A Fast Stochastic Inversion to Leverage the Full Power of the Stochastic Process

If geostatistical inversion benefits are widely recognized for geo-energy studies, their current implementations make it difficult to be incorporated into routines. The Fast Stochastic Inversion (FSI) presented in this paper is a new methodology leveraging different methods to speed up the stochastic process while preserving the quality and the high-resolution content of the result. It is based on the spectral merging method which enables the integration in the frequency domain of results from deterministic inversion with stochastic simulations from well data avoiding the need to run a multitude of iterations of simulations for each location. The simulations in FSI are run using the turning band methodology, which is faster than sequential Gaussian simulations (SGS), outputting unconditional simulations which are then conditioned by kriging. The result quality is similar to SGS, however the computation times are significantly reduced as presented in an example. As a consequence, FSI can be used more easily than standard geostatistical inversions to enable geoscientists to perform comprehensive uncertainty assessment.