摘要
正向计算零件形位公差的测点位置与零件定位基准之间的敏感度,定义损失函数为敏感度趋于零的均方误差,反向计算损失函数较于定位基准的梯度值,根据梯度方向迭代优化零件的定位设计。结果表明:根据零件的测点位置可以实现GD&T定位系统的自动优化,满足不同工厂对定位敏感度的设计标准。
The sensitivity between the measuring point position of the part geometric tolerance and the part positioning datum was calculated in the forward direction.The loss function was defined as the mean square error whose sensitivity tends to zero.The gradient value of the loss function relative to the positioning datum was inversely calculated.The positioning design of the part was optimized iteratively according to gradient direction.The results show that the auto optimization of GD&T positioning system can be realized according to the position of measuring points of parts, and the design criteria of different factories for positioning sensitivity can be met.
作者
夏文艺
XIA Wenyi(Ford Motor Research&Engineering(Nanjing)Co.,Ltd.,Nanjing Jiangsu 211100,China)
出处
《汽车零部件》
2023年第2期64-67,共4页
Automobile Parts
关键词
机器学习
梯度下降
定位设计
自动优化
machine learning
gradient decent
positioning design
auto optimize
