Study On Adaptive Subtraction Of Different Types Of Diffraction Multiples

Formation of multiple wave has to meet certain conditions, for example, on thesea surface or seafloor, reflection coefficient is strong on these reflective surface,when wave impedance of medium is strong, to form multiples recorded in theseismic record. When the reflected wave goes back to the ground, due to the waveimpedance of the interface of sea and air or interface of ground and the air are bothstrong, the reflected wave is propagated down from the free surface. When itencounter reflection interface, it is reflected again and reflected on free surface, maybe repeated propagation on these two interfaces, and finally form the surface-relatedmultiples. The multiple, as a kind of coherent noise, always disturbs the seismicacquisition, processing and interpretation work. In particular, in marine seismic data,the energy of the surface-related multiples is the strongest and has the most influence.The requirements of seismic data increase more and more in oil industry, and thatalso leads the high demand of the multiple suppression technology. As a result, theconventional filter-type approaches are very difficult to meet the need.Surface-related multiples are common in marine seismic data, which is anobvious coherent interference and its existence seriously affects the seismic imagingauthenticity and reliability, and seriously influence seismic interpretation work. Howto suppress the multiples and retain energy of the primary wave with high quality isso important for a series of follow-up work. Reflection multiples and diffractionmultiples exist in the seismic data in the same time, because energy of diffractionmultiples is weaker than that of reflection multiples, diffraction multiples has a moreobvious nonstationary characteristics, conventional methods are difficult toeffectively eliminate it, thus its exist will badly affect the seismic data processing,and even cause a great interpretation error. Therefore, according to conditions andcharacteristics of reflection multiples and diffraction multiples, we separatereflection waves and diffraction waves in the seismic data and process themseparately, the authenticity and reliability of the seismic imaging will be improved. The surface-related multiple elimination(SRME) is a process that removes theinfluence of the surface reflectivity from the data. It is feedback-iteration approach,belonged to the prediction-subtraction approaches, and based on the free surfacemodel and layer surface model. An important property of the proposed multipleelimination process is that no knowledge of the subsurface is required. The SRMEmethod is more effective than those simple conventional filter-type multiplesuppression approaches, where the moveout properties of primaries and multiples arevery similar, as well as for situations with a complex multiple-generating system.In this work, tests of two kinds of models are used, we do wave field separationusing three methods: the plane wave destruction filter, structure-enhancingtime-varying median filtering, and structure-enhancing triangular smoothing filter.We choose a better effective approach to combine with the prediction of SRMEmethod and nonstationary adaptive attenuation method. Then, we do predictionusing SRME, and predict three kinds of multiples: reflection-reflection multiples,reflection-diffraction multiples, and diffraction-diffraction multiples, then match thedifferent types of multiples using nonstationary adaptive attenuation method,especially for nonstationary diffraction multiples, and finally reflection multiples anddiffraction multiples will be eliminated from seismic wave field. Using fiveprocessing strategies of adaptive matching in this work, separation of reflectionwaves and diffraction waves can influence suppression effects of multiple, whencomparing nonstationary adaptive attenuation with conventional method of leastsquare. Numerical tests shows that processing reflection waves and diffractionseparately will make reflection multiples better be suppressed. Because ofnonstationary characteristics of diffraction multiples, the nonstationary adaptiveattenuation method can suppress diffraction multiples with different types moreeffectively.