Header menu link for other important links
Predicting Critical Parameters of a Declining Oil Well - A Data Driven Approach
Published in Institute of Physics
Volume: 107
Issue: 1
Pages: 10701 - 10720
Machine learning is an artificial intelligence application in which accessible data is processed or assisted in the analysis of any given dataset using algorithms. Machine learning is increasingly being employed in numerous oil and gas sector activities for better forecasts, ranging from litho-stratigraphic identification to pipeline network maintenance and service, and hence greatly improving safety. Using properly designed Neural Network algorithms, various parameters of a well under production has been predicted with reasonable accuracy. After a well is drilled and completed with casing, perforations are done to establish communication between the well and reservoir followed by production from the well. There are various production parameters, some of which are critical as they determine the life and health of a well, while others are general parameters which give details of the features and characteristics of the well as well as the reservoir. These parameters are obtained either directly from field measurements in real time with specific equipment or some are calculated with complex empirical formulas. In this paper, we have taken a machine learning methodology by designing neural networks models to predict the decline curve and further determine some of production parameters. © The Electrochemical Society
About the journal
JournalECS Transactions
PublisherInstitute of Physics
Open AccessNo