Research On Fiber Bragg Grating Wavelength Demodulation Technology Based On DSP

In some specific engineering applications, the existing optical fiber grating demodulation systems are difficult to be used, due to some limitations, such as high cost, large volume, high power consumption and so on. So the key problem about the optical fiber grating demodulation system is how to realize low cost, miniaturization and low power consumption when people use it in practical engineering.This thesis introduces the coupled mode theory and the principles of FBG(Fiber Bragg Gratting) as pressure, temperature and strain sensors, and expounds the advantages and disadvantages of several demodulation technologies. Through the reaserch on the working principle and the characteristics of output signal of the F-P (Fabry-Perot) tunable filter, an optical fiber grating demodulation system based on DSP(Digital Signal Processing)and CPLD(Complex Programmable Logic Device) is put forward. The thesis is divided into two parts, hardware design and software design. In the part of hardware design, ADS8556is selected to achieve a five channel parallel acquisition module for signal acquisition; EPM240T100C8N which is a type of CPLD, and DAC712are chosen to achieve a sawtooth scanning voltage generation module which is used as the F-P filter scan driver, a synchronization pulse signal is generated to drive the D/A converter and A/D converter to work synchronously; IS61LV25616which is a type of SRAM (Static Random Access Memory) is selected to achieve external expansion of DSP to complete datas storage; a DSP minimum system is designed by using TMS320F2812as a core controller to control the data acquisition, data storage, then to complete the data processing. In the software design part, firstly, the A/D acquisition module is programmed to control signal acquisition, then the mapped address space of SRAM is allocated so that data can be stored accurately; CPLD is programmed to generate input signal of D/A converter and sync pulse signal. After data acquisition and data storage, a further data processing is needed to obtain the Bragg wavelength. In this thesis, a Gaussian fitting algorithm is programmed to complete the data processing, but it is found that it is very slow in data processing, when it is used in practical application. Therefore, this thesis, by observing and analysing the signal of the reflected spectra, proposes a new peak detection algorithm which is programmed by C language. In comparison with the former algorithms, it is found that the new algorithm can detect the peaks precisely, at the same time it can also effectively improves data processing speed of the system.