Tunnel safety state evaluation and prediction method based on lining damage evolution

Abstract

The invention relates to a tunnel safety state evaluation and prediction method based on lining damage evolution in the technical field of tunnel engineering. The method comprises the following steps: acquiring tunnel lining crack data, inputting the tunnel lining crack data into a TF-Transform deep learning analysis model, outputting the tunnel lining crack data to obtain damage characteristics, summarizing a damage evolution mechanism and a development rule of a cracked tunnel, and establishing a cracked tunnel safety evaluation index system; a PL-VIKOR comprehensive evaluation model is constructed based on a crack tunnel safety evaluation index system and an improved subjective and objective combination weighting method based on normal distribution goodness of fit; establishing a multi-dimensional crack lining monitoring system, and monitoring real-time tunnel section data; and predicting a tunnel section development result by adopting an LSTM model based on real-time tunnel section data, and evaluating a future safety state of the tunnel in combination with a PL-VIKOR comprehensive evaluation model. And real-time and objective evaluation of the safety state of the cracked tunnel structure is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tunnel engineering, in particular to a tunnel safety state evaluation and prediction method based on lining damage evolution. BACKGROUND

[0002] Layered soft rock is a typical rock mass widely distributed in the rugged mountainous areas of western China, which has significant rock anisotropy characteristics. Under the environment of high ground stress, layered soft rock is prone to induce significant non-uniform large deformation, and such deformation is one of the core causes of serious diseases such as cracking, spalling and even collapse of the lining structure of the operating tunnel.

[0003] At present, the damage evolution mechanism and development law of the layered soft rock cracking and damage tunnel under high ground stress in the operation period are still lack of systematic and in-depth research. In engineering practice, the traditional treatment method has obvious limitations for the existing lining cracking and damage diseases, and it is difficult to meet the long-term treatment needs under complex geological conditions. On the one hand, the existing cracking and damage identification and evaluation highly depend on periodic manual inspection, which relies on manual experience, has low efficiency and insufficient coverage, and is difficult to capture the early germination and rapid development stage of damage, resulting in delayed treatment decision and missing the best intervention opportunity. On the other hand, traditional treatment schemes such as grouting reinforcement, embedding, and lining, are often based on fixed standards or empirical formulas, without fully considering the differences in key factors such as ground stress state, bedding occurrence, rock mass mechanical properties, damage evolution stage, and environmental load (such as groundwater) of specific tunnel sections. This "one-size-fits-all" treatment method cannot accurately match the complex and variable actual working conditions, resulting in unstable treatment effect, disease recurrence, and even resource waste.

[0004] Therefore, in the related art, there is an urgent need for a method that can improve the accuracy and response efficiency of tunnel safety state evaluation and prediction. SUMMARY

[0005] Therefore, in the related art, there is an urgent need for a method that can improve the accuracy and response efficiency of tunnel safety state evaluation and prediction.

[0006] In a first aspect, the present application provides a tunnel safety state evaluation and prediction method based on lining damage evolution. The method comprises: Collecting tunnel lining cracking and damage data based on a lining structure-stratum composite simulation test system, inputting the data into a TF-Transformer deep learning analysis model to output damage characteristics, and summarizing the damage evolution mechanism and development law of the cracking and damage tunnel based on the damage characteristics to establish a cracking and damage tunnel safety evaluation index system. A PL-VIKOR comprehensive evaluation model is constructed based on a crack tunnel safety evaluation index system and an improved subjective and objective combination weighting method based on normal distribution goodness of fit. A multi-dimensional crack lining monitoring system is established to monitor real-time tunnel section data. A pre-trained LSTM model is used to predict the development of the tunnel section based on the real-time tunnel section data, and the future safety state of the tunnel is evaluated based on the development of the tunnel section and the PL-VIKOR comprehensive evaluation model.

[0007] Optionally, in an embodiment of the present application, the crack tunnel safety evaluation index system includes structural mechanics characteristic index, tunnel geological condition index, surrounding rock engineering characteristic index and lining crack characteristic index.

[0008] Optionally, in an embodiment of the present application, the improved subjective and objective combination weighting method based on normal distribution goodness of fit includes: An improved AHP method is used to calculate the subjective weight, and an improved entropy weight method is used to calculate the objective weight. The combination weighting is performed based on the normal distribution goodness of fit of the sample evaluation result sequence.

[0009] Optionally, in an embodiment of the present application, the improved AHP method is used to calculate the subjective weight, and the improved entropy weight method is used to calculate the objective weight, which includes: The AHP method is improved based on the quasi-optimal transfer matrix, and the entropy weight method is improved based on the relative proportion.

[0010] Optionally, in an embodiment of the present application, the combination weighting is performed based on the normal distribution goodness of fit of the sample evaluation result sequence, which includes: The normal distribution goodness of fit under different combination coefficients of each method is calculated, and the final combination weighting result is determined based on the normal distribution goodness of fit.

[0011] Optionally, in an embodiment of the present application, the PL-VIKOR comprehensive evaluation model includes: The group utility value and individual regret value are calculated based on the safety evaluation index data, index weight, positive ideal solution and negative ideal solution. The cosine angle value is calculated based on the safety evaluation index data. The benefit ratio is calculated based on the group utility value, individual regret value and cosine angle value. The safety grade of the monitoring section is determined based on the benefit ratio.

[0012] Optionally, in an embodiment of the present application, the method further includes: Intelligent disposal is performed based on the future safety state of the tunnel.

[0013] The above-mentioned tunnel safety status assessment and prediction method based on lining damage evolution first collects tunnel lining crack damage data based on the lining structure-stratum complex simulation test system, inputs the data into the TF-Transformer deep learning analysis model to output damage characteristics, and summarizes the damage evolution mechanism and development law of the cracked tunnel based on the damage characteristics, and establishes a cracked tunnel safety evaluation index system; then, a PL-VIKOR comprehensive evaluation model is constructed based on the cracked tunnel safety evaluation index system and an improved subjective and objective combined weighting method based on the normal distribution goodness of fit; then, a multi-dimensional cracked lining monitoring system is established to monitor real-time tunnel section data; finally, a pre-trained LSTM model is used based on the real-time tunnel section data to predict the tunnel section development results, and the future safety status of the tunnel is evaluated based on the tunnel section development results combined with the PL-VIKOR comprehensive evaluation model. In other words, by establishing a multi-dimensional cracked lining monitoring system, a cracked tunnel safety evaluation index system, and a comprehensive evaluation method for cracked tunnels, and combining it with a prediction and early warning model calculated by LSTM, based on the monitoring data, the evolution trend of the safety status of lining structures with potential risks can be predicted prospectively. This can timely and accurately grasp the status of the monitored section tunnel and achieve real-time and objective evaluation of the safety status of the cracked tunnel structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 1 is a flow chart of a method for evaluating and predicting tunnel safety status based on lining damage evolution in one embodiment; Figure 2 A schematic diagram of a lining structure-stratum composite simulation test system in one embodiment; Figure 3 Schematic diagram of the structure of the TF-Transformer deep learning analysis model in one embodiment; Figure 4 A schematic diagram of the structure of an LSTM tunnel monitoring data prediction model in one embodiment; Figure 5 The figure is a flowchart of the specific steps of a method for evaluating and predicting the safety status of a tunnel based on lining damage evolution in one embodiment. DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0016] In one embodiment, Figure 1 As shown in FIG, a tunnel safety status assessment and prediction method based on lining damage evolution is provided, which includes the following steps: S101: Based on the lining structure-stratum complex simulation test system, tunnel lining crack damage data is collected and input into the TF-Transformer deep learning analysis model to output damage characteristics. Based on the damage characteristics, the damage evolution mechanism and development law of the cracked tunnel are summarized, and a safety evaluation index system for cracked tunnels is established.

[0017] In the embodiment of the present application, first, based on Figure 2 The lining structure-stratum complex simulation test system shown in the figure collects tunnel lining crack damage data. The monitoring instruments used include earth pressure cells, strain gauges, displacement meters, crack distance meters, acoustic emission test devices and piezoelectric ceramic equipment. The tunnel bedding angle is simulated by polytetrafluoroethylene film, and the damage condition is comprehensively measured by the acoustic emission test device and piezoelectric ceramic equipment.

[0018] Afterwards, the data was input into the TF-Transformer deep learning analysis model to analyze the damage characteristics including the damage location and damage degree, and the damage evolution mechanism and development law of the cracked tunnel were summarized based on the damage location and damage degree. Figure 3 As shown in Figure 1, the structure of the TF-Transformer deep learning analysis model includes a feature segmenter module and a Transformer module. The feature segmenter can convert features into embeddings, which are then processed by the Transformer module. The Transformer module has a process including normalization, multi-head self-attention mechanism and feedforward. Among them, the feature segmenter module can convert the input feature x into an embedding T∈Rk×d. Given a feature The embedding of is calculated as follows:

[0019] in, is the jth eigenvalue, is the learnable bias vector corresponding to the j-th feature, It is a conversion function related to feature type.

[0020] In summary:

[0021]

[0022]

[0023] in, is a numerical eigenvalue, is a learnable feature scaling vector, is a learnable feature bias vector, It is the embedding result of numerical features; is the embedding result of categorical features, is a learnable feature bias vector, is an embedding lookup operation, where is a one-hot vector for the corresponding class, is a learnable embedding matrix; is a token matrix, is the total number of features, stack[⋅] denotes a vertical stack, and is a matrix that concatenates all feature embeddings in order.

[0024] The TF-Transformer deep learning analysis model includes two steps of pre-training and application in specific application. After pre-training, the model only needs to input the bedding angle, strain data, crack width depth, and initial crack type to judge the damage condition of the tunnel lining, so as to better summarize the damage evolution mechanism and development law of the cracked tunnel, and thus establish a safety evaluation index system of the cracked tunnel.

[0025] In an embodiment of the present application, the safety evaluation index system of the cracked tunnel includes structural mechanics characteristic indexes, tunnel geological condition indexes, surrounding rock engineering characteristic indexes, and lining crack characteristic indexes.

[0026] S103: Construct a PL-VIKOR comprehensive evaluation model based on the safety evaluation index system of the cracked tunnel and the improved subjective and objective combination weighting method based on the goodness of normal distribution fitting.

[0027] In the embodiments of the present application, the improved subjective and objective combination weighting method based on the goodness of normal distribution fitting is used to determine the weight of each safety evaluation index, and a PL-VIKOR comprehensive evaluation model is constructed.

[0028] Specifically, in an embodiment of the present application, the improved subjective and objective combination weighting method based on the goodness of normal distribution fitting includes: S201: Calculate the subjective weight by using the improved analytic hierarchy process (AHP), and calculate the objective weight by using the improved entropy weight method (EWM).

[0029] S203: Perform combination weighting based on the method of goodness of normal distribution fitting of the sample evaluation result sequence.

[0030] In the embodiments of the present application, the improved analytic hierarchy process (AHP) and the improved entropy weight method (EWM) are used to calculate the subjective and objective weights, respectively, and the method of goodness of normal distribution fitting of the sample evaluation result sequence is used to perform combination weighting based on the subjective and objective weights. Optionally, the combination weighting method includes the multiplication addition method, the linear weighting method, and the game theory combination weighting method.

[0031] Specifically, in an embodiment of the present application, the improved analytic hierarchy process (AHP) and the improved entropy weight method (EWM) are used to calculate the subjective and objective weights, respectively, and the method of goodness of normal distribution fitting of the sample evaluation result sequence is used to perform combination weighting based on the subjective and objective weights. Optionally, the combination weighting method includes the multiplication addition method, the linear weighting method, and the game theory combination weighting method. The analytic hierarchy process is improved based on the quasi-optimal transfer matrix, and the entropy weight method is improved based on the relative weight.

[0032] In one embodiment of the present application, for the calculation of subjective weights, first, a judgment matrix is ​​constructed: , the judgment matrix is ​​the judgment matrix obtained by each expert comparing the evaluation indicators pairwise according to the 1-9 scaling method; is the number of evaluation indicators Indicates the The indicator is relative to the The importance of each indicator. Then, the antisymmetric matrix is ​​constructed based on the reciprocal matrix A. ,in Afterwards, the optimal transfer matrix is ​​constructed based on the antisymmetric matrix B ,in Finally, the quasi-optimal transfer matrix of matrix A is calculated based on the optimal transfer matrix C. ,in The matrix can obtain the weight value at one time without the need for consistency check, which effectively avoids the blindness of adjusting the judgment matrix.

[0033] For the calculation of objective weights, first, the data is standardized. In order to eliminate the impact of the differences in the dimensions of different indicators, the positive indicators (the larger the better) and negative indicators (the smaller the better) in the data are standardized. is the value of the i-th evaluation indicator of the j-th target, The value after standardization is calculated as follows: Positive indicators:

[0034] Negative indicators:

[0035] in, is the original index value, indicating that the jth index is in the The value of the samples; is the minimum value of the indicator; is the maximum value of the indicator.

[0036] Afterwards, the entropy value of each evaluation index is calculated The proportion of indicators in all indicators , which is calculated as follows: ,

[0037] in, is the sample proportion, indicating that the i-th sample is in the ; The index entropy value is used to reflect the discrete degree of the index data.

[0038] Then, the calculation formula of the entropy weight of each evaluation index is as follows:

[0039] wherein, is the average value of all entropy values other than 1; N is a precision coefficient, and here N = 41.27.

[0040] In an embodiment of the present application, the method for fitting goodness of normal distribution based on the sample evaluation result sequence is combined weighting, which comprises: calculating the normal distribution fitting goodness under different combination coefficients of each method, and determining the final combined weighting result based on the normal distribution fitting goodness.

[0041] In an embodiment of the present application, the normal distribution fitting goodness under different combination coefficients of each method is calculated by enumeration method under the condition of combination coefficient change (change interval is 0.001), the fitting goodness of different weighting methods is compared and selected by fitting goodness calculation, and the final combined weighting result is calculated. Specifically, the weighting value when the fitting goodness of the multiplication addition method, the linear weighting method and the game theory combined weighting method is the largest is selected. Then, the three values are compared, the maximum value of the three is selected, and the final combined weighting result is determined. The normal distribution fitting goodness calculation method is as follows: first, n samples are arranged in ascending order, and are obtained.

[0042] wherein is the sample mean, S is the sample standard deviation,

[0043] Then, the sample empirical distribution function is calculated, and the formula is as follows:

[0044] Then, the statistical quantity is constructed according to the maximum value of the deviation between the sample empirical distribution function and the standard normal distribution function . The statistical quantity obeys the Kolmogorov distribution, and the expression of the distribution function is as follows:

[0045] Finally, the expression value of the fitting goodness is calculated, which represents With the probability of a result more extreme (i.e. less likely to conclude that the two distribution functions are similar) than the sample observations obtained under the approximation.

[0046] In an embodiment of the present application, the PL-VIKOR comprehensive evaluation model comprises: S301: calculating group utility value and individual regret value based on safety evaluation index data, index weight, positive ideal solution and negative ideal solution.

[0047] S303: calculating cosine angle value based on safety evaluation index data.

[0048] S305: calculating benefit ratio based on the group utility value, individual regret value and cosine angle value.

[0049] S307: determining the safety level of the monitoring section based on the benefit ratio.

[0050] In an embodiment of the present application, by using the PL-VIKOR comprehensive evaluation method, a scientific and quantitative safety evaluation index system is established based on structural mechanical characteristics, tunnel geological conditions, surrounding rock engineering properties and lining crack damage characteristics, and a constructed sample and a measured sample are established to realize real-time and objective evaluation of the safety state of the crack tunnel structure. Specifically, the measured sample is obtained from the monitoring data, and the constructed sample is established according to the index system and its benchmark. The benchmark is divided into equal parts, and the divided values are taken as input values, i.e. the constructed sample. For example, the benchmark of index 1 is [0-0.25], [0.25-0.5], [0.5-0.75] and [0.75-1.0], and the benchmark of index 2 is [0-0.3], [0.3-0.5], [0.5-0.8] and [0.8-1.0]. If 9 constructed samples are divided, the values of the 9 constructed samples are (0, 0), (0.125, 0.15), (0.25, 0.25) and so on. The calculation formula is as follows: first, the positive ideal solution and the negative ideal solution are determined.

[0051] With the positive ideal solution as a reference:

[0052] With the negative ideal solution as a reference:

[0053] wherein, , is the group utility value and the individual regret value of the ith sample with the positive ideal solution as a reference; , is the group utility value and the individual regret value of the ith sample with the negative ideal solution as a reference; is the positive ideal solution of the jth evaluation index, is the negative ideal solution of the jth evaluation index, denotes the normalized index data.

[0054] Based on the safety evaluation index data, index weight, positive ideal solution, and negative ideal solution, the group utility value and individual regret value are calculated: Based on Aggregation function:

[0055] Take p = 1, and calculate the evaluation section Group utility under n indexes:

[0056] wherein, is the group utility value; is the index weight, which is obtained by AHP+EWM combined weighting; is the normalized gap, which represents the relative distance between the current value and the optimal value.

[0057] Take p = ∞, and calculate the evaluation section Individual regret under n indexes:

[0058] wherein, is the individual regret value.

[0059] Then, the cosine angle value is calculated based on the safety evaluation index data:

[0060] wherein, is the index value of the ith sample, is the positive ideal solution vector.

[0061] Finally, the benefit ratio is calculated based on the group utility value, individual regret value, and cosine angle value, and the safety level of the monitoring section is determined based on the benefit ratio.

[0062]

[0063] wherein, denotes the group benefit weight, denotes the individual regret weight, is the fluctuation similarity weight, which can be adjusted according to the different emphasis of the organizer on each sub-item. The Q value is used to sort the constructed sample and the measured sample, thereby determining the safety level of each monitoring section.

[0064] S105: Establish a multi-dimensional crack lining monitoring system to monitor real-time tunnel section data.

[0065] In the embodiments of the present application, a multi-dimensional crack and damage lining monitoring system is established, surface strain gauges and surface crack meters are installed at the crack and damage tunnel parts, a comprehensive monitoring network covering the apparent form and internal state of the lining structure is constructed, and dynamic and comprehensive perception of the crack and damage development process is realized to provide real-time and reliable data support for structure safety evaluation and disposal decision.

[0066] In S107, a pre-trained LSTM model is used to predict the tunnel section development result based on the real-time tunnel section data, and a PL-VIKOR comprehensive evaluation model is used to evaluate the future safety state of the tunnel based on the tunnel section development result.

[0067] In the embodiments of the present application, a pre-trained LSTM model is used to predict the tunnel section development result based on the real-time tunnel section data, and a PL-VIKOR comprehensive evaluation model is used to evaluate the future safety state of the tunnel based on the tunnel section development result output by the LSTM model. The evolution trend of the lining structure safety state with potential risks is predicted in advance, so that early warning and early disposal are realized.

[0068] The LSTM model structure has good memory function and can automatically extract deep feature information in time series, and shows good adaptability in prediction tasks involving complex time-varying data such as tunnel engineering. In the long short-term memory network (LSTM), the structure of the repeating module is more complex, which is composed of four neural network layers and interacts in a specific way. As shown in FIG. 1, the calculation of a single neuron in LSTM includes two parts: the update of the neural network state and the calculation of the output value. Three gating mechanisms are included in the LSTM neuron, which are input gate, forget gate and output gate. These gating functions jointly regulate the transmission process of input value, memory value and output value, and the LSTM calculation model is as follows: Figure 4 (1) The forget gate is used to control the amount of information that needs to be forgotten in the neural network state at the current time. The calculation process of the forget gate is as follows:

[0069] In the formula, y f represents the output of the forget gate, h t-1 represents the hidden state at the previous time, σ represents the sigmoid activation function, W f represents the weight matrix of the forget gate, b f represents the bias term of the forget gate, and x t represents the input data at the current time.

[0070] (2) The input gate is composed of two parts, which are the input value of the input gate and the new candidate input value: ​​​​​​​

[0071]

[0072] where the output value of tanh function ranges from -1 to 1, is the candidate memory content at the current time.

[0073] The input gate is used to filter the information of the input layer, and its calculation process is as follows:

[0074] where, represents the updated state of the neural network at the current time, is the state at the previous time.

[0075] (3) The calculation process of the output gate and the hidden state is as follows:

[0076]

[0077] where, is the output gate value, controls the final output quantity, and is the hidden state at the current time, which is also the LSTM output at this time.

[0078] To improve the efficiency of model training and prediction accuracy, the monitored data needs to be preprocessed before establishing the database for training and testing. Preprocessing not only accelerates the training speed of the neural network model, but also effectively prevents the problem of gradient explosion in the training process. The tunnel monitoring data is normalized to make its numerical value distributed in the range of [0, 1]. The normalization of input data uses the min-maxScaler function in the Python deep learning library scikit-learn, and its calculation formula is as follows:

[0079] where x is the original data value, is the minimum value in the original data, is the maximum value in the original data, is the normalized data value.

[0080] During the normalization process, the maximum and minimum values of the training set samples need to be recorded. After the model training is completed, the output results of the model need to be processed by inverse normalization to restore the true predicted values.

[0081] Meanwhile, in order to evaluate the accuracy of the LSTM network model in predicting monitoring data, root mean square error (RmSE) and coefficient of determination (R2) are selected as evaluation indexes of the prediction model.

[0082]

[0083]

[0084] wherein n is the number of samples, is the i-th actual value, is the i-th predicted value, is the root mean square error. is the coefficient of determination.

[0085] In an embodiment of the present application, the method further comprises: based on the future safety state of the tunnel, performing hierarchical intelligent treatment.

[0086] In an embodiment of the present application, according to the future safety state of the tunnel, targeted hierarchical intelligent treatment is performed, and specific measures are shown in Table 1. Table 1

[0087] In the above-mentioned tunnel safety state evaluation and prediction method based on lining damage evolution, first, the tunnel lining crack damage data is collected based on the lining structure-stratum composite simulation test system, which is input into the TF-Transformer deep learning analysis model to output the damage characteristics, and the damage evolution mechanism and development law of the cracked tunnel are summarized based on the damage characteristics, and a cracked tunnel safety evaluation index system is established; then, based on the cracked tunnel safety evaluation index system and the improved subjective and objective combination weighting method based on the goodness of fit of normal distribution, a PL-VIKOR comprehensive evaluation model is constructed; then, a multi-dimensional cracked lining monitoring system is established to monitor real-time tunnel section data; finally, based on the real-time tunnel section data, a pre-trained LSTM model is used to predict the tunnel section development result, and based on the tunnel section development result, the future safety state of the tunnel is evaluated in combination with the PL-VIKOR comprehensive evaluation model. That is, by establishing a multi-dimensional cracked lining monitoring system, a cracked tunnel safety evaluation index system, and a cracked tunnel comprehensive evaluation method, and combining a prediction and early warning model calculated by an LSTM, the evolution trend of the safety state of the lining structure with potential risks can be predicted in advance according to the monitoring data, the state of the monitoring section tunnel can be accurately grasped in time, and real-time and objective evaluation of the safety state of the cracked tunnel structure can be realized.

[0088] The following illustrates the specific implementation steps of the tunnel safety state evaluation and prediction method based on lining damage evolution of the present application with a specific embodiment. As shown in Figure 5 The tunnel crack lining damage evolution test system is composed of an analysis model based on TF-Transformer deep learning and a lining structure-stratum composite simulation test system. Through the damage evolution test system, the damage evolution mechanism and development law of the cracked tunnel are summarized and induced to provide theoretical judgment for the establishment of the cracked tunnel safety evaluation index system. For example, theoretically, 300 kN will crack, so 0-300 kN is set as the first level of structure safety, and 900 kN will completely destroy, so 900-1200 kN is set as the fourth level of structure insecurity. A multi-dimensional cracked tunnel lining monitoring system is established, which can provide data support for the establishment of the cracked tunnel safety evaluation index system through the monitoring of the site lining. For example, if the maximum stress monitored is 1000 kN, the upper limit of the evaluation index benchmark can be set to 1200 kN (including some redundancy). The improved AHP-EWM combined weighting method based on the goodness of normal distribution fitting can assign appropriate weights to the cracked tunnel safety evaluation index system, so as to evaluate the safety state of the cracked tunnel in time through the PL-VIKOR comprehensive evaluation method. The monitoring data provided by the multi-dimensional cracked tunnel lining monitoring system are preprocessed based on the min-maxScaler function in the Python deep learning library scikit-learn, and the development results of the monitoring data are predicted through the establishment of an LSTM model. The predicted data are provided to the PL-VIKOR comprehensive evaluation model to realize the prediction of the future safety state of the cracked tunnel, and targeted disposal measures are proposed according to the safety state prediction.

[0089] It should be understood that although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other sequences. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or steps or stages in other steps.

[0090] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties.

[0091] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0092] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0093] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A tunnel safety status assessment and prediction method based on lining damage evolution, characterized by: The method comprises: Using a lining structure-stratum complex simulation test system, tunnel lining crack damage data was collected and input into the TF-Transformer deep learning analysis model to output damage characteristics. Based on these damage characteristics, the damage evolution mechanism and development law of cracked tunnels were summarized, and a safety evaluation index system for cracked tunnels was established. A PL-VIKOR comprehensive evaluation model was constructed based on the safety evaluation index system of cracked tunnels and an improved subjective and objective combined weighting method based on the goodness of fit of the normal distribution. Establish a multi-dimensional cracked lining monitoring system to monitor real-time tunnel cross-section data; Based on the real-time tunnel section data, a pre-trained LSTM model is used to predict the tunnel section development results, and based on the tunnel section development results combined with the PL-VIKOR comprehensive evaluation model, the future safety status of the tunnel is evaluated.

2. The tunnel safety status assessment and prediction method based on lining damage evolution according to claim 1 is characterized in that: The crack damage tunnel safety evaluation index system includes structural mechanics characteristic index, tunnel geological condition index, surrounding rock engineering characteristic index and lining crack damage characteristic index.

3. The tunnel safety status assessment and prediction method based on lining damage evolution according to claim 1 is characterized in that: The improved subjective and objective combined weighting method based on the normal distribution goodness of fit includes: The improved analytic hierarchy process is used to calculate the subjective weight, and the improved entropy weight method is used to calculate the objective weight; The combination weighting is performed based on the normal distribution fitting goodness of the sample evaluation result sequence.

4. The tunnel safety status assessment and prediction method based on lining damage evolution according to claim 3 is characterized in that: The method of calculating the subjective weight by using the improved analytic hierarchy process and the method of calculating the objective weight by using the improved entropy weight method includes: The analytic hierarchy process is improved based on the quasi-optimal transfer matrix, and the entropy weight method is improved based on the relative weight.

5. The tunnel safety status assessment and prediction method based on lining damage evolution according to claim 3 is characterized in that: The method for performing combination weighting based on the normal distribution goodness of fit of the sample evaluation result sequence includes: The normal distribution fitting goodness of fit under different combination coefficients of each method is calculated, and the final combination weighting result is determined based on the normal distribution fitting goodness of fit.

6. The tunnel safety status assessment and prediction method based on lining damage evolution according to claim 1 is characterized in that: The PL-VIKOR comprehensive evaluation model includes: Calculate group utility value and individual regret value based on safety evaluation index data, index weight, positive ideal solution and negative ideal solution; Calculate the cosine angle value based on the safety evaluation index data; Calculating a benefit ratio based on the group utility value, the individual regret value, and the cosine angle value; The safety level of the monitoring section is determined based on the benefit ratio.

7. The tunnel safety status assessment and prediction method based on lining damage evolution according to claim 1 is characterized in that: The method further comprises: Perform hierarchical intelligent disposal based on the future safety status of the tunnel.

Images