Uncertainty Analysis using Design of Experiments to Assess the Development Plan of a North Sea Field

Abstract

This study assessed the impact of static and dynamic variables in EUR and NPV in the development plan of a North Sea offshore field with 81 m of water, light oil crude of 40.5 API and 510 SCF/STB of GOR comprised of sandstones from a shallow marine environment in anticline structure separated to the northeast and southwest by a pair of normal faults. The analysis is conducted through the application of different experimental design techniques and the preparation of a comparison between them. Uncertainty analysis has been prepared to characterize the appropriate range for nine variables that affect the oil recovery and net present value of the field development. A Folded Plackett-Burman design was prepared to screen the initial nine variables; the linear regression results show that the oil water contact, permeability anisotropy and net to gross are the significant variables. Also, the residual analysis demonstrated that the proxy equation should be improved to have better predictability in the non-sampled space. In consequence, a D-Optimal and a Central Composite experimental design were prepared for the three significant variables. The regression results show better coefficient correlation and lower least square errors in the D-Optimal design using a full quadratic model and confirmed the oil water contact as the most significant variable of the field. Finally, Monte Carlo Simulation was performed in the proxy model from the D-Optimal design, which resulted in an expected value ultimate recovery of 357 MMSTB. The paper presents an exciting workflow to analyze different experimental design techniques, compare them and use the most suitable to prepare the development plan of a field.