Virtual multiphase flow metering using diverse neural network ensemble and adaptive simulated annealing
作者：
Highlights：
• Development of data-driven virtual flow meter (VFM) using diverse neural network ensembles.
• Adaptive simulated annealing is used for pruning and combining strategy selection.
• VFM can provide real-time monitoring for fields with common metering infrastructure.
• Achieved 4.7% and 2.5% mean absolute errors for gas and liquid flow rate estimations.
• The proposed method outperforms standard stacking and bagging techniques.
摘要
•Development of data-driven virtual flow meter (VFM) using diverse neural network ensembles.•Adaptive simulated annealing is used for pruning and combining strategy selection.•VFM can provide real-time monitoring for fields with common metering infrastructure.•Achieved 4.7% and 2.5% mean absolute errors for gas and liquid flow rate estimations.•The proposed method outperforms standard stacking and bagging techniques.
论文关键词：Neural network,Ensemble method,Simulated annealing,Multiphase flow,Virtual flow meter,Soft sensor
论文评审过程：Received 7 March 2017, Revised 4 October 2017, Accepted 5 October 2017, Available online 6 October 2017, Version of Record 13 October 2017.
论文官网地址：https://doi.org/10.1016/j.eswa.2017.10.014
