一种汽液两相分析方法、装置、介质和设备
By correcting the liquid phase area through grayscale image processing and neural network models, the error in liquid holdup calculation caused by uneven liquid film distribution is solved, enabling accurate liquid holdup measurement of various flow patterns during nuclear reactor flood cooling, and suitable for real-time monitoring and analysis of complex flow patterns.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHONGQING UNIV
- Filing Date
- 2025-03-10
- Publication Date
- 2026-07-17
AI Technical Summary
In the process of nuclear reactor flood cooling, the non-uniform distribution of the liquid film phase interface causes refraction, which affects the accuracy of liquid content calculation in traditional visualization data processing. Especially under narrow channels and high temperature conditions, invasive equipment disrupts the flow pattern, and non-invasive methods have failed to effectively handle complex mixed flow patterns.
By combining grayscale image processing with a neural network model, the liquid phase area is corrected through an elliptical structure model, the liquid and vapor phase regions are extracted based on a gradient algorithm, and deep learning is used to identify dispersed droplets and calculate the total liquid phase area to obtain an accurate liquid content.
It significantly improves the accuracy of liquid content calculation, can identify multiple flow patterns in narrow rectangular channels, is suitable for complex mixed flow patterns, provides non-invasive real-time monitoring and analysis, and improves the accuracy of flow and heat transfer parameters.
Smart Images
Figure CN120125563B_ABST