Uncertainty Assessment of Burgan Marrat Carbonate Reservoir using Response Surface Modelling to Plan Early-Stage Field Development

Burgan Marrat, a deep carbonate reservoir was transferred from exploration to development team for an accelerated production of the newly discovered oil. This multi-billion barrel reservoir is spread over 450 km2, has more than 40 faults, 8 compartments with large variation in oil-water contact and reservoir/fluid characteristics. The objective of this work is to understand the key uncertainties and quantify their impact on the reservoir offtake rate and oil recovery by conducting uncertainty assessment. An interdisciplinary team identified the key uncertainty parameters expected to have significant impact on the reservoir development. The range and probability distribution law for each parameter was set considering the uncertainties due to limited measurements or variation in interpretations. A Response Surface Model (RSM) was created to evaluate the uncertainties by using a base dynamic model and applying an appropriate experimental design, which allowed to efficiently study the uncertainty space with a feasible number of simulations. Using the RSM, the primary effects and interaction between parameters were quantified to rank the uncertainties based on their impact on field production. Key uncertainty parameters were identified including eight OWCs, six fault transmissibilities, horizontal and vertical permeability multipliers, and porosity multiplier. Latin Hypercube was found to be the appropriate Experimental Design for the study considering 17 parameters and the need of building a reliable RSM that includes interactions between them. The design recommended 155 simulation cases, which were prepared and submitted automatically by the software. Multi-time Responses were analyzed qualitatively to identify the top 5 uncertainties having material impact on field production over 20 years considering 6 existing wells and 30 new well locations. The RSM quantitative evaluation showed three parameters (OWC2, OWC4 and OWC1) having a total effect on the response higher than 10%; followed by PERMX and OWC3 with less than 5%. The other 12 parameters have total effects less than 2%, and the interactions effect is less than 0.5% for any interaction between two parameters. Contrary to the intuition, none of the faults proved impact on the reservoir production. The results prove very useful to make a right development and appraisal strategy in early life of the reservoir. The new well locations can be ranked and prioritized to optimize the development and effectively appraise the areas with high risks. Uncertainty assessment has value throughout the life of the reservoir. However, this study indicates that its application in early life of the reservoir can bring immense value. An uncertainty analysis on the reservoir production helps in decision-making regarding the number of wells and their locations to reach a target production by managing the risks.