FPGA-based distributed optical fiber vibration positioning and suppression method and system

Abstract

The application provides a distributed optical fiber vibration positioning and suppression method and system based on FPGA, relates to the field of optical fiber sensing technology, and comprises the following steps: receiving a Raman scattering light signal of an optical fiber sensing unit, performing empirical mode decomposition and Hilbert transformation through FPGA, constructing a time-frequency feature map to calculate an energy peak time difference, determining a vibration source position by using a hyperbolic positioning method, calculating optimal driving parameters based on vibration propagation characteristics, and cooperatively controlling multiple suppression nodes to generate interference waveforms to block the propagation of vibration energy. The application can realize accurate positioning and effective suppression of optical fiber vibration and improve the performance of a distributed optical fiber sensing system.

Description

Technical Field

[0001] This invention relates to the field of fiber optic sensing technology, and in particular to a distributed fiber optic vibration localization and suppression method and system based on FPGA. Background Technology

[0002] Distributed fiber optic vibration detection and suppression systems have broad application prospects in smart grids, oil pipeline safety monitoring, transportation infrastructure protection, and border security. This technology uses optical fiber as the sensing medium to achieve real-time detection, location, and suppression of vibration signals along the pipeline. Utilizing the Raman scattering principle, the distributed fiber optic sensing system uses the fiber itself as a sensing element to effectively sense external vibrations and convert them into measurable optical signals. With the development of programmable logic devices such as FPGAs, the computational power and real-time performance of distributed fiber optic vibration location and suppression systems have been significantly improved.

[0003] Traditional fiber optic vibration signal processing methods typically employ fixed basis function analysis methods such as Fourier transform. These methods have limited ability to extract time-frequency features from non-stationary signals, making it difficult to accurately characterize the transient characteristics of vibration signals, thus affecting positioning accuracy and suppression effectiveness. Existing technologies for vibration source localization often use single time delay estimation algorithms, failing to fully consider the multipath effects and dispersion phenomena of vibration signals during spatial propagation. This results in significant positioning deviations in complex environments, failing to provide accurate location information for subsequent vibration suppression. Most existing vibration suppression methods employ passive damping or single-point active suppression strategies, lacking precise modeling of vibration energy propagation paths and attenuation characteristics. This prevents coordinated control of multiple suppression nodes, leading to poor suppression effects and increasing the risk of system instability or secondary vibration interference. Summary of the Invention

[0004] This invention provides a distributed optical fiber vibration localization and suppression method and system based on FPGA, which can solve the problems in the prior art.

[0005] A first aspect of the present invention provides a distributed optical fiber vibration localization and suppression method based on FPGA, comprising:

[0006] The system receives Raman scattered light signals returned by sensing units on distributed optical fibers and obtains digital vibration signals through analog-to-digital conversion.

[0007] The FPGA performs empirical mode decomposition on the digital vibration signal to obtain intrinsic mode components, and extracts the instantaneous features of the intrinsic mode components through Hilbert transform. Based on the instantaneous features, a Hilbert-Huang time-frequency feature map is constructed. The propagation delay of the optical fiber vibration is determined by calculating the energy peak time difference in the Hilbert-Huang time-frequency feature map. Based on the propagation delay of the optical fiber vibration, the spatial coordinates of the optical fiber vibration source are calculated using the hyperbolic positioning method.

[0008] Based on the spatial coordinates, multiple preset suppression nodes around the optical fiber vibration source are activated. The FPGA calculates the propagation path and attenuation period of the vibration energy in space according to the vibration propagation characteristics collected by each suppression node. Based on the propagation path and the attenuation period, the FPGA calculates the optimal driving timing and optimal driving parameters for each suppression node. The optimal driving timing and optimal driving parameters are used to coordinately control the multiple suppression nodes to generate interference waveforms, block the propagation path of the vibration energy, and suppress the optical fiber vibration in the target area.

[0009] The FPGA performs empirical mode decomposition on the digital vibration signal to obtain intrinsic mode components, and extracts the instantaneous features of the intrinsic mode components through Hilbert transform. Based on the instantaneous features, a Hilbert-Huang time-frequency feature map is constructed, including:

[0010] The FPGA identifies local extrema in the digital vibration signal. Based on these local extrema, it constructs an upper and lower envelope using cubic spline interpolation. It then calculates the mean values ​​of the upper and lower envelopes to obtain a mean signal. The mean signal is subtracted from the digital vibration signal to obtain candidate intrinsic modal components. The FPGA determines whether the number of extrema of the candidate intrinsic modal components is equal to or differs from the number of zero-crossing points by 1. If so, the candidate intrinsic modal component is identified as such.

[0011] The intrinsic mode components are subjected to the Hilbert transform to obtain the Hilbert transformed signal. The intrinsic mode components and the Hilbert transformed signal are then combined to construct an analytic signal in complex form. Based on the analytic signal, the signal envelope is calculated to obtain the instantaneous amplitude, and the phase angle is calculated and the time derivative is obtained to obtain the instantaneous frequency.

[0012] Calculate the local rate of change of the instantaneous amplitude and the instantaneous frequency, and determine the width of the adaptive window on the time-frequency plane based on the local rate of change; after locally smoothing the time-frequency energy distribution of the intrinsic mode components using the adaptive window, perform local contrast enhancement based on the Laplace operator to obtain the Hilbert-Huang time-frequency feature map.

[0013] The fiber vibration propagation delay is determined by calculating the energy peak time difference in the Hilbert-Huang time-frequency characteristic spectrum. Based on the fiber vibration propagation delay, the spatial coordinates of the fiber vibration source are calculated using the hyperbolic positioning method, including:

[0014] The time-domain energy curve is obtained by integrating the Hilbert-Huang time-frequency feature map along the time dimension. The local mean and local standard deviation of the time-domain energy curve are calculated and combined to obtain an adaptive threshold. Multiple energy peak moments are extracted from the time-domain energy curve using the adaptive threshold. The time difference between the energy peak moments is calculated to obtain the fiber vibration propagation delay.

[0015] A predetermined number of sensing units are selected on the distributed optical fiber. The sensing unit that detects the vibration signal earliest is set as the focal sensing unit. Other sensing units are paired with the focal sensing unit in sequence to form multiple sets of sensing unit pairs. Based on the vibration propagation delay of the optical fiber, a hyperbola is constructed with the two sensing units of each set of sensing unit pairs as the focal points.

[0016] The intersection points of the multiple sets of sensing units with their respective hyperbolas are determined by solving the system of equations. A search region is constructed with the intersection points as the center. The spatial coordinates of the optical fiber vibration source are obtained by solving the minimum mean square error criterion within the search region.

[0017] Based on the spatial coordinates, multiple preset suppression nodes around the optical fiber vibration source are activated. The FPGA calculates the propagation path and attenuation period of the vibration energy in space according to the vibration propagation characteristics collected by each suppression node, including:

[0018] Based on the spatial coordinates, a spherical monitoring domain is constructed around the optical fiber vibration source. The Voronoi diagram partitioning algorithm is used to divide the spherical monitoring domain into multiple sub-regions. Activation signals are sent to the suppression nodes in each sub-region according to the constraints of maximizing coverage and minimizing overlap.

[0019] The vibration propagation characteristics are collected using the activated suppression node. The FPGA extracts the propagation path feature coefficients from the vibration propagation characteristics and uses a ray tracing algorithm to calculate the reflection coefficient and refraction coefficient of each propagation medium interface. The propagation path feature coefficients are set as fitness evaluation indicators, and the reflection coefficient and refraction coefficient are used as path constraints. An adaptive particle swarm optimization algorithm is constructed to calculate the propagation path of the vibration energy.

[0020] Based on the propagation path, the decay characteristics of the vibration energy over time are extracted to obtain the time decay factor, and the loss of the vibration energy during spatial propagation is calculated to obtain the spatial decay factor. A bidirectional long short-time memory algorithm is used to establish the dynamic response matrix of the time decay factor and the spatial decay factor. The periodic solution of the dynamic response matrix is ​​solved by the eigenvalue decomposition method to obtain the decay period of the vibration energy.

[0021] Based on the propagation path and the attenuation period, the FPGA calculates the optimal driving timing and optimal driving parameters for each suppression node, and uses the optimal driving timing and optimal driving parameters to collaboratively control multiple suppression nodes to generate interference waveforms, including:

[0022] A phase difference constraint is established between adjacent suppression nodes based on the propagation path, and a timing constraint is established between adjacent suppression nodes based on the decay period. The phase difference constraint and the timing constraint are combined to form an optimization objective. The optimal driving timing of each suppression node is calculated by minimizing the optimization objective.

[0023] The driving amplitude of each suppression node is calculated based on the time decay factor and the spatial decay factor. The deviation between the actual energy and the target energy at each suppression node is calculated. The compensation coefficient is determined based on the deviation. The driving amplitude is dynamically corrected by iteratively adjusting the compensation coefficient to obtain the optimal driving parameters of each suppression node.

[0024] The FPGA allocates the optimal driving parameters to each of the suppression nodes, establishes a global clock synchronization triggering mechanism based on the optimal driving timing, calculates the superposition coefficient of adjacent waveforms based on the phase difference of each of the suppression nodes, determines the waveform superposition method according to the superposition coefficient and the global clock synchronization triggering mechanism, and uses the waveform superposition method to collaboratively control each of the suppression nodes to generate the interference waveform.

[0025] A global clock synchronization triggering mechanism is established based on the optimal driving timing. The superposition coefficient of adjacent waveforms is calculated based on the phase difference of each suppression node. The waveform superposition method is determined according to the superposition coefficient and the global clock synchronization triggering mechanism, including:

[0026] Obtain the system reference clock signal, system reference clock period, and clock compensation delay constant from the FPGA;

[0027] The optimal driving timing is divided by the system reference clock period and superimposed on the clock compensation delay constant to obtain the trigger count value of each suppression node. The system reference clock signal is multiplied by a phase-locked loop to obtain a synchronization clock signal. The synchronization clock signal and the trigger count value are input to the trigger timer of each suppression node to generate the global clock synchronization trigger mechanism.

[0028] The phase difference and distance between adjacent suppression nodes are obtained, and the cosine function value of the phase difference is multiplied by the exponential decay function value of the distance to obtain the superposition coefficient between the adjacent waveforms; the amplitude correction factor of the adjacent waveforms is calculated based on the energy value and energy decay rate of each suppression node.

[0029] Based on the global clock synchronization triggering mechanism, the superposition timing of adjacent waveforms is determined by the superposition coefficient, the waveform amplitude of each suppression node is adjusted according to the amplitude correction factor, and the waveform superposition mode is obtained through the superposition timing and the waveform amplitude.

[0030] A second aspect of the present invention provides an FPGA-based distributed optical fiber vibration localization and suppression system, comprising:

[0031] The first unit is used to receive the Raman scattered light signal returned by the sensing unit on the distributed optical fiber, and obtain the digital vibration signal through analog-to-digital conversion;

[0032] The second unit is used by the FPGA to perform empirical mode decomposition on the digital vibration signal to obtain intrinsic mode components, and to extract the instantaneous features of the intrinsic mode components through Hilbert transform. Based on the instantaneous features, a Hilbert-Huang time-frequency feature spectrum is constructed. The fiber optic vibration propagation delay is determined by calculating the energy peak time difference in the Hilbert-Huang time-frequency feature spectrum. Based on the fiber optic vibration propagation delay, the spatial coordinates of the fiber optic vibration source are calculated using the hyperbolic positioning method.

[0033] The third unit is used to activate multiple preset suppression nodes around the optical fiber vibration source based on the spatial location coordinates. The FPGA calculates the propagation path and attenuation period of the vibration energy in space according to the vibration propagation characteristics collected by each suppression node. Based on the propagation path and the attenuation period, the FPGA calculates the optimal driving timing and optimal driving parameters of each suppression node. The optimal driving timing and optimal driving parameters are used to coordinately control the multiple suppression nodes to generate interference waveforms, block the propagation path of the vibration energy, and suppress the optical fiber vibration in the target area.

[0034] A third aspect of the embodiments of the present invention,

[0035] An electronic device is provided, comprising:

[0036] processor;

[0037] Memory used to store processor-executable instructions;

[0038] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0039] Fourth aspect of the present invention,

[0040] A computer-readable storage medium is provided, having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0041] The beneficial effects of this application are as follows:

[0042] This invention employs empirical mode decomposition combined with Hilbert transform to extract the instantaneous features of vibration signals, constructs a Hilbert-Huang time-frequency feature spectrum, determines the vibration propagation delay by calculating the energy peak time difference, and then uses the hyperbolic positioning method to accurately calculate the spatial coordinates of the vibration source, significantly improving the accuracy and reliability of vibration source positioning.

[0043] This invention activates preset suppression nodes around the vibration source location coordinates, calculates the propagation path and attenuation period of vibration energy in real time through FPGA, and adaptively generates the optimal driving timing and parameters, thereby achieving precise suppression of vibration interference and effectively ensuring the stable operation of the distributed optical fiber system.

[0044] This invention utilizes the parallel computing capabilities of FPGA to achieve integrated vibration signal processing, positioning calculation, and suppression control. The system has a fast response speed and can respond to vibration interference in complex environments in real time. It is suitable for various application scenarios such as earthquake monitoring, pipeline protection, and boundary security, and has strong practicality and promotional value. Attached Figure Description

[0045] Figure 1 This is a flowchart illustrating the distributed optical fiber vibration localization and suppression method based on FPGA according to an embodiment of the present invention.

[0046] Figure 2 A schematic diagram of the calculation process for suppressing node activation and vibration energy propagation. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0049] Figure 1 This is a flowchart illustrating the FPGA-based distributed optical fiber vibration localization and suppression method according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0050] The system receives Raman scattered light signals returned by sensing units on distributed optical fibers and obtains digital vibration signals through analog-to-digital conversion.

[0051] The FPGA performs empirical mode decomposition on the digital vibration signal to obtain intrinsic mode components, and extracts the instantaneous features of the intrinsic mode components through Hilbert transform. Based on the instantaneous features, a Hilbert-Huang time-frequency feature map is constructed. The propagation delay of the optical fiber vibration is determined by calculating the energy peak time difference in the Hilbert-Huang time-frequency feature map. Based on the propagation delay of the optical fiber vibration, the spatial coordinates of the optical fiber vibration source are calculated using the hyperbolic positioning method.

[0052] Based on the spatial coordinates, multiple preset suppression nodes around the optical fiber vibration source are activated. The FPGA calculates the propagation path and attenuation period of the vibration energy in space according to the vibration propagation characteristics collected by each suppression node. Based on the propagation path and the attenuation period, the FPGA calculates the optimal driving timing and optimal driving parameters for each suppression node. The optimal driving timing and optimal driving parameters are used to coordinately control the multiple suppression nodes to generate interference waveforms, block the propagation path of the vibration energy, and suppress the optical fiber vibration in the target area.

[0053] In one optional implementation, the FPGA performs empirical mode decomposition on the digital vibration signal to obtain intrinsic mode components, and extracts the instantaneous features of the intrinsic mode components through Hilbert transform. Based on these instantaneous features, a Hilbert-Huang time-frequency feature map is constructed, including:

[0054] The FPGA identifies local extrema in the digital vibration signal. Based on these local extrema, it constructs an upper and lower envelope using cubic spline interpolation. It then calculates the mean values ​​of the upper and lower envelopes to obtain a mean signal. The mean signal is subtracted from the digital vibration signal to obtain candidate intrinsic modal components. The FPGA determines whether the number of extrema of the candidate intrinsic modal components is equal to or differs from the number of zero-crossing points by 1. If so, the candidate intrinsic modal component is identified as such.

[0055] The intrinsic mode components are subjected to the Hilbert transform to obtain the Hilbert transformed signal. The intrinsic mode components and the Hilbert transformed signal are then combined to construct an analytic signal in complex form. Based on the analytic signal, the signal envelope is calculated to obtain the instantaneous amplitude, and the phase angle is calculated and the time derivative is obtained to obtain the instantaneous frequency.

[0056] Calculate the local rate of change of the instantaneous amplitude and the instantaneous frequency, and determine the width of the adaptive window on the time-frequency plane based on the local rate of change; after locally smoothing the time-frequency energy distribution of the intrinsic mode components using the adaptive window, perform local contrast enhancement based on the Laplace operator to obtain the Hilbert-Huang time-frequency feature map.

[0057] The FPGA performs empirical mode decomposition (EMD) on the input digital vibration signal to identify local extrema, including local maxima and local minima. These extrema are identified by scanning the entire signal sequence and comparing the amplitudes of adjacent sampling points. A sampling point is marked as a local maximum when its amplitude is greater than that of its two adjacent sampling points, and as a local minimum when its amplitude is less than that of its two adjacent sampling points. Based on the identified local maxima, a cubic spline interpolation algorithm is used to construct the upper envelope. This algorithm ensures that the curve value at each maximum point is equal to the original signal value, and that the curve has continuous first and second derivatives at each point. Similarly, a lower envelope is constructed based on the local minima. In practical applications, when the signal contains 500 sampling points, including 23 local maxima and 22 local minima, the cubic spline interpolation algorithm can generate smooth upper and lower envelopes, effectively reflecting the overall trend of the signal.

[0058] The mean values ​​of the upper and lower envelopes at each sampling point are calculated to obtain the mean signal. This mean signal reflects the local mean level of the original vibration signal and represents the slowly changing components of the signal. Subtracting the mean signal from the original digital vibration signal yields candidate intrinsic mode components (IMCs). These candidate components are then verified to check if they meet the definition conditions of an IMC: the total number of extrema and zero-crossings in the candidate component is counted. When the total number of extrema and zero-crossings are equal to or differ by 1, the candidate component is confirmed as a valid IMC. In practical cases, when a candidate component contains 45 extrema and 44 zero-crossings, it meets the conditions of an IMC and is confirmed as an IMC.

[0059] For the confirmed intrinsic mode components (IMCs), a Hilbert transform is performed. Essentially, the Hilbert transform rotates the phase of the original signal by 90 degrees, generating signal components orthogonal to the original signal. The FPGA implements a Hilbert transform algorithm based on the Fast Fourier Transform (FFT). This algorithm first performs an FFT on the IMCs, then multiplies the positive frequency components by a complex unit *i* and the negative frequency components by a complex unit *-i* in the frequency domain, and finally performs an inverse FFT to obtain the Hilbert-transformed signal. Using the IMCs as the real part and the Hilbert-transformed signal as the imaginary part, a complex analytic signal is constructed. Based on this analytic signal, the signal envelope at each time point is calculated, i.e., the magnitude of the analytic signal, to obtain the instantaneous amplitude. Simultaneously, the phase angle of the analytic signal is calculated, and the time derivative of the phase angle sequence is obtained to obtain the instantaneous frequency. In a specific experimental case, when the IMCs exhibit a frequency modulation characteristic with a gradually increasing frequency over a specific time period, the calculated instantaneous frequency accurately reflects this trend, gradually increasing from an initial 25 Hz to 75 Hz.

[0060] To construct a high-resolution time-frequency characteristic map, the local rates of change of instantaneous amplitude and instantaneous frequency are calculated. For instantaneous amplitude, the relative change in amplitude between adjacent time points is calculated; for instantaneous frequency, the frequency change between adjacent time points is calculated. These local rates of change reflect the rate of change of the signal's local characteristics in the time-frequency plane. Based on these rates of change, a narrower window is used in high-rate-of-change regions to preserve detailed features, while a wider window is used in low-rate-of-change regions to improve energy concentration. In practical applications, the window width is typically dynamically adjusted by ±50% based on the local rate of change, from a basic window width.

[0061] A defined adaptive window is used to locally smooth the time-frequency energy distribution of the intrinsic mode components. A weighted averaging method is employed, averaging the energy values ​​of each point and its neighborhood on the time-frequency plane, with the weights inversely proportional to the distance from the point to the center point. This smoothing effectively suppresses random noise and artifacts in the time-frequency distribution, improving the recognizability of the true energy patterns. After smoothing, the Laplace operator calculates the approximate second derivative value at each point on the time-frequency plane, reflecting the local curvature of the energy distribution. The original energy distribution is combined with the Laplace operator calculation results in a specific ratio to enhance the contrast between the energy concentration region and the surrounding region, making the time-frequency characteristics more prominent.

[0062] Through the above processing steps, the FPGA successfully constructed a Hilbert-Huang time-frequency characteristic spectrum. This spectrum uses time and frequency as coordinate axes and energy density as the value, intuitively displaying the changes of each frequency component in the vibration signal over time. This spectrum has high time-frequency resolution and good energy concentration, accurately reflecting the temporal location, frequency characteristics, and energy distribution of vibration events, providing reliable characteristic data for subsequent vibration localization and suppression.

[0063] In one optional implementation, the fiber vibration propagation delay is determined by calculating the energy peak time difference in the Hilbert-Huang time-frequency characteristic spectrum. Based on the fiber vibration propagation delay, the spatial location coordinates of the fiber vibration source are calculated using the hyperbolic positioning method, including:

[0064] The time-domain energy curve is obtained by integrating the Hilbert-Huang time-frequency feature map along the time dimension. The local mean and local standard deviation of the time-domain energy curve are calculated and combined to obtain an adaptive threshold. Multiple energy peak moments are extracted from the time-domain energy curve using the adaptive threshold. The time difference between the energy peak moments is calculated to obtain the fiber vibration propagation delay.

[0065] A predetermined number of sensing units are selected on the distributed optical fiber. The sensing unit that detects the vibration signal earliest is set as the focal sensing unit. Other sensing units are paired with the focal sensing unit in sequence to form multiple sets of sensing unit pairs. Based on the vibration propagation delay of the optical fiber, a hyperbola is constructed with the two sensing units of each set of sensing unit pairs as the focal points.

[0066] The intersection points of the multiple sets of sensing units with their respective hyperbolas are determined by solving the system of equations. A search region is constructed with the intersection points as the center. The spatial coordinates of the optical fiber vibration source are obtained by solving the minimum mean square error criterion within the search region.

[0067] The FPGA performs integration along the time dimension on the Hilbert-Huang time-frequency characteristic spectrum, transforming the two-dimensional time-frequency energy distribution into a one-dimensional time-domain energy curve. During integration, for each time point, the energy values ​​of all frequency components at that time point are summed to generate the energy value for that time point, thus forming a complete time-domain energy curve. This integration operation is essentially a projection of the signal energy onto the time axis, highlighting the temporal characteristics of the vibration signal. In practical applications, when the original Hilbert-Huang time-frequency characteristic spectrum has a resolution of 1000×500 points, the time-domain energy curve obtained after integration contains 1000 sampling points, each point representing the total energy level at the corresponding time.

[0068] The sliding window technique is used to calculate the local statistical characteristics of the time-domain energy curve. The window size is typically set to 5% to 10% of the total curve length. For each time point, the mean of all energy values ​​within the window centered at that point is calculated as the local mean; simultaneously, the standard deviation of the energy values ​​within the window is calculated as the local standard deviation. The weighted sum of the local mean and the local standard deviation is used as the adaptive threshold for that time point. The weighting coefficients are set according to the system's desired detection sensitivity; typically, the weight of the local mean is 1.0, and the weight of the local standard deviation is 2.5. Using this adaptive thresholding method, energy peaks can still be accurately identified even when the background noise level varies. In a test case, when the background noise varied by 50% in different fiber optic segments, this method still maintained a peak detection accuracy of over 95%.

[0069] When the energy value at a point on the curve exceeds the corresponding adaptive threshold and is greater than the energy values ​​of its adjacent points, that point is marked as an energy peak point. The timestamps of all identified energy peak points are recorded and arranged in chronological order. The time difference between adjacent peak times is calculated, which is the fiber optic vibration propagation delay. This delay data directly reflects the propagation characteristics of the vibration signal in the optical fiber and is the foundational data for subsequent positioning calculations. In practical applications, a single vibration event typically generates multiple delay values. For example, when there are 10 sensing units distributed on the optical fiber, 9 valid delay measurements will be generated.

[0070] A predetermined number of sensor units are selected on the distributed optical fiber to participate in the positioning calculation. The number of sensor units is usually determined based on the fiber length and the required positioning accuracy, with a typical configuration of 2-5 sensor units per kilometer of fiber. The sensor unit that detects the vibration signal earliest is set as the focal sensor unit, which is usually the sensor unit closest to the vibration source. Other sensor units are paired with the focal sensor unit in sequence to form multiple sensor unit pairs. Each sensor unit pair contains the focal sensor unit and another sensor unit. The vibration signal detection time difference between the two pairs is a key parameter for constructing the positioning model. In a certain actual deployment, 8 sensor units were evenly distributed along a 1500-meter-long optical fiber. When a vibration event occurred, the sensor unit numbered 3 detected the signal earliest and was set as the focal sensor unit. It was then paired with the other 7 sensor units to form 7 sensor unit pairs.

[0071] Based on fiber optic vibration propagation delay data, a hyperbola is constructed for each pair of sensing units. A hyperbola is a quadratic curve where the distance difference from any point to the two foci is constant. In fiber optic vibration localization, the two sensing units in a pair serve as the two foci of the hyperbola. The characteristic constant of the hyperbola is the time difference between the vibration signals detected by the two sensing units multiplied by the vibration propagation velocity in the fiber. The vibration propagation velocity is a key parameter of the system and is usually determined through pre-calibration; a typical value in silica fiber is approximately 5000 m / s. The trajectory of each hyperbola on a two-dimensional plane is calculated using the hyperbolic equation. In practical applications, when the distance between the two sensing units is 300 meters and the detection time difference is 0.05 seconds, the constructed hyperbola represents the set of all locations where the vibration source is located, and the distance difference from these locations to the two sensing units is exactly 250 meters.

[0072] After constructing multiple hyperbolas, these hyperbolas are solved simultaneously to find their intersection. Theoretically, if all measurements are accurate, these hyperbolas should intersect at a single point, i.e., the true location of the vibration source. However, due to errors in actual measurements, these hyperbolas usually do not intersect precisely at one point, but rather form a cluster of intersection points. The distribution area of ​​these intersection points is determined, and a search area is constructed centered on the location with the highest intersection point density. The size of the search area depends on the expected positioning accuracy of the system and computational resources; a typical configuration is a square area with sides of 10-50 meters. In one experiment, five hyperbolas were constructed, with their intersection points distributed within an area with a diameter of approximately 15 meters. A 30-meter × 30-meter square search area was constructed centered on the location with the highest intersection point density.

[0073] A uniformly distributed grid of candidate points is set within the search area, with a grid density typically of 1-4 points per square meter. For each candidate point, the theoretical time difference from that point to each sensor unit is calculated and compared with the actual measured time difference to calculate the mean square error. Specifically, for each pair of sensor units, the distance difference from the candidate point to the two sensor units is divided by the vibration propagation velocity to obtain the theoretical time difference; the theoretical time difference is subtracted from the actual measured time difference to obtain the error; the sum of the squares of the corresponding errors for all sensor unit pairs is calculated to obtain the total mean square error; all candidate points are iterated through to find the point with the smallest mean square error, and its coordinates are the spatial coordinates of the vibration source determined by the system. In a positioning test, 1600 candidate points were set within a 40m × 40m search area, and the mean square error value was determined to be 0.000025 seconds through minimum mean square error calculation. 2 The optimal point was determined, and its coordinates were identified as the location of the vibration source. Actual verification showed that the positioning error was less than 5 meters.

[0074] By combining the aforementioned technologies, the FPGA-based distributed fiber optic vibration localization system can efficiently and accurately calculate the spatial coordinates of fiber optic vibration sources. Utilizing the high-precision time information provided by the Hilbert-Huang time-frequency characteristic spectrum and combining it with the geometric advantages of the hyperbolic positioning method, precise localization of vibration sources in complex environments is achieved. This localization method not only possesses high precision but also exhibits excellent anti-interference capabilities and real-time performance, meeting various requirements in practical application scenarios.

[0075] In one optional implementation, based on the spatial coordinates, multiple preset suppression nodes around the optical fiber vibration source are activated. The FPGA calculates the propagation path and attenuation period of the vibration energy in space according to the vibration propagation characteristics collected by each suppression node, including:

[0076] Based on the spatial coordinates, a spherical monitoring domain is constructed around the optical fiber vibration source. The Voronoi diagram partitioning algorithm is used to divide the spherical monitoring domain into multiple sub-regions. Activation signals are sent to the suppression nodes in each sub-region according to the constraints of maximizing coverage and minimizing overlap.

[0077] The vibration propagation characteristics are collected using the activated suppression node. The FPGA extracts the propagation path feature coefficients from the vibration propagation characteristics and uses a ray tracing algorithm to calculate the reflection coefficient and refraction coefficient of each propagation medium interface. The propagation path feature coefficients are set as fitness evaluation indicators, and the reflection coefficient and refraction coefficient are used as path constraints. An adaptive particle swarm optimization algorithm is constructed to calculate the propagation path of the vibration energy.

[0078] Based on the propagation path, the decay characteristics of the vibration energy over time are extracted to obtain the time decay factor, and the loss of the vibration energy during spatial propagation is calculated to obtain the spatial decay factor. A bidirectional long short-time memory algorithm is used to establish the dynamic response matrix of the time decay factor and the spatial decay factor. The periodic solution of the dynamic response matrix is ​​solved by the eigenvalue decomposition method to obtain the decay period of the vibration energy.

[0079] like Figure 2 As shown, the method includes:

[0080] After determining the location of the vibration source, a spherical monitoring domain is constructed based on its spatial coordinates. The radius of the monitoring domain is determined according to the estimated propagation range of the vibration energy, typically set to 20 to 100 meters. Within this spherical monitoring domain, a Voronoi diagram partitioning algorithm is applied to divide the entire monitoring space into multiple non-overlapping sub-regions. In the specific implementation, the FPGA uses the known locations of suppression nodes as the points for generating the Voronoi diagram. Each sub-region contains one suppression node and its surrounding spatial points. The distance from these spatial points to the suppression node is less than the distance to any other suppression node. The Voronoi diagram partitioning algorithm is implemented incrementally, starting from an initial triangulation, sequentially inserting each suppression node, and updating the partitioning results. In a practical application, when there are 16 suppression nodes distributed within the monitoring domain, the Voronoi diagram partitioning algorithm divides the monitoring domain into 16 irregular polyhedral sub-regions, each with an average volume of approximately 523 cubic meters.

[0081] Coverage is defined as the ratio of the union of the sensing ranges of all activated nodes to the volume of the entire monitoring domain, while overlap is defined as the ratio of the intersection to the union of the sensing ranges of multiple nodes. Based on the constraints of maximizing coverage and minimizing overlap, activation signals are sent to the suppressed nodes in each sub-region. A greedy algorithm selects the optimal combination of activated nodes, ensuring a monitoring domain coverage rate of over 95% while controlling the overlap of sensing ranges between nodes to below 20%. The FPGA sends activation commands to the selected suppressed nodes via an embedded wireless communication module. These commands include parameters such as node ID, sampling rate, and data upload cycle. In a practical deployment, 12 nodes were activated from 16 candidate nodes, achieving a monitoring domain coverage rate of 97.8% and an average overlap of 17.3% between nodes.

[0082] Each suppression node is equipped with a triaxial accelerometer and an acoustic sensor, with a sampling rate set to 1000Hz and 16-bit quantization. Once activated, the suppression node collects vibration propagation characteristic data, including the propagation velocity, amplitude attenuation rate, and phase change of the vibration wave in different directions. The data is transmitted in real-time to the FPGA processing module via a low-power wireless network. The FPGA extracts propagation path characteristic coefficients from these vibration propagation characteristics, including the ratio of propagation velocities in each direction, the directionality coefficient of amplitude attenuation, and the phase delay coefficient. The extraction process uses a combination of wavelet transform and spectral analysis to perform time-frequency decomposition on the original signal, identifying propagation characteristics from each frequency band. In a certain test scenario, the ratio of propagation velocities in the x, y, and z directions extracted from the collected vibration data was 1.0:0.85:0.72, indicating a significant difference in the propagation velocities in the horizontal and vertical directions.

[0083] Based on the extracted propagation characteristics, a ray tracing algorithm was used to calculate the reflection and refraction coefficients of the interfaces of various propagation media. The ray tracing algorithm simulates the propagation of vibration energy as light rays from the vibration source. When the vibration rays encounter interfaces of different propagation media, the reflection and refraction are calculated. The algorithm uses an iterative approach, discretizing the space into a 50cm×50cm×50cm grid. At each grid node, the direction and energy of the vibration rays are calculated. For the reflection coefficient calculation, the relationship between the incident angle and the reflection angle of the vibration wave, as well as the energy ratio of the reflected wave to the incident wave, are analyzed. For the refraction coefficient, the change in the propagation direction of the vibration wave in different media and the energy transfer efficiency are analyzed. In a test environment containing complex structures such as concrete walls and metal supports, the average reflection coefficient of the concrete interface was calculated to be 0.35, and the reflection coefficient of the metal support interface was 0.72, indicating that most of the vibration energy is reflected at the metal interface.

[0084] An adaptive particle swarm optimization (PSO) algorithm was constructed to calculate the propagation path of vibration energy. The PSO algorithm initialized with 100 particles, each representing a propagation path. Each particle contained multiple dimensional parameters representing the direction and length of each segment along the path. The characteristic coefficients of the propagation path were set as fitness evaluation indicators, while the reflection and refraction coefficients served as path constraints. During the algorithm's iterative optimization, particles updated their velocity and position based on their individual optimal position and the global optimal position. Simultaneously, the inertia weight, cognitive coefficient, and social coefficient were dynamically adjusted based on real-time acquired propagation characteristics. The fitness function was designed as the Euclidean distance between the measured propagation characteristics and the theoretical model's predicted values; a smaller distance indicated that the path more closely matched the actual propagation situation. After 50 iterations, the algorithm converged to the optimal solution, determining the main propagation paths of the vibration energy. In a certain experimental scenario, vibration energy was identified as propagating from the source point along three main paths: a direct path and two reflected paths, with energy distribution ratios of 65%, 25%, and 10%, respectively.

[0085] Based on the calculated propagation path, the attenuation characteristics of vibration energy over time are further extracted. FPGA analysis is used to analyze the change trend of vibration signal amplitude recorded at each suppression node over time. An exponential fitting method is employed to extract the time attenuation factor, which represents the proportion of vibration energy attenuation per unit time and is typically influenced by the material's internal damping characteristics. Simultaneously, the loss of vibration energy during spatial propagation is calculated to obtain the spatial attenuation factor. The spatial attenuation factor represents the attenuation law of vibration energy as the propagation distance increases and is typically related to the geometric diffusion and material absorption of the propagation medium. In actual measurements, when the vibration source is continuously excited, the observed time attenuation factor of vibration energy in a 15-meter concrete structure is 0.12 / second, indicating an energy attenuation of 12% per second; the spatial attenuation factor is 0.08 / meter, indicating an energy attenuation of 8% per meter of propagation distance.

[0086] The bidirectional Long Short-Term Memory (LSTM) network contains 100 hidden units. The input layer receives decay data in both time and space dimensions, processes temporal information through two hidden layers (forward and backward), and the output layer generates a dynamic response matrix. Network training employs stochastic gradient descent with a batch size of 32 and a learning rate of 0.001, reaching a stable state after 2000 training epochs. The trained network can accurately predict vibration energy levels at different times and spatial locations. A dynamic response matrix with temporal and spatial decay factors is established using the bidirectional LSM algorithm. The dimension of the dynamic response matrix is ​​the number of sampling points on the time axis multiplied by the number of spatial grids, and each matrix element represents the vibration energy value at a specific spatiotemporal point. In a test case containing 10 seconds of data and 1000 spatial points, a 10000×10000-dimensional dynamic response matrix is ​​generated, accurately describing the spatiotemporal distribution of vibration energy.

[0087] Eigenvalue decomposition (EVD) decomposes a high-dimensional matrix into a combination of eigenvalues ​​and eigenvectors. Eigenvalues ​​represent the system's natural frequencies, and eigenvectors represent the corresponding vibration modes. The FPGA performs Singular Value Decomposition on the dynamic response matrix, extracting the top 10 largest singular values ​​and their corresponding left and right singular vectors. By analyzing the time components of these singular vectors, the system's periodic characteristics are identified. The periodic solution is typically represented by the sinusoidal component in the eigenvector, and its period is the decay period of the vibration energy. In analyzing the vibration propagation characteristics of a building structure, three main decay periods were identified: 0.28 seconds, 0.56 seconds, and 1.12 seconds, corresponding to the reciprocals of the structure's first, second, and third natural frequencies, respectively. This decay period information provides important reference for subsequent vibration suppression strategies.

[0088] Using the methods described above, the FPGA-based distributed fiber optic vibration localization and suppression system comprehensively analyzed the propagation path and attenuation period of vibration energy in space, laying a theoretical foundation for achieving precise vibration suppression. It integrates multiple techniques such as spatial partitioning, propagation characteristic analysis, path optimization, and dynamic response modeling, enabling accurate characterization of vibration propagation laws in complex environments.

[0089] In one optional implementation, based on the propagation path and the attenuation period, the FPGA calculates the optimal driving timing and optimal driving parameters for each suppression node, and uses the optimal driving timing and optimal driving parameters to collaboratively control multiple suppression nodes to generate interference waveforms, including:

[0090] A phase difference constraint is established between adjacent suppression nodes based on the propagation path, and a timing constraint is established between adjacent suppression nodes based on the decay period. The phase difference constraint and the timing constraint are combined to form an optimization objective. The optimal driving timing of each suppression node is calculated by minimizing the optimization objective.

[0091] The driving amplitude of each suppression node is calculated based on the time decay factor and the spatial decay factor. The deviation between the actual energy and the target energy at each suppression node is calculated. The compensation coefficient is determined based on the deviation. The driving amplitude is dynamically corrected by iteratively adjusting the compensation coefficient to obtain the optimal driving parameters of each suppression node.

[0092] The FPGA allocates the optimal driving parameters to each of the suppression nodes, establishes a global clock synchronization triggering mechanism based on the optimal driving timing, calculates the superposition coefficient of adjacent waveforms based on the phase difference of each of the suppression nodes, determines the waveform superposition method according to the superposition coefficient and the global clock synchronization triggering mechanism, and uses the waveform superposition method to collaboratively control each of the suppression nodes to generate the interference waveform.

[0093] Based on the calculated propagation path, phase difference constraints describe the phase relationship of the vibration wave between different suppression nodes, ensuring that the suppressed waveforms generated by each node can form an effective interference mode. FPGA analyzes the propagation time of the vibration wave from the source point to each suppression node, converting the propagation time difference between adjacent suppression nodes into a phase difference. The phase difference calculation considers the frequency characteristics of the vibration wave, establishing phase difference constraints between adjacent suppression nodes. For a vibration signal with a dominant frequency of 50Hz, the phase difference between two suppression nodes 5 meters apart is approximately 57 degrees. Simultaneously, timing constraints are established between adjacent suppression nodes based on the decay period. These timing constraints ensure that each suppression node is activated in a specific time sequence, generating a waveform opposite to the original vibration wave. The design of the timing constraints considers the decay period of the vibration energy; for example, when the detected vibration decay period is 0.3 seconds, the activation time difference between adjacent suppression nodes should be controlled within 0.1 seconds to ensure the suppression effect.

[0094] The optimization objective is composed of phase difference constraints and timing constraints. The FPGA minimizes this objective using the Lagrange multiplier method. The optimization process uses the driving timing of the suppression nodes as independent variables and the weighted sum of phase difference and timing constraints as the objective function. The weight coefficients are dynamically adjusted according to the vibration characteristics. A typical configuration is a phase difference constraint weight of 0.7 and a timing constraint weight of 0.3. The FPGA executes the gradient descent method with an initial step size of 0.05, gradually adjusting the driving timing of each suppression node until the objective function converges to a local minimum. In practical applications, when 12 suppression nodes are deployed, the optimization algorithm converges after 50 iterations. The calculated optimal driving timing shows that the time difference between the first and last triggered suppression nodes is 0.085 seconds, and the triggering order of each node is highly consistent with the propagation path of the vibration wave.

[0095] The FPGA uses the vibration source location and energy as a reference point to calculate the theoretical amplitude of the vibration energy when it reaches each suppression node along the propagation path. The calculation process takes into account the geometric diffusion loss and material absorption loss of the vibration during propagation. For a suppression node 15 meters away from the vibration source, when the time decay factor is 0.15 / second and the spatial decay factor is 0.1 / meter, the vibration energy at this node is theoretically about 30% of the energy at the source point. The FPGA calculates the deviation between the actual energy and the target energy at each suppression node. The actual energy is obtained by measuring the sensor on the suppression node. The target energy is the energy level that should be at this location under ideal suppression conditions, which is usually set to 1.2 times the background noise level.

[0096] The compensation coefficient reflects the difference between the theoretical model and the actual environment, and is used to correct the driving amplitude. The FPGA adopts an adaptive control strategy, with the initial compensation coefficient set to 1.0. It is continuously optimized through an iterative adjustment process. In each iteration, the FPGA measures the suppression effect, calculates the residual vibration energy, and adjusts the compensation coefficient based on the residual energy. The adjustment formula is the current compensation coefficient plus the difference between the residual energy and the target energy multiplied by the learning rate, typically 0.2. Through 5-10 iterations, the compensation coefficient can be optimized to a stable value. For example, in a complex structural environment, the final determined compensation coefficient range is 0.85 to 1.35, reflecting the differences in vibration propagation characteristics at different locations. By dynamically correcting the driving amplitude through the compensation coefficient, the optimal driving parameters for each suppression node are obtained.

[0097] The FPGA allocates the calculated optimal driving parameters to each suppression node. This allocation process is implemented via a high-speed serial communication interface, using time-division multiple access (TDMA) to send control commands to each node. These commands include parameters such as driving amplitude, frequency, phase, and trigger timestamp. The data packet size is 64 bytes, and the transmission rate is 10 Mbps. To ensure accurate command execution, cyclic redundancy check (CRC) is used for data integrity verification, and an error retransmission mechanism guarantees 99.99% communication reliability. In one test, the FPGA completed parameter allocation to 12 suppression nodes within 50 milliseconds, achieving a 100% communication success rate.

[0098] The FPGA-based distributed fiber optic vibration localization and suppression system achieves precise vibration suppression using the aforementioned method. Multiple suppression nodes work collaboratively to generate precise interference waveforms, directionally canceling the original vibration in both space and time. Key technologies include propagation-characteristic-based drive timing optimization, energy-balanced drive parameter calculation, a high-precision clock synchronization mechanism, and an intelligent waveform superposition strategy. The integrated application of these technologies enables the system to address vibration problems in complex environments, providing effective solutions for fields such as precision instrument protection and structural vibration control.

[0099] In one optional implementation, a global clock synchronization triggering mechanism is established based on the optimal driving timing, and the superposition coefficient of adjacent waveforms is calculated based on the phase difference of each suppression node. The waveform superposition method is determined according to the superposition coefficient and the global clock synchronization triggering mechanism, including:

[0100] Obtain the system reference clock signal, system reference clock period, and clock compensation delay constant from the FPGA;

[0101] The optimal driving timing is divided by the system reference clock period and superimposed on the clock compensation delay constant to obtain the trigger count value of each suppression node. The system reference clock signal is multiplied by a phase-locked loop to obtain a synchronization clock signal. The synchronization clock signal and the trigger count value are input to the trigger timer of each suppression node to generate the global clock synchronization trigger mechanism.

[0102] The phase difference and distance between adjacent suppression nodes are obtained, and the cosine function value of the phase difference is multiplied by the exponential decay function value of the distance to obtain the superposition coefficient between the adjacent waveforms; the amplitude correction factor of the adjacent waveforms is calculated based on the energy value and energy decay rate of each suppression node.

[0103] Based on the global clock synchronization triggering mechanism, the superposition timing of adjacent waveforms is determined by the superposition coefficient, the waveform amplitude of each suppression node is adjusted according to the amplitude correction factor, and the waveform superposition mode is obtained through the superposition timing and the waveform amplitude.

[0104] The FPGA provides a system reference clock signal with a frequency of 100MHz and a clock stability better than ±1ppm. The system reference clock period is 10 nanoseconds, and the clock compensation delay constant is set to 1000 clock cycles to compensate for communication and processing delays. The optimal drive timing is divided by the system reference clock period and added to the clock compensation delay constant to obtain the trigger count value of each suppression node. For example, when the optimal drive timing of a suppression node is 0.025 seconds, its trigger count value is 2,500,000 + 1,000 = 2,501,000 clock cycles. The FPGA's built-in phase-locked loop multiplies the system reference clock signal by 4 to obtain a 400MHz synchronous clock signal, improving timing accuracy. The synchronous clock signal and the trigger count value are input to the trigger timer of each suppression node to generate a global clock synchronization trigger mechanism. This mechanism ensures that each suppression node is activated at a precise time, with a timing accuracy better than 1 microsecond.

[0105] The FPGA acquires the phase difference and distance information between adjacent suppression nodes. Between adjacent suppression nodes A and B, when the measured phase difference is 65 degrees and the distance is 7.5 meters, the system calculates the cosine function value of the phase difference as 0.423. The exponential decay function value of the distance is calculated in negative exponential form. When the decay coefficient is 0.1 / meter, the exponential decay value corresponding to the 7.5-meter distance is 0.472. Multiplying the cosine function value of the phase difference with the exponential decay function value of the distance yields the superposition coefficient of 0.2 between adjacent waveforms. This superposition coefficient represents the degree of influence of the suppression waveform generated by node B on the vibration at the location of node A.

[0106] Based on the energy value and energy decay rate of each suppression node, the FPGA calculates the amplitude correction factor for adjacent waveforms. The amplitude correction factor takes into account the attenuation of vibration energy during propagation between nodes, as well as the energy output capability of each node itself. For a suppression node with an energy value of 200mJ and an energy decay rate of 18% / meter, when there is another node located at 4.5 meters within its influence range, the amplitude correction factor is calculated to be 0.78, indicating that the theoretical amplitude needs to be reduced by 22% to avoid over-suppression. The FPGA performs similar calculations on all adjacent node pairs in the network to generate a complete amplitude correction matrix.

[0107] Based on a global clock synchronization triggering mechanism, the FPGA uses a superposition coefficient to determine the superposition timing of adjacent waveforms. This superposition timing describes the temporal sequence and overlap of waveforms generated by different suppression nodes. The relative trigger times of the waveforms are adjusted according to the superposition coefficient; a larger superposition coefficient results in a smaller trigger time difference between adjacent waveforms, enhancing the superposition effect. For adjacent node pairs with a superposition coefficient of 0.2, their trigger time difference is set to 80% of the theoretical propagation time. That is, when the theoretical propagation time is 15 milliseconds, the actual trigger time difference is 12 milliseconds. Simultaneously, the waveform amplitude of each suppression node is adjusted according to an amplitude correction factor. This adjustment process considers the mutual influence between nodes to ensure that the total energy after superposition achieves the optimal suppression effect. In a certain test scenario, the originally calculated suppression node drive amplitude of 1.25A was adjusted to 0.975A using a correction factor of 0.78, avoiding vibration rebound caused by over-suppression.

[0108] By superimposing timing and waveform amplitude, the FPGA determines the waveform superposition method. This method includes information such as the triggering sequence, phase relationship, and energy distribution of each suppression node, forming a complete suppression strategy. A predictive-correction control mode is employed, predicting future vibration states based on a vibration propagation model and generating suppression waveforms in advance. Simultaneously, real-time feedback data corrects prediction errors, dynamically adjusting the suppression strategy. The FPGA converts the waveform superposition method into specific drive signals, which are then output to the actuators of each suppression node via a digital-to-analog converter.

[0109] This invention relates to an FPGA-based distributed optical fiber vibration localization and suppression system, the system comprising:

[0110] The first unit is used to receive the Raman scattered light signal returned by the sensing unit on the distributed optical fiber, and obtain the digital vibration signal through analog-to-digital conversion;

[0111] The second unit is used by the FPGA to perform empirical mode decomposition on the digital vibration signal to obtain intrinsic mode components, and to extract the instantaneous features of the intrinsic mode components through Hilbert transform. Based on the instantaneous features, a Hilbert-Huang time-frequency feature spectrum is constructed. The fiber optic vibration propagation delay is determined by calculating the energy peak time difference in the Hilbert-Huang time-frequency feature spectrum. Based on the fiber optic vibration propagation delay, the spatial coordinates of the fiber optic vibration source are calculated using the hyperbolic positioning method.

[0112] The third unit is used to activate multiple preset suppression nodes around the optical fiber vibration source based on the spatial location coordinates. The FPGA calculates the propagation path and attenuation period of the vibration energy in space according to the vibration propagation characteristics collected by each suppression node. Based on the propagation path and the attenuation period, the FPGA calculates the optimal driving timing and optimal driving parameters of each suppression node. The optimal driving timing and optimal driving parameters are used to coordinately control the multiple suppression nodes to generate interference waveforms, block the propagation path of the vibration energy, and suppress the optical fiber vibration in the target area.

[0113] A third aspect of the present invention provides an electronic device, comprising:

[0114] processor;

[0115] Memory used to store processor-executable instructions;

[0116] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0117] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0118] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for distributed optical fiber vibration location and suppression based on FPGA, characterized in that, The method comprises: receiving a Raman scattered light signal returned by a sensing unit on a distributed optical fiber, and obtaining a digital vibration signal through analog-to-digital conversion; performing empirical mode decomposition on the digital vibration signal by an FPGA to obtain an intrinsic mode component, extracting an instantaneous feature of the intrinsic mode component through Hilbert transform, and constructing a Hilbert-Huang time-frequency feature map based on the instantaneous feature; determining an optical fiber vibration propagation time delay by calculating a time difference of energy peaks in the Hilbert-Huang time-frequency feature map, and calculating spatial position coordinates of an optical fiber vibration source by using a hyperbolic positioning method based on the optical fiber vibration propagation time delay, comprising: integrating the Hilbert-Huang time-frequency feature map along a time dimension to obtain a time domain energy curve, calculating a local mean and a local standard deviation of the time domain energy curve to obtain an adaptive threshold, extracting a plurality of energy peak times from the time domain energy curve by using the adaptive threshold, and calculating a time difference between the energy peak times to obtain the optical fiber vibration propagation time delay; selecting a preset number of sensing units on the distributed optical fiber, setting a sensing unit that detects a vibration signal earliest as a focal point sensing unit, and sequentially pairing the other sensing units with the focal point sensing unit to form a plurality of sensing unit pairs; based on the optical fiber vibration propagation time delay, constructing a hyperbola with the two sensing units of each sensing unit pair as focal points; simultaneously solving a plurality of hyperbolas corresponding to the plurality of sensing unit pairs to determine the intersection positions of the plurality of hyperbolas, and constructing a search region with the intersection positions as the center, and solving the spatial position coordinates of the optical fiber vibration source in the search region by using a least mean square error criterion; based on the spatial position coordinates, activating a plurality of suppression nodes around the optical fiber vibration source, the FPGA calculating a propagation path and a decay period of vibration energy in space according to the vibration propagation characteristics collected by each suppression node; based on the propagation path and the decay period, the FPGA calculates an optimal driving time sequence and an optimal driving parameter of each suppression node, and uses the optimal driving time sequence and the optimal driving parameter to cooperatively control a plurality of suppression nodes to generate an interference waveform, block the propagation path of the vibration energy, and suppress the optical fiber vibration in a target region.

2. The method of claim 1, wherein, The FPGA performs empirical mode decomposition on the digital vibration signal to obtain an intrinsic mode component, and extracts an instantaneous feature of the intrinsic mode component through Hilbert transform, and constructs a Hilbert-Huang time-frequency feature map based on the instantaneous feature, comprising: the FPGA identifies local extreme points in the digital vibration signal, constructs an upper envelope line and a lower envelope line based on the local extreme points, and calculates a mean value of the upper envelope line and the lower envelope line to obtain a mean value signal; subtracts the mean value signal from the digital vibration signal to obtain a candidate intrinsic mode component, and judges whether the number of extreme points and the number of zero-crossing points of the candidate intrinsic mode component are equal or differ by 1, and if so, determines that it is the intrinsic mode component; performing the Hilbert transform on the intrinsic modal component to obtain a Hilbert transform signal, constructing the intrinsic modal component and the Hilbert transform signal into an analytic signal in complex form, calculating a signal envelope based on the analytic signal to obtain an instantaneous amplitude, calculating a phase angle and a time derivative to obtain an instantaneous frequency; calculating local variation rates of the instantaneous amplitude and the instantaneous frequency, determining a width of an adaptive window on a time-frequency plane according to the local variation rates, performing local contrast enhancement based on a Laplacian operator on time-frequency energy distribution of the intrinsic modal component after local smoothing by using the adaptive window, and obtaining the Hilbert-Huang time-frequency feature map.

3. The method of claim 1, wherein, based on the spatial position coordinates, activating a plurality of suppression nodes pre-set around the optical fiber vibration source, and the FPGA calculating a propagation path and an attenuation period of vibration energy in space according to vibration propagation characteristics collected by each suppression node; based on the spatial position coordinates, constructing a spherical monitoring domain around the optical fiber vibration source, dividing the spherical monitoring domain into a plurality of sub-regions by using a Voronoi diagram division algorithm, and sending an activation signal to the suppression nodes in each sub-region according to constraint conditions of maximum coverage and minimum overlap; collecting the vibration propagation characteristics by using the activated suppression nodes, extracting propagation path feature coefficients from the vibration propagation characteristics by the FPGA, and calculating reflection coefficients and refraction coefficients of each propagation medium interface by using a ray tracing algorithm; setting the propagation path feature coefficients as fitness evaluation indexes, setting the reflection coefficients and the refraction coefficients as path constraint conditions, and constructing an adaptive particle swarm optimization algorithm to calculate the propagation path of the vibration energy. based on the propagation path, extracting a time attenuation factor of the vibration energy with time to obtain an attenuation period, calculating a spatial attenuation factor of the vibration energy in the spatial propagation process, establishing a dynamic response matrix of the time attenuation factor and the spatial attenuation factor by using a bidirectional long short-term memory algorithm, and solving a periodic solution of the dynamic response matrix by eigenvalue decomposition to obtain the attenuation period of the vibration energy.

4. The method of claim 1, wherein, based on the propagation path and the attenuation period, the FPGA calculates the optimal driving time sequence and the optimal driving parameters of each suppression node, and utilizes the optimal driving time sequence and the optimal driving parameters to cooperatively control a plurality of suppression nodes to generate interference waveforms, including: based on the propagation path, establishing a phase difference constraint between adjacent suppression nodes, based on the attenuation period, establishing a time sequence constraint between adjacent suppression nodes, combining the phase difference constraint and the time sequence constraint into an optimization target, and calculating the optimal driving time sequence of each suppression node by minimizing the optimization target; calculating the driving amplitude of each suppression node according to the time attenuation factor and the spatial attenuation factor, calculating the deviation of the actual energy from the target energy at each suppression node, determining a compensation coefficient based on the deviation, dynamically modifying the driving amplitude by iteratively adjusting the compensation coefficient, and obtaining the optimal driving parameters of each suppression node; The FPGA assigns the optimal driving parameters to each of the suppression nodes, establishes a global clock synchronization trigger mechanism based on the optimal driving timing, calculates superposition coefficients of adjacent waveforms based on phase differences of each of the suppression nodes, determines a waveform superposition mode according to the superposition coefficients and the global clock synchronization trigger mechanism, and controls each of the suppression nodes to generate the interference waveform by using the waveform superposition mode.

5. The method of claim 4, wherein, The global clock synchronization trigger mechanism is established based on the optimal driving timing, the superposition coefficients of adjacent waveforms are calculated based on phase differences of each of the suppression nodes, and the waveform superposition mode is determined according to the superposition coefficients and the global clock synchronization trigger mechanism, including: obtaining a system reference clock signal, a system reference clock period and a clock compensation delay constant from the FPGA; dividing the optimal driving timing by the system reference clock period and superimposing the clock compensation delay constant to obtain a trigger count value of each of the suppression nodes, frequency multiplying the system reference clock signal by using a phase-locked loop to obtain a synchronization clock signal, inputting the synchronization clock signal and the trigger count value into a trigger timer of each of the suppression nodes to generate the global clock synchronization trigger mechanism; obtaining a phase difference and a distance between adjacent suppression nodes, multiplying a cosine function value of the phase difference by an exponential decay function value of the distance to obtain a superposition coefficient between the adjacent waveforms, and calculating an amplitude correction factor of the adjacent waveforms according to an energy value and an energy decay rate of each of the suppression nodes; based on the global clock synchronization trigger mechanism, the superposition coefficient is used to determine a superposition timing of adjacent waveforms, the amplitude correction factor is used to adjust a waveform amplitude of each of the suppression nodes, and the waveform superposition mode is obtained through the superposition timing and the waveform amplitude.

6. A distributed optical fiber vibration location and mitigation system based on FPGA for implementing the method according to any one of claims 1-5, characterized in that, including: a first unit configured to receive a Raman scattered light signal returned by a sensing unit on a distributed optical fiber, and obtain a digital vibration signal through analog-to-digital conversion; a second unit configured to perform empirical mode decomposition on the digital vibration signal by using an FPGA to obtain an intrinsic mode component, extract an instantaneous feature of the intrinsic mode component by using a Hilbert transform, construct a Hilbert-Huang time-frequency feature map based on the instantaneous feature, calculate an energy peak time difference in the Hilbert-Huang time-frequency feature map to determine an optical fiber vibration propagation time delay, and calculate spatial position coordinates of an optical fiber vibration source by using a hyperbolic positioning method based on the optical fiber vibration propagation time delay; a third unit configured to activate a plurality of suppression nodes around the optical fiber vibration source based on the spatial position coordinates, and calculate a propagation path and a decay period of vibration energy in space according to vibration propagation characteristics collected by each of the suppression nodes by using the FPGA; based on the propagation path and the decay period, the FPGA calculates optimal driving timing and optimal driving parameters of each of the suppression nodes, controls a plurality of suppression nodes to generate an interference waveform by using the optimal driving timing and the optimal driving parameters, blocks the propagation path of the vibration energy, and suppresses optical fiber vibration in a target area.

7. An electronic device, comprising: including: a processor; a memory for storing processor-executable instructions; The processor is configured to invoke the instructions stored in the memory to execute the method in any one of claims 1 to 5.

8. A computer-readable storage medium having stored thereon computer program instructions, wherein, The computer program instructions, when executed by the processor, implement the method in any one of claims 1 to 5.

Images