Positioning Error Analysis Based On Dual M-Z Optic Fiber Disturbance Sensing System

With the progress of society and the continuous development of economy,the optical fiber sensing technology has been rapidly developed since the 1970 s.The dual Mach-Zehnder(M-Z)interference optical fiber sensing system has the advantages of simple structure,high sensitivity,easy signal processing,convenient demodulation,low cost and so on.It has been widely studied and applied in the fields of perimeter security,pipeline leak detection,bridge,dam and large facility construction projects.Dual M-Z optical fiber sensor positioning system uses external intrusion disturbance to cause the phase change of interference signal,and then demodulates the change of light intensity through interference,which is transformed into electrical signal by photodetector.By using cross-correlation algorithm,the time delay difference between the two interference signals is calculated,so as to realize the location of disturbance intrusion signal.I have built a positioning system of double M-Z optical fiber disturbance sensing in the laboratory,analyzed the error of the positioning results,and carried out research and discussion in several aspects.The main work is as follows:First of all,through reading and investigating a large number of literature,I understood the development status of interferometric fiber sensing technology,mastered the working principles of several relatively common interferometric fiber sensing systems,and learned the cross-correlation algorithm for calculating the time difference between two interference optical signals in depth.Secondly,a double M-Z optical fiber sensing positioning experimental system was built in the laboratory,and a large number of data were collected and processed.Through the analysis of the experimental results,it is found that there is always a large positioning error under the existing conditions.According to theoretical analysis,the positioning results of the dual M-Z fiber optic sensing positioning system are obtained by the cross-correlation calculation of the two interference signals,and it is found that the correlation directly affects the positioning results of the system.The higher the correlation between the two signals,the smaller the positioning error.The correlation between the two signals is determined by many factors,and the causes of the errors are also many aspects.The focus of this paper is the positioning error analysis of the dual M-Z fiber disturbance sensing system.Then,positioning error further analysis of the experimental system,based on 156 experimental data of the different characteristics of the interference signal waveform display can be divided into Phase transition point,type and the starting point of the other types,the former two types by statistics in according to the proportion of the total result,discovered that these two types of data error is bigger.Furthermore,the reasons for Phase transition point and the starting point of the interference signal waveforms are analyzed.Both of these two kinds of waveforms weaken the correlation of the interference signal and lead to large positioning error.Finally,the signal demodulation processing method of double M-Z fiber sensing system is improved.Through the simulation experiment calculation,it is found that the traditional direct cross-correlation algorithm has the problem of circular boundary,resulting in large positioning error,so we try to improve the algorithm,and adopt the window cross-correlation algorithm to reduce the boundary effect and reduce the error brought by the algorithm.According to the characteristics of type data with Phase transition point,the calculation method of window number and manual method are used to deal with the data of starting point problem.Through comparison,the positioning error of 156 experimental data calculated by using the window cross-correlation algorithm is reduced from 6.48% to 5.78%,the positioning error of Phase transition point data is reduced to 2.21%,and the positioning error of starting point type data is reduced to 4.58%.The total positioning error increased from6.48% to 3.93%.