Reservoir simulation history matching typically involves a process whereby reservoir rock, fluid properties and well-bore conditions are modified in some systematic way to match observed production rates and pressures. The degree to which initial reservoir and well parameters need to be modified can often be attributed to the amount and quality of hard data available and the integration of that data into the geocellular model. All too often, however, reservoir engineers find that they must modify reservoir parameters significantly in order to obtain a history match. Historically, porosity cut-offs have often been applied to fine-grid geocellular models during upscaling to obtain the reservoir simulation-scale model. Porosity cut-offs have typically been selected by inspection of the core-derived porosity- permeability (P&P) relationship(s) and selecting the porosity that falls below an a priori specified permeability cut-off. During subsequent history matching, a disconnection between upscaled geocellular model and history-matched permeabilities becomes apparent. Permeability modifiers of two or more often need be applied across the reservoir to match reservoir performance. Discrepancies between well-log-derived porosities and core-measured porosities due to the scale (averaging) effects of the logs have been long recognized and can be accounted for in various ways (Klein et al., 2006). However, the discrepancy in permeabilities between the initial upscaled geocellular and history-matched reservoir simulation models appears to be different. In 2007, Delfiner provided a very simple, yet elegant, explanation for this discrepancy and one potential solution to the problem. Simple examples are utilized to show how errors in the porosity-permeability transform can propagate throughout the model building and history matching process, resulting in a potentially significant underestimation of OOIP and recovery. In particular, the potential impacts of the porosity-permeability transform on net pay definition, liquid permeability, initial water saturation, relative permeability, fractional flow, production rate and ultimate recovery are discussed along with how the implementation of the suggested solution by Delfiner (2007), together with other approaches, can help mitigate these errors.
展开▼
