Study On Periodical Error In Fiber Fabry-Perot Strain Sensing System

After two decades of development, Extrinsic Fiber Fabry-Perot Interferometer (EFPI) strain sensor was demonstrated to be the best candidate for long time health monitoring on bridges due to it's simplicity, stability and accuracy, etc. Theoretically, the EFPI strain sensor has a good linearity and very high accuracy, but there are still some errors in a practical EFPI strain sensing system. Especially, when the sensor was continuously loaded, the output of the sensing system showed a periodical fluctuation error, and the amplitude of this fluctuation error is in the range of several micro-strain (με). This problem has not been solved yet. It reduced the reliability of our EFPI strain sensing system; furthermore, it became an obstacle which restricted the use and spread of the EFPI strain sensing system in health monitoring for concrete bridges. This research work is mainly focused on the periodical fluctuation error in the EFPI strain sensing system.Firstly, the principle and demodulation method of EFPI strain sensor were briefly introduced. Then, a practical EFPI strain sensing was analyzed in detail and the discrete gap transform (DGT) demodulation method was carefully studied. Two possible reasons of above-mentioned periodical error, the disturbance of high order propagation modes and the influence of spectral background noise, were put forward.Secondly, the number of propagation modes and their power distribution were calculated. In the EFPI cavity, the LP01 mode and LP11 mode can be approximated to fundamental (zero order) Gauss beam and first order Hermite-Gauss beam respectively. After passing through the cavity, the phase difference between them can be calculated based on the Gauss approximation. The theoretical analyzed results shows that the'mode issue'is not the reason of above-mentioned periodical error but an important reason of the random error of the sensing system. It's one of the theoretical limitations of the accuracy of our EFPI strain sensing systemThirdly, the influence of spectral background noise was studied. Both theoretically analyzed and computer stimulated results demonstrated that the spectrum noise is the main reason of periodical fluctuation error. An improved DGT algorithm with wavelet denoising and a background subtraction was developed, and the experimental results demonstrated the validity of the improved algorithm.Finally, an implementation example of the improved EFPI strain sensing syste was given. As a subsystem of Shibanpo Yangtze River Bridge Health Monitoring System, the strain sensing system with 108 EFPI strain sensors showed a very good stability and reliability during the passed eight months, and some useful data were collected through it.