薄壁变形机匣件限位检测机理分析与预测

为了研究带有一定公差的薄壁机匣件二次装夹变形规律,以约束表面和相关表面为研究对象,建立了机匣件约束表面与相关表面变形关系的数学分析模型。使用有限元法模拟变形机匣件的限位装夹过程,将模拟产生的数据作为BP神经网络的训练样本,借助神经网络的非线性映射能力建立了装夹变形的预测模型。通过对某航空材料7075机匣件作试验分析,验证了所构建的三维有限元模型与神经网络预测模型较为合理,能够为机匣件的加工制造提供反馈,为机匣件的装配分析提供良好的校验。

Mechanism Analysis and Prediction of Restraint Measuring Method of Thin-Walled Deformed Engine Casing

In order to study the deformation law of the secondary clamping of thin-walled engine casing with certain tolerances,the mathematical analysis model of the deformed relationship between the constraint surface of the casing and the related surface,being taken as the research object,is established.The finite element method is used to simulate the restraint clamping process of the deformed casing.The data generated by the simulation is used as the training sample of BP neural network.The prediction model of the clamping deformation is established by the nonlinear mapping ability of the neural network.Through the experimental analysis of aeronautical material 7075 casing,it is verified that the constructed three-dimensional finite element model and neural network prediction model are reasonable,which can provide feedback for the machining and manufacturing of the casing,and provide good verification for the assembly analysis of the casing.