Abstract Flow behavior of a metal during hot deformation is influenced by physical phenomena such as work hardening, dynamic recovery and dynamic recrystallization. Effects of these phenomena on flow stress can be expressed through various semi-empirical models. To find unknown parameters of these semi-empirical models, one can use a process such as system identification in the way that the difference between experimental data and model output become minimized. Genetic algorithm is one of reliable and flexible methods in this category that has gained extensive application in different fields of science; so, in this research, Genetic algorithm was used to model flow stress of API-X70 microalloyed steel considering mentioned metallurgical phenomena during hot torsion test. Accuracy of the developed models for dynamic recovery and recrystallization was evaluated through statistical methods. Results showed a good agreement between the developed models and experimental data and also indicated that these models are very suitable for predicting flow stress.

中文翻译：

---

API-X70微合金钢热流动行为的遗传算法建模与实验对比

摘要 金属在热变形过程中的流动行为受加工硬化、动态回复和动态再结晶等物理现象的影响。这些现象对流动应力的影响可以通过各种半经验模型来表达。为了找到这些半经验模型的未知参数，可以使用系统识别等过程，以使实验数据和模型输出之间的差异最小化。遗传算法是该类别中可靠且灵活的方法之一，已在不同科学领域得到广泛应用；因此，在本研究中，考虑到热扭转试验中提到的冶金现象，使用遗传算法对 API-X70 微合金钢的流动应力进行建模。通过统计方法评估了动态恢复和再结晶模型的准确性。结果表明开发的模型和实验数据之间有很好的一致性，也表明这些模型非常适合预测流动应力。