ABSTRACT Response surface modeling is a well-grounded method to identify empirical relationships between factors involved in engineering experiments. This paper presents a novel mathematical/statistical approach to design an optimal predictive model to evaluate hardness by eddy current method. A new approach has been presented to find an optimal frequency in eddy current testing which maximizes detectability and accuracy of the measured values. In order to determine the hardness values of AISI D2 tool steel during tempering, the eddy current tests were conducted by applying different values of operating frequency in the range 50–1000 Hz. Unknown heat treatment conditions estimated from eddy current tests creates a link that provides hardness estimation. The results showed that coupling the eddy current method to the proposed model acts as a high accurate hardness measuring system for the samples with unknown heat treatment conditions.

中文翻译：

---

涡流法表征D2工具钢硬度的预测模型：一种统计优化方法

摘要 响应面建模是确定工程实验中涉及的因素之间的经验关系的一种有充分根据的方法。本文提出了一种新颖的数学/统计方法来设计最佳预测模型，以通过涡流法评估硬度。提出了一种新方法来寻找涡流检测中的最佳频率，从而最大限度地提高测量值的可检测性和准确性。为了确定 AISI D2 工具钢在回火过程中的硬度值，通过在 50-1000 Hz 范围内应用不同的工作频率值进行涡流测试。根据涡流测试估计的未知热处理条件创建了一个提供硬度估计的链接。