LBM computational domain planning method based on image recognition

Abstract

The invention discloses an LBM computational domain planning method based on image recognition, which surpasses the limitation that a mathematical expression required by a traditional specified computational domain expresses a computational domain and boundary conditions, and can divide the computational domain into an integer matrix containing predefined label information based on any closed image (engineering image or actual experimental image). According to the method, after a closed image is input, a plurality of regions are automatically divided through a series of preprocessing, pixel points in the regions can be selected, a drainage basin and a wall surface region are planned according to a predefined integer label, pixel points near the edge of the image are selected to automatically identify a boundary entrance and exit, and planning is also carried out according to a predefined integer. Finally, a matrix with various predefined integer information is output, and different regions and entrances and exits are distinguished by different integers, so that corresponding boundary conditions can be defined according to the predefined integers when the LBM method is used.

Description

technical field [0001] The MATLAB-based image recognition method of the present invention surpasses the limitation of the calculation field of the Lattice Boltzmann method (LBM) represented by traditional mathematical expressions, and can divide the calculation area into an integer matrix based on any closed image. The invention belongs to the field of computational fluid dynamics. Background technique [0002] The Lattice Boltzmann Method (Lattice Boltzmann Method) is a new computational fluid dynamics (CFD) method that was born in the past 30 years. The huge amount of calculation brought by the equation has made great progress in multiphase flow, heat transfer phase transition, magnetic fluid and lattice growth. It is an important computational fluid dynamics method. [0003] With the development of LBM, the models suitable for various force fields, phase transitions and boundary precision are becoming more and more perfect. In the programming process, the division of c...

AI Technical Summary

Problems solved by technology

In the programming process, the division of computational domains, wall boundaries, and inlet and outlet boundaries all require limited mathematical expressions to achieve. However, for many practical conditions, including pore flow, lattice growth, and microchannel flow in actual experiments, etc. , its boundaries are usually complex and difficult to represent with simple mathematical expressions, which limits the complexity of the LBM model to a certain extent

Images