Stepwise uncertainty reduction in time-lapse seismic interpretation using multi-attribute analysis

Abstract

Recently, time-lapse seismic (4D seismic) has been steadily used to demonstrate the relation between field depletion and 4D seismic response, and, subsequently, to provide more efficient field management. A key component of reservoir monitoring is the knowledge of fluid movement and pressure variation. This information is vital in assisting infill drilling and as a reliable source of data to update reservoir models, and, consequently, in helping to improve model-based reservoir management and decision-making processes. However, in practice, varying levels of uncertainty are inherent in the 4D seismic interpretation of reservoirs that uses a multipart production regime. The complex nature of some 4D seismic signals emphasizes the role of the competing effects of geology, rock and fluid interactions. Hence, a reliable 4D interpretation requires an interdisciplinary approach that entails data analysis and insights from geophysics, engineering and geology. In this study, a stepwise workflow was introduced to reduce the uncertainties in the 4D seismic interpretation and to identify the improvements required in order to perform better reservoir surveillance. In parallel, the workflow demonstrates the use of engineering data analysis in conducting a consistent interpretation, and encompasses the 3D and 4D seismic attributes with engineering data analysis. This study was carried out in a Brazilian heavy-oil offshore field where production started in 2013. The field experienced intense production activity up to 2016, making the deep-water field an ideal candidate to explore the challenges in interpreting complex 4D signals. Beyond these challenges, a significant understanding of reservoir behaviour is obtained and improvements to the reservoir simulation model are suggested that could assist reservoir engineers with data assimilation applications.