Time Domain Acoustic Equation Reverse Time Migration And Constructing Angle Gathers

In recent years, with the rapid development of computer technology and theimprovement of the complex structural reservoir imaging requirements, reverse timemigration gets more attention from the academia and industry for not restricted bystratal dips and better fidelity, and became a research hotspot of the currentexploration geophysics. Although reverse time migration is a method with higherprecision, the excessive calculation, large buffer memory and low-frequency noiselimited the development of reverse time migration.This paper summarizes the current storage strategy which are frequently usedaim at problems of the storage content in reverse time migration. Based on theefficient boundary storage strategy, we redefine the efficient boundary area and givethe modified efficient boundary storage strategy, which can reduce storage of reversetime migration efficiently without increasing any additional computation, byanalyzing the finite difference scheme. The result of test with Marmousi model andSEG/EAGE salt dome model shows the effectiveness and practical applicability of themodified efficient boundary storage strategy.Low-frequency noise affects the imaging quality of reverse time migrationseriously. Choosing the appropriate low-frequency noise attenuation method is animportant aspect involved necessarily during reverse time migration procedure. First,we analyze the generation mechanism of low and medium frequency noise in reversetime migration, as well as analyze and classify all kinds of low-frequent noiseattenuation methods, we also conclude advantages and disadvantages of them.Decomposing the full wave field into up-going wave, down-going wave, left-goingwave and right-going wave with the Poynting vector realizes imaging with the normalized wave field decomposition cross-correlation imaging condition based onthe Poynting vector, and therefore attenuating the low-frequent noise and improve theimaging accuracy. The trial with Marmousi model and SEG/EAGE model shows theeffectiveness of the decomposition of up-going wave, down-going wave, left-goingwave and right-going wave with the using of the Poynting vector in the process of thewave field propagation. And the result of the test illustrates the normalized wave fielddecomposition cross-correlation imaging condition based on the Poynting vector canattenuate the low-frequent noise better and acquire higher imaging accuracy of reversetime migration.There are other kind of structural false appearance in the result of reverse timemigration imaging, besides the low-frequent noise above. In this paper, according tothe condition that reverse time migration can image accurately, which means thereceiver back propagation wave field and shot forward propagating wave field aresame in the propagation path and the time. Reverse time migration dose reverseextrapolation to seismic records received by different receiving direction, differentlength and different offset from the detection point, getting the reverse propagatingwave field of the receiver, and comparing with the shot forward propagating wavefield, analyzing the reason why the shot forward propagating wave field and thereceiver back propagation wave field differences existed. According to the analysis ofthe model data result, we know that the seismic data can not be received in alldirections, resulting in new wave field during the reverse extrapolation process, andfalse appearance of the reverse time migration imaging. In addition, the influence ofchannel spacing on the reverse time migration imaging is less than which of arraylength.Although the reverse time migration is good at improving complex structuralimaging quality, the method itself requires high velocity accuracy. The commonimage gather, which can provide velocity information for depth migration processing,is a kind of important output gather of pre-stack depth migration, In this paper, wecarry out the method of picking up angel domain common image gather in the reversetime migration process, and this method can calculate reflection angle with Poynting vector. Finally, the angle domain common image gather can be acquired directlyaccording to the given imaging condition. With less additional calculation and storagecontent, this method is propitious to migration velocity analysis. The simple modeltrial tested and verified the sorting of the angle when Poynting vectors pick up angledomain common image gathers. The trial result of picking up angle gather with thismethod on SEG/EAGE2D salt dome shows the effectiveness of this method.