1.Key Laboratory of Advanced Transducers and Intelligent Control System, Ministry of Education and Shanxi Province, Taiyuan 030024, China 2.Institute of Optoelectronic Engineering, College of Physics and Optoelectronics, Taiyuan University of Technology, Taiyuan 030024, China
Fund Project:Project supported by the National Natural Science Foundation of China(Grant Nos. 61527819, 61875146) and the Shanxi Scholarship Council of China (Grant Nos. 2016–036, 2017–052).
Received Date:04 January 2019
Accepted Date:18 February 2019
Available Online:01 May 2019
Published Online:20 May 2019
Abstract:The high-precision structural health monitoring of large civil structures and materials are increasingly demanded with widely using the distributed fiber sensors. A Brillouin optical correlation domain analysis for millimeter-levelhigh spatial resolution sensing using broadband chaotic laser is proposed and demonstrated. Through the analysis of the influence of polarization state and feedback strength on the chaotic laser, we experimentally achieve a broadband chaotic laser with a spectrum over 7.5 GHz in –3 dB which means that the theoretical spatial resolution is 3 mm, and we also successfully measure the distribution of fiber Brillouin gain spectrum with a temperature over 300 m measurement range with 7.05 mm spatial resolution, which is the first time that the sensor system based on chaotic laser has achieved the measurement with millimeter-level. However, there is still a difference in spatial resolution between the experimental and theoretical values. We can find that the chaotic laser has a time-delay feature; besides, with the broadening of chaotic laser, the threshold of stimulated Brillouin scattering in optical fibers increases while the Brillouin gain will weaken if the pump power is not enough here, and the cross-correlation peak of chaotic laser will narrow. All these problems cause the Brillouin gain signal to be easily submerged by noise, so the performance of the chaotic Brillouin optical correlation domain analysis system will decrease ultimately. Therefore, we also propose an optimization of Brillouin optical correlation domain analysis system by introducing the time-gated scheme into pump branch. It is obvious that the peak power of the pump wave is heightened by more than 9.5 dB after being amplitude-modulated by a square pulse with a pulse width of greater than acoustic phonon lifetime, and the signal-to-back ground noise ratio of the gain spectrum is improved effectively in theory; the cross correlation between chaotic pump wave and probe waveis locked within a pulse duration time, and the residual stimulated Brillouin scattering interactions existing outside the central correlation peak can be largely inhibited. In this optimized setup, the performance of the distributed temperature sensing is improved to 3.12 mm spatial resolution, which corresponds well to the theoretical value. The improved chaotic Brillouin optical correlation domain analysis technology will have a great potential application in high-precision structural health monitoring of large civil structures. Keywords:broadband chaotic laser/ millimeter-level spatial resolution/ Brillouin optical correlation domain analysis/ temperature measurement
$\frac{{{\rm{d}}N\left( t \right)}}{{{\rm{d}}t}} = \frac{I}{{qV}} - \frac{1}{{{\tau _n}}}N\left( t \right) - G\left( t \right){\left| {E\left( t \right)} \right|^2},\tag{2b}$
$G\left( t \right) = \frac{{G\left[ {N\left( t \right) - N} \right]}}{{1 + \varepsilon {{\left| {E\left( t \right)} \right|}^2}}}.\tag{2c}$
实验选择DFB-LD在最佳反馈强度0.12且偏振匹配态b时的输出作为BOCDA系统光源, 此时混沌激光的–3 dB线宽约为7.5 GHz, 系统的理论空间分辨率约为3 mm. 设置探测光的扫频范围为10.55—10.75 GHz, 扫频步进为1 MHz, 获得待测光纤末端布里渊增益谱(Brillouin gain spectrum, BGS)随温度变化的结果如图3所示. 显然, 与室温下(23 ℃)测量得到的布里渊增益曲线(蓝色)相比, 加热到55 ℃时的布里渊增益曲线(红色)产生了约32 MHz的频移, 和布里渊频移(Brillouin frequency shift, BFS)对温度的灵敏度(1 MHz/℃)相吻合, 且此时测量增益谱的信号背景噪声比约为2.28 dB. 图 3 不同温度下待测光纤末端的布里渊增益谱 Figure3. The BGS at different temperature end of FUT
进一步测量得到待测光纤沿线布里渊频移的分布如图4所示. 其中, 图4(a)为300 m待测光纤沿线布里渊频移的整体分布图, 可以看到在室温区(23 ℃)布里渊频移稳定在约10.653 GHz, 光纤末端加热区(269 m处, 55 ℃)见图中红色曲线标示, 布里渊频移发生明显变化; 图4(b)为加热位置附近的局部放大图, 其中加热区(20 cm)内平均布里渊频移约为10.685 GHz, 布里渊频移变化量约为32 MHz, 与前述实验结果一致, 而且布里渊频移的标准差(standard deviation, Std)[18]约为 ± 1.8 MHz即测量温度误差约为 ± 1.8 ℃, 说明系统具有较好的测量准确性. 图 4 待测光纤沿线布里渊频移分布图 (a)整条光纤沿线的布里渊频移分布; (b)加热位置附近的局部放大图 Figure4. The map of BFS distribution along the FUT: (a) Measured along the entire FUT; (b) the local enlargement near heated zone
根据上述实验结果解调出待测光纤沿线布里渊频移的分布曲线如图5所示. 由于BOCDA系统的实验空间分辨率可以用上升沿和下降沿10%—90%所对应的光纤长度的平均值来表示[8], 如图中上升沿和下降沿对应的光纤长度分别为6.62 mm和7.48 mm, 取其平均值为7.05 mm. 所以基于宽线宽混沌激光BOCDA系统的空间分辨率约为7.05 mm, 突破了毫米量级分辨率的分布式温度测量. 图 5 待测光纤沿线布里渊频移分布曲线 Figure5. Measured distribution of the Brillouin frequency shift along the FUT
24.2.基于时间门控的宽线宽混沌BOCDA技术 -->
4.2.基于时间门控的宽线宽混沌BOCDA技术
上述基于宽线宽混沌激光的BOCDA系统, 其空间分辨率的理论值为3 mm, 但实验测量值仅有7.05 mm与理论值相比存在较大的误差. 从误差来源分析, 考虑到本文中所使用的光源为宽线宽混沌激光, 与窄线宽的混沌激光相比其在光纤中的受激布里渊散射阈值变大[22], 这里受激布里渊散射的阈值可表示为[23]
实验设置调制脉冲的持续时间?τ = 100 ns, 周期为T = 1.5 μs, 幅值电压为2.8 V, 此时被调制EOM的消光比达到最大约为20 dB[18]. 脉冲调制前后泵浦光的时序变化如图7所示, 其中蓝色曲线为脉冲调制前泵浦光的时序, 红色曲线为脉冲调制后的泵浦光时序, 可以看到脉冲调制后泵浦光的峰值功率提高了约9.5 dB, 理论上有效提高了测量增益谱的信号背景噪声比. 图 7 脉冲调制前后泵浦光时序图 Figure7. The time series of the chaotic pump waves (red) and pulse amplitude-modulated (blue)
图8所示为引入时间门控技术前后待测光纤中发生受激布里渊散射的示意图. 图8(a)为前述系统待测光纤中发生受激布里渊散射示意图, 可以看到两路光在待测光纤中间位置产生稳定的相关峰(即受激布里渊散射的声波场), 但由于混沌信号在外腔时延处发生弱幅自相关导致在主峰附近存在残余次峰[17], 并在受激布里渊散射放大时会引起额外的噪声并沿着光纤不断积累, 最终导致增益信号被噪声淹没而影响测量精度. 图8(b)为引入脉冲调制后两路光在待测光纤中发生受激布里渊散射示意图, 由于脉冲调制使两路光的相互作用被限制于脉冲持续时间内, 非中心峰放大和非零基底噪声被有效抑制. 图 8 引入时间门控技术前(a)后(b)两路光在待测光纤中发生受激布里渊散射示意图 Figure8. The schematic diagram of SBS in the previous system (a) and the time-gated system (b)
此时待测光纤中布里渊增益谱随温度变化的测量结果如图9所示. 因为在泵浦路中引入时间门控技术后, 测量的布里渊增益谱是泵浦光光谱和与待测光纤中布里渊谱的卷积, 经脉冲调制的泵浦光光谱在一定程度上被展宽, 导致布里渊增益谱被展宽, 最终布里渊频移量出现了1 MHz的测量偏差[18], 如图9(a)所示, 此时的布里渊频移量约为33 MHz. 但该偏差对温度的测量结果没有影响, 图9(b)所示为加入时间门控技术前后待测光纤中混沌布里渊频移随温度的变化关系, 可以看出加入时间门控技术后, 系统的温度系数由原来的1.03 MHz/℃变为1.09 MHz/℃, 所以33 MHz的频移量与32 ℃的实际温差相匹配. 同时相比于前述系统此时测量增益谱的信号背景噪声比由约2.28 dB提升为4.55 dB. 图 9 混沌布里渊增益谱和温度的关系 (a)待测光纤中随温度变化的布里渊增益谱; (b)加入时间门控技术前后待测光纤中随温度变化的布里渊频移量 Figure9. The relationship of the Chaotic BGS with temperature: (a) Temperature-dependence of the BGS in the FUT; (b) that of the BFS in the chaotic BOCDA systems with (blue) and without (red) the time-gated scheme
进一步得到此时系统中待测光纤沿线布里渊频移的分布如图10所示. 图10(a)为布里渊频移沿130 m待测光纤的整体分布情况, 可以看到, 在室温区(23 ℃)布里渊频移稳定在约10.653 GHz, 在加热区(125 m处, 55 ℃)布里渊频移发生明显变化. 图10(b)为加热位置附近的局部放大图, 在加热区(约3 m)内布里渊频移约为10.686 GHz, 且此时布里渊频移的标准差约为 ± 1.7 MHz. 图 10 待测光纤沿线布里渊频移分布图 (a)整条光纤沿线的布里渊频移分布; (b)加热位置附近的局部放大图 Figure10. The map of BFS distribution along the FUT: (a) Measured along the entire FUT:(b) the local enlargement near heated zone
同样根据上述实验结果解调出待测光纤沿线布里渊频移的分布曲线如图11所示. 图中上升沿和下降沿对应的光纤长度分别为3.06 mm和3.15 mm, 取其平均值得到此时系统的空间分辨率为3.12 mm. 如前所述, 本文所提出的基于宽线宽混沌激光BOCDA系统的理论空间分辨率约为3 mm, 可见在系统中引入时间门控技术后有效提高了系统的测量精度, 使系统的实验空间分辨率达到了其理论值. 图 11 优化后系统中待测光纤沿线布里渊频移分布曲线 Figure11. Measured distribution of the Brillouin frequency shift along the FUT in the setup after optimization