论文部分内容阅读
针对关节臂式坐标测量机的测量误差问题,对关节臂式坐标测量机的误差源进行了分析。利用GA-BP神经网络对关节臂式坐标测量机的单点测量误差进行了模型的建立与预测分析,避免了复杂的数学关系推导。应用HEXAGON-Infinite2.0型关节臂式坐标测量机,在不同的误差参数输入的条件下测量标准锥窝得到单点测量误差,获得模型的训练样本和测试样本,并对模型进行了预测和验证。结果表明,GA-BP神经网络可以对关节臂式坐标测量机进行单点测量误差的预测,GA-BP神经网络模型预测的相对误差的平均值为5%,具有较高的预测精度。
According to the measurement error of articulated arm CMM, the error source of articulated arm CMM is analyzed. GA-BP neural network was used to establish and predict the single-point measurement error of the articulated arm CMM, thus avoiding the complicated derivation of the mathematical relation. HEXAGON-Infinite2.0 articulated arm coordinate measuring machine was used to measure the standard cone under the condition of different error parameters to obtain single-point measurement error. The training samples and test samples were obtained and the model was predicted and validated . The results show that the GA-BP neural network can predict the error of single-point measurement of the arm-arm CMM. The average relative error of the GA-BP neural network model prediction is 5%, with a high prediction accuracy.