Gas Reserves Estimation by Monte Carlo Simulation and Chemical Flooding Optimization Using Experimental Design and Response Surface Methodology

Reserves estimation contains considerable uncertainties especially in a fluvial environment with many lateral discontinuous reservoirs. With this type of reservoir, the difficulty in determining the reservoir properties is relatively high since there are numerous possible values for the outcomes, which make reserves estimation problematic. A Monte Carlo simulation is one method to handle these uncertainties. By using this method associated with the concept of analogy, the reliability of reserves estimation can be increased. In this study, this method was used to estimate reserves in actual reservoirs in the Gulf of Thailand. A Monte Carlo simulation was conducted by sampling values from the probability distribution functions (PDFs) of the reservoir properties, and then the sampled values were used to calculate values of reserves using the volumetric reserves equation. The process was repeated until a PDF of reserves was generated. This method required the fitting PDFs of reservoir properties, which were constructed using appropriate and sufficient data. For the geological area with insufficient data, we used a “t-test for two population means with unknown and unequal variances” to analyze the data from different categories when they could be combined. In this study, two sampling methods: basic Monte Carlo (MC) and Latin Hypercube (LH), were applied and compared for a particular case of reserves estimation. Also, with the t-test, guidelines for data selection were created. Using these guidelines and Monte Carlo simulation, our results showed that the LH method converged to the final results at a smaller number of iterations compared to the MC method. Most importantly, the values of reserves obtained from the study were close to the actual reserves. Therefore, it could be concluded that reserves estimation will be more accurate with the appropriate data set and Monte Carlo simulation. The second part of this report concerns chemical flooding optimization using experimental design and response surface methodology. Chemical flooding has been known as one of the most technically successful enhanced oil recovery method. This method can recover high percentage of residual oil; however, there is a critical issue in terms of its high cost and high uncertainty to be considered. Many process parameters significantly affect the chemical flooding process such as surfactant slug size, surfactant concentration, polymer drive size, and polymer concentration. Therefore, intensive study in the flooding process should be concerned. The concept of experimental design and response surface methodology was employed in this study to determine an optimal chemical flooding design in terms of maximizing NPV. The associated risks can be mitigated effectively with the application of this method.