摘要
文章针对S700K容易出现时间轴平移和扭曲的8种故障模式和正常模式所对应的功率曲线,提出一种改进动态时间规整的道岔故障诊断方法。首先,将功率曲线样本集分为参考模板集和测试模板集;其次,计算待识别功率曲线和各参考模板的规整距离;最后,将规整距离最小的参考模板对应的故障号作为待识别功率曲线的最大可能故障号。仿真结果表明,该方法可适用于时间轴平移和扭曲的功率曲线,具有较高的准确率和较快的诊断速度。
This paper proposes an improved dynamic time warping turnout fault diagnosis method for the power curves corresponding to the eight fault modes and normal modes that are prone to time axis translation and distortion in S700K.Firstly,the power curve sample set is divided into the reference template set and test template set.Secondly,the warping distance between the power curve to be identified and each reference template is calculated.Finally,the fault number corresponding to the reference template with the smallest warping distance is taken as the maximum possible fault number of the power curve to be identified.The simulation results show that the method can be applied to power curves with time axis translation and distortion,and it has high accuracy and fast diagnosis speed.
作者
欧阳鑫锋
孔令刚
OUYANG Xinfeng;KONG Linggang(Key Laboratory of Opt-Electronic Technology and Intelligent Control of Ministry of Education,Lanzhou Jiaotong University,Lanzhou 730070,China)
出处
《现代信息科技》
2023年第20期136-139,143,共5页
Modern Information Technology