论文部分内容阅读
为了实现对重庆市轻轨轨道梁锚固螺杆的故障检测,提出了一种基于支持向量数据描述的锚固螺杆故障诊断方法,该方法只需要正常螺杆样本,且不需要对原始数据进行特征提取,就可以建立单值分类器,解决了缺少故障螺杆样本的难题。通过与常见的三种单值分类方法比较,表明SVDD分类器具有很好的分类效果和计算效率,能较好地区分正常螺杆和非正常螺杆,为轻轨锚固螺杆故障检测提供了新的诊断方法。
In order to detect the failure of the anchorage screw in the light rail orbit of Chongqing, a new method to diagnose the problem of the anchoring screw based on the support vector data is proposed. The method only needs the normal screw sample and does not need to extract the feature of the original data. The establishment of a single value classifier, solves the problem of missing screw samples. Compared with the common three kinds of single value classification methods, it shows that the SVDD classifier has a good classification effect and computational efficiency, can distinguish between normal screw and non-normal screw, which provides a new diagnostic method for light rail anchoring screw fault detection .