Predicting bottomhole pressure in vertical multiphase flowing wells using artificial neural networks

Over the years, accurate prediction of pressure drop has been of vital importance in vertical multiphase flowing oil wells in order to design an effective production string and optimum production strategy selection. Various scientists and researchers have proposed correlations and mechanistic models for this purpose since 1950, most of which are widely used in the industry. But in spite of recent improvements in pressure prediction techniques, most of these models fail to provide the desired accuracy of pressure drop, and further improvement is still needed.This study presents an artificial neural network (ANN) model for prediction of the bottomhole flowing pressure and consequently the pressure drop in vertical multiphase flowing wells. The model was developed and tested using field data covering a wide range of variables. A total of 413 field data sets collected from Iran fields were used to develop the ANN model. These data sets were divided into training, validation and testing sets in the ratio of 4:1:1. The results showed that the research model significantly outperforms all existing methods and provides predictions with higher accuracy, approximately 3.5% absolute average percent error and 0.9222 correlation coefficient.

Research Highlights
► ANN is utilized to predict bottom hole pressure in vertical multiphase flowing wells.
► Best network structure is developed by comparing different networks’ error analysis.
► Proposed model outperforms available empirical correlations and mechanistic models.
► Trend analysis confirms that ANN model correctly simulates actual physical process.