Multiphase Flowrate Measurement With Multimodal Sensors and Temporal Convolutional Network

Accurate multiphase flow measurement is vital in monitoring and optimizing various production processes. Deep learning has as of late arose a promising approach for assessing flowrate dependent on customary meters. In this paper, we propose multi-modal sensor Temporal Convolution Network (TCN) based method to predict the volumetric oil/gas two-phase flows. The flowrates liquid gas phase vary from 0.96 - 6.13 $\text{m}^{{3}}$ /h 5.5 121.2 /h, respectively. sequential sensing data are simultaneously collected Venturi tube dual-plane Electrical Capacitance Tomography (ECT) pilot-scale facility. reference derived single-phase flowmeters. Z-score First-Difference (FD) pre-processing methods employed manipulate instantaneous time series data. pre-processed utilized training TCN model. Experimental results reveal that model can effectively provide guidance estimation demonstrate effectiveness combining sensors prediction under complex conditions.