Distributed optical fiber acoustic wave sensing track safety monitoring system

Abstract

The application relates to the field of rail transit safety monitoring, and discloses a distributed optical fiber acoustic wave sensing rail safety monitoring system, which comprises a sensing and collecting module, a signal processing module, a feature extraction module, an event identification module, an event positioning module and an early warning management module; the sensing and collecting module captures a rail vibration signal based on a phase optical time domain reflection technology to generate time sequence data; the signal processing module performs noise suppression and modal decomposition on the collected vibration signal to extract an abnormal signal; the feature extraction module extracts time-frequency features and envelope features of the vibration signal; the event identification module classifies the types of rail abnormal events in combination with a deep learning model; the event positioning module determines the rail position of the abnormal event; and the early warning management module generates multistage early warning signals according to the identification result. The application has high precision, real-time performance, robustness and expansibility, and provides reliable guarantee for the operation safety of rail transit.

Description

Technical Field

[0001] This invention relates to the field of rail transit safety monitoring technology, specifically a distributed fiber optic acoustic wave sensing rail safety monitoring system. Background Technology

[0002] As a crucial component of modern urban transportation, rail transit's operational safety directly impacts the convenience and security of public travel. However, with the rapid expansion of rail transit networks, the complexity of the track operating environment has significantly increased, and abnormal events caused by track vibrations (such as track loosening, illegal construction, and foreign object intrusion) pose an increasingly prominent threat to track safety. Traditional track safety monitoring technologies primarily rely on fixed-point sensors or manual inspections. These methods suffer from poor real-time performance, limited monitoring range, and high deployment costs, making it difficult to meet the demands of modern rail transit for efficient safety monitoring.

[0003] While existing track safety monitoring systems have made some progress in certain areas, such as using accelerometers for vibration monitoring or identifying track intrusions through video surveillance, significant technical shortcomings remain. Firstly, point sensors have limited monitoring range, making it difficult to achieve full track coverage. Furthermore, their signal acquisition is significantly affected by environmental noise, compromising monitoring accuracy. Secondly, traditional video surveillance technology relies on image processing algorithms for identifying track vibration events. This method is poorly suited to low-light environments such as tunnels and curves, and its ability to monitor dynamic vibration signals is insufficient. In addition, existing systems often lack precise location and classification capabilities for abnormal events, failing to generate timely warnings and guide maintenance work. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a distributed fiber optic acoustic wave sensing track safety monitoring system, which enables high-precision real-time monitoring of abnormal vibration events along the track, intelligent identification and accurate location of event types, and generation of real-time early warning signals to ensure the safe operation of rail transit.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a distributed fiber optic acoustic wave sensing track safety monitoring system, comprising:

[0006] The sensing and acquisition module is used to capture vibration signals of optical fibers along the track based on phase light time-domain reflectometry and generate time-series vibration data.

[0007] The signal processing module is used to denoise the vibration data and decompose the abnormal vibration signal.

[0008] The feature extraction module is used to extract time-frequency features and envelope features from the abnormal vibration signal;

[0009] The event recognition module is used to classify track vibration events based on extracted features and identify event types.

[0010] The event location module is used to determine the specific location of an event on the track based on the time delay of the optical signal;

[0011] The early warning management module is used to generate early warning signals and visualize information based on the identified event type and location information.

[0012] Preferably, the sensing acquisition module includes:

[0013] A pulsed light source is used to inject high-power optical pulses along an optical fiber;

[0014] The signal acquisition unit is used to acquire backscattered light signals in the optical fiber.

[0015] The data conversion unit is used to convert the collected scattered light signals into track vibration data.

[0016] Preferably, the signal processing module includes:

[0017] Noise suppression unit, used to filter out background noise in vibration data;

[0018] The mode decomposition unit is used to decompose vibration signals through variational mode decomposition and extract abnormal vibration signals.

[0019] Preferably, the feature extraction module includes:

[0020] The frequency decomposition unit is used to perform multi-band decomposition of abnormal vibration signals using wavelet packet decomposition and extract the energy features corresponding to each frequency band.

[0021] The envelope analysis unit is used to calculate the instantaneous envelope of the abnormal vibration signal based on the Hilbert transform and extract the dynamic change characteristics of the vibration amplitude.

[0022] The feature integration unit is used to combine the extracted frequency band energy features and vibration amplitude dynamic change features into a feature vector, which is then transmitted to the event recognition module.

[0023] Preferably, the frequency decomposition unit extracts the frequency band energy features through the following steps:

[0024] The abnormal vibration signal was decomposed into multi-band signals using wavelet packet decomposition technology. Each band signal was calculated using a recursive formula.

[0025]

[0026]

[0027] in, For the first Layer, First Decomposed signal of each node, and These are the low-pass and high-pass coefficients of the decomposition filter, respectively. Represents discrete sampling points of a time series. Represents the time variable of the signal;

[0028] The energy characteristics of the decomposed signal for each frequency band are calculated using the following formula:

[0029]

[0030] in, For the first Energy characteristics of each frequency band Indicates the first Layer, First The signal decomposed at node i is at the... The value at each sampling point.

[0031] Preferably, the envelope analysis unit extracts the dynamic change characteristics of vibration amplitude through the following steps:

[0032] For abnormal mode signals Perform Hilbert transform to calculate the instantaneous signal:

[0033]

[0034] in, Represents modal signals The Hilbert transform result, Time-domain representation of abnormal vibration signals Represents the integral variable. Represents a time variable. It is a mathematical constant;

[0035] Use original signal and its Hilbert transform result Calculate the vibration envelope The formula for calculating the vibration envelope is:

[0036]

[0037] in, It represents the instantaneous envelope of the vibration signal.

[0038] Preferably, the event recognition module includes:

[0039] The feature input unit is used to receive the frequency features and time-domain features output by the feature extraction module and combine them into a multi-dimensional feature vector.

[0040] The deep learning classification unit classifies and identifies multidimensional feature vectors based on convolutional neural networks and long short-term memory networks.

[0041] The classification output unit is used to generate type labels and confidence scores for track vibration events based on the deep learning classification results, and then transmits the results to the event localization module.

[0042] Preferably, the event location module includes:

[0043] The time delay calculation unit is used to calculate the time delay difference of optical signal propagation caused by abnormal vibration events in the optical fiber.

[0044] The position calculation unit is used to calculate the specific location of the abnormal vibration event on the track based on the time delay difference and the refractive index of the optical fiber.

[0045] The event calibration unit is used to map the calculated location to the orbital geographic coordinate system, providing event location results for orbital points.

[0046] Preferably, the early warning management module includes:

[0047] An alarm generation unit is used to generate multi-level alarm signals based on event type and location information.

[0048] Visualization units are used to display the type, location, and severity of track events in real time;

[0049] Data storage unit, used to record historical event information, supporting querying and analysis.

[0050] This invention also provides a distributed fiber optic acoustic sensing method for track safety monitoring, comprising the following steps:

[0051] Vibration signals along the track were acquired using phase-optical time-domain reflectometry.

[0052] Noise suppression is performed on the collected vibration signals, and abnormal vibration signals are extracted through mode decomposition.

[0053] Frequency band energy features are extracted using wavelet packet decomposition, and the signal envelope is calculated based on Hilbert transform.

[0054] The extracted features are input into a deep learning model to classify and identify types of track anomaly events.

[0055] Calculate the specific location of orbital events based on optical signal delay difference;

[0056] Early warning signals are generated based on the event type and location.

[0057] This invention provides a distributed fiber optic acoustic wave sensing track safety monitoring system. It has the following beneficial effects:

[0058] 1. This invention uses phase-time reflectometry to collect optical fiber vibration signals along the track in real time. Through the high-sensitivity collaboration of pulse light source and photodetector, it can accurately capture weak vibration changes and avoid interference from external environmental noise, thereby ensuring high precision and high signal-to-noise ratio of vibration signal acquisition.

[0059] 2. This invention effectively removes background noise from track vibration signals and separates vibration components related to abnormal events through wavelet denoising and mode decomposition methods, ensuring that the processed signal has good representativeness and robustness. The signal processing module is designed to adapt to complex track operating environments and improve the accuracy of abnormal signal identification.

[0060] 3. The feature extraction module of this invention utilizes wavelet packet decomposition and Hilbert transform techniques to simultaneously acquire the frequency features and time-domain envelope features of the vibration signal, forming a comprehensive feature vector. This multi-dimensional feature information can accurately characterize the properties of abnormal track vibrations, providing high-quality input data for subsequent classification and recognition.

[0061] 4. The event recognition module of this invention employs a deep learning algorithm combining convolutional neural networks and long short-term memory networks, which can accurately classify track anomaly event types. The classification model can not only extract the spatial distribution pattern of vibration signals, but also capture dynamic changes in the time series, effectively improving the accuracy and robustness of track event recognition.

[0062] 5. This invention enables high-precision location of track vibration events through accurate calculation and geographic mapping of optical signal delay time. The event location module can quickly determine the specific abnormal points on the track and provide the track's geographic coordinate information, providing crucial support for rapid response and handling of abnormal events.

[0063] 6. The early warning management module of this invention generates real-time early warning signals and performs hierarchical processing based on event identification and location results. Early warning information is presented intuitively through a visual interface, showing the type, location, and severity of track anomalies, supporting parallel monitoring of multiple events and subsequent query analysis. This module's design significantly improves the response efficiency of track safety monitoring. Attached Figure Description

[0064] Figure 1 This is a schematic diagram of the system architecture of the present invention;

[0065] Figure 2 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0066] The technical solutions in 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.

[0067] Please see the appendix Figure 1 This invention provides a distributed fiber optic acoustic wave sensing track safety monitoring system, which aims to identify, locate, and warn of abnormal events along the track by real-time monitoring and analysis of track vibration signals.

[0068] like Figure 1 As shown, the distributed fiber optic acoustic wave sensing track safety monitoring system of the present invention may include a sensor acquisition module, a signal processing module, a feature extraction module, an event recognition module, an event location module, and an early warning management module. The following provides a detailed description of each module of the system.

[0069] In this embodiment, the sensing and acquisition module is used to capture the vibration signal of the optical fiber along the track in real time based on phase optical time domain reflectometry (Φ-OTDR) and convert the captured vibration signal into analyzable time-series vibration data. This module monitors the intensity changes of the backscattered light signal in the optical fiber to sense the phase changes of the light signal caused by vibration, thereby obtaining track vibration information.

[0070] As an alternative, the sensing acquisition module includes the following functional units: a pulse light source unit, an optical signal receiving unit, a data conversion unit, and a timing data output unit. These units are connected by circuits to realize signal transmission and processing.

[0071] Specifically, the pulsed light source unit is used to inject high-power pulsed light into the optical fiber deployed along the track. Exemplarily, this light source can be a pulsed laser with a wavelength range of 1.3 μm to 1.55 μm to ensure high sensitivity and low loss characteristics in the optical fiber communication band. In one possible implementation, the pulse width can be adjusted between 10 ns and 100 ns, depending on the spatial resolution requirements of the system.

[0072] The optical signal receiving unit receives backscattered optical signals caused by vibrations in the optical fiber. It should be noted that the intensity of the backscattered optical signal is directly related to the phase change caused by vibrations in the fiber. The backscattered signal is converted into an electrical signal by a photodetector and further transmitted to the data processing module for analysis. As an example, a PIN photodiode or avalanche photodiode can be used as the photodetector to improve signal detection sensitivity.

[0073] In one possible implementation, the phase change of the vibration signal It can be expressed by the following formula:

[0074]

[0075] in:

[0076] Indicates the wavelength of the injected light wave;

[0077] Indicates the refractive index of the optical fiber;

[0078] Indicates the location of the optical fiber The length change caused by vibration.

[0079] The data conversion unit digitizes the electrical signals output by the photodetector and generates time-series data corresponding to the track vibration signals. Specifically, the data conversion unit includes an analog-to-digital converter (ADC) and a data buffer, used to convert continuous analog signals into discrete digital signals, and to store and transmit the signals. In some embodiments, the data sampling frequency can be set from 10 kHz to 1 MHz to meet the acquisition requirements of different track vibration signals.

[0080] For example, the time-series data output unit outputs the processed vibration signal data to the signal processing module. The output data may include the time, space, and intensity information of the vibration signal, and the format may be a time-series point series or a spectrum data. It should be noted that, in order to improve the accuracy of the data, the time-series data can be initially suppressed by filtering algorithms to suppress high-frequency noise.

[0081] Understandably, the sensor acquisition module can also include various extended functions. For example:

[0082] In some embodiments, to further improve spatial resolution, the optical fiber can be segmented and labeled, with each segment ranging in length from 10m to 50m.

[0083] In another possible implementation, to ensure the stability of signal transmission, the system can add optical amplifiers at key nodes of the optical fiber to compensate for optical losses in the signal.

[0084] It is important to emphasize that the sensor acquisition module is designed based on the technical solution of this invention, ensuring high-sensitivity capture of track vibration signals. Through the collaboration of a high-power pulsed light source and a high-precision photodetector, this module achieves accurate capture of fiber optic vibration signals along the track, providing high-quality vibration time-series data for subsequent modules.

[0085] In this embodiment, the signal processing module is used to denoise the vibration time-series data provided by the sensing acquisition module and decompose the abnormal vibration signal to provide a high-quality input signal for the subsequent feature extraction module. Through precise noise suppression and signal decomposition methods, this module can effectively remove background noise and irrelevant signals from the track vibration signal and extract the vibration features corresponding to the abnormal event.

[0086] Alternatively, the signal processing module includes a noise suppression unit and a mode decomposition unit, which respectively perform noise reduction and mode decomposition tasks on the vibration signal. These two units work together to ensure the accuracy and robustness of the signal processing.

[0087] Specifically, the noise suppression unit employs wavelet denoising to perform multi-resolution decomposition of the vibration signal to remove high-frequency noise and background interference. In some embodiments, the wavelet denoising process includes the following steps: First, an appropriate wavelet basis function (such as Daubechies or Symlet) is selected to decompose the input signal and obtain signal components at different resolutions; then, high-frequency components are suppressed using a soft thresholding function, and finally, the denoised signal is reconstructed. It should be noted that the key to wavelet denoising lies in the threshold selection. In some embodiments, the threshold can be dynamically adjusted according to the signal-to-noise ratio to optimize the denoising effect.

[0088] In one possible implementation, the wavelet-denoised signal can be represented as:

[0089]

[0090] in:

[0091] This represents the signal after noise reduction;

[0092] Represents the coefficients after wavelet decomposition;

[0093] Represents the soft threshold function;

[0094] Indicates the number of decomposition layers.

[0095] Understandably, wavelet denoising can not only effectively remove high-frequency noise, but also preserve the low-frequency information of the signal, thereby preserving the characteristics of the track vibration signal to the greatest extent.

[0096] The mode decomposition unit further decomposes the denoised signal to extract abnormal vibration signals. Specifically, in this embodiment, variational mode decomposition (VMD) is used to decompose the signal into several mode components with narrowband characteristics. The core principle of VMD is to optimize the spectrum partitioning and concentrate the energy of different frequency bands of the signal into independent modes.

[0097] As one implementation method, the objective function of mode decomposition can be expressed as:

[0098]

[0099] in:

[0100] Indicates the first One modal signal;

[0101] Indicates the first The center frequency of each mode;

[0102] Indicates the total number of modes;

[0103] It is a unit impulse function;

[0104] It is a time variable.

[0105] For example, the VMD solution process includes the following steps: First, the initial modal signal and center frequency are set through spectrum initialization; then, the modal signal and frequency are gradually adjusted using an iterative optimization algorithm to make the spectrum distribution between modes optimal; finally, the output modal signal contains the modal components corresponding to the abnormal vibration signal.

[0106] It should be noted that, in some embodiments, a penalty term can be introduced to constrain the amplitude range of the modal signals in order to improve the stability of mode decomposition. Furthermore, the total number of modes can be dynamically adjusted according to the characteristics of the actual track vibration signal. This is to avoid over- or under-modal decomposition.

[0107] In one possible implementation, the anomalous signal obtained after modal decomposition can be used in a subsequent feature extraction module. It is understood that the low-frequency components of the modal signal typically correspond to the normal operating vibrations of the track, while the mid-to-high-frequency components may be directly related to anomalous events. Therefore, spectral analysis of the modal signal can further improve the accuracy of anomalous event identification.

[0108] As an extension, this module may also include a signal enhancement unit for amplifying weak anomalous events in the vibration signal. In some embodiments, the signal enhancement unit employs an adaptive filtering method, dynamically adjusting the filter parameters according to the frequency characteristics of the anomalous signal to enhance the signal strength in a specific frequency band.

[0109] In this embodiment, the feature extraction module extracts time-frequency features and envelope features from the abnormal vibration signal separated by the signal processing module, which serve as the feature vector input to the event recognition module. Through this module's processing, key features of abnormal events can be effectively extracted, providing high-precision data support for subsequent classification and recognition.

[0110] Alternatively, the feature extraction module includes a frequency decomposition unit and an envelope analysis unit, used to extract frequency and time-domain features of the signal, respectively. These two units work together to ensure the extracted features are representative and robust.

[0111] Specifically, the frequency decomposition unit employs wavelet packet decomposition technology to decompose the vibration signal into multiple frequency bands. Wavelet packet decomposition is a multi-resolution analysis method that can simultaneously extract low-frequency and high-frequency components of a signal through recursive decomposition. In one possible implementation, the calculation process of wavelet packet decomposition is as follows:

[0112] First, an appropriate wavelet basis (such as Daubechies or Symlet) is selected to decompose the abnormal vibration signal. The recursive formula for wavelet packet decomposition is:

[0113]

[0114]

[0115] in:

[0116] Indicates the first Layer, First The signal of each node;

[0117] and These are the coefficients of the low-pass filter and the high-pass filter, respectively;

[0118] Indicates the index of the discrete sampling point;

[0119] Represents a time variable.

[0120] Through the above decomposition, signal components in different frequency bands can be obtained. It should be noted that the frequency range corresponding to each component is related to the signal's sampling frequency and the number of decomposition levels. In some embodiments, the frequency range of the components can be calibrated through spectral analysis.

[0121] After decomposition, the frequency decomposition unit calculates the energy characteristics of each frequency band signal. The formula for calculating the energy characteristics is:

[0122]

[0123] in:

[0124] Indicates the first Energy characteristics of each frequency band;

[0125] Indicates the first Layer, First Node at the The value of each sampling point.

[0126] It is understandable that the output characteristics of the frequency decomposition unit... It characterizes the energy distribution of the signal in different frequency ranges, providing key frequency dimension features for subsequent classification models.

[0127] The envelope analysis unit is responsible for extracting the time-domain features of the abnormal vibration signal. Specifically, in this embodiment, the envelope analysis unit calculates the instantaneous envelope of the signal using the Hilbert transform. The definition formula for the Hilbert transform is:

[0128]

[0129] in:

[0130] Indicates an abnormal vibration signal;

[0131] This represents the result of the Hilbert transform of the signal;

[0132] and These are time variables and integral variables, respectively;

[0133] It is a mathematical constant.

[0134] Through the original signal Hilbert transform results Calculate the instantaneous envelope of the signal:

[0135]

[0136] in:

[0137] This represents the instantaneous envelope signal.

[0138] It should be noted that instantaneous envelope The characteristics of vibration signal amplitude variation over time are described. As one possible implementation, the output of the envelope analysis unit can be the statistical characteristics of the envelope signal (such as average value, peak value, and rate of change) to enhance the representativeness of the time-domain characteristics.

[0139] For example, the output features of the frequency decomposition unit and the envelope analysis unit can be combined into a feature vector. ,in:

[0140] Indicates frequency characteristics;

[0141] Represents time-domain features.

[0142] Understandably, the feature vectors from the feature extraction module provide rich time-frequency information for the event recognition module, thereby improving the accuracy of abnormal event classification.

[0143] To adapt to the needs of different orbital environments, the feature extraction module can also be expanded. For example:

[0144] In some embodiments, the feature extraction module may incorporate an adaptive feature selection algorithm to dynamically adjust the extracted feature set based on the statistical characteristics of the track vibration signal.

[0145] In another possible implementation, frequency domain and time domain features can be fused with other signal features (such as phase features or spatial features) to further enhance the representativeness of the feature vectors.

[0146] In this embodiment, the event recognition module analyzes the feature vectors output by the feature extraction module to identify the type of track vibration event. By introducing deep learning technology, the event recognition module can accurately classify track abnormal events (such as track loosening, illegal construction, foreign object intrusion, etc.) and output the event type and confidence level.

[0147] As an alternative, the event recognition module includes a feature input unit, a deep learning classification unit, and a classification output unit. These units work together to form a complete recognition process, from the input of feature vectors to the output of classification results.

[0148] Specifically, the feature input unit is used to receive the multidimensional feature vector generated by the feature extraction module. It should be noted that the dimensions and number of input features can be dynamically adjusted according to the complexity of the track vibration signal.

[0149] The deep learning classification unit employs an architecture combining convolutional neural networks (CNNs) and long short-term memory networks (LSTMs) to classify and recognize feature vectors. As an example, CNNs are used to extract the spatial distribution patterns of feature vectors, and their convolution calculation formula is:

[0150]

[0151] in:

[0152] This represents the value of the feature map after convolution;

[0153] Indicates the weights of the convolution kernel;

[0154] This represents the value at the corresponding position in the input feature map;

[0155] For bias;

[0156] For example, ReLU is an activation function.

[0157] Understandably, the role of convolutional neural networks is to extract local features from multi-dimensional feature vectors, thereby improving the accuracy of event classification.

[0158] In one possible implementation, a Long Short-Term Memory (LSTM) network is used to capture the time-series patterns of the feature vectors. The state update formula for the LSTM is as follows:

[0159]

[0160] in:

[0161] Indicates the current hidden state;

[0162] This indicates the hidden state at the previous moment;

[0163] Indicates the current input features;

[0164] Represents the weight matrix;

[0165] Indicates bias;

[0166] For activation functions (such as Sigmoid).

[0167] It should be noted that time series modeling using LSTM can capture the dynamic changes in track vibration signals, thereby improving the ability to determine the type of event.

[0168] The classification output unit generates event type labels and confidence scores based on the results of the deep learning model. Alternatively, event type labels may include common event types such as track loosening, illegal construction, and foreign object intrusion, while the confidence score quantifies the reliability of the classification results. In some embodiments, the classification output unit may use a Softmax function to calculate the probability of each event type.

[0169]

[0170] in:

[0171] Representing the eigenvector Belongs to event type The probability of;

[0172] This indicates the deep learning model's understanding of event types. The original predicted value.

[0173] In one possible implementation, the classification output unit can also set alarm thresholds based on confidence scores. For example, when the confidence score for a certain event type exceeds a set value, an alarm signal is triggered and transmitted to the early warning management module.

[0174] Understandably, the design focus of the event recognition module is on accuracy and real-time performance. To further improve classification accuracy, some embodiments can extend the deep learning classification unit. For example, multi-model ensemble techniques can be introduced to perform a weighted average of the prediction results from multiple classification models, thereby obtaining a more stable classification result.

[0175] In this embodiment, the event location module is used to determine the specific location of a track vibration event in the optical fiber based on the time delay signal acquired by the optical fiber sensor, and maps this location to the geographic coordinate system of the track, providing accurate spatial information for subsequent early warning and management. Through precise time delay calculation and geographic mapping, this module can achieve real-time location of abnormal events.

[0176] As an alternative, the event location module includes a time delay calculation unit, a location calculation unit, and a geographic mapping unit. These units work in coordination to gradually complete the trajectory location calibration of the event.

[0177] Specifically, the time delay calculation unit is used to calculate the time delay difference in optical signal propagation caused by a vibration event in the optical fiber. For example, the formula for calculating the time delay difference is as follows:

[0178]

[0179] in:

[0180] This indicates the time it takes for the backscattered light signal to reach the photodetector;

[0181] This indicates the time it takes for the pulsed light signal to be emitted from the light source;

[0182] This indicates the time delay of the optical signal propagation.

[0183] It should be noted that the time delay difference Time delay is a key characteristic of optical signal propagation in optical fiber, directly reflecting the change in the length of the optical signal's propagation path. As one possible implementation, the time delay calculation unit can accurately capture this time delay using a high-speed data acquisition card. and This ensures high accuracy in delayed calculations.

[0184] The location calculation unit calculates the specific location of the vibration event within the optical fiber based on the time delay difference and the physical characteristics of the fiber. Specifically, the event location... It can be calculated using the following formula:

[0185]

[0186] in:

[0187] Indicates the distance of the event within the optical fiber;

[0188] The speed of light;

[0189] The refractive index of the optical fiber;

[0190] This represents the time delay difference in the propagation of the optical signal.

[0191] Understandably, since the optical signal needs to travel back and forth in the optical fiber, the path length is... Therefore, the denominator in the formula includes a factor of 2. It should be noted that the refractive index of optical fiber... It usually depends on the material of the optical fiber (e.g., quartz optical fiber). It is usually around 1.45, and can be obtained through experimental calibration.

[0192] In one possible implementation, to further improve positioning accuracy, the position calculation unit can use interpolation algorithms to refine the time delay data. For example, linear interpolation or cubic spline interpolation can be used to fit the delay variation of the optical signal, thereby generating higher resolution position information between sampling points.

[0193] The geographic mapping unit is used to map the calculated fiber optic locations. The event location results of the track points are generated by mapping the data to the geographic coordinate system of the track. Specifically, the geographic mapping unit can convert the length coordinates in the optical fiber into the actual geographic coordinates of the track based on the segment calibration information of the optical fiber deployment. In some embodiments, the segment length of the optical fiber can be between 10 meters and 50 meters, and each segment of the optical fiber is associated with the track position through calibration data.

[0194] Understandably, in complex track environments, such as tunnels or curves, the fiber optic cable's deployment path may deviate from the actual track path. Therefore, the geographic mapping unit can incorporate correction algorithms to adjust the positioning results based on the track's geometry. For example, the fiber optic cable's position can be calibrated using CAD drawings of the track or GPS positioning data to ensure the accuracy of the positioning results.

[0195] In this embodiment, the early warning management module generates early warning signals based on the event type output by the event identification module and the location information provided by the event location module, and then visualizes and manages the early warning information. This module allows track anomalies to be presented to monitoring personnel in an intuitive way, along with corresponding response suggestions to ensure track operation safety.

[0196] As an alternative, the early warning management module includes an alarm generation unit, an information visualization unit, and a data storage unit. These units collaborate through data interfaces to achieve real-time early warning and information traceability.

[0197] Specifically, the alarm generation unit generates corresponding multi-level warning signals based on the event type and confidence level provided by the event recognition module and the track position provided by the event location module. For example, the multi-level warning signals can be divided into low-level, medium-level, and high-level warnings. The warning classification can be based on the following logic:

[0198] If the confidence score of the event recognition module If the value exceeds a set threshold (e.g., 0.8) and the event type is a high-risk event (e.g., loose track or illegal construction), an advanced warning will be triggered.

[0199] If the confidence score is between 0.5 and 0.8, and the event type is a medium-risk event, then a medium-level warning is triggered.

[0200] If the confidence score is below 0.5, or the event type is a low-risk event (such as small foreign object intrusion), a low-level warning is triggered.

[0201] It should be noted that the alarm generation unit can also generate location-based early warning signals by combining the location information provided by the event location module. For example, in the case of advanced early warning, the specific location and geographical coordinates of the track can be marked in the early warning signal to quickly locate abnormal events.

[0202] The information visualization unit is used to display early warning signals and related information in real time on the monitoring platform. As one possible implementation, the visualization interface includes the display of event type, location information, and early warning level. Specifically:

[0203] Event types can be displayed using text labels (such as "loose tracks" or "illegal construction");

[0204] Location information can be visually displayed using markers on the track plan;

[0205] Warning levels can be distinguished by color coding (e.g., red for high-level warning, yellow for medium-level warning, and green for low-level warning).

[0206] Understandably, to improve monitoring efficiency, the information visualization unit can also support the simultaneous display of multiple events. In some embodiments, the system can overlay track position markers of different events to form a heat map of track anomalies, thereby helping monitoring personnel quickly identify areas of concentrated risk.

[0207] The data storage unit records all early warning information, including event type, location information, warning level, and timestamp. This information is stored in the database as logs, supporting subsequent queries and analysis. For example, the search function can be used to query all high-level early warning records within a specific time period to analyze the safety status of track operations.

[0208] As an extension, the early warning management module can also interface with external response systems. In some embodiments, when an advanced early warning is triggered, the system can automatically send alarm information to track maintenance personnel, along with the event type and location data. Furthermore, the system can also link with the train dispatching system to temporarily restrict train speeds in high-risk areas to reduce the probability of safety accidents.

[0209] In one possible implementation, to improve the reliability of the early warning signal, the early warning management module can also incorporate multi-signal fusion technology. For example, it can combine the intensity information of the vibration signal and the confidence score of the event classification to calculate a comprehensive early warning index. :

[0210]

[0211] in:

[0212] Indicates the intensity of the vibration signal (e.g., the maximum value of the envelope);

[0213] The confidence score indicates the classification of the event.

[0214] and These are weighting coefficients, which can be adjusted according to actual needs.

[0215] In summary, this invention combines phase-optical time-domain reflectometry (Φ-OTDR) with deep learning algorithms to achieve high-precision real-time monitoring of vibration signals along the track and identification of abnormal events. This invention can accurately capture abnormal signals in track vibration, identify various abnormal events such as track loosening, illegal construction, and foreign object intrusion, and provide real-time early warning and location services, providing reliable protection for the safe operation of rail transit. It features high precision, real-time performance, and wide adaptability.

[0216] Accordingly, please refer to the appendix. Figure 2 The present invention also provides a distributed fiber optic acoustic sensing method for track safety monitoring, comprising the following steps:

[0217] S1. Vibration signals along the track are collected using phase-optic time-domain reflectometry.

[0218] First, the method is based on phase-domain reflectometry, which uses a sensing module to capture vibration signals from optical fibers along the track. A pulsed light source injects high-power pulsed light into the optical fibers deployed along the track. Vibrations in the fibers cause changes in the backscattered light signal. The optical signal is converted into an electrical signal by a photodetector, generating time-series data corresponding to the vibration signal.

[0219] S2. Noise suppression is performed on the collected vibration signals, and abnormal vibration signals are extracted through mode decomposition;

[0220] The acquired vibration signals undergo noise suppression and decomposition by the signal processing module. A multi-resolution noise reduction algorithm removes background noise while preserving abnormal vibration information. Subsequently, a mode decomposition method is used to decompose the signal into several modal components, extracting vibration signals related to the abnormal event.

[0221] S3. Extract frequency band energy features using wavelet packet decomposition and calculate the signal envelope based on Hilbert transform;

[0222] The processed abnormal vibration signal enters the feature extraction module for further extraction of time-frequency and envelope features. Wavelet packet decomposition is used to calculate the energy distribution of the signal in different frequency bands, generating frequency features. Simultaneously, based on envelope analysis, the time-domain features of the signal are obtained, forming a complete feature vector, which serves as input data for subsequent classification models.

[0223] S4. Input the extracted features into a deep learning model to classify and identify the types of track anomaly events;

[0224] Feature vectors are input into a deep learning classification model in the event recognition module to identify track anomaly event types. The classification model combines convolutional neural networks and long short-term memory networks to extract spatial distribution patterns and time series patterns of features, respectively, and finally outputs the event type and confidence score. The classification results can distinguish between various anomaly events such as track loosening, illegal construction, and foreign object intrusion.

[0225] S5. Calculate the specific location of the orbital event based on the optical signal delay difference;

[0226] After classification, the event location module calculates the event's precise location in the orbit using the difference in propagation time delay of the optical signal. By mapping the fiber optic length coordinates to the orbit's geographic coordinate system, spatial location information of the event is generated, providing precise location data for subsequent early warning signals.

[0227] S6. Generate early warning signals based on event type and location;

[0228] Finally, the early warning management module generates early warning signals based on the categorized event types and location information, and displays the relevant information through a visualization platform. The tiered processing of early warning signals intuitively reflects the severity of the event and supports subsequent response decisions. Simultaneously, all early warning information is recorded in a database for subsequent queries and statistical analysis of track safety status.

[0229] This method constructs an efficient and accurate track safety monitoring process through the coordinated work of data collection, processing, feature extraction, classification and identification, positioning and early warning. It has significant advantages in real-time performance and robustness, providing reliable protection for rail transit safety.

[0230] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A distributed fiber optic acoustic sensing track safety monitoring system, characterized in that, include: The sensing and acquisition module is used to capture vibration signals of optical fibers along the track based on phase light time-domain reflectometry and generate time-series vibration data. The signal processing module is used to denoise the vibration data and decompose the abnormal vibration signal. The feature extraction module is used to extract time-frequency features and envelope features from the abnormal vibration signal; The event recognition module is used to classify track vibration events based on extracted features and identify event types. The event location module is used to determine the specific location of an event on the track based on the time delay of the optical signal; The early warning management module is used to generate early warning signals and visualize information based on the identified event type and location information; The feature extraction module includes: The frequency decomposition unit is used to perform multi-band decomposition of abnormal vibration signals using wavelet packet decomposition and extract the energy features corresponding to each frequency band. The envelope analysis unit is used to calculate the instantaneous envelope of the abnormal vibration signal based on the Hilbert transform and extract the dynamic change characteristics of the vibration amplitude. The feature integration unit is used to combine the extracted frequency band energy features and vibration amplitude dynamic change features into a feature vector and transmit it to the event recognition module. The frequency decomposition unit extracts frequency band energy features through the following steps: The abnormal vibration signal was decomposed into multi-band signals using wavelet packet decomposition technology. Each band signal was calculated using a recursive formula. ； ； in, For the first Layer, First Decomposed signal of each node, and These are the low-pass and high-pass coefficients of the decomposition filter, respectively. Represents discrete sampling points of a time series. Represents the time variable of the signal; The energy characteristics of the decomposed signal for each frequency band are calculated using the following formula: ； in, For the first Energy characteristics of each frequency band Indicates the first Layer, First The signal decomposed at node i is at the... The values ​​at each sampling point; The envelope analysis unit extracts the dynamic change characteristics of vibration amplitude through the following steps: For abnormal mode signals Perform Hilbert transform to calculate the instantaneous signal: ； in, Represents modal signals The Hilbert transform result, Time-domain representation of abnormal vibration signals Represents the integral variable. Represents a time variable. It is a mathematical constant; Use original signal and its Hilbert transform result Calculate the vibration envelope The formula for calculating the vibration envelope is: ； in, It represents the instantaneous envelope of the vibration signal.

2. The distributed fiber optic acoustic wave sensing track safety monitoring system according to claim 1, characterized in that, The sensing acquisition module includes: A pulsed light source is used to inject high-power optical pulses along an optical fiber; The signal acquisition unit is used to acquire backscattered light signals in the optical fiber. The data conversion unit is used to convert the collected scattered light signals into track vibration data.

3. The distributed fiber optic acoustic wave sensing track safety monitoring system according to claim 1, characterized in that, The signal processing module includes: Noise suppression unit, used to filter out background noise in vibration data; The mode decomposition unit is used to decompose vibration signals through variational mode decomposition and extract abnormal vibration signals.

4. The distributed fiber optic acoustic wave sensing track safety monitoring system according to claim 1, characterized in that, The event recognition module includes: The feature input unit is used to receive the frequency features and time-domain features output by the feature extraction module and combine them into a multi-dimensional feature vector. The deep learning classification unit classifies and identifies multidimensional feature vectors based on convolutional neural networks and long short-term memory networks. The classification output unit is used to generate type labels and confidence scores for track vibration events based on the deep learning classification results, and then transmits the results to the event localization module.

5. The distributed fiber optic acoustic wave sensing track safety monitoring system according to claim 1, characterized in that, The event location module includes: The time delay calculation unit is used to calculate the time delay difference of optical signal propagation caused by abnormal vibration events in the optical fiber. The position calculation unit is used to calculate the specific location of the abnormal vibration event on the track based on the time delay difference and the refractive index of the optical fiber. The event calibration unit is used to map the calculated location to the orbital geographic coordinate system, providing event location results for orbital points.

6. The distributed fiber optic acoustic wave sensing track safety monitoring system according to claim 1, characterized in that, The early warning management module includes: An alarm generation unit is used to generate multi-level alarm signals based on event type and location information. Visualization units are used to display the type, location, and severity of track events in real time; Data storage unit, used to record historical event information, supporting querying and analysis.

7. A distributed fiber optic acoustic sensing method for track safety monitoring, applied to the system described in any one of claims 1-6, characterized in that, Includes the following steps: Vibration signals along the track were acquired using phase-optical time-domain reflectometry. Noise suppression is performed on the collected vibration signals, and abnormal vibration signals are extracted through mode decomposition. Frequency band energy features are extracted using wavelet packet decomposition, and the signal envelope is calculated based on Hilbert transform. The extracted features are input into a deep learning model to classify and identify types of track anomaly events. Calculate the specific location of orbital events based on optical signal delay difference; Early warning signals are generated based on the event type and location.

Images