Mass Flow Rate Measurement of Gas-Liquid Two-Phase Flow Using Multi-Sensor Data Fusion and Soft Computing Model

This article presents a novel method for measuring the mass flow rate of gas-liquid two-phase flow based on the multi-sensor data fusion and soft computing model. A multi-sensor system comprising a throat-extended Venturi tube (TEVT) and a dual-modality electrical sensor (DMES) has been developed for gas-liquid two-phase flow measurement. Soft computing models are employed to address the intricate non-linear mapping between the measurement data and flow parameters. Initially, flow regimes are identified based on the time-domain features of the multi-sensor data using a support vector machine (SVM). Subsequently, mass quality is derived from the multi-differential pressure fluctuations and the eigenvalue sequence of the normalized electrical matrices, employing a hybrid neural network comprising a convolution neural network and a deep neural network (DNN). Ultimately, gas/liquid over-reading (OR) is predicted via extreme gradient boosting (XGBoost) using multi-differential pressure ratios. The gas and liquid mass flow rates are subsequently derived from the preceding results. The proposed method addresses the issue that the parameters measurement of gas-liquid two-phase flow is significantly influenced by the flow regimes, and achieves accurate flow rate measurement under the diverse flow regimes. Experimental validation confirms the method’s effectiveness and superior performance compared to conventional approaches.