Research On Online Identification Method Of Gas-liquid Two-phase Flow Pattern Based On Embedded System

In human life and production,multiphase flow is widely used in chemical industry,petroleum,metallurgy,environmental protection and petroleum pipeline.Especially the research scholars at home and abroad pay more and more attention to the research of multiphase flow,which makes the research of multiphase flow become a research topic of scholars,the most special of which is gas-liquid two-phase flow.However,the phase interface between gas phase and liquid phase changes greatly with the change of flow rate and flow rate.In particular,the problem of the transition between flow patterns,coupled with the lack of technology and theory,the imperfection of experimental related equipment,has a great impact on the flow pattern acquisition and flow pattern recognition of gas-liquid two-phase flow.To solve this problem,the method of combining embedded system with image is put forward.By installing the virtual machine Vxworks in the Window system,then creating a new virtual machine,installing the Linux system,con Fig.and select the hardware compatibility Workstation 15.x,linux operating system,ubuntu64 bit,number of processors,memory size,I/ O controller type,network connection,etc.Finally,through transplantation,you can know that there is a shared folder between the embedded core development boards of the Ubuntu system,and you can share files with each other.At the same time,you can also download and install a series of software such as Python and Opencv.The language uses Python,and Opencv is used because it involves a lot of functions,and you can directly call the function library for free.It provides users with a large number of function libraries for image processing,and can quickly write programs and run them for image processing.deal with.This paper uses a high-definition camera to take a picture of the flow of a horizontal pipeline gas-liquid two-phase flow,which achieves non-contact,and is transmitted to the embedded system in the computer through the image collector to perform image grayscale,image noise reduction,and histogram equalization.According to image processing technology,laminar flow,wavy flow and bubbly flow can be identified according to image processing technology.Finally,VGG16 convolutional neural network can also be used to explain the flow law of gas-liquid two-phase flow more accurately.To realize the automatic recognition of different flow patterns.Therefore,the combination of embedded system and image processing technology has attracted more and more attention.