基于相对误差的最小二乘法在切削力预测中的应用

对比分析高斯最小二乘法和基于相对误差的最小二乘法,后者的相对误差均值更小,适合保证实际值更小时的精度,更适合应用实际工程分析。在机械加工研究中,切削力通过应力、应变或测力仪传感器测量获得,使用成本较高,因此,提出了一种基于数控机床伺服参数的测量方法。测量得到电流环伺服参数TCMD和切削力,采用基于相对误差的最小二乘法,两者之间的关系模型,并将此关系模型应用于切削力的预测中。试验结果表明切削力可通过伺服参数进行预测,基于相对误差的最小二乘法的预测模型具有较高的精度。

APPLICATION OF RELATIVE ERROR BASED LEAST SQUARE METHOD IN CUTTING FORCE PREDICTION

Comparative analyzing Gauss least square method and relative error based least squares method, the latter method has smaller mean relative error and is more suitable for practical engineering application analysis. In the study of mechanical processing, cutting force is obtained by stress, strain sensors or dynamometer, which is of higher cost; this paper proposes a cutting force measuring method based on CNC machine tool servo parameters. Measure and obtain current loop servo parameters TCMD and cutting force, build the relationship model through relative error based least square method, and apply this model for cutting force prediction. The experiment results show that the cutting force can be predicted by using the servo parameters, and the relative error based least square prediction model has high precision.