| As the strategic passage of the traffic construction,efficiently and safely constructing of tunnel engineering has been gotten attention seriously.However,due to the fact that the geological condition of the tunnel surrounding rock is complex and changeable,unfavorable geologic bodies such as fault,fracture zone,karst and others can easily cause the geological disasters during construction such as water and mud gushing,rock failure,et al..Therefore,a vital method,i.e.tunnel space advanced prediction to guarantee the safety and efficiency of tunnel construction is carrying out,which can study the geologic condition in front of tunnel face accurately and prepare for the precaution in advance.The advanced prediction technology has been started in 1970 s,and the seismic reflection method is widely used in the engineering field attributed to the advantage of high precision and far-distance detecting.However,as the complex geological condition around tunnel space which makes the seismic wave field extremely complicated,it is difficult to identify and obtain a reliable imaging result.At present,the research on the tunnel seismic prediction(TSP)is mainly in two aspects: i.e.forward modeling and imaging.But there are still series issues need to a further discuss and learn: 1)for the forward modeling of tunnel space,the common method is finite difference algorithm,nevertheless this algorithm is often difficult to deal with the special arc free boundary,which cannot help to obtain the tunnel space seismic wave field with real response characteristics.2)Most of the research on TSP imaging is Kirchhoff migration or diffraction scan migration.However,reverse time migration(RTM)which is generally used and has the highest accuracy is mainly based on two-dimensional space,which has the shortage on the research in the 3D tunnel space RTM and survey layout.3)Till now,for the multi-component seismic data obtained from TSP,the main processing flow is based on the acoustic offset or the scalar offset of the P-wave and S-wave separation.Due to the complexity of the tunnel seismic data,it is difficult to get the pure P-wave or pure S-wave data through the wave field separation from seismic records,and the remaining energy can cause disturbing to RTM.Aiming at the problems that are worth further studying in TSP,we have developed finite element(FE)elastic wave field simulation and p-wave RTM which is based on elastic wave reverse time extrapolation in the 3D tunnel space.In this paper,the main work and achievements are shown as follows:(1)Based on two-order displacement elastic wave equation,FE equation of 3D elastic wave field is established using virtual work principle.And we have presented the implementation process of explicit Newmark finite difference algorithm with second order accuracy.The storage efficiency for sparse matrix was compared for row compression storage CSR and one-dimensional variant-banded storage.The FE simulation of 3D elastic wave field is realized under loading explosion source in unstructured grid,which provides the basis for the study of RTM.(2)Based on the theories of perfectly matched layer(PML),FE equation of split PML boundary condition based on the second order displacement wave equation and Non-splitting convolutional perfectly matched layer boundary condition(CFS-UPML)are derived,which is verified the correctness of the FE simulation of surface wave by comparing with analytical solution.The absorption effect and the stability of the calculation for a long time of two boundary conditions are compared under the environment with surface wave,grazing angle of incident wave and near surface high Poisson’s ratio media.Finally,the correctness of the FE surface wave simulation in the condition with 3D inclined free surface is verified.The results show that FE method with flexible mesh,automatically satisfying the free boundary conditions and other advantages can obtain more accurate surface wave records,and wave field simulation adapt to the complex environment;CFS-UPML boundary conditions on grazing angle of the incident wave absorption effect and in the high Poisson’s ratio media long time stability have more advantages,but the computational efficiency is lower than the PML boundary.(3)According to the structure characteristics of common unfavorable geological bodies(lithologic interface,cave)discovered in tunnel processing,three-dimensional tunnel geological geophysical models are built.At meantime,3D tunnel space elastic wave full wave field simulation is realized,and analyzes the field characteristics,which enhances the systematical understanding of wave propagation characteristics in tunnel space,and provides a theoretical basis for the identification of tunnel seismic reflection information.Analysis results show: when the shot points are located at tunnel’s sidewall,S-wave reflection energy generated by lithologic interface is stronger than P-wave reflection relatively;While the detector disposed on the tunnel sidewall near dipping interface,obtained P-wave reflection energy is stronger than the other tunnel sidewall,which is more conducive to P-wave reflection imaging;When cave is filling with air,detectors gain strongest reflected wave energy than that filled with water,and cave filling with soft soil produces weakest reflected wave energy.(4)After the analyzing the basic theory of elastic wave RTM,the key steps(wave field separation,imaging conditions and noise suppression for imaging)in RTM are discussed and analyzed in detail.Then feasibility of the tunnel space RTM is verified.Results: the method based on Helmholtz decomposition can be used to decouple P-wave and S-wave in the tunnel wave field effectively;The implementation of cross correlation imaging condition based on the boundary memory strategy can effectively solve the problem of memory consumption in 3D RTM;The method based on Poynting vector and Laplace filter theories can effectively eliminate the low frequency noise interference in RTM.By discussing,the technological process for three-dimensional elastic p-wave TSP based on multi-component seismic data-elastic wave field extrapolation-p-and s-waves field separation-elastic p-wave imaging are established.(5)Based on the research of tunnel space 3D FE modeling and elastic wave RTM imaging,tunnel space 3D p-wave RTM imaging based on elastic wave reverse time extrapolation is realized.Factors(i.e.Survey layout and noise)affecting the imaging results is discussed by using the models which have interface with 60°dip angle.The results show that(1)the methods based on reverse time extrapolation can obtain an efficient tunnel space imaging results;(2)two-sided survey layouts have a preferable performance applied to the condition with dip interface,and one array can satisfy the requirement of imaging accuracy showing in the theoretical imaging records.In this research another conclusion by discussing the impacts of noise demonstrates that(1)when ratio of noise is no more than 50%,the imaging profiles can be considered as efficient;(2)the energy of false lineups increases constantly while the ratio is growing;(3)only when the ratio is greater than 70%,the false lineups can disturb the imaging results.This demonstrates that the RTM methods adopted in this research have a better function in noise immunity;(6)According to the 3D TSP flow mentioned above,imaging research for tunnel space unfavorable geologic bodies(whose angle between the bodies and the tunnel axis is 90 degree)is developed.The results show: the 3D elastic p-wave RTM imaging methods based on multi-component seismic data and elastic wave reverse time extrapolation can get a preferable performance when unfavorable geologic bodies are in front of the tunnel face.By comparing the imaging results for dipping interface and caves located at side face,the imaging energy of geologic bodies whose angle between the bodies and the tunnel axis is 90 degree is weaker relatively.Therefore,in the case of high background noise,it is hard to get accurate imaging results when geologic bodies whose angle between the bodies and the tunnel axis is 90 degree.(7)Through the imaging results of noise records,it shows that the method adopted in this paper has a certain noise immunity.The feasibility and effectiveness of the methods are demonstrated by real TSP data based on discussed survey layout and imaging methods.In this paper,the innovative achievements are mainly show as follow: The absorbing boundary condition based on two-dimensional split PML is developed in three-dimensional space efficiently,and tunnel space FE full wave field modeling based on the second order displacement wave equation for split PML boundary condition and CFS-UPML are realized.Three-dimensional tunnel space RTM is realized effectively and key factors affecting the imaging results is discussed and analyzed systematically.The survey layouts satisfied the requirement for RTM are built.Through theoretical simulation and case study,the technological flow for three-dimensional elastic p-wave TSP based on multi-component seismic data-elastic wave field extrapolation-p-and s-waves field separation-elastic p-wave imaging are established. |
PDF Full Text Request