Gas-Liquid Two-Phase Flow Measurement Models for Coriolis Mass Flowmeters Based on SVM and ANN

• Main Authors: Zhuang, Xuan, Sun, Lijun, Zhao, Zihui, Shao, Xin
• Format: Conference Proceeding
• Language: English
• Subjects: Artificial neural network; Artificial neural networks; Coriolis mass flowmeter; Gas-liquid two-phase flow; Generalization ability; Measurement uncertainty; Search methods; Solid modeling; Support vector machine; Support vector machines; Training; Volume measurement
• ISSN: 2161-2927
• DOI: 10.23919/CCC58697.2023.10239752

The flow measurement under the conditions of gas-liquid two-phase flow is an urgent problem to be solved. Coriolis mass flowmeters are introduced into the two-phase flow measurement due to its superior performance in single-phase flow measurement. Due to the large error in the two-phase flow measurement using Coriolis mass flowmeters, this paper uses the Support Vector Machine (SVM) based on the PSO-grid search method and Back Propagation - Artificial Neural Network (BP-ANN) based on PSO algorithm to model the liquid mass flowrate correction and the gas volume fraction (GVF) prediction respectively. Besides, this paper employs the backward stepwise selection method combined with the correlation analysis to select the appropriate input variables. Two indicators, accuracy and generalization ability, are used to evaluate the quality of the model. According to the experimental results, the appropriate model is selected, and the effectiveness of the model is verified. Finally based on the Baker flow regime map, the reason why a better model can be obtained from the training set after excluding some flow points is analyzed.