ABSTRACT The aim in this article is to evaluate microstructural changes, hardness variations, and wear behavior of H13 hot work tool steel as a function of austenitizing and tempering temperature using nondestructive magnetic hysteresis loop method. To obtain different microstructural characteristics in the H13 specimens, austenitizing and tempering temperatures were varied in the range of 1,050–1,100°C and 200–650°C, respectively. The microstructural features, hardness, and wear loss were characterized using X-ray diffraction/metallographic examinations, hardness measurements, and a pin-on-disk wear tester, respectively. The relations between features obtained from the conventional methods and parameters extracted from the magnetic hysteresis loops were established. Results demonstrate that the proposed nondestructive method is able to assess the wear behavior of the heat treated H13 tool steels. Besides, a standard Generalized Regression Neural Network (GRNN) was trained with a training dataset and then used to estimate the hardness of a given sample with its measured values of magnetic parameters. Experimental results indicate that, if the training dataset has sufficient samples, the proposed method will have a very high accuracy to estimate hardness of the sample, nondestructively.

中文翻译：

---

使用磁滞回线法对热处理 H13 工具钢的显微组织和机械性能进行无损表征

摘要 本文的目的是使用无损磁滞回线法评估 H13 热作工具钢的显微组织变化、硬度变化和磨损行为随奥氏体化和回火温度的变化。为了在 H13 试样中获得不同的显微组织特征，奥氏体化和回火温度分别在 1,050-1,100°C 和 200-650°C 的范围内变化。分别使用 X 射线衍射/金相检验、硬度测量和销盘式磨损测试仪表征显微结构特征、硬度和磨损损失。建立了从传统方法获得的特征与从磁滞回线中提取的参数之间的关系。结果表明，所提出的无损方法能够评估热处理 H13 工具钢的磨损行为。此外，标准的广义回归神经网络 (GRNN) 使用训练数据集进行训练，然后使用其磁参数测量值来估计给定样本的硬度。实验结果表明，如果训练数据集有足够的样本，所提出的方法将具有非常高的准确度，非破坏性地估计样本的硬度。