Affiliations: Department of Electromechanical Engineering, Osaka
Electro-Communication University, Japan. | Department of Mechanical Engineering, Kansai
University, Japan.
Abstract: It is an important challenge to analyze a three-dimensional thermal
flow field in engineering, science, and agriculture. For such an analysis, it
is essential to measure physical quantities such as temperature and velocity
over the entire thermal flow field. This paper presents a measurement system
based on color image processing for temperature and velocity vector
distributions in a three-dimensional thermal flow field. Flow visualization is
accomplished by the use of thermo-sensitive liquid crystal tracers. An
algorithm for the color-to-temperature transformation using a multi-layer
feed-forward neural network is applied to three-dimensional natural convection
in a rotating cylindrical cell. Two-dimensional temperature distributions in a
slit plane are obtained by using the algorithm. A three-dimensional temperature
distribution is consequently constructed by interpolating the two-dimensional
distributions using the B-spline function. In addition, the Spatio-Temporal
correlation method is applied to the natural convection to obtain a
three-dimensional velocity vector distribution.
Keywords: three-dimensional measurement, color image processing, liquid crystal, velocity vector distribution, temperature distribution