Tunnel support structure safety monitoring method and system based on rock mass-support feedback

Abstract

The application provides a tunnel supporting structure safety monitoring method and system based on rock mass-support feedback, and relates to the technical field of geotechnical engineering tunnel construction safety monitoring. The method comprises the following steps: collecting supporting reaction force data through a pressure sensor and collecting deformation data through a displacement sensor to generate multivariate monitoring data; extracting reaction force change characteristics and deformation change characteristics through an energy analysis algorithm and calculating energy accumulation values and energy release values of surrounding rock; analyzing the dynamic changes of the energy accumulation values and the energy release values through a stage identification algorithm to generate stage identification results; and converting the stage identification results into surrounding rock stability state information through a state generation algorithm, so that real-time perception of the surrounding rock state, dynamic evaluation of energy and intelligent identification of deformation stages are achieved, the defects of a passive and static traditional supporting system are overcome, and the safety and adaptability of tunnel supporting are significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of geotechnical engineering tunnel construction safety monitoring, and in particular to a tunnel support structure safety monitoring method and system based on rock mass-support feedback. BACKGROUND

[0002] There are more and more large-span tunnel projects in the fields of transportation, water and electricity, etc. Traditional support systems, such as ordinary mortar anchor rods, steel arches, and sprayed concrete, are mostly passive and static in design. Their support characteristics are fixed after installation, which has significant drawbacks. First, the traditional support cannot perceive the changes in the surrounding rock state, leading to over-supporting or under-supporting, making it difficult to achieve the right support state. Second, the essence of hard rock disaster is energy instability. The traditional support system lacks an effective energy dissipation mechanism and cannot actively absorb and convert the huge energy released by the surrounding rock, but only passively withstands it. Finally, the support system and the surrounding rock are one-way, and there is no closed-loop "perception-analysis-response" intelligent feedback loop, which cannot dynamically adjust according to different stages of surrounding rock deformation.

[0003] Therefore, it is an urgent need and a frontier direction to develop a stress self-adaptive intelligent support system that can cooperate with the deformation of the surrounding rock, intelligently adjust its own characteristics, and actively dissipate energy to solve the stability control problem of hard rock large-span tunnels. SUMMARY

[0004] To solve the above problems in the prior art, the present application provides a tunnel support structure safety monitoring method based on rock mass-support feedback, which comprises:

[0005] S1: Based on the sensor network arranged on the support structure, the support reaction force data is collected by the pressure sensor, and the deformation data is collected by the displacement sensor, and the support reaction force data and the deformation data are subjected to data fusion processing to generate multi-element monitoring data;

[0006] S2: Based on the multi-element monitoring data, the reaction force change characteristics and the deformation change characteristics are extracted by the energy analysis algorithm, and based on the reaction force change characteristics and the deformation change characteristics, the energy accumulation value and the energy release value of the surrounding rock are calculated, and the energy state parameters are generated by combination;

[0007] S3: Based on the energy state parameters, the dynamic changes of the energy accumulation value and the energy release value are analyzed by the stage recognition algorithm, and the current deformation stage of the surrounding rock is recognized based on the dynamic changes, and a stage recognition result is generated;

[0008] S4: The stage recognition result is converted into a surrounding rock stability state description by a state generation algorithm to generate surrounding rock stability state information.

[0009] Compared with the prior art, the present application has the following advantages:

[0010] The tunnel supporting structure safety monitoring method based on rock mass-support feedback solves the problems of the traditional supporting system in the background technology through the synergistic effect of steps S1 to S4. In S1, based on the sensor network arranged on the supporting structure, the supporting reaction force data is collected through the pressure sensor, and the deformation data is collected through the displacement sensor, and the supporting reaction force data and the deformation data are subjected to data fusion processing to generate multi-element monitoring data. This step realizes real-time and multi-dimensional perception of the surrounding rock state, overcomes the disadvantage that the traditional supporting system cannot perceive the change of the surrounding rock state, and provides a reliable data basis for subsequent analysis. In S2, based on the multi-element monitoring data, the reaction force change characteristics and the deformation change characteristics are extracted through the energy analysis algorithm, and based on the reaction force change characteristics and the deformation change characteristics, the energy accumulation value and the energy release value of the surrounding rock are calculated, and the energy state parameter is generated by combination. This step converts the physical data into energy indicators, quantifies the energy dynamics of the surrounding rock, solves the problem of insufficient energy control capability of the traditional supporting system, and provides a key basis for energy management. In S3, based on the energy state parameter, the dynamic changes of the energy accumulation value and the energy release value are analyzed through the stage recognition algorithm, and the current deformation stage of the surrounding rock is recognized based on the dynamic changes to generate a stage recognition result. This step realizes intelligent analysis of the deformation behavior of the surrounding rock, can accurately judge the stability state of the surrounding rock, and lays a foundation for dynamically adjusting the supporting strategy. In S4, the stage recognition result is converted into a surrounding rock stability state description through a state generation algorithm to generate surrounding rock stability state information. This step converts the analysis result into an operable state output to form a complete "perception-analysis-response" closed-loop feedback.

[0011] The entire method realizes a coherent process from data collection, energy analysis, stage recognition to state generation through the close cooperation of the four steps, ensures that the supporting system can respond to the change of the surrounding rock in real time and accurately, thereby optimizing the supporting effect, avoiding over-supporting or under-supporting, and improving the energy dissipation capability, and finally enhancing the stability and safety of the tunnel project. BRIEF DESCRIPTION OF DRAWINGS

[0012] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings needed to be used in the specific embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0013] Figure 1 The figure shows the flowchart of the tunnel supporting structure safety monitoring method based on rock mass-support feedback provided by an embodiment of the present application.

[0014] Figure 2 Fig. 1 shows a structural schematic diagram of a tunnel support structure safety monitoring system based on rock mass-support feedback according to an embodiment of the present application. DETAILED DESCRIPTION

[0015] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be clearly and completely described below. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0016] The specific embodiments of the present application will be described below.

[0017] Embodiment 1

[0018] As shown in the drawings, the present application provides a tunnel support structure safety monitoring method based on rock mass-support feedback, comprising: Figure 1 S1: based on the sensor network arranged on the support structure, collecting support reaction force data through the pressure sensor and collecting deformation data through the displacement sensor, and performing data fusion processing on the support reaction force data and the deformation data to generate multi-element monitoring data;

[0019] S2: based on the multi-element monitoring data, extracting the reaction force change characteristics and the deformation change characteristics through the energy analysis algorithm, and based on the reaction force change characteristics and the deformation change characteristics, calculating the energy accumulation value and the energy release value of the surrounding rock, and combining to generate the energy state parameter;

[0020] S3: based on the energy state parameter, analyzing the dynamic changes of the energy accumulation value and the energy release value through the stage recognition algorithm, and based on the dynamic changes to identify the current deformation stage of the surrounding rock, and generating the stage recognition result;

[0021] S4: converting the stage recognition result into the surrounding rock stability state description through the state generation algorithm, and generating the surrounding rock stability state information.

[0022]

[0023] ​In the implementation of the tunnel support structure safety monitoring method based on rock mass-support feedback, first, data collection is performed through the sensor network arranged on the support structure. The pressure sensor is used to monitor the support reaction force data in real time, and the displacement sensor is used to capture the deformation data. These data are integrated into multi-element monitoring data through data fusion processing, ensuring the consistency of information in time and space. Data fusion processing aligns and matches the raw data from different sensors, eliminating errors caused by differences in collection time or location, thereby forming a complete and reliable monitoring foundation. Multi-element monitoring data not only contains the original values of the reaction force and deformation, but also enhances the multidimensionality and usability of the data through fusion algorithms, providing solid input for subsequent analysis. This step relies on efficient sensor deployment strategies and fusion algorithms, such as timestamp alignment and spatial interpolation methods, to ensure the accuracy and real-time performance of data in dynamic environments. The technical effect lies in improving the comprehensiveness and reliability of monitoring. Traditional methods often rely on a single data type, which can easily overlook the correlation between reaction force and deformation. However, this method can more comprehensively reflect the surrounding rock state through data fusion processing, reducing the risk of misjudgment. Data fusion processing integrates multi-source information, avoiding one-sided analysis caused by isolated data, thereby providing high-quality input for energy analysis algorithms and indirectly enhancing the response accuracy and stability of the entire monitoring system. Through this step, the system can early identify abnormal changes in surrounding rock, laying the foundation for subsequent energy calculation and stage identification, and ultimately achieving dynamic optimization of support strategies.

[0024] The synergy of the entire process is that the data collection and fusion of S1 step provides a unified data source for the energy analysis of S2, ensuring the coherent execution of subsequent algorithms, and the generation of multi-element monitoring data makes it possible to visualize and analyze the surrounding rock state, improving the efficiency of engineering management. This method effectively overcomes the shortcomings of insufficient perception in traditional support through real-time, multi-dimensional data collection and processing, enabling the support system to more accurately adapt to changes in surrounding rock, thereby improving the safety and durability of tunnel engineering. In summary, this method achieves comprehensive monitoring through sensor networks and data fusion processing, enhances the accuracy and consistency of data, provides a reliable foundation for energy analysis, improves system response speed, and optimizes support decisions through multi-source information integration, ultimately achieving dynamic management of the surrounding rock state.

[0025] In some implementations, the energy analysis algorithm includes a data preprocessing sub-algorithm and a feature extraction sub-algorithm, and S2 includes:

[0026] S2.1: Based on the multi-element monitoring data, through the data preprocessing sub-algorithm, filter processing is performed on the support reaction force data, and denoising processing is performed on the deformation data, generating clean monitoring data;

[0027] S2.2: Based on the cleaning monitoring data, the feature extraction sub-algorithm extracts the support reaction force change characteristic value from the support reaction force data and the deformation change characteristic value from the deformation data, and generates a feature data set;

[0028] S2.3: Based on the feature data set, the energy calculation sub-algorithm calculates the energy accumulation value using the reaction force change characteristics and the energy release value using the deformation change characteristics, and combines the energy accumulation value and the energy release value to generate an energy state parameter.

[0029] In the implementation of the energy analysis algorithm, the data preprocessing sub-algorithm first processes the multivariate monitoring data, performs filtering processing on the support reaction force data to eliminate high-frequency interference, and at the same time implements denoising processing on the deformation data to remove fluctuations caused by environmental factors, thereby generating clean monitoring data. By removing noise and outliers, the quality and usability of the data are improved, making the subsequent feature extraction more accurate. The feature extraction sub-algorithm extracts the reaction force change characteristic value from the support reaction force data based on the clean monitoring data, such as calculating the mean, variance or trend index to reflect the dynamic change of the reaction force; Similarly, the deformation change characteristic value is extracted from the deformation data, such as displacement rate or acceleration, to capture the core features of the deformation behavior. These characteristic values combine to generate a feature data set, which provides input for the energy calculation sub-algorithm. The energy calculation sub-algorithm calculates the energy accumulation value using the reaction force change characteristics, representing the potential energy stored in the surrounding rock, and calculates the energy release value using the deformation change characteristics, representing the energy dissipated by the surrounding rock, and finally combines to generate the energy state parameter. The technical effect lies in that through data preprocessing and feature extraction, the algorithm can extract key indicators from raw data, reducing the impact of noise on analysis, thereby improving the accuracy and reliability of the energy state parameter. The data preprocessing sub-algorithm eliminates random errors in the data through filtering and denoising, enabling the feature extraction sub-algorithm to more effectively identify the essential changes in reaction force and deformation, and the energy calculation sub-algorithm accurately quantifies the energy dynamics based on clean features, avoiding energy miscalculations caused by poor data quality in traditional methods.

[0030] The synergy of this process is reflected in the fact that data preprocessing ensures the purity of input data, feature extraction converts data into analyzable features, and energy calculation maps these features into energy indicators, forming a complete chain from data cleaning to parameter generation. Through the step-by-step execution of the energy analysis algorithm, the system can monitor the energy state of the surrounding rock in real time, providing a scientific basis for stage identification, thereby supporting the adaptive adjustment of the support system. In summary, this algorithm improves data quality through preprocessing and feature extraction, accurately calculates energy accumulation and release values, enhances the reliability of state parameters, realizes real-time monitoring of energy dynamics, optimizes the analysis process through step-by-step processing, and ultimately improves the response accuracy and safety of tunnel support.

[0031] In some implementations, S2.1 comprises:

[0032] S2.1.1: Based on the multi-element monitoring data, high-frequency noise removal is performed on the support reaction force data to obtain first filtered data;

[0033] S2.1.2: Based on the first filtered data, low-frequency drift correction is performed on the deformation data to obtain second filtered data;

[0034] S2.1.3: Based on the second filtered data, abnormal points in the support reaction force data and the deformation data are identified and removed to generate clean monitoring data.

[0035] In the implementation of S2.1, based on the multi-element monitoring data, first, high-frequency noise removal is performed on the support reaction force data to eliminate high-frequency interference signals generated during sensor acquisition through digital filtering technology, obtaining first filtered data. This process ensures the smoothness and stability of the reaction force data. High-frequency noise removal uses a low-pass filter or similar algorithm to effectively filter out short-term fluctuations and retain the long-term trend of the reaction force data, thereby reducing the risk of false positives. Next, based on the first filtered data, low-frequency drift correction is performed on the deformation data to compensate for baseline drift that may be caused by temperature or mechanical factors in the displacement sensor, obtaining second filtered data. Low-frequency drift correction adjusts the deformation data by referencing a reference value or historical data to ensure that it accurately reflects the behavior of the surrounding rock. Then, based on the second filtered data, abnormal points in the support reaction force data and the deformation data are identified and removed, such as using statistical methods like Z-score or cluster analysis to detect values that deviate from the normal range and are removed, finally generating clean monitoring data. The technical effect is that, through high-frequency noise removal and low-frequency drift correction, interference factors in the data are effectively controlled, and abnormal point removal further improves the purity of the data, making the clean monitoring data more closely reflect the actual surrounding rock state. High-frequency noise removal prevents short-term interference from masking the true reaction force changes, low-frequency drift correction avoids the influence of system errors on deformation data, and abnormal point removal eliminates deviations caused by occasional factors, ensuring the accuracy and consistency of the data in subsequent analysis.

[0036] The synergistic effect of this step is that noise removal and drift correction handle errors in different frequency bands, and abnormal point removal serves as the final quality control. The combination of the three forms a complete data cleaning process, providing reliable input for feature extraction and energy calculation. Through this implementation, clean monitoring data can more accurately reflect the dynamic behavior of the surrounding rock, reducing false positives caused by data quality issues, thereby improving the robustness and credibility of the entire monitoring system. In summary, this step removes data noise and drift through multi-level filtering and correction, identifies and removes outliers, generates high-quality clean data, enhances the accuracy of monitoring, and optimizes the data flow through systematic processing, ultimately supporting accurate assessment of the energy state of the surrounding rock.

[0037] In some implementations, S2.2 comprises:

[0038] S2.2.1: Based on the clean monitoring data, calculate the statistical features of the support reaction force data through the reaction force feature extraction sub-algorithm, and obtain the reaction force change feature values;

[0039] S2.2.2: Based on the clean monitoring data, calculate the statistical features of the deformation data through the deformation feature extraction sub-algorithm, and obtain the deformation change feature values;

[0040] S2.2.3: Merge the reaction force change feature values and the deformation change feature values into a unified dataset through the feature combination sub-algorithm, and generate the feature dataset.

[0041] In the implementation of S2.2, based on the clean monitoring data, the reaction force feature extraction sub-algorithm calculates the statistical features of the support reaction force data, such as mean, standard deviation or peak value index, to obtain the reaction force change feature values. These feature values quantify the fluctuations and trends of the reaction force in the time series, thus reflecting the pressure changes of the surrounding rock on the support structure. The reaction force feature extraction sub-algorithm uses sliding window or time series analysis method to update the feature values in real time, ensuring that it can capture the dynamic evolution of the reaction force. At the same time, based on the clean monitoring data, the deformation feature extraction sub-algorithm calculates the statistical features of the deformation data, such as the cumulative value, rate of change or distribution pattern of displacement, to obtain the deformation change feature values. These values describe the size and speed of the surrounding rock deformation, providing key inputs for energy analysis. The deformation feature extraction sub-algorithm uses similar methods to ensure that the feature values accurately represent the core patterns of deformation behavior. Then, through the feature combination sub-algorithm, the reaction force change feature values and the deformation change feature values are merged into a unified dataset to generate the feature dataset. The feature combination sub-algorithm uses data standardization or dimension reduction techniques, such as principal component analysis, to integrate different dimension features into a consistent format, facilitating subsequent processing. The technical effect lies in that the reaction force feature extraction and deformation feature extraction capture key indicators of surrounding rock state from different dimensions, and the feature combination integrates these indicators into a comprehensive dataset, improving the comprehensiveness and efficiency of analysis. The reaction force change feature values highlight the dynamic characteristics of the reaction force through statistical calculation, while the deformation change feature values focus on the evolution law of deformation. The feature combination sub-algorithm eliminates the redundancy and inconsistency between features through unified processing, making the feature dataset more suitable for the energy calculation sub-algorithm.

[0042] The synergy of this step is reflected in the parallel execution of the reaction force feature extraction and the deformation feature extraction, which respectively optimize the analysis of their respective data domains, while the feature combination serves as a bridge to integrate the two into a coherent whole, ensuring that the subsequent calculation of the energy state parameters is based on multi-dimensional features. Through this implementation, the feature data set can fully reflect the relationship between the reaction force and the deformation of the surrounding rock, laying the foundation for accurate assessment of the energy state, thereby enhancing the intelligent response capability of the support system. In summary, this step generates detailed feature values by extracting statistical features of the reaction force and deformation, and after combination, forms a unified data set, improving the multi-dimensionality and usability of the data, optimizing the input of energy analysis, and enhancing the system coordination through feature integration, ultimately achieving accurate monitoring of the surrounding rock state.

[0043] In some implementations, the stage recognition algorithm includes a trend analysis sub-algorithm, a change rate calculation sub-algorithm, and an amplitude recognition sub-algorithm, and S3 includes:

[0044] S3.1: Based on the energy state parameters, through the trend analysis sub-algorithm, analyze the trend of the energy accumulation value and the energy release value, and generate a first deformation stage identifier;

[0045] S3.2: Based on the energy state parameters, through the change rate calculation sub-algorithm, calculate the change rate of the energy accumulation value and the energy release value, and generate a second deformation stage identifier;

[0046] S3.3: Based on the energy state parameters, through the amplitude recognition sub-algorithm, determine the amplitude of the energy accumulation value and the energy release value, and generate a third deformation stage identifier;

[0047] S3.4: Integrate the first deformation stage identifier, the second deformation stage identifier, and the third deformation stage identifier through the comprehensive sub-algorithm to generate a stage recognition result.

[0048] In the implementation of the stage recognition algorithm, the trend analysis sub-algorithm analyzes the long-term trends of the energy accumulation value and the energy release value based on the energy state parameters, determines the change direction of these parameters within a preset time window (e.g., a 10-minute window) through calculating the moving average or linear regression analysis, and generates the first deformation stage identification (such as "stable", "slow accumulation" or "slow release"). The trend analysis sub-algorithm can identify the slow accumulation or release pattern of energy, providing a basis for early warning. The change rate calculation sub-algorithm calculates the change rate of the energy accumulation value and the energy release value based on the energy state parameters, such as obtaining the instantaneous change speed through the first-order differentiation or difference method, and generates the second deformation stage identification (such as "low-speed change", "accelerating accumulation" or "accelerating release"). This identification reflects the acceleration or deceleration characteristics of energy dynamics, helping to judge whether the surrounding rock has entered an active deformation period. The change rate calculation sub-algorithm ensures real-time monitoring of the rate of energy change, avoiding missing rapid evolution events. The amplitude recognition sub-algorithm determines the amplitude of the energy accumulation value and the energy release value based on the energy state parameters, i.e., through calculating the peak value or fluctuation range (such as standard deviation) to evaluate the intensity of energy change, and generates the third deformation stage identification (such as "low amplitude", "medium amplitude" or "high amplitude"). This identification is used to identify the intensity of surrounding rock deformation, distinguishing between minor fluctuations and major events. The amplitude recognition sub-algorithm uses threshold comparison (e.g., the energy accumulation value exceeding 20% of the historical average value is considered high amplitude) or statistical distribution analysis to ensure the objectivity of amplitude evaluation. Finally, the stage recognition result is generated by integrating the first deformation stage identification, the second deformation stage identification and the third deformation stage identification through the comprehensive sub-algorithm. The comprehensive sub-algorithm uses logic rules or machine learning models for fusion: in the logic rule implementation, a preset rule library is used (e.g., if the first deformation stage identification is "stable", the second deformation stage identification is "low-speed change", and the third deformation stage identification is "low amplitude", the stage recognition result is "stable stage"; if the first deformation stage identification is "slow accumulation", the second deformation stage identification is "accelerating accumulation", and the third deformation stage identification is "high amplitude", the stage recognition result is "unstable stage"); in the machine learning model implementation, a classification model (such as a decision tree or a support vector machine) is used, which is trained by historical data, with the feature vector of the first deformation stage identification, the second deformation stage identification and the third deformation stage identification as input, and the deformation stage classification (such as "stable", "gradual change" or "acceleration") as output. The comprehensive sub-algorithm selects logic rules or machine learning models according to engineering requirements to ensure the accuracy and robustness of the stage recognition result. The technical effect lies in that trend analysis, change rate calculation and amplitude recognition analyze energy dynamics from different angles, and the comprehensive sub-algorithm integrates these information to generate a more reliable stage recognition result, improving the accuracy and timeliness of surrounding rock state judgment.The trend analysis sub-algorithm captures the long-term behavior of energy, the rate calculation sub-algorithm focuses on short-term dynamics, and the amplitude identification sub-algorithm evaluates the strength of changes. These three sub-algorithms complement each other, covering multiple dimensions of energy analysis. The comprehensive sub-algorithm avoids the limitations of a single indicator by fusing the results, thereby generating a comprehensive and robust stage recognition result.

[0049] The synergy of this step is reflected in the division of labor among the various sub-algorithms, which respectively handle trend, rate, and amplitude information. The comprehensive sub-algorithm serves as a summary point, ensuring the consistency and operability of the output results, and providing a scientific basis for support adjustment. Through this implementation, the stage recognition algorithm can dynamically track the deformation stage of surrounding rock, support the implementation of adaptive support strategies, and ultimately improve the safety and stability of tunnel engineering. In summary, this algorithm generates multi-dimensional indicators through trend, rate, and amplitude analysis, and after integration, it obtains accurate stage recognition results, enhances the comprehensiveness of state judgment, realizes dynamic monitoring, optimizes decision support through comprehensive processing, and ultimately improves the response efficiency and reliability of the support system.

[0050] In some implementations, S3.1 includes:

[0051] S3.1.1: Through the trend analysis sub-algorithm, calculate the change trend of the energy accumulation value within the preset time window to obtain accumulation trend information; and calculate the change trend of the energy release value within the preset time window to obtain release trend information;

[0052] S3.1.2: When the accumulation trend information and the release trend information meet the preset stability condition, generate a first deformation stage identifier through the identifier generation sub-algorithm.

[0053] In the implementation of the S3.1 step, the trend analysis sub-algorithm calculates the change trend of the energy accumulation value within a preset time window based on the energy state parameter, analyzes the long-term evolution direction of the energy accumulation value through the moving average or linear regression method, and obtains the accumulation trend information, which reflects the sustained change of the energy storage of the surrounding rock. At the same time, the trend analysis sub-algorithm calculates the change trend of the energy release value within a preset time window, evaluates the dynamic behavior of the energy release value using a similar method, and obtains the release trend information, which reveals the evolution mode of the energy dissipation of the surrounding rock. The selection of the preset time window depends on the engineering environment and monitoring requirements, and is usually set according to historical data or experience to ensure the timeliness and accuracy of the trend analysis. When the accumulation trend information and the release trend information meet the preset stability conditions, for example, the energy accumulation value presents a stable or slowly rising trend, and the energy release value remains relatively stable, the identification generation sub-algorithm automatically generates a first deformation stage identification, which is used to indicate that the surrounding rock is in a stable deformation stage. The preset stability conditions are realized through threshold comparison or logical rules, for example, judging whether the absolute value of the trend slope is lower than a certain threshold, to ensure the reliability of the identification generation. The technical effect lies in that the trend analysis sub-algorithm can early identify the stable state of the energy of the surrounding rock through long-term trend calculation, avoiding misjudgment caused by short-term fluctuations, thereby providing accurate stage input for the support system. The trend analysis of the energy accumulation value and the energy release value captures the macro behavior of the energy of the surrounding rock, the accumulation trend information reflects the potential risk of energy storage, and the release trend information shows the efficiency of energy dissipation. When both meet the stability conditions, it indicates that the deformation of the surrounding rock is within a controllable range, and the identification generation sub-algorithm generates a first deformation stage identification accordingly, ensuring the timeliness and accuracy of the stage identification result.

[0054] The synergy of this step lies in that the trend analysis sub-algorithm closely links with the energy state parameter, converting the energy dynamics into quantifiable trend information, while the identification generation sub-algorithm automatically outputs the identification based on the preset conditions, forming a coherent process from trend analysis to stage identification, supporting subsequent state generation. Through this implementation, the system can monitor the long-term change of the energy of the surrounding rock in real time, providing a scientific basis for support adjustment and reducing the risk of over-supporting or under-supporting. In summary, this step generates accumulation and release trend information through trend analysis, automatically generates a first deformation stage identification when the stability conditions are met, improves the accuracy and early warning capability of stage identification, realizes long-term monitoring of the energy dynamics of the surrounding rock, and optimizes the identification generation through preset conditions, ultimately enhancing the response efficiency and stability of the support system.

[0055] In some implementations, S3.2 includes:

[0056] S3.2.1: through the change rate calculation sub-algorithm, calculate the change rate of the energy accumulation value to obtain the accumulation change rate; and calculate the change rate of the energy release value to obtain the release change rate;

[0057] S3.2.2: When the accumulation rate and the release rate meet the preset acceleration condition, a second deformation stage identifier is generated by an identifier generation sub-algorithm.

[0058] In the implementation of step S3.2, the rate calculation sub-algorithm calculates the rate of change of the energy accumulation value based on the energy state parameter, which quantifies the acceleration or deceleration characteristics of the energy storage of the surrounding rock. At the same time, the rate calculation sub-algorithm calculates the rate of change of the energy release value, which reflects the dynamic strength of the energy dissipation of the surrounding rock. The rate calculation sub-algorithm uses real-time data streams to ensure the timeliness of the calculation results, such as updating the rate value through a sliding window to adapt to the rapid evolution of the surrounding rock state. When the accumulation rate and the release rate meet the preset acceleration condition, such as the accumulation rate exceeding a certain threshold and the release rate rising synchronously, the identifier generation sub-algorithm automatically generates a second deformation stage identifier, which is used to indicate that the surrounding rock enters an accelerated deformation or rapid energy release stage. The preset acceleration condition is realized by comparing the rate with historical benchmarks or theoretical models, such as using statistical methods to determine abnormal change points to ensure the accuracy of the identifier generation. The technical effect is that the rate calculation sub-algorithm can quickly capture the short-term dynamics of the energy of the surrounding rock by calculating the rate of change of the energy accumulation value and the energy release value in real time, and timely identify the risk of accelerated deformation, thereby providing key warnings for the support system. The accumulation rate reveals the acceleration trend of energy storage, and the release rate shows the sharp change of energy dissipation. When both meet the acceleration condition, it indicates that the surrounding rock may enter an unstable state, and the identifier generation sub-algorithm generates a second deformation stage identifier accordingly, ensuring the response capability of the stage identification result to rapid evolution events.

[0059] The synergistic effect of this step is that the rate calculation sub-algorithm is directly related to the energy state parameter, which converts energy dynamics into rate indicators, and the identifier generation sub-algorithm automatically outputs identifiers based on preset conditions, forming an efficient process from rate analysis to stage identification, supporting subsequent decision-making. Through this implementation, the system can dynamically track the instantaneous changes of the energy of the surrounding rock, provide data support for emergency response, and prevent the overloading of the support structure. In summary, this step generates accumulation and release rates through rate calculation, automatically generates a second deformation stage identifier when the acceleration condition is met, enhances the sensitivity of stage identification to rapid changes, realizes real-time monitoring of the dynamics of the energy of the surrounding rock, and optimizes the generation of identifiers through preset conditions, ultimately improving the safety and adaptability of the support system.

[0060] In some implementations, S1 includes:

[0061] S1.1: Real-time acquisition of support reaction force data based on pressure sensors arranged on the support structure to obtain a reaction force data set;

[0062] S1.2: Real-time acquisition of deformation data based on displacement sensors arranged on the support structure to obtain a deformation data set;

[0063] S1.3: Time alignment and spatial matching of the reaction force data set and the deformation data set through a data fusion algorithm to generate multi-element monitoring data.

[0064] In the implementation of S1, real-time acquisition of support reaction force data is based on pressure sensors arranged on the support structure, which are installed at key positions of anchor cables or support members and continuously monitor the reaction force value through electrical signals or wireless transmission methods to obtain a reaction force data set containing the size, direction, and timestamp information of the reaction force. At the same time, real-time acquisition of deformation data is based on displacement sensors arranged on the support structure, which are placed on the surface of surrounding rock or the support interface and capture deformation through laser ranging or strain gauge technology to obtain a deformation data set recording the displacement, rate, and spatial distribution of deformation. The arrangement of pressure sensors and displacement sensors needs to consider spatial coverage and representativeness to ensure that the data comprehensively reflects the state of surrounding rock, such as uniformly arranging sensors on the vault and sidewall of the tunnel. Time alignment and spatial matching of the reaction force data set and the deformation data set through a data fusion algorithm ensures the consistency of data on a unified time axis through synchronization timestamps or interpolation methods, and associates data to specific locations through coordinate mapping or gridding technology to generate multi-element monitoring data. The data fusion algorithm uses multi-source information integration methods such as Kalman filtering or data association algorithms to eliminate conflicts and redundancies between sensors and improve the overall quality of data. The technical effect lies in that real-time acquisition of pressure sensors and displacement sensors provides raw data of reaction force and deformation, and the data fusion algorithm integrates multi-source information through time alignment and spatial matching to generate consistent and reliable multi-element monitoring data, laying a foundation for subsequent energy analysis. The reaction force data set directly reflects the stress state of the support structure, and the deformation data set reveals the geometric changes of the surrounding rock. The data fusion algorithm unifies both in time and space dimensions, avoiding inconsistencies in data caused by acquisition differences, thereby generating high-quality multi-element monitoring data to ensure the accuracy and continuity of subsequent analysis.

[0065] The synergy of this step is reflected in the fact that the pressure sensor and the displacement sensor collect data in parallel, covering different aspects of the surrounding rock state, and the data fusion algorithm serves as a bridge to integrate heterogeneous data into a unified format, forming a complete chain from collection to fusion, supporting the calculation of subsequent energy state parameters. Through this implementation, the system can comprehensively perceive the surrounding rock dynamics, reduce the problem of data islands, and improve the overall efficiency of the monitoring system. In summary, this step generates a set of reaction force and deformation data through real-time collection by sensors, and generates multi-element monitoring data through time alignment and spatial matching by data fusion algorithms, enhancing the comprehensiveness and consistency of the data, achieving multi-dimensional perception of the surrounding rock state, and optimizing the data quality through fusion processing, ultimately providing reliable input for intelligent support.

[0066] In some implementations, S4 includes:

[0067] S4.1: Convert the stage identification result into a text format through a report generation sub-algorithm to generate a text report;

[0068] S4.2: Map the text report to graphical elements through a visualization sub-algorithm to generate graphical display data;

[0069] S4.3: Based on the text report and the graphical display data, integrate the text and graphical information through a state generation algorithm to generate surrounding rock stability state information.

[0070] In the implementation of S4, the stage recognition results are converted into text format by a report generation sub-algorithm, which translates the identifiers and parameters in the stage recognition results into structured text descriptions based on natural language processing or template filling techniques, generating a text report that includes the surrounding rock deformation stage, energy state summary, and recommended measures. The text report adopts a standardized format to ensure clear and readable information, facilitating quick understanding of the surrounding rock state by engineers. Meanwhile, the text report is mapped to graphical elements through a visualization sub-algorithm, which uses chart generation libraries or graphics engines to convert key data in the text into line charts, heat maps, or three-dimensional models, generating graphical display data that presents the surrounding rock deformation trend and energy dynamics in a visual form. The visualization sub-algorithm focuses on the intuitiveness and interactivity of graphical elements, such as color coding to represent risk levels, enhancing user experience. Based on the text report and graphical display data, a state generation algorithm integrates the text and graphical information. The state generation algorithm uses information fusion techniques (such as weight-based data fusion or decision tree models), which specifically include: parsing key parameters (such as deformation stage identifiers and energy values) from the text report and extracting visual features (such as trend line slopes or heat map intensities) from the graphical display data; then, combining the text and graphical information through fusion rules (for example, if the text report indicates "accelerating deformation stage" and the graphical display shows a continuous increase in energy release value, a "high risk" state is generated) to generate surrounding rock stability state information. The surrounding rock stability state information provides comprehensive state assessment and early warning prompts, such as outputting in a structured report format, including a text summary and graphical references. The state generation algorithm ensures the consistency and operability of the output information through logical verification (such as consistency checks to ensure that the text and graphical information do not conflict). The technical effect lies in the report generation sub-algorithm converting the stage recognition results into readable text, facilitating recording and communication, the visualization sub-algorithm enhancing information expression through graphical elements, and the state generation algorithm integrating both to generate comprehensive surrounding rock stability state information, improving decision support capabilities. The text report describes the stage recognition results in detail in written form, ensuring accurate information transmission, the graphical display data highlights key trends through visual means, improving information absorption efficiency, and the state generation algorithm integrates text and graphics, forming a multi-modal output, ensuring the comprehensiveness and practicality of the surrounding rock stability state information.

[0071] The synergistic effect of this step is reflected in the fact that the report generation sub-algorithm and the visualization sub-algorithm handle text and graphical outputs respectively, while the status generation algorithm acts as a consolidation point, integrating the two into a coherent whole, forming a closed-loop process from result transformation to status generation, supporting engineering applications. Through this implementation, the system can provide intuitive and detailed status information, facilitating timely action by personnel and enhancing the effectiveness of tunnel management. In summary, this step generates text reports through the report generation sub-algorithm, generates graphical data displays through the visualization sub-algorithm, and integrates and generates surrounding rock stability status information through the status generation algorithm. This improves the readability and usability of the information output, realizes multimodal expression of status assessment, and optimizes decision support through integration, ultimately improving the response efficiency of the support system and the safety of the project.

[0072] Example 2

[0073] like Figure 2 As shown, in a second aspect, the present invention proposes a tunnel support structure safety monitoring system based on rock mass-support feedback. The system employs the method provided in any of the above embodiments, and the system includes:

[0074] The data acquisition module is used to collect support reaction force data through pressure sensors and deformation data through displacement sensors based on a sensor network deployed on the support structure. The support reaction force data and deformation data are then fused to generate multi-dimensional monitoring data.

[0075] The energy analysis module is used to extract reaction force change characteristics and deformation change characteristics based on multi-source monitoring data through energy analysis algorithms, and to calculate the energy accumulation value and energy release value of the surrounding rock based on the reaction force change characteristics and deformation change characteristics, and to combine them to generate energy state parameters;

[0076] The stage identification module is used to analyze the dynamic changes of energy accumulation and energy release values ​​based on energy state parameters and stage identification algorithms, and to identify the current deformation stage of the surrounding rock based on the dynamic changes, generating stage identification results.

[0077] The state generation module is used to convert the stage identification results into a description of the surrounding rock stability state through a state generation algorithm, thereby generating surrounding rock stability state information.

[0078] This system corresponds to the method provided in Embodiment 1 above, and will not be described in detail here.

[0079] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions recorded in the above embodiments can still be modified, or some or all of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the technical solutions of the embodiments of the present application.

Claims

1. A method for safety monitoring of tunnel support structures based on rock mass-support feedback, characterized in that, include: S1: Based on the sensor network deployed on the support structure, pressure sensors collect support reaction force data, and displacement sensors collect deformation data. The support reaction force data and deformation data are fused together to generate multi-dimensional monitoring data. S2: Based on multivariate monitoring data, the reaction force change characteristics and deformation change characteristics are extracted through energy analysis algorithms. Based on these characteristics, the energy accumulation and energy release values ​​of the surrounding rock are calculated, and combined to generate energy state parameters. The energy analysis algorithm includes a data preprocessing sub-algorithm and a feature extraction sub-algorithm. S2 includes: S2.1: Based on multi-source monitoring data, the support reaction force data is filtered and the deformation data is denoised through a data preprocessing sub-algorithm to generate clean monitoring data. S2.2: Based on cleaning monitoring data, the feature extraction sub-algorithm extracts the reaction force change feature value from the support reaction force data and the deformation change feature value from the deformation data to generate a feature dataset; S2.3: Based on the feature dataset, the energy calculation sub-algorithm calculates the energy accumulation value using the reaction force change characteristics and the energy release value using the deformation change characteristics. The energy accumulation value and the energy release value are combined to generate the energy state parameters. S3: Based on energy state parameters, the dynamic changes of energy accumulation and energy release values ​​are analyzed through a stage identification algorithm, and the current deformation stage of the surrounding rock is identified based on the dynamic changes, generating stage identification results; S4: Through the state generation algorithm, the stage identification results are converted into a description of the surrounding rock stability state, generating surrounding rock stability state information.

2. The method for safety monitoring of tunnel support structures based on rock mass-support feedback according to claim 1, characterized in that, S2.1 includes: S2.1.1: Based on multi-source monitoring data, high-frequency noise is removed from the support reaction force data to obtain the first filtered data; S2.1.2: Based on the first filtered data, perform low-frequency drift correction on the deformed data to obtain the second filtered data; S2.1.3: Based on the second filtered data, identify and remove abnormal points in the support reaction force data and deformation data to generate cleaning monitoring data.

3. The method for safety monitoring of tunnel support structures based on rock mass-support feedback according to claim 1, characterized in that, S2.2 includes: S2.2.1: Based on the cleaning monitoring data, the statistical characteristics of the support reaction force data are calculated through the reaction force feature extraction sub-algorithm to obtain the reaction force change characteristic value; S2.2.2: Based on cleaning monitoring data, the statistical characteristics of the deformation data are calculated through the deformation feature extraction sub-algorithm to obtain the deformation change feature value; S2.2.3: The feature combination algorithm is used to merge the reaction force change feature values ​​and deformation change feature values ​​into a unified dataset to generate a feature dataset.

4. The method for safety monitoring of tunnel support structures based on rock mass-support feedback according to claim 1, characterized in that, The stage identification algorithm includes a trend analysis sub-algorithm, a rate of change calculation sub-algorithm, and an amplitude identification sub-algorithm. S3 includes: S3.1: Based on the energy state parameters, the trend analysis sub-algorithm is used to analyze the trend of energy accumulation and energy release values ​​and generate the first deformation stage identifier; S3.2: Based on the energy state parameters, the rate of change of energy accumulation and energy release values ​​is calculated through the rate of change calculation sub-algorithm to generate the second deformation stage identifier; S3.3: Based on the energy state parameters, the amplitude of the energy accumulation value and the energy release value are determined through the amplitude recognition sub-algorithm, and the third deformation stage identifier is generated; S3.4: Integrate the first deformation stage identifier, the second deformation stage identifier, and the third deformation stage identifier through the comprehensive sub-algorithm to generate the stage identification result.

5. The method for safety monitoring of tunnel support structures based on rock mass-support feedback according to claim 4, characterized in that, S3.1 includes: S3.1.1: Through the trend analysis sub-algorithm, calculate the changing trend of energy accumulation value within a preset time window to obtain accumulation trend information; and calculate the changing trend of energy release value within a preset time window to obtain release trend information; S3.1.2: When the accumulated trend information and the released trend information meet the preset stability conditions, the first deformation stage identifier is generated through the identifier generation sub-algorithm.

6. The method for safety monitoring of tunnel support structures based on rock mass-support feedback according to claim 4, characterized in that, S3.2 includes: S3.2.1: Calculate the rate of change of energy accumulation value using the rate of change calculation sub-algorithm to obtain the accumulation rate of change; and calculate the rate of change of energy release value to obtain the release rate of change. S3.2.2: When the accumulated rate of change and the released rate of change meet the preset acceleration conditions, the second deformation stage identifier is generated through the identifier generation sub-algorithm.

7. The method for safety monitoring of tunnel support structures based on rock mass-support feedback according to claim 1, characterized in that, S1 includes: S1.1: Based on pressure sensors deployed on the support structure, support reaction force data is collected in real time to obtain a reaction force dataset; S1.2: Based on displacement sensors deployed on the support structure, deformation data is collected in real time to obtain a deformation dataset; S1.3: Through data fusion algorithms, the reaction force dataset and deformation dataset are time-aligned and spatially matched to generate multi-dimensional monitoring data.

8. The method for safety monitoring of tunnel support structures based on rock mass-support feedback according to claim 1, characterized in that, S4 include: S4.1: Convert the stage recognition results into text format and generate a text report using the report generation sub-algorithm; S4.2: The text report is mapped to graphical elements through a visualization sub-algorithm to generate graphically displayed data; S4.3: Based on text reports and graphical display data, the stability status information of the surrounding rock is generated by integrating text and graphical information through a state generation algorithm.

9. A tunnel support structure safety monitoring system based on rock mass-support feedback, characterized in that, The system employs the method described in any one of claims 1 to 8, the system comprising: The data acquisition module is used to collect support reaction force data through pressure sensors and deformation data through displacement sensors based on a sensor network deployed on the support structure. The support reaction force data and deformation data are then fused to generate multi-dimensional monitoring data. The energy analysis module is used to extract reaction force change characteristics and deformation change characteristics based on multi-source monitoring data through energy analysis algorithms, and to calculate the energy accumulation value and energy release value of the surrounding rock based on the reaction force change characteristics and deformation change characteristics, and to combine them to generate energy state parameters; The stage identification module is used to analyze the dynamic changes of energy accumulation and energy release values ​​based on energy state parameters and stage identification algorithms, and to identify the current deformation stage of the surrounding rock based on the dynamic changes, generating stage identification results. The state generation module is used to convert the stage identification results into a description of the surrounding rock stability state through a state generation algorithm, thereby generating surrounding rock stability state information.

Images