An improved cuckoo search algorithm for inverse heat conduction and heat convection

Abstract

A competitive improved mode of cuckoo search algorithm (CSA) named step enhanced cuckoo search algorithm (SECSA) in terms of enhanced search step length is proposed to solve inverse heat transfer problems (IHTPs) in the present research. In SECSA, a traction factor is introduced and hence the global and local search is balanced. Further, an adjustment strategy is suggested, to adaptive step size scaling factor and adaptive discovery probability. It is optimized to constrain the randomness of the original algorithm, strengthen the information exchange between populations, and improve computational efficiency and robustness. Therefore the performance of the original CSA is improved. SECSA along with collocation spectral method (CSM) is utilized to solve conduction and convective inverse heat transfer problem (IHTP) of reconstructing boundary conditions, optimizing thermophysical parameters as well as identifying the heat flux, in both cylindrical coordinate and in Cartesian coordinate. For the irregular regions, body fitted coordinate transformation technique is adopted to transform the irregular regions into the regular ones, based on the theory of adaptive coordinate transformation and the regularization of complex variable functions. Results are analyzed both qualitatively and quantitatively in the considerations of population quantity, expansion degree and number of measuring points. The statistical results show the potential performance of improved CSs in IHTPs. The convergence speed, calculation accuracy and universality are significantly improved.