Rail transit leaky cable detection method and device, electronic equipment and storage medium

Abstract

The invention relates to a rail transit leaky coaxial cable detection method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a radio frequency signal in a leaky coaxial cable, and carrying out the coupling and adjustment of the radio frequency signal, so as to convert the radio frequency power into a voltage signal. And determining a standing-wave ratio and a signal attenuation amount according to the voltage signal, carrying out wavelet threshold noise reduction on the vibration acceleration of each edge node, and generating a data packet in combination with temperature and humidity data. And calling a bidirectional long-short-term memory network of an attention mechanism to calculate a health index corresponding to the data packet, and predicting a health index change trend in a set time period. And comparing the health index and the health index change trend with a set threshold value so as to give an alarm when the set threshold value is not met. According to the invention, the state and the aging reason of the leaky coaxial cable are evaluated from multiple dimensions, the problem that the service life of the leaky coaxial cable is difficult to comprehensively evaluate through single parameter monitoring is solved, the detection precision of the state of the leaky coaxial cable is improved, and the requirements of rail transit for real-time performance and low operation cost can be better met.

Description

Technical Field

[0001] This invention relates to the field of track inspection technology, and in particular to a method, apparatus, electronic device, and storage medium for detecting leaky cables in rail transit. Background Technology

[0002] Leaky coaxial cables are a critical infrastructure for ensuring the continuity of wireless communication in enclosed spaces such as subways and railways. Their performance degradation directly affects the reliability of train control, passenger information systems, and public safety.

[0003] Currently, traditional methods for monitoring leaky cables mainly rely on regular manual inspections or single standing wave ratio (VSWR) tests. These methods typically only monitor electrical performance parameters such as VSWR, making it difficult to comprehensively reflect performance degradation caused by the combined effects of environmental temperature and humidity changes, train vibrations, and other mechanical and external factors. This makes it challenging to accurately assess the remaining lifespan of the leaky cable. Furthermore, strong electromagnetic interference from traction power supply exists along rail transit lines, making the acquisition circuits of traditional monitoring devices susceptible to interference and data distortion. Monitoring accuracy is difficult to guarantee in complex electromagnetic environments. In addition, existing solutions are mostly centralized, with all raw data uploaded to a central server. This requires high transmission bandwidth and results in significant data processing delays, making it difficult to meet the high real-time requirements of rail transit for rapid fault location (typically requiring second-level response). Summary of the Invention

[0004] Therefore, it is necessary to provide a method, device, electronic equipment, and storage medium for detecting leaky cables in rail transit that can comprehensively, accurately, and in real time monitor the status of leaky cables and provide intelligent early warning, in order to address the above-mentioned technical problems.

[0005] This invention provides a method for detecting leaky cables in rail transit systems, the method comprising: The radio frequency (RF) signal in the leaky cable is acquired, and the RF signal is coupled and adjusted to convert the RF power into a voltage signal. The standing wave ratio and signal attenuation are determined based on the voltage signal, and wavelet threshold noise reduction is performed on the vibration acceleration of each edge node. Data packets are generated by combining temperature and humidity data. The bidirectional long short-term memory network with an attention mechanism is invoked to calculate the health index corresponding to the data packet and predict the trend of health index change within a set time period. The health index and its trend are compared with a set threshold to issue an alarm when the set threshold is not met.

[0006] In one embodiment, acquiring the radio frequency signal in the leaky cable and coupling and adjusting the radio frequency signal to convert the radio frequency power into a voltage signal includes: The radio frequency signals are coupled in the forward and reverse directions respectively by directional couplers, and the attenuator is called to adjust the coupled radio frequency signals so as to send them to the radio frequency detector to convert the radio frequency power into forward and reverse detection voltages. Calculate the ratio of the forward and reverse detection voltages to obtain the standing wave ratio (SWR), and calculate the difference between the actual forward detection voltage and the calibrated detection voltage to obtain the signal attenuation.

[0007] In one embodiment, the step of acquiring the radio frequency signal in the leaky cable and coupling and adjusting the radio frequency signal to convert the radio frequency power into a voltage signal further includes: The DC voltage signal from the radio frequency detector is received, and the voltage difference between the positive and negative terminals of the DC voltage signal is amplified by an amplifier while the interference voltage at the positive and negative terminals of the DC voltage signal is suppressed to obtain the amplified differential signal. The differential signal is sent to a digital-to-analog converter for sampling to eliminate interference voltage in the standing wave ratio and signal attenuation.

[0008] In one embodiment, the step of determining the standing wave ratio and signal attenuation based on the voltage signal, performing wavelet threshold noise reduction on the vibration acceleration of each edge node, and generating a data packet by combining temperature and humidity data includes: Temperature and humidity data and vibration acceleration in the leaky cable detection area are collected by a temperature and humidity sensor and a triaxial accelerometer, respectively. The temperature and humidity data and vibration acceleration are time-aligned and labeled with device names to obtain an initial data stream. The initial data stream is sent to an edge node, where the edge node performs format verification on the initial data stream.

[0009] In one embodiment, the step of determining the standing wave ratio and signal attenuation based on the voltage signal, performing wavelet threshold noise reduction on the vibration acceleration of each edge node, and generating a data packet by combining temperature and humidity data further includes: The acceleration time-domain signal is extracted from the formatted and verified initial data stream, and the wavelet basis function is called to perform wavelet packet decomposition on the acceleration time-domain signal to obtain multi-level detail coefficients and approximation coefficients. Extract the first layer of detail coefficients from the multi-layer detail coefficients, calculate the kurtosis value of the first layer of detail coefficients, and set an adaptive threshold based on the kurtosis value; Wherein, the first layer detail coefficient is any layer detail coefficient among the multi-layer detail coefficients, and the adaptive threshold is inversely proportional to the kurtosis value.

[0010] In one embodiment, the step of determining the standing wave ratio and signal attenuation based on the voltage signal, performing wavelet threshold noise reduction on the vibration acceleration of each edge node, and generating a data packet by combining temperature and humidity data further includes: The threshold function is called to shrink the first layer detail coefficients, and the shrunken first layer detail coefficients are reconstructed by wavelet reconstruction with the approximate coefficients of the corresponding layer to obtain the denoised acceleration time domain signal. Vibration features are extracted from the noise-reduced acceleration time-domain signal, and the vibration features, standing wave ratio, signal attenuation, and temperature and humidity data are packaged into the data package.

[0011] In one embodiment, the bidirectional long short-term memory network that invokes the attention mechanism calculates the health index corresponding to the data packet and predicts the trend of health index changes within a set time period, including: The bidirectional long short-term memory network of the attention mechanism is trained under supervision using pre-stored historical data and labeled fault parameters. Mean squared error is introduced as the loss function and Adam optimizer to obtain the trained health assessment model. The health assessment model is invoked to calculate the health index corresponding to the data packet at each monitoring point, and the health index is input into a linear regression model or an LSTM prediction model to predict the trend of health index changes within a set time period.

[0012] The present invention also provides a rail transit cable leakage detection device for implementing the rail transit cable leakage detection method described in any of the above claims, the device comprising: A signal conversion module is used to acquire radio frequency signals in the leaky cable and couple and adjust the radio frequency signals to convert radio frequency power into voltage signals; The data integration module is used to determine the standing wave ratio and signal attenuation based on the voltage signal, perform wavelet threshold noise reduction on the vibration acceleration of each edge node, and generate data packets by combining temperature and humidity data. The model detection module is used to call the bidirectional long short-term memory network with attention mechanism to calculate the health index corresponding to the data packet and predict the trend of health index change within a set time period. The health early warning module is used to compare the health index and its changing trend with a set threshold, so as to issue an alarm when the set threshold is not met.

[0013] The present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the rail transit cable leakage detection method as described above.

[0014] The present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, implements the rail transit cable leakage detection method as described above.

[0015] The aforementioned method, device, electronic equipment, and storage medium for detecting leaky cables in rail transit acquire radio frequency (RF) signals from the leaky cable and couple and adjust these signals to convert RF power into voltage signals. Then, the standing wave ratio (VSWR) and signal attenuation are determined based on the voltage signals, and wavelet threshold noise reduction is applied to the vibration acceleration of each edge node. Data packets are generated by combining temperature and humidity data. Subsequently, a bidirectional long short-term memory (LSTM) network with an attention mechanism is invoked to calculate the health index corresponding to the data packets and predict the trend of health index changes within a set time period. Finally, the health index and its trend are compared with a set threshold to trigger an alarm when the threshold is not met. This invention comprehensively evaluates the condition and causes of aging of leaky cables from multiple dimensions, solving the problem that traditional single-parameter monitoring is difficult to comprehensively assess the lifespan of leaky cables. Furthermore, by combining differential acquisition circuits with adaptive electromagnetic interference suppression algorithms, the accuracy of key electrical parameter monitoring can still be guaranteed even in strong interference environments, effectively improving the detection accuracy of leaky cable conditions and better meeting the requirements of rail transit for high real-time performance and low operating costs. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is one of the flowcharts of the rail transit leaky cable detection method provided by the present invention; Figure 2 This is a schematic diagram of the overall process of leaky cable detection in a specific embodiment of the present invention for the rail transit leaky cable detection method; Figure 3 This is the second flowchart illustrating the rail transit cable leakage detection method provided by the present invention. Figure 4 This is the third flowchart illustrating the rail transit cable leakage detection method provided by the present invention. Figure 5 This is the fourth flowchart illustrating the rail transit cable leakage detection method provided by the present invention. Figure 6 The fifth flowchart illustrates the method for detecting leaky cables in rail transit provided by this invention. Figure 7 This is the sixth flowchart illustrating the rail transit cable leakage detection method provided by the present invention. Figure 8 This is the seventh flowchart illustrating the rail transit cable leakage detection method provided by the present invention. Figure 9This is a schematic diagram of the structure of the rail transit cable leakage detection device provided by the present invention; Figure 10 This is a diagram of the internal structure of the electronic device provided by the present invention. Detailed Implementation

[0018] 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, 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.

[0019] The following is combined Figures 1 to 10 The present invention describes a method, apparatus, electronic device, and storage medium for detecting leaky cables in rail transit systems.

[0020] like Figure 1 As shown, in one embodiment, a method for detecting leaky cables in rail transit includes the following steps: Step S110: Obtain the radio frequency signal in the leaky cable, and couple and adjust the radio frequency signal to convert the radio frequency power into a voltage signal.

[0021] Specifically, the data acquisition controller acquires the radio frequency signal in the leaky cable, and couples the radio frequency signal in the forward and reverse directions through a directional coupler. Then, the weak signal coupled out is adjusted by a precision attenuator and sent to a zero-frequency stable radio frequency detector to convert the radio frequency power into a DC voltage signal.

[0022] Combination Figure 2 As shown in the specific embodiment, the rail transit leaky cable detection method provided by the present invention uses monitoring terminals (assuming model LTU-100) fixedly installed on leaky coaxial cables on or near equipment mounting plates on the tunnel sidewall at intervals of 250 meters. Each monitoring terminal is connected to an edge processing node (assuming model EN-200) located at the midpoint of the section or in the station equipment room via RS-485 bus or low-power industrial Ethernet. Approximately 15 edge processing nodes along the entire line are connected to a cloud management platform server cluster deployed in the Operation Control Center (OCC) via a dedicated rail transit transmission network (such as an industrial ring network). After power-on, the monitoring terminals begin periodic operation, with the edge processing nodes continuously polling all monitoring terminals within their jurisdiction, collecting and processing data; the processed feature data and alarm information are pushed to the cloud platform in real time; the platform performs comprehensive analysis, storage, and display, and issues alarms via sound and light, SMS, and maintenance work order systems when warning conditions are met.

[0023] In this embodiment, the core hardware of the monitoring terminal is a low-power, high-precision data acquisition microcontroller unit (MCU). First, a pair of highly directional couplers (e.g., 20dB coupling) are used to couple the radio frequency (RF) signal from the leaky cable in both forward and reverse directions. The weak coupled signal is adjusted by a precision attenuator and then sent to a zero-frequency stable RF detector (e.g., ADL5513) to convert the RF power into a DC voltage signal. The VSWR is obtained by calculating the ratio of the forward to reverse detection voltages; the signal attenuation is obtained by comparing the difference (in dB) between the current forward detection voltage and the initial calibration voltage.

[0024] Furthermore, the DC voltage signal output by the detector (containing the real signal and common-mode electromagnetic interference) is fed into a precision instrumentation amplifier (such as INA188) with a high common-mode rejection ratio (CMRR > 120dB @ 50Hz). This amplifier amplifies the voltage difference between the two input terminals (positive and negative signals) while greatly suppressing the common interference voltage (generated by a strong electromagnetic field). The amplified differential signal is then sent to the MCU's high-precision ADC (analog-to-digital converter) for sampling. Actual measurements show that this circuit can stably control the measurement error of the effective signal within 0.1dB even under the background of strong pulse interference of -45dBm (approximately 5.6mV) generated by the train traction system.

[0025] Step S120: Determine the standing wave ratio and signal attenuation based on the voltage signal, perform wavelet threshold noise reduction on the vibration acceleration of each edge node, and generate a data packet by combining temperature and humidity data.

[0026] Specifically, the server calculates the standing wave ratio (SWR) and signal attenuation based on the voltage signal obtained in step S110. The SWR is the ratio of the forward to the reverse detection voltage, and the signal attenuation is the difference between the current forward detection voltage and the initial calibration voltage. Wavelet thresholding is then applied to the vibration acceleration at each monitoring point to extract vibration features. Finally, the extracted vibration features, along with temperature and humidity data, SWR, and signal attenuation, are packaged into a concise data packet.

[0027] Combination Figure 2 As shown, in a specific embodiment, the rail transit leaky cable detection method provided by the present invention uses an integrated digital temperature and humidity sensor (such as SHT35) to detect leaky cables via I... 2 The MCU communicates via the C-bus, with a measurement accuracy of ±0.2°C and ±2%RH. Simultaneously, a triaxial MEMS accelerometer (such as the ADXL357) with a range of ±10g is used, communicating with the MCU via an SPI interface. This accelerometer is directly fixed to a leaky cable clamp or a nearby rigid structure to sense vibrations caused by passing trains.

[0028] The digital output signals of the aforementioned sensors have strong anti-interference capabilities, but their power supply lines are also filtered and isolated to prevent power supply noise from introducing interference. The MCU's internal firmware coordinates the operation of each module, synchronously acquiring all parameters at a 1-second interval by default. The acquired raw data is timestamped and labeled with a device ID, and then sent to the edge processing node via the communication interface.

[0029] In this embodiment, after receiving the raw data stream, the edge processing node first performs format verification, and then applies an improved wavelet thresholding denoising algorithm to the time-series signal (mainly vibration acceleration signal). During the denoising process, firstly, a frame of acceleration time-domain signal sampled at 100Hz is zero-mean normalized, and the db4 wavelet basis function is selected for 5-level wavelet packet decomposition to obtain detail coefficients and approximation coefficients at each level from low to high frequency. Then, adaptive thresholding is performed. Traditional global thresholds (such as general thresholds) are prone to distortion in impact signals (such as transient vibrations caused by loose joints). This example introduces a layered adaptive threshold. For the j-th level detail coefficient, the kurtosis value of the j-th level detail coefficient is calculated. This kurtosis value effectively reflects the strength of the impact component in the signal, and its expression is: ,

[0030] In the formula, Let J be the detail coefficients for the j-th layer; Let be the kurtosis value of the detail coefficients at the j-th layer; , are the mean and standard deviation of the detail coefficients for the j-th layer, respectively.

[0031] Then, based on the calculated kurtosis value, an adaptive threshold is designed, expressed as: ,

[0032] In the formula, For the length of detail coefficients; The adaptive threshold for the detail coefficients of the j-th layer; , Both are adjustment factors, with values ​​of 0.8 and 0.2 respectively. In this formula, when the kurtosis of a certain layer signal is large (significant impact characteristics), the threshold is automatically reduced to retain more coefficients of that layer (i.e., impact characteristics); when the kurtosis is small (mainly noise), the threshold is relatively increased to filter out more noise.

[0033] Next, a continuously differentiable threshold function is used to shrink the coefficients to avoid the discontinuities of the hard threshold function and the constant deviation of the soft threshold function. The processed detail coefficients are then reconstructed using wavelet analysis with the unprocessed approximate coefficients to obtain the denoised time-domain signal. This processing improves the signal-to-noise ratio of the vibration signal from the original 40-50 dB to over 65 dB. The processed vibration characteristic values ​​(such as RMS and peak value), along with other calibrated and formatted parameters (VSWR, attenuation, temperature, and humidity), are combined into a streamlined data package and uploaded to the cloud.

[0034] Step S130: Invoke the bidirectional long short-term memory network with attention mechanism to calculate the health index corresponding to the data packet, and predict the trend of health index change within a set time period.

[0035] Specifically, a bidirectional long short-term memory network with an attention mechanism is used as the core prediction model. The model is trained based on historical data and labeled fault parameters. The trained model is then used to calculate the health index corresponding to the data packet. Finally, linear regression and LSTM are used to predict the trend of health index changes in the future.

[0036] Combination Figure 2 As shown in the specific embodiment, the rail transit leaky cable detection method provided by this invention involves a cloud platform receiving data from all edge nodes and writing it into a time-series database (such as InfluxDB) and a relational database. Each data point includes a device ID, time, location coordinates, and multiple parameter values. Then, a Bidirectional Long Short-Term Memory (Attention-BiLSTM) network with an attention mechanism is used as the core prediction model. The model input is set to a time window (such as the past 24 hours), and the aligned multi-parameter time-series sequence is determined based on the aforementioned time-series database and relational database. In the model structure, the BiLSTM layer is used to learn the long-term dependencies between parameters; the Attention layer automatically calculates the weights of the parameters' influence on the current state at different times, focusing on points of abnormal change; and the fully connected layer outputs a comprehensive health index between 0 and 100.

[0037] During model training, supervised training is performed using historical normal data, data prior to known faults, and manually labeled fault data. The loss function is mean squared error (MSE), and the optimizer is Adam. After training, a health assessment model is obtained. Then, for each monitoring point, the platform reads its latest multi-parameter features every second and inputs them into the trained health assessment model to calculate the current health index in real time. Finally, the health index sequence for the most recent period is input into a linear regression model or a simple LSTM prediction model to predict the trend of health index changes over a future period (e.g., 24 hours).

[0038] Step S140: Compare the health index and its trend with a set threshold to issue an alarm when the set threshold is not met.

[0039] Specifically, the health index and its changing trend obtained in step S130 are compared with the set threshold, and alarms are issued according to the degree of fault based on the comparison results.

[0040] Combination Figure 2 As shown in the specific embodiment, the rail transit cable leakage detection method provided by the present invention sets two threshold levels for alarm activation, namely: Warning threshold (e.g., health index < 80): When the health index falls below this threshold or the predicted trend is that it will quickly fall below this threshold, the platform will trigger an "attention" warning to remind maintenance personnel to pay attention. Alarm threshold (e.g., health index < 60): When the health index is below this threshold, or when multiple adjacent monitoring points show abnormalities at the same time, the platform determines it as a fault. Combined with the analysis of abrupt changes in signal attenuation, the platform can accurately locate the source of the fault (error ≤ 5 meters).

[0041] In this embodiment, the entire process from alarm detection to information release via API call to the SMS gateway and work order system is completed within 30 seconds. The alarm information is simultaneously displayed on the platform's GIS map, with the faulty section highlighted and flashing.

[0042] In another embodiment, taking a loose leaky cable connector as an example: Second T: The train passes by, and monitoring terminal A collects an abnormally increased vibration peak value. At the same time, due to impedance mismatch caused by loose connection, its standing wave ratio and attenuation show slight jumps. The ambient temperature and humidity remain unchanged.

[0043] At time T+1: The edge node receives the raw data from terminal A. The wavelet denoising algorithm accurately preserves the vibration and impact characteristics, calculates the high RMS value, and uploads it along with the electrical parameter change values.

[0044] At T+3 seconds: The cloud platform receives data, and the health assessment model identifies a strong correlation between "abnormally increased vibration" and "synchronous deterioration of electrical parameters," determining that the mechanical structure problem is causing a decline in electrical performance, and the health index drops rapidly from 85 to 62.

[0045] At time T+5: The platform compares the data from adjacent terminals B and C and finds that the attenuation of terminal B has also increased slightly, while terminal C is normal. Based on the attenuation gradient, the platform accurately locates the fault point approximately 3 meters downstream of terminal A (i.e., the connector location).

[0046] At T+28 seconds: The platform generates an alarm work order containing "fault location (composed of line and mileage), fault type (loose connector), health index (62), and recommended measures (tightening check)" and pushes it to the mobile APP of relevant maintenance personnel and the work team scheduling system. Finally, the entire time from the occurrence of the fault to the response is controlled within 30 seconds.

[0047] The aforementioned method for detecting leaky cables in rail transit acquires radio frequency (RF) signals from the leaky cable and couples and adjusts these signals to convert RF power into voltage signals. Then, based on the voltage signals, the standing wave ratio (VSWR) and signal attenuation are determined, and wavelet thresholding is applied to the vibration acceleration of each edge node for noise reduction. Data packets are then generated by combining temperature and humidity data. Subsequently, a bidirectional long short-term memory (LSTM) network with an attention mechanism is invoked to calculate the health index corresponding to the data packets and predict the trend of health index changes within a set time period. Finally, the health index and its trend are compared with a set threshold to trigger an alarm when the threshold is not met. This method comprehensively evaluates the condition and causes of aging of the leaky cable from multiple dimensions, solving the problem that traditional single-parameter monitoring is insufficient for comprehensively assessing the cable's lifespan. Furthermore, by combining differential acquisition circuitry with an adaptive electromagnetic interference suppression algorithm, the accuracy of key electrical parameter monitoring is still guaranteed even in strong interference environments, effectively improving the detection accuracy of the leaky cable condition and better meeting the requirements of rail transit for high real-time performance and low operating costs.

[0048] like Figure 3 As shown, in one embodiment, the rail transit leaky cable detection method provided by the present invention includes the following steps in step S110: Step S111: The RF signals are coupled in the forward and reverse directions respectively through the directional coupler, and the attenuator is called to adjust the coupled RF signals so as to send them to the RF detector to convert the RF power into forward and reverse detection voltages.

[0049] Step S112: Calculate the ratio of the forward and reverse detection voltages to obtain the standing wave ratio, and calculate the difference between the actual forward detection voltage and the calibrated detection voltage to obtain the signal attenuation.

[0050] like Figure 4 As shown, in one embodiment, the rail transit cable leakage detection method provided by the present invention further includes the following steps after step S110: Step S410: Receive the DC voltage signal from the radio frequency detector, and amplify the voltage difference between the positive and negative terminals of the DC voltage signal through an amplifier, while suppressing the interference voltage at the positive and negative terminals of the DC voltage signal, to obtain the amplified differential signal.

[0051] Step S420: The differential signal is sent to the digital-to-analog converter for sampling to eliminate interference voltage in the standing wave ratio and signal attenuation.

[0052] like Figure 5 As shown, in one embodiment, the rail transit cable leakage detection method provided by the present invention includes the following steps in step S120: Step S121: Collect temperature and humidity data and vibration acceleration of the leaky cable detection area using a temperature and humidity sensor and a triaxial accelerometer, respectively, and perform time alignment and device name labeling on the temperature and humidity data and vibration acceleration to obtain the initial data stream.

[0053] Step S122: Send the initial data stream to the edge node, and perform format verification on the initial data stream through the edge node.

[0054] like Figure 6 As shown, in one embodiment, the rail transit cable leakage detection method provided by the present invention further includes the following steps in step S120: Step S123: Extract the acceleration time-domain signal from the formatted and verified initial data stream, and call the wavelet basis function to perform wavelet packet decomposition on the acceleration time-domain signal to obtain multi-level detail coefficients and approximation coefficients.

[0055] Step S124: Extract the first layer detail coefficient from the multi-layer detail coefficients, calculate the kurtosis value of the first layer detail coefficients, and set an adaptive threshold based on the kurtosis value.

[0056] The first level detail coefficient is any level detail coefficient among the multi-level detail coefficients, and the adaptive threshold is inversely proportional to the kurtosis value.

[0057] like Figure 7 As shown, in one embodiment, the rail transit cable leakage detection method provided by the present invention further includes the following steps in step S120: Step S125: Call the threshold function to shrink the first layer detail coefficients, and then perform wavelet reconstruction on the shrunken first layer detail coefficients and the approximate coefficients of the corresponding layer to obtain the denoised acceleration time domain signal.

[0058] Step S126: Extract vibration features from the noise-reduced acceleration time-domain signal, and package the vibration features, standing wave ratio, signal attenuation, and temperature and humidity data into a data package.

[0059] like Figure 8 As shown, in one embodiment, the rail transit cable leakage detection method provided by the present invention includes the following steps in step S130: Step S131: Supervised training of the bidirectional long short-term memory network of the attention mechanism is performed using pre-stored historical data and labeled fault parameters. Mean squared error is introduced as the loss function and Adam optimizer to obtain the trained health assessment model.

[0060] Step S132: Call the health assessment model to calculate the health index corresponding to the data packets of each monitoring point, and input the health index into the linear regression model or LSTM prediction model to predict the trend of health index change within a set time period.

[0061] The following describes the rail transit leaky cable detection device provided by the present invention. The rail transit leaky cable detection device described below can be referred to in correspondence with the rail transit leaky cable detection method described above.

[0062] like Figure 9 As shown, in one embodiment, a rail transit cable leakage detection device includes a signal conversion module 910, a data integration module 920, a model detection module 930, and a health warning module 940.

[0063] The signal conversion module 910 is used to acquire the radio frequency signal in the leaky cable and couple and adjust the radio frequency signal to convert the radio frequency power into a voltage signal.

[0064] The data integration module 920 is used to determine the standing wave ratio and signal attenuation based on the voltage signal, and to perform wavelet threshold noise reduction on the vibration acceleration of each edge node, and generate data packets by combining temperature and humidity data.

[0065] The model detection module 930 is used to call the bidirectional long short-term memory network with attention mechanism to calculate the health index corresponding to the data packet and predict the trend of health index change within a set time period.

[0066] The health warning module 940 is used to compare the health index and the trend of the health index with the set threshold, so as to issue an alarm when the set threshold is not met.

[0067] In this embodiment, the signal conversion module 910 of the rail transit leaky cable detection device provided by the present invention is specifically used for: The RF signals are coupled in both forward and reverse directions using directional couplers, and the attenuator is used to adjust the coupled RF signals so that they can be sent to the RF detector to convert the RF power into forward and reverse detection voltages.

[0068] Calculate the ratio of the forward and reverse detection voltages to obtain the standing wave ratio (SWR), and calculate the difference between the actual forward detection voltage and the calibrated detection voltage to obtain the signal attenuation.

[0069] In this embodiment, the rail transit cable leakage detection device provided by the present invention further includes a signal amplification module, used for: The DC voltage signal from the radio frequency detector is received, and the voltage difference between the positive and negative terminals of the DC voltage signal is amplified by an amplifier while the interference voltage at the positive and negative terminals of the DC voltage signal is suppressed to obtain the amplified differential signal.

[0070] The differential signal is sent to a digital-to-analog converter for sampling to eliminate interference voltage in the standing wave ratio and signal attenuation.

[0071] In this embodiment, the data integration module 920 of the rail transit cable leakage detection device provided by the present invention is specifically used for: Temperature and humidity data and vibration acceleration in the leaky cable detection area are collected by temperature and humidity sensors and triaxial accelerometers, respectively. The temperature and humidity data and vibration acceleration are time-aligned and labeled with device names to obtain the initial data stream.

[0072] The initial data stream is sent to the edge node, where it is formatted and validated.

[0073] In this embodiment, the data integration module 920 of the rail transit leaky cable detection device provided by the present invention is further used for: The acceleration time-domain signal is extracted from the formatted and verified initial data stream, and wavelet basis functions are called to perform wavelet packet decomposition on the acceleration time-domain signal to obtain multi-level detail coefficients and approximation coefficients.

[0074] Extract the first layer of detail coefficients from the multi-layer detail coefficients, calculate the kurtosis value of the first layer of detail coefficients, and set an adaptive threshold based on the kurtosis value.

[0075] The first level detail coefficient is any level detail coefficient among the multi-level detail coefficients, and the adaptive threshold is inversely proportional to the kurtosis value.

[0076] In this embodiment, the data integration module 920 of the rail transit leaky cable detection device provided by the present invention is further used for: The threshold function is called to shrink the first layer detail coefficients, and the shrunken first layer detail coefficients and the corresponding approximation coefficients are reconstructed by wavelet to obtain the denoised acceleration time domain signal.

[0077] Vibration features are extracted from the denoised acceleration time-domain signal, and the vibration features, standing wave ratio, signal attenuation, and temperature and humidity data are packaged into a data package.

[0078] In this embodiment, the model detection module 930 of the rail transit leaky cable detection device provided by the present invention is specifically used for: The bidirectional long short-term memory network with attention mechanism is trained under supervision using pre-stored historical data and labeled fault parameters. Mean squared error is introduced as the loss function and Adam optimizer to obtain the trained health assessment model.

[0079] The health assessment model is called to calculate the health index corresponding to the data packets of each monitoring point, and the health index is input into the linear regression model or LSTM prediction model to predict the trend of health index changes within a set time period.

[0080] Figure 10 This example illustrates a schematic diagram of the physical structure of an electronic device, which can be a smart terminal. Its internal structure diagram can be as follows: Figure 10 As shown. The electronic device includes a processor, internal memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for detecting leaky cables in rail transit, which includes: The radio frequency (RF) signal in the leaky cable is acquired, and the RF signal is coupled and adjusted to convert the RF power into a voltage signal. The standing wave ratio and signal attenuation are determined based on the voltage signal, and wavelet threshold noise reduction is performed on the vibration acceleration of each edge node. Data packets are generated by combining temperature and humidity data. The bidirectional long short-term memory network with attention mechanism is invoked to calculate the health index corresponding to the data packet and predict the trend of health index change within a set time period. The health index and its trend are compared with a set threshold to issue an alarm when the threshold is not met.

[0081] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the electronic device to which the present invention is applied. A specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0082] On the other hand, the present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, implements a method for detecting cable leakage in rail transit, the method comprising: The radio frequency (RF) signal in the leaky cable is acquired, and the RF signal is coupled and adjusted to convert the RF power into a voltage signal. The standing wave ratio and signal attenuation are determined based on the voltage signal, and wavelet threshold noise reduction is performed on the vibration acceleration of each edge node. Data packets are generated by combining temperature and humidity data. The bidirectional long short-term memory network with attention mechanism is invoked to calculate the health index corresponding to the data packet and predict the trend of health index change within a set time period. The health index and its trend are compared with a set threshold to issue an alarm when the threshold is not met.

[0083] In another aspect, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and when the processor executes the computer instructions, it implements a rail transit cable leakage detection method, the method comprising: The radio frequency (RF) signal in the leaky cable is acquired, and the RF signal is coupled and adjusted to convert the RF power into a voltage signal. The standing wave ratio and signal attenuation are determined based on the voltage signal, and wavelet threshold noise reduction is performed on the vibration acceleration of each edge node. Data packets are generated by combining temperature and humidity data. The bidirectional long short-term memory network with attention mechanism is invoked to calculate the health index corresponding to the data packet and predict the trend of health index change within a set time period. The health index and its trend are compared with a set threshold to issue an alarm when the threshold is not met.

[0084] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory.

[0085] By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0086] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0087] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for detecting leaky cables in rail transit systems, characterized in that, The method includes: The radio frequency (RF) signal in the leaky cable is acquired, and the RF signal is coupled and adjusted to convert the RF power into a voltage signal. The standing wave ratio and signal attenuation are determined based on the voltage signal, and wavelet threshold noise reduction is performed on the vibration acceleration of each edge node. Data packets are generated by combining temperature and humidity data. The bidirectional long short-term memory network with an attention mechanism is invoked to calculate the health index corresponding to the data packet and predict the trend of health index change within a set time period. The health index and its trend are compared with a set threshold to issue an alarm when the set threshold is not met.

2. The method for detecting leaky cables in rail transit according to claim 1, characterized in that, The step of acquiring the radio frequency signal in the leaky cable and coupling and adjusting the radio frequency signal to convert the radio frequency power into a voltage signal includes: The radio frequency signals are coupled in the forward and reverse directions respectively by directional couplers, and the attenuator is called to adjust the coupled radio frequency signals so as to send them to the radio frequency detector to convert the radio frequency power into forward and reverse detection voltages. Calculate the ratio of the forward and reverse detection voltages to obtain the standing wave ratio (SWR), and calculate the difference between the actual forward detection voltage and the calibrated detection voltage to obtain the signal attenuation.

3. The method for detecting leaky cables in rail transit according to claim 2, characterized in that, The process of acquiring the radio frequency (RF) signal in the leaky cable, coupling and adjusting the RF signal to convert the RF power into a voltage signal, further includes: The DC voltage signal from the radio frequency detector is received, and the voltage difference between the positive and negative terminals of the DC voltage signal is amplified by an amplifier while the interference voltage at the positive and negative terminals of the DC voltage signal is suppressed to obtain the amplified differential signal. The differential signal is sent to a digital-to-analog converter for sampling to eliminate interference voltage in the standing wave ratio and signal attenuation.

4. The method for detecting leaky cables in rail transit according to claim 1, characterized in that, The process of determining the standing wave ratio and signal attenuation based on the voltage signal, performing wavelet threshold noise reduction on the vibration acceleration of each edge node, and generating a data packet by combining temperature and humidity data includes: Temperature and humidity data and vibration acceleration in the leaky cable detection area are collected by a temperature and humidity sensor and a triaxial accelerometer, respectively. The temperature and humidity data and vibration acceleration are time-aligned and labeled with device names to obtain an initial data stream. The initial data stream is sent to an edge node, where the edge node performs format verification on the initial data stream.

5. The method for detecting leaky cables in rail transit according to claim 4, characterized in that, The process of determining the standing wave ratio and signal attenuation based on the voltage signal, performing wavelet threshold noise reduction on the vibration acceleration of each edge node, and generating a data packet by combining temperature and humidity data also includes: The acceleration time-domain signal is extracted from the formatted and verified initial data stream, and the wavelet basis function is called to perform wavelet packet decomposition on the acceleration time-domain signal to obtain multi-level detail coefficients and approximation coefficients. Extract the first layer of detail coefficients from the multi-layer detail coefficients, calculate the kurtosis value of the first layer of detail coefficients, and set an adaptive threshold based on the kurtosis value; Wherein, the first layer detail coefficient is any layer detail coefficient among the multi-layer detail coefficients, and the adaptive threshold is inversely proportional to the kurtosis value.

6. The method for detecting leaky cables in rail transit according to claim 5, characterized in that, The process of determining the standing wave ratio and signal attenuation based on the voltage signal, performing wavelet threshold noise reduction on the vibration acceleration of each edge node, and generating a data packet by combining temperature and humidity data also includes: The threshold function is called to shrink the first layer detail coefficients, and the shrunken first layer detail coefficients are reconstructed by wavelet reconstruction with the approximate coefficients of the corresponding layer to obtain the denoised acceleration time domain signal. Vibration features are extracted from the noise-reduced acceleration time-domain signal, and the vibration features, standing wave ratio, signal attenuation, and temperature and humidity data are packaged into the data package.

7. The method for detecting leaky cables in rail transit according to claim 1, characterized in that, The bidirectional long short-term memory network that invokes the attention mechanism calculates the health index corresponding to the data packet and predicts the trend of health index changes within a set time period, including: The bidirectional long short-term memory network of the attention mechanism is trained under supervision using pre-stored historical data and labeled fault parameters. Mean squared error is introduced as the loss function and Adam optimizer to obtain the trained health assessment model. The health assessment model is invoked to calculate the health index corresponding to the data packet at each monitoring point, and the health index is input into a linear regression model or an LSTM prediction model to predict the trend of health index changes within a set time period.

8. A cable leakage detection device for rail transit, characterized in that, For implementing the rail transit cable leakage detection method according to any one of claims 1 to 7, the apparatus comprises: A signal conversion module is used to acquire radio frequency signals in the leaky cable and couple and adjust the radio frequency signals to convert radio frequency power into voltage signals; The data integration module is used to determine the standing wave ratio and signal attenuation based on the voltage signal, perform wavelet threshold noise reduction on the vibration acceleration of each edge node, and generate data packets by combining temperature and humidity data. The model detection module is used to call the bidirectional long short-term memory network with attention mechanism to calculate the health index corresponding to the data packet and predict the trend of health index change within a set time period. The health early warning module is used to compare the health index and its changing trend with a set threshold, so as to issue an alarm when the set threshold is not met.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the rail transit leaky cable detection method according to any one of claims 1 to 7.

10. A computer storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the rail transit leaky cable detection method according to any one of claims 1 to 7.

Images