CO₂ Gas-liquid Two-phase Flow Measurement Using Coriolis Flowmeters Incorporating Neural Networks

With the development of industrialization in the world after the industrial revolution,the demand for energy has been increasing,which has led to the excessive emission of greenhouse gases represented by CO2.Under this background,CCS(carbon capture and storage)technology has emerged as the times require.emphasized.During the long-distance transport of CO2 in liquid form,a part of CO2 is converted into a gaseous state due to changes in pressure and temperature,so that it exists in the form of gas-liquid two-phase flow.In order to accurately know the fluid state within the pipeline and ensure safe transportation,real-time monitoring and measurement of the mass flow rate and gas volume fraction of the fluid in the pipeline is essential.In this paper,a prediction system of mass flow and gas volume fraction of CO2 gas-liquid two-phase flow based on Coriolis flowmeter and neural network is proposed after the research progress of CCS technology at home and abroad,and various measuring equipment and measuring technical means are known.In order to solve the problem that BP neural network is easy to fall into local minimum,genetic algorithm is introduced to optimize the initial weight of BP neural network.The parameters measured by Coriolis flowmeter and the measured values of differential pressure flowmeter in the experiment are taken as the input of neural network model to be selected.The most suitable input variables are selected through calculation,and then the structural parameters of neural network are determined according to the empirical formula to complete the model construction.Two series of tests were carried out for liquid mass flow rates from 250 kg/h to 2716 kg/h and GVF(gas volume fractions)from 0 to 30%,namely Test Ⅰ and Test Ⅱ.The fluid temperature during the test was approximately 20°C.A total of 142 data sets were collected from Experiment Ⅰ for training in neural network models,and 53 data sets were recorded from Experiment Ⅱ to test the performance of the neural network.The analysis of the test results shows the feasibility and effectiveness of the neural network combined with the Coriolis flowmeter for CO2 gas-liquid two-phase flow parameter measurement.