Nature inspired algorithms for the solution of inverse heat transfer problems applied to distinct unsteady heat flux orientations in cylindrical castings

One of the most important parameters in the foundry design of castings is the heat transfer coefficient at the casting/mold surface (h), which significantly affects the solidification cooling rate and defines the as-cast microstructure and, consequently, its properties. The inverse heat transfer problem (IHTP) is a needed task to be solved in many casting processes to evaluate this coefficient. In this work, four different nature inspired algorithms [genetic algorithm (GA), evolutionary strategy (ES), artificial immune system (AIS) and particle swarm algorithm (PSO)] have been integrated to a numerical model of solidification to solve the IHTP problem with a view to evaluating the transient profile of h during solidification of three different cylindrical casting setups (outward, inward and upward). The aim is to reduce the number of iterations to achieve successful results by the association between metaheuristics and different sets of h values. Therefore, an objective function was modeled to match the temperatures predicted by the numerical model to those of experimental measurements. Sets of 10, 20, 50, 100 and 200 h values used to compose its transient profile have been evaluated by the four metaheuristics approaches aiming to characterize each casting situation.