Tunnel safety monitoring method based on multiple sensors

Abstract

The present invention discloses a tunnel safety monitoring method based on multiple sensors, the method comprising signal acquisition, establishing a data set, establishing a tunnel safety monitoring model, and tunnel safety monitoring. The present invention belongs to the technical field of tunnel safety monitoring, and specifically refers to a tunnel safety monitoring method based on multiple sensors. The present scheme extracts signal information through singular spectrum decomposition and residual signal processing; retains the phase information of the original signal by using Hilbert transform and phase sequence calculation; makes judgments based on the complexity of the residual signal, thereby improving the adaptability and efficiency of signal decomposition; introduces adjustable parameters when designing the activation function, and adopts an exponential decay function within the negative range to effectively map negative inputs to a smaller output range, thereby avoiding the gradient explosion problem; designs an initial movement strategy by balancing random search and refined search, and randomly selects a portion of individuals in the population for random movement; and improves the search convergence speed and quality.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnel safety monitoring, and in particular to a tunnel safety monitoring method based on multiple sensors. Background Art

[0002] Tunnel safety monitoring utilizes sensors, data processing technology, and model building and optimization to monitor and evaluate tunnel structures, environments, and operational status in real time. Its primary purpose is to identify potential safety hazards within tunnels and implement timely preventive and control measures to ensure safe tunnel operation. However, conventional tunnel safety monitoring methods suffer from excessive noise in the original signal, excessive decomposition during signal decomposition, and poor signal information retention. Furthermore, conventional tunnel safety monitoring methods suffer from instability in the model learning process due to improperly designed model activation functions, as well as slow model convergence and low accuracy during parameter search due to improper model parameter settings. Summary of the Invention

[0003] To address the above-mentioned issues and overcome the shortcomings of the prior art, the present invention provides a tunnel safety monitoring method based on multiple sensors. This method addresses the problems of excessive noise content in the original signal, over-decomposition during the signal decomposition process, and poor signal information retention in conventional tunnel safety monitoring methods. This method extracts signal information through singular spectrum decomposition and residual signal processing. Hilbert transform and phase sequence calculation are used to preserve the phase information of the original signal. A decision is made based on the complexity of the residual signal, effectively avoiding over-decomposition of unnecessary signals and improving the adaptability and efficiency of signal decomposition. Furthermore, in conventional tunnel safety monitoring methods, problems such as instability in the model learning process caused by improper model activation function design, slow model convergence, and low accuracy in parameter search due to improper model parameter setting are addressed. This method introduces adjustable parameters into the activation function design and uses an exponential decay function within the negative range, making the model learning process more flexible and effectively mapping negative inputs to a smaller output range, thus avoiding the gradient explosion problem. An initial movement strategy is designed by balancing random search and refined search, and a portion of individuals in the population are randomly selected for random movement. This improves the search convergence speed and quality, thereby improving model accuracy and the speed of model establishment.

[0004] The technical solution adopted by the present invention is as follows: The present invention provides a tunnel safety monitoring method based on multiple sensors, which includes the following steps:

[0005] Step S1: signal acquisition;

[0006] Step S2: Create a data set;

[0007] Step S3: Establishing a tunnel safety monitoring model;

[0008] Step S4: Tunnel safety monitoring.

[0009] Furthermore, in step S1, the signal acquisition is based on sensors collecting historical tunnel signal data; the historical tunnel signal data includes environmental parameters, structural parameters, gas concentration, traffic flow, sound signals and tunnel safety status; and the tunnel safety status is used as a data label.

[0010] Furthermore, in step S2, the establishment of the data set specifically includes the following steps:

[0011] Step S21: pre-decomposition; decompose the signal to be decomposed based on singular spectrum decomposition, set the number of components to 1, and take the only single singular spectrum component obtained by decomposition as the result; iterate k times to obtain the component; it is expressed as follows:

[0012] ;

[0013] Where, SSC k is the singular spectrum component obtained after k iterations; y is the signal to be decomposed; SSD(·) is the singular spectrum decomposition process;

[0014] Step S22: Obtain the residual signal. Based on the residual signal, calculate the error between the residual signal and the original signal in the kth iteration. An error threshold is set in advance. When the error is lower than the error threshold, the decomposition is completed and go to step S24; otherwise, go to step S23. The formula used is as follows:

[0015] ;

[0016] Where res is the residual signal, SSC i1 is the singular spectral component at the i1th iteration;

[0017] Step S23: Evaluate the complexity; a complexity threshold is pre-set and a determination is made after each signal decomposition. When the complexity of the singular spectrum component is higher than the complexity threshold, the residual signal is determined to be a useless signal; otherwise, the residual signal is regarded as the original signal for the next decomposition; specifically, the following steps are included:

[0018] Step S231: Hilbert transform, the formula used is as follows:

[0019] ;

[0020] ;

[0021] Where Z(i2) is the transformed complex sequence; Y(i2) is the signal to be decomposed; H[·] is the Hilbert transform; j is the imaginary unit; A(i2) is the amplitude part; g(i2) is the phase angle; is the value of the signal to be processed at the time delay τ;

[0022] Step S232: Calculate the phase sequence using the following formula:

[0023] ;

[0024] Where θ(·) is the phase sequence element, i2 is the phase sequence element index;

[0025] Step S233: Reconstruct the phase space. Based on the permutation entropy, the phase space is reconstructed using the phase information of each signal. The formula used is as follows:

[0026] ;

[0027] Where S(j1) is the reconstructed component in the reconstructed phase space, j1 is the starting position index; m is the reconstruction parameter, which indicates the number of phase angles contained in each reconstructed component;

[0028] Step S234: Arrange the reconstructed components in ascending order of value, extract the position index of each element in the original reconstructed sequence in ascending order, and record the position index as l, thereby obtaining the corresponding position sequence; and calculate the probability of occurrence of the same position sequence; the formula used is as follows:

[0029] ;

[0030] Where, P i2 is the probability of occurrence of the i2th identical position sequence; Num i2 is the number of identical position sequences; N is the total number of reconstructed sequences;

[0031] Step S235: Evaluate signal complexity using the following formula:

[0032] ;

[0033] Where PG is the evaluation value of signal complexity; p is the normalization parameter; q is the expansion parameter used to increase the sensitivity to extreme probabilities; is a weight parameter used to adjust the probability weighting; sign(·) is a sign function;

[0034] Step S24: Divide the data set; obtain a data set based on the decomposed signal and divide it into a training set and a test set.

[0035] Furthermore, in step S3, the tunnel safety monitoring model is established by establishing a BP neural network based on the data set divided in step S2, and optimizing the model parameters; specifically, the following steps are included:

[0036] Step S31: Initialization, initializing the connection weights and biases of neurons;

[0037] Step S32: Forward propagation, first calculate the input of the hidden layer neurons, and then perform nonlinear transformation on the input through the activation function to obtain the output of the hidden layer neurons; it is expressed as follows:

[0038] ;

[0039] ;

[0040] Where u i is the input of the i-th hidden layer neuron; ω mi is the connection weight from the mth neuron in the input layer to the ith neuron in the hidden layer; b is the bias of the hidden layer; v i is the output of the i-th neuron in the hidden layer; f(·) is the sigmoid activation function; M is the total number of neurons; x m is the input of the mth neuron in the input layer;

[0041] Step S33: Calculate the output; the formula used is as follows:

[0042] ;

[0043] ;

[0044] ;

[0045] Where u j is the input of the jth neuron; ω ij is the connection weight from the i-th neuron in the hidden layer to the j-th neuron in the output layer; is the threshold of the i-th neuron in the output layer; g(·) is the activation function; x is the activation function input; α and is the activation function parameter that controls the shape of the function;

[0046] Step S34: Back propagation, according to the error back propagation, reverse update weights and thresholds; when the model converges or reaches the maximum number of training times, the tunnel safety monitoring model training is completed;

[0047] Step S35: Model parameter optimization, optimizing the initial parameters of the neural network and the activation function parameters; specifically including:

[0048] Step S351: Optimization initialization: establishing a search space based on the initial parameters of the neural network and the activation function parameters; randomly initializing the optimization population, and using the prediction accuracy of the tunnel safety monitoring model trained based on the position of the optimization individual on the test set as the fitness value of the optimization individual;

[0049] Step S352: Design the initial movement strategy and select the top 10% individuals with the best fitness as the preferred individuals; balance the random search and refined search, and the formula used is as follows:

[0050] ;

[0051] Where, and are the positions of the dth dimension of the i-th optimized individual at the t+1th iteration and the tth iteration respectively; r1 is a random number in the range of 0 to 1; f i is the fitness value of the optimized individual; f best is the maximum fitness value of the population; is the position of the bi-th preferred individual in the d-th dimension at the t-th iteration; BI is the total number of preferred individuals; is the historical optimal position of the optimized individual in the dth dimension at the tth iteration; β is a random parameter; γ is the movement determination threshold;

[0052] Step S353: Random movement: randomly select 10% of the individuals in the population for random movement. The formula used is as follows:

[0053] ;

[0054] Where, is the optimized individual position after random movement; r2 is a random number in the range of 0 to 1, independent of r1; far(·) is the individual position farthest from the target individual;

[0055] Step S354: Determination: a fitness threshold is set in advance. When there is an optimized individual whose fitness threshold is higher than the fitness threshold, the tunnel safety monitoring model is established. If the maximum number of iterations is reached, return to step S351; otherwise, continue the iterative search.

[0056] Furthermore, in step S4, the tunnel safety monitoring is to collect real-time tunnel signal data based on sensors for the established tunnel safety monitoring model, and the real-time tunnel signal data includes environmental parameters, structural parameters, gas concentration, traffic flow and sound signals; after signal decomposition, the data is input into the tunnel safety monitoring model; and tunnel safety monitoring is realized based on the output of the tunnel safety monitoring model.

[0057] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0058] (1) In view of the problems of excessive noise content in the original signal, over-decomposition and poor signal information retention in the signal decomposition process in general tunnel safety monitoring methods, this scheme extracts signal information through singular spectrum decomposition and residual signal processing; retains the phase information of the original signal by using Hilbert transform and phase sequence calculation; and makes judgments based on the complexity of the residual signal, effectively avoiding over-decomposition of unnecessary signals and improving the adaptability and efficiency of signal decomposition.

[0059] (2) In view of the problems in general tunnel safety monitoring methods such as improper design of model activation function leading to unstable model learning process, improper setting of model parameters leading to slow model convergence and low accuracy when searching parameters, this scheme introduces adjustable parameters when designing activation function and adopts exponential decay function in the negative range to make the model learning process more flexible, effectively map negative input to a smaller output range, and avoid gradient explosion problem; design the initial movement strategy by weighing random search and refined search, and randomly select a part of individuals in the population for random movement; improve the search convergence speed and quality; and thus improve the accuracy and establishment speed of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A schematic diagram of the flow of the tunnel safety monitoring method based on multiple sensors provided by the present invention;

[0061] Figure 2 Schematic diagram of the process of step S2;

[0062] Figure 3 Schematic diagram of the process of step S3;

[0063] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0064] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0065] In the description of the present invention, it should be understood that terms such as "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they should not be understood as limiting the present invention.

[0066] Example 1, see Figure 1 The present invention provides a tunnel safety monitoring method based on multiple sensors, which includes the following steps:

[0067] Step S1: Signal acquisition, collecting historical tunnel signal data based on sensors;

[0068] Step S2: Establishing a data set, extracting signal information through singular spectrum decomposition and residual signal processing, and retaining the phase information of the original signal by using Hilbert transform and phase sequence calculation, thereby completing the establishment of the data set;

[0069] Step S3: Establishing a tunnel safety monitoring model by introducing adjustable parameters when designing the activation function and adopting an exponential decay function within a negative range; designing an initial movement strategy by balancing random search and refined search, and randomly selecting a portion of individuals in the population for random movement to optimize the model parameters; thereby establishing a tunnel safety monitoring model;

[0070] Step S4: Tunnel safety monitoring.

[0071] Example 2, see Figure 1 This embodiment is based on the above embodiment. In step S1, the historical tunnel signal data includes environmental parameters, structural parameters, gas concentration, traffic flow, sound signals and tunnel safety status; the tunnel safety status is used as a data label.

[0072] Example 3, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. In step S2, establishing a data set specifically includes the following steps:

[0073] Step S21: pre-decomposition; decompose the signal to be decomposed based on singular spectrum decomposition, set the number of components to 1, and take the only single singular spectrum component obtained by decomposition as the result; iterate k times to obtain the component; it is expressed as follows:

[0074] ;

[0075] Where, SSC kis the singular spectrum component obtained after k iterations; y is the signal to be decomposed; SSD(·) is the singular spectrum decomposition process;

[0076] Step S22: Obtain the residual signal. Based on the residual signal, calculate the error between the residual signal and the original signal in the kth iteration. An error threshold is set in advance. When the error is lower than the error threshold, the decomposition is completed and go to step S24; otherwise, go to step S23. The formula used is as follows:

[0077] ;

[0078] Where res is the residual signal, SSC i1 is the singular spectral component at the i1th iteration;

[0079] Step S23: Evaluate the complexity; a complexity threshold is pre-set and a determination is made after each signal decomposition. When the complexity of the singular spectrum component is higher than the complexity threshold, the residual signal is determined to be a useless signal; otherwise, the residual signal is regarded as the original signal for the next decomposition; specifically, the following steps are included:

[0080] Step S231: Hilbert transform, the formula used is as follows:

[0081] ;

[0082] ;

[0083] Where Z(i2) is the transformed complex sequence; Y(i2) is the signal to be decomposed; H[·] is the Hilbert transform; j is the imaginary unit; A(i2) is the amplitude part; g(i2) is the phase angle; is the value of the signal to be processed at the time delay τ;

[0084] Step S232: Calculate the phase sequence using the following formula:

[0085] ;

[0086] Where θ(·) is the phase sequence element, i2 is the phase sequence element index;

[0087] Step S233: Reconstruct the phase space. Based on the permutation entropy, the phase space is reconstructed using the phase information of each signal. The formula used is as follows:

[0088] ;

[0089] Where S(j1) is the reconstructed component in the reconstructed phase space, j1 is the starting position index; m is the reconstruction parameter, which indicates the number of phase angles contained in each reconstructed component;

[0090] Step S234: Arrange the reconstructed components in ascending order of value, extract the position index of each element in the original reconstructed sequence in ascending order, and record the position index as l, thereby obtaining the corresponding position sequence; and calculate the probability of occurrence of the same position sequence; the formula used is as follows:

[0091] ;

[0092] Where, P i2 is the probability of occurrence of the i2th identical position sequence; Num i2 is the number of identical position sequences; N is the total number of reconstructed sequences;

[0093] Step S235: Evaluate signal complexity using the following formula:

[0094] ;

[0095] Where PG is the evaluation value of signal complexity; q is an expansion parameter used to increase the sensitivity to extreme probabilities; is a weight parameter used to adjust the probability weighting; sign(·) is a sign function;

[0096] Step S24: Divide the data set; obtain a data set based on the decomposed signal and divide it into a training set and a test set.

[0097] By performing the above operations, the present scheme addresses the problems of excessive noise content in the original signal, over-decomposition during the signal decomposition process, and poor ability to retain signal information in general tunnel safety monitoring methods. This scheme extracts signal information through singular spectrum decomposition and residual signal processing; retains the phase information of the original signal by using Hilbert transform and phase sequence calculation; and makes judgments based on the complexity of the residual signal, effectively avoiding over-decomposition of unnecessary signals and improving the adaptability and efficiency of signal decomposition.

[0098] Example 4, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. In step S3, a tunnel safety monitoring model is established by establishing a BP neural network based on the data set divided in step S2, and optimizing the model parameters. Specifically, the following steps are included:

[0099] Step S31: Initialization, initializing the connection weights and biases of neurons;

[0100] Step S32: Forward propagation, first calculate the input of the hidden layer neurons, and then perform nonlinear transformation on the input through the activation function to obtain the output of the hidden layer neurons; it is expressed as follows:

[0101] ;

[0102] ;

[0103] Where u i is the input of the i-th hidden layer neuron; ω mi is the connection weight from the mth neuron in the input layer to the ith neuron in the hidden layer; b is the bias of the hidden layer; v i is the output of the i-th neuron in the hidden layer; f(·) is the sigmoid activation function; M is the total number of neurons; x m is the input of the mth neuron in the input layer;

[0104] Step S33: Calculate the output; the formula used is as follows:

[0105] ;

[0106] ;

[0107] ;

[0108] Where u j is the input of the jth neuron; ω ij is the connection weight from the i-th neuron in the hidden layer to the j-th neuron in the output layer; is the threshold of the i-th neuron in the output layer; g(·) is the activation function; x is the activation function input; α and is the activation function parameter that controls the shape of the function;

[0109] Step S34: Back propagation, according to the error back propagation, reverse update weights and thresholds; when the model converges or reaches the maximum number of training times, the tunnel safety monitoring model training is completed;

[0110] Step S35: Model parameter optimization, optimizing the initial parameters of the neural network and the activation function parameters; specifically including:

[0111] Step S351: Optimization initialization: establishing a search space based on the initial parameters of the neural network and the activation function parameters; randomly initializing the optimization population, and using the prediction accuracy of the tunnel safety monitoring model trained based on the position of the optimization individual on the test set as the fitness value of the optimization individual;

[0112] Step S352: Design the initial movement strategy and select the top 10% individuals with the best fitness as the preferred individuals; balance the random search and refined search, and the formula used is as follows:

[0113] ;

[0114] Where, and are the positions of the dth dimension of the i-th optimized individual at the t+1th iteration and the tth iteration respectively; r1 is a random number in the range of 0 to 1; f i is the fitness value of the optimized individual; f best is the maximum fitness value of the population; is the position of the bi-th preferred individual in the d-th dimension at the t-th iteration; BI is the total number of preferred individuals; is the historical optimal position of the optimized individual in the dth dimension at the tth iteration; β is a random parameter; γ is the movement determination threshold;

[0115] Step S353: Random movement: randomly select 10% of the individuals in the population for random movement. The formula used is as follows:

[0116] ;

[0117] Where, is the optimized individual position after random movement; r2 is a random number in the range of 0 to 1, independent of r1; far(·) is the individual position farthest from the target individual;

[0118] Step S354: Determination: a fitness threshold is set in advance. When there is an optimized individual whose fitness threshold is higher than the fitness threshold, the tunnel safety monitoring model is established. If the maximum number of iterations is reached, return to step S351; otherwise, continue the iterative search.

[0119] By performing the above operations, the problems of improper design of model activation functions leading to unstable model learning process, improper setting of model parameters leading to slow model convergence and low accuracy when searching parameters in general tunnel safety monitoring methods are solved. This scheme introduces adjustable parameters when designing the activation function and adopts an exponential decay function within the negative range to make the model learning process more flexible, effectively mapping negative inputs to a smaller output range and avoiding the gradient explosion problem; designs the initial movement strategy by balancing random search and refined search, and randomly selects a part of the population for random movement; improves the search convergence speed and quality; and thereby improves the accuracy and establishment speed of the model.

[0120] Example 5, see Figure 1 This embodiment is based on the above embodiment. In step S4, tunnel safety monitoring is to collect real-time tunnel signal data based on sensors for the established tunnel safety monitoring model. The real-time tunnel signal data includes environmental parameters, structural parameters, gas concentration, traffic flow and sound signals; after signal decomposition, it is input into the tunnel safety monitoring model; and tunnel safety monitoring is realized based on the output of the tunnel safety monitoring model.

[0121] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0122] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

[0123] The present invention and its embodiments are described above. This description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if a person skilled in the art is inspired by this and, without departing from the purpose of the present invention, designs structures and embodiments similar to this technical solution without inventiveness, they shall fall within the scope of protection of the present invention.

Claims

1. A tunnel safety monitoring method based on multiple sensors, characterized by: The method comprises the following steps: Step S1: signal acquisition; Step S2: Establishing a data set, extracting signal information through singular spectrum decomposition and residual signal processing, and retaining the phase information of the original signal by using Hilbert transform and phase sequence calculation, thereby establishing a data set; Step S3: Establishing a tunnel safety monitoring model by introducing adjustable parameters when designing the activation function and adopting an exponential decay function within a negative range; designing an initial movement strategy by balancing random search and refined search, and randomly selecting a portion of individuals in the population for random movement to optimize the model parameters; thereby establishing a tunnel safety monitoring model; Step S4: tunnel safety monitoring; In step S3, the tunnel safety monitoring model is established by establishing a BP neural network based on the data set divided in step S2 and optimizing the model parameters. Specifically, the steps include: Step S31: Initialization, initializing the connection weights and biases of neurons; Step S32: Forward propagation, first calculate the input of the hidden layer neurons, and then perform nonlinear transformation on the input through the activation function to obtain the output of the hidden layer neurons; it is expressed as follows: ； ； Where u i is the input of the i-th hidden layer neuron; ω mi is the connection weight from the mth neuron in the input layer to the ith neuron in the hidden layer; b is the bias of the hidden layer; v i is the output of the i-th neuron in the hidden layer; f(·) is the sigmoid activation function; M is the total number of neurons; x m is the input of the mth neuron in the input layer; Step S33: Calculate the output; the formula used is as follows: ； ； ； Where u j is the input of the jth neuron; ω ij is the connection weight from the i-th neuron in the hidden layer to the j-th neuron in the output layer; is the threshold of the i-th neuron in the output layer; g(·) is the activation function; x is the activation function input; α and is the activation function parameter that controls the shape of the function; Step S34: Back propagation, according to the error back propagation, reverse update weights and thresholds; when the model converges or reaches the maximum number of training times, the tunnel safety monitoring model training is completed; Step S35: Model parameter optimization.

2. The tunnel safety monitoring method based on multiple sensors according to claim 1 is characterized by: In step S2, the establishment of the data set specifically includes the following steps: Step S21: pre-decomposition; decompose the signal to be decomposed based on singular spectrum decomposition, set the number of components to 1, and take the only single singular spectrum component obtained by decomposition as the result; iterate k times to obtain the component; it is expressed as follows: ； Where, SSC k is the singular spectrum component obtained after k iterations; y is the signal to be decomposed; SSD(·) is the singular spectrum decomposition process; Step S22: Obtain the residual signal. Based on the residual signal, calculate the error between the residual signal and the original signal in the kth iteration. An error threshold is set in advance. When the error is lower than the error threshold, the decomposition is completed and go to step S24; otherwise, go to step S23. The formula used is as follows: ； Where res is the residual signal, SSC i1 is the singular spectral component at the i1th iteration; Step S23: Evaluate complexity; Step S24: Divide the data set; obtain a data set based on the decomposed signal and divide it into a training set and a test set.

3. The tunnel safety monitoring method based on multiple sensors according to claim 2 is characterized by: In step S23, the complexity evaluation is performed by presetting a complexity threshold, and making a judgment after each signal decomposition. When the complexity of the singular spectrum component is higher than the complexity threshold, the residual signal is judged as a useless signal; otherwise, the residual signal is regarded as the original signal of the next decomposition. Specifically, the following steps are performed: Step S231: Hilbert transform, the formula used is as follows: ； ； Where Z(i2) is the transformed complex sequence; Y(i2) is the signal to be decomposed; H[·] is the Hilbert transform; j is the imaginary unit; A(i2) is the amplitude part; g(i2) is the phase angle; is the value of the signal to be processed at the time delay τ; Step S232: Calculate the phase sequence using the following formula: ； Where θ(·) is the phase sequence element, i2 is the phase sequence element index; Step S233: Reconstruct the phase space. Based on the permutation entropy, the phase space is reconstructed using the phase information of each signal. The formula used is as follows: ； Where S(j1) is the reconstructed component in the reconstructed phase space, j1 is the starting position index; m is the reconstruction parameter, which indicates the number of phase angles contained in each reconstructed component; Step S234: Arrange the reconstructed components in ascending order of value, extract the position index of each element in the original reconstructed sequence in ascending order, and record the position index as l, thereby obtaining the corresponding position sequence; and calculate the probability of occurrence of the same position sequence; the formula used is as follows: ； Where, P i2 is the probability of occurrence of the i2th identical position sequence; Num i2 is the number of identical position sequences; N is the total number of reconstructed sequences; Step S235: Evaluate signal complexity using the following formula: ； Where PG is the evaluation value of signal complexity; q is an expansion parameter used to increase the sensitivity to extreme probabilities; is a weight parameter used to adjust the probability weighting; sign(·) is a sign function.

4. The tunnel safety monitoring method based on multiple sensors according to claim 3 is characterized by: The model parameter optimization in step S35 is to optimize the initial parameters of the neural network and the activation function parameters; specifically, it includes: Step S351: Optimization initialization: establishing a search space based on the initial parameters of the neural network and the activation function parameters; randomly initializing the optimization population, and using the prediction accuracy of the tunnel safety monitoring model trained based on the position of the optimization individual on the test set as the fitness value of the optimization individual; Step S352: Design the initial movement strategy and select the top 10% individuals with the best fitness as the preferred individuals; balance the random search and refined search, and the formula used is as follows: ； Where, and are the positions of the dth dimension of the i-th optimized individual at the t+1th iteration and the tth iteration respectively; r1 is a random number in the range of 0 to 1; f i is the fitness value of the optimized individual; f best is the maximum fitness value of the population; is the position of the bi-th preferred individual in the d-th dimension at the t-th iteration; BI is the total number of preferred individuals; is the historical optimal position of the optimized individual in the dth dimension at the tth iteration; β is a random parameter; γ is the movement determination threshold; Step S353: Random movement: randomly select 10% of the individuals in the population for random movement. The formula used is as follows: ； Where, is the optimized individual position after random movement; r2 is a random number in the range of 0 to 1, independent of r1; far(·) is the individual position farthest from the target individual; Step S354: Determination: a fitness threshold is set in advance. When there is an optimized individual whose fitness threshold is higher than the fitness threshold, the tunnel safety monitoring model is established. If the maximum number of iterations is reached, return to step S351; otherwise, continue the iterative search.

5. The tunnel safety monitoring method based on multiple sensors according to claim 1 is characterized by: In step S4, the tunnel safety monitoring is to collect real-time tunnel signal data based on the established tunnel safety monitoring model based on sensors. The real-time tunnel signal data includes environmental parameters, structural parameters, gas concentration, traffic flow and sound signals; After signal decomposition, it is input into the tunnel safety monitoring model; Tunnel safety monitoring is achieved based on the output of the tunnel safety monitoring model.

6. The tunnel safety monitoring method based on multiple sensors according to claim 1 is characterized by: In step S1, the signal acquisition is based on the collection of historical tunnel signal data by sensors; the historical tunnel signal data includes environmental parameters, structural parameters, gas concentration, traffic flow, sound signals and tunnel safety status; Use the tunnel security status as a data tag.

Images