Research On Flow Regime Identification Based On Wavelet Analysis And Neural Network For Electrical Resistance Tomography System

Two-phase flow exists in modern industrial production extensively, influencing development of industry in extent degree. Two-phase flow system is a nonlinear and dynamical system, Two-phase fluid has complex flow characteristic, there are effect of interface and relative velocity in phases, and this makes the measurement of two-phase flow parameters become more difficult. In two-phase flow system, flow regime is an important measurement parameter, the accurate identification of flow regime is the basis of measuring the other parameters of two-phase flow accurately, so the intelligent identification of flow regime is one important role of two-phase flow research. The research of the paper is based on 12-electrical resistance tomography system and flow regime of oil-water two-phase flow, and the application of wavelet analysis and neural network in two-phase flow regime identification are studied.This paper summarized the significance of the measurement technology of two-phase flow in science research and industry based on extensive reading, at the same time the development process, current research situation and the crucial problem of resolution of the identification method of two-phase flow regime are analyzed and summarized in detail, the shortcoming of flow regime identification which is based on image reconstruction is pointed out. A method of flow regime identification which is based on extracting feature of measurement data is proposed. The problem of extracting feature which is based on measurement data of electrical resistance tomography system is researched by the method of wavelet analysis. Taking the common flow regime as the research object, the method of extracting feature of measurement data which is based on wavelet packet analysis is researched. Do lots of studying on effectively making use of statistic feature of wavelet energy of signal and containing feature of flow regime. And the extracted data will be taken as input information of radial basis function neural network, both the modeling and the simulation are made for neural network.As to the current method of flow regime identification based on electrical resistance tomography system, a comprehensive and deep research is made. On the basis of numerous simulations, a method of flow regime identification of neural network which is based on extracting feature of measurement data is proposed, the experiment result of simulation is analyzed and summarized deeply. Through research work of the paper, an effective mean is provided for flow regime identification of two-phase flow.