Passive learning to address nonstationarity in virtual flow metering applications

作者：

Highlights：

• Passive learning is applied to steady-state VFMs.

• Six VFM types are investigated for 10 petroleum wells.

• Frequent calibration is key to improve the long-term performance.

• VFMs with physical considerations are advantageous with little data.

摘要

•Passive learning is applied to steady-state VFMs.•Six VFM types are investigated for 10 petroleum wells.•Frequent calibration is key to improve the long-term performance.•VFMs with physical considerations are advantageous with little data.

论文关键词：Virtual flow metering,Nonstationarity,Passive learning,Online learning,Periodic batch learning,Neural networks

论文评审过程：Received 8 February 2022, Revised 5 July 2022, Accepted 1 August 2022, Available online 12 August 2022, Version of Record 17 August 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.118382