论文部分内容阅读
以铣削加工为对象，研究了多刃切削加工过程的刀具状态监测问题，从系统的角度分析了刀具状态的多传感器监测原理，并以此为依据确定了采用声发射（ＡＥ）传感器和动态切削力传感器可有效地监测加工过程。文中提出了一种多传感器信号的特征提取方法，该方法利用偏最小二乘法计算样本矩阵的本征值，根据置信因子确定特征维数。为验证该方法的有效性，建立了一个铣削加工实验系统，实验结果表明，该方法可在多种切削条件下获得较高的识别率。
In this paper, the problem of cutter status monitoring in multi-blade cutting is studied. The multi-sensor monitoring principle of cutter status is analyzed from the angle of system. Based on this, the AE sensor and dynamic cutting The force sensor effectively monitors the process. In this paper, a method for feature extraction of multisensor signals is proposed. This method uses the partial least squares method to calculate the eigenvalues of the sample matrix, and determines the feature dimension according to the confidence factor. In order to verify the effectiveness of the method, a milling experiment system is established. The experimental results show that this method can obtain high recognition rate under a variety of cutting conditions.