Open AccessPrediction of shale oil production based on Prophet algorithm
Author(s) -
Wan Xiaolong,
Yong-ling Zou,
Juan Wang,
Weina Wang
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/2009/1/012056
Subject(s) - oil shale , petroleum engineering , production (economics) , shale gas , oil production , algorithm , shale oil , reservoir simulation , artificial neural network , computer science , environmental science , geology , artificial intelligence , paleontology , economics , macroeconomics
The large-scale volume fracturing development of shale oil horizontal wells, and the production is affected by seasonal cycle and emergencies, resulting in the complexity of production prediction. Aiming at the problem of poor actual effect or heavy workload of shale oil production prediction by classical reservoir engineering, reservoir numerical simulation and other methods, a new shale oil production prediction method based on Prophet algorithm is proposed It is easier to find the inherent law from a small amount of data. In this paper, we use the production data of production wells in Huanjiang A reservoir in 2015 to predict the production, and compare the prediction results with long-term and short-term memory neural network (LSTM) and ARPS production decline model. The results show that the prediction accuracy of Prophet algorithm is higher, and it is more accurate for the production of complex shale oil.