基于支持向量机的铁磁构件裂纹识别

Identification of Crack in Metal Component Based on Support Vector Machine

鉴于支持向量机的优秀特性,将其应用在金属磁记忆检测中,构造基于梯度值的裂纹标识量,实现对裂纹的定位和裂纹程度的标定,并通过有限元仿真,计算裂纹和裂纹程度及对应的漏磁信号,得到训练样本和测试样本,并训练支持向量机。仿真结果表明:应用支持向量机实现铁磁构件裂纹识别是可行的。

In the light of excellence of support vector machine, we apply it in metal magnetic memory detection. When we consrruct marking quantum for crack based on gradient, accomplish locating for crack and calibration for crack extent, we can get training sample and testing sample by simulating based on finite element and train support vector machine. Simulation result manifest it is feasible that realize crack identificaton applying support vector machine.