UNCERTAINTY ANALYSIS IN RESERVOIR MODELING USING DESIGN OF EXPERIMENT

• Main Author: SAYYIDI (NIM : 22214010), MUHAMMAD
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/23336
• Institution: Institut Teknologi Bandung

In Indonesia. oil and gas industry has been operated more than 100 years with decreasing production by time due to the absent of big hydrocarbon reserves invention. Through the development of technology, some unproven oil fields could be improved to be proven. Subsurface modeling includes static and dynamic modeling is a common example of these technologies. <br /> <br /> <br /> <br /> Complexity, variation of reservoir parameters and data limitation will be a challenge in subsurface modeling since it gives an unclear description of real reservoir ondition. Thus, accuracy level of created model dan estimated reserved are depends on the used data. <br /> <br /> <br /> <br /> The objective of this study is to build alternative models from limited data using uncertainty analysis. The alternative models should be can representative with future reservoir condition by adding number of wells or production data which validated by history matching. <br /> <br /> <br /> <br /> The methods work through the following steps. First, build the basic model through integrated geophysical, geological and petrophysical analysis. Afterwards, uncertainty analysis is applied using several parameters. Experimental design produces some alternative models derived from the basic. These models will be calibrated or validated with history matching. <br /> <br /> <br /> <br /> Uncertainty parameters that used in this study are porosity, well top (marker), water saturation and net to gross. These parameters affect the dsitribution of rock properties and the reserves. The range value of these parameters are determined based on statistical data analysis, analogy, and subjectivity appraisement. Design level 3 Box-Behenken used as the experimental design methods to develop alternative methods with various parameter combinations. <br /> <br /> <br /> <br /> Sensitivity analysis on uncertainty parameters delivers the most influence parameter to reserves estimation result. Alternative model derived from limited data could predict the probability of reservoir condition in the future with additional data. As a proof, 40 of 113 alternative models have matched with the model which done with data addition. <br /> <br /> <br /> <br /> History matching in experimental design priority model was done using parameters that be applied for the basic model. Compatibility level of the alternative model is affected by the reserves and uncertainty parameters combination. The result show that 90% of alternative models have tolerance range about 5%.