Multi-domain Feature Fusion Damage Assessment Of Lamb Wave Structural Response For Composite Plates

Composite materials are widely used in modern industrial manufacturing due to their light weight,high strength and chemical resistance.However,the complex damage mechanism of composite materials makes it very difficult to diagnose the structural damage,especially to assess the damage degree.Therefore,it is of great scientific and engineering importance to establish an effective damage monitoring and assessment system for the timely detection of safety hazards in order to address the difficult damage assessment problems faced by structural health monitoring(SHM)studies of composite materials.In this paper,based on the study of the propagation mechanism of Lamb wave in composite structures,by analyzing the multi-domain feature parameters of the extracted damage response signals and enriching the damage-sensitive information sources,this thesis pro poses the use of sparse arrays combined with machine learning methods to realize the fusion of damage features,establish damage diagnosis models,and then realize the localization and degree assessment of composite material damage.The main research of this paper is divided into the following sections:(1)The basic methodological theory of Lamb wave-based SHM technology is discussed,including the propagation process and dispersion characteristics of Lamb waves,the principle of damage monitoring based on piezoelectric sensors,and the main methods of damage degree assessment.(2)The influence of typical damage of composite material on Lamb wave response signal in the time domain,frequency domain,and the time-frequency domain are analyzed,and the mechanism and method of extracting damage sensitive information by multi-domain feature parameters characterization are given.The damage feature vectors are extracted and constructed by multi-domain signal analysis for response signals with different damage degrees.(3)The damage assessment method of composite materials based on the sparse array and machine learning is studied and designed.Firstly,a sparse sensor array method is designed for real-time signal acquisition;Secondly,based on Fisher linear discriminant analysis,the optimal path monitoring scheme is proposed to improve the monitoring efficiency of the sparse array,and the data is further reduced in dimensionality by combining with principal component analysis(PCA);Finally,a damage assessment method based on support vector machine(SVM)is studied to achieve multi-domain damage feature fusion.A sparse array-based multi-domain feature fusion damage diagnosis model is established,including the damage localization model,and the damage degree assessment model.The damage is first localized and then evaluated.Experimental studies were conducted on structural objects of glass fiber reinforced epoxy resin composite plates.The classification evaluation and experimental verification of typical quantitative delamination damage of different degrees show that the multi-domain feature fusion damage diagnosis technique based on sparse arrays proposed in this study can accurately achieve the localization and quantitative damage degree evaluation of artificial delamination damage of composite plate structures.The research results provide an innovative approach to the difficult problems faced by SHM of composite materials,which has potential engineering application prospects.