摘要
模锻件各部位受应力和温度作用的差异性会形成不同的晶粒尺寸。以Cr-Co-Mo-Ni齿轮钢模锻件为对象,结合某锻造厂的实际模锻工艺参数,利用DEFORM软件中的神经网络技术,建立了Cr-Co-Mo-Ni齿轮钢晶粒尺寸和峰值应力的预测模型,并将计算结果与工业试验结果进行了验证对比。结果表明,该模型对于晶粒尺寸的预测最大误差为6.56%,模型精度较高,能够较好地用于模锻过程不同工艺参数下对晶粒尺寸的预测,继而为改善模锻件晶粒尺寸均匀性提供重要的理论基础。
Because of the otherness from stress and temperature, the grain sizes at different position of die forging would be different. Based on an actual die forging process in a certain forge plant, the Cr-Co-Mo-Ni gear steel was regarded as the object of study. A neural network prediction model from DEFORM software for grain size and peak stress of Cr-Co-Mo-Ni gear steel was established. The verification results from industrial test showed that the maximum error of prediction model is 6.56%, which had a high accuracy and could precast the grain size under different process parameters in die forging process. This research could provide an important theoretical basis for optimizing the grain size uniformity of die forging.
出处
《工业加热》
CAS
2014年第6期11-15,共5页
Industrial Heating
基金
国家自然科学基金(51074021)