基于支持向量机的贯流风叶叶片粘连缺陷诊断

Diagnosis of Cross-flow Fan Leaf Blade Adhesion Defects Based on Support Vector Machine

为了提高贯流风叶叶片粘连缺陷诊断的准确率和鲁棒性,提出了一种基于支持向量机的贯流风叶叶片粘连缺陷诊断方法。该方法以线性核函数为内积核函数,在追求分类间隔最大化的前提下,建立了叶片粘连缺陷诊断数学模型。仿真和实际测试结果表明,即使在使用较少的训练样本的情况下,该模型仍能达到较高的叶片粘连缺陷诊断率,效果优于传统的诊断方法,为贯流风叶叶片粘连缺陷诊断提供了新的途径。

A defects diagnosis method based on support vector machine is bustness of cross-flow fan leaf blade adhesion defects diagnosis. The linear kernel function. The mathematical model of cross-flow fan leaf blade adhesion defects for better accuracy and rois used as the inner product kernel diagnosis is established under the premise of pursuing maximum interval in classification. The simulation and actual test results show that the model reaches a higher accuracy of leaf blade adhesion defects diagnosis even with less training samples and has better performance than the traditional defects diagnosis methods.