三坐标测量机基本运动误差的相关性及建模

Correlation of Fundamental Motion Errors of CMM and Its New Modeling Method Modeling

针对现有几何误差软件补偿过程中采用线性插值法推算模型中的1 8项基本运动误差存在的问题 ,从理论上分析了构件沿各坐标轴运动 1 8项基本误差的直线度误差与转角误差的内在关系。提出用 4层前向神经网络建立三坐标测量机 1 8项基本运动误差模型 ,并以国产 SZC- 65 4型三坐标测量机为例进行基本运动误差神经网络建模研究 ,提高了误差补偿精度。

In this paper,it is firstly demonstrated theoretically the relationship between straightness error and angular error in eighteen fundamental errors, of CMM component along each axis, in accordance with some drawbacks of linear interpolation of the eighteen fundamental errors for the current geometric error software compensation Based on the conclusion, a new scheme is put forward, which utilizes four layer feed-forward neural networks in modeling eighteen items of fundamental errors of the CMM Finally type SZC-654 CMM made in China is taken for example to conduct research on modeling fundamental motion errors by means of neural networks to improve the error compensation accuracy of the CMM