Design of Fiber Bragg Grating Demodulation System Based on DSP

0 Introduction

Fiber Bragg grating sensor (FBGS) is a functional fiber sensor using fiber Bragg grating (FBG) as a sensitive element. It can be used to directly detect temperature and strain, as well as indirect measurement of many other physical and chemical quantities related to temperature and strain. In the application research of fiber Bragg grating sensors, wavelength demodulation is an important aspect. At present, the main obstacle restricting the application of fiber Bragg grating sensors is the demodulation of sensing signals. The wavelength demodulation methods mainly include spectrometer, slant edge filtering method, tunable filtering method, interference scanning method, matched grating method, etc. However, among these methods, the cost of spectrometer is relatively high, the resolution of slant edge filtering method is relatively small, the interferometer does not have good repeatability, and the scanning period of tunable filter is relatively long. Therefore, in recent years, the matched grating method has become more and more popular. To this end, this paper introduces a simple, inexpensive system design method for detecting fiber Bragg grating sensors by demodulation of two parallel matched gratings.

1 Dual grating matching principle The schematic diagram of

the dual grating matching system is shown in Figure 1. The light emitted by the broadband light source enters the sensing FBG through a 3 dB coupler. After being reflected by the FBG, it enters two matching gratings. The corresponding two photodetectors obtain optical signals related to their corresponding wavelengths, which are then converted into electrical signals by the photodetectors and enter the signal acquisition and processing circuit to extract useful signals. Finally, the subsequent signal processing system realizes data acquisition and processing.



In Figure 1, PD1 and PD2 are photodetectors, and the optical power P detected by the photodetectors is:



Where I ₁ (λ) and I ₂ (λ) are the reflection power spectrum density functions of the sensing grating and the matching grating, respectively. The reflection power spectrum functions of both can be approximated by Gaussian functions:



Where I ₀ is the peak value of the reflection spectrum intensity; λs is the wavelength value corresponding to the reflection spectrum intensity of I ₀ ; △λs is the 3 dB bandwidth of the reflection spectrum. In general, the size of the optical power detected by the photodetector is proportional to the convolution size of the reflection spectrum of the sensing grating and the matching grating. The smaller the difference between the central wavelength λc of the sensing grating and the central wavelength λp of the matching grating, the larger the corresponding convolution value. When △λ is greater than the threshold △λmin, the convolution value is too small and demodulation may not continue. Therefore, the demodulation range will be limited.

The ordinary matching method has only one sensing grating and one matching grating, corresponding to only one △λ. When △λ≥△λmin, the demodulation system will not be able to continue demodulation. For the dual-grating matching demodulation system, the central wavelengths of the sensing grating and the two parallel matching gratings are approximately equal, but there is a slight difference. The relationship between the three is: λp1<λc<λp2, λp1 and λp2 represent the central wavelengths of the two matching gratings respectively. λc is the central wavelength of the sensing grating. When the sensing grating is under external stress, △λ1=|λc-λp1|, △λ2=|λc-λp2|; when λc increases, △λ1 increases and △λ2 decreases; when λc decreases, △λ1 decreases and △λ2 increases. Figure 2 shows the relationship between △λ1, △λ2 and λc, where △λmin is the minimum value that the photodetector can detect. Therefore, according to Figure 2, in theory, there is always at least one photodetector that can detect the available light signal in the dual grating matching demodulation system.



2 Design of demodulation system based on DSP

2.1 System hardware design

The light reflected by the matching grating is incident on the photodetector (PD) and can be converted into an electrical signal. The photoelectric conversion part and the signal acquisition part mainly complete the acquisition of the PD output electrical signal, and the collected signal is then converted into a digital signal for processing by the DSP. The DSP mainly completes the interpolation operation and peak search of the data, and feeds back the processing results to the DSP. The DSP controls the stepper motor according to the feedback signal to complete the next demodulation work. The system hardware block diagram is shown in Figure 3.

In order to achieve high-precision data sampling, this system uses a 12-bit successive approximation analog/digital conversion chip AD1674 with a parallel microcomputer interface launched by AD Company in the United States to realize the system's analog-to-digital conversion. AD1674 has a built-in sample-and-hold (SHA), a 10V reference voltage source, a clock source, and a temporary storage/tri-state output buffer that can be directly interfaced with the microprocessor bus.

This system uses TMS320VC5402 as the main control chip. This fixed-point DSP chip can realize the processing of fiber grating sensor signals, the control and display of stepper motors, etc. The chip has powerful data calculation and processing functions. Its RPT and MAC instructions can realize multiplication and accumulation operations in a single instruction cycle. Its flexible circular buffer and efficient C language enable TMS320VC5402 to easily realize data circular addressing and convolution operations, thereby achieving high-speed demodulation. 2.2 System Software Design The software part of the DSP system mainly consists of initialization program, linear interpolation subroutine or curve fitting subroutine, display program, driver, interrupt service program and other parts. The A/D conversion and serial communication code can be implemented in the interrupt service program. The initialization program is used to complete the initialization of resources such as DSPI/O port, internal A/D converter, serial port, interrupt, etc. In order to coordinate the control of A/D conversion and stepper motor, the DSP can send a control signal to control the stepper motor so that the digital signal obtained after A/D conversion corresponds to the number of steps added to the matching grating. The program of the display part can convert this digital signal into a digital quantity that directly represents stress through algebraic transformation, and then dynamically realize stress display through table lookup. When the reflection wavelength of the matching grating overlaps with the central wavelength of the fiber grating reflection wave, the photoelectric conversion outputs a pulse signal and requests an interrupt to the DSP. Then the DSP executes the interrupt service program to read the digital quantity converted by the internal A/D converter of the DSP into the DSP and save it, and finally send it to the host computer through the serial port and then interrupt and return. 3 Analysis of experimental results The experiment shows that when the weight mass increases from 0 g to 60 g, the central wavelength of FBG1 attached to the cantilever beam drifts by 0.716 nm. Figure 4 shows the curve of the central wavelength of FBG1 changing with the weight mass. As shown in Figure 4, the change in the central wavelength of FBG on the cantilever beam has a good linear relationship with the weight mass applied to the free end of the cantilever beam, and has high sensitivity. In the experiment, increasing tensile stress can be applied to the cantilever beam by increasing the weight mass. The reflected light signals of the two matching gratings are received by their corresponding photodetectors respectively. The analog voltage signal output by the photodetector is processed by a series of processes of the signal processing system with DSP as the core to obtain the size of the external physical quantity sensed by the sensing fiber Bragg grating. When the voltage signal value of PD1 after processing is 5 V, the corresponding points are A and C, that is, there are two corresponding wavelength values ​​of fiber Bragg gratings. Therefore, for this 5 V voltage, the demodulation system cannot directly determine whether the mass of the corresponding cantilever beam load is the mass corresponding to point A or point C. For the dual grating matching demodulation system, there are often two matching gratings and two corresponding photodetectors. In addition to PD1, there is also PD2. The system can determine the mass of the weight added when generating a 5V voltage through the voltage values ​​corresponding to PD1 and PD2. In fact, the change in the central wavelength of the matching grating can be obtained through the calculation and processing of the DSP system, thereby obtaining the change in the central wavelength of the sensing grating. For the dual grating matching demodulation system, the corresponding sensing grating can take the double sides of the reflection spectrum, thereby expanding the measurement range of the sensing grating. The curve of the voltage change with stress after the output of the photodetector passes through the signal conditioning circuit is shown in Figure 5. 4 Conclusion The dual grating matching demodulation system is a new method based on the matching method and improved. It inherits the advantages of the matching method such as simple structure, low cost, and easy implementation. At the same time, the dual grating matching demodulation system also solves the problem of limited measurement range caused by photodetectors and the dual value problem in the matching method. The demodulation system pastes the matching grating on a special cantilever beam and uses DSP for processing, which not only improves the response speed, but also improves the accuracy and stability of demodulation. The possibility of matching grating damage due to excessive stress is reduced. The demodulation experiment using tension as the measured value of the system proves that the system has good linearity, demodulation accuracy, speed and sensitivity.