基于刀具振动模式特征的深孔圆度误差预测方法

Prediction Method for Roundness Error of Machined Deep-hole Based on Vibration Behaviors of Cutting Tools

依据深孔加工的实际工作特点,提出了以刀具振动信号时频信息为输入特征的加工孔圆度误差预测方法。为了构建刀具振动特征与深孔加工圆度误差之间的映射关系,利用改进的模糊聚类技术,并将其引入到标准线性支持向量机算法中,使得刀具振动模式的模糊输入空间划分问题转化成初始输入空间的初值问题,实现了在规则数较少的情况下,仍具有较好的圆度误差辨识算法、辨识精度及泛化能力。以此为基础,构建了深孔钻削加工孔圆度误差的预测模型,并结合大量的实验数据,验证了提出的加工孔圆度误差预测模型的有效性与可行性。

On the basis of feature of machined deep hole,an intelligent recognition method is proposed to predict the roundness error of machined hole by using the time-frequency domain information of machining tool cutting vibration. In order to establish the relationship of tool cutting vibration and roundness error,the partitioning problem of the fuzzy input-space of vibration mode is converted into the initial value of input space by introducing the modified fuzzy clustering algorithm into support vector machine,so that the higher accuracy and generalization ability of the proposed method are guaranteed simultaneously. On this basis,a prediction model for roundness error of machined deep-hole is built. A series of experiments were conducted with different cutting parameters and electric currents acting on vibration suppression instrument. The experimental results show that the proposed method is effective and feasible.