Quantitative Nondestructive Testing Of Mechanical Properties In Ferromagnetic Materials Based On Incremental Permeability

Nowadays,an increasing number of new non-destructive testing technologies for ferromagnetic materials have occurred,owing to the wide use of these materials in industries.The method of incremental permeability(MIP),one of the new methods,is based on permeability,which has a keen connection with the material properties such as residual life,tensile strength and fatigue resistance.Thus it can be used in the testing and evaluation of the mechanical properties in ferromagnetic materials.In this project,part of the mechanical properties was quantificationally tested based on incremental permeability.Firstly,a testing system based on incremental permeability was designed.In the system,the type of exciting signal and signal collection were designed,including the structure and the parameters of the pick-up probe.What's more,two modules,the constant current source module and the lock-in amplifier one,were designed and produced with the type of circuit boards to meet the need of signal detection.Then a large number of MIP signals were obtained after the experiments using the system having been conducted.In the period of signal analysis,a deep research about the feature extraction of MIP signal has been done.Thus,characteristics were extracted from both butterfly curve and impedance plane,which was initially used in eddy current testing.And then,preprocessing methods for the data was studied and executed in the paper.Finally,neural network models were established by the processed data,aiming for testing the mechanical properties in ferromagnetic materials based on MIP quantificationally.It can be found that when the specimen set is complete,the model made by BP neural network had the ability to predicate the mechanical properties wanted precisely.However,when it comes to an uncompleted specimen set,BP neural network had a poor performance.In that case,the network optimized by PSO showed a better performance,including a higher precision and stability.