On the Nature-Inspired Algorithms Applied to Characterize Heat Transfer Coefficients

Inverse Heat Conduction Problems (IHCP) are known as “reverse engineering” problems, due to the reversal of a cause-effect sequence, in the field of heat transfer analysis. An inverse problem means that some of the initial, boundary conditions or material properties are not fully specified as determined from the measured temperature profiles at some specific locations. The inverse problems in most situations are likely to be ill-posed. Solutions of the inverse problem are very sensitive to measurement errors, i.e. small errors in the measured data values can produce very large errors in solutions. In general, the exclusivity and stability of an inverse problem solution is not guaranteed. In recent years, the inverse problems have been studied extensively due to their applications in various engineering disciplines. In this work, an inverse analysis for the reconstruction of local coordinate and a time-varying Heat Transfer Coefficient, in two-dimensional cylindrical coordinates is investigated. The inverse heat conduction analysis is based on the applications of a Nature-Inspired approaches. Transient temperature measurements at multi-locations in the body of the work piece, obtained by the solution of the direct heat transfer problem, served as the virtual experimental data required to solve the inverse analysis. The fitness function which is defined by the quadratic residual between the measurements and the calculated temperatures is minimized. The Nature-Inspired algorithms have been parallelized and implemented on a GPU architecture. Numerical results are demonstrated that the determination of Heat Transfer Coefficient functions can be performed by using the novel computational method, as well as, the GPU implementation; provide a less time consuming and accurate estimation.