Data-Assimilation Approach Applied to Stress Matching in Buzios Field Geomechanics

Creating a robust geomechanical model of a petroleum reservoir is a complex and challenging effort. This undertaking requires integrating diverse datasets from multiple disciplines, including geophysics, petrophysics, geology, laboratory testing, and numerical simulations. In this regard, the Buzios field presents substantial challenges, with its impressive dimensions encompassing 852 km2, average net pay of 400 m, a water depth of 2000 m, and a significant salt thickness of 2000 m situated just below a 1000 m post-salt sediment interval. Given the scale of operations, any geomechanical issues can result in substantial financial losses, underscoring the critical need for a highly reliable geomechanical model. To achieve this reliability, accurately matching the stress history is paramount. This process can range from fully manual adjustments to fully automated methods. This paper explores the optimization of model parameters through the Ensemble Smoother - Multiple Data Assimilation (ES-MDA) technique, which iteratively updates model parameters while incorporating both data and model uncertainties. By generating multiple model realizations, we assess the sensitivity of results to different parameter combinations, ultimately identifying the configuration that yields the best match. Our findings demonstrate the effectiveness of ES-MDA in enhancing the reliability of geomechanical models, thereby supporting more informed decision-making in the management of the Buzios field.