Electrical Capacitance-Based Flow Regimes Identification—Multiphase Experiments and Sensor Modeling

Abstract

Electrical capacitance tomography (ECT) has been used in flow regime identification studies. Usage of raw measurements for direct time series analysis of the raw ECT data removes the need for image processing, leading to easier and faster recognition of flow regimes with non-invasive ECT sensor arrays. Based on raw ECT data, a simple algorithm that can be implemented online and provide a priori knowledge of the flow regimes is presented in this paper. In the approach proposed by us, the time series of each inter-electrode normalized capacitance measurements are taken over a suitable duration. Four mathematical operations are considered for data reduction: averaging, computation of standard deviation (SD), and high-pass and low-pass filtering. Eigenvalues of the matrices have been used for recognition of spatial features, free of artefacts caused by rotations. At this stage, the methodology categorizing flowing into stratified smooth, stratified wavy, annular, slug, or plug regimes has been tested with a significant data set of measurements of fully developed air-water flows in horizontal 56-mm diameter pipe. For selected flow regimes, four criteria, based on eigenvalues of matrices formed by raw data from the array of sensors, are tested and validated using both experimental and adequate numerical models of the sensor arrays. Based on the results presented in this paper, the approach using “frame-by-frame” time series analysis is capable of giving valuable process insight using non-intrusive sensing techniques and can be easily adapted to other sensing modalities.