Horizontal wells application in heterogeneous reservoirs is challenging as assumes high expenditures in conditions of problems with such wells modeling. The model matched to history data can show inaccuracy in horizontal wells performance forecast. In this paper the issue is investigated with using of integrated approach.
Uncertainties related to reservoir complexity are considered as possible sources of erroneous prediction. Reservoir study coupled with the review of analytical solutions is conducted to identify the uncertain parameters influencing horizontal well performance. Their ranges of variation obtained from research data are included into qualitative analysis of uncertainties impact. It is performed on the model with estimation of oil production as a main decision criterion. All the uncertainties are incorporated into simulation process with using of Monte Carlo technique and automatic history matching. Non-unique nature of the model is shown and the significance of performing multivariant prediction for horizontal wells is proved.
As a result the modeling procedure with uncertainties inclusion is proposed and methods of uncertainties reduction are recommended in order to increase model reliability. This will allow uncertainties management and more founded decision making.