Vibration Detection Of Composite Pultruded Multi-cavity Plates Based On Wavelet Packet Transform

With the progress of the times and the vigorous development of science and technology,the composite materials play a very important role in daily life.Especially in the fields of aerospace,automobile industry,machinery manufacturing,medicine,civil engineering,energy and electronic industry.Nowdays for the simple composite material profile processing more use of pultrusion molding process,this process method under the action of the traction machine continuous fiber resin infiltration and through the molding mold heating curing molding.In the pultrusion production process,temperature,traction speed,traction,pressure and other production factors have a certain impact on the product,these unstable factors will cause the product curing incomplete,partial defects.For pultrusion technology products mostly by experienced technicians in front of the mold with experience to prevent defects in advance,but the poor stability of test methods,at the same time also have no way to guarantee production of finished product whether there is a defect,is not only a certain effects on the yield,will also affect the reliability of the product,more serious defects will cause the production line fully parking,cause huge economic losses.Therefore,it is necessary to find a method for online detection of multi-cavity profile,so that the production line can form a closed loop control and ensure the stable production.In this paper,the composite material pultrusion multi-cavity plate is tested by percussion detection technology.1.This paper introduces the research status of composite material forming process and non-destructive testing technology at home and abroad.2.The principle of knock detection technology and the data acquisition system based on Lab VIEW are introduced,and the experimental scheme is designed based on the principle of knock detection technology.3.The vibration signal preprocessing,Fourier transform and feature extraction based on wavelet packet transform are introduced.By eliminating polynomial and five-point cubic smoothing preprocessing method,the vibration signal is more smooth.Then the energy spectrum of the vibration signal of the composite pultrusion multi-cavity plate is analyzed by wavelet packet transform,and the energy ratio is taken as the eigenvector of the composite pultrusion multi-cavity plate,and the eigenvector is normalized.4.This paper introduces the support vector machine(SVM),takes the feature vector as the input of SVM,and then classifes and identifies the normal and fault composite pultrusion multi-cavity plates.The normal and fault composite pultrusion multi-cavity plates were classified and identified by support vector machine,and the recognition rate reached86.36%.The results show that this method can be used for online testing of pultrusion forming process,which overcomes the problem of low reliability due to product lag and is believed tohave good practicability in practical application.5.The fault classification effect of composite pultrusion multi-cavity plate is analyzed with different support vector machines(SVM)for the sensor installation position.When the sensor is installed in the middle of the composite pultrusion multi-cavity plate,the classification effect of the failure of the composite pultrusion multi-cavity plate and the normal composite pultrusion multi-cavity plate reaches 55%,and the classification effect of the failure of the cavity beam plane and the failure of the cavity force bar of the composite pultrusion multi-cavity plate reaches 75%.When the sensor is installed at the left-most end of the composite pultrusion multi-cavity plate,the classification effect of the failure of the composite pultrusion multi-cavity plate and the normal composite pultrusion multi-cavity plate reaches 85%,and the classification effect of the failure of the beam plane of the composite pultrusion multi-cavity plate and the failure of the pultrusion multi-cavity plate reaches 100%.The results show that the sensor installed on the two ends of the composite pultrusion multi-cavity plate has the best classification effect,and the sensor installed on the middle of the composite pultrusion multi-cavity plate has the worst classification effect.6.The work of this research is summarized and the prospect of this research is presented.The innovation of this research lies in that there is no research scholars to conduct real-time detection of composite pultrusion multi-cavity plates.