光时域反射的事件分析算法研究

发布时间：2018-07-10 20:24

本文选题：光时域反射 + 短时傅里叶变换　；参考：《上海交通大学》2015年硕士论文

【摘要】：光时域反射(OTDR)是目前较为实用的一种光纤后向散射测量技术,它通过测量光脉冲沿光纤传输时所产生的后向信号,主要包括Rayleigh、Raman、Brillouin等散射信号类型及Fresnel反射,以获取光纤链路的特性,被广泛应用于无源光网络(PON)的实时监测和故障定位中。随着PON的日益庞大和复杂,研究OTDR的精准高效的事件分析算法显得尤为重要,它可减少网络维护时间和人力开销,以确保网络正常稳定地运行。传统的事件定位算法包括两点法与最小二乘法、Gabor变换法、小波变换法等。两点法与最小二乘法相结合是早期的事件定位算法中较为成熟的一种,它原理简单,容易编程实现,但事件判别易受噪声影响,精度差。针对于此,Gabor变换法作为一种时频分析的解决方案,可有效提高检测精度,但其窗函数固定,只具有单一分辨率。相比而言,小波变换法通过对OTDR数据做离散二进小波变换的多分辨率分析更为可取,但计算复杂度过高,不适合在现有的PON监测系统中采用。因此,本文对比研究了不同算法定位OTDR事件的性能优劣,基于计算复杂度和算法性能的折中考虑,提出了短时傅里叶变换(STFT)与模型相关匹配相结合的方法,有效地提高了事件判别的准确度和对噪声的容限度。另外,传统OTDR事件分析是基于点对点的光纤线路进行的,而在实际PON中,线路基本上都是复杂的点对多点结构,因而除了对事件定位算法提出更高要求以外,还需要合理的故障告警算法与之匹配,才能自动地及时告警故障点位置。本文提出的点对多点的故障告警算法,周期性地获取参考和测量两条迹线,设置故障点优先级,针对断点、裂缝、接头松动、熔接损耗等不同情况完成故障告警流程。最后,本文基于实验室自主研发的硬件测试平台,设计了相应的基于Code::Blocks的软件系统,包括算法的C/C++实现,利用wxWidgets插件用户界面设计,迹线显示及操作,故障信息告警等重要内容。该系统界面友好,流程合理,鲁棒性及实用性较强,能够很好地满足用户在线监控PON的需求。
[Abstract]:Optical time domain reflection (OTDR) is a more practical optical fiber backscatter measurement technique. It is widely used in passive optical network (PON) by measuring the backward signals produced by optical pulses along the optical fiber, mainly including the types of Rayleigh, Raman, and Brillouin scattering signals and Fresnel reflection. In time monitoring and fault location, with the increasing and complex of PON, it is very important to study the accurate and efficient event analysis algorithm of OTDR. It can reduce the network maintenance time and human cost to ensure the normal and stable operation of the network. The traditional event location algorithm includes two points method and the least square method, the Gabor transform method, the wavelet transform. The combination of two points method and the least square method is a mature one in the early event location algorithm. It has simple principle and easy programming, but the event discrimination is easily affected by noise and the precision is poor. As a solution of time frequency analysis, Gabor transform can effectively improve the detection precision, but its window function is fixed, only the window function is fixed, only the window function is fixed, only the window function is fixed, only the window function is fixed, only the window function is fixed, only the window function is fixed, only the window function is fixed, only It has a single resolution. In comparison, the wavelet transform method is preferable to multiresolution analysis of discrete two into wavelet transform by OTDR data, but the computational complexity is too high and is not suitable for the existing PON monitoring system. Therefore, this paper compares the performance of different algorithms to locate OTDR events, based on Computational complexity and The method of combining the short time Fourier transform (STFT) with the model correlation matching is proposed, which effectively improves the accuracy of the event discrimination and the tolerance to the noise. In addition, the traditional OTDR event analysis is based on the point to point fiber line, and in the actual PON, the line is basically a complex point pair. In addition to the higher requirements of the event location algorithm, in addition to the higher requirements for the event location algorithm, a reasonable fault warning algorithm is needed to automatically alarm the fault location in time. The point to multipoint fault warning algorithm proposed in this paper is to periodically obtain two trace lines, set the priority of the fault point, and crack the breakpoints. In the end, the corresponding software system based on Code:: Blocks, including the C/C++ implementation of the algorithm, the design of the user interface of the wxWidgets plug-in, the trace display and operation of the trace, the fault information alarm and so on, is designed based on the hardware test platform of the independent research and development of the laboratory. The system has friendly interface, reasonable process, strong robustness and practicability, and it can well meet the needs of users online monitoring PON.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN253

【相似文献】