A study on the method of joint inversion of multiscale seismic data
多尺度地震资料联合反演方法研究

The conventional 3D surface seismic impedance inversion has the shortcomings of multiple solutions and low resolution. It is possible to overcome these shortcomings with plenty of multiscale seismic data in reservoir geophysics. Based on the Bayesian theorem, a new likelihood function is built according to the joint probability distribution. The surface seismic data, the VSP data, and the crosswell seismic data are then integrated into the joint inversion framework for multiscale seismic data, and the work flow of the joint inversion is completed and improved. The model testing results show that the high frequency information of the small-scale seismic data is introduced into the joint inversion system, which effectively improves the resolution of the inversion. In the real data processing, the global feature of the large-scale seismic data is preserved in the inversion result with apparently higher resolution. The multiplicity of solutions in conventional inversion is reduced and the characteristics consistency among multiscale seismic data is improved. These demonstrate the feasibility of the joint inversion method in practical applications.