论文部分内容阅读
为了解决直升机动部件疲劳损伤定位问题,提出了基于多输出支持向量回归算法的声发射源平面定位方法。以声发射信号的多个时域参数作为输入,破损点的平面坐标(x,y)作为输出,用支持向量回归机逼近输入输出之间的非线性映射关系,然后利用支持向量回归机的泛化推广能力,实现声发射源的平面定位。通过碳纤维材料试件断铅定位试验结果表明:该方法有效的实现了声发射源的平面定位,并且在收敛速度和定位精度上优于RBF神经网络。
In order to solve the problem of fatigue damage localization of helicopter moving parts, a planar acoustic emission source localization method based on multiple output support vector regression is proposed. Taking the multiple time-domain parameters of acoustic emission signal as input and the plane coordinates (x, y) of the damaged point as output, the support vector regression machine is used to approximate the nonlinear mapping relationship between input and output. Then, Promote the ability to achieve the location of acoustic emission sources. The results of the lead-lead positioning test by carbon fiber specimen show that this method can effectively achieve the planar location of acoustic emission sources and is superior to RBF neural network in convergence rate and positioning accuracy.