Study On Development Of Automated Eddy Current System And Defect Characterization Method For Quality Control Of Valve Push Rod

The valve putter,one of the members of valve-train,act as a transmitter that pass the thrust of camshaft to rocker mechanism through the valve lifter and the valve putter.The valve putter in manufacturing tends to appearing many problems such as quenching crack,low hardness,round beating,which will reduce its mechanical properties.While working the valve putter with slender shape is easy to be affected by cycle effect of alternating force like reciprocating jacking as well as pressure transmitted from cam shaft and prone to warping distortion or even broken,so the product quality of the valve putter plays an important role in safe car driving.At present,the valve putter is usually detected by the manual.There are some deficiencies in it,such as low efficiency,easy to leak,hard to trace test results,etc.Aiming at the product quality problem of the valve putter existing in the manual detection,this paper proposed a new method that consists of the application of eddy current method to detect cracks and hardness of the valve putter and the application of an inductive displacement sensor to measure circle run-out and length of the valve putter.In the paper a set of valve putter automatic detection system,which can accurately detect the parameters of surface cracks,circle run-out,length and ball hardness of the valve putter,is designed to achieve automatically sorting of qualified and unqualified valve putter products and to create a good economic benefits for enterprises and users.Its main achievements and innovation points are as follows:First of all,length and ball hardness of the valve putter,a set of dedicated valve putter automatic detection system which mainly include mechanical structure,electrical control,detecting instrument and PC software is designed according to detection requirements of surface crack length,circle run-out.100 roots are randomly selected from the qualified valve putter with manual detection on the spot,and 5 roots are identified to be unqualified by the system with duplicate confirmation.Test results confirmed the reliability of the automatic detection system.Then,based on Maxwell equations and the finite element theory,the finite element model of eddy current testing signal in the valve putter is created.The influence law of the ball head nose lift and the crack parameters on the eddy current testing signal is simulated,which provides the theoretical basis for optimizing the system parameters setting and experimental research.Furthermore,the influence law of excitation frequency on the crack signal is studied from the aspects of sensitivity,defect signal feature distribution and classification accuracy.It is found that the excitation frequency of the maximum defect is the best.Based on this,an excitation frequency optimization method based on experiment is proposed.The results provide theoretical support for the excitation frequency setting of the valve putter automatic detection system.Finally,a flaw signal feature extraction algorithm based on PCA and KPCA is proposed,and a defect classifier based on ANN and SVM is constructed.The experimental results show that KPCA-SVM has the best classification performance.The research results of this paper provide a theoretical basis for the automatic detection of the valve putter,specify the next research direction at the same time,which has important theoretical and application value.