Reduction of edge effect using response surface methodology and artificial neural network modeling of a spur gear treated by induction with flux concentrators

The aim of the study is to determine the effect of each parameter involved in the induction heating process on the final temperature distribution and case depth dispersion of a spur gear placed between two other gears having identical shapes and acting as a flux concentrator using two different approaches. The purpose of flux concentrators is to adjust the heat distribution in the part at the end of the heating process and to produce a better case depth between the middle and the edge of the gear. Mechanical properties of the gear could be improved by minimizing the edge effect at the tooth; thus, the optimization of temperature gradient between the middle and the edge plan by varying geometrical and machine parameters was studied. Two structured and comprehensive approaches to design an efficient model based on analysis of variance (ANOVA) and artificial neural networks (ANN) for the estimation of quality and the prediction of temperature profiles and edge effect was developed. The obtained results demonstrate that the statistical model was able to predict accurately the behavior of temperature and case depth distribution. In the final phase, several experimental tests were conducted on the induction machine to validate the simulation results and the prediction model.

本研究的目的是采用两种不同方法，确定感应加热过程中涉及的每个参数对放置在另外两个形状相同且作为磁通集中器的齿轮之间的正齿轮的最终温度分布和硬化层深度离散度的影响。磁通集中器的目的是在加热过程结束时调整零件中的热量分布，并在齿轮中部和边缘之间产生更好的硬化层深度。通过最小化齿部的边缘效应可以改善齿轮的机械性能；因此，研究了通过改变几何参数和机器参数来优化中部和边缘平面之间的温度梯度。开发了两种结构化且全面的方法，用于基于方差分析（ANOVA）和人工神经网络（ANN）设计一个有效的模型，以评估质量以及预测温度分布和边缘效应。所得结果表明，统计模型能够准确预测温度和硬化层深度分布的行为。在最后阶段，在感应机上进行了几次实验测试，以验证模拟结果和预测模型。