Gas–Solid Two-Phase Flow Pattern Identification Based on Artificial Neural Network and Electrostatic Sensor Array

Abstract

A method for gas–solid two-phase flow pattern identification in horizontal pneumatic conveying pipelines is proposed based on an electrostatic sensor array (ESA) and artificial neural network (ANN). The ESA contains eight identical arc shaped electrodes. Numerical simulation is conducted to discuss the contributions of the electrostatic signals to the flow patterns according to the error recognition rate, and the results show that the amplitudes of the output signals from each electrode of the ESA can give important information on the particle distribution and further infer the flow patterns. In experiments, the average values and standard deviations of the eight output signals’ amplitudes are respectively extracted as the inputs of the ANN to identify four kinds of flow patterns in a pneumatic conveying pipeline, which are fully suspended flow, stratified flow, dune flow and slug flow. Results show that for any one of those two input values, the correct rates of the ANN model are all 100%.