Productivity Optimization of Oil Wells Using a New 3D Finite-Element Wellborre Inflow Model and Artificial Neural Network
The most common method of completing a well involves cementing a steel casing in the well bore and then using shaped charges to perforate the casing and penetrate the producing zone. The productivity of these wells is influenced by several factors, which include the length of the individual perforations, casing entrance hole diameter, perforation shot density, phase angle between the perforations, and the degree of damage inside and around the perforations. This paper presents application of a new wellbore inflow model to quantify the effects of these individual parameters on productivity. This model incorporates the cone-shaped perforation geometry with a tapered tip that has been observed in the laboratory for years. For the first time, the asymmetric, spiral distribution of perforations around a wellbore is modeled using a full 3-dimensional finite element model with over 30,000 elements in each perforation layer instead of the 2-dimensional and quasi 3-dimensional models used in the past. In addition, the wellbore inflow model is used to study the effect of reservoir anisotropy and dip angle of the bedding planes on perforation design over a wide range of shot densities and phasing. Productivity results from the new 3D model are compared with previous models to demonstrate the improvements to the inflow predictions. The effect of reservoir anisotropy on perforation design is also studied over a wide range of shot densities and phasing. Results that highlight the effect of reservoir anisotropy are presented in this work. The wellbore inflow model results are also used to develop a neural network algorithm that closely matches the finite element simulation results. This neural network provides an efficient method of evaluating the primary factors that can influence a perforation design and make it possible to optimize the flow performance of cased and perforated wells. Based on the wellbore inflow model results and the neural network algorithm, the PerfProTM software product was developed which is used to optimize perforation design of cased and perforated wells. Finally, we demonstrate practical application of PerfPro using a field example.