Research On Acoustic Emission Detection Technology Of Aluminum Plate With Holes Based On Improved Time Difference Mapping Method

In recent years,guided wave-based structural health monitoring(SHM)technology has continued to develop.Acoustic emission(AE)detection technology is an important non-destructive testing(NDT)method.Due to its characteristics of detecting dynamic defects and real-time monitoring of the safety of active board structural parts,the SHM technology has good application prospects.The aluminum plate structure with holes is widely used in the aerospace industry,automobile industry,shipbuilding industry,national defense industry and other fields.However,various defects may appear in its processing,manufacturing and transportation applications,which may pose hidden dangers to personal and property safety.So it is necessary to carry out non-destructive testing.Time difference mapping(TDM)method is a new type of AE source localization method.When using this method,there is no need to consider problems of wave propagation speed and various modes existing in the signal.The traditional Time Difference Mapping(T-TDM)method still has room for improvement in signal arrival time calculation,signal preprocessing and sound source localization algorithm.This paper takes the complex aluminum plate with holes as the research object,optimizes the T-TDM method,proposes an improved Time Difference Mapping(I-TDM)method,and it was used to conduct AE detection research on complex aluminum plate.The main research work is as follows:(1)Research on the calculation method of signal arrival time based on Akaike Information Criterion(AIC).When the traditional threshold method is used to calculate the arrival time of the signal,it is easy to introduce errors.In this paper,the method of identifying the signal arrival time in the T-TDM method is improved.The AIC method is introduced to identify the arrival time of the AE signal.(2)Research on AE signal preprocessing methods based on data cleaning,data scaling processing and data correlation analysis.In order to establish a more accurate time difference mapping database,it is necessary to introduce related data preprocessing methods in the TDM method.In the data cleaning,this paper introduces the Grobbs criterion to deal with abnormal data.The missing data is supplemented by the fitting and filling method.Data scaling and correlation analysis were performed.(3)Research on TDM localization algorithm based on Support Vector Regression(SVR),Bayesian Ridge Regression(BRR)and Multilayer Neural Network(MNN).The T-TDM method uses clustering analysis for sound source localization,which is cumbersome and easy to introduce errors.In order to improve the localization accuracy,this paper studied the sound source localization method based on the SVR model,BRR model and MNN model.It is verified by experiments that the three models have good localization performance.The three models were further fused into the SVR-BRR-MNN model.The results showed that the SVR-BRR-MNN model has an ideal localization performance.Compared with the single model,the fusion model has stronger generalization ability.So the algorithm based on the SVRBRR-MNN model is finally introduced as the sound source localization algorithm in I-TDM method.(4)Research on AE detection of complex aluminum plate.In this paper,the AE detection study was carried out on the aluminum plate with holes.The localization performance of the SVR model,BRR model and MNN model in the aluminum plate with holes was compared.The effect of the time difference grid size and sensor array layout to the I-TDM method was studied.When performing AE source localization research on aluminum plates with holes,the average absolute errors of algorithm localization based on SVR model,BRR model,MNN model and SVR-BRR-MNN model are 3.5mm,8.9mm,5.1mm and 3.6mm respectively.The relative errors are 0.49%,1.26%,0.72% and 0.51% respectively.When using grids of 100mm?100mm,50mm?50mm and 25mm?25mm,the localization errors are 11.3mm,3.8mm and 3.6mm respectively.Considering the difference in localization accuracy and workload,the grid size of 50mm?50mm may be considered when using the I-TDM method in the future.The localization errors of the selected regular and irregular sensor arrays are both 3.8mm.Whether the rules are arranged in the form of sensor arrays does not affect the localization performance of the I-TDM method.