Agradecimentos: This work was carried out in association with the ongoing project registered as BG-07, “Reduction of uncertainties through the incorporation of 4D seismic data in the modeling of the reservoir” (UNICAMP/BG Brazil/ ANP) funded by BG E&P Brasil Ltda. (Shell Brasil Petróleo Ltda....

Agradecimentos: This work was carried out in association with the ongoing project registered as BG-07, “Reduction of uncertainties through the incorporation of 4D seismic data in the modeling of the reservoir” (UNICAMP/BG Brazil/ ANP) funded by BG E&P Brasil Ltda. (Shell Brasil Petróleo Ltda. Subsidiary) under the ANP R&D levy as “Investment commitment to research and development.” The authors thank the ETLP team from Heriot-Watt University, especially M.-D. Mangriotis, V. Omofoma, A. Briceño, and Y. Zhen for the helpful discussions. The authors also acknowledge UNISIM, DE-FEM-UNICAMP, and CEPETRO for supporting this work and CGG and Schlumberger for the software licenses. We also thank Statoil (operator of the Norne field), ENI, and Petoro for the release of the Norne data and the Centre for Integrated Operations at NTNU for cooperation and coordination of the Norne Case. The views expressed in this paper are the views of the authors and do not necessarily reflect the views of Statoil and the Norne license partners

Time-lapse (or 4D) seismic attributes are extensively used as inputs to history matching workflows. However, this integration can potentially bring problems if performed incorrectly. Some of the uncertainties regarding seismic acquisition, processing, and interpretation can be inadvertently...

Time-lapse (or 4D) seismic attributes are extensively used as inputs to history matching workflows. However, this integration can potentially bring problems if performed incorrectly. Some of the uncertainties regarding seismic acquisition, processing, and interpretation can be inadvertently incorporated into the reservoir simulation model yielding an erroneous production forecast. Very often, the information provided by 4D seismic can be noisy or ambiguous. For this reason, it is necessary to estimate the level of confidence on the data prior to its transfer to the simulation model process. The methodology presented in this paper aims to diagnose which information from 4D seismic that we are confident enough to include in the model. Two passes of seismic interpretation are proposed: the first, intended to understand the character and quality of the seismic data and, the second, to compare the simulation-to-seismic synthetic response with the observed seismic signal. The methodology is applied to the Norne field benchmark case in which we find several examples of inconsistencies between the synthetic and real responses and we evaluate whether these are caused by a simulation model inaccuracy or by uncertainties in the actual observed seismic. After a careful qualitative and semiquantitative analysis, the confidence level of the interpretation is determined. Simulation model updates can be suggested according to the outcome from this analysis. The main contribution of this work is to introduce a diagnostic step that classifies the seismic interpretation reliability considering the uncertainties inherent in these data. The results indicate that a medium to high interpretation confidence can be achieved even for poorly repeated data

Fechado