论文部分内容阅读
针对工件-夹具系统的误差分离问题,基于径向基神经网络算法建立了夹具误差的分离和识别算法。根据工件位姿变化对测试点位移数据的函数关系,对测试数据进行处理,并使用径向基神经网络方法分别建立夹紧力-测试数据和测试数据-夹紧力的拟合模型,实现了定位误差与夹紧误差的分离,计算出工件的位姿变化量和夹紧力的大小,从而能够为误差补偿或者故障诊断提供数据支持。使用该算法对实验数据进行分离与识别,夹紧力和位姿变化量的预测误差分别控制在10%和13%以内。
Aiming at the error separation problem of workpiece-fixture system, a separation and identification algorithm of fixture error is established based on RBF neural network algorithm. According to the variation of position and orientation of the workpiece as a function of the displacement data of the test point, the test data are processed and the fitting models of clamping force-test data and test data-clamping force are established respectively using RBF neural network method, The separation of the positioning error and the clamping error calculates the variation of the pose and the magnitude of the clamping force of the workpiece so as to provide data support for error compensation or fault diagnosis. The algorithm was used to separate and identify the experimental data, and the prediction errors of the clamping force and the change of posture were controlled within 10% and 13% respectively.