Research On Interference Fit Condition Monitoring Method Based On Laser Ultrasonic

Interference assembly is a way to realize assembly through interference fit between workpieces.Because of its simple structure and good neutrality,it is widely used in various mechanical equipment.The quality of interference assembly determines the service performance of equipment.Due to the effects of oscillation,impact and load during service,there will be stress concentration and eccentric load,resulting in loose interference fit and even fretting cracks at the edge.Therefore,it is necessary to test the quality of interference fit.In this thesis,the connection state of interference fit is evaluated by laser ultrasonic method,mainly including the assembly state of interference fit,namely whether it is loose;and detect the contact state and edge crack of the joint surface.And through the method of deep learning,small cracks are detected and identified.Firstly,the thesis discusses the theoretical basis of ultrasonic guided wave detection of interference fit state.According to the propagation characteristics of guided wave through the interface,the idea and technique of laser ultrasound detection of interference-fit state were proposed.Determine different excitation receiving methods,explore the relationship between guided wave energy and interference based on the wave energy dissipation method,so as to realize mutual verification between the results of different detection methods.Secondly,according to the theoretical analysis,the experimental process is designed,the laser ultrasonic interference assembly quality inspection platform is built,and the samples and parts are prepared.The ultrasonic guided wave is excited by laser on the surface of aluminum plate to achieve the purpose of non-contact excitation.The power spectrum density(PSD)is used to calculate the guided wave power under different interference amounts,and an evaluation method of interference assembly state based on guided wave energy is proposed.The micro crack at the contact edge is identified by obtaining the waveform propagation diagram of the whole field,and the crack is located by coupling the scanning area and crack position.The relationship curve between guided wave power and interference is obtained by experimental verification.This method can not only obtain the interference assembly quality,but also realize the identification and location of micro cracks through the obtained whole field waveform propagation diagram,which provides a new method for the state detection of interference assembly structure and the detection of micro cracks.Finally,the deep learning method is introduced to intelligently detect cracks with different sizes.The crack images detected by guided wave visualization technology are labeled to construct the data set.The One-stage target detection algorithm-YOLOv5 neural network is introduced as a deep learning model for crack detection.The feasibility of the model is verified by training experiments,and the detection and recognition of small cracks are realized.In summary,through the experimental study,the relationship curve model between interference and guided wave energy is established,and the propagation law of ultrasonic guided wave in interference fit structure is revealed.The feasibility of ultrasonic visualization for edge crack detection of interference fit is verified.The intelligent detection of interference fit edge cracks is realized through deep learning.The research results show that the ideas and methods proposed in this thesis have reference effect on the health monitoring of interference fit structure.