Gas and liquid flow measurement of wet gas based on vortex meter-liquid film sensor and neural network

A new measurement solution based on vortex meter-conductance dual-modality system and neural network is proposed to achieve online measurement of both gas and liquid flow rates. A dual-conductance ring sensor is developed to obtain the information of liquid film thickness, disturbance wave velocity and frequency, which is optimized by using numerical method. The gas and liquid flow rates are predicted by an error back propagation neural network (BPNN) with optimized parameters, and different liquid film parameters are tested. It is found that the prediction accuracy is best when the input parameters are apparent gas flow rate, disturbance wave frequency, liquid film average thickness and density ratio, the percentage errors (PEs) of gas and liquid phase are within
$\pm$
1.0% and
$\pm 4.0\%$
, respectively. The prediction accuracy will decrease when the input parameter includes disturbance wave velocity, the maximum values of PEs for gas and liquid phase are 1.82% and 9.49%, respectively. Finally, detail discussion is given about the different prediction results.