Research On Demodulation System And Algorithm Of Optical Fiber Sensor Network

In each field of today's society,intelligence is the general trend of future development.All kinds of Io T products in daily life,unmanned technology of automobile and aircraft,radar system,equipment structure monitoring,safety detection of basic buildings,industrial production automation and its robot system all need more intelligent sensing technology.Therefore,the sensing accuracy and speed of the sensor are required to be higher,and the sensor network with wider coverage and higher coverage density is needed.Taking intelligent robots in industrial production as an example,intelligent robots need to replace humans in extreme environments to carry out a lot of work with high risk,high repetition rate,great difficulty and long time without interruption.This requires a sensor network which can sense multiple physical parameters with high precision under extreme conditions.And based on the technology of optical fiber sensing network system with its small size,high sensitivity,anti-electromagnetic interference,corrosion resistance,electrical insulation,high reliability,easy reuse and unique advantages is widely applied in the similar field of intellisense and aerospace,petrochemical industry,electric power system,the structure detection,security monitoring,industrial production,and other places.According to the development trend of the current intelligent sensing field,optical fiber sensing technology is developing towards the direction of networking,dynamic,array and multi-parameter quantization.In order to improve the detection range,frequency,sensitivity and resolution of the sensor system,it has become the focus of current research to explore the reuse mechanism of the new sensor network and the dynamic demodulation technology of the new generation of large-capacity sensor network.Therefore,this paper starts with the multiplexing technology of optical fiber sensor and the corresponding demodulation system,and combines the current mature optical fiber sensing mechanism and signal processing algorithm technology.Three new signal demodulation methods for quasi-distributed passive fiber grating sensor network distributed active fiber grating sensor network and time-frequency analysis fiber vibration sensor system are designed respectively.The three demodulation systems studied in this paper are as follows:1.In this paper,a spectral demodulation method based on rational cubic B-spline interpolation and Hu invariant moments retrieval(IMR)algorithm for quasi-distributed fiber grating sensor networks is proposed.Based on the proposed method,36 data points of FBG spectra are only need to be sampled for accurate demodulation of four FBG sensor network.The demodulation error of this method is within ± 0.5 pm.The average demodulation time required by the IMR algorithm is about ?0.02 s,which supports real-time measurements at a scanning frequency of about 50 Hz.2.This paper presents a demodulation method of frequency division multiplexing sensor network based on variational mode decomposition(VMD)algorithm and distributed multi-longitudinal mode beat technology.Based on the time-domain characteristics of optical signals,the method uses variational mode decomposition algorithm to decompose the beat signals generated in each sensing cavity from the multiplexed beat signals containing multi-mode components,and preserves the frequency domain information of the beat signals in each cavity well.This method solves the problem of demodulation when the frequency spectrum of beat signals in each cavity is overlapped and increases the number of reusable sensing cavities greatly.3.In this paper,a noise separation and frequency analysis method of optical fiber vibration sensing system based on time-varying filtering empirical mode decomposition(TVF-EMD)algorithm and multi-longitudinal mode beat frequency transmission technology is proposed.The TVF-EMD algorithm was used to separate the vibration information of the original signal with the strong noise mode components.According to the frequency domain characteristics of each mode,the vibration mode and noise mode were identified.Then the time-frequency characteristics of the vibration modes were analyzed by the Hilbert transform(HT).Therefore,this method not only plays a good noise reduction effect,but also can analyze the frequency changes at each time point.