High-precision fiber bragg grating demodulation method and system

Abstract

The invention provides a high-precision fiber grating demodulation method and system, and belongs to the technical field of optical fiber sensing. The method comprises the following steps: firstly, carrying out Kalman filtering denoising to obtain a denoised voltage signal; secondly, establishing an FP etalon wavelength reference sequence; and thirdly, based on the denoised FBG reflection wave voltage signal, accurately positioning the peak voltage and the sampling position of the FBG reflection peak. And finally, based on the FBG reflection peak value voltage and the wavelength-voltage reference sequence, through the processes of adjacent reference point matching and linear interpolation calculation, calculating to obtain the accurate FBG reflection wavelength. The system is realized based on a high-precision fiber grating demodulation system and comprises a signal generation module, an optical module, a signal acquisition module and an upper computer processing module. The method can solve the problems of large nonlinear error, obvious temperature drift interference, weak noise suppression and the like in the traditional demodulation technology, and is suitable for a fiber bragg grating sensing system in the scenes of fluid pressure detection, structural health monitoring and the like.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of optical fiber sensing, and relates to a high-precision fiber grating demodulation method and system, in particular to a fiber grating demodulation method and system based on an FP etalon and Kalman filtering, which can be widely applied to industrial monitoring, environmental monitoring, structural health monitoring and the like, and realizes accurate sensing and demodulation of physical quantities such as fluid pressure, temperature and strain. BACKGROUND

[0002] With the rapid development of optical fiber sensing technology, fiber grating (FBG) has become a core device in the modern sensing field due to its high sensitivity, anti-electromagnetic interference and remote measurement advantages. In the application of fiber grating (FBG), the demodulation technology is the key link to determine the measurement accuracy and reliability, and the core requirement is to realize accurate identification and calculation of the FBG reflection wavelength.

[0003] In the existing demodulation technology, the traditional spectrum analysis method and the intensity modulation method have the problems of low sensitivity and weak anti-interference ability. Although the demodulation technology based on a tunable FP filter can improve the resolution, it is affected by the inherent hysteresis characteristics of the piezoelectric ceramic core element of the FP filter, and is prone to nonlinear response, which leads to wavelength calculation error. At the same time, environmental temperature fluctuations will cause the filter characteristic of the FP filter to drift (temperature drift problem), which seriously damages the stability of the demodulation system. In addition, the noise signal introduced in the photoelectric conversion process will interfere with the extraction of the reflection spectrum characteristics, further reducing the demodulation accuracy.

[0004] To solve the above problems, Chinese invention patent CN103196473B proposes a multi-channel high-precision demodulation device, which dynamically compensates the signal amplitude through an electrically controlled adjustable optical attenuator, thereby improving the signal-to-noise ratio. However, this patent does not design a compensation mechanism for the coupling effect of piezoelectric ceramic hysteresis and temperature drift, and the noise suppression only relies on power adjustment, which does not optimize from the spectrum feature level, and cannot solve the problem of nonlinear error accumulation. In terms of temperature drift interference, Chinese invention patent CN110686708A proposes a variable temperature environment suppression scanning nonlinear optical fiber grating sensing demodulation system, which realizes nonlinear suppression through double reference calibration of F-P etalon and long optical path difference interferometer, which improves the temperature drift influence to a certain extent, but it does not establish a special noise signal suppression module, lacks a solution to the spectral interference introduced in the photoelectric conversion process, and does not involve the cooperative compensation logic of hysteresis and noise.

[0005] In summary, how to provide a high-precision fiber grating demodulation algorithm that can specifically optimize the limitations of mainstream wavelength calibration schemes, can simultaneously solve the problems of piezoelectric ceramic nonlinear response, FP filter temperature drift interference and photoelectric conversion noise suppression, and is suitable for complex dynamic scenes to realize accurate wavelength calculation in a wide spectral range, has also become a problem to be solved. SUMMARY

[0006] In view of the problems of nonlinear error, temperature drift interference, noise influence and poor adaptability of the calibration scheme existing in the existing fiber grating demodulation technology, the application provides a fiber grating demodulation method and system based on an FP standard device and Kalman filtering, which can realize high-precision FBG wavelength calculation and meet the actual application requirements of the fiber grating sensing system.

[0007] In order to achieve the above-mentioned purpose, the technical scheme adopted by the application is:

[0008] A high-precision fiber grating demodulation method, which realizes accurate demodulation of FBG reflected wavelength through three-level processing procedures of noise suppression, wavelength calibration and linear calculation, in combination with the signal acquisition characteristics of the hardware system, and specifically comprises the following steps:

[0009] Firstly, Kalman filtering denoising is performed on the photoelectric conversion voltage signals collected by the analog-to-digital conversion chip (ADC), including the FBG reflected wave voltage signal and the FP standard device reference voltage signal, to obtain the FBG reflected wave voltage signal and the FP standard device reference voltage signal after denoising, and the specific steps are as follows:

[0010] Step 1.1. Separate the FBG reflected wave voltage signal and the FP standard device reference voltage signal from the photoelectric conversion voltage signal collected by the ADC chip as the processing object of Kalman filtering, and provide a clear data basis for subsequent model construction:

[0011] The photoelectric conversion voltage signal collected by the ADC chip is taken as the processing object, which contains two voltage signals, namely the FBG reflected wave voltage signal and the FP standard device reference voltage signal, and both of them are mixed with circuit thermal noise and electromagnetic interference; by the characteristic that the number of wave peaks of the FP standard device reference voltage signal is much more than that of the FBG reflected wave voltage signal, the FBG reflected wave voltage signal and the FP standard device reference voltage signal are separated from the original data collected by the ADC chip, and the FBG reflected wave voltage signal and the FP standard device reference voltage signal are subjected to Kalman filtering denoising treatment respectively, so as to ensure that the model construction is matched with the signal characteristics.

[0012] Step 1.2. For the FBG reflected wave voltage signal and the FP standard device reference voltage signal determined in step 1.1, a dynamic state space model is constructed to provide a mathematical framework for iterative filtering, and the specific steps are as follows:

[0013] Step 1.2.1: For the separated FBG reflected wave voltage signal and FP etalon reference voltage signal, a first-order linear state equation is used to describe the temporal correlation of the voltage signals. At the same time, a Gaussian noise model based on experimental data fitting is introduced to construct the observation equation, forming the state transition equation as shown in formula (1):

[0014] (1)

[0015] Where, x k Let x be the state variables of the FBG reflected wave voltage signal and the FP etalon reference voltage signal at time k; A is the state transition matrix, taken as A=I, where I represents the identity matrix; k-1 The voltage signal of the FBG reflected wave and the reference voltage signal of the FP etalon at time k-1 are the state quantities. The noise is a process noise that follows a Gaussian distribution with a mean of 0 and a variance of 0.01.

[0016] Step 1.2.2: Based on the state variables of the FBG reflected wave voltage signal and the FP etalon reference voltage signal obtained in Step 1.2.1, and the raw voltage data acquired by the ADC chip, the observation equation shown in Equation (2) is constructed by quantizing the noise interference in the acquisition process:

[0017] (2)

[0018] in, The voltage values ​​of the FBG reflected wave voltage signal and the FP etalon reference voltage signal, which are actually acquired by the ADC chip at time k, are the original acquired values ​​of the FBG reflected wave voltage signal and the FP etalon reference voltage signal after separation in step 1.1. H represents the observation matrix, and H=I is taken to ensure that the observed values ​​are directly related to the state variables. The noise observed follows a Gaussian distribution with a mean of 0 and a variance of 0.02.

[0019] Formulas (1) and (2) together constitute a dynamic state-space model. The state transition equation describes the inherent variation law of the FBG reflected wave voltage signal and the FP etalon reference voltage signal. The observation equation is related to the state variables and original acquisition values ​​of the FBG reflected wave voltage signal and the FP etalon reference voltage signal. The combination of the two provides a complete mathematical framework for the prediction-update iteration of Kalman filtering, ensuring that the noise suppression process conforms to the temporal characteristics of the signal and the noise distribution characteristics.

[0020] Step 1.3: Based on the dynamic state-space model constructed in Step 1.2, execute the prediction-update iteration process to achieve noise suppression. Specifically:

[0021] Step 1.3.1: By preliminarily predicting the current state of the FBG reflected wave voltage signal and the FP etalon reference voltage signal through time-series correlation, a basis can be provided for subsequent error correction. Therefore, the optimized state quantities of the FBG reflected wave voltage signal and the FP etalon reference voltage signal at time k-1 are used to estimate the theoretical state values ​​of the FBG reflected wave voltage signal and the FP etalon reference voltage signal at time k. The corresponding state prediction equation is:

[0022] (3)

[0023] in, Let A be the predicted state variable at time k; A = I (same as the state transition matrix in step 1.2). This is the updated state at time k-1.

[0024] Step 1.3.2, by accumulating historical errors and process noise, an error metric can be provided for the calculation of filter gain, resulting in the prediction error covariance equation used to quantify the uncertainty of predicted state variables:

[0025] (4)

[0026] in, Let the prediction error covariance at time k be the value used to describe... The range of deviation from the actual state. A T Let A be the transpose of A; Let Q be the error covariance updated at time k-1; Q represents the process noise covariance matrix, Q=0.01I.

[0027] Step 1.3.3, to determine the optimal correction weights, we can balance the prediction error and the observation error, and obtain the filter gain calculation equation for dynamically adjusting the correction weights of the observations on the predicted state:

[0028] (5)

[0029] in, H is the filter gain at time k; T R is the transpose of H in step 1.2; R represents the observation noise covariance matrix, R=0.02I.

[0030] Step 1.3.4 involves weighting the residuals using the filter gain to achieve accurate correction of the predicted state and eliminate noise interference. This requires combining observed values ​​to correct the predicted state, resulting in the optimized FBG reflected voltage signal and FP etalon reference voltage signal state variables at time k. The corresponding state update equation is:

[0031] (6)

[0032] in, The updated FBG reflected wave voltage signal and FP etalon reference voltage signal state quantities are given at time k. To predict residuals, which is to quantify the deviation between predicted and observed values.

[0033] Step 1.3.5 adjusts the prediction error covariance by filtering the gain to reflect the uncertainty of the updated FBG reflected voltage signal and FP etalon reference voltage signal state quantities, providing an error basis for the prediction at the next time step. Finally, the error covariance update equation for updating the error range of the optimized FBG reflected voltage signal and FP etalon reference voltage signal state quantities at time k is obtained as follows:

[0034] (7)

[0035] in, Let I be the error covariance updated at time k; I is the identity matrix; the other parameters are the same as before.

[0036] Finally, the above iterative process is repeated time-by-time for all the acquired data of the FBG reflected wave voltage signal and the FP etalon reference voltage signal to eliminate noise components and retain the characteristic peak information of the FBG reflected wave voltage signal and the FP etalon reference voltage signal.

[0037] Step 1.4: Based on the iterative filtering results of Step 1.3, output the denoised voltage signal to solve the noise interference problem of the photoelectric conversion signal.

[0038] The iterative optimization results of the FBG reflected wave voltage signal and the FP etalon reference voltage signal are summarized, and the denoised FBG reflected wave voltage signal and FP etalon reference voltage signal are generated respectively. This process can improve the signal-to-noise ratio of the FBG reflected wave voltage signal and the FP etalon reference voltage signal, and provide a high-quality signal foundation for the subsequent establishment of the FP etalon wavelength reference sequence and the localization of the FBG reflection peak.

[0039] The second step is to establish the FP etalon wavelength reference sequence. Based on the denoised FP etalon reference voltage signal output from the first step, a stable wavelength-voltage reference sequence is established to provide a calibration benchmark for subsequent FBG reflection wavelength calculation, as detailed below:

[0040] Step 2.1: Based on the denoised FP etalon reference voltage signal obtained in Step 1, an adaptive threshold peak finding algorithm is used to filter effective transmission peak voltage values, providing a peak data basis for subsequent wavelength calculations.

[0041] Iterate through the denoised FP etalon reference voltage signal data from the first step and calculate the average voltage of the FP etalon reference voltage signal. with standard deviation The overall distribution characteristics of the FP etalon reference voltage signal are quantified; a threshold is set based on the 3σ criterion to effectively eliminate the interference of residual small noise points on peak identification. The FP etalon reference voltage signal data is traversed point by point. If the voltage value corresponding to a certain sampling point is greater than the voltage values ​​corresponding to the previous and next sampling points, then that sampling point is a local maximum. All local maxima points greater than a threshold T are selected, and the formula for calculating the threshold T is:

[0042] (8)

[0043] in, and These are the mean and standard deviation of the FP etalon reference voltage signal, respectively.

[0044] During the screening process, the sampling position corresponding to each local maximum point is recorded simultaneously. With voltage value The local maxima obtained through the above steps are the effective transmission peaks. Finally, the sampling positions of all effective transmission peaks are output. With corresponding voltage value .

[0045] Step 2.2: Based on the effective transmission peak data obtained in Step 2.1 and the known parameters of the FP etalon, calculate the wavelength interval between adjacent transmission peaks to provide a quantitative basis for wavelength calibration. The known parameters of the FP etalon are the free spectral range and the center wavelength of the FP etalon; specifically:

[0046] The effective transmission peak sampling position output in step 2.1 Voltage value Using the core data and the known parameters of the FP standard etalon, wavelength spacing is calculated. The calculation expression is as follows:

[0047] (9)

[0048] Where λ is the center wavelength of the FP etalon; c is the speed of light; and FSR is the free spectral range.

[0049] Substituting the known parameters of the FP etalon, we get This value represents the wavelength interval between adjacent effective transmission peaks. The final output shows the wavelength interval between adjacent effective transmission peaks. .

[0050] Step 2.3: Based on the effective transmission peak voltage value obtained in Step 2.1 and the wavelength interval obtained in Step 2.2, construct the wavelength-voltage reference sequence to complete the construction of the FP etalon wavelength reference system. Specifically:

[0051] From the effective transmission peak voltage values ​​in step 2.1, select the value that corresponds to the center wavelength of the FP etalon. The corresponding transmission peak voltage value is used as the reference peak voltage value, and its voltage value is With wavelength Using the reference peak as a reference, the wavelength of each effective transmission peak is calculated sequentially for the effective transmission peaks on both sides of the reference peak. The calculation expression is:

[0052] (10)

[0053] in, This is the current transmission peak number. The reference peak number;

[0054] The wavelength of each effective transmission peak Its corresponding voltage value One-to-one correspondence, forming a wavelength-voltage reference sequence. , where n is the total number of effective transmission peaks.

[0055] The final output is a complete wavelength-voltage reference sequence, providing a stable wavelength calibration benchmark for the linear calculation of the FBG reflection peak wavelength in the fourth step.

[0056] The third step is precise location of the FBG reflection peak. Based on the denoised FBG reflection wave voltage signal obtained in the first step, the peak voltage and sampling position of the FBG reflection peak are precisely located, providing core characteristic parameters for subsequent wavelength calculation, as detailed below:

[0057] Step 3.1: Based on the denoised FBG reflected wave voltage signal obtained in the first step, the sampling interval is divided and a cubic spline fitting function is constructed to achieve a continuous description of the voltage signal.

[0058] Using the denoised FBG reflected wave voltage signal output from the first step as the processing object, the signal is divided into several intervals, with each interval consisting of 5 consecutive sampling points. A cubic spline function is constructed for each interval, satisfying the conditions of voltage continuity, first derivative continuity, and second derivative continuity at the interval endpoints. These functions are then concatenated to form a fitting function covering the entire voltage spectrum. The calculation expression is:

[0059] (11)

[0060] (12)

[0061] (13)

[0062] (14)

[0063] (15)

[0064] in, This represents the fitted function. j is the interval index. , , , This represents the fitting coefficient. , Represents the x and y coordinates of points within the interval. for exist The first right derivative at; for exist The first left derivative at; for exist The second right derivative at point; for exist The second left derivative at point;

[0065] Formulas (12) and (13) are the continuity conditions for the endpoints of the interval, formula (14) is the continuity condition for the first derivative, and formula (15) is the continuity condition for the second derivative.

[0066] Step 3.2: Based on the fitting function constructed in Step 3.1, the least squares method is used to optimize the accuracy, ensuring that the fitted curve closely matches the original signal. Specifically:

[0067] With step 3.1 To optimize the object, the objective function is constructed as follows:

[0068] (16)

[0069] Where E represents the objective function, used to quantify the deviation between the model output and the target value; m represents the number of sampling points, taken as m = 5. express The corresponding fitted value; express The corresponding voltage value.

[0070] Iterative adjustment of fitting coefficients , , , To make E ≤ 0.005V, output the optimized fitting function. .

[0071] Step 3.3: Based on the optimized fitting function in Step 3.2, extract the peak sampling position and voltage value of the FBG reflection peak. Specifically:

[0072] right Find the first derivative, and calculate its expression as follows:

[0073] (17)

[0074] make The peak sampling position is obtained by solving. ;Will Substitution To obtain the peak voltage Output and This provides characteristic parameters for the fourth step of wavelength linear solution.

[0075] The fourth step is linear interpolation calculation of the FBG reflection wavelength. Based on the peak-to-peak voltage of the FBG reflection output in the third step and the wavelength-voltage reference sequence established in the second step, the accurate FBG reflection wavelength is calculated through a process of adjacent reference point matching and linear interpolation calculation, as follows:

[0076] Step 4.1, based on the peak-to-peak voltage of the FBG reflection from step 3. Compared with the wavelength-voltage reference sequence in the second step, adjacent reference points are selected, specifically:

[0077] Traverse the reference sequence constructed in the second step Find with Two adjacent reference voltages , and corresponding wavelength , ,in ;

[0078] Step 4.2, based on the adjacent reference points from step 4.1 and the third step... The wavelength is calculated using linear interpolation, and the corresponding calculation expression is:

[0079] (18)

[0080] in, This indicates the FBG reflection wavelength that needs to be demodulated. Indicates and Adjacent and in The reference voltage value on the left. Indicates and Adjacent and in The reference voltage value on the right. , They represent , The corresponding reference wavelength.

[0081] The FBG reflection wavelength was calculated. Demodulation is complete.

[0082] A high-precision fiber Bragg grating demodulation system implements the core steps of the aforementioned fiber Bragg grating demodulation method, namely Kalman filtering noise reduction, FP etalon wavelength reference sequence establishment, FBG reflection peak accurate positioning, and FBG reflection wavelength linear interpolation calculation, by dividing the system into modules and coordinating their execution. The high-precision fiber Bragg grating demodulation system implements the aforementioned fiber Bragg grating demodulation method and includes a signal generation module, an optical module, a signal acquisition module, and a host computer processing module; specifically:

[0083] The signal generation module includes a signal generator. The signal generator produces a sawtooth wave voltage signal that acts on a tunable FP filter.

[0084] The optical module includes an ASE broadband light source, a tunable FP filter, an FP etalon, a 1×2 coupler, and a circulator. The ASE broadband light source, after being modulated by the tunable FP filter, is split into two paths by the 1×2 coupler: a sensing light signal and a reference light signal. The sensing light signal is injected into the FBG sensing array via the circulator and then connected to the signal acquisition module. The reference light signal is input to the FP etalon and then connected to the signal acquisition module.

[0085] The signal acquisition module includes two photodetectors, an analog-to-digital converter (ADC), and an STM32 microcontroller. Specifically: the first photodetector is connected to the FBG sensor, converting the FBG reflected light signal into an FBG reflected voltage signal. The second photodetector is connected to the FP etalon, converting the FP etalon reference light signal into an FP etalon reference voltage signal; the ADC converts the continuous FBG reflected voltage signal and the FP etalon reference voltage signal into discrete digital signals; and the STM32 microcontroller sends the digital signals converted by the ADC to the host computer processing module.

[0086] The host computer processing module, serving as the core execution carrier of the aforementioned fiber Bragg grating demodulation method, receives data transmitted from the signal acquisition module and performs the entire data processing flow in the sequence of Kalman filtering denoising, wavelength reference sequence establishment, reflection peak localization, and wavelength linear interpolation calculation. Furthermore, the host computer processing module includes a data storage unit and a visualization unit. The data storage unit saves the denoised signal, fitted curve data, and demodulation results in CSV format. It supports historical data backtracking; the visualization unit dynamically displays the voltage signal curve, fitting curve and demodulation wavelength in real time, intuitively presenting the processing process and results, which is convenient for operators to monitor and debug.

[0087] The beneficial effects of this invention are as follows:

[0088] This invention establishes a stable wavelength reference sequence using the FP etalon and combines it with a linear interpolation algorithm to effectively eliminate the nonlinearity and temperature drift error of the tunable FP filter, achieving high-precision demodulation with performance superior to the traditional acetylene chamber method and reference grating method. Furthermore, it uses a Kalman filter algorithm to dynamically optimize filter parameters based on the photoelectric conversion noise characteristics, improving the signal-to-noise ratio and ensuring peak identification accuracy in weak signal scenarios. Simultaneously, relying on the temperature drift resistance of the FP etalon and the dynamic noise suppression capability of the Kalman filter, the algorithm can adapt to complex industrial environments without requiring additional nonlinear calibration or temperature compensation circuits. Attached Figure Description

[0089] Figure 1 This is an overall flowchart of the demodulation method of the present invention.

[0090] Figure 2 This is a comparison of the voltage waveforms before and after Kalman filtering.

[0091] Figure 3 This is a graph showing the wavelength correspondence between the transmission spectrum of the FP etalon and the reflection spectrum of the FBG.

[0092] Figure 4 This describes the interrelationships between the modules in the demodulation system of this invention.

[0093] Figure 5 A schematic diagram of the hardware system architecture adapted to the algorithm. Detailed Implementation

[0094] The present invention will be further described below with reference to specific implementation examples.

[0095] A high-precision fiber Bragg grating demodulation method, comprising the following steps, utilizes a three-stage processing flow of noise suppression, wavelength calibration, and linear calculation, combined with the signal acquisition characteristics of the hardware system, to achieve accurate demodulation of the FBG reflection wavelength:

[0096] The first step is Kalman filtering for noise reduction. Based on the photoelectric conversion voltage signal acquired by a 16-bit ADC at a sampling rate of 200KSPS, specifically including 1000 data points each for the FBG reflected wave voltage signal and the FP etalon reference voltage signal, Kalman filtering is performed on both signals to obtain the denoised FBG reflected wave voltage signal and FP etalon reference voltage signal, as detailed below:

[0097] Step 1.1: Separate the FBG reflected wave voltage signal and the FP etalon reference voltage signal from the photoelectric conversion voltage signal acquired by the ADC chip as the processing objects for Kalman filtering, providing a clear data foundation for subsequent model construction:

[0098] The photoelectric conversion voltage signal acquired by the ADC chip is used as the processing object. This signal contains two voltage signals: 1000 data points each for the FBG reflected wave voltage signal and the FP etalon reference voltage signal. The amplitudes of both the FBG reflected wave voltage signal and the FP etalon reference voltage are between 0 and 3.3V. The FP etalon reference voltage signal has approximately 20 peaks, while the FBG reflected wave voltage signal has only one peak. Based on this characteristic, the FBG reflected wave voltage signal and the FP etalon reference voltage signal are separated from the raw data acquired by the ADC chip. Kalman filtering is then performed on the FBG reflected wave voltage signal and the FP etalon reference voltage signal to denoise them, ensuring that the model construction matches the signal characteristics.

[0099] Step 1.2: Based on the FBG reflected wave voltage signal and the FP etalon reference voltage signal determined in Step 1.1, a dynamic state-space model is constructed to provide a mathematical framework for iterative filtering. Specifically:

[0100] Step 1.2.1: For the separated FBG reflected wave voltage signal and FP etalon reference voltage signal, a first-order linear state equation is used to describe the temporal correlation of the voltage signals. At the same time, Gaussian noise with a mean of 0 and a variance of 0.01 is introduced. The state transition matrix A is taken as the second-order identity matrix I to construct the observation equation, forming the state transition equation as shown in formula (19):

[0101] (19)

[0102] in, , Let be the voltage state quantity of the FBG reflected wave at time k. Let be the reference voltage state quantity of the FP standard etalon at time k; N represents a Gaussian distribution.

[0103] Step 1.2.2: For the state variables of the FBG reflected wave voltage signal and the FP etalon reference voltage signal obtained in Step 1.2.1, and the original voltage data collected by the ADC chip, the noise interference in the acquisition process is quantized to Gaussian noise with a mean of 0 and a variance of 0.02. The observation matrix H is taken as a two-dimensional identity matrix I to ensure that the observed values ​​are directly related to the state variables, and the observation equation as shown in formula (20) is constructed:

[0104] (20)

[0105] in, , This represents the voltage value of the FBG reflected wave voltage signal actually acquired by the ADC chip at time k. The voltage value of the FP etalon reference voltage signal is the original acquisition value of the FBG reflected wave voltage signal and the FP etalon reference voltage signal after separation in step 1.1.

[0106] Formulas (19) and (20) together constitute a dynamic state-space model, in which the state transition equation describes the inherent change law of the FBG reflected wave voltage signal and the FP etalon reference voltage signal, and the observation equation is related to the state variables and original acquisition values ​​of the FBG reflected wave voltage signal and the FP etalon reference voltage signal. The combination of the two provides a complete mathematical framework for the prediction-update iteration of Kalman filtering, ensuring that the noise suppression process conforms to the temporal characteristics of the signal and the noise distribution characteristics.

[0107] Step 1.3: Based on the dynamic state-space model constructed in Step 1.2, execute the prediction-update iterative process to achieve noise suppression. Specifically, taking the state variables of the FBG reflected wave voltage signal and the FP etalon reference voltage signal at time 199 as an example to predict and update the state variables of the FBG reflected wave voltage signal and the FP etalon reference voltage signal at time 200:

[0108] Step 1.3.1: By preliminarily predicting the current state of the FBG reflected wave voltage signal and the FP etalon reference voltage signal through time-series correlation, a basis can be provided for subsequent error correction. Therefore, the optimized state quantities of the FBG reflected wave voltage signal and the FP etalon reference voltage signal at time 199 are adopted. Estimate the theoretical state values ​​of the FBG reflected wave voltage signal and the FP etalon reference voltage signal at time 200. Based on the dynamic state-space model determined in step 1.2 and the experimentally obtained FBG reflected wave voltage signal and FP etalon reference voltage signal data, it can be concluded that... , combined Calculated according to formula (3) .

[0109] Step 1.3.2, by accumulating historical errors and process noise, an error metric can be provided for the filter gain calculation. The variance between the estimated and true values ​​of the FBG reflected wave voltage signal at point 199, measured experimentally, is 0.008, and the variance between the estimated and true values ​​of the FP etalon reference voltage signal is 0.006. Therefore, the error covariance updated based on time 199 is used to calculate the error. Combining Process noise covariance According to formula (4), the prediction error covariance used to quantify the uncertainty of the predicted state variables is obtained. .

[0110] Step 1.3.3: To determine the optimal correction weights, the filter gain used to dynamically adjust the correction weights of the observed values ​​on the predicted state is obtained by balancing the prediction error and the observation error. Based on the prediction error covariance obtained in step 1.3.2 , combined Observation noise covariance The filter gain at time 200 is calculated using formula (5). .

[0111] Step 1.3.4 involves weighting the residuals using the filter gain to achieve accurate correction of the predicted state and eliminate noise interference. This requires combining observed values ​​to correct the predicted state, resulting in the optimized FBG reflected wave voltage signal and FP etalon reference voltage signal state quantities at time 200. The FBG reflected wave voltage signal acquired by the ADC at time 200, combined with the FP etalon reference voltage signal data obtained in the experiment, was analyzed. Calculate the predicted residual Then through Weighted correction, the optimized state variables at time 200 are calculated according to formula (6). .

[0112] Step 1.3.5 adjusts the prediction error covariance by adjusting the filter gain to reflect the uncertainty of the updated FBG reflected voltage signal and FP etalon reference voltage signal state quantities, providing an error basis for the prediction at the next time step. Finally, the error covariance of the updated state quantities at time 200 after optimization is obtained. Based on the results obtained in step 1.3.3 The result obtained in step 1.3.2 Calculated using formula (7) .

[0113] Finally, the above iterative process is repeated time-by-time for 1000 data points of the FBG reflected wave voltage signal and the FP etalon reference voltage signal to eliminate noise components and retain the characteristic peak information of the FBG reflected wave voltage signal and the FP etalon reference voltage signal.

[0114] Step 1.4: Based on the iterative filtering results of Step 1.3, output the denoised voltage signal to solve the noise interference problem of the photoelectric conversion signal.

[0115] The above iterative process is repeated time-by-time for 1000 sampling points each of the FBG reflected wave voltage signal and the FP etalon reference voltage signal, ultimately yielding the denoised FBG reflected wave voltage signal and the FP etalon reference voltage signal. Figure 2Comparison of voltage waveforms before and after Kalman filtering. Figure 2 The blue dashed line represents the original FBG voltage waveform, and the red solid line represents the waveform after noise reduction, providing a visual verification of the filtering effect.

[0116] The second step is to establish the FP etalon wavelength reference sequence. Based on the denoised FP etalon reference voltage signal output from the first step, a stable wavelength-voltage reference sequence is established to provide a calibration benchmark for subsequent FBG reflection wavelength calculation, as detailed below:

[0117] Step 2.1: Based on the 1000-point voltage data of the denoised FP etalon reference voltage signal obtained in Step 1, an adaptive threshold peak-finding algorithm is used to filter the effective transmission peak voltage values, providing a peak data basis for subsequent wavelength calculations.

[0118] Iterate through the 1000-point FP etalon reference voltage signal data after the first step of noise reduction, and calculate the voltage mean of the FP etalon reference voltage signal. with standard deviation The quantization of the FP standard reference voltage signal's overall distribution characteristics; based on 3 The criteria set a threshold to effectively eliminate the interference of residual minute noise points on peak identification. The threshold is calculated according to formula (8). By iterating through 1000 points of FP etalon reference voltage signal data, if the voltage value corresponding to a certain sampling point is greater than the voltage values ​​corresponding to the previous and next sampling points, then that sampling point is a local maximum point. After iterating through 1000 points of FP etalon reference voltage signal data according to the above steps, 22 local maximum points were selected.

[0119] During the screening process, the sampling position corresponding to each local maximum point is recorded simultaneously. With voltage value The local maxima obtained through the above steps are the effective transmission peaks. Finally, the sampling positions of all effective transmission peaks are output. With corresponding voltage value .

[0120] Step 2.2: Based on the effective transmission peak data obtained in Step 2.1 and the known parameters of the FP etalon, calculate the wavelength interval between adjacent transmission peaks to provide a quantitative basis for wavelength calibration. The known parameters of the FP etalon are the free spectral range and the center wavelength of the FP etalon; specifically:

[0121] The effective transmission peak sampling position output in step 2.1 Voltage value Using the core data and combined with the known parameters of the FP etalon, the center wavelength of the FP etalon is... Free spectral range speed of light The wavelength interval is calculated according to formula (9). .

[0122] Step 2.3, based on the 22 effective transmission peak voltage values ​​obtained in Step 2.1 and the wavelength interval obtained in Step 2.2. Construct a wavelength-voltage reference sequence to complete the construction of the FP standard etalon wavelength reference system. Specifically:

[0123] From the 22 effective transmission peak voltage values ​​in step 2.1, select the center wavelength of the FP etalon. The corresponding transmission peak voltage value is used as the reference peak voltage value, and its voltage value is With wavelength Using the reference peak as a reference, the wavelength of each effective transmission peak is calculated sequentially for the effective transmission peaks on both sides of the reference peak. Calculated according to formula (10) ,wait, .

[0124] The wavelength of each effective transmission peak Its corresponding voltage value One-to-one correspondence, forming a wavelength-voltage reference sequence. .

[0125] The final output is a complete wavelength-voltage reference sequence, providing a stable wavelength calibration benchmark for the linear calculation of the FBG reflection peak wavelength in the fourth step.

[0126] The third step is precise location of the FBG reflection peak. Based on the denoised FBG reflection wave voltage signal obtained in the first step, the peak voltage and sampling position of the FBG reflection peak are precisely located, providing core characteristic parameters for subsequent wavelength calculation, as detailed below:

[0127] Step 3.1: Based on the denoised FBG reflected wave voltage signal obtained in the first step, the sampling interval is divided and a cubic spline fitting function is constructed to achieve a continuous description of the voltage signal.

[0128] Using 1000 data points of the denoised FBG reflected wave voltage signal output from the first step as the processing object, the data is divided into 200 intervals, with each interval consisting of 5 consecutive sampling points. A cubic spline function is constructed for each interval, satisfying the conditions of voltage continuity, first derivative continuity, and second derivative continuity at the interval endpoints. These functions are then concatenated to form a fitting function covering the entire voltage spectrum. The results were obtained from formulas (11) to (15). :

[0129] Step 3.2: Based on the fitting function constructed in Step 3.1, the least squares method is used to optimize the accuracy, ensuring that the fitted curve closely matches the original signal. Specifically:

[0130] With step 3.1 To optimize the object, the objective function is constructed according to formula (16), and the fitting coefficients are iteratively adjusted. , , , To make E ≤ 0.005V, output the optimized fitting function.

[0131] Step 3.3: Based on the optimized fitting function in Step 3.2, extract the peak sampling position and voltage value of the FBG reflection peak. Specifically:

[0132] right Find the first derivative, which is calculated according to formula (17). .

[0133] make The peak sampling position is obtained by solving. ;Will Substitution To obtain the peak voltage Output and This provides characteristic parameters for the fourth step of wavelength linear solution.

[0134] The fourth step is linear interpolation calculation of the FBG reflection wavelength. Based on the peak-to-peak voltage of the FBG reflection output in the third step and the wavelength-voltage reference sequence established in the second step, the accurate FBG reflection wavelength is calculated through a process of adjacent reference point matching and linear interpolation calculation, as follows:

[0135] Step 4.1, based on the peak-to-peak voltage of the FBG reflection from step 3. Compared with the wavelength-voltage reference sequence in the second step, adjacent reference points are selected, specifically:

[0136] Traverse the reference sequence constructed in the second step Find with Two adjacent reference voltages , and corresponding wavelength , ;

[0137] Step 4.2, based on the adjacent reference points from step 4.1 and the third step... The wavelength is calculated by linear interpolation and obtained according to formula (18). ;

[0138] The FBG reflection wavelength was calculated. Demodulation is completed; the entire process of the high-precision fiber Bragg grating demodulation method is as follows: Figure 1 As shown.

[0139] This embodiment also provides a high-precision fiber Bragg grating demodulation system, which modularly carries and collaboratively executes the core steps of the above-mentioned fiber Bragg grating demodulation method, including Kalman filtering noise reduction, FP etalon wavelength reference sequence establishment, FBG reflection peak accurate positioning, and FBG reflection wavelength linear interpolation calculation. The high-precision fiber Bragg grating demodulation system implements the above-mentioned fiber Bragg grating demodulation method and includes a signal generation module, an optical module, a signal acquisition module, and a host computer processing module; specifically:

[0140] The signal generation module mentioned above: such as Figure 4 and Figure 5 This includes a signal generator. The signal generator produces a sawtooth wave voltage signal that acts on a tunable FP filter.

[0141] The optical module mentioned above: such as Figure 4 and Figure 5 The system includes an ASE broadband light source, a tunable FP filter, an FP etalon, a 1×2 coupler, and a circulator. The ASE broadband light source, after being modulated by the tunable FP filter, is split into two paths by the 1×2 coupler: a sensing light signal and a reference light signal. The sensing light signal is injected into the FBG sensing array via the circulator and then connected to the signal acquisition module. The reference light signal is input to the FP etalon and then connected to the signal acquisition module.

[0142] The signal acquisition module mentioned above: such as Figure 4 and Figure 5 It includes two photodetectors, an analog-to-digital converter (ADC), and an STM32 microcontroller. Specifically: the first photodetector is connected to the FBG sensor, converting the FBG reflected light signal into an FBG reflected voltage signal. The second photodetector is connected to the FP etalon, converting the FP etalon reference light signal into an FP etalon reference voltage signal; the ADC converts the continuous FBG reflected voltage signal and the FP etalon reference voltage signal into discrete digital signals; the STM32 microcontroller sends the digital signals converted by the ADC to the host computer processing module.

[0143] The host computer processing module: such as Figure 4 and Figure 5 As the core execution carrier of the aforementioned fiber Bragg grating demodulation method, it receives data transmitted from the signal acquisition module and performs the entire data processing flow in the sequence of Kalman filtering denoising, wavelength reference sequence establishment, reflection peak localization, and wavelength linear interpolation calculation. Furthermore, the host computer processing module includes a data storage unit and a visualization unit. The data storage unit saves the denoised signal, fitted curve data, and demodulation results in CSV format. It supports historical data backtracking; the visualization unit dynamically displays the voltage signal curve, fitting curve and demodulation wavelength in real time, intuitively presenting the processing process and results, which is convenient for operators to monitor and debug.

[0144] The above embodiments are merely illustrative of the implementation methods of the present invention, but should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the protection scope of the present invention.

Claims

1. A high-precision fiber grating demodulation method, characterized in that, The high-precision fiber grating demodulation method realizes accurate demodulation of FBG reflected wavelength through three-level processing procedures of noise suppression, wavelength calibration and linear calculation, and specifically includes the following steps. First step, Kalman filter denoising; The collected photoelectric conversion voltage signal includes FBG reflected wave voltage signal and FP etalon reference voltage signal, and the FBG reflected wave voltage signal and the FP etalon reference voltage signal are subjected to Kalman filter denoising processing respectively to obtain the FBG reflected wave voltage signal and the FP etalon reference voltage signal after denoising processing; Specifically: Step 1.1, separate the FBG reflected wave voltage signal and the FP etalon reference voltage signal from the photoelectric conversion voltage signal collected from the ADC chip; Step 1.2, for the FBG reflected wave voltage signal and the FP etalon reference voltage signal determined in step 1.1, a dynamic state space model is constructed; Step 1.3, based on the dynamic state space model constructed in step 1.2, a prediction-update iteration process is performed to realize noise suppression; Step 1.4, based on the iteration filtering result of step 1.3, the iteration optimization results of the FBG reflected wave voltage signal and the FP etalon reference voltage signal are summarized to generate the FBG reflected wave voltage signal and the FP etalon reference voltage signal after denoising respectively; Second step, establish the FP etalon wavelength reference sequence; Based on the denoised FP etalon reference voltage signal output in the first step, a stable wavelength-voltage reference sequence is established to provide a calibration benchmark for subsequent FBG reflected wavelength calculation; Specifically: Step 2.1, based on the denoised FP etalon reference voltage signal obtained in the first step, an adaptive threshold peak searching algorithm is used to screen effective transmission peak voltage values; Step 2.2, based on the effective transmission peak data obtained in step 2.1 and the known parameters of the FP etalon, the wavelength interval of adjacent transmission peaks is calculated to provide a quantitative basis for wavelength calibration; Step 2.3, based on the effective transmission peak voltage values obtained in step 2.1 and the wavelength interval obtained in step 2.2, a wavelength-voltage reference sequence is constructed to complete the establishment of the FP etalon wavelength reference system; Third step, accurate positioning of FBG reflection peak; Based on the denoised FBG reflected wave voltage signal obtained in the first step, the peak voltage and sampling position of the FBG reflection peak are accurately positioned to provide core feature parameters for subsequent wavelength calculation; Step 3.1, based on the denoised FBG reflected wave voltage signal obtained in the first step, the sampling interval is divided and a cubic spline fitting function is constructed to realize continuous description of the voltage signal: Step 3.2, based on the fitting function constructed in step 3.1, the least squares method is used to optimize the accuracy to ensure that the fitting curve fits the original signal; Step 3.3, based on the optimized fitting function in step 3.2, the peak sampling position and voltage value of the FBG reflection peak are extracted; Fourth step, FBG reflected wavelength linear interpolation calculation; Based on the FBG reflection peak voltage output in the third step and the wavelength-voltage reference sequence established in the second step, the accurate FBG reflected wavelength is calculated through the procedures of adjacent reference point matching and linear interpolation calculation.

2. The high-precision fiber grating demodulation method according to claim 1, characterized in that, The step 1.1 specifically is: The photoelectric conversion voltage signal collected by the ADC chip is taken as the processing object, and the signal contains two voltage signals, namely the FBG reflected wave voltage signal and the FP etalon reference voltage signal; By the characteristic that the number of wave peaks of the FP etalon reference voltage signal is much more than that of the FBG reflected wave voltage signal, the FBG reflected wave voltage signal and the FP etalon reference voltage signal are separated from the original data collected by the ADC chip; The FBG reflected wave voltage signal and the FP etalon reference voltage signal are respectively subjected to Kalman filtering denoising processing, so as to ensure that the model construction matches the signal characteristics.

3. The high-precision fiber grating demodulation method according to claim 2, characterized in that, The step 1.2 is specifically: Step 1.2.1, for the separated FBG reflected wave voltage signal and FP etalon reference voltage signal, a first-order linear state equation is used to describe the time sequence correlation of the voltage signal, and a Gaussian noise model based on experimental data fitting is introduced to construct an observation equation, so as to form a state transition equation as shown in formula (1): （1）； wherein x k is the state quantity of the FBG reflected wave voltage signal and the FP etalon reference voltage signal at the kth moment; A is a state transition matrix, and A = I is taken, wherein I represents a unit matrix; x k-1 is the state quantity of the FBG reflected wave voltage signal and the FP etalon reference voltage signal at the (k-1)th moment; is the process noise; Step 1.2.2, for the state quantity of the FBG reflected wave voltage signal and the FP etalon reference voltage signal obtained in step 1.2.1, and the original voltage data collected by the ADC chip, the noise interference in the quantization collection process is constructed to form an observation equation as shown in formula (2): （2）； wherein, is the voltage value of the FBG reflection wave voltage signal actually collected by the ADC chip at the kth moment and the voltage value of the FP standard reference voltage signal; H represents an observation matrix, and H = I is taken to ensure that the observation value is directly related to the state quantity; is the observation noise; Formulas (1) and (2) jointly constitute a dynamic state space model, and the combination of the two provides a complete framework for the prediction-update iteration of Kalman filtering.

4. The high-precision fiber grating demodulation method according to claim 3, characterized in that, The step 1.3 is specifically: Step 1.3.1, the current FBG reflected wave voltage signal and FP etalon reference voltage signal state are preliminarily predicted through time sequence correlation, and the state quantity of the FBG reflected wave voltage signal and the FP etalon reference voltage signal after optimization at the k-1 time is used to estimate the theoretical state value of the FBG reflected wave voltage signal and the FP etalon reference voltage signal at the k time, and the corresponding state prediction equation is: （3）； wherein, is the predicted state quantity at the kth time instant; A = I, the state transition matrix of step 1.2; is the updated state quantity at the k-1th time instant; Step 1.3.2, the accumulated historical error and the process noise are used to provide an error measurement basis for the calculation of the filtering gain, and a prediction error covariance equation for quantifying the uncertainty of the predicted state quantity is obtained: （4）； wherein, is the prediction error covariance at the kth time instant, describing the deviation range from the true state; A T is the transpose matrix of A; is the updated error covariance at the k-1th time instant; Q represents the process noise covariance matrix, Q = 0.01I; Step 1.3.3, in order to determine the optimal correction weight, the filtering gain calculation equation for dynamically adjusting the correction weight of the observation value to the predicted state is obtained: （5）； wherein, Hkis the filter gain at the kth time instant; H T R is the transpose matrix of H in step 1.2; R represents the observation noise covariance matrix, R = 0.02I; Step 1.3.4, the residual is weighted through the filtering gain, and the predicted state is accurately corrected; the observation value is combined to correct the predicted state, and the state quantity of the FBG reflected wave voltage signal and the FP etalon reference voltage signal after optimization at the k time is obtained, and the corresponding state update equation is: （6）； wherein, is the FBG reflected wave voltage signal at the kth time instant and the FP etalon reference voltage signal state quantity, is the prediction residual, i.e. the deviation of the quantized predictor from the observed value; Step 1.3.5, the prediction error covariance is adjusted through the filtering gain, the uncertainty of the updated FBG reflected wave voltage signal and FP etalon reference voltage signal state quantity is reflected, an error basis is provided for the next time prediction, and an error covariance update equation for updating the error range of the FBG reflected wave voltage signal and FP etalon reference voltage signal state quantity after optimization at the k time is obtained: （7）； wherein, is the updated error covariance at the kth time instant; I is the identity matrix; the remaining parameters are the same as before. Finally, the iteration process is repeated for all collected data of the FBG reflected wave voltage signal and the FP etalon reference voltage signal at each time, and the noise components are eliminated while the characteristic peak information of the FBG reflected wave voltage signal and the FP etalon reference voltage signal is retained.

5. The high-precision fiber grating demodulation method according to claim 4, characterized in that, The second step is: The step 2.1 is specifically: Traverse the first step denoising after the FP standard reference voltage signal data, calculate the voltage mean value of the FP standard reference voltage signal With standard deviation , Quantify the overall distribution characteristics of the FP standard reference voltage signal; The threshold is set based on the 3σ criterion; The FP standard reference voltage signal data is traversed sample by sample, if the voltage value corresponding to a sample point is greater than the voltage values corresponding to the previous sample point and the next sample point of the sample point, then the sample point is a local maximum point; all local maximum points greater than the threshold T are screened out; the calculation formula of the threshold T is: （8）； wherein with are the voltage mean and standard deviation of the FP etalon reference voltage signal, respectively; The sampling position corresponding to each local maximum point is recorded synchronously during the screening process The voltage value The local maximum point obtained through the above steps is the effective transmission peak, and the sampling position of all effective transmission peaks is finally output The corresponding voltage value ; The step 2.2 is specifically: Effective transmission peak sampling position output by step 2.1 , voltage value The core data, combined with the known parameters of the FP etalon, performs wavelength interval calculation, and the calculation expression is: （9）； Wherein, λ is the center wavelength of the FP standard; c is the speed of light; FSR is the free spectral range; Substitute the known parameters of the FP etalon , the value is the wavelength interval of adjacent effective transmission peaks; finally output the wavelength interval of adjacent effective transmission peaks ; The step 2.3 is specifically: From the effective transmission peak voltage value of step 2.1, select the effective transmission peak voltage value corresponding to the FP etalon center wavelength as the reference peak voltage value, and the voltage value is ; take the wavelength as the reference, and calculate the wavelength of each effective transmission peak on both sides of the reference peak in turn, and the calculation expression is: （10）； wherein is the current transmission peak order number, is the reference peak order number; the wavelength of each effective transmission peak the corresponding voltage value one-to-one correspondence, forming a wavelength-voltage reference sequence n is the total number of effective transmission peaks Finally, the complete wavelength-voltage reference sequence is output, which provides a stable wavelength calibration reference for the fourth step of linear calculation of the FBG reflection peak wavelength.

6. The high-precision fiber grating demodulation method according to claim 5, characterized in that, The third step is: The step 3.1 is specifically: The FBG reflection wave voltage signal after the noise removal in the first step is taken as a processing object, and is divided into several intervals according to 5 continuous sampling points as an interval; a cubic spline function is constructed for each interval, and the voltage continuity, first derivative continuity and second derivative continuity conditions of the interval endpoints are met, so as to splice and form a fitting function covering the whole voltage spectrum , and the calculation expression is: （11）； （12）； （13）； （14）； （15）； in, This represents the fitted function; j is the interval index. , , , Represents the fitting coefficient; , Represents the x and y coordinates of the points in the interval; for exist The first right derivative at; for exist The first left derivative at; for exist The second right derivative at point; for exist The second left derivative at that point; The step 3.2 is specifically: Step 3.1 To optimize the object, the objective function is constructed as: （16）； Wherein, E represents the objective function, for quantifying the deviation of model output and target value; m represents the number of sampling points, taking m = 5; Indicates The corresponding fitting value; Indicates The corresponding voltage value; Iteratively adjusting the fitting coefficients , , , , such that E < 0.005 V, outputting the optimized fitting function ; The step 3.3 is specifically: For the first derivative, the expression is calculated as: （17）； Let , the peak sampling position is solved; substitute into , the peak voltage is obtained, output and , which are the characteristic parameters for the fourth step of wavelength linear solution.

7. The high-precision fiber grating demodulation method according to claim 6, characterized in that, The fourth step is specifically: Step 4.1, peak voltage of FBG reflection peak based on third step With the wavelength-voltage reference sequence of the second step, the adjacent reference points are screened, in particular: traversing the reference sequence constructed in the second step , finding two reference voltages adjacent to , , and corresponding wavelengths , wherein ; Step 4.2, based on the adjacent reference points of step 4.1 and the third step The wavelength is calculated by linear interpolation, and the corresponding calculation expression is: （18）； wherein, represents the FBG reflection wavelength to be demodulated; represents the reference voltage value adjacent to and on the left of ; represents the reference voltage value adjacent to and on the right of ; , respectively represent the corresponding reference wavelengths; , ; The FBG reflection wavelength is finally calculated and demodulation is completed.

8. A high-precision fiber grating demodulation system, characterized in that, The high-precision fiber grating demodulation system realizes the high-precision fiber grating demodulation method of any one of claims 1-7, comprising a signal generation module, an optical module, a signal acquisition module, and an upper computer processing module.

9. The high-precision fiber grating demodulation system according to claim 8, characterized in that, The high-precision fiber grating demodulation system is specifically: The signal generation module comprises a signal generator; the signal generator generates a sawtooth wave voltage signal acting on the tunable FP filter; The optical module comprises an ASE broadband light source, a tunable FP filter, an FP standard, a 1×2 coupler, and a circulator; the ASE broadband light source is modulated by the tunable FP filter, and then divided into two paths by the 1×2 coupler, which are respectively the sensing light signal and the reference light signal: the sensing light signal is connected with the signal acquisition module after being injected into the FBG sensing array by the circulator, and the reference light signal is connected with the signal acquisition module after being input into the FP standard; The signal acquisition module comprises two photodetectors, an analog-to-digital conversion chip ADC, and a master chip; specifically: the first photodetector is connected with the FBG sensor to convert the FBG reflection wave optical signal into an FBG reflection wave voltage signal; the second photodetector is connected with the FP standard to convert the FP standard reference light signal into an FP standard reference voltage signal; the analog-to-digital conversion chip ADC converts the continuous FBG reflection wave voltage signal and the FP standard reference voltage signal into discrete digital signals; the master chip sends the digital signals converted by the analog-to-digital conversion chip ADC to the upper computer processing module; The upper computer processing module is the core execution carrier of the above-mentioned fiber grating demodulation method, and is used for receiving the data transmitted by the signal acquisition module.

10. The high-precision fiber grating demodulation system according to claim 9, characterized in that, The host computer processing module further comprises a data storage unit and a visualization unit; the data storage unit saves the denoised signal, the fitting curve data and the demodulation result in CSV format , supports historical data backtracking; the visualization unit dynamically displays the voltage signal curve, the fitting curve and the demodulation wavelength through a real-time curve, and presents the processing process and result.

Images