Quantification and analysis of uncertainties in reservoir modeling using multiple-point geostatistics

---

Tezin Türü: Doktora

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Petrol ve Doğal Gaz Mühendisliği Bölümü, Türkiye

Tezin Onay Tarihi: 2012

Öğrenci: MOHAMED MOHIELDIN FADLELMULA FADLELSEED

Danışman: SERHAT AKIN

Özet:

This study analyzed and quantified uncertainties of reservoirs modeled using multiple-point geostatistics (MPG). The uncertainty types analyzed herein are training image (TI) and hard data (porosity) uncertainties. Aiming at studying the impact of TI uncertainty, this study provides a tool to parameterize TIs having channel structure by a mathematical (Sine) function so that a TI is a function of four parameters. These parameters are channels’ number, waves’ number in each channel, amplitude level of waves, and Z-direction slices’ number. These parameters are used to generate 2D and 3D TIs to remodel a reservoir utilizing a proposed MPG methodology. Analysis of cumulative oil production values showed that TI having 5 Z-direction slices and 3 channels with 2 medium or low amplitude level waves or more produced representative and reliable reservoir models. Thus, these values are set as thresholds of the TI’s parameters. Additionally, increasing the number of channels and waves of a TI decreased the uncertainty of the simulated reservoir. However, increasing the number of Z slices beyond 5 and the amplitude level had no effect on the uncertainty. Analysis of the original and recoverable oil in place (OOIP and ROIP) values showed that the effect of channel number and amplitude level are random. However, the number of waves is directly proportional to OOIP and ROIP values. Moreover, utilization of the thresholds defined decreased the uncertainty range of OOIP and ROIP prediction. Finally, the investigation of hard data uncertainty revealed that porosity data uncertainty has great impact on the simulated reservoir.