【摘要】 在金属塑性成形过程的应变路径控制中 ,控制模型的建立是一关键。文中所采用的应变路径控制模型是建立在材料物性参数 m值实时辨识前提下以数值模拟和神经网络技术为基础的控制模型。根据上一步加载材料的实际应变增量、所加应力增量以及材料所在的应力状态识别出材料物性参数 m值 ,再根据材料加载后的应力状态、目标应变增量以及识别所得 m值 ,由训练好的识别应力增量的人工神经网络产生应加的载荷增量。在 MTS试验设备上进行了薄壁管拉扭试验 ,通过试验过程中的实际应变路径与想要得到的目标应变路径的比较 ,验证了在正确识别材料物性参数前提下 ,基于数值模拟和神经网络的应变路径控制方法的正确性更多还原

【Abstract】 Real time control of deformation path is key to precision metal forming. To implement real time control, we propose using ANN(artificial neural networks) and numerical simulation in our new approach for controlling deformation that is based on identification of material parameter m . To our best knowledge, the control model for our new approach does not exist and has to be established first. Section 2 discusses in much detail our new approach for controlling deformation path. Fig.2 shows the schematical diagram of the control model for our new approach; it shows the relationships among four things: (1) present stress state, (2) stress increments identified by ANN, (3) identified value of m , (4) MTS (Material Test System) testing machine. Using the control model in Fig.2 and performing iterative calculations, we can achieve real time control of deformation path. We did testing to see how effective is our new approach in achieving real time control in precision metal forming. Fig.5 shows the designed deformation path (indicted by dots) as compared with deformation path achieved by our approach for real time control (indicted by solid squares). The comparison shows that our new approach gives fairly good results.更多还原