Seismic inversion works as one of pragmatic and effective approaches to estimate the formation information of reservoir. In this study, one robust Bayesian inversion method incorporating the mix-domain convolution with model bounding constraints is proposed. The time-domain response and partial frequency components are all utilized in our proposed work. Firstly, the objective updated function is yielded with Bayesian inference in joint time and frequency domain. Besides, the sparse constraint is incorporated into the final objective function to improve robustness of inversion. Conventional seismic inversion methods always don't incorporate the lower and upper bounding constraints on model parameters, due to which unrealistic predicted results may arise. To get rid of this problem, the transformation function used for lower and upper bounding constraints on P-wave impedance is introduced in our method as it renders to reduce the unrealistic parameters and enhances the reliability of prediction effectively. In addition, the synthetic examples demonstrate the effectiveness and robustness of the proposed method. Finally, one field case is implemented carefully and the estimated parameters with bounding constraints at the borehole-side location preserve a high degree of agreements with real logging data.
Presentation Date: Wednesday, October 19, 2016
Start Time: 10:45:00 AM
Location: Lobby D/C
Presentation Type: POSTER