Robust AVO inversion for elastic modulus and its application in fluid factor calculation

Abstract

Fluid factor based on Biot-Gassmann equation plays an important role in reservoir fluid discrimination. We present a method to calculate fluid factor without density information and only with elastic modulus (P-wave and Swave modulus) inverted by nonlinear AVO inversion based on a novel zoeppritz approximation equation. A practical and robust nonlinear AVO inversion technique is developed in Bayesian framework. The objective is to obtain the maximum a posteriori (MAP) solution for Pwave modulus, S-wave modulus and density. Gaussian and Cauchy distributions are explored for the likelihood and a priori probability distributions. The optimization problem results in weak nonlinear solutions, and the introduction of low frequency constraint and statistic probability information to the objective function enables the inversion more stable and less sensitive to initial model. Tests on synthetic data show that all the parameters can be estimated perfectly when no noise exists, and the estimated P-wave and S-wave modulus still be reasonable when the S/N ratio is as low as 1:2, even with rather smooth initial model parameters. However, the estimated density gives much bias just when the S/N ratio is 2:1 while utilizing the same constraint for all parameters. The inversion method has also been tested on a real data set. The inverted results show good agreement with well logs except the uncertain density information, and the fluid factor calculated from the inverted P-wave and S-wave modulus also show good agreement with drilling result.