Research ArticleGas-water two-phase flow characterization with Electrical Resistance Tomography and Multivariate Multiscale Entropy analysis

- Gas-water two-phase flow is measured with an Electrical Resistance Tomography.
- Multivariate Multiscale Sample Entropy is extracted from multiple information by ERT.
- MSampEn is sensitive to flow velocity change and flow pattern transition on different scales.
- Flow pattern transition is analyzed on an “average MSampEn versus MSampEn at scale 10” plot.

Flow behavior characterization is important to understand gas-liquid two-phase flow mechanics and further establish its description model. An Electrical Resistance Tomography (ERT) provides information regarding flow conditions at different directions where the sensing electrodes implemented. We extracted the multivariate sample entropy (MSampEn) by treating ERT data as a multivariate time series. The dynamic experimental results indicate that the MSampEn is sensitive to complexity change of flow patterns including bubbly flow, stratified flow, plug flow and slug flow. MSampEn can characterize the flow behavior at different direction of two-phase flow, and reveal the transition between flow patterns when flow velocity changes. The proposed method is effective to analyze two-phase flow pattern transition by incorporating information of different scales and different spatial directions.