Prediction of Two-Phase Pressure Drop Using Artificial Neural Network

doi:10.21608/erjm.2019.66275

Cited by 1 publication

(1 citation statement)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The expert system has been validated using a new set of data (144 points), and a Visual Basic Application (VBA) for excel has been developed to automate the selection process of the best correlation for the different flowing conditions and well configuration. El-Kadi et al (2019) developed an ANN model to predict the pressure drop in different oil and gas fields. They used 7581 points and divided the well into segments to increase the accuracy of the pressure drop prediction.…”

Section: Introductionmentioning

confidence: 99%

Expert solution for effects of input parameters on multiphase flow correlations

El-Moniem¹,

El-Banbi

2021

JER is an international, peer-reviewed journal that publishes f

View full text Add to dashboard Cite

Oil and gas production represents an essential source of energy. Optimization of oil and gas production systems requires accurate calculation of pressure drop in tubing and flowlines. Many empirical correlations and mechanistic models exist to calculate pressure drop in tubing and flowlines. Previous work has shown that some correlations provide more accurate results under certain flow conditions, PVT data, and well configurations than others. However, the effects of errors in input data on the selection of which correlations to use have not been investigated. This paper studies different multiphase flow correlations to determine the effects of their input parameters on (1) the accuracy of calculated pressure drop and (2) the selection of best correlation. A database consisting of 33 oil wells and 32 gas wells was selected, and a commercial software was used to build different well models. A total of 715 well models were constructed and used to investigate the effects of errors in correlations inputs on both the calculated bottomhole pressure and the selection of best correlation(s). The methodology was based on perturbing the values of the selected input parameters and calculating the new predicted bottomhole flowing pressure. Then, the effects of error in input parameters on how the calculated bottomhole pressure was different from observed data were quantified. The effect of this error in input parameters was also checked against the algorithm that selects the best correlation(s). It was found that errors in input GOR have the greatest effects for oil wells, while gas specific gravity and the tubing roughness are the most effective parameters for gas wells. The results were integrated into a rule-based expert system. A new set of data, consisting of 220 cases from 10 new oil wells and 10 new gas wells, was used to validate the expert system. The expert system was found to predict the best correlation(s) with a success rate of 80%, and it also identifies the input parameters whose error would affect the value of calculated bottomhole pressure significantly. Finally, the rules of the expert system were programmed into a VBA-Code to ease its use.

show abstract

Section: Introductionmentioning

confidence: 99%