Underground engineering construction scheme dynamic optimization method, system and computer readable storage medium

Abstract

The application discloses a kind of underground engineering construction scheme dynamic optimization method, system and computer readable storage medium, including integration geometry information, geology parameter space distribution, supporting structure parameter and history monitoring data constructs initial digital twin model;In construction process, the posterior distribution of surrounding rock mechanics parameters is dynamically revised based on Bayesian inversion method using real-time monitoring data;Define trigger threshold and comprehensive trigger index, when inversion parameter deviation exceeds threshold, automatically trigger scheme dynamic adjustment;According to the weight of the comprehensive trigger index dynamic adjustment geology adaptability index, make the scheme optimization result more close to the current actual geological conditions;While outputting the adjusted scheme, automatically generate transition path and standardized construction instruction and issue to field management system, form "monitoring→inversion→adjustment→construction" full-automatic closed loop.The application improves the construction safety and scheme adaptability of tunnel reconstruction and expansion project.

Description

Technical Field

[0001] This invention relates to the field of underground engineering construction technology, specifically to a method, system, and computer-readable storage medium for dynamic optimization of underground engineering construction schemes. Background Technology

[0002] Underground engineering faces significant challenges, including complex geological conditions and high uncertainty regarding the state of existing structures. Currently, engineering practice primarily relies on geological data obtained during the exploration phase to pre-determine construction plans, supplemented by monitoring and early warning systems during construction. However, existing technologies suffer from the following systemic shortcomings:

[0003] The disconnect between static decision-making and dynamic construction: Traditional methods determine the construction plan once the geological survey is completed. However, the actual surrounding rock parameters revealed during underground engineering construction often deviate significantly from the survey report (such as elastic modulus, cohesion, and internal friction angle). When the actual geological conditions are worse than the design expectations, the initially determined excavation methods and support parameters may no longer be applicable. The existing system lacks an automatic triggering mechanism for dynamically adjusting the plan based on real-time monitoring data, leading to increased construction safety risks or project delays.

[0004] Predictive and early warning systems lack a closed-loop decision-making process: In recent years, some studies have introduced digital twin technology into underground engineering, using monitoring data and numerical simulations to predict deformation and provide early warnings of risks. However, existing digital twin systems, after issuing an early warning, still rely on human experience to decide how to adjust the plan (such as whether to change construction methods, increase anchor density, or add temporary supports), and cannot automatically generate a transition path from the current construction state to the optimal solution or structured construction instructions. This "early warning—human decision-making—manual execution" model results in a delayed response (usually several hours to several days), making it difficult to effectively control the accumulation of risks.

[0005] The disconnect between geological parameter inversion and scheme selection: Although a few studies have involved the inversion of surrounding rock parameters based on monitoring data (such as using Bayesian methods or neural networks), the inversion results are only used to correct parameter values ​​in the numerical model and do not form a closed-loop coupling with the scheme selection model. Specifically, the geological parameters obtained from the inversion do not directly participate in the dynamic scoring of candidate schemes, the weight of the geological suitability index cannot be adaptively adjusted according to the actual parameter deviation, and the selection results cannot be automatically converted into construction instructions. Therefore, existing technologies cannot achieve integrated, fully automated operation of "monitoring → inversion → adjustment → construction".

[0006] In summary, there is an urgent need for a dynamic optimization method and system for underground engineering construction schemes that can dynamically correct geological parameters based on real-time monitoring data, automatically trigger scheme adjustments, and directly generate construction instructions based on the adjusted schemes. This would solve the problems of static decision-making, delayed early warning response, and separation of inversion and optimization in existing technologies, thereby improving the construction safety and scheme adaptability of tunnel reconstruction and expansion projects. Summary of the Invention

[0007] The purpose of this invention is to provide a method, system, and computer-readable storage medium for dynamic optimization of underground engineering construction schemes, so as to at least solve the problem of the lack of an automatic triggering mechanism for dynamically adjusting schemes based on real-time monitoring data.

[0008] To address the aforementioned technical problems, in a first aspect, the present invention provides a method for dynamic optimization of underground engineering construction schemes, comprising the following steps:

[0009] S1: Obtain basic data of underground engineering, determine initial geological parameters based on the basic data, construct a spatial distribution model of the initial geological parameters using a random field model, and integrate them into an initial digital twin model;

[0010] S2: Collect real-time monitoring data of underground engineering, use the statistical characteristics of the initial geological parameters as the prior distribution, construct the likelihood function based on the difference between the real-time monitoring data and the finite element prediction value, establish a Bayesian inversion model, solve the posterior distribution of the current geological parameters, obtain the optimal estimate of the current geological parameters, and update it to the digital twin model.

[0011] S3: Calculate the deviation vector between the current geological parameters and the initial geological parameters and its corresponding comprehensive trigger index, and compare the comprehensive trigger index with a preset threshold; when the comprehensive trigger index exceeds the preset threshold or any deviation value in the deviation vector exceeds the corresponding exclusive threshold, the dynamic adjustment process of the trigger scheme is activated.

[0012] S4: Obtain multiple alternative solutions and re-evaluate them using the TOPSIS method with dynamic weights. This includes: calculating each evaluation index based on the inverted geological parameters, constructing a decision matrix and performing forwarding and normalization, and calculating the initial weights of each index using the entropy weight method; then dynamically adjusting the weights of each index based on the comprehensive triggering index, and calculating the closeness of each solution based on the adjusted weights; the solution with the highest closeness is output as the optimal solution.

[0013] S5: Identify the current construction status, generate a transition path from the current construction status to the optimal solution, convert the transition path into structured construction instructions and issue them to the site management system, and simultaneously feed the site construction data back to the digital twin model in real time, forming a complete closed-loop iteration.

[0014] Furthermore, step S1 specifically includes:

[0015] S11: Basic data for the underground engineering project; the tunnel basic data includes BIM model, geological survey data, design parameters and historical monitoring data;

[0016] S12: Extract geometric information, support structure parameters, and historical monitoring data sequences from the tunnel foundation data;

[0017] S13: Construct a spatial distribution model for initial geological parameters. Based on the statistical characteristics of geological exploration data, a random field model is used to describe the spatial variability of geological parameters along the longitudinal direction of the tunnel. The random field model uses an exponential correlation function, and the correlation distance is determined based on geological statistical experience in tunnel engineering.

[0018] S14: Integrate the geometric information, support structure parameters, historical monitoring data sequences, and initial geological parameter spatial distribution into an initial digital twin model.

[0019] Furthermore, step S2 specifically includes:

[0020] S21: Collect real-time monitoring data of the tunnel at fixed time intervals during construction;

[0021] S22: Construct a Bayesian inversion model, using the statistical characteristics of geological exploration data as the prior distribution, and construct a likelihood function based on the difference between real-time monitoring data and finite element prediction values; the posterior distribution of the Bayesian inversion model is proportional to the product of the prior distribution and the likelihood function.

[0022] S23: The Markov chain Monte Carlo method is used to solve the posterior distribution and obtain the optimal estimate of the geological parameters at the current time.

[0023] S24: Update the inverted geological parameters to the digital twin model and iterate the digital twin model in real time.

[0024] Furthermore, the likelihood function in step S22 is constructed based on the assumption that the monitoring error follows a normal distribution, and the likelihood function is expressed as:

[0025] ;

[0026] in, For real-time monitoring data These are finite element prediction values. M represents the standard deviation of the monitoring error, and M represents the number of data types being monitored.

[0027] Furthermore, in step S23, solving for the posterior distribution using the Markov chain Monte Carlo method specifically includes:

[0028] S231: Initialize parameters ;

[0029] S232: Set the iteration count R, burn-in period From the proposed distribution using a Gaussian distribution Generate candidate samples ,in ;

[0030] S233: Calculate the probability of acceptance:

[0031] S234: Based on probability accept ,otherwise ;

[0032] S235: Discard the results of the first D iterations and take the mean of the remaining samples as the optimal estimate. :

[0033] .

[0034] Furthermore, the deviation vector in step S3 is:

[0035] ;

[0036] ;

[0037] in, The deviation vector is n; n is the number of geological parameters. For the first The relative deviation of each geological parameter; For the first Inversion values ​​of geological parameters; For the first Initial values ​​for each geological parameter;

[0038] The formula for calculating the comprehensive trigger index is:

[0039] ;

[0040] or,

[0041]

[0042] in, To form a comprehensive trigger index, Indicates taking the deviation vector The maximum value of each element in the list. For the first The weights of each parameter, and satisfying .

[0043] Further, step S4 includes:

[0044] S41: Calculate the evaluation indicators for each scheme based on the inverted geological parameters;

[0045] S42: Construct a decision matrix, positively process cost-type indicators, and keep the original values ​​of benefit-type indicators; and normalize all indicators using Euclidean norm;

[0046] S43: The initial weights of each index are calculated using the entropy weight method;

[0047] S44: Then, based on the comprehensive trigger index, the weights of each indicator are dynamically adjusted. The adjustment formula is as follows:

[0048] ;

[0049] ;

[0050] in, For the first The indicator is based on the adjusted composite trigger index. The weight of each indicator, To remove indicators Other indicators, and

[0051] To be respectively with and The initial weights of the corresponding indicators, Where N is the amplification factor, and N is the total number of indicators;

[0052] S45: Determine the positive and negative ideal solutions, and calculate the weighted Euclidean distance from each solution to the positive and negative ideal solutions based on the weights of each index.

[0053] S46: Calculate the proximity of each solution based on the weighted Euclidean distance, and output the solution with the highest proximity as the optimal solution; the formula for calculating the proximity is:

[0054] ;

[0055] in, Let be the weighted Euclidean distance from the i-th solution to the positive ideal solution. Let be the weighted Euclidean distance from the i-th solution to the negative and positive ideal solutions.

[0056] Furthermore, in step S5, the transition path is generated as follows:

[0057] If the current work method differs from the optimal solution, the transition path includes work stoppage, reinforcement, and work method switching procedures.

[0058] If the current support parameters differ from those of the optimal solution, the transition path includes sequentially executed transition steps, process parameters, and expected duration.

[0059] The construction instructions include instruction type, triggering reason, current status, adjusted plan, transition process, process parameters and risk control elements.

[0060] Secondly, the present invention provides a dynamic optimization system for underground engineering construction schemes, comprising:

[0061] The model layer includes a digital twin model management module, a Bayesian inversion calculation module, and a finite element calculation engine. The digital twin model management module is used to store, update, and retrieve geometric information, spatial distribution of geological parameters, support structure parameters, and monitoring data. The Bayesian inversion calculation module is used to construct a Bayesian inversion model, solve for the posterior distribution of geological parameters, and output the optimal estimate. The finite element calculation engine is used to provide finite element prediction values ​​for Bayesian inversion and to provide surrounding rock stability simulation results for scheme optimization.

[0062] The framework layer includes a data acquisition and preprocessing module, a trigger judgment module, a scheme optimization engine, and an instruction generation module. The data acquisition and preprocessing module is used to acquire and preprocess real-time monitoring data. The trigger judgment module is used to calculate the deviation vector and comprehensive trigger index of geological parameters to complete the scheme adjustment trigger judgment. The scheme optimization engine is used to realize dynamic re-optimization of schemes based on the improved TOPSIS method and output the optimal scheme. The instruction generation module is used to generate transition paths and convert them into structured construction instructions.

[0063] The application layer includes a data acquisition module, a dynamic optimization module, and a construction implementation module. The data acquisition module is used to acquire basic data and transmit it to the model layer. The dynamic optimization module is used to call the functions of the framework layer and the model layer to realize dynamic optimization of the entire process of the solution. The construction implementation module is used to issue construction instructions to the site management system and collect construction site data to feed back to the model layer.

[0064] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps mentioned in the first aspect.

[0065] The beneficial effects of this invention are as follows: By constructing an initial digital twin model to integrate geometric information, spatial distribution of geological parameters, support structure parameters, and historical monitoring data, the posterior distribution of surrounding rock mechanical parameters is dynamically corrected using real-time monitoring data based on the Bayesian inversion method during construction, solving the problem of scheme failure caused by deviation between exploration data and actual geological conditions; by defining trigger thresholds and comprehensive triggering indices, the scheme is automatically dynamically adjusted when the deviation of inversion parameters exceeds the threshold, significantly shortening the response time compared to manual decision-making and avoiding risk accumulation; the weight of geological adaptability indicators is dynamically adjusted according to the comprehensive triggering index, making the scheme optimization result more consistent with the current actual geological conditions; while outputting the adjusted scheme, a transition path and standardized construction instructions are automatically generated and sent to the site management system, forming a fully automatic closed loop of "monitoring → inversion → adjustment → construction", significantly improving the construction safety and scheme adaptability of tunnel reconstruction and expansion projects. Attached Figure Description

[0066] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, use the same reference numerals to denote the same or similar parts. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0067] Figure 1 This is a flowchart of a dynamic optimization method for underground engineering construction schemes according to an embodiment of the present invention;

[0068] Figure 2 This is a block diagram of the overall architecture of a dynamic optimization system for underground engineering construction schemes according to an embodiment of the present invention. Detailed Implementation

[0069] Firstly, such as Figure 1 As shown, this invention discloses a dynamic optimization method for underground engineering construction schemes, comprising the following steps:

[0070] S1: Obtain basic data of underground engineering, determine initial geological parameters based on the basic data, construct a spatial distribution model of the initial geological parameters using a random field model, and integrate them into an initial digital twin model;

[0071] S2: Collect real-time monitoring data of underground engineering, use the statistical characteristics of the initial geological parameters as the prior distribution, construct the likelihood function based on the difference between the real-time monitoring data and the finite element prediction value, establish a Bayesian inversion model, solve the posterior distribution of the current geological parameters, obtain the optimal estimate of the current geological parameters, and update it to the digital twin model.

[0072] S3: Calculate the deviation vector between the current geological parameters and the initial geological parameters and its corresponding comprehensive trigger index, and compare the comprehensive trigger index with a preset threshold; when the comprehensive trigger index exceeds the preset threshold or any deviation value in the deviation vector exceeds the corresponding exclusive threshold, the dynamic adjustment process of the scheme is triggered, and step S4 is executed.

[0073] S4: Obtain multiple alternative solutions and re-evaluate them using the TOPSIS method with dynamic weights. This includes: calculating each evaluation index based on the inverted geological parameters, constructing a decision matrix and performing forwarding and normalization, and calculating the initial weights of each index using the entropy weight method; then dynamically adjusting the weights of each index based on the comprehensive triggering index, and calculating the closeness of each solution based on the adjusted weights; the solution with the highest closeness is output as the optimal solution.

[0074] S5: Identify the current construction status, generate a transition path from the current construction status to the optimal solution, convert the transition path into structured construction instructions and issue them to the site management system, and simultaneously feed the site construction data back to the digital twin model in real time, forming a complete closed-loop iteration.

[0075] According to one embodiment of this application, step S1 specifically includes:

[0076] S11: Basic data for the approach tunnel; the basic data for the tunnel includes the BIM model (IFC format), geological survey data (PDF format), design parameters, and historical monitoring data within a specified time range;

[0077] S12: Extract geometric information, support structure parameters, and historical monitoring data sequences from the tunnel foundation data; geometric information includes the tunnel axis, cross-sectional profile, support structure layout, and geological geometry.

[0078] S13: Construct a spatial distribution model for initial geological parameters. Based on the statistical characteristics of geological exploration data, a random field model is used to describe the spatial variability of geological parameters along the tunnel longitudinal direction. The random field model uses an exponential correlation function, and the correlation distance is determined based on geological statistical experience in tunnel engineering.

[0079] S14: Extract geometric information Support structure parameters Historical monitoring data sequences Spatial distribution of initial geological parameters Integrated into the initial digital twin model

[0080] .

[0081] According to one embodiment of this application, step S2 specifically includes:

[0082] S21: Collect real-time monitoring data of the tunnel at fixed time intervals during construction;

[0083] S22: Construct a Bayesian inversion model, using the statistical characteristics of geological exploration data as the prior distribution, and construct a likelihood function based on the difference between real-time monitoring data and finite element prediction values; the posterior distribution of the Bayesian inversion model is proportional to the product of the prior distribution and the likelihood function.

[0084] S23: The Markov chain Monte Carlo method is used to solve the posterior distribution and obtain the optimal estimate of the geological parameters at the current time.

[0085] S24: Update the inverted geological parameters to the digital twin model and iterate the digital twin model in real time.

[0086] According to one embodiment of this application, the likelihood function in step S22 is constructed based on the assumption that the monitoring error follows a normal distribution, and the likelihood function is expressed as:

[0087] ;

[0088] in, For real-time monitoring data These are finite element prediction values. M represents the standard deviation of the monitoring error, and M represents the number of data types being monitored.

[0089] According to one embodiment of this application, in step S23, the Metropolis-Hastings algorithm of the Markov chain Monte Carlo method is used to solve for the posterior distribution, specifically including:

[0090] S231: Initialize parameters ;

[0091] S232: Set the number of iterations Burning period From the proposed distribution using a Gaussian distribution Generate candidate samples ,in ;

[0092] S233: Calculate the probability of acceptance: ;

[0093] S234: Based on probability accept ,otherwise ;

[0094] S235: Discard the results of the first D iterations and take the mean of the remaining samples as the optimal estimate. :

[0095] .

[0096] This application constructs an initial digital twin model, integrating geometric information, spatial distribution of geological parameters, support structure parameters, and historical monitoring data. During construction, based on the Bayesian inversion method, the posterior distribution of surrounding rock mechanical parameters is dynamically corrected using real-time monitoring data (solved using the Markov chain Monte Carlo method), and the inverted parameters are updated to the digital twin model. Compared to traditional methods that rely solely on static parameters from the exploration phase, this invention continuously calibrates the geological model as construction progresses, ensuring that subsequent optimal solutions are always based on the current actual geological conditions, effectively avoiding decision-making errors caused by variability in surrounding rock parameters.

[0097] According to one embodiment of this application, the deviation vector in step S3 is:

[0098] ;

[0099] ;

[0100] in, Let n be the deviation vector, which is the vector consisting of the relative deviations between the inverted geological parameters and the initial geological parameters; n is the number of geological parameters. For the first The relative deviation of a geological parameter is the ratio of the absolute difference of the parameter to the initial geological parameter. For the first Inversion values ​​of geological parameters; For the first Initial values ​​for each geological parameter;

[0101] The formula for calculating the comprehensive trigger index is:

[0102] Maximum value form: ;

[0103] or,

[0104] Weighted sum form: ;

[0105] in, To form a comprehensive trigger index, Indicates taking the deviation vector The maximum value of each element in the table is used to calculate the comprehensive trigger index in the form of the maximum value. It can be applied to scenarios where an early warning is triggered if any indicator deviates beyond the limit; For the first The weights of each parameter are determined based on engineering experience using the analytic hierarchy process (AHP) or expert scoring, and satisfy the following conditions: The comprehensive trigger index is calculated through a weighted sum. It can be applied to scenarios that comprehensively assess the cumulative impact of multiple indicators.

[0106] This application defines a parameter deviation vector and a comprehensive triggering index (taking the maximum value or weighted sum of the deviations of each parameter), and sets a specific threshold. When the deviation between the inverted geological parameters and the initial geological parameters exceeds the preset threshold, the system automatically triggers dynamic re-optimization of the scheme without human intervention. This mechanism significantly shortens the response time compared to manual decision-making, enabling timely adjustments to the construction scheme in the early stages of geological condition deterioration, avoiding risk accumulation, and significantly improving construction safety.

[0107] According to one embodiment of this application, step S4 includes:

[0108] S41: Calculate the evaluation indicators for each scheme based on the inverted geological parameters;

[0109] S42: Constructing the Decision Matrix (where m is the number of options and k is the number of indicators), positively process the cost-related indicators: Keep the original values ​​for efficiency indicators. And the Euclidean norm is used to normalize all indicators: ;

[0110] S43: The initial weights of each indicator are calculated using the entropy weight method. The calculation includes:

[0111] 1) Calculate the weight of the indicators: (when season );

[0112] 2) Calculate information entropy: ;

[0113] 3) Calculate the entropy weight: ;

[0114] in, For the first Entropy weights (initial weights) of each indicator;

[0115] S44: Then, based on the comprehensive trigger index, the weights of each indicator are dynamically adjusted. The adjustment formula is as follows:

[0116] ;

[0117] ;

[0118] in, For the first The indicator is based on the adjusted composite trigger index. The weight of each indicator, To remove indicators Other indicators, and To be respectively with and The entropy weight of the corresponding indicator The amplification factor is 0.3 to 0.8 (the value ranges from 0.3 to 0.8, determined according to the importance level of the project), and N is the number of indicators.

[0119] S45: Determine the positive ideal solution and the negative ideal solution, positive ideal solution Negative ideal solution The weighted Euclidean distance from each solution to the positive and negative ideal solutions is calculated based on the weights of each indicator. The calculation formula is as follows:

[0120] ;

[0121] ;

[0122] S46: Calculate the proximity of each solution based on the weighted Euclidean distance, and output the solution with the highest proximity as the optimal solution; the formula for calculating the proximity is:

[0123] ;

[0124] in Let be the weighted Euclidean distance from the i-th solution to the positive ideal solution. Let be the weighted Euclidean distance from the i-th solution to the negative and positive ideal solutions; Let be the similarity of the i-th scheme. The larger the value, the better the solution.

[0125] This application employs an improved TOPSIS method during the scheme re-optimization stage. In this method, the weight of the geological adaptability index is dynamically adjusted based on a comprehensive trigger index, while the weights of other indices are normalized proportionally. When the actual geological parameters deviate significantly, the weight of the geological adaptability index automatically increases, thus ensuring that the scheme optimization results place greater emphasis on the degree of matching with the current geological conditions (e.g., prioritizing stronger support schemes). Compared to the existing approach of separating inversion and optimization with fixed weights, this achieves adaptive optimization where "the greater the deviation, the more the decision emphasizes geological adaptability," improving the scientific rigor and adaptability of the decision-making process.

[0126] According to one embodiment of this application, step S5 specifically includes:

[0127] S51: Identify the current construction status, including the length of excavated section, the length of supported section, and the parameters of implemented support;

[0128] S52: Generate a transition path from the current state to the optimal solution. The transition path is generated based on the differences in construction methods and support parameters between the current construction state and the optimal solution. If the construction methods are different, the transition path includes work stoppage, reinforcement, and construction method switching procedures. If only the support parameters are different, the transition path includes parameter adjustment procedures. The transition path also includes the transition procedures to be executed in sequence, process parameters, and the expected duration.

[0129] S53: Based on a predefined template library, the transition path is converted into a structured construction instruction. The construction instruction includes the instruction type, triggering reason, current status, adjusted plan, transition process, process parameters and risk control elements.

[0130] S54: The instructions are sent to the construction site management system, and the construction data is fed back to the digital twin model in real time. After the model status is updated, a continuous iterative closed-loop optimization process is formed, thereby completing the complete closed loop of monitoring, inversion, adjustment and construction.

[0131] After outputting the optimal solution, this application further identifies the current construction status (excavated length, supported section length, current construction method, etc.), automatically generates a transition path including the sequence of procedures, process parameters, and estimated duration, and converts it into standardized construction instructions based on a predefined template library (including instruction type, triggering reason, adjusted solution, risk control elements, etc.). Finally, it is issued to the site management system via API. Construction execution data is fed back to the digital twin model in real time, completing a full closed loop. Compared with the shortcomings of existing digital twin systems that only provide early warnings and require manual decision-making, this invention achieves end-to-end automation from monitoring data to construction instructions, significantly improving construction response efficiency and engineering adaptability.

[0132] Secondly, such as Figure 2 As shown, the present invention provides a dynamic optimization system for underground engineering construction schemes, comprising:

[0133] The model layer includes a digital twin model management module, a Bayesian inversion calculation module, and a finite element calculation engine. The digital twin model management module is used to store, update, and retrieve geometric information, spatial distribution of geological parameters, support structure parameters, and monitoring data. The Bayesian inversion calculation module is used to construct a Bayesian inversion model, solve for the posterior distribution of geological parameters, and output the optimal estimate. The finite element calculation engine is used to provide finite element prediction values ​​for Bayesian inversion and to provide surrounding rock stability simulation results for scheme optimization.

[0134] The framework layer includes a data acquisition and preprocessing module, a trigger judgment module, a scheme optimization engine, and an instruction generation module. The data acquisition and preprocessing module is used to acquire and preprocess real-time monitoring data. The trigger judgment module is used to calculate the deviation vector and comprehensive trigger index of geological parameters to complete the scheme adjustment trigger judgment. The scheme optimization engine is used to realize dynamic re-optimization of schemes based on the improved TOPSIS method and output the optimal scheme. The instruction generation module is used to generate transition paths and convert them into structured construction instructions.

[0135] The application layer includes a data acquisition module, a dynamic optimization module, and a construction implementation module. The data acquisition module is used to acquire basic data and transmit it to the model layer. The dynamic optimization module is used to call the functions of the framework layer and the model layer to realize dynamic optimization of the entire process of the solution. The construction implementation module is used to issue construction instructions to the site management system and collect construction site data to feed back to the model layer.

[0136] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps mentioned in the first aspect.

[0137] The embodiments of the technical solution of the present invention will be described in detail below with reference to the accompanying drawings. This embodiment takes a highway tunnel reconstruction and expansion project as an example to illustrate the present invention in detail. The highway tunnel is a shallow-buried tunnel with a surrounding rock grade of IV. The reconstruction and expansion target is to widen the two-way four-lane roadway to a two-way six-lane roadway. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and are therefore only examples, and should not be used to limit the scope of protection of the present invention.

[0138] Example 1

[0139] This embodiment will be described in conjunction with Figures 1 and 2:

[0140] Step S1: Initial digital twin model construction (corresponding to the method flow in Figure 1, and the application layer data import module and model layer digital twin model management module in Figure 2)

[0141] S11: Geological survey report (PDF format), BIM model (IFC format), support design parameters and historical monitoring data for the past 5 years for the tunnel;

[0142] S12: Extract geometric information, including the control points of the three-dimensional spatial curve of the tunnel axis, cross-sectional profile parameters, support structure layout parameters and stratum geometric information, as shown in Table 2.

[0143] S13: Construct a spatial distribution model for the initial geological parameters. The statistical characteristics of the initial geological parameters are extracted as shown in Table 1. An exponential correlation function is used to describe the spatial variability of the parameters. The correlation function formula is:

[0144] ;

[0145] in, The distance between the two points is the mileage difference, and the relevant distance is taken as... (Determined based on geological statistics experience of similar tunnel projects);

[0146] Random Field A Gaussian random field with zero mean and unit variance satisfies:

[0147] ;

[0148] The tunnel is laid longitudinally at equal intervals. Discretized The initial geological parameters at each point are distributed along the longitudinal direction of the tunnel as follows: The discretized initial geological parameter vector is obtained:

[0149] ;

[0150] S14: Extract the support structure parameters as shown in Table 3, extract historical monitoring data and construct a time series:

[0151] ;

[0152] in, For monitoring time, For the settlement of the vault, To facilitate the convergence of surrounding areas, For the surrounding rock stress, For the support structure to bear the force.

[0153] S15: Integration into the initial digital twin model:

[0154] ;

[0155] in, For geometric information, The initial spatial distribution of geological parameters, For the parameters of the support structure, This is historical monitoring data.

[0156] Table 1 Statistical Characteristics of Initial Geological Parameters

[0157]

[0158] Note: Mean Statistical values ​​taken from the site survey report; standard deviation Based on the determination of the coefficient of variation of similar engineering parameters, the coefficient of variation... .

[0159] Table 2 Geometric Information Extraction Table

[0160]

[0161] Table 3 Support Structure Parameter Table

[0162]

[0163] Step S2: Dynamic correction of geological parameters based on Bayesian inversion (corresponding to the Bayesian inversion calculation module and finite element calculation engine in the model layer and the data acquisition and preprocessing module in the framework layer in Figure 1, and the method flow in Figure 2).

[0164] Construction up to chainage At that time, the monitoring data on crown settlement and surrounding convergence showed significant deviations from the initial finite element predictions, prompting the initiation of dynamic correction of geological parameters:

[0165] The data from the dome settlement monitoring show the cumulative settlement. The surrounding area converges. , compared with the initial finite element prediction value ( There is a significant deviation.

[0166] S21: Collect real-time monitoring data and form a real-time monitoring data vector: ,in This represents the measured value of the arch settlement. The measured values ​​represent the perimeter convergence values; the initial finite element prediction values ​​for the crown settlement and perimeter convergence are respectively... ;

[0167] S22: Construct a Bayesian inversion model where the posterior distribution is proportional to the product of the prior distribution and the likelihood function.

[0168] ;

[0169] The prior distribution assumes a Gaussian distribution:

[0170] ;

[0171] The mean Covariance Determined based on the statistical characteristics in Table 1;

[0172] The likelihood function is constructed based on the assumption that the monitoring error follows a normal distribution:

[0173] ;

[0174] in, These are measured values. To predict displacement, the standard deviation of the monitoring error is...

[0175] The calculation yields: , ;

[0176] S23: Solving for the posterior distribution using the Metropolis-Hastings algorithm with the Markov chain Monte Carlo method:

[0177] 1) Initialize parameters ;

[0178] 2) Set the number of iterations R=10000, and the burn-in period From the proposed distribution using a Gaussian distribution Generate candidate samples ,in ;

[0179] 3) Calculate the acceptance probability: ;

[0180] 4) Based on probability accept ,otherwise ;

[0181] 5) Discard the results of the first 2000 iterations and take the mean of the remaining samples as the optimal estimate:

[0182] ;

[0183] Optimal estimates of geological parameters after inversion: , ;

[0184] S24: Inverted geological parameters Update the digital twin model to obtain the updated model: ,in This is a sequence of historical monitoring data after integrating real-time monitoring data.

[0185] Step S3: Dynamic adjustment of the scheme trigger (corresponding to the framework layer trigger judgment module in Figure 1, and the method flow in Figure 2)

[0186] S31: Calculate the deviation vector, which is a vector composed of relative deviations, according to the formula.

[0187] The calculation yields:

[0188]

[0189] ;

[0190] S32: Define the comprehensive trigger index. This embodiment uses the maximum value form:

[0191]

[0192] If a weighted sum is used, the weight vector... (Determined by the Analytic Hierarchy Process), then: ;

[0193] S33: Set specific trigger thresholds for key parameters as shown in Table 4, including elastic modulus deviation. This triggers the dynamic re-optimization process of the solution.

[0194] Table 4 Trigger Threshold Table

[0195]

[0196] Step S4: Dynamic re-optimization of the scheme based on inversion parameters (corresponding to the scheme optimization engine in the framework layer of Figure 1, and the method flow in Figure 2).

[0197] S41: The alternative solution library contains 3 solutions:

[0198] Option A: Step-by-step excavation, with anchor bolt spacing of 1.2m;

[0199] Option B: CD method excavation, anchor bolt spacing 0.8m, with the addition of a temporary invert arch;

[0200] Option C: CRD method excavation, anchor spacing 0.6m, temporary support densification.

[0201] S42: Based on inversion parameters The evaluation indicators for each scheme are calculated, as shown in Table 5. Geological suitability is calculated using the following formula:

[0202] ;

[0203] in, For the currently inverted elastic modulus, cohesion, and internal friction angle, The corresponding initial value is used; the coefficient 5 is only to normalize the benchmark value and make it easier for engineers to understand intuitively.

[0204] S43: Constructing the decision matrix Corresponding schemes A, B, and C, There are 7 corresponding indicators; for cost-related indicators (construction cost, construction period, traffic disruption), the inverse method is used for positive transformation; for benefit-related indicators (geological compatibility, construction safety risk, structural durability, environmental impact), the original values ​​are maintained. The matrix after positive transformation is as follows: ;

[0205] Using Euclidean norm pairs Normalize:

[0206] ;

[0207] After normalization, it satisfies

[0208] ;

[0209] S44: The objective weights are calculated using the entropy weight method, and the entropy weight vector is obtained by calculating the following steps:

[0210] ;

[0211] S45: Dynamically adjust the weight of geological compatibility index; the project importance level is Level 1, and an amplification factor is applied. Adjust according to the formula:

[0212] ;

[0213] The weights of other indicators are adjusted proportionally:

[0214] ;

[0215] Adjusted weights The weights sum to approximately 1;

[0216] S46: Determine the positive and negative ideal solutions: ;

[0217] Calculate the weighted Euclidean distance:

[0218] ,

[0219] ;

[0220] Calculate proximity: get Sort by proximity in descending order, and output the optimal solution as solution B.

[0221] Table 5 Comparison of Evaluation Indicators for Alternative Solutions

[0222]

[0223] Step S5: Adjust the instruction generation and construction closed loop (corresponding to the instruction generation module of the framework layer in Figure 1, the construction implementation module of the application layer, and the method flow in Figure 2).

[0224] S51: Identify current construction status: Excavated length Length of supported section The current construction method is the step method, and the current anchor spacing is 1.2m;

[0225] S52: Generate a transition path. Because the optimal solution differs from the current construction method, the transition path includes work stoppage, reinforcement, and construction method switching procedures, specifically:

[0226] 1) Cease the current bench excavation operation;

[0227] 2) Construct a temporary invert arch (C30 concrete, 300mm thick) in the area from K1+185 to K1+200 behind the tunnel face, with an estimated duration of 24 hours;

[0228] 3) For the K1+080~K1+200 section, the spacing of the anchor bolts in the reinforcement system will be reduced from 1.2m to 0.8m, with an estimated duration of 48 hours;

[0229] 4) Switch to CD method excavation, advance the left pilot tunnel 3m first, then the right pilot tunnel follows, and advance 0.8m in a cycle.

[0230] S53: Based on the predefined template library, standardized construction instructions are generated as shown in Table 6;

[0231] S54: Construction instructions are issued to the construction site management system via API. Construction personnel execute the instructions and simultaneously feed back the construction progress and on-site monitoring data to the digital twin model management module at the model layer in real time to update the digital twin model, forming a complete closed loop of "monitoring → inversion → adjustment → construction → monitoring".

[0232] Table 6 Examples of Construction Instructions

[0233]

[0234] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A dynamic optimization method for underground engineering construction schemes, characterized in that, Includes the following steps: S1: Obtain basic data of underground engineering, determine initial geological parameters based on the basic data, construct a spatial distribution model of the initial geological parameters using a random field model, and integrate them into an initial digital twin model; S2: Collect real-time monitoring data of underground engineering, use the statistical characteristics of the initial geological parameters as the prior distribution, construct the likelihood function based on the difference between the real-time monitoring data and the finite element prediction value, establish a Bayesian inversion model, solve the posterior distribution of the current geological parameters, obtain the optimal estimate of the current geological parameters, and update it to the digital twin model. S3: Calculate the deviation vector between the current geological parameters and the initial geological parameters and its corresponding comprehensive triggering index, and compare the comprehensive triggering index with a preset threshold; when the comprehensive triggering index exceeds the preset threshold or any deviation value in the deviation vector exceeds the corresponding exclusive threshold, the dynamic adjustment process of the scheme is triggered, and step S4 is executed. S4: Obtain multiple alternative solutions and re-evaluate the alternative solutions using the TOPSIS method combined with dynamic weights, including: calculating each evaluation index based on the inverted geological parameters, constructing a decision matrix and completing forwarding and normalization, and calculating the initial weights of each index using the entropy weight method. Then, based on the comprehensive trigger index, the weights of each indicator are dynamically adjusted, and the closeness of each scheme is calculated based on the adjusted weights; the scheme with the highest closeness is output as the optimal scheme. S5: Identify the current construction status, generate a transition path from the current construction status to the optimal solution, convert the transition path into structured construction instructions and issue them to the site management system, and simultaneously feed the site construction data back to the digital twin model in real time, forming a complete closed-loop iteration.

2. The dynamic optimization method for underground engineering construction schemes according to claim 1, characterized in that, Step S1 specifically includes: S11: Basic data for the underground engineering project; the tunnel basic data includes BIM model, geological survey data, design parameters and historical monitoring data; S12: Extract geometric information, support structure parameters, and historical monitoring data sequences from the tunnel foundation data; S13: Construct a spatial distribution model for initial geological parameters. Based on the statistical characteristics of geological exploration data, a random field model is used to describe the spatial variability of geological parameters along the longitudinal direction of the tunnel. The random field model uses an exponential correlation function, and the correlation distance is determined based on geological statistical experience in tunnel engineering. S14: Integrate the geometric information, support structure parameters, historical monitoring data sequences, and initial geological parameter spatial distribution into an initial digital twin model.

3. The dynamic optimization method for underground engineering construction schemes according to claim 1, characterized in that, Step S2 specifically includes: S21: Collect real-time monitoring data of the tunnel at fixed time intervals during construction; S22: Construct a Bayesian inversion model, using the statistical characteristics of geological exploration data as the prior distribution, and construct a likelihood function based on the difference between real-time monitoring data and finite element prediction values; the posterior distribution of the Bayesian inversion model is proportional to the product of the prior distribution and the likelihood function. S23: The Markov chain Monte Carlo method is used to solve the posterior distribution and obtain the optimal estimate of the geological parameters at the current time. S24: Update the inverted geological parameters to the digital twin model and iterate the digital twin model in real time.

4. The dynamic optimization method for underground engineering construction schemes according to claim 3, characterized in that, The likelihood function in step S22 is constructed based on the assumption that the monitoring error follows a normal distribution, and the likelihood function is expressed as: ； in, To monitor data in real time, These are finite element prediction values. M represents the standard deviation of the monitoring error, and M represents the number of data types being monitored.

5. The dynamic optimization method for underground engineering construction schemes according to claim 4, characterized in that, In step S23, the Markov chain Monte Carlo method is used to solve for the posterior distribution, specifically including: S231: Initialize parameters ; S232: Set the iteration count R, burn-in period From the proposed distribution using a Gaussian distribution Generate candidate samples ,in ; S233: Calculate the probability of acceptance: ; S234: Based on probability accept ,otherwise ; S235: Discard the results of the first D iterations and take the mean of the remaining samples as the optimal estimate. : 。 6. The dynamic optimization method for underground engineering construction schemes according to claim 1, characterized in that, The deviation vector in step S3 is: ； ； in, The deviation vector is n; n is the number of geological parameters. For the first The relative deviation of each geological parameter; For the first Inversion values ​​of geological parameters; For the first Initial values ​​for each geological parameter; The formula for calculating the comprehensive trigger index is as follows: ； or, ； in, To form a comprehensive trigger index, This indicates taking the maximum value of each element in the deviation vector. Let g be the weight of the g-th parameter, and satisfy... .

7. The dynamic optimization method for underground engineering construction schemes according to claim 6, characterized in that, Step S4 includes: S41: Calculate the evaluation indicators for each scheme based on the inverted geological parameters; S42: Construct a decision matrix, positively process cost-type indicators, and keep the original values ​​of benefit-type indicators; and normalize all indicators using Euclidean norm; S43: The initial weights of each index are calculated using the entropy weight method; S44: Then, based on the comprehensive trigger index, the weights of each indicator are dynamically adjusted using the following formula: ； ； in, Let the weight of the j-th indicator be the weight of the j-th indicator after adjustment based on the comprehensive trigger index. To remove indicators Other indicators, and To be respectively with and The initial weights of the corresponding indicators, Where N is the amplification factor, and N is the total number of indicators; S45: Determine the positive and negative ideal solutions, and calculate the weighted Euclidean distance from each solution to the positive and negative ideal solutions based on the adjusted weights; S46: Calculate the proximity of each scheme based on the weighted Euclidean distance, and output the scheme with the highest proximity as the optimal scheme; the formula for calculating the proximity is: ； in, Let be the weighted Euclidean distance from the i-th solution to the positive ideal solution, and be... Let be the weighted Euclidean distance from the i-th solution to the negative and positive ideal solutions.

8. The dynamic optimization method for underground engineering construction schemes according to claim 7, characterized in that, In step S5, the transition path is generated as follows: If the current work method differs from the optimal solution, the transition path includes work stoppage, reinforcement, and work method switching procedures. If the current support parameters differ from those of the optimal solution, the transition path includes sequentially executed transition steps, process parameters, and expected duration. The construction instructions include instruction type, triggering reason, current status, adjusted plan, transition process, process parameters and risk control elements.

9. A dynamic optimization system for underground engineering construction schemes, characterized in that, include: The model layer includes a digital twin model management module, a Bayesian inversion calculation module, and a finite element calculation engine; The digital twin model management module is used to store, update, and retrieve geometric information, spatial distribution of geological parameters, support structure parameters, and monitoring data; the Bayesian inversion calculation module is used to construct a Bayesian inversion model, solve for the posterior distribution of geological parameters, and output the optimal estimate; the finite element calculation engine is used to provide finite element prediction values ​​for Bayesian inversion and to provide surrounding rock stability simulation results for scheme optimization. The framework layer includes a data acquisition and preprocessing module, a trigger judgment module, a scheme optimization engine, and an instruction generation module; the data acquisition and preprocessing module is used to acquire real-time monitoring data and perform preprocessing. The trigger judgment module is used to calculate the deviation vector of geological parameters and the comprehensive trigger index to complete the scheme adjustment trigger judgment; The scheme optimization engine is used to dynamically re-optimize schemes based on the improved TOPSIS method and output the optimal scheme; the instruction generation module is used to generate transition paths and convert them into structured construction instructions; The application layer includes a data acquisition module, a dynamic optimization module, and a construction implementation module. The data acquisition module is used to acquire basic data and transmit it to the model layer. The dynamic optimization module is used to call the functions of the framework layer and the model layer to realize dynamic optimization of the entire process of the solution. The construction implementation module is used to issue construction instructions to the site management system and collect construction site data to feed back to the model layer.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method of claim 1.

Images