考虑垂直距离分级的光纤防外破识别算法研究

Research on the Identification Algorithm of Optical Fiber Anti-Breakthrough Considering Vertical Distance Classification

为进一步加强光缆的运行保护措施,提出了一种基于分布式光纤振动系统的光缆外破事件器械识别与垂直距离判断的双指标方案。通过实时采集光纤振动信号,分析时频域信号的多重特征,提取受影响光纤段、频率能量占比、频率峰值及分布等关键性特征,送入训练好的宏流形学习模型下,同时识别光缆外破事件的器械类型及垂直距离,并对垂直距离进行分级预警,高效开启对应的及时响应措施。案例结果表明,本方案在应用中的识别率高达90%,实现外破类型和垂直距离分级的双识别,大大提高了光缆外破事件的防御举措。

In order to further strengthen the protection measures for the operation of the optical cable,a dual-index scheme based on the distributed optical fiber vibration system for the equipment identification and vertical distance judgment of the optical cable external damage event is proposed.The fiber vibration signals in real time are acquired.The multiple features of time-frequency domain signals are analyzed.The key features such as the affected fiber section,frequency energy ratio,frequency peak value and distribution are extracted.The type of equipment and vertical distance of the incident of external optical cable breakage,the grading early warning of the vertical distance,and the effective activation of corresponding timely response are measured.The case shows that this scheme not only meets the recognition accuracy rate,but also ensures the recognition efficiency of the system,which has high practical application value.