Probabilistic Reserve Estimation Constrained by Limited Production Data: An Integrated Approach

Abstract

Large uncertainties in structure and facies had been recognized in a major gas field in Pakistan after early production. The conventional reserve estimation methods had failed in providing a reliable estimate of gas-in-place (GIIP). It was possible to get a good history match of one-year production data using a wide range of GIIP through a slight and acceptable adjustment of porosity and permeability. The resulting possible range of GIIP could easily vary by a factor of 1.5. Structural uncertainties did not warrant volumetric estimates either. Material balance technique was questionable due to non-uniform drainage of the reservoir. Clearly these deterministic techniques of reserve estimation were not applicable at this stage of production considering the complexities of the reservoir. A probabilistic technique was therefore developed that addressed both static and dynamic uncertainties in an integrated approach while honoring the available production history. Combined treatment of static and dynamic uncertainties also ensured a better coverage of the entire sample space, thus making the probabilistic approach more reliable. Latin Hypercube Sampling (LHS) helped minimizing the number of simulation runs while providing a reasonable coverage of the sample space. Yet we ended up with almost 1500 simulation runs. The process of history matching, ranking and keeping track of all these simulation runs demanded an innovative workflow. A number of software tools were used to automate and optimize this process. Out of 1500 simulation runs, the 200 best runs having minimum objective function through history matching were selected. These runs were later used for production forecasting, for providing a range of reserves, and for sensitivity analysis to identify the most influential variables. Structure and NTG were identified as the two most critical variables for *GIIP while residual gas saturation was identified as an additional sensitive variable for reserves. Different geostatistical realizations had little impact on GIIP or reserves.