Multi-Wave Stochastic Inversion Physical Parameters Prediction Method Driven by Linearized Rock Physics

Pre-stack seismic inversion is an effective means of using seismic data to achieve prediction of physical parameters of underground media. However, pure PP wave inversion suffers from high inversion multiplicity solutions and limited prediction accuracy. Therefore, we propose a multi-wave stochastic inversion physical parameters prediction method driven by linearized rock physics. First, this method establishes a robust relationship between elastic and physical parameters based on a statistical rock physics model. Second, the linearized AVO approximations for PP wave and SH-SH wave expressed by porosity, clay volume, and water saturation are derived. Numerical simulations of the reflection characteristics of the two model interfaces show that both new formulations have high accuracy. On this basis, we construct a joint inversion equation of PP wave and SH-SH wave for porosity, clay volume, and water saturation. A stochastic inversion method based on the ensemble smoother with multiple data assimilation (ES-MDA) for the multi-wave joint physical parameters is proposed. Stanford VI-E model tests indicate that the physical parameters obtained from the joint PP wave and SH-SH wave inversion have higher identification accuracy and smaller relative errors compared to the pure PP wave inversion. Moreover, field data tests demonstrate the good practicality of the method for seismic prediction of underground media physical parameters.