Damage Identification Of Concrete Specimens Based On Acoustic Emission Multi-parameter Analysis

Concrete structures are susceptible to damage in service,and early defects are often not easily detected and,if left untreated,will inevitably affect the safety and stability of the overall project operation.Acoustic emission,as a dynamic,real-time non-destructive testing technique,is extremely sensitive to the signal of crack initiation and expansion within the structure,and plays an important role in damage assessment.This paper takes concrete material as the object,around structural damage analysis and identification,and conducts four-point bending continuous loading acoustic emission tests on plain concrete beams and four-point bending graded loading acoustic emission tests on reinforced concrete beams to study the change characteristics of various acoustic emission parameters in the process of concrete material damage evolution,the main results are as follows:（1）In order to study the real-time change law of acoustic emission parameters during specimen loading,an improved acoustic emission ba-value algorithm is proposed.According to the characteristics of the negative correlation between the ba-value and the stress,it is believed that the ba-value can be used as a parameter index to judge whether serious damage occurs to the solid concrete beam under continuous loading conditions.（2）In order to study the damage evolution process of the specimen,based on the theory of acoustic emission rate process,the polynomial fitting of the cumulative number of acoustic emission events corresponding to the force level and acoustic emission is performed,and the damage process of the plain concrete beam and reinforced concrete beam is quantitatively expressed.According to the change law of the probability density function,the specimen destruction process can be divided into three damage stages.（3）In order to verify the rationality of the division of damage stages,a BP neural network method based on genetic algorithm optimization is proposed,and the sample databases of plain concrete beams and reinforced concrete beams are established with acoustic emission parameters as the objects,and the identification and classification are carried out,and the results show that the acoustic emission parameters of each damage stage have certain regular characteristics.（4）In order to study the characteristics of acoustic emission events of different failure types during specimen loading,a method based on Gaussian hybrid model（GMM）for RA-AF scatter clustering is proposed,which realizes the distinction between tensile and shear failure events of prime concrete beam and reinforced concrete beam during damage failure.Through the two-dimensional and three-dimensional probability density cloud maps,the characteristics of the two destructive types of acoustic emission events under each stress level can be intuitively reflected.