Header menu link for other important links
Enabling Oil Production Forecasting using Machine Learning
D. Patwardhan Bikash Kumar Parhi Samarth,
Published in Springer Science and Business Media Deutschland GmbH
Volume: 1367
Pages: 452 - 464
Machine learning is defined as an application of artificial intelligence where available information is used through algorithms to process or assist the processing of any set of data. There is growing evidence of machine learning being used in many oil and gas industry operations for better predictions, right from facies identification to pipeline maintenance, thereby enhancing safety significantly. The rate of a production well decreases with time and this relationship is known as the decline curve. Decline curve analysis is a graphical method used for analyzing oil and gas production rates and forecasting future performance, thereby assisting in overall field performance management. Arp’s decline curve analysis is usually the most common method which includes a comprehensive set of equations defining the exponential, harmonic and hyperbolic declines which can only be applied in case of a stabilized production trend. This paper focuses on predicting the decline curve by using appropriately designed neural networks and machine learning algorithms irrespective of production trend. The feature vectors are geological location of the well, production data, pressure data, and operating constraints. Fine tuning these parameters lead us to predict the performance of a well with reasonable certainty. © 2021, Springer Nature Singapore Pte Ltd.
About the journal
JournalData powered by Typeset10th International Advanced Computing Conference (Springer)
PublisherData powered by TypesetSpringer Science and Business Media Deutschland GmbH
Open AccessNo