysis of the finite element for magnetic flux leakage field for crack defect, to find out the crack parameter namely the depth, the width, the ratio of depth and width, the space between parallel crack etc. For the influence of the crack flux leakage field, and analyze exterior examination condition, such as sensor liftoff, the height of air. Aim at the actual circumstance for Magnetic flux leakage testing engineering, sets up the experiment system for the crack Magnetic flux leakage testing and pass experiment to identify the analytical result for spouse dipole model and finite element analysis.With applying ANSYS to carry though the simulation and analysis of the finite element, pick up characteristic quantity of the magnetic flux leakage field for crack defect, According to the simulation analysis data, pick up characteristic quantity of the magnetic flux leakage field for crack defect, and research the qualitative and quantitative connection between shape parameter of the crack defect and characteristic quantity of magnetic flux leakage field. With the quantitative relation foundation of characteristic parameter and characteristic quantity, found the geometry parameter of the crack defect to predict BP neural network model, and synthetically use the data which the finite element analysis gain and the data which the Magnetic flux leakage testing of the crack defect gain during the crack analysis as the training stylebook of BP network, equip BP neural network, analyze and make sure the parameter characteristic of the equivalent depth and width of the defect, and identify the dependability of the network to the forecast of the crack defect.