Application of genetic algorithm and simulated anneling methaeuristics in a steel process

• Autor(a) principal: Pimenta, Cristie Diego
• Data de Publicação: 2021
• Outros Autores: Silva, Messias Borges, Marins, Fernando Augusto Silva
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Revista Produção Online
• Texto Completo: https://www.producaoonline.org.br/rpo/article/view/4200
• topic: Heat Treatment. SAE 9254. Response Surface Metodology; Genetic algorithms. Meta-heuristic.; Tratamento Térmico. SAE 9254. Metodologia de Superfície de Resposta. Algorítmos Genéticos. Meta-heurística.

The aim of this article is to show the application of Genetic and Simulated Anneling algorithms to optimize statistical predictions for the process of tempering heat treatment in steel wires. This statistical modeling may be able to replace the process used for the preparation of tempering and tempering furnaces, which is traditionally carried out by means of adjustments made from the result of the mechanical hardness property, tested in the laboratory and required to meet customer specifications. We sought to understand the influence of the input variables (factors) and their effects on the mechanical property hardness, in SAE 9254 steel wires, for the 2.00mm diameter, used for the manufacture of valve and clutch springs for the follow-up automobile. The main input variables of the process were investigated and for that, the Quadratic Multiple Regression and the Response Surface Methodology (RSM) were used. For the optimization of the statistical model, the methodologies Genetic Algorithm (AG) and the Simulated Annealing Meta-heuristic were used. The results revealed that it is possible to obtain good results if this statistical model is used and if the statistical model is optimized through the techniques applied in this article. If the methodologies are applied correctly, this could bring scientific advances that could provide the automation of this process, and consequently this could contribute to the increase in productivity and product quality.