在主动测量中的基于支持向量机尺寸判定的研究

Research on Size Determination Based on Support Vector Machine in Active Measurement

针对磨加工主动测量仪信号点到尺寸时的工件尺寸与信号点预设值由于圆度误差而存在偏差的问题,提出了基于支持向量机的圆度误差修正的尺寸判定方法。通过交叉验证参数寻优方法和最小二乘支持向量机理论搭建了圆度误差预测模型,实验分析模型的平均相对误差为MRE=0.0095,均方误差为MSE=0.0050;基于圆度误差的修正方法使到尺寸判定偏差绝对值最大不超过0.5μm,有效减弱了圆度误差对尺寸判定的影响,证明了基于圆度误差尺寸判定的合理性和可行性。

In view of the problem of the deviation between the workpiece size in the determination of the dimension of the active measuring instrument for grinding and the signal point preset value due to roundness error.Proposed a roundness error correction method for determining the size based on the support vector machine.The roundness error prediction model is established by cross validation parameter optimization method and least squares support vector machine theory.The experimental results show that the average relative error(MRE)equals 0.0095 and the mean square error(MSE)equals 0.0050.The correction method based on roundness error does not exceed 0.5μm.The effect of roundness error on size determination is effectively reduced,and the rationality and feasibility of roundness error size determination are proved.