ANALISIS KETIDAKPASTIAN DALAM PERHITUNGAN VOLUME MINYAK AWAL MENGGUNAKAN METODA GEOSTATISTIK DAN PEMODELAN RESERVOIR TIGA DIMENSI

• Main Author: Muldi , Devi
• Format: Theses
• Language: Indonesia
• Subjects: Geologi, hidrologi & meteorologi
• Online Access: https://digilib.itb.ac.id/gdl/view/7596
• Institution: Institut Teknologi Bandung

Uncertainty is due to our lack of understanding of the reservoirs. It exists in raw data measurement, data interpretation, conceptual geological model, structural model, petrophysical model, etc. It is prevalent in every condition and at every phase of oil and gas exploration, appraisal, and production development. In the life of an oil and gas field, there is only limited amount of hard data that can be used to characterize complex subsurface. Even if the data are perfect, they are most often indirect measurements of the variables of interest. For those reasons mentioned above, there are always uncertainties with the subsurface data. Therefore, an uncertainty model is required to relate these data to what is being estimated (i.e OOIP). This study demonstrated a 3D geostatistical reservoir modeling and integrated static uncertainty analysis in assessing volumetric original oil in place (OOIP) in Area 7 Duri Field, Indonesia. The distribution of reservoir properties was performed in two steps: simulate first the rocktypes as the framework, and then simulate the reservoir petrophysical properties conditional on the rocktypes. The 3D Jacta uncertainty reservoir model incorporates the following identified uncertain elements: volume of shale model (Vsh), porosity model, permeability model, net-to-gross (NTG) model, water saturation (Sw) model, and fluid contacts. Model requirement is to construct a sgrid with 25m x 25m x 2ft average cell dimension for the Pertama reservoir. Top, base and proportional intermediate layering in the sgrid was based on lithostratigraphic well markers. Vsh and Phie variograms were modeled separately for each rocktypes and sand/shale layer. Perm and Sw variograms are assumed to have same variogram as those of Vsh. Vsh and Phie property were simulated using Sequential Gaussian Simulation using well data as conditioning data. To account for cross-correlation between petrophysical properties such as porosity and permeability, sequential modeling approach was used to populate permeability and water saturation. The output of Vsh and porosity modeling serves as an input for subsequent perm and Sw population steps. Permeability was populated using Sequential Gaussian Simulation with collocated cokriging (SGSCC) algorithm with previously generated 3D Phie model as soft data. Sw was distributed using SGSCC algorithm with previously generated 3D permeability model as soft data. The correlation coefficients required by SGSCC were calculated from bivariate cross plots of hard and soft data per rocktypes and shale/shale regions. Gocad/Jacta software enables a nested simulation, in which one uncertain variable is simulated dependent to other uncertain variables. For each Vsh simulation, there are four Phie realizations, for each Phie realization, there are four permeability realizations, and so forth. The realizations related to resulted low case, best case and high case OOIP are retrieved. Variation in their respective reservoir porosity and water saturation model can be explored. The result of uncertainty analysis enables the team to identify of key subsurface uncertainties and to capture the spread of possibilities in OOIP numbers.