基于机器学习的环形锻件轧制毛坯智能推演设计方法

Intelligent deduction design method of ring forging rolling blank based on machine learning

针对环形锻件轧制成形工艺方案关键设计难点中的毛坯设计问题,结合环件轧制毛坯设计理论与机器学习算法,建立了基于BP神经网络算法的轧环毛坯尺寸智能推演模型和基于支持向量机算法的冲孔毛坯尺寸智能推演模型,利用从企业收集到的历史工艺案例数据对推演模型进行学习训练,通过对比不同推演模型结构下的工艺设计结果确定了最终的轧制毛坯智能推演方案,并利用有限元模拟和物理实验验证了智能推演方案的可靠性。

Aiming at the blank design problem in the key design difficulties of the ring forging rolling process scheme,combined with the ring rolling blank design theory and machine learning algorithm,an intelligent deduction model of the ring rolling blank size based on the BP neural network algorithm and an intelligent deduction model of punching blank size based on the support vector machine were established.The deduction model was learned and trained by using historical process case data collected from the enterprise,and the final intelligent deduction shceme of rolling blank was determined by comparing the process design results with different deduction model structures.The reliability of intelligent deduction scheme was verified by finite element simulation and physical experiment.