【摘要】 利用ANSYS的APDL语言建立磨床部件的参数化模型 ,计算出磨床部件的动态特性 ,快速采样得到BP神经网络模型的学习样本 ,建立基于BP神经网络的动态分析模型 ,将磨床部件结构参数与其动态特性之间的关系反映为神经网络模型的网络输入与网络输出之间的数学关系 ,从而方便地、快速地对磨床部件进行了动态灵敏度分析 .结果表明 ,在BP神经网络模型上进行磨床结构优化要比在有限元模型上方便、快速 ,该方法特别适用于对大型复杂结构的优化设计计算更多还原

【Abstract】 With the help of parametrization model, the dynamic properties of the grinder part were calculated and the train sample of the BP neural network was obtained quickly. The dynamic analysis models based on the neural network are built up. The relationship between dimensions and dynamic properties of the grinder’s part is defined as the mathematic relationship between inputs and outputs of the BP neural network. Thus the analysis of dynamic sensitivity for grinder’s part can be carried out conveniently and quickly. This method is particularly suitable for calculation of design optimization for large and complex parts.更多还原