Moment Tensor Inversion Optimization And Its Application In Uniaxial Compression Of Jointed Sandstone

The study of rock failure mechanism is of great significance to the prevention and control of geological disasters in mines,slopes,tunnels and other geotechnical engineering.Moment tensor analysis based on acoustic emission technique can not only reflect the space-time evolution of micro cracks inside rock,also can obtain key information such as sliding direction,rupture source type,and so on.This method provides basis for the study of rock failure mechanism.It has been widely used in laboratory rock mechanics experiment and geotechnical engineering.Initial amplitude of P wave and AE localization is the basis of moment tensor analysis.In this paper,based on the project of national natural foundation of China,"study on the mechanism and constitutive model of macro and micro damage degradation of jointed rock mass under the wet and dry cycle"(51978106).In view of the shortcomings of the commonly used methods at present,the improved grey Wolf algorithm(BGWO)and the improved AR-BIC algorithm based on data preprocessing are proposed to optimize the solution of moment tensor and apply them to the analysis of the failure process of joint sandstone under uniaxial compression.The main research contents and achievements of this paper are as follows:(1)Both the traditional double-sequence AR-AIC and AR-BIC algorithms cause P wave arrival time recognize delay because of the complex amplitude characteristics of AE signals.The excessively short noise sequences in single sequence AR-AIC and ARBIC algorithms are covered by signal due to the low amplitude of the signal,resulting in recognize distortion of P wave arrival time.In this paper,an improved ar-bic algorithm based on data preprocessing is proposed.Meanwhile,taking into account the variance surge effect and introducing a reduction factor,a p-wave time recognition algorithm(ar-vic)based on the variance surge effect of autoregressive model is proposed.The accuracy of both algorithms reaches more than 98%,and the efficiency is 50% higher than the traditional methods.(2)Gray Wolf algorithm(GWO)is introduced and improved for AE localization.Combining with AE test,the residual value of travel-time equations is determined as the measurement standard of localization error.Aiming at the poor local search ability of GWO,an improved grey Wolf algorithm(BGWO)based on the elimination system of population memory was proposed.The results of an example show that,among the six positioning algorithms,BGWO,BGSA,GWO,GSA,Geiger and LS,BGWO has advantages in positioning accuracy,search speed and stability.(3)By combining all-phase phase difference method with Wavelet-EMD algorithm,more accurate delay estimation than threshold method is theoretically obtained.However,due to the inhomogeneity of rock medium and the reflection and refraction of joint surface,the results of time delay estimation are not ideal when the main frequency of the signal received by different sensors from the same source varies greatly.(4)Using the results of optimized moment tensor,and connecting with the stress-strain curve,time sequence correlation and AE characteristic parameters,analyzes the failure process of joints and sandstone under uniaxial compression,found that pressure dense phase is given priority to with shear fracture of random distribution of micro cracks,micro cracks after entering elastic stage first gathered in the weak position on both ends of the prefabricated crack,then gradually expand to the specimen ends,gradually fusion penetration,macroscopic failure happens in the end.It is also found that the mixed fracture in the whole shear failure is mainly concentrated in the macroscopic crack formation process,and the mixed fracture plays a role in promoting the crack fusion and transfixion.