A method and system for safety assessment of underground tunnel construction

Abstract

The present invention relates to an underground tunnel construction safety assessment method and system, which relates to the field of underground engineering, and includes: interactive target scenes, obtaining multidimensional assessment targets; based on a multidimensional assessment target interactive knowledge base, obtaining an assessment indicator set, wherein the assessment indicator set includes multiple assessment indicator groups, and the multiple assessment indicator groups correspond to the multidimensional assessment targets one by one; obtaining historical construction safety assessment records, parsing the historical construction safety assessment records, and configuring the assessment weight of the multidimensional assessment targets; according to the assessment weight, constructing a safety assessment model, and combining the historical construction safety assessment records with an integrated learning method to train the safety assessment model; based on the assessment indicator set, real-time collection of construction monitoring information of the target scene, and inputting the construction monitoring information into the safety assessment model for assessment analysis.

Description

Technical Field

[0001] The present invention relates to the field of underground engineering, and in particular to a method and system for evaluating the safety of underground tunnel construction. Background Art

[0002] In the field of underground tunnel construction, construction safety assessment is a key link to ensure the smooth progress of the project and the safety of personnel. Traditional safety assessment methods mainly rely on manual experience judgment and simple monitoring of some key indicators, such as the detection of gas concentration in the tunnel and local structural stability. On the one hand, manual experience judgment is highly subjective, and the assessment results of different assessors for the same construction scene may vary greatly, and it is difficult to fully consider the mutual influence of many complex factors in the underground tunnel construction process. On the other hand, only monitoring some key indicators cannot form a multi-dimensional and systematic assessment of construction safety, and it is easy to miss some potential safety hazards, resulting in inaccurate and incomplete assessment results. Summary of the invention

[0003] The present invention aims to solve the technical problem in the prior art that the subjectivity and one-sidedness of assessment affect the accuracy of assessment, and provides an underground tunnel construction safety assessment method and system to solve the problem.

[0004] The technical solution of the present invention to solve the above technical problems is as follows:

[0005] In a first aspect, the present invention provides a method for safety assessment of underground tunnel construction:

[0006] Interactive target scenarios, obtain multi-dimensional evaluation targets.

[0007] Based on the multidimensional evaluation target interactive knowledge base, an evaluation indicator set is obtained, wherein the evaluation indicator set includes multiple evaluation indicator groups, and the multiple evaluation indicator groups correspond one-to-one to the multidimensional evaluation targets.

[0008] Obtain historical construction safety assessment records, parse the historical construction safety assessment records, and configure the assessment weight of the multi-dimensional assessment target.

[0009] According to the assessment severity, a safety assessment model is constructed, and the safety assessment model is trained in combination with the historical construction safety assessment records and an integrated learning method.

[0010] Based on the evaluation index set, the construction monitoring information of the target scene is collected in real time, and the construction monitoring information is input into the safety assessment model for evaluation and analysis.

[0011] In a second aspect, the present invention provides an underground tunnel construction safety assessment system:

[0012] The target scene interaction module is used to interact with the target scene and obtain multi-dimensional evaluation targets.

[0013] The evaluation indicator set acquisition module is used to acquire an evaluation indicator set based on the multidimensional evaluation target interactive knowledge base, wherein the evaluation indicator set includes multiple evaluation indicator groups, and the multiple evaluation indicator groups correspond one-to-one to the multidimensional evaluation targets.

[0014] The assessment weight configuration module is used to obtain historical construction safety assessment records, parse the historical construction safety assessment records, and configure the assessment weight of the multi-dimensional assessment target.

[0015] The safety assessment model construction and training module is used to construct a safety assessment model according to the assessment severity, and train the safety assessment model in combination with the historical construction safety assessment records and an integrated learning method.

[0016] The real-time evaluation and analysis module is used to collect the construction monitoring information of the target scene in real time based on the evaluation index set, and input the construction monitoring information into the safety assessment model for evaluation and analysis.

[0017] The beneficial effects of the present invention are as follows: by acquiring target scene information, a multidimensional evaluation target is determined; through an interactive knowledge base, an evaluation indicator set is extracted according to the multidimensional evaluation target, the evaluation indicator set is composed of multiple evaluation indicator groups, and each group of indicators corresponds to the multidimensional evaluation target one by one; historical construction safety assessment records are collected, the record data are parsed, and the evaluation weights of the multidimensional evaluation targets are configured; based on the evaluation weights, a safety assessment model is constructed, and the model is trained by applying an integrated learning method in combination with the historical construction safety assessment records; construction monitoring information of the target scene is collected in real time, and based on the extracted evaluation indicator set, the construction monitoring information is input into the trained safety assessment model for analysis and evaluation, thereby achieving a technical effect of multidimensional, systematic and accurate evaluation of underground tunnel construction safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of a process flow of an underground tunnel construction safety assessment method provided by the present invention;

[0019] Figure 2 A schematic structural diagram of an underground tunnel construction safety assessment provided by the present invention.

[0020] In the accompanying drawings, the components represented by the reference numerals are described as follows:

[0021] Target scenario interaction module 11, evaluation indicator set acquisition module 12, evaluation heavy configuration module 13, security assessment model construction and training module 14, real-time assessment analysis module 15. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0023] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0024] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0025] Embodiment 1:

[0026] like Figure 1 As shown, an embodiment of the present invention provides a method for safety assessment of underground tunnel construction.

[0027] Interactive target scenarios, obtain multi-dimensional evaluation targets.

[0028] Specifically, the multidimensional assessment target refers to multiple assessment dimensions related to the safety assessment of the target scenario, including, for example, structural safety, environmental safety, equipment safety, personnel safety, construction process compliance, etc.

[0029] Specifically, the target scenario refers to the specific areas or stages that need to be focused on and evaluated during the underground tunnel construction process, such as the initial support construction stage, the tunnel boring machine operation area, the construction of complex geological sections, the temporary support removal process, etc. The multi-dimensional evaluation targets corresponding to different stages are not the same.

[0030] By obtaining multi-dimensional evaluation targets through interactive target scenarios, the evaluation is no longer limited to a single dimension or a few key indicators, but can comprehensively consider various factors in the underground tunnel construction process, such as geological conditions, construction technology, equipment operating status, personnel operating specifications and other dimensions, thereby providing a more comprehensive and systematic basis for the subsequent acquisition of evaluation indicators and model construction, effectively overcoming the one-sidedness of traditional methods.

[0031] Based on the multidimensional evaluation target interactive knowledge base, an evaluation indicator set is obtained, wherein the evaluation indicator set includes multiple evaluation indicator groups, and the multiple evaluation indicator groups correspond one-to-one to the multidimensional evaluation targets.

[0032] Specifically, the multidimensional evaluation target interactive knowledge base is a database containing multidimensional targets and corresponding knowledge related to the safety assessment of underground tunnel construction. By interacting with the knowledge base, detailed information and evaluation indicators related to each evaluation target can be obtained; the evaluation indicator set is a set of multiple evaluation indicator groups, which are used to quantify and evaluate specific standards for underground tunnel construction safety, such as geological stability, reliability of construction equipment, personnel operating specifications, etc., and each evaluation indicator group corresponds to a specific multidimensional evaluation target.

[0033] Specifically, based on the determined multidimensional evaluation objectives, the multidimensional evaluation objective interaction knowledge base is interacted with to obtain the evaluation indicator group corresponding to each multidimensional evaluation objective, thus forming a complete evaluation indicator set. Each evaluation indicator group contains multiple specific evaluation indicators, and each evaluation indicator represents a quantitative analysis of the security or compliance of a specific dimension.

[0034] For example, multidimensional evaluation targets may include geological stability, reliability of construction equipment, personnel operating specifications, etc. By interacting with the knowledge base, evaluation indicators related to geological stability (such as rock strength, groundwater level changes, etc.), evaluation indicators related to construction equipment reliability (such as equipment failure rate, maintenance records, etc.), and evaluation indicators related to personnel operating specifications (such as safety training records, operating error rates, etc.) can be obtained. These evaluation indicator groups together constitute a comprehensive evaluation indicator set, which provides a basis for the subsequent construction and training of safety assessment models.

[0035] Through the interaction between multi-dimensional evaluation objectives and the knowledge base, a set of evaluation indicators that accurately corresponds to them can be obtained to ensure that the obtained evaluation indicator group is both comprehensive and targeted, covering various key indicators required for underground tunnel construction safety assessment, and providing strong support for accurate assessment of construction safety.

[0036] Obtain historical construction safety assessment records, parse the historical construction safety assessment records, and configure the assessment weight of the multi-dimensional assessment target.

[0037] Specifically, the historical construction safety assessment record contains the safety assessment data and results conducted during the past underground tunnel construction process, illustratively including geological conditions, operating status of construction equipment, personnel operation records, safety accident records, etc. The historical construction safety assessment record is used to provide a reference for current and future construction safety assessments.

[0038] Specifically, assessment weight refers to the quantitative configuration of the importance of multi-dimensional assessment targets. By analyzing historical construction safety assessment records, the weight of each assessment target in the overall safety assessment can be determined. For example, geological conditions may have a greater impact on construction safety, so their assessment weight is higher; while some minor construction details may have a smaller impact, and their assessment weight is lower.

[0039] By analyzing the equipment failure records in historical data, appropriate weights can be assigned to the reliability assessment targets of construction equipment, which in turn helps to build a more scientific and reasonable safety assessment model to ensure that the assessment results can fully and accurately reflect the actual situation of construction safety.

[0040] In some embodiments, obtaining historical construction safety assessment records, parsing the historical construction safety assessment records, and configuring the assessment weight of the multi-dimensional assessment target include:

[0041] Based on the multidimensional assessment target, the historical construction safety assessment records are divided; the division results are traversed, and the target assessment feature data of each assessment target is analyzed and obtained; the target assessment feature data are merged to obtain the assessment severity coefficient set corresponding to the multidimensional assessment target; based on the severity coefficient set, the relative assessment severity coefficient of the multidimensional assessment target is calculated, and the output is the assessment severity.

[0042] Specifically, the historical construction safety assessment records extracted from the construction safety database include: assessment time dimension, that is, the time stamp of each record, which is convenient for tracking changes in safety conditions over time; assessment objectives, which are used to clarify each assessment objective, such as structural safety, environmental safety, equipment safety, etc.; assessment indicators, that is, the specific indicators of each objective and its corresponding historical assessment data; assessment conclusions, that is, the overall result of each assessment, such as qualified, unqualified or a specific risk level.

[0043] Specifically, the target assessment characteristic data are specific characteristics related to each assessment target, which are used to describe and quantify the performance of the assessment target in historical construction. For example, for the assessment target of geological conditions, the target assessment characteristic data may include rock strength, groundwater level changes, frequency of geological disasters, etc.

[0044] Specifically, the evaluation weight coefficient set is a set of multiple evaluation weight coefficients, each corresponding to an evaluation target. The evaluation weight coefficient is used to quantify the importance of each evaluation target and is calculated by integrating the target evaluation feature data. The relative evaluation weight coefficient is a coefficient obtained by normalizing the evaluation weight coefficient set, which is used to indicate the relative importance of each evaluation target in the overall evaluation.

[0045] Specifically, firstly, based on the classification rules of multidimensional assessment targets, historical assessment records are divided into different subsets; for example, historical records for structural safety targets include indicator data such as surrounding rock pressure, support displacement and lining stress; historical records for environmental safety targets include different categories such as groundwater level, temperature and humidity, so as to analyze each assessment target separately; then, each classified record is analyzed in detail to extract characteristic data related to the assessment target, including: difficulty characteristics of the complexity of the work related to the assessment target and the technical challenges; accuracy characteristics indicating the precision of monitoring, evaluation or analysis of the assessment target in historical records; frequency characteristics of the target in historical construction assessment records, reflecting its importance or typicality; the potential impact of the assessment target on construction safety, reflecting its importance and relevance, etc.

[0046] Specifically, through data fusion technology, the extracted feature data are integrated to calculate the assessment weight coefficient of each assessment target, where data fusion can adopt weighted average, fuzzy logic, neural network and other methods; then, the assessment weight coefficient set is normalized, and the relative assessment weight coefficient of each assessment target is calculated. The relative assessment weight coefficient reflects the deviation of each assessment target relative to the average weight coefficient, which is used for the subsequent construction of the security assessment model to ensure that the weight distribution of the assessment model is reasonable.

[0047] By acquiring and parsing historical construction safety assessment records, the assessment weight of multi-dimensional assessment targets is configured. This process abandons the traditional method of relying on manual experience and judgment. Through objective analysis of historical data, the importance of each assessment target can be determined more accurately, which provides a more scientific and reasonable weight distribution for the construction of the safety assessment model, and further improves the objectivity and accuracy of the assessment results.

[0048] According to the assessment severity, a safety assessment model is constructed, and the safety assessment model is trained in combination with the historical construction safety assessment records and an integrated learning method.

[0049] Specifically, the security assessment model is an assessment model obtained by training a plurality of benchmark assessment models with basic assessment capabilities through an integrated learning method, wherein the assessment weight is used to define the structure of the benchmark assessment model for constructing the security assessment model, including the proportion of computing power consumption of a class of benchmark models corresponding to each assessment target; in other words, the above-mentioned security assessment model includes multiple groups of benchmark assessment models, each group of benchmark assessment models corresponds to an assessment target, and targeted training is performed through corresponding sample data to obtain specific assessment capabilities under the assessment target. In addition, the number of models included in each group of benchmark assessment models is constrained by the assessment weight, thereby ensuring that the performance overhead of each group of benchmark assessment models is consistent with the importance of the corresponding assessment target, which in turn helps to improve the adaptability and assessment efficiency of the model, and prioritizes the assessment accuracy of assessment targets with high attention and importance.

[0050] Specifically, the above-mentioned multiple groups of benchmark evaluation models with specific evaluation capabilities are combined through an integrated learning method. Exemplarily, the combination method includes: voting on the prediction results of multiple groups of benchmark evaluation models, and selecting the prediction results of the majority of models as the final output; taking a weighted average of the prediction results of each benchmark evaluation model, and adjusting the weight according to the evaluation severity, etc.

[0051] In the above steps, a safety assessment model is constructed according to the assessment weight, and is trained in combination with historical construction safety assessment records and ensemble learning methods. The ensemble learning method can give full play to the advantages of multiple learning algorithms. Through in-depth mining and analysis of historical data, the model can learn the complex laws and patterns in the safety assessment of underground tunnel construction, thereby improving the assessment accuracy and generalization ability of the model, enabling it to better adapt to the safety assessment needs under different construction scenarios and conditions, and effectively solving the problems of strong subjectivity and inaccurate assessment results of existing methods.

[0052] In some embodiments, a safety assessment model is constructed according to the assessment severity, and the safety assessment model is trained by combining the historical construction safety assessment records with an integrated learning method, before that, the steps include:

[0053] Based on the machine learning model structure, multiple benchmark evaluation models are constructed; the computing power consumption of each benchmark evaluation model is evaluated, and the evaluation results are associated with the multiple benchmark evaluation models and stored, and output as a benchmark model library.

[0054] Specifically, the benchmark assessment model is a generalized security assessment model with a simpler structure and higher efficiency, but average accuracy. The benchmark model library is a collection that stores multiple benchmark assessment models and their assessment results, which is used to quickly compare and select model structures and improve the efficiency of model construction.

[0055] Specifically, first, select a variety of common machine learning model structures, such as decision trees, support vector machines, neural networks, etc., and build benchmark evaluation models respectively. These models will serve as references for subsequent evaluation and selection; then, train and test each benchmark evaluation model, record its computing power consumption during training and prediction, including CPU and GPU usage, running time, etc., and associate the computing power consumption evaluation results of each benchmark evaluation model with its model structure to form a benchmark model library for quickly querying and comparing the performance and resource consumption of different model structures, helping to select the most suitable model structure;

[0056] The above-mentioned evaluation results of computing power consumption can help select a model structure that can run efficiently in an environment with limited resources, thereby improving the practicality and deployment efficiency of the model. At the same time, the benchmark model library also provides a platform for fast query and comparison, which can quickly select the model structure that best suits the current task and reduce the time cost of model construction and evaluation.

[0057] In some embodiments, a safety assessment model is constructed according to the assessment severity, and the safety assessment model is trained by combining the historical construction safety assessment records with an integrated learning method, including:

[0058] The actual evaluation computing power of the target scenario is obtained to determine the upper limit of resources that can be used for safety evaluation; based on the evaluation weight, the expected allocated computing power of each evaluation target in the multi-dimensional evaluation target is calculated and determined; using the benchmark model library as the selection space and the expected allocated computing power as the upper limit constraint, a random selection is performed to obtain multiple benchmark model groups; based on the historical construction safety evaluation records, an integrated training of multiple benchmark model groups is performed to generate the safety evaluation model.

[0059] Specifically, the actual evaluation computing power refers to the actual computing resources that can be used for security evaluation in the target scenario. This includes the number of available CPUs and GPU cores, memory size, etc. The determination of the actual evaluation computing power helps to reasonably allocate resources and ensure the efficiency of the model training and evaluation process; the expected allocation computing power is the computing resources that should be allocated to each evaluation target calculated based on the evaluation weight, which is used to ensure that resource allocation matches the importance of the evaluation target and improve resource utilization efficiency.

[0060] Specifically, first, the actual computing resources available for safety assessment in the target scenario are obtained through the system resource monitoring tool, including the number of CPU and GPU cores and memory size, so as to determine the resource upper limit and ensure that the model training and evaluation process will not exceed the available resources; then, according to the evaluation weight of each evaluation target, the computing resources to be allocated are calculated. The expected allocation computing power of each evaluation target is proportional to its evaluation weight. In other words, targets with higher evaluation weight are allocated more computing power to ensure that resource allocation matches the importance of the evaluation target; then, with the benchmark model library as the selection space, according to the expected allocation computing power of each evaluation target, several benchmark models are randomly selected from the model library to form multiple benchmark model groups, and it is ensured that the computational complexity of each benchmark model group does not exceed the upper limit of the allocation computing power of the corresponding evaluation target. Preferably, each benchmark model group contains multiple different benchmark evaluation models to ensure diversity; finally, multiple benchmark model groups are independently trained using historical construction safety assessment records, and the final safety assessment model is generated by combining multiple trained benchmark models through ensemble learning methods such as voting, stacking, and boosting. Ensemble training can improve the accuracy and robustness of the model.

[0061] Through the above method, under the constraints of actual computing resources, the evaluation can be used to heavily optimize the model selection and training process to achieve the construction of an efficient evaluation model for underground tunnel construction safety, ensuring that the model runs efficiently in an environment with limited resources while improving the accuracy and reliability of the evaluation results.

[0062] In some implementations, performing integrated training of multiple benchmark model groups based on the historical construction safety assessment records includes:

[0063] Based on the historical construction safety assessment records, multiple baseline model groups are initially trained indiscriminately to obtain a first model set; the division results of the historical construction safety assessment records are used as the training data selection space to perform specialized training on multiple first model groups in the first model set respectively; and the first model groups after multiple specialized trainings are integrated to obtain the safety assessment model.

[0064] Specifically, indiscriminate initial training refers to the unified preliminary training of multiple benchmark model groups without optimization for specific evaluation targets. The purpose is to allow all benchmark models to learn on the same data set and lay the foundation for subsequent specialized training; specialized training refers to further training on the specific feature data of each evaluation target on the basis of indiscriminate initial training, so as to improve the adaptability and accuracy of models in different groups to specific evaluation targets.

[0065] Specifically, the first model set is a plurality of model groups obtained after indiscriminate initial training, and each model group contains a plurality of baseline models that have undergone preliminary training; the division results refer to the results after the historical construction safety assessment records are divided according to multi-dimensional assessment objectives, and each division result contains data related to a specific assessment objective.

[0066] Specifically, first, use the complete historical construction safety assessment records (including assessment target feature data and assessment result data) to conduct unified preliminary training for multiple benchmark model groups, so that all benchmark models can learn on the same data set to ensure that each model has a basic performance level; for each model group in the first model set, use the corresponding division results to perform specialized training to improve the model's adaptability and accuracy to specific assessment targets. For example, for geological condition assessment targets, use data related to geological conditions for specialized training; for construction equipment assessment targets, use data related to construction equipment for specialized training.

[0067] Specifically, through ensemble learning methods, multiple first model groups that have undergone specialized training are integrated, such as voting, stacking, and boosting techniques, to generate a final security assessment model. By integrating multiple specialized trained models and combining the prediction results of multiple models to generate the final security assessment model, the robustness and generalization ability of the model can be improved, the error of a single model can be reduced, and the accuracy and reliability of the assessment results can be improved.

[0068] Based on the evaluation index set, the construction monitoring information of the target scene is collected in real time, and the construction monitoring information is input into the safety assessment model for evaluation and analysis.

[0069] Specifically, the construction monitoring information of the target scene is collected in real time based on the evaluation indicator set and input into the safety assessment model for evaluation and analysis. This real-time and dynamic evaluation method can timely capture various changes and potential risks in the underground tunnel construction process, provide real-time safety warnings and decision-making support for construction personnel and managers, and help take measures in advance to prevent the occurrence of safety accidents, further ensuring the safety and reliability of underground tunnel construction.

[0070] In some embodiments, collecting construction monitoring information of the target scene in real time and inputting the construction monitoring information into the safety assessment model for assessment and analysis includes:

[0071] According to the evaluation indicator set, the monitoring sensor status of the target scene is checked and adaptively adjusted; the collection frequency and collection parameters are set corresponding to the evaluation indicator set, and the monitoring sensor is activated for sensor collection; the sensor collection result is transmitted to the security assessment model to obtain the security assessment result, wherein the security assessment result at least includes the security level, unsafe location, and unsafe category.

[0072] Specifically, first, according to the requirements of the evaluation indicator set, check the working status of all monitoring sensors in the target scene, including the accuracy, sensitivity, online status, etc. of the sensors. For problems found, such as sensor damage, inaccurate data, etc., make necessary adaptive adjustments, such as calibrating sensors, replacing damaged sensors, adjusting the layout of sensors, etc., so as to ensure that the collected data is reliable and effective; then, according to the specific requirements of the evaluation indicator set, set appropriate collection frequency and collection parameters for each evaluation indicator. For example, for geological condition monitoring, a higher collection frequency and high-precision measurement parameters may be required; then, activate the monitoring sensor to collect data according to the set frequency and parameters, and transmit the construction monitoring information collected by the sensor to the safety assessment model in real time. The safety assessment model evaluates and analyzes the input data and generates safety assessment results, including safety level, the overall safety status of the current construction scene, such as "safe", "warning", "dangerous"; unsafe location, that is, the specific location of hidden dangers in the area; unsafe category, that is, the type of hidden danger, such as collapse risk, equipment failure, gas leakage, etc.

[0073] The above steps, by checking and adjusting the monitoring sensors, ensure that the collected data is reliable and valid, provide an accurate data basis for safety assessment, set appropriate collection frequency and parameters, ensure that key information in the construction process can be collected in real time, and potential safety problems can be discovered in time. The collected data is transmitted to the safety assessment model in real time, and the assessment results are generated quickly, so that construction management personnel can understand the construction safety status in a timely manner and take necessary measures to prevent safety accidents.

[0074] Further, based on the multi-dimensional evaluation target interactive knowledge base, an evaluation indicator set is obtained, and then the method further includes:

[0075] Performing principal component analysis on the evaluation indicator set, identifying key evaluation indicators and redundant evaluation indicators, and outputting the results of the indicator analysis; streamlining the evaluation indicator set according to the results of the indicator analysis, removing the redundant evaluation indicators, and retaining the key evaluation indicators.

[0076] Specifically, principal component analysis (PCA) is used to analyze the evaluation indicator set to identify key evaluation indicators that have a greater impact on the evaluation results and redundant evaluation indicators that have a smaller impact, and the output is an indicator analysis result. Exemplarily, the indicator analysis result includes the contribution rate of each evaluation indicator and a mark of whether it is a key evaluation indicator; then, based on the indicator analysis result, redundant evaluation indicators with lower contribution rates are removed, and key evaluation indicators with higher contribution rates are retained to generate a streamlined evaluation indicator set for subsequent security assessment model construction and training.

[0077] Through the above-mentioned principal component analysis, it is possible to identify key evaluation indicators that have a greater impact on the evaluation results, ensure the accuracy and reliability of the evaluation model, and at the same time remove redundant evaluation indicators that have a smaller impact on the evaluation results, simplify the evaluation model, and improve the model's operating efficiency and interpretability; through the streamlined set of evaluation indicators, a more efficient and accurate security assessment model can be built, improving the practicality and deployment efficiency of the model.

[0078] The method of the present invention can realize multi-dimensional interaction of evaluation targets, thereby obtaining a comprehensive set of evaluation indicators, overcoming the one-sidedness of traditional methods. By parsing historical construction safety assessment records and configuring assessment weights, an accurate data basis is provided for the construction of a safety assessment model, reducing the subjective errors caused by manual experience judgment. By training the safety assessment model with the help of an integrated learning method, the assessment accuracy and generalization ability of the model can be effectively improved, enabling it to better adapt to the safety assessment needs of different underground tunnel construction scenarios. Real-time collection of construction monitoring information and input into the model for evaluation and analysis realizes dynamic and real-time monitoring of construction safety, timely discovers potential safety hazards, and provides a more reliable and systematic guarantee for underground tunnel construction safety.

[0079] An underground tunnel construction safety assessment method provided by an embodiment of the present invention has at least the following technical effects:

[0080] By acquiring the target scene information, the multi-dimensional evaluation target is determined; through the interactive knowledge base, the evaluation indicator set is extracted according to the multi-dimensional evaluation target, and the evaluation indicator set consists of multiple evaluation indicator groups, and each group of indicators corresponds to the multi-dimensional evaluation target one by one; historical construction safety assessment records are collected, the record data is analyzed, and the evaluation weights of the multi-dimensional evaluation targets are configured; according to the evaluation weights, a safety assessment model is constructed, and the model is trained by applying an integrated learning method in combination with the historical construction safety assessment records; the construction monitoring information of the target scene is collected in real time, and based on the extracted evaluation indicator set, the construction monitoring information is input into the trained safety assessment model for analysis and evaluation, thereby achieving the technical effect of multi-dimensional, systematic and accurate evaluation of underground tunnel construction safety.

[0081] Embodiment 2:

[0082] like Figure 2 As shown, based on the same inventive concept as the underground tunnel construction safety assessment method provided in the first embodiment, the embodiment of the present invention further provides an underground tunnel construction safety assessment system, including:

[0083] A target scene interaction module 11 is used to interact with the target scene and obtain a multi-dimensional evaluation target;

[0084] An evaluation indicator set acquisition module 12 is used to acquire an evaluation indicator set based on the multidimensional evaluation target interactive knowledge base, wherein the evaluation indicator set includes a plurality of evaluation indicator groups, and the plurality of evaluation indicator groups correspond one-to-one to the multidimensional evaluation targets;

[0085] An assessment weight configuration module 13 is used to obtain historical construction safety assessment records, parse the historical construction safety assessment records, and configure the assessment weight of the multi-dimensional assessment target;

[0086] A safety assessment model construction and training module 14 is used to construct a safety assessment model according to the assessment severity, and train the safety assessment model in combination with the historical construction safety assessment records and an integrated learning method;

[0087] The real-time evaluation and analysis module 15 is used to collect the construction monitoring information of the target scene in real time based on the evaluation index set, and input the construction monitoring information into the safety evaluation model for evaluation and analysis.

[0088] In some embodiments, the evaluation reconfiguration module 13:

[0089] A historical record division unit, used for dividing the historical construction safety assessment records based on the multi-dimensional assessment target;

[0090] A target evaluation feature data analysis unit, used to traverse the division results, analyze and obtain target evaluation feature data of each evaluation target;

[0091] An evaluation weight coefficient set acquisition unit, used for fusing the target evaluation feature data to acquire an evaluation weight coefficient set corresponding to the multi-dimensional evaluation target;

[0092] The relative evaluation severity coefficient calculation unit is used to calculate the relative evaluation severity coefficient of the multidimensional evaluation target based on the severity coefficient set, and output it as the evaluation severity.

[0093] In some embodiments, the safety assessment model building and training module 14 includes a baseline model building unit for:

[0094] Build multiple benchmark evaluation models based on the machine learning model structure;

[0095] The computing power consumption of each benchmark evaluation model is evaluated, and the evaluation results are associated with multiple benchmark evaluation models and stored, and output as a benchmark model library.

[0096] In some embodiments, the safety assessment model building and training module 14 further includes:

[0097] An actual evaluation computing power acquisition unit is used to obtain the actual evaluation computing power of the target scenario and determine the upper limit of resources that can be used for security evaluation;

[0098] An expected allocated computing power calculation unit, used to calculate and determine the expected allocated computing power of each evaluation target in the multi-dimensional evaluation target based on the evaluation weight;

[0099] A benchmark model group selection unit, configured to use the benchmark model library as a selection space and the expected allocated computing power as an upper limit constraint to perform random selection to obtain a plurality of benchmark model groups;

[0100] The safety assessment model generating unit is used to perform integrated training of multiple benchmark model groups based on the historical construction safety assessment records to generate the safety assessment model.

[0101] In some embodiments, the security assessment model generation unit in the security assessment model construction and training module 14 includes:

[0102] An indifferent initial training unit, configured to perform indifferent initial training on the plurality of the benchmark model groups based on the historical construction safety assessment records to obtain a first model set;

[0103] A specialized training unit, configured to use the division result of the historical construction safety assessment record as a training data selection space, and perform specialized training on a plurality of first model groups in the first model set respectively;

[0104] The safety assessment model integration unit is used to integrate the first model groups after multiple specialized training to obtain the safety assessment model.

[0105] In some embodiments, the real-time evaluation and analysis module 15 includes:

[0106] A monitoring sensor inspection and adjustment unit, used to inspect the status of the monitoring sensors of the target scene and make adaptive adjustments according to the evaluation indicator set;

[0107] A collection frequency and parameter setting unit, used to set the collection frequency and collection parameters corresponding to the evaluation indicator set, and activate the monitoring sensor for sensor collection;

[0108] The safety assessment result acquisition unit is used to transmit the sensor collection result to the safety assessment model to obtain the safety assessment result, wherein the safety assessment result at least includes the safety level, unsafe position, and unsafe category.

[0109] Furthermore, the system also includes:

[0110] A principal component analysis unit, used to perform principal component analysis on the evaluation index set, identify key evaluation indexes and redundant evaluation indexes, and output the result of the index analysis;

[0111] The evaluation indicator set simplification element is used to simplify the evaluation indicator set according to the indicator analysis result, remove the redundant evaluation indicators, and retain the key evaluation indicators.

[0112] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0113] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0114] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0115] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0116] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0117] Although preferred embodiments of the present invention have been described, additional changes and modifications may occur to these embodiments once those skilled in the art understand the basic inventive concepts.

[0118] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention belong to the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. A method for safety assessment of underground tunnel construction, characterized in that: The method comprises: Interactive target scenarios to obtain multi-dimensional evaluation targets; Based on the multidimensional evaluation target interactive knowledge base, obtaining an evaluation indicator set, wherein the evaluation indicator set includes a plurality of evaluation indicator groups, and the plurality of evaluation indicator groups correspond one-to-one to the multidimensional evaluation targets; Obtaining historical construction safety assessment records, parsing the historical construction safety assessment records, and configuring the assessment weight of the multi-dimensional assessment target; According to the assessment severity, a safety assessment model is constructed, and the safety assessment model is trained by combining the historical construction safety assessment records with an integrated learning method; Based on the evaluation index set, the construction monitoring information of the target scene is collected in real time, and the construction monitoring information is input into the safety assessment model for evaluation and analysis; Wherein, according to the assessment severity, a safety assessment model is constructed, and the safety assessment model is trained by combining the historical construction safety assessment records with an integrated learning method, before that, it includes: Based on the machine learning model structure, multiple benchmark evaluation models are constructed; the computing power consumption of each benchmark evaluation model is evaluated, and the evaluation results are associated with the multiple benchmark evaluation models and stored, and output as a benchmark model library; According to the assessment severity, a safety assessment model is constructed, and the safety assessment model is trained by combining the historical construction safety assessment records with an integrated learning method, including: The actual evaluation computing power of the target scenario is obtained to determine the upper limit of resources that can be used for safety evaluation; based on the evaluation weight, the expected allocated computing power of each evaluation target in the multi-dimensional evaluation target is calculated and determined; using the benchmark model library as the selection space and the expected allocated computing power as the upper limit constraint, a random selection is performed to obtain multiple benchmark model groups; based on the historical construction safety evaluation records, an integrated training of multiple benchmark model groups is performed to generate the safety evaluation model.

2. A method for safety assessment of underground tunnel construction as claimed in claim 1, characterized in that: Obtaining historical construction safety assessment records, parsing the historical construction safety assessment records, and configuring the assessment severity of the multi-dimensional assessment target, including: Dividing the historical construction safety assessment records based on the multi-dimensional assessment target; Traverse the partitioning results and analyze and obtain the target assessment feature data of each assessment target; The target evaluation feature data is integrated to obtain an evaluation weight coefficient set corresponding to the multi-dimensional evaluation target; Based on the severity coefficient set, the relative assessment severity coefficient of the multidimensional assessment target is calculated and output as the assessment severity.

3. A method for safety assessment of underground tunnel construction as claimed in claim 2, characterized in that: Based on the historical construction safety assessment records, multiple benchmark model groups are integrated and trained, including: Based on the historical construction safety assessment records, a plurality of the benchmark model groups are trained in an indiscriminate manner to obtain a first model set; Using the division results of the historical construction safety assessment records as a training data selection space, performing specialized training on multiple first model groups in the first model set respectively; Integrate the first model groups after multiple specialized training to obtain the security assessment model.

4. A method for safety assessment of underground tunnel construction as claimed in claim 3, characterized in that: Based on the evaluation index set, the construction monitoring information of the target scene is collected in real time, and the construction monitoring information is input into the safety assessment model for evaluation and analysis, including: According to the evaluation indicator set, checking the monitoring sensor status of the target scene and making adaptive adjustments; Setting the collection frequency and collection parameters for the evaluation indicator set, and activating the monitoring sensor for sensor collection; The sensing acquisition result is transmitted to the safety assessment model to obtain a safety assessment result, wherein the safety assessment result at least includes a safety level, an unsafe location, and an unsafe category.

5. The underground tunnel construction safety assessment method according to claim 1, characterized in that: Based on the multidimensional evaluation target interactive knowledge base, an evaluation indicator set is obtained, wherein the evaluation indicator set includes multiple evaluation indicator groups, and the multiple evaluation indicator groups correspond to the multidimensional evaluation targets one by one, and then, it also includes: Performing principal component analysis on the evaluation indicator set, identifying key evaluation indicators and redundant evaluation indicators, and outputting the indicator analysis results; The evaluation indicator set is simplified according to the indicator analysis result, the redundant evaluation indicators are removed, and the key evaluation indicators are retained.

6. An underground tunnel construction safety assessment system, characterized in that: The system is used to execute the underground tunnel construction safety assessment method according to any one of claims 1 to 5, and the system comprises: The target scene interaction module is used to interact with the target scene and obtain multi-dimensional evaluation targets; An evaluation indicator set acquisition module, used to acquire an evaluation indicator set based on the multidimensional evaluation target interactive knowledge base, wherein the evaluation indicator set includes a plurality of evaluation indicator groups, and the plurality of evaluation indicator groups correspond one-to-one to the multidimensional evaluation targets; An assessment weight configuration module, used to obtain historical construction safety assessment records, parse the historical construction safety assessment records, and configure the assessment weight of the multi-dimensional assessment target; A safety assessment model construction and training module, used to construct a safety assessment model according to the assessment severity, and train the safety assessment model in combination with the historical construction safety assessment records and an integrated learning method; The real-time evaluation and analysis module is used to collect the construction monitoring information of the target scene in real time based on the evaluation index set, and input the construction monitoring information into the safety assessment model for evaluation and analysis.

Images