一种基于ANN的FGH96合金动态流动行为的电路模型

An Analog-Circuit Model of PM96 Superalloy Plasticity Behaviors for Real-time Control during Hot Deformation

在锻造领域 ,多采用数字模型进行工艺设计和过程控制 ,但数字模型无法完成材料成型过程的实时控制。因此 ,有必要建立一种适合于实时控制的材料动态流动行为模型 ,以提高生产效率和锻件质量。通过热模拟试验 ,对热等静压态 FGH96合金的高温流动特性进行了研究 ,用 BP( Back Propagation)网络建立了 FGH96合金隐式本构关系模型 ,利用模拟电路的快速反应与易于控制等特点 ,建立了基于 ANN的 FGH96合金的动态本构模型。该模型具有较高的预测精度 ,能很好地反映材料成型过程的动态流动行为 。

Production efficiency and quality of forgings made from difficult to deform materials suffer from the lack of real time control of its bulk forming process. At the present time, numerical models (such as numerical simulation based on FEM), are adopted broadly in technological design and process control by engineers in forging field. This model, however, cannot implement real time control of materials forming process. It is thus necessary to establish a dynamic flow behaviors model for materials fit for real time control of materials deformation processing to increase production efficiency and forging qualities. In this paper, the hot deformation behaviors of PM96 superalloy have been characterized in the temperature range 1 000~1 100℃ and strain rate range 0.001~0.1 s -1 using hot compressive simulation experiments on hot isostatic pressing (HIP) material. The implicit neural constitutive model (INCM) of PM96 superalloy was obtained by using BP artificial neural network (ANN). The dynamic constitutive model (DCM) was implemented by using analog circuit approach. Testing results show that the INCM based ANN and the DCM based analog circuit have high predictive precision and can well describe the dynamic behaviors of PM96 superalloy during deformation. The DCM can be applied to real time control of materials deformation processing.