Perpustakaan Digital ITB

The use of statistical science is very important in research in the oil industry, particularly Design of Experiment (DOE) that can make experiments in the research could be done in more efficient way. There are various DOE methods that have been commonly used to date, some of them are Factorial Method, Response Surface Method (RSM), Mixture Design, and Taguchi design. Of those several methods, the author would like to observe which method is most recommended to a case study. The case study is an uncertainty analysis of the water injection impact of a naturally fractured reservoir. Experiments were carried out by using Eclipse software for the simulation and also Minitab software to generate the case parameter combination. Parameters that is observed on this study is fracture transmissibility, matrix transmissibility, fracture pore volume, matrix pore volume and also water injection rate. Based on previous study, those parameter is suspected to affect naturally fractured reservoir gas production, and it is proven on this study case. The result of this study case is recommending Plackett-Burman method for data with 2 level variance and Taguchi method for data with 3 level variance.