Étude de sensibilité du profil de dureté des engrenages traités thermiquement par induction en fonction des paramètres machine

Abstract

Nowadays, induction heat treatment processes are increasingly used in automotive and aeronautics to improve the service performance of various mechanical components such as gears. Indeed, by its selective heating, the process can generate a surface hardness profile and produce residual compressive stresses favorable to the fatigue behavior without generating excessive distortion. The surface induction hardening is performed using an electromagnetic heating followed by rapid quenching. The surface hardness, the case depth and the compressive residual stress generated during the induction hardening ar machine dependent parameters (power machine, frequency, preheating and heating times, cooling rate and design of the inductor and the shower). The vast complexity of the phenomena and the current state of the art are such that there is no empirical or numerical models able to predict the hardness profile and residual stress with an good enough accuracy to be used during development cycle and industrial design. Therefore, manufacturers must establish the parameters for induction heat treatment by trial and error. This method is costly and time consuming. Therefore, this project involves a study of sensitivity of the hardness profile of cylindrical bars and spur gears heated by induction using finite element models (axisymmetric and 2D) combined with experimental tests. This sensitivity study is considered as the basic step towards the development of prediction models. The study mainly aims at evaluating the effect of material properties and machine parameters on the hardness profile. As the induction heating is very fast, it is reasonable to assume that the material properties are different from those measured under thermodynamic equilibrium conditions. For this reason, the study attempts to measure the effect of variations in material properties on the surface temperature using the axisymmetric model. The results show that the relative magnetic permeability is the property that most significantly influences surface températures and the variation of this property has a large effect on the hardness profile. The effects of specific heat and electrical conductivity are rather low, while the thermal conductivity has a negligible effect on the model developed. Moreover, the variation in temperature of austenitizing margins have limited effects on the developed model. Therefore, the use of material properties at thermodynamic equilibrium was sufficient to establish models for predicting trends. Moreover, study of global sensitivity profile of hardness as a function of machine parameters allowed, in the first place, to compare the two cases of heat to medium and high frequency, by applying them to a disc and a gear and determine the effect of the imposed current density and heating time on the hardness profile using the results of the simulations. It appears that the edge effect is smaller in the MF heating case. In addition, the 2D model can confirm that the MF power heat more the tooth root, while the HF power heat the tooth tip region. Finally, the regression models developed were used to estimate surface temperatures and depths hardened with time of heating and power received by the part. The experimental tests, performed on discs and gears with the same exterior dimensions, have confirmed a clear correlation between hardness profiles measured and those obtained by simulation. Moreover, since the power transmitted by the machine to the coil is not known and it is difficult to measure the current in the inductor, the issue was to find a method leading to a match between simulation and experimentation. Indeed, tests have allowed the calibration of developed models by evaluating the power ratio between the power received by the part and that provided by the machine. Thus, the calibrated models are able to predict global trends of the hardness profile as a function of machine parameters even if the material properties measured at thermodynamic equilibrium are used. Finally, a local sensitivity analysis was performed by varying the machine parameters slightly around the nominal values leading to a definite hardness profile. Based on the statistical tools of variance analysis, this study showed that the hardness profile is very sensitive to small variations of the power machine and the heating time in the case of the disc. With regard to the gears and considering the dual frequency sequential heating mode, HF power and HF heating time are the parameters that most affect the hardness profile. Although the study focused on the specific geometry and a given material, it is clear that the models developed can be used to develop mechanical components and they help reduce the development time. The greatest contribution of this work has been to bridge the gap between theory, simulation and practice.Therefore, it laid the groundwork to achieve a more general and robust model capable of supporting the interactions between the machine parameters, variables, dimensions and material properties. The accuracy of the model could be greatly improved by considering inductor current measurment during heating and material properties caracterization in thermodynamic nonequilibrium conditions.

Date	9 May 2011
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Philippe Bocher (Supervisor) & Jean Brousseau (Co-supervisor)