Research On Signal Demodulation Methods Of Optical Fiber EFPI Acoustic Wave Sensor Based On FPGA

The fiber-optic EFPI acoustic sensor based on MEMS thin film has the advantages of compact structure,small volume,passive metal free,high sensitivity and anti electromagnetic interference,so it has great application potential in the fields of high voltage transformer monitoring,industrial abnormal sound warning,military acoustic radar and security interception.At present,for the intensity demodulation scheme of fiber-optic EFPI acoustic sensor,the feedback mechanism is generally used to adjust the wavelength so that the demodulation point is within the linear range,But there are some defects,such as complex feedback regulation,poor real-time demodulation,etc.In addition,the demodulation stability and reliability cannot be guaranteed when dealing with the key problems such as interference spectrum temperature drift and poor probe consistency,which seriously limits the batch practical ability of fiber-optic EFPI acoustic sensor.In this thesis,we mainly focus on the research of fiber-optic EFPI acoustic sensor based on MEMS silicon nitride film.The main work and achievements are as follows:First,we introduce the structure and working mechanism of fiber-optic EFPI acoustic sensor based on MEMS Thin film,focusing on the analysis of the interference spectrum characteristics of fiber-optic MEMS acoustic sensor,and the interference spectrum of the sensor under different cavity length is measured by experiment.The maximum contrast of interference fringes is about 16 dB when the cavity length is about100 um,this optimum initial cavity length is chosen as the cavity length of the sensor.Second,we propose a three wavelength adaptive intensity demodulation algorithm which is proved to be effective in solving the problem of static work point blanking or serious distortion caused by temperature drift and poor consistency of sensor probe.At the same time,it ensures that the worst demodulation sensitivity is not less than 86.6%of the best demodulation sensitivity,which greatly improves the demodulation stability and practicability of the fiber-optic EFPI acoustic sensor.Third,we design and construct a fiber-optic EFPI acoustic sensor system based on three wavelength adaptive intensity demodulation,including the selection and matching of regular parameters and noise analysis of photoelectric devices such as light source,dense wavelength multiplexer and photodetector,design and manufacture low-speedad7606 and high-speed ad9226 ADC data acquisition cards.According to the FPGA core board,the sound card,network card and other peripheral equipment are configured.The integrated assembly structure of the system is optimized,and a complete set of miniaturized,practical and productized optical fiber acoustic detection equipment is formed.Fourth,we design and implement a digital signal demodulation and processing scheme based on Xilinx zynq heterogeneous SOC,including the FPGA implementation of three wavelength adaptive intensity demodulation algorithm and the development of ARM human-computer interaction software.Among them,an oversampling module based on decimation filter is designed to improve the SNR of the system by about 6dB compared with direct sampling;In the realization of FIR digital filter,a serial structure design method based on the idea of time division multiplexing is proposed,which greatly reduces the cost of hardware digital signal processing unit of high-order digital filter;A high-speed data exchange mode based on AXI DMA is adopted in the communication between FPGA and arm.The whole system has complete functions from signal demodulation to human-computer interaction,and the resource consumption and power consumption are greatly optimized.Fifth,according to the experimental data,the demodulation sensitivity of the fiber-optic MEMS acoustic sensor is as high as 830.5mV/Pa(1kHz),the signal-to-noise ratio is 74.57dB(1kHz/80.5dBA),the dynamic range is 0.54mpa-0.635pa(1kHz/THD<3%),and other key parameters are measured,such as anti-interference spectrum temperature drift demodulation experiment and random sampling of sensor probe Demodulation experiment.In the temperature drift demodulation experiment,the worst demodulation effect of the three wavelength adaptive intensity demodulation method in the temperature range of 10?-60? is 85.9% of the best demodulation effect.In the random sampling demodulation experiment of sensor probe,the worst demodulation effect of each probe is 84.68% of the best demodulation effect,which is basically consistent with the theoretical value,further verifying the three wavelength adaptive intensity demodulation method stability and reliability.