2.25CrlMo0.25V钢热变形组织演变模型开发

发布时间：2018-03-23 05:14

本文选题：2.25Cr-1Mo-0.25V钢　切入点：动态再结晶模型　出处：《太原科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：大型锻件常被用为大型机械设备的核心部件。恶劣的工作环境对材料的宏观力学性能和微观组织状态都提出了极高的要求。掌握这类大型锻件在热变形过程中微观组织演变规律,对实现锻造过程中对微观组织的预测和工艺参数的制定具有重要的指导性意义。2.25Cr-1Mo-0.25V钢是石化工业中的大型核心设备——加氢反应器的主要制造材料。本文以该材料为对象,对其高温热变形后微观组织的演变进行了分析和建模。本文在Gleeble-1500D热模拟试验机上以40μm、70μm、120μm三种不同平均晶粒尺寸的2.25Cr-1Mo-0.25V钢作为试样,采用不同的热变形参数,进行热压缩试验。通过对试验获得的不同热变形参数下的应力应变曲线以及变形后试样的微观组织,阐述了2.25Cr-1Mo-0.25V钢热变形组织演变规律,讨论了热变形参数对变形后试样平均晶粒尺寸、动态再结晶体积分数和晶粒尺寸的影响。然后以热压缩试验获得的试验数据为基础,在讨论了BP神经网络的算法、网络结构以及模拟退火算法相关参数后,建立了基于模拟退火算法优化的BP神经网络的2.25Cr-1Mo-0.25V钢的动态再结晶模型,包括4-17-11-1的平均晶粒尺寸模型、3-14-1的临界应变模型、4-14-7-1的动态再结晶体积分数模型及4-16-12-1的动态再结晶晶粒尺寸模型。对各个模型的训练结果进行分析,各模型训练样本相对误差最大仅为4.71%,测试样本最大仅为5.74%,说明了各模型具有较好的准确性和泛化性能。将建立的各个模型,通过二次开发的形式嵌入DEFORM有限元软件中,实现了对2.25Cr-1Mo-0.25V钢热变形过程中动态再结晶过程的模拟。通过模拟结果与一组2.25Cr-1Mo-0.25V钢圆柱体镦粗试验试验数据的比较,验证了所建立动态再结晶模型的准确性。
[Abstract]:Large forgings are often used as the core components of large mechanical equipment. The poor working environment requires very high macroscopic mechanical properties and microstructures of materials. Master this kind of large forgings in the process of hot deformation. The law of the evolution of the organization, It is of great guiding significance to predict the microstructure in forging process and to formulate process parameters. Steel 2.25Cr-1Mo-0.25V is the main manufacturing material of hydrogenation reactor, which is a large core equipment in petrochemical industry. The evolution of microstructure after hot deformation at high temperature was analyzed and modeled. In this paper, three kinds of 2.25Cr-1Mo-0.25V steel with different average grain size, 40 渭 m, 70 渭 m and 120 渭 m, were used as samples on Gleeble-1500D thermal simulation machine, and different thermal deformation parameters were adopted. Thermal compression test was carried out. Through the stress-strain curves under different thermal deformation parameters obtained from the test and the microstructure of the specimens after deformation, the evolution law of hot deformation microstructure of 2.25Cr-1Mo-0.25V steel was described. The effects of thermal deformation parameters on average grain size, dynamic recrystallization volume integral number and grain size of deformed samples are discussed. Based on the experimental data obtained from thermal compression test, the algorithm of BP neural network is discussed. The dynamic recrystallization model of 2.25Cr-1Mo-0.25V steel based on BP neural network optimized by simulated annealing algorithm is established after the network structure and the parameters of simulated annealing algorithm. The critical strain model of 4-17-11-1, the critical strain model of 3-14-1, the dynamic recrystallization volume integral number model of 4-14-7-1 and the dynamic recrystallization grain size model of 4-16-12-1, are included. The training results of each model are analyzed. The maximum relative error of each model training sample is only 4.71, and the maximum of test sample is only 5.74, which shows that each model has better accuracy and generalization performance. Each model will be embedded in DEFORM finite element software through the form of secondary development. The dynamic recrystallization process of 2.25Cr-1Mo-0.25V steel during hot deformation was simulated, and the accuracy of the dynamic recrystallization model was verified by comparing the simulation results with that of a group of 2.25Cr-1Mo-0.25V steel cylinder upsetting test data.
【学位授予单位】：太原科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TG142.1

【相似文献】