Research On Data Processing Of Optical Fiber Sensing System Based On BOTDA

With the development of social industry level,distributed optical fiber sensing system has been favored by more and more researchers due to its unique advantages such as small size,light weight,strong anti-interference ability,low cost,high voltage resistance,high sensitivity,corrosion resistance,good electrical insulation performance and fully distributed sensing.By embedding the optical fiber sensor on the surface or in the body of the measured object,real-time monitoring of parameter changes can be realized,which is suitable for self-detection and self-diagnosis of intelligent structure and material performance improvement.And this can also meet the needs of large-scale networking and intelligence of human society.It completely overcomes the defect that the point-type sensor is difficult to comprehensively monitor the measured field,and has the advantage that the traditional sensor does not have,so it shows a very attractive prospect in the fault diagnosis and accident early warning in many fields such as energy,electric power,aerospace,construction,communication,transportation,security and defense,military and so on.The basic working principle of optical fiber sensing are under the external environmental factors such as temperature,light is vulnerable to the impact of these external factors(or modulation),leading to corresponding change such as intensity,phase,frequency,polarization mode etc.By detecting these changes,you can gain the the change of temperature or strain,so as to realize the sensing.The distributed optical fiber sensing system based on Brillouin Optical Time Domain Analysis(BOTDA)is a sensing system based on the mechanism of frequency shift changes of the stimulated Brillouin scattering effect in the optical fiber caused by external temperature or strain.This paper mainly proposes a data processing method for optical fiber sensing system based on BOTDA,and verifies the validity of the scheme through temperature measurement experiments.Current BOTDA system,in the accuracy of measurement,spatial resolution,sensing distance of more,is relatively mature,and this paper mainly focuses on the rapid processing of sensor data.Without any increase in the system complexity,we aim at exploring the new way of data processing to reduce the data processing steps and time,avoiding the use of a large number of accumulative average denoising,reducing fitting times and the influence of the parameters set in the process of denoising.On the whole,this paper introduces the distributed optical fiber sensing system based on Rayleigh,Raman,three kinds of brillouin scattering effect and the basic structure,emphatically elaborated the basic principle of distributed optical fiber sensing system based on stimulated Brillouin scattering effect optical time-domain analysis(BOTDA),gives a description of three coupled wave equations and its numerical solution,the analysis of threshold to the relation of the fiber length,and introduce the key components of the system,at the same time independently set up a set of optical fiber sensing system based on BOTDA.We made the FPGA pulse generator and the labVIEW-based wavelet filter.The acquisition signals processed by polarization and those not processed by polarization are compared and the system test is completedBy comparing the Brillouin scattering spectrum and image processing,we find that the Brillouin scattering spectrum has image characteristics.Image processing has the global nature,through image processing can reduce the system load and delay caused by a large amount of cumulative average and multiple mutation position fitting in the conventional optical fiber sensing demodulation.On the edge of the image processing is image grayscale change obviously,and the Brillouin gain spectrum of the morphological changes of gray level changes are similar to the edge.According to the principle of edge detection,we can use or improved Sobel operators to identify the sharpening edge,and then to extract the Brillouin frequency shift.We emphatically studied the use of Sobel operators in feature extraction of the Brillouin frequency in our article,the formula of shortening the time of measurement is deduced compared with the cumulative average,and the enhanced Sobel operators are verified by experiment in the feature extraction of Brillouin gain spectrum,and also verified the enhanced sobel operators have the secondary denoising ability,and we achieved the spatial resolution of the 10 m along 14 km optical fiber,whose maximum standard deviation is about 1.3 ?.Compared with the traditional Brillouin scattering spectrum fitting,the image processing method has a better global property,which avoids the repeated fitting operation when mutation occurs in multiple positions,and greatly reduces the steps and processing time of data processing.In the process,we also propose to apply principal component analysis to the denoising of Brillouin scattering spectrum.Through experimental analysis,it is found that in the current commonly used wavelet denoising,different wavelet bases have different effects on the denoising effect,resulting in large errors after sobel operator processing,and the selection of wavelet basis is also difficult.Since the PCA processing process is not affected by the human parameter setting and the processing result is only related to the data itself,the denoising method is not affected by the parameter setting,and it also avoids the cumulative average of a large number of data.The method through the way of "dimension" to analyze the main ingredient in data,and use main ingredients to achieve image reconstruction,the dimension of the data to rapidly restore the Brillouin gain spectrum,and then Brillouin frequency shifts achieve feature extraction by the method of edge detection based on sobel operators,and the effectiveness of the method is verified by experiment,which has realized the spatial resolution of the 10 m along the 14 km optical fiber and the maximum standard deviation is about 1.3 ?.A cross-sensitive measurement system based on double-peak Brillouin scattering spectrum is also presented.Based on the principle of microwave photonics,the system uses the principle that two gain spectra are generated by pumping lights of two different wavelengths in the optical fiber,and each gain spectrum generates different Brillouin frequency shifts under the same temperature or strain,so as to realize the cross-sensitive measurement of temperature and strain.Under the existing experimental equipment and conditions,the simulation and experimental spectral analysis are carried out.According to the the same mechanism,by adjusting the interval of wavelengths of two pumps to achieve the Brillouin gain spectrum superposition,the new Brillouin gain spectrum width will be obviously increased,in the micro-vibration measurement based on slope method,the vibration measurement range can be increased,we simulated this process.