A Tiered Workflow for Probabilistic Estimation of Pressure and Saturation Changes from 4D Seismic Data

Abstract

Summary This presentation will showcase a practical workflow to estimate reservoir pressure and saturation changes from 4D seismic monitoring data. The workflow gradually evolves from a qualitative fast-track estimation to more robust quantitative estimations that utilize machine learning models, Bayesian statistics and stochastic sampling. The final solution integrates data from repeated well logs, reservoir simulations and 4D seismic data to provide a most likely estimation to the changes in dynamic reservoir properties and their associated uncertainties. This probabilistic solution can then provide multiple realizations of pressure and saturation distributions that match the 4D seismic data equally well. It can also provide conditional probability ranges that can be used to aid in decision making for reservoir management. The workflow is showcased with two real case applications from sandstone reservoirs in the North Sea.