Multidimensional inverse radiation heat transfer analysis using Monte Carlo methods

The use of Monte Carlo methods for recovering temperature distributions in multidimensional geometries from spectral emission data is investigated. The Monte Carlo global optimization methods are used to obtain the temperature distribution. The governing equation of the emerging radiation is solved by a Monte Carlo simulation. The technique is verified using calculated and experimental emission data for one-dimensional and multidimensional problems by comparing recovered temperature distributions with predicted ones.; A physical model is formulated for a three-dimensional geometry containing absorbing, emitting, and scattering medium and having opaque or transparent boundaries. The Monte Carlo simulation consists of tracing all spectral energy bundles emitted from the physical system for an assumed temperature distribution. The Monte Carlo optimization method is used to vary the assumed temperature distribution which is expressed in a form of Legendre polynomials in order that the one which best matches the measured and calculated spectral emissions can be found.; The technique is first applied to the one-dimensional problem. Two different cases are studied. In the first case, a temperature distribution and the spectral emissions can be determined analytically. The technique is evaluated using calculated emission data with or without random data error. In the second case, the experimental data are obtained from the literature. The technique is evaluated using the existing experimental data. Both results show the validity of the Monte Carlo approach for solving the one-dimensional inverse problem.; The application of the method to two- and three-dimensional problems is investigated by using numerically calculated temperature distributions and emission data. The effects of number of bundles, element size, random data error, and variations of optical properties are also studied. The inversion results indicate that the technique can accurately recover the temperature profile for the multidimensional problem.; A two-dimensional experiment is conducted to provide emisison data and temperature measurements from an opaque surface. A three-dimensional heat transfer analysis is performed to predict the temperature distribution on the surface from the measured ambient temperature and total heat flux. Comparison of the recovered, predicted, and measured temperature distributions shows that the Monte Carlo technique works successfully for the multidimensional problem using experimental emission data.