基于分布式光纤传感器的周界报警系统的研究

发布时间：2018-12-09 20:06

【摘要】：目前,中国机场多采用围栏、专人巡逻、红外等手段防止外来入侵,这些传统手段成本高、探测距离短、实时性差、漏报率高,无法满足机场安全运行的需求。分布式光纤传感系统作为一种高性能的周界报警系统,具有灵敏度高,抗干扰能力强,传输距离长等优点,是一种非常理想的机场周界报警系统。基于此,本课题研制了一套基于相位敏感光时域反射仪(Φ-OTDR)和马赫增德(Mach-Zehnder)联合探测的机场周界报警系统。系统采用Φ-OTDR系统实现对扰动的定位,采用Mach-Zehnder实现对扰动的识别。本文工作主要有以下几方面:第一,研究了基于相干探测Φ-OTDR的传感原理。基于对Φ-OTDR理论公式的详细推导,自行设计并搭建了一套Φ-OTDR实验平台,对系统各器件的选取进行了分析。第二,针对Φ-OTDR后向散射信号弱的问题,本文采用累加平均和小波联合去噪的方法,采用真实的试验数据,验证了联合去噪的有效性。针对传统定位算法易出现虚警的问题,本文采用防区定位算法,对该算法进行了场外实验,实验发现该算法定位准确率达到了80%以上。第三,结合Φ-OTDR的硬件系统,本文自行设计了一款Φ-OTDR软件系统。该软件系统采用生产者消费者模式将数据采集和数据处理分离,解决了两者之间的数据等待问题。此外,该软件系统实现了累加平均、小波去噪算法和防区定位算法。第四,由于Φ-OTDR系统只能对扰动进行定位,为增加对扰动的识别功能,本课题结合使用了Mach-Zehnder系统。利用已有的Mach-Zehnder系统,本文对Mach-Zehnder系统的扰动识别算法进行研究。针对马赫增德信号的非平稳特性,本文提出了基于互补集合经验模态分解(CEEMD)奇异值熵和多核支持向量机(SVM)的信号特征提取与识别方法。试验表明,此方法的识别率达到了90%。
[Abstract]:At present, Chinese airports mostly use fencing, special patrol, infrared and other means to prevent foreign invasion. These traditional methods have the advantages of high cost, short detection distance, poor real-time performance and high rate of missing reports, which can not meet the needs of airport safe operation. As a high performance perimeter alarm system, distributed optical fiber sensing system has the advantages of high sensitivity, strong anti-interference ability and long transmission distance. It is an ideal airport perimeter alarm system. Based on this, an airport perimeter alarm system based on phase sensitive optical time domain reflectometer (桅-OTDR) and Mach-Zehnder is developed. 桅-OTDR system is used to locate disturbance and Mach-Zehnder is used to identify disturbance. The main work of this paper is as follows: firstly, the sensing principle based on coherent detection 桅-OTDR is studied. Based on the detailed derivation of 桅-OTDR theoretical formula, a set of 桅-OTDR experimental platform is designed and built, and the selection of each device in the system is analyzed. Secondly, aiming at the weakness of 桅-OTDR backscattering signal, the method of cumulative average and wavelet joint denoising is adopted in this paper, and the real experimental data are used to verify the effectiveness of joint denoising. In order to solve the problem of false alarm in the traditional localization algorithm, this paper presents an off-site experiment on this algorithm, and finds that the accuracy of the algorithm is more than 80%. Thirdly, combining 桅-OTDR hardware system, this paper designs a 桅-OTDR software system. The software system uses producer and consumer model to separate data collection from data processing, and solves the problem of data waiting between them. In addition, the software system realizes the accumulative average, wavelet denoising algorithm and area location algorithm. Fourthly, because 桅-OTDR system can only locate the disturbance, in order to increase the recognition function of disturbance, the Mach-Zehnder system is used in this paper. Based on the existing Mach-Zehnder system, the disturbance identification algorithm of Mach-Zehnder system is studied in this paper. In view of the non-stationary characteristics of Maherzonde signal, this paper presents a signal feature extraction and recognition method based on complementary set empirical mode decomposition (CEEMD) singular value entropy and multi-core support vector machine (SVM). The experiment shows that the recognition rate of this method is 90%.
【学位授予单位】：中国民航大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212;TP277

【参考文献】