Caracterização experimental e modelagem estatística das curvas de resfriamento do aço aisi 4340 em diferentes tratamentos térmicos

Abstract

The limited number of quantitative studies on the influence of cooling rate on the thermal, mechanical, and magnetic properties of AISI 4340 steel motivated the development of this research. This work aimed to experimentally characterize and statistically model the cooling curves of the material under annealing and normalizing conditions, correlating thermal parameters with microstructural transformations and resulting material properties. The methodology involved heat treatments under distinct cooling conditions (inside the furnace and in still air), real-time temperature monitoring, microstructural analysis, Rockwell hardness testing, and magnetic characterization using a Hall-effect sensor. Cooling curves were modeled using nonlinear regression in RStudio software. Two distinct models were selected for each condition: a bi-exponential thermal decay model (ExpDec2) for the annealing process and a double exponential model (Von Bertalanffy) with asymptotic behavior for the normalizing process. Both models showed excellent fit quality, with coefficients of determination (R²) above 0.999 and mean absolute error values below 1 °C, validating their applicability for describing thermal behavior. Microstructural analysis showed that the annealed sample contained 71.2% ferrite and 28.8% pearlite, while the normalized sample exhibited 21.9% retained austenite, 36.8% bainite, and 41.3% martensite. These differences were reflected in the hardness values: 59.3 HRB (annealed), 76.0 HRB (normalized), and 72.0 HRB (as-received). Magnetic testing revealed the highest magnetic flux density in the annealed sample (725.28 Gauss at 2 V), with less than 0.2% variation among all conditions. It is concluded that the cooling rate significantly influences phase formation and the mechanical behavior of AISI 4340 steel, whereas magnetic properties are sensitive to microstructural changes. The statistical modeling applied proved to be an effective tool for predicting thermal behavior and optimizing heat treatment parameters in industrial applications