Abstract In the kinetic studies of martensitic phase transformation, the propagation velocity of the phase boundary is considered as the speed of transformation. But there is no interface at the beginning of phase transformation, we cannot track the development of the interface. Here we propose a new mathematical model that describes the rate of martensitic transformation by the data-driven method, which is based on the study of the change in area fraction in each phase. In this paper, the K-means clustering method is used to classify martensite and austenite through advanced quantitative differential interference contrast microscopy system. After classification, the clustering results are in good agreement with microscopic observations. The phase transformation rate for a set of tensile tests under various strain rates is determined. Our work provides a useful reference for modeling the strain-driven martensitic transformation rate.

中文翻译：

---

通过数据驱动方法研究应变驱动的马氏体相变速率

摘要 在马氏体相变动力学研究中，相界的传播速度被认为是相变的速度。但是相变开始时没有界面，我们无法跟踪界面的发展。在这里，我们提出了一个新的数学模型，该模型通过数据驱动的方法描述了马氏体转变速率，该模型基于对每个相中面积分数变化的研究。本文采用K-means聚类方法，通过先进的定量微分干涉对比显微系统对马氏体和奥氏体进行分类。分类后的聚类结果与微观观察结果吻合较好。确定了在各种应变率下进行的一组拉伸试验的相变率。