Soft Sensing Model Of Two-phase Flow Based On Visibility Graph And Phase-space Complex Network

Two-phase flow exists widely in nature and some modern industrial production fields and is closely related to human life and production.The pneumatic conveying of ore sludge,coal powder,flour and other materials and the flow of materials in oil Wells are typical applications of two-phase flow.The study of two-phase flow can effectively serve the dynamic adjustment and optimization of the development scheme in industrial production and provide important guidance information for industrial energy saving and production increase.However,in the actual flow of two-phase flow,the shape and distribution of phase interface are randomly variable in time and space,resulting in the complex and diverse flow patterns of two-phase flow,which brings great difficulties to the accurate measurement of dynamic parameters of two-phase flow.Taking two-phase flow as the research object,this paper aims to solve the problem of dynamic measurement of gas-solid two-phase flow and gas-liquid two-phase flow and reveal the mechanism of the change of flow pattern of two-phase flow.The main work of this paper is as follows:Firstly,this paper designs and optimizes the electrostatic sensor from the point of view of measuring the dynamic parameters of gas-solid two-phase flow.The finite element simulation software COMSOL is used to simulate the electrostatic sensor and optimize the width of the two annular electrodes on the electrostatic sensor and the distance between the two electrodes.Then,dynamic experiments were carried out on the experimental platform of gas-solid two-phase flow to verify the performance of the sensor.Meanwhile,dynamic data of gas-solid two-phase flow were collected to provide data support for the subsequent measurement of dynamic parameters of gas-solid two-phase flow.After that,the four-sector sensors were designed and optimized and the data were obtained on the gas-liquid two-phase flow dynamic platform,which provided data support for subsequent identification of gas-liquid two-phase flow patterns and measurement of dynamic parameters.Then,a visibility graph network is constructed for the data of gas-solid two-phase flow,and the change of mass flow rate of gas-solid two-phase flow is analyzed by complex network method.At the same time,the aggregation coefficient entropy is used to analyze the network,and the flow state of gas-solid two-phase flow is described.In addition,we also propose a convolutional neural network model,which can realize dynamic measurement of mass flow in gas-solid two-phase flow,to solve the difficult problem of mass flow measurement in practical engineering.The results show that the proposed model has good performance under different working conditions.Finally,the collected gas-liquid two-phase flow data is used to construct a joint recursive network,and the average node degree is calculated to represent the evolution of gas-liquid two-phase flow patterns.The results show that this method can reveal the evolution mechanism of flow structure.In addition,a densely-connected neural network based on joint recursive network is proposed to measure the flow parameters of gas-liquid two-phase flow.The mean absolute percentage error(MAPE)of the model is less than2.8%,and the mean square error(MSE)is less than 0.0023,both of which are better than the comparison method.The experimental results show that the proposed soft sensing model can measure the dynamic parameters of two-phase flow well,and can also provide a new way for the measurement of dynamic parameters in industrial production.