Research On Composite Material Crack Identification Method Based On Natural Frequency And Neural Network

With the progress of science and technology and the development of society,’ energy conservation and environmental protection ’ has become a topic of widespread concern in society.T700 epoxy fabric composite has become an ideal material for lightweight rail transit due to its advantages of light weight,high strength,strong designability,corrosion resistance and good molding process.However,due to the influence of the manufacturing process and overload bearing,the composite material is prone to cracks during use.In order to avoid the occurrence of safety accidents,it is particularly important to monitor the health status and identify the early crack state of composite materials.Based on the dynamic theory,this paper studies the crack identification of composite structures.Firstly,by optimizing the material parameters of the finite element model of the composite material,the simulation model approximates the simulation test model.Further,the corresponding relationship between natural frequency and crack condition is constructed by simulation analysis.Finally,the neural network is established by using the natural frequency to realize the reverse identification of the crack condition.The main contents include:(1)Theoretical study on vibration characteristics of T700 composites.Based on the dynamic theory,the relationship between the stiffness and natural frequency of the composite material is studied,and the natural frequency is used to characterize the vibration characteristics of the structure.(2)The finite element approximation model is established based on the experimental data.Based on the modal characteristics measured by the test,the material parameters of the composite material are optimized to make the modal parameters output by the simulation model approach the modal parameters measured by the test,so as to realize the approximate simulation of the test model with the simulation model.(3)Construct the corresponding relationship between identification index and fracture condition.By solving the finite element simulation model under different crack positions and depths,the variation law between the characterization vector and the crack condition is analyzed,and the natural frequency is used to characterize the crack condition.(4)Research on crack identification based on neural network.The neural network is used to construct the natural frequency database under different crack conditions to realize the reverse identification of cracks,and the anti-noise analysis of the training grid is carried out considering the influence of environmental noise.