Open AccessIntelligent coalbed methane production management and control technology based on reinforcement learning algorithm
Author(s) -
Hongli Wang,
Suian Zhang,
Bin Liu
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 - 10.1088/1742-6596/1894/1/012029
Subject(s) - coalbed methane , reinforcement learning , petroleum engineering , coal , production (economics) , engineering , computer science , algorithm , coal mining , environmental science , artificial intelligence , waste management , economics , macroeconomics
The production control of coalbed methane wells has long been viewed as the most challenging step in its development process. For human engineers, they rely too much on previous experience. For artificial intelligence, there is no complete frame to use. Here we proposed a system with reinforcement learning algorithm to CBM production control optimization that used a proxy model to simulate the gas and water seepage in coal seam, and a ‘value networks’ to evaluate gas and water production capability and three control policy mode: bottomhole pressure (BHP) regression model, BHP reduction rate mode, BHP table to select moves. The system achieved a 20.99% and 38.14% increment in cumulative gas and water production, respectively.