Abnormal Judgment of Blade Machining Process Based on SVDD

Abstract

Abstract Aiming at the problem of stability monitoring of blade processing, this paper proposes a method to judge the stability of blade machining process. The measured point deviation of the blade pattern can represent the processing quality of the blade. Firstly, the dimension of the normalized deviation is reduced by the kernel principal component analysis, and appropriate kernel functions and parameters can be determined. Eigenvectors of samples are trained by support vector data description, and the parameters of model are determined with false alarm rate. In addition, the radius of hyperspheres are calculated to obtain control limits. The example data is applied to verify the feasibility of the method. The experimental results show that abnormal fluctuation of blade machining process can be effectively detected by the method.