基于神经网络的精密磨床床身结构优化研究

Abstract

变形和固有频率是评价精密机床结构刚性的重要参数。为了提高磨床的整体刚性,提出一种利用bP神经网络来优化精密磨床床身结构的方法。以自主开发的高精度平面磨床2Mk1760为研究对象,利用有限元分析软件AnSyS对床身进行静力分析和模态分析。为了减小精密磨床的变形和振动,将部分关键参数作为神经网络的输入,以此预计床身尺寸参数与变形量和固有频率之间的关系,从而获得磨床床身的优化尺寸。分析结果表明,通过神经网络优化后的床身变形量相比初始结构减小了36.51%,而固有频率增加了11.96%。

Deformation and inherent frequency are key parameters to evaluate structure rigidity for precision machine tool.In order to improve rigidity for grinding machine,this study proposes a method that using neural network to optimize the structure of precision grinding machine bed.Take precision grinding machine tool( 2MK1760) which is self developed for an example,the static structure and dynamic modal analysis is carried out by using finite element method( ANSYS).In order to alleviate the influence of deformation and vibration on precision grinding machine bed,some parameters are served as input of BP neural work model,which estimate the relationship between deformation and inherent frequency.According to analysis results,an optimization is obtained to determine the dimension of precision grinding machine bed.The results indicate that the deformation after neural network optimization compared with initial structure has decreased by 36.51%,meanwhile,the natural frequency has increased by 11.96%.