Thesis (M.Eng.)--Chulalongkorn University, 2007

In this research, the author uses the statistical methods for assessing the various sources of uncertainties in material balance equation application, to compare how well the result of each uncertainty analysis method and to study how they affect the amount of original oil in-place calculation. Following those statistical methods, the author has brought several statistical methods from the relevant papers and researches regarding the uncertainty analysis method. Firstly, the Monte Carlo simulation procedure has been used to generate random variables in the material balance equation. Then the author also applied other uncertainty analysis methods for instance the first-order approximation based on Taylor’s series expansion, response surface method and experimental design. The statistical results of traditional Monte Carlo simulation will be set as the reference point in order to compare to other methods. In addition, the predictive result of each method will be also compared to the base case scenario. From the statistical result of each method, we can conclude that the statistical result of the Box-Behnken experimental design with second-order response surface is reasonably accurate. In addition, the predictive result of the Box-Behnken experimental design with second-order response surface can provide the nearest predictive result to the base case scenario. Moreover the Box-Behnken experimental design with second-order response surface can provide both statistical and predictive result at optimum accuracy.