| 60 | 0 | 35 |
| 下载次数 | 被引频次 | 阅读次数 |
为解决车轴锻件尺寸大、工序长，导致其径向锻造过程中晶粒尺寸预测和控制难，而采用有限元计算则会耗费大量时间的问题，采用Forge有限元软件开展了车轴径向锻造过程全流程模拟，获得了混合非时序数据和时序数据组合工艺参数的输入特征量和以晶粒尺寸及位置坐标为输出特征量的数据集，应用时序神经网络构建了车轴径向锻造晶粒尺寸演变预测模型。经训练和评估后的模型可有效预测训练数据外的车轴径向锻造过程晶粒尺寸演变，结果与有限元模拟结果相近。因而，该模型可满足工程领域快速预测径向锻造过程晶粒尺寸演变的需求。
Abstract：To solve the problem that it is difficult to predict and control the grain size of axle forgings during radial forging process because of the large size and long working procedure, and the finite element calculation consumes a lot of time, Forge finite element software was used to simulate the whole process of radial forging of axle, and the data set of the input features of the process parameters of the combination of mixed non-temporal data and temporal data and the output features of the grain size and position coordinate were obtained. The prediction model of the grain size evolution of axle radial forging was constructed by using temporal neural network. After training and evaluation, the model can effectively predict the grain size evolution of axle radial forging process outside the training data, and the results are similar to the finite element simulation results. Therefore, this model can meet the need of engineering field to predict grain size evolution in radial forging process quickly.
[1] 陶竑宇，付传锋.中国高速铁路几个重要零件的精锻成形[J].兵器装备工程学报，2016,37(6):119-123.TAO Hongyu,FU Chuanfeng.Precision forging of several important parts of China high-speed railway[J].Journal of Ordnance Equipment Engineering,2016,37(6):119-123.
[2] 徐方，董湘怀，王新宝，等.矩形截面件径向锻造工艺优化[J].锻压技术，2019,44(4):1-9.XU Fang,DONG Xianghuai,WANG Xinbao,et al.Optimization of radial forging process for rectangular section parts [J].Forging & Stamping Technology,2019,44(4):1-9.
[3] BAMBACH M,IRAN M,SIZOVA I,et al.A soft sensor for property control in multi-stage hot forming based on a level set formulation of grain size evolution and machine learning[J].Advances in Industrial and Manufacturing Engineering,2021,2:100041.
[4] PETRIK J,BAMBACH M.Deep forge:Leveraging AI for microstructural control in metal forming via model predictive control[J].Journal of Manufacturing Processes,2024,121:193-204.
[5] 赵志光.30Cr2Ni2Mo钢热变形微观组织演变及其神经网络模型的研究[D].太原：太原科技大学，2023.ZHAO Zhiguang.Study on microstructure evolution andits neural network models of 30Cr2Ni2Mo steel in hot deformations[D].Taiyuan:Taiyuan University of Science and Technology,2023.
[6] 孔德瑜，晏洋，张浩，等.基于GA-BP神经网络晶粒尺寸预测模型的轮端轮毂锻造工艺优化[J].精密成形工程，2024,16(3):44-51.KONG Deyu,YAN Yang,ZHANG Hao,et al.Optimization of wheel end hub forging process based on GA-BP neural network grain size prediction model [J].Journal of Netshape Forming Engineering,2024,16(3):44-51.
[7] 马进.某新型车轴钢快锻成形过程晶粒尺寸演变及预测研究[D].上海：上海交通大学，2021.MA Jin.Study on evolution and prediction of grain size during rapid forging of a new type of axle steel [D].Shanghai:Shanghai Jiao Tong University,2021.
[8] 骆俊廷，赵静启，杨哲懿，等.基于Deform软件二次开发和BP神经网络的TA15多向锻造微观组织预报[J].航空学报，2021,42(12):327-344.LUO Junting,ZHAO Jingqi,YANG Zheyi,et al.Microstructure prediction of TA15 multi-direction forging based on Deform software secondary development and BP neural network [J].Acta Aeronautica et Astronautica Sinica,2021,42(12):327-344.
[9] 曹中源.车轴用EA4T钢微观组织演化模型的构建及其在锻造工艺设计中的应用[D].上海：上海交通大学，2019.CAO Zhongyuan.Construction of microstructure evolution model of EA4T steel for axle and its application in forging process design [D].Shanghai:Shanghai Jiao Tong University,2019.
[10] ABUEIDDA D W,KORIC S,SOBH N A,et al.Deep learning for plasticity and thermo-viscoplasticity[J].International Journal of Plasticity,2021,136:102852.
[11] MOZAFFAR M,BOSTANABAD R,CHEN W,et al.Deep learning predicts path-dependent plasticity[J].Proceedings of the National Academy of Sciences,2019,116(52):26414-26420.
基本信息:
DOI：
中图分类号:U270.6;TG316;TP183
引用信息:
[1]胡志文,任明明,向华等.基于时序神经网络的车轴径向锻造晶粒尺寸演变预测[J].塑性工程学报,2025,32(08):33-39.
基金信息:
山西省重点实验室项目(202204010931026); 太原市揭榜挂帅项目(2024TYJB0108)