A Stable And Fast Approach For Measuring Brillouin Frequency Shift
Qing Bai,Yu Zhang,Yu Wang,Dong Wang,Baoquan Jin,Yuncai Wang
DOI: https://doi.org/10.1109/piers-fall.2017.8293415
2017-01-01
Abstract:In a conventional scheme of a Brillouin Optical Time Domain Reflector (BOTDR), an electro-optic modulator (EOM) or an acoustic optical modulator (AOM) is generally utilized to modulate the continuous light into the pulse probe light. However, the fluctuation of the working point will result in the instability of the extinction ratio when an EOM is selected, and a slow modulation speed will bring about the inability of modulating a narrow pulse when an AOM is chosen. In view of the above issue, a scheme for stably modulating the pulse probe light based on a semiconductor optical amplifier (SOA) is proposed. Meanwhile, the scheme integrates a pulse signal source with a synchronous trigger output to simplify the hardware structure of the system. Firstly, a SOA optic pulse modulator is used in the optic link to produce the pulse probe light into the sensing fiber, and coherent detection is used to obtain the beat signal of the Brillouin scattering. Moreover, the microwave frequency-sweeping method is adopted to acquire the variation curves of power with the fiber length at different frequency points. Then, the Brillouin Frequency Shift (BFS) distribution along the sensing fiber can be figured out by processing the acquired original data. Specially, in order to speed the data process, a software kernel of underlying Levenberg-Marquardt algorithm is developed based on C language, and the MFC message response mechanism and multi-thread characteristics are used to build a man-machine interface. Finally, the strain distribution along the fiber can be calculated based on the proportionality coefficient of BFS change versus strain change. The experiment results show that the BFS measurement accuracy of +/- 3.74 MHz is achieved over a 10.7-kilometer sensing fiber with the spatial resolution of 1.2 m, meanwhile the data processing time is reduced by about 45 times compared to that when using the LabVIEW, which indicates the stable and fast BFS measurement is realized.