Urban tunnel limit invasion prevention intelligent early warning method based on distributed acoustic sensing

Abstract

The invention discloses an urban tunnel limit invasion prevention intelligent early warning method based on distributed acoustic sensing, and relates to the field of underground monitoring and rail safety early warning, and the method specifically comprises the steps: inputting a data record of a distributed optical fiber; the records are subjected to standard size windowing, spectral amplitude solving, range intercepting, data truncation and normalization; training a convolutional auto-encoder to reconstruct data, and performing encoding dimension reduction on the data by using an encoder part; training a random forest classifier by using the feature vectors subjected to real-time data preprocessing and dimension reduction; continuous data is preprocessed, and after dimension reduction of the data by a convolution encoder, detection and classification are carried out by using a random forest; and performing post-processing based on signal space-time continuity screening on a detection result to obtain an accurate detection and classification result. According to the invention, the construction signals in the distributed acoustic sensing data are monitored in real time by using the convolution auto-encoder and the random forest algorithm, detection and accurate classification are carried out with high temporal-spatial resolution, and intelligent early warning of track construction based on distributed optical fibers is realized.

Description

Technical Field

[0001] The invention relates to the field of underground monitoring and track safety early warning, and in particular to an urban tunnel intrusion prevention intelligent early warning method based on distributed acoustic sensing. Background Art

[0002] Limit intrusion refers to an object that exceeds the permitted outline dimension line within the limit of safe operation of rail transit. Limit intrusion accidents can affect train operation and threaten the safety of passengers. In serious cases, they may cause major safety accidents that endanger people's lives and property safety. Construction equipment operating above subway tunnels is the main potential threat of intrusion into tunnels. Therefore, timely detection and grasp of construction dynamics above tunnels will be an important measure to protect rail safety.

[0003] Traditional rail inspections are done manually. High-frequency manual inspections can effectively prevent foreign objects from intruding during construction, but the labor cost is too high. Therefore, it is urgent to develop a new technology that can overcome the limitations of manual inspections and achieve real-time tunnel construction monitoring at low labor and material costs. Summary of the invention

[0004] In order to solve the above-mentioned problems, the present invention provides an urban tunnel intrusion prevention and limit intelligent early warning method based on distributed acoustic sensing, which utilizes convolutional autoencoders and random forest algorithms to monitor construction signals in distributed acoustic sensing data in real time, and can detect and accurately classify them with high temporal and spatial resolution, thereby realizing intelligent early warning of track construction based on distributed optical fiber.

[0005] To achieve the above object, the present invention provides an urban tunnel anti-intrusion intelligent early warning method based on distributed acoustic sensing, which specifically includes the following steps: Step S1: inputting vibration data recorded by distributed optical fiber acoustic wave sensing, the vibration data including waveforms detected by all channels; Step S2: Implement sliding windowing and data preprocessing on the continuous distributed acoustic sensor monitoring real-time data using a standardized process; Step S3: Reconstruct the spectrum by training the convolutional autoencoder using continuous data; Step S4: using the encoder to preprocess and compress the real-time data and reduce its dimension, and train the random forest classifier; Step S5: pre-process the real-time data stream and use the encoder to reduce the dimension and then use random forest to classify to obtain the detection result.

[0006] Preferably, the standard window size is determined according to the propagation distance, duration and detection accuracy required of the engineering equipment signal to be detected.

[0007] Preferably, in step S2, data preprocessing includes performing a time axis fast Fourier transform on the single windowed data to the frequency domain, intercepting the required frequency domain range, taking the amplitude, and truncating and normalizing the data using percentiles.

[0008] Preferably, in step S3, the convolutional autoencoder includes an encoder part and a decoder part, the encoder part and the decoder part are composed of multiple convolutional layers with the same total number of layers and symmetry, the residual is set to the mean square error loss between the input and output, and the network is trained with a dynamic learning rate.

[0009] Preferably, in step S4, manually picked noise, subway and construction signal sliding windows and preprocessed real-time data are input into the encoder part, and the encoded feature vector is used to train a detection and classification integrated random forest classifier.

[0010] Preferably, in step S5, the convolutional autoencoder and the random forest classifier are connected, the continuous data sliding window in the real-time data stream and the two-dimensional batch data obtained by preprocessing are flattened and input into the network to obtain the prediction result, and the prediction result is projected back to the two-dimensional space-time domain.

[0011] Preferably, it includes a distributed acoustic sensing device connected to the track optical fiber and a matching mobile workstation for data storage and data processing.

[0012] Preferably, continuity constraint parameters are set according to the duration and propagation distance of the detected construction signal, and detection results with spatiotemporal continuity weaker than a threshold are identified as discrete false detections and removed.

[0013] Therefore, the present invention adopts the above-mentioned urban tunnel anti-intrusion intelligent early warning method based on distributed acoustic sensing, which has the following beneficial effects: (1) In the present invention, distributed acoustic sensing technology uses optical fiber as a sensor and uses the phase of backscattered Rayleigh light to measure the axial strain rate. Distributed acoustic sensing technology can transform tens of kilometers of optical cables into dense sensors arranged at meter intervals, and can perform sampling at a sampling rate far exceeding that of traditional seismic instruments.

[0014] (2) The present invention utilizes the space-time continuity constraint to filter out erroneous recognition and improve the accuracy; and outputs the category, location and time of the construction signal detected.

[0015] (3) The present invention uses convolutional autoencoders and random forest algorithms to monitor construction signals in distributed acoustic sensor data in real time, which can detect and accurately classify them with high temporal and spatial resolution, thus realizing intelligent early warning of track construction based on distributed optical fiber.

[0016] (4) The random forest in the present invention is an integrated learning method for classification and regression, which is a voting machine composed of multiple decision trees. After the two-dimensional frequency domain distributed acoustic sensor data is reduced in dimension by a convolutional autoencoder, a relatively stable classification result is obtained by using the random forest algorithm, thereby realizing the detection and classification of construction signals, and the computational efficiency is high, and real-time monitoring can be achieved.

[0017] (5) The convolutional autoencoder in the present invention is a symmetric neural network composed of multiple convolutional layers, consisting of an encoder and a decoder, and is mainly used in data denoising, data dimensionality reduction and other fields. Fast Fourier transform of distributed acoustic sensor data to the frequency domain can effectively erase time information and solve the translation invariance problem of convolutional neural networks.

[0018] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A method flow chart of an urban tunnel anti-intrusion intelligent early warning method based on distributed acoustic sensing according to the present invention; Figure 2 The waveform domain and frequency domain characteristics of the day and night noise signals recorded by the distributed optical fiber in the present invention; Figure 3 It is the waveform domain and frequency domain characteristics of the day and night subway signals recorded by the distributed optical fiber in the present invention; Figure 4 It is the waveform domain and frequency domain characteristics of various construction signals recorded by the distributed optical fiber in the present invention; Figure 5 It is the structure and flow chart of the convolutional autoencoder combined with random forest in the present invention; Figure 6 This is a diagram showing the reconstruction effect of the convolutional autoencoder on the spectrum in the present invention; Figure 7 This is an effect diagram of using continuity to constrain the detection results in the present invention. DETAILED DESCRIPTION

[0020] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0021] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0022] The words "include" or "comprises" and the like used in the present invention mean that the elements before the word include the elements listed after the word, and do not exclude the possibility of also including other elements. The orientation or position relationship indicated by the terms "inside", "outside", "upper", "lower", etc. is based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation of the present invention. When the absolute position of the described object changes, the relative position relationship may also change accordingly. In the present invention, unless otherwise clearly specified and limited, the terms "attachment" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral body; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0023] like Figure 1 As shown, an urban tunnel intrusion prevention and intelligent early warning method based on distributed acoustic sensing includes a distributed acoustic sensing device connected to a track optical fiber and a matching mobile workstation for data storage and data processing; Use the trained convolutional encoder network and random forest network to achieve data dimension reduction and feature vector detection and classification; Use space-time continuity constraints to filter out false recognition and improve accuracy; output the type, location and time of the detected construction signal.

[0024] The standard window size is determined based on the propagation distance, duration and detection accuracy required of the engineering equipment signal to be detected.

[0025] The continuity constraint parameters are set according to the duration and propagation distance of the detected construction signal, and the detection results with spatiotemporal continuity weaker than the threshold are identified as discrete false detections and removed, thereby reducing the false alarm rate of the system.

[0026] The specific steps include: Step S1: inputting vibration data recorded by distributed optical fiber acoustic wave sensing, the vibration data including waveforms detected by all channels; Step S2: Implement sliding windowing and data preprocessing on the continuous distributed acoustic sensor monitoring real-time data using a standardized process; In step S2, data preprocessing includes performing a time axis fast Fourier transform on the single windowed data to the frequency domain, intercepting the required frequency domain range, taking the amplitude, and truncating and normalizing the data using percentiles.

[0027] Step S3: Reconstruct the spectrum by training the convolutional autoencoder using continuous data; In step S3, the convolutional autoencoder includes an encoder part and a decoder part. The encoder part and the decoder part are composed of multiple convolutional layers with the same total number of layers and symmetry. The residual is set to the mean square error loss between the input and output, and the network is trained with a dynamic learning rate.

[0028] Step S4: using the encoder to preprocess and compress the real-time data and reduce its dimension, and train the random forest classifier; In step S4, the manually picked noise, subway and construction signal sliding windows and preprocessed real-time data are input into the encoder part, and the encoded feature vector is used to train the integrated detection and classification random forest classifier.

[0029] Step S5: Preprocess the real-time data stream and use the encoder to reduce the dimension and then use the random forest to classify to obtain the detection result. In step S5, the convolutional autoencoder and the random forest classifier are connected, the continuous data sliding window in the real-time data stream and the two-dimensional batch data obtained by preprocessing are flattened and input into the network to obtain the prediction result, and the prediction result is projected back to the two-dimensional spatiotemporal domain.

[0030] Example The construction signal detection method applied in the tunnel periphery specifically includes the following steps: The seismic waveform recorded by the distributed optical fiber is windowed by sliding, and fast Fourier transformed to the frequency domain to intercept the response frequency band where the construction signal is located; Use the spectrum of a large amount of continuous data to train the convolutional autoencoder to obtain a reliable encoder as a data feature extraction tool; Use an encoder to reduce the dimension of the spectrum graph to be detected to obtain a compressed feature vector; Manually identify construction events and create labels. After data augmentation, the labels are divided into training and validation sets, with 0 to n representing noise, subway signals, and different construction categories. Compress the labeled event spectrum into feature vectors and train the random forest model; Connect the continuous data to the convolution encoder and random forest network to get the prediction results.

[0031] According to this embodiment, a tunnel periphery construction signal detection system based on distributed fiber optic acoustic sensor frequency domain coding dimensionality reduction and random forest algorithm is proposed, including: implementing windowing and data preprocessing on continuous distributed acoustic sensor monitoring data in a standardized process; connecting the trained convolution encoder and the random forest network, flattening the two-dimensional batch data obtained by sliding the window and inputting it into the network to obtain a prediction result, and projecting the prediction result back to the two-dimensional space-time domain; using the space-time continuity of the construction signal to constrain the prediction result, remove discrete false detections, and output the category, position and time of the construction signal obtained by the detection.

[0032] Distributed acoustic sensing technology uses optical fiber as a sensor and uses the phase of backscattered Rayleigh light to measure the axial strain rate. Distributed acoustic sensing technology can transform tens of kilometers of optical cables into dense sensors arranged at meter intervals, and can sample at a sampling rate far exceeding that of traditional seismic instruments. Communication optical cables and reserved optical cables are often laid in operating subway tunnels, often hung on iron frames on the tunnel walls along the tunnel. These optical cables can be connected to a distributed acoustic sensor interrogator at one end, thus becoming a dense vibration sensor.

[0033] Field experiments in the Hefei Metro have shown that even in the case of poor optical cable-tunnel wall coupling such as wall-mounted, distributed acoustic sensing can still demodulate and obtain high-fidelity tunnel wall vibration signals. Figure 2 As shown in the figure, the background noise signal recorded by the distributed optical fiber in the experiment, the waveform is shown on the top, the spectrum is shown on the bottom, and the noise characteristics on the left and right are daytime and nighttime respectively. The background noise level during the day is significantly higher than that at night, and the "common mode noise" generated by the vibration of the device interrogator itself is clearly visible in the night noise. Figure 3 As shown in the figure, the subway operation signals recorded by the distributed optical fiber in the experiment are the signals generated by the subway trains running in the upward and downward directions on the left and right. Since the optical cable is hung on the wall of the subway tunnel in the downward direction, the wind pressure signal generated by the subway operation can be recorded. Figure 4 As shown in the figure, the four types of construction machinery signals recorded by distributed optical fiber in the experiment are from excavators, road rollers, crushers, and drilling machines on the upper left, upper right, lower left, and lower right. It can be seen that distributed acoustic sensing records can accurately reflect the frequency range, spatial distribution and other characteristics of various construction signals.

[0034] In the data preprocessing stage, the continuous data of the distributed optical fiber is downsampled to 200 Hz in time and multi-channel superimposed to 10 meters in space. The data is windowed with 10 channels and 20 seconds, and data augmentation is achieved by adjusting the step size. The continuous data is fast Fourier transformed on the time axis to obtain the spectrum, retaining only the amplitude of the 0-80 Hz part to obtain 10×1600 two-dimensional data. The data is truncated and normalized with the fifth percentile of the spectrum as the lower bound and the ninety-fifth percentile as the upper bound.

[0035] like Figure 5 As shown in the figure, the upper part shows the structure of the convolutional autoencoder neural network, which consists of 7 convolutional layers and 7 transposed convolutional layers symmetrical to the former. The first half is the encoder and the second half is the decoder. The encoder compresses and encodes the input data of 1×10×1600 to obtain a feature vector of size 8×2×13=208, and the decoder restores the feature vector to data of the original size. The training goal is to make the residual difference between the output of the decoder and the input of the encoder as small as possible, so that the encoded feature vector can represent the main features of the input data. Figure 6 As shown in the figure, the convolutional autoencoder reconstructs the truncated spectrum. The good consistency between the input and output illustrates the reliability of the autoencoder coding compression.

[0036] All manually labeled data are preprocessed and then input into the autoencoder to obtain their corresponding feature vectors. Use the feature vectors and corresponding labels to train Figure 5 The random forest model shown in the lower middle part obtains a feature vector classifier to determine whether the feature vector belongs to noise or a certain type of construction signal. The subway signal can be output as a separate class or merged with the noise into one class. The preprocessor, convolutional encoder and random forest classifier are connected to obtain a detector that can directly detect and classify construction signals by sliding windows on the original two-dimensional data. The detection precision and recall of the first model on the test set are shown in Table 1. The test results show that at least 89.23% of the construction alarms made by the artificial intelligence model are real alarms, and 70.62% of the manually detected events are also detected by the model. The model considers the different destructive capabilities of the four types of construction on the ground and sets different weight balancing strategies. For example, the signal of the excavator is not significant and the invasiveness is not strong, so the model tries to improve its precision as much as possible and reduce the requirement for its recall; the drilling machine has the greatest invasive risk, and it is necessary to ensure the highest possible recall. The precision rates of the four types of construction machinery, namely excavators, road rollers, crushers and drilling machines, were 76.65%, 92.71%, 87.39% and 76.63% respectively, and the recall rates were 25.55%, 73.85%, 71.39% and 81.84% respectively. This shows that the model can effectively monitor and warn the construction situation on the ground.

[0037] Table 1 ;

[0038] In practical applications, the early warning system requires a lower false alarm rate. The post-processing stage of the model can add space-time continuity constraints to reduce false alarms. Combined with the duration and propagation distance of the construction signal, the continuity parameter requires that the result of the sliding window detection has corresponding continuity in space and time, otherwise it will be judged as a false detection. Figure 7 As shown in the figure, the effect of continuity constraint. For a certain section of distributed acoustic sensor recording data, this algorithm is used for detection after sliding the window, where the horizontal axis corresponds to distance, the vertical axis corresponds to time, and the blocks of different colors represent the various types of construction signals detected. The upper figure is the original output, and the lower figure is the output after screening with the spatiotemporal continuity constraint. Table 2 shows the test results of a single event after adding the continuity constraint. At this time, the precision rate of the model for construction signals is increased to 99.34%, which greatly reduces the false detection rate and improves practicality. After testing, this algorithm can be connected to the distributed acoustic sensor data stream for real-time detection.

[0039] Table 2 ;

[0040] Therefore, the present invention adopts the above-mentioned urban tunnel intrusion prevention and limit intelligent early warning method based on distributed acoustic sensing, uses convolutional autoencoder and random forest algorithm to monitor the construction signals in distributed acoustic sensing data in real time, detects and accurately classifies with high temporal and spatial resolution, and realizes intelligent early warning of track construction based on distributed optical fiber.

[0041] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. An urban tunnel anti-intrusion intelligent early warning method based on distributed acoustic sensing, characterized in that: The specific steps include: Step S1: inputting vibration data recorded by distributed optical fiber acoustic wave sensing, the vibration data including waveforms detected by all channels; Step S2: Implement sliding windowing and data preprocessing on the continuous distributed acoustic sensor monitoring real-time data using a standardized process; Step S3: Reconstruct the spectrum by training the convolutional autoencoder using continuous data; Step S4: using the encoder to preprocess and compress the real-time data and reduce its dimension, and train the random forest classifier; Step S5: pre-process the real-time data stream and use the encoder to reduce the dimension and then use random forest to classify to obtain the detection result.

2. According to the method of claim 1, the method is characterized by: The standard window size is determined based on the propagation distance, duration and detection accuracy required of the engineering equipment signal to be detected.

3. According to the method of claim 1, the method is characterized by: In step S2, data preprocessing includes performing a time axis fast Fourier transform on the single windowed data to the frequency domain, intercepting the required frequency domain range, taking the amplitude, and truncating and normalizing the data using percentiles.

4. According to the method of claim 3, the method is characterized by: In step S3, the convolutional autoencoder includes an encoder part and a decoder part. The encoder part and the decoder part are composed of multiple convolutional layers with the same total number of layers and symmetry. The residual is set to the mean square error loss between the input and output, and the network is trained with a dynamic learning rate.

5. According to the method of claim 4, the method is characterized by: In step S4, the manually picked noise, subway and construction signal sliding windows and preprocessed real-time data are input into the encoder part, and the encoded feature vector is used to train the integrated detection and classification random forest classifier.

6. According to the method of claim 5, the method is characterized by: In step S5, the convolutional autoencoder and the random forest classifier are connected, the continuous data sliding window in the real-time data stream and the two-dimensional batch data obtained by preprocessing are flattened and input into the network to obtain the prediction result, and the prediction result is projected back to the two-dimensional spatiotemporal domain.

7. A system for an urban tunnel anti-intrusion intelligent early warning method based on distributed acoustic sensing according to any one of claims 1 to 6, characterized in that: It includes distributed acoustic sensing equipment connected to the track optical fiber and a matching mobile workstation for data storage and data processing.

8. The distributed optical fiber construction signal detection system based on autoencoder and random forest according to claim 7, characterized in that: The continuity constraint parameters are set according to the duration and propagation distance of the detected construction signal, and the detection results with spatiotemporal continuity weaker than the threshold are identified as discrete false detections and removed.

Images