Abstract
Inflow performance relationship (IPR) curves have been extensively studied in the petroleum engineering community; from the classic Vogel's method for vertical wells to modified models for horizontal wells. Previous works have indicated that the performance of horizontal wells can be different from that in vertical wells due to their more complex geometries. Hence, this encourages us to aim for modified IPR correlations.
Modeling the production performance in a horizontal well requires an understanding of the parameters that may affect the fluid flow geometry and well/reservoir interface. In this work, numerical modeling was employed to study the performance of horizontal wells under different well/reservoir conditions. A new IPR relationship was proposed for horizontal wells producing from solution-gas drive reservoirs along with a modified absolute open-flow potential (AOFP) which was proposed initially by Kabir. This was to introduce new terms that accounted for the effect of bubble-point pressure and recovery factor which were found to be significant. The developed model was tested against field data, and it was also evaluated and compared with the current IPR curves.
In order to generate an IPR curve for a horizontal well at the specified recovery factor, simulation models were constructed and run for each bottomhole pressure. Then, computer codes were used to extract the results from each results file. Next, the coefficicents and parameters of the equations were obtained from non-linear regression and curve fitting. Then, IPR curves from the analytical model were generated at specified well/reservoir conditions, and at a certain recovery factors. Finally, the results from the model were compared to the simulation model and field data.
This work confirmed that the IPR curves in horizontal wells producing from a saturated reservoir may need modified relationships due to more cmplex flow geometry and well/reservoir interactions. The developed three-parameter IPR relationship presented reasonable accuracy as was compared to the currenmt models and field/simulation data. Also, the new AOFP equation showed a reasonable error of 3% compared to the simulation results.
