Distributed intelligent monitoring system and method based on optical fiber communication

Abstract

The invention discloses a distributed intelligent monitoring system and method based on optical fiber communication, and particularly relates to the technical field of optical fiber communication. Brillouin optical time domain analysis is combined with an adaptive laser power regulation and control algorithm, the detection optical power is dynamically adjusted according to the nonlinear characteristic of an optical fiber, nondestructive Kalman filtering is combined with a dynamic Bayesian network, real-time state estimation is carried out on collected data, time sequence analysis is carried out based on a UKF predicted value, and the detection optical power is obtained. An improved signal deconvolution algorithm is introduced in a BOTDA demodulation process to effectively separate SBS-induced false signals, and signal processing is adaptively optimized through a dynamic noise weight adjustment model, so that the accuracy of data demodulation is improved, and the accuracy of SBS-induced false signals is improved. Through fusion of optical signal regulation and control, intelligent data processing and an automatic early warning mechanism, the stability, precision and predictive capacity of the optical fiber monitoring system in a complex environment are improved, and the optical fiber monitoring system is suitable for safety monitoring of key infrastructures.

Description

Technical Field

[0001] The present invention relates to the technical field of optical fiber communication, and in particular to a distributed intelligent monitoring system and method based on optical fiber communication. Background Art

[0002] With the rapid development of fiber-optic communication technology, distributed intelligent monitoring systems based on fiber-optic transmission have been widely used in fields such as industrial safety, infrastructure health monitoring, smart grids, and intelligent transportation. Compared with traditional wireless monitoring methods, fiber-optic communication offers the advantages of high speed, low latency, high anti-interference capabilities, and ultra-long-distance transmission, making it particularly suitable for large-scale, long-distance, and high-precision monitoring applications. In practical applications, distributed fiber-optic sensing technologies (such as DTS, DAS, and BOTDA) combined with artificial intelligence algorithms can provide real-time monitoring and prediction of physical quantities such as environmental changes, object vibration, temperature, and pressure, thereby enabling early fault detection, anomaly alarms, and remote operation and maintenance.

[0003] The existing technology has the following shortcomings:

[0004] In distributed fiber-optic sensing systems (such as BOTDA), optical signals transmitted through the fiber can be distorted by nonlinear Brillouin scattering, leading to false alarms or data misinterpretations. This problem is particularly severe in scenarios involving long-distance fiber monitoring (>50 km), simultaneous sampling at multiple points, or high optical power. Specifically, when the fiber is subjected to stress, temperature changes, or vibration, the Brillouin scattering signal undergoes a frequency shift. However, in highly nonlinear environments, the fiber can generate self-stimulated Brillouin scattering (SBS), which can amplify the measured signal and even overlap with the actual monitoring signal, causing false detections or misinterpretations. Summary of the Invention

[0005] The purpose of the present invention is to provide a distributed intelligent monitoring system and method based on optical fiber communication to solve the shortcomings of the background technology.

[0006] In order to achieve the above object, the present invention provides the following technical solution: a distributed intelligent monitoring method based on optical fiber communication, comprising:

[0007] A Brillouin optical time-domain analysis system is used to obtain real-time status information along the optical fiber, and an adaptive laser power control algorithm is used to dynamically adjust the detection optical power according to the nonlinear characteristics of the optical fiber.

[0008] A lossless Kalman filter is used to estimate the real-time state of the collected data. Combined with a dynamic Bayesian network, the future state estimate calculated by the UKF filter is used as input to perform time series prediction of the physical state along the optical fiber.

[0009] During the demodulation process of the BOTDA system, an improved signal deconvolution algorithm is introduced to separate the false signal components caused by self-excited Brillouin scattering, and a dynamic noise weight adjustment model is used to adaptively adjust the signal processing weight according to environmental parameters.

[0010] An intelligent abnormality classification warning mechanism is adopted to automatically determine the fault level and trigger the corresponding alarm strategy based on the fiber status assessment results.

[0011] Preferably, the adaptive laser power control algorithm specifically includes: calculating the Brillouin gain factor to evaluate the impact of the current optical power on the signal gain, and adjusting the power if the gain factor exceeds a threshold; setting a Brillouin scattering threshold, and reducing the pump light power if the gain factor exceeds the threshold, otherwise maintaining the current power level.

[0012] Preferably, the lossless Kalman filter specifically includes:

[0013] A lossless transformation method is used to generate Sigma points to approximate the state distribution of nonlinear systems;

[0014] Calculate the weighted mean to estimate the optimal value of the current state and update the system error estimate based on the weighted covariance;

[0015] Combined with the prediction covariance calculation, the system state estimation is optimized and the measurement update process is adjusted.

[0016] Preferably, the dynamic Bayesian network time series prediction specifically includes:

[0017] Calculate state transition probabilities and establish state change models based on historical data;

[0018] Use the maximum expectation algorithm to optimize state prediction and adapt it to changes in different environmental factors;

[0019] Calculate the predicted value for the next k steps and output the physical state trend along the optical fiber.

[0020] Preferably, the improved signal deconvolution algorithm specifically includes:

[0021] Calculate the Brillouin gain spectrum and analyze the real physical state signal along the optical fiber;

[0022] The Wiener-Hopf blind deconvolution method was used to separate the false signals induced by SBS;

[0023] An iterative optimization mechanism is adopted to adjust the deconvolution model based on error feedback to enhance signal quality.

[0024] Preferably, the dynamic noise weight adjustment model specifically includes:

[0025] Calculate the basic noise model and build a noise signature library based on historical data;

[0026] Adopting environmental noise measurement model to evaluate the impact of environmental factors on noise in real time;

[0027] Calculate the dynamic noise weight coefficient and adaptively adjust the measurement covariance matrix according to environmental parameters to optimize signal demodulation accuracy.

[0028] Preferably, the intelligent abnormality classification warning mechanism specifically includes:

[0029] Calculate anomaly scores to evaluate temperature, strain, and vibration data along the fiber;

[0030] Set abnormality classification thresholds and classify abnormal situations into four levels: normal, warning, severe, and emergency;

[0031] According to the abnormality level, different levels of alarm strategies are triggered.

[0032] Preferably, the alarm strategy specifically includes: low-level alarm: sending SMS / email to notify operation and maintenance personnel, increasing data sampling frequency; intermediate alarm: the system automatically records abnormal areas, generates maintenance work orders, and arranges manual inspections; high-level alarm: triggering automatic safety mechanisms.

[0033] The present invention also provides a distributed intelligent monitoring system based on optical fiber communication, including a data acquisition module, a time sequence prediction module, a signal demodulation and optimization module and an intelligent early warning module;

[0034] Data acquisition module: uses a Brillouin optical time-domain analysis system to obtain real-time status information along the optical fiber, and adopts an adaptive laser power control algorithm to dynamically adjust the detection light power according to the nonlinear characteristics of the optical fiber;

[0035] Time series prediction module: uses lossless Kalman filtering to perform real-time state estimation of collected data. Combined with dynamic Bayesian networks, it uses the future state estimate calculated by UKF filtering as input to perform time series prediction of the physical state along the optical fiber.

[0036] Signal demodulation and optimization module: During the demodulation process of the BOTDA system, an improved signal deconvolution algorithm is introduced to separate the false signal components caused by self-excited Brillouin scattering, and a dynamic noise weight adjustment model is used to adaptively adjust the signal processing weight according to environmental parameters;

[0037] Intelligent early warning module: Adopts intelligent abnormality classification early warning mechanism, automatically determines the fault level and triggers the corresponding alarm strategy based on the fiber status assessment results.

[0038] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0039] 1. The present invention combines Brillouin optical time-domain analysis (BOTDA) with an adaptive laser power control algorithm. The present invention can dynamically adjust the detection optical power according to the nonlinear characteristics of the optical fiber to suppress the SBS effect and improve the stability of long-distance (>50km) monitoring. At the same time, a lossless Kalman filter (UKF) combined with a dynamic Bayesian network (DBN) is used to accurately estimate the real-time status data along the optical fiber and predict the future status based on historical data, thereby improving the intelligence level of the monitoring system. In addition, during the BOTDA signal demodulation process, an improved signal deconvolution algorithm is introduced to separate the false signals induced by SBS, and a dynamic noise weight adjustment model is used to adaptively optimize signal processing according to environmental parameters (temperature gradient, mechanical vibration, etc.), further improving the demodulation accuracy.

[0040] 2. This invention automatically evaluates fiber optic monitoring data through an intelligent anomaly classification and early warning mechanism. It then categorizes the data into four levels: normal, early warning, severe, and emergency, based on the severity of the anomaly. This system triggers corresponding alarm policies, such as low-level SMS / email notifications, mid-level maintenance scheduling, and high-level automatic safety measures (such as pipeline shutdowns and grid load adjustments). This method not only significantly improves the real-time performance, accuracy, and anti-interference capabilities of the fiber optic monitoring system, but also reduces false alarms and missed alarms, lowering maintenance costs. It is suitable for long-term safety monitoring of critical infrastructure such as oil pipelines, power transmission lines, bridges, and tunnels, and has broad application value in areas such as industrial intelligent monitoring, infrastructure health management, and disaster early warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0042] Figure 1 Flow chart of the method of the present invention.

[0043] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0044] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0045] Example 1, please refer to Figure 1 As shown, the distributed intelligent monitoring method based on optical fiber communication described in this embodiment includes:

[0046] A Brillouin optical time-domain analysis system is used to obtain real-time status information along the optical fiber, and an adaptive laser power control algorithm is used to dynamically adjust the detection optical power according to the nonlinear characteristics of the optical fiber.

[0047] A lossless Kalman filter is used to estimate the real-time state of the collected data. Combined with a dynamic Bayesian network, the future state estimate calculated by the UKF filter is used as input to perform time series prediction of the physical state along the optical fiber.

[0048] During the demodulation process of the BOTDA system, an improved signal deconvolution algorithm is introduced to separate the false signal components caused by self-excited Brillouin scattering, and a dynamic noise weight adjustment model is used to adaptively adjust the signal processing weight according to environmental parameters.

[0049] An intelligent abnormality classification warning mechanism is adopted to automatically determine the fault level and trigger the corresponding alarm strategy based on the fiber status assessment results.

[0050] This method uses a Brillouin optical time-domain analysis (BOTDA) system to perform high-precision monitoring of the optical fiber status. It also introduces an adaptive laser power control algorithm to dynamically adjust the detection optical power based on the nonlinear characteristics of the optical fiber to suppress the self-stimulated Brillouin scattering (SBS) effect and improve signal stability and reliability. The following are the specific implementation steps:

[0051] Select an appropriate optical fiber as the sensing medium (such as single-mode fiber or doped fiber), ensuring its Brillouin gain coefficient and bending radius are suitable for long-distance measurement. Set the initial laser operating wavelength (such as 1550nm) and pulse width (such as 10ns-50ns). The pulse width determines the spatial resolution (generally 110m). Set the initial detection optical power (such as 10dBm) to avoid excessive initial power, which may cause SBS gain to exceed the limit.

[0052] Initial system calibration is performed using a reference standard fiber, and its Brillouin Frequency Shift (BFS) characteristics are measured. The fiber is heated using a temperature control device (such as a thermostat) and the BFS changes at different temperatures are measured to construct a temperature-stress curve.

[0053] Pump light is generated using a tunable narrow-linewidth laser (such as a DFB laser), and its pulse width is controlled using an intensity modulator (IM). An erbium-doped fiber amplifier (EDFA) is used to boost the pump light power and improve signal strength. A polarization controller (PC) optimizes the fiber polarization state to reduce polarization-induced measurement errors. A continuous-wave probe light (Stokes light) is input at the other end, and a frequency step-sweep method is used to gradually change the probe light frequency to obtain the Brillouin gain spectrum at different locations.

[0054] The Brillouin scattering signal is received by a high-precision photodetector (APD or InGaAs detector) and converted into an electrical signal. A high-bandwidth data acquisition card (such as a 1GSps sampling rate ADC) is used for signal acquisition and the data is stored in an FPGA or high-speed memory module. Fourier transform (FFT) and wavelet transform are used to reduce noise on the raw signal to improve signal quality.

[0055] The Lorentzian fitting algorithm is used to curve fit the Brillouin gain spectrum and calculate the Brillouin frequency shift at each sampling point. Combined with the temperature-stress relationship curve, the temperature and stress field distribution along the optical fiber are calculated.

[0056] Since high-power pump light can induce the SBS effect, resulting in signal distortion and even self-oscillation, this method adopts an adaptive laser power control algorithm to dynamically adjust the detection light power to ensure measurement accuracy.

[0057] Calculate the current SBS gain factor G through the Brillouin gain spectrum B , the formula is:

[0058] Where: g B is the Brillouin gain coefficient (m / W), which depends on the fiber material; P p is the current pump light power (W); Δv B is the Brillouin gain bandwidth (MHz).

[0059] Set the Brillouin scattering threshold G B ,th, if G B >G B , th, indicating that the optical power is too high and needs to be reduced; otherwise, maintain the current power; calculate the nonlinear Brillouin scattering gain Γ in combination with the fiber length L B , if the conditions are met, the power adjustment mechanism is triggered:

[0060] Among them, Γ B is the preset safety threshold.

[0061] Adopting proportional-integral-derivative (PID) control algorithm to dynamically adjust the pump light power Pp , the formula is:

[0062]

[0063] Among them: K p , K i , K d is the PID control parameter, and adaptive learning optimization is adopted; when G B When it is higher, reduce P p When G B When it is low, moderately increase P p .

[0064] In the next round of BFS calculation, the detection optical power is further optimized based on the real-time optical power-Brillouin gain curve. If the SBS effect is still strong, the polarization state of the optical fiber is adjusted or the optical fiber doping is increased to lower the SBS threshold (such as using germanium-doped fiber).

[0065] In a distributed optical fiber monitoring system, the physical state along the optical fiber can be expressed as a state vector: t =[T t , S t , V t ] T ; Where: T t is the temperature of a certain position of the optical fiber (Temperature), S t is the strain at a certain position of the optical fiber, V t is the vibration intensity (Vibration) at a certain position of the optical fiber, and T is the matrix transpose.

[0066] The state evolution equation of the system can be expressed as a nonlinear dynamic system: t+1 =f(x t ,u t )+w t ; Where: f is a nonlinear state transfer function, for example: Among them, α, β, γ, δ, η, ξ are environmental parameters, u t is the control input, such as external disturbances in the environment (such as wind speed and pressure changes); w t ~N(0,Q) is the process noise, which obeys Gaussian distribution and has a covariance matrix of Q. The measurement model is: t =h(x t )+v t ; Where: h is the measurement function, for example:

[0067] Among them, λ1, λ2, are the error coefficients of the measurement system, v t ~N(0, R) is the measurement noise, and the covariance matrix is R.

[0068] Due to the nonlinear characteristics of the system, the unlossy Kalman filter (UKF) is used for state estimation. The specific calculation steps are as follows:

[0069] UKF uses UnscentedTransform to generate a set of Sigma points χ i To approximate the state distribution: χ0 = x t ,

[0070] Where: L is the state dimension (in this example, L = 3); P t is the state error covariance matrix, k is a hyperparameter, and is generally taken as κ=0.

[0071] Perform a nonlinear transformation on each Sigma point: Calculate the weighted mean: Compute the forecast covariance:

[0072] Where,

[0073] is the prior error covariance matrix, which represents the estimate of the uncertainty of the system state in the prediction step, with a dimension of L×L, where L is the dimension of the state variable and W i The weight of the Sigma point is used to weight the contribution of different Sigma points. The weight is usually predefined according to the lossless transformation method. is the predicted value of the i-th Sigma point, which represents the estimated state after the nonlinear state transfer function at time t+1, with a dimension of L×1.

[0074] Define the state transition probability of DBN: Where: h t is a latent variable (such as environmental interference, material fatigue); P(h t |x t ) is the prior probability of the latent variable; P(x t+1 |h t ) is the probability distribution of future states.

[0075] The DBN model is trained by the maximum expectation algorithm (EM algorithm) and the state estimation value predicted by UKF is used As input: Finally calculate the prediction for the next k steps: x t+k =argmaxP(x t+k |x 1：t ).

[0076] Normal state: predicted temperature T t+k and strain S t+kIn the safe range, it means that the optical fiber is not affected by abnormal stress; Abnormal state: the predicted temperature T t+k Beyond the safety range or strain S t+k If the safety range is exceeded, optical fiber breakage may occur.

[0077] In distributed fiber-optic sensing systems based on Brillouin optical time-domain analysis (BOTDA), self-stimulated Brillouin scattering (SBS) can introduce nonlinear gain effects, leading to the inclusion of spurious signal components in the measured signal, affecting the precise measurement of physical quantities such as temperature and strain. Traditional signal processing methods (such as FFT spectrum analysis) have difficulty distinguishing true signals from SBS-induced spurious signals. Therefore, an improved signal deconvolution algorithm and a dynamic noise weight adjustment model are needed to enhance signal demodulation accuracy.

[0078] By using deconvolution technology, the signal distortion caused by the SBS effect in the BOTDA system is removed, and the true Brillouin gain signal and the false signal components are separated. In the BOTDA system, the Brillouin gain spectrum B(f) can be expressed as: Where: g B is the Brillouin gain coefficient, P p is the pump light power, Δv B is the Brillouin gain bandwidth;

[0079] H(f) is the system response function, which describes the actual Brillouin frequency shift behavior of the optical fiber; n(f) is the noise, including environmental noise and false signals induced by SBS.

[0080] Due to the nonlinear effect induced by SBS, the measured signal B meas (f) contains a convolution term: B meas (f)=B(f)*G SBS (f)+n(f); where: G SBS (f) is the SBS nonlinear response function, which causes signal distortion; * represents the convolution operation.

[0081] To eliminate the signal distortion caused by SBS, we use the improved Wiener-Hopf blind deconvolution method to reconstruct the true signal H(f): in:

[0082] represents the estimated true Brillouin gain spectrum, B meas (f) is the nonlinear gain model induced by SBS (which can be determined experimentally), and λ is the regularization parameter to prevent numerical instability.

[0083] Because G SBS (f) May be affected by environmental factors, so iterative optimization is used:

[0084] Set initial As an estimate;

[0085] Calculate the deconvolution And calculate the error:

[0086] According to the error e (k) renew The expression is: η is the learning rate, which controls the iteration step size and determines the speed of SBS error correction. Usually, η is set to a small value (such as 0.01 to 0.1) to prevent numerical divergence.

[0087] Iterate until convergence (error e (k) below the threshold).

[0088] Since the characteristics of SBS-induced noise vary with environmental factors (temperature gradient, mechanical vibration, and fiber aging), a dynamic noise weight adjustment model is adopted to adaptively adjust the signal processing weights and optimize the demodulation accuracy.

[0089] In the UKF estimation process, the state update formula is: t+1 =f(x t )+w t ; Among them: the process noise covariance matrix Q needs to be adaptively adjusted to adapt to changes in environmental noise. Define the dynamic noise weight: Q t =αQ base +(1-α)Q env ;Q base As the basic noise model, Q is set based on experimental data. env is the measured value of environmental noise, and α is the adaptive adjustment coefficient.

[0090] In a distributed fiber optic sensing (DOFS) system, fiber conditions (such as temperature, strain, and vibration) may become abnormal due to environmental changes or equipment failures. Therefore, an intelligent, graded, and early warning mechanism for abnormalities can automatically determine the fault level and trigger the appropriate alarm strategy based on fiber condition assessment results. This improves the monitoring system's real-time response capabilities, reduces false alarms and missed alarms, and enhances maintenance efficiency.

[0091] BOTDA (Brillouin Optical Time Domain Analysis) or DAS (Distributed Acoustic Sensing) is used to obtain the temperature, strain, vibration and other physical conditions along the optical fiber. Undestructed Kalman filtering (UKF) is used to perform signal denoising to obtain a more accurate optical fiber state estimate x t , calculate the abnormal score A of each monitoring point t , the expression is:

[0092] Where: T norm 、Snorm 、V norm is the reference value of temperature, strain and vibration under normal operating conditions, T max 、S max 、V max is the maximum allowable value of the abnormal state; T min 、T min 、T min is the minimum allowable value of the abnormal state; w1, w2, w3 are weighting coefficients (which can be optimized through data training).

[0093] Classification based on anomaly score:

[0094] Among them: A low 、A mid 、A high is the preset abnormal threshold.

[0095] Normal state: The fiber status is stable and no alarm is required; data continues to be collected regularly and stored in the database for trend analysis.

[0096] Early warning status triggers a low-level alarm (such as SMS / email notification) to alert operation and maintenance personnel to possible abnormal trends. The system enters enhanced monitoring mode and increases the data sampling frequency.

[0097] In a serious state, a medium-level alarm is triggered. The system automatically records the abnormal area and marks it as a high-risk area. The remote operation and maintenance system recommends arranging personnel to conduct on-site inspections and generates maintenance work orders in the background system.

[0098] In an emergency, an emergency alarm is triggered (such as on-site sound and light alarm, mobile phone push notification, automatic phone call), and automatic safety measures are immediately implemented.

[0099] Example 2, please refer to Figure 2 As shown, the distributed intelligent monitoring system based on optical fiber communication described in this embodiment includes a data acquisition module, a time series prediction module, a signal demodulation and optimization module, and an intelligent early warning module;

[0100] Data acquisition module: uses a Brillouin optical time-domain analysis system to obtain real-time status information along the optical fiber, and adopts an adaptive laser power control algorithm to dynamically adjust the detection light power according to the nonlinear characteristics of the optical fiber;

[0101] Time series prediction module: uses lossless Kalman filtering to perform real-time state estimation of collected data. Combined with dynamic Bayesian networks, it uses the future state estimate calculated by UKF filtering as input to perform time series prediction of the physical state along the optical fiber.

[0102] Signal demodulation and optimization module: During the demodulation process of the BOTDA system, an improved signal deconvolution algorithm is introduced to separate the false signal components caused by self-excited Brillouin scattering, and a dynamic noise weight adjustment model is used to adaptively adjust the signal processing weight according to environmental parameters;

[0103] Intelligent early warning module: Adopts intelligent abnormality classification early warning mechanism, automatically determines the fault level and triggers the corresponding alarm strategy based on the fiber status assessment results.

[0104] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0105] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0106] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0107] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A distributed intelligent monitoring method based on optical fiber communication, characterized in that: include: A Brillouin optical time-domain analysis system is used to obtain real-time status information along the optical fiber, and an adaptive laser power control algorithm is used to dynamically adjust the detection optical power according to the nonlinear characteristics of the optical fiber. A lossless Kalman filter is used to estimate the real-time state of the collected data. Combined with a dynamic Bayesian network, the future state estimate calculated by the UKF filter is used as input to perform time series prediction of the physical state along the optical fiber. During the demodulation process of the BOTDA system, an improved signal deconvolution algorithm is introduced to separate the false signal components caused by self-excited Brillouin scattering, and a dynamic noise weight adjustment model is used to adaptively adjust the signal processing weight according to environmental parameters. An intelligent abnormality classification warning mechanism is adopted to automatically determine the fault level and trigger the corresponding alarm strategy based on the fiber status assessment results.

2. The distributed intelligent monitoring method based on optical fiber communication according to claim 1, characterized in that: The adaptive laser power control algorithm specifically includes: calculating the Brillouin gain factor to evaluate the impact of the current optical power on the signal gain. If the gain factor exceeds a threshold, the power is adjusted; setting a Brillouin scattering threshold. If the gain factor exceeds the threshold, the pump light power is reduced; otherwise, the current power level is maintained.

3. The distributed intelligent monitoring method based on optical fiber communication according to claim 1, characterized in that: The lossless Kalman filter specifically includes: A lossless transformation method is used to generate Sigma points to approximate the state distribution of nonlinear systems; Calculate the weighted mean to estimate the optimal value of the current state and update the system error estimate based on the weighted covariance; Combined with the prediction covariance calculation, the system state estimation is optimized and the measurement update process is adjusted.

4. The distributed intelligent monitoring method based on optical fiber communication according to claim 3, characterized in that: The dynamic Bayesian network time series prediction specifically includes: Calculate state transition probabilities and establish state change models based on historical data; Use the maximum expectation algorithm to optimize state prediction and adapt it to changes in different environmental factors; Calculate the predicted value for the next k steps and output the physical state trend along the optical fiber.

5. The distributed intelligent monitoring method based on optical fiber communication according to claim 4, characterized in that: The improved signal deconvolution algorithm specifically includes: Calculate the Brillouin gain spectrum and analyze the real physical state signal along the optical fiber; The Wiener-Hopf blind deconvolution method was used to separate the false signals induced by SBS; An iterative optimization mechanism is adopted to adjust the deconvolution model based on error feedback to enhance signal quality.

6. The distributed intelligent monitoring method based on optical fiber communication according to claim 5, characterized in that: The dynamic noise weight adjustment model specifically includes: Calculate the basic noise model and build a noise signature library based on historical data; Adopting environmental noise measurement model to evaluate the impact of environmental factors on noise in real time; Calculate the dynamic noise weight coefficient and adaptively adjust the measurement covariance matrix according to environmental parameters to optimize signal demodulation accuracy.

7. The distributed intelligent monitoring method based on optical fiber communication according to claim 6, characterized in that: The intelligent abnormality classification warning mechanism specifically includes: Calculate anomaly scores to evaluate temperature, strain, and vibration data along the fiber; Set abnormality classification thresholds and classify abnormal situations into four levels: normal, warning, severe, and emergency; According to the abnormality level, different levels of alarm strategies are triggered.

8. The distributed intelligent monitoring method based on optical fiber communication according to claim 7, characterized in that: The alarm strategy specifically includes: low-level alarm: sending SMS / email to notify operation and maintenance personnel, increasing the data sampling frequency; intermediate alarm: the system automatically records abnormal areas, generates maintenance work orders, and arranges manual inspections; high-level alarm: triggering automatic safety mechanisms.

9. A distributed intelligent monitoring system based on optical fiber communication, for implementing the distributed intelligent monitoring method based on optical fiber communication according to any one of claims 1 to 8, characterized in that: Including data acquisition module, time series prediction module, signal demodulation and optimization module and intelligent early warning module; Data acquisition module: uses a Brillouin optical time-domain analysis system to obtain real-time status information along the optical fiber, and adopts an adaptive laser power control algorithm to dynamically adjust the detection light power according to the nonlinear characteristics of the optical fiber; Time series prediction module: uses lossless Kalman filtering to perform real-time state estimation of collected data. Combined with dynamic Bayesian networks, it uses the future state estimate calculated by UKF filtering as input to perform time series prediction of the physical state along the optical fiber. Signal demodulation and optimization module: During the demodulation process of the BOTDA system, an improved signal deconvolution algorithm is introduced to separate the false signal components caused by self-excited Brillouin scattering, and a dynamic noise weight adjustment model is used to adaptively adjust the signal processing weight according to environmental parameters; Intelligent early warning module: Adopts intelligent abnormality classification early warning mechanism, automatically determines the fault level and triggers the corresponding alarm strategy based on the fiber status assessment results.

Images