Research On Structural Defect Imaging Based On Lamb Ultrasonic Guided Waves

With the development of society,there have been increasing concerns about the safety and reliability of the key structures of the large infrastructure,aerospace spacecraft.The non-destructive structure health monitoring and defect imaging technology based on ultrasonic guided wave have received more and more attention.Structural Health Monitoring is a kind of interdisciplinary emerging technology,it can obtain health status of the engineering structure and equipment structure through the sensor arrays.It is of great significance to ensure the safety of key structures,to avoid the occurrence of major catastrophic accidents and to ensure the safety of people's lives and property.Based on the research background of National Natural Science Foundation of China,aiming at the research direction of nondestructive structural health monitoring and defect imaging technology based on ultrasonic guided waves,this paper focuses on the theories of elastic waves,the extraction of flaw signals,minimum variance defect imaging,the sensor layout optimization based on Bayesian probability distribution,the method of guided wave imaging adaptive parameter estimation and compensation,analysis of algorithm and experimental verification,etc.The research has achieved good results and obtained some innovative research achievements.The research work and main contributions are as follows:(1)Based on the analysis of Lamb wave dispersion and multimodal propagation characteristics,this paper describes the main technical problems about dispersion,multimodal and baseline signals acquisition for guided wave defect imaging,analyzes the technical characteristics and shortcomings of traditional imaging methods,puts forward an improved imaging method based on the envelope curves and the position weight coefficient.An evaluation indicator of the guided waves image effect is also put forward in this paper.(2)Aim at the minimum variance imaging method,based on the minimum variance distortionless response,this paper deduces the optimization condition of the weight coefficient from the imaging principle and uses the Lagrangian multiplier method to optimize the weighting factor by the space-time cross-correlation matrix decomposition.The diagonal loading for weight coefficient is used to enhance robustness against modeling error.In this paper scattering characteristics of the guided waves also are studied in order to improve the image of interference effect.The complex degree of solving the pixels is reduced by introducing proper window function and vectoring guided wave variables,which could improve the real-time performance and make it fit software solving.(3)Based on the Bayesian Risk Decision-making Analysis,Optimal Sensor Placement Algorithm is studied in this paper.According to the relationship between the costs in the structural health monitoring and states including the predicted state and the real damage state,the summation of the global state domain is transformed into a discrete region that can be processed independently,which reduces the processing difficulty.The cost function is simplified by dividing states into "no damage" and "damage" binary state.The performance indicator of the exciting-receiving channel is established and its solution relation is confirmed.Based on the above researches,the improved genetic algorithm is used to realize PZT(Piezoelectric Ceramic Transducer,its abbreviation is PZT)sensor array position optimization.(4)The important parameter estimation algorithm of guided waves imaging is studied in this paper.Firstly,the generalized model of ultrasonic guided waves propagation is deduced in the frequency domain.The phase noise and logarithmic amplitude noise of the guided waves detection system are studied by using the probability distribution function.The probability distribution of logarithmic amplitude noise and phase noise is established Function,an estimate of the noise variance is established from the measurement signals.The distance estimation could be obtained by projecting a pre-measured propagation distance onto two or more standard orthogonal vectors,the other model parameters are solved by using Moore-Penrose generalized inverse matrix in zero space of the A matrix.Based on the integer constraint condition of model,an improved nonlinear search is used to solve the zero space coefficient,so the all parameter estimations of the guided wave model are obtained.The numerical simulation and experimental study show the effectiveness of the algorithm.(5)The combination of parameter estimation and minimum variance imaging is realized in this paper.First,the algorithm of frequency mapping domain into the wavenumber domain is used to realize the dispersion compensation.The parameter estimation algorithm is used to compensate the transfer function which is susceptible to environmental impact.Therefore,the interferences are isolated in the baseline subtraction operation,which avoid artifacts caused by environmental changes,without the need to pre-establish baseline signal database under same environment.This paper shows the effectiveness of the parameter estimation algorithm to the minimum variance imaging by means of the time domain signal coincidence contrast,the distance domain estimation error waveform coincidence contrast,the wave number domain coincidence contrast and the frequency domain signal characteristic analysis.An image processing method of pixel block is explored.The average threshold is used to improve the local detail for the image quality.(6)Aiming at minimum variance imaging algorithm,the simulation test analyses are carried out,at the same time,method validation,experimental comparison and analysis work are developed.The simulation and experimental results show that the proposed algorithms in this paper are feasible and effective,this research work reflects the significance of good technical methods and provides an important technical support for the practical application of the relevant theoretic.