Geomechanical Model from Subseismic Resolution Data

Abstract

Abstract This article demonstrates the extension of subseismic-driven earth elastic model into geomechanic domain. The geomechanical model is becoming uncertain as the impact of subseismic uncertainty. The detail and the uncertainty of the geomechanical model need to be quantitatively defined. An innovative stochastic partial stacked seismic inversion is proposed to be combined with the geomechanical model built from well to address the problem mentioned. The stochastic inversion uses the energy attribute to control the misfit between the synthetic and true seismic through a Bayesian inference. The prior model of inversion is generated through a geometrical and geophysical perturbation of reflectivity model derived from apparent seismic amplitude. The perturbation is constrained by assigned wavelet, tuning and spectral analysis, and locally varying anisotropy to ensure the adequacy of both vertical and lateral resolution. A geomechanical low frequency model is built by calibrating the low frequency model predicted from the relative geomechanical properties (which is the extension of the relative elastic impedance) inverted from the primary inversion with the geomechanical low frequency model derived from the well log data. The absolute value of the geomechanical property is then estimated based on the frequency content matching between the calibrated relative geomechanical properties and the well log geomechanical model. The related uncertainty of the inverted geomechanical properties is assessed by employing the Markov Chain Monte Carlo (MCMC) sampling methodology. The proposed method was test to predict the geomechanical model of a fractured volcanic rock basement of Malay Basin. The proposed combination approach between well log geomechanic and stochastic inversion based elastic impedance shows the critical challenge in modeling the geomechanic from subseismic resolution seismic data and promises a good solution. Compared with conventional deterministic inversion, the proposed stochastic inversion provides much better resolution both in lateral and vertical direction, and geologically more consistent. The detail inverted geomechanical properties, inherited by the detail inverted elastic properties, can be used to study the fracture of the potential fractured basement hydrocarbon reservoir.