Research On Demodulation Technique Of Fiber Bragg Grating Sensor Based On Simulated Annealing Algorithm

FBG (Fiber Bragg Grating) sensor, a kind of functional fiber optic sensor which made of fiber Bragg grating sensitive characteristics, is the focus of optical fiber sensor technique field. FBG sensor is a kind of wavelength modulated sensor, and it has been widely used for the superiorities of resistance to electromagnetic interference, measurement range, corrosion resistance, good stability, easy multiplexing, as well as the advantages of other sensors. Therefore, how to accurately measure the shift of Bragg wavelength becomes the key problem in the research on the FBG sensor.In this thesis, the mechanism of FBG sensor, several typical demodulation method and FBG sensor multiplexing technique, the advantages and disadvantages of which are summarized, are analyzed in details. After research on the principle, advantages and disadvantages of traditional simulated annealing algorithm, a modified algorithm is proposed to improve its behavior. The demodulation method based on F-P tunable filter, using the improved the simulated annealing algorithm and spectral shape multiplexing technique is presented. The theoretical feasibility of the demodulation method is verified by the theory and MATLAB simulation.According to the actual situation and specific simulated annealing algorithm requirement, the right ARM chip (S3C2440A) is selected. The core algorithm and the control software are programmed in C language. The PC interface is programmed in VC++. The communication between PC and lower level computer is completed via serial port (RS-232), and the demodulation results can be received and displayed in superior machine in the same way.The improved simulated annealing algorithm is used on the ARM9 (S3C2440A) to demodulate the simulated signal data. Experimental results show that spectral shape multiplexing technique and the improved simulated annealing algorithm can still demodulate the center wavelengths even when the spectrum of one FBG partly or wholly overlaps with others'. Therefore, the method can save spectral bandwidth and increase the number of multiplexed FBG sensors effectively. Meanwhile, the improved simulated annealing algorithm is faster than the traditional simulated annealing algorithm, and can process the data faster.