Multiscale Voltage Reconstruction With Attention-Based Network for Volume Fraction Prediction of Industrial Oil–Water Two-Phase Flow by EIT
Hao Yu,Zhixi Zhang,Yang Gao,Jiabin Jia
DOI: https://doi.org/10.1109/tim.2022.3169557
IF: 5.6
2022-05-07
IEEE Transactions on Instrumentation and Measurement
Abstract:Oil–water two-phase flow as a typical two-phase flow type widely exists in various industrial processes and the accurate measurement of oil volume fraction (VF) plays a significant role in transporting and separating oil–water mixture in the processes. Electrical impedance tomography (EIT) as a merging technology with the advantages of noninvasive, low cost, and real-time measurement is widely applied in the industrial field to measure the VF for different types of two-phase flows. However, the measurement process of taking homogeneous reference voltages is time-consuming and costly. To cope with the problem, in the article, by establishing an end-to-end mapping between measurement voltages and VF, we propose an attention UNet-fully connected (AU-FC) architecture. Relying on the attention mechanism, the reconstructed voltages having a strong correlation or a weak correlation with VF are highlighted or suppressed, respectively. Oil–water two-phase flow experiment was conducted in the NEL facility to collect EIT voltage data. Compared with six state-of-the-art and existing machine learning methods, the proposed method performs better in predicting VF. The results indicate that the proposed AU-FC architecture can accurately and real-time predict the VF of oil–water two-phase flow, which improves the application potential of EIT combined with deep learning method in the industrial field.
engineering, electrical & electronic,instruments & instrumentation