Performance Analysis of Metaheuristic Optimization Algorithms in Estimating the Interfacial Heat Transfer Coefficient on Directional Solidification

Abstract

In this paper is proposed an evaluation of ten metaheuristic optimization algorithms applied on the inverse optimization of the Interfacial Heat Transfer Coefficient (IHTC) coupled on the solidification phenomenon. It was considered an upward directional solidification system for Al-7wt.% Si alloy and, for IHTC model, a exponential time function. All thermophysical properties of the alloy were considered constant. Scheil Rule was used as segregation model ahead phase-transformation interface. Optimization results from Markov Chain Monte Carlo method (MCMC) were considered as reference. Based on average, quantiles 95% and 5%, kurtosis, average iterations and absolute errors of the metaheuristic methods, in relation to MCMC results, the Flower Pollination Algorithm (FPA) and Moth-Flame Optimization (MFO) presented the most appropriate results, outperforming the other methods in this particular phenomenon, based on these metrics. The regions with the most probable values for parameters in IHTC time function were also determined.