Research And Application Of High-precision Deblending Method For Simultaneous Source Seismic Data

Seismic acquisition is a trade-off between economic and quality.Traditional seismic data acquisition,whether 2D or 3D seismic acquisition,requires large enough source interval to avoid mutual interference between two adjacent sources,which requires great time and cost.The simultaneous source technology explodes multiple sources of different positions by a certain way to acquire a mutual interfered record that breaks the limit of the conventional seismic acquisition theory that there must be a large temporal or spatial interval to avoid overlap from adjacent sources.The simultaneous source technology can achieve several times the acquisition density of conventional source acquistion in the same time,so that high-quality seismic data can be obtained,or at the same density of acquisition,the acquisition time will be greatly reduced and the acquisition cost will be reduced.There are two imaging methods about the blended data at this stage: perform the migration processing directly on the blended data without any pre-separation,or recovery the blended data to individual shot data,which is called "deblending",then apply standard migration processing to these deblended data.This paper focuses on the deblending of the simultaneous source seismic data.Because the deblending problem is a typical under set problem in mathematics,we can not get the result by solving the inverse of the matrix directly.We can obtain the result by the least square method are called pseudo-deblending.The pseudo-deblending data contains a lot of blending noise,and it can not weaken the interference from the adjacent excitation source signal in the seismic record.The blending noise is different in different seismic gather,the common shot point gather exists in coherent form,and it shows random pulse form in non-common shot gather.The non-coherent intensity of the blending noise directly affects the separation accuracy of the results.In this paper,the parameters of the simultaneous source acquistion system are setted by considering the factors of these three aspects: the number of source in simultaneous source,the delay time range of a single source in simultaneous source and the position distribution of single source in simultaneous source.Reducing the random delay range between single sources in the simultaneous source can effectively improve the efficiency of the acquisition,but this also has a negative impact on the deblending of the simultaneous source data.The blending noise in the non-common-source domain data is significantly more concentrated after pseudo-separation,which is difficult to suppress when the random delay time range is small.In this paper,we propose a method to separate the simultaneous source seismic data based on the trilateral filter and compare with the iterative multi-level median filtering method.When the time delay range is large,the two methods can both get a good separation result.When the time delay range is small,the method presented in this paper can be more effective and retain more detail information.The actual data calculation results show that this method can also suppress other random noise to a certain extent.The problem of deblending can be considered as a linear problem for solving Ax =b,where x is the deblending result,b indicates the actual observed seismic records,and A represents the blending process.This problem can be solved by the sparse constraint inversion method.The 3D blending data contains more information which on the one hand introduce more strongly blending noise,on the other hand makes it more difficult to construct the blending operator.Given these two problems,this paper deblends the 3D blending data in Radon domain with the sparse constrained inversion which can get higher precision of separation result,and using the GPS time excited by the source to blending and pseudo-deblending the results of the last iterative separation at the common receiver point gather by a long record can process the blending data one receiver by one receiver iteratively rather than the whole data and this can also avoid the construction of deblending operator.In this paper,we use modeled data and real data to test the feasibility of our method.