本文選題：光時(shí)域反射 + 短時(shí)傅里葉變換�。� 參考：《上海交通大學(xué)》2015年碩士論文

【摘要】：光時(shí)域反射(OTDR)是目前較為實(shí)用的一種光纖后向散射測量技術(shù),它通過測量光脈沖沿光纖傳輸時(shí)所產(chǎn)生的后向信號,主要包括Rayleigh、Raman、Brillouin等散射信號類型及Fresnel反射,以獲取光纖鏈路的特性,被廣泛應(yīng)用于無源光網(wǎng)絡(luò)(PON)的實(shí)時(shí)監(jiān)測和故障定位中。隨著PON的日益龐大和復(fù)雜,研究OTDR的精準(zhǔn)高效的事件分析算法顯得尤為重要,它可減少網(wǎng)絡(luò)維護(hù)時(shí)間和人力開銷,以確保網(wǎng)絡(luò)正常穩(wěn)定地運(yùn)行。傳統(tǒng)的事件定位算法包括兩點(diǎn)法與最小二乘法、Gabor變換法、小波變換法等。兩點(diǎn)法與最小二乘法相結(jié)合是早期的事件定位算法中較為成熟的一種,它原理簡單,容易編程實(shí)現(xiàn),但事件判別易受噪聲影響,精度差。針對于此,Gabor變換法作為一種時(shí)頻分析的解決方案,可有效提高檢測精度,但其窗函數(shù)固定,只具有單一分辨率。相比而言,小波變換法通過對OTDR數(shù)據(jù)做離散二進(jìn)小波變換的多分辨率分析更為可取,但計(jì)算復(fù)雜度過高,不適合在現(xiàn)有的PON監(jiān)測系統(tǒng)中采用。因此,本文對比研究了不同算法定位OTDR事件的性能優(yōu)劣,基于計(jì)算復(fù)雜度和算法性能的折中考慮,提出了短時(shí)傅里葉變換(STFT)與模型相關(guān)匹配相結(jié)合的方法,有效地提高了事件判別的準(zhǔn)確度和對噪聲的容限度。另外,傳統(tǒng)OTDR事件分析是基于點(diǎn)對點(diǎn)的光纖線路進(jìn)行的,而在實(shí)際PON中,線路基本上都是復(fù)雜的點(diǎn)對多點(diǎn)結(jié)構(gòu),因而除了對事件定位算法提出更高要求以外,還需要合理的故障告警算法與之匹配,才能自動(dòng)地及時(shí)告警故障點(diǎn)位置。本文提出的點(diǎn)對多點(diǎn)的故障告警算法,周期性地獲取參考和測量兩條跡線,設(shè)置故障點(diǎn)優(yōu)先級,針對斷點(diǎn)、裂縫、接頭松動(dòng)、熔接損耗等不同情況完成故障告警流程。最后,本文基于實(shí)驗(yàn)室自主研發(fā)的硬件測試平臺(tái),設(shè)計(jì)了相應(yīng)的基于Code::Blocks的軟件系統(tǒng),包括算法的C/C++實(shí)現(xiàn),利用wxWidgets插件用戶界面設(shè)計(jì),跡線顯示及操作,故障信息告警等重要內(nèi)容。該系統(tǒng)界面友好,流程合理,魯棒性及實(shí)用性較強(qiáng),能夠很好地滿足用戶在線監(jiān)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.
【學(xué)位授予單位】：上海交通大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2015
【分類號】：TN253

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 母國光,梁俊忠,王君慶,方志良;一種新的調(diào)頻條紋分析算法[J];光學(xué)學(xué)報(bào);1987年08期

2 朱述龍;快速近似主成分分析算法[J];遙感學(xué)報(bào);1999年01期

3 李海量;采用混合方程與外代數(shù)解法的電路分析算法[J];上海海運(yùn)學(xué)院學(xué)報(bào);1986年01期

4 吳龍華;對DPIV相關(guān)分析算法的優(yōu)化[J];水動(dòng)力學(xué)研究與進(jìn)展(A輯);2004年05期

5 楊立雄;GLR(K)文法及其分析算法[J];南昌大學(xué)學(xué)報(bào)(理科版);1985年01期

6 付松林,李新,楊留記;規(guī)范LR(K)分析算法的優(yōu)化[J];西北大學(xué)學(xué)報(bào)(自然科學(xué)版);1985年01期

7 岳博,焦李成;兩階段混合因子分析算法[J];自然科學(xué)進(jìn)展;2002年02期

8 趙冬玲;馮艷若;潘正運(yùn);;基于網(wǎng)格的K-均值聚類分析算法的研究[J];科技通報(bào);2014年07期

9 汪文琦;董強(qiáng);商少凌;吳t熻，

本文編號：2114600