基于支持向量回归的OPGW应变分析

OPGW strain analysis based on support vector regression

介绍了使用布里渊散射测试光纤应变的原理,针对BOTDR(布里渊光时域反射仪)技术难以捕捉光纤微应变的不足,提出了一种基于支持向量回归的OPGW(光纤架空地线复合缆)应变分析方法。该方法对光纤段在不同拉力条件下的频移变化进行建模,建立频移向量并运用核函数思想将向量映射至高维空间,通过支持向量回归对频移向量进行回归分析。实验证明了该算法的有效性,并显示了该算法在OPGW异常检测中的巨大潜力。

This paper introduces the principle of testing fiber strain by using the Brillion scattering. In view of the deficiency of the BOTDR technology in monitoring fiber micro-strains, it presents a Support Vector Regression （SVR）-based method for Optical Power Ground Wire （OPGW） strain analysis. In this method, it builds a model for the frequency shift variations of fiber sections under different tensile forces, maps the vector to high-dimensional space using the kernel function and makes regression analysis by using the SVR method. Experimental results prove the effectiveness of this algorithm and demonstrate its huge potential in OPGW anomaly detection.