Full-period management method and system for tunnel construction organization optimization based on BIM (Building Information Modeling)

Abstract

The invention discloses a BIM-based full-period management method and system for tunnel construction organization optimization, and belongs to the technical field of construction organization and cost management. The management method comprises the following steps: establishing a three-dimensional geometric model and a parameterized model of a tunnel based on a BIM technology; performing work decomposition on a construction task, and calculating a construction activity time parameter; calculating the cost of materials, machinery and manpower of each construction unit, and summarizing the total cost of the project; optimizing a construction period and resource allocation under a multi-constraint condition by adopting a genetic algorithm; predicting future process time consumption and construction progress based on an LSTM time sequence prediction model; and integrating dynamic data of the construction site, and updating the BIM model in real time. Through combination with dynamic data, efficient organization and scientific management of the whole tunnel construction process are realized, the construction efficiency can be improved, the engineering cost can be reduced, and the scientificity and rationality of project decision making are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of construction organization and cost management, and in particular relates to a full-cycle management method and system for optimizing tunnel construction organization based on BIM. Background Art

[0002] Tunnel projects have a long construction period, strong uncontrollability, and great difficulty in progress control. How to reasonably plan the construction sequence, divide the construction sections, and allocate resources to minimize the project duration and cost under certain conditions is an important issue to ensure the efficient and safe progress of tunnel construction. Traditional tunnel construction organization and cost management usually rely on experience and manual calculations, which are easily affected by various complex variables and uncertain factors, resulting in construction delays, cost overruns, and increased construction safety risks.

[0003] In recent years, with the rapid development of Building Information Modeling (BIM) technology and intelligent optimization algorithms, engineers and project managers have gradually started to use these advanced tools to improve the scientificity and accuracy of tunnel construction organization and cost management. Although certain achievements have been made in the application of BIM technology and optimization algorithms in tunnel construction, most of the existing methods focus on single-dimensional construction planning or static cost analysis, resulting in defects such as low efficiency of tunnel construction organization, inaccurate cost management, or insufficient optimization of construction plans.

[0004] For example, the patent application with the application number 2022116240122 in China discloses a BIM-based project progress monitoring method and its system. This application case builds a BIM model of the project to be monitored and an artificial intelligence algorithm model, and compares it with the prediction of the artificial intelligence LSTM algorithm model, so as to judge and predict various factors existing in the construction process, avoid the risks existing in the construction as much as possible, and minimize the factors affecting the construction.

[0005] For another example, the patent application case with the Chinese patent application number 2024109319730 discloses a method for optimizing the construction progress of a building based on BIM technology. First, a three-dimensional digital BIM model containing information about various parts of the building is constructed, and data on the construction site, including construction progress, resource usage, and environmental conditions, are collected in real time using Internet of Things (IoT) sensors, RFID tags, drones, and 3D laser scanners. Subsequently, the collected data are integrated with the BIM model, and big data technology is used to process and analyze the data. Then, a construction progress prediction model is constructed based on machine learning and deep learning technologies, and a resource optimization scheduling algorithm combining genetic algorithms and ant colony algorithms is designed, and the construction path is optimized using graph theory and path planning algorithms. Then, a construction progress monitoring platform integrating the BIM model, real-time data, and optimization algorithms is adopted to construct an adjustment mechanism and a visualization function to display the construction progress and resource usage in real time. Finally, risk assessment is carried out using the BIM model and real-time data.

[0006] However, the above patent application case involves a wide variety of data types and a large amount of data, making the construction progress monitoring relatively difficult. At the same time, it does not involve the whole-process construction management and faces the problem of being difficult to close the loop to guide the deviation between design and construction. Summary of the Invention

[0007] Aiming at the problems of low efficiency of tunnel construction organization, inaccurate cost management, or insufficient optimization of construction plans in the prior art, the present invention provides a full-cycle management method and system for optimizing tunnel construction organization based on BIM. By integrating multi-dimensional parameter information related to construction and combining genetic algorithms to comprehensively optimize the construction period, resources, and costs under multiple constraints. Subsequently, a dual-channel LSTM architecture is used to perform spatial-temporal association between BIM model parameters and real-time construction data, so as to systematically predict the key parameters of tunnel construction and the impact of different strategies on the project progress and costs, effectively improving the scientificity and rationality of construction organization and management, and significantly enhancing the efficiency of tunnel construction organization.

[0008] To achieve the above object, the technical solution provided by the present invention is as follows:

[0009] The first aspect of the present invention provides a full-cycle management method for optimizing tunnel construction organization based on BIM, including:

[0010] S1. Extract basic information and parameters related to construction from engineering documents, and establish a three-dimensional geometric model and a parametric model of the tunnel based on BIM technology;

[0011] S2. Decompose the construction tasks, divide the construction sections, construction processes, and construction units, and calculate the time parameters of construction activities;

[0012] S3. Calculate the cost of materials, machinery, and labor for each construction unit based on the quantities of work derived from the tunnel BIM model, and summarize the total project cost.

[0013] S4. Optimize the construction period and resource allocation under multiple constraints using the genetic algorithm, and give the optimal construction schedule and resource allocation plan with the lowest cost and a reasonable construction period.

[0014] S5. Based on the construction history ledger data and the BIM model, predict the future process duration and construction progress using the LSTM time series prediction system.

[0015] S6. Integrate the dynamic data on the construction site, update the BIM model in real time, and thus dynamically track and adjust the construction process.

[0016] According to the construction management method described in the first aspect of the present invention, the sources of project information in S1 include, but are not limited to, geological exploration reports, design and construction drawings, special construction plans, and construction organization designs. The basic information and parameters related to construction extracted from the project information include, but are not limited to, tunnel geological information (including rock layer distribution and geological joints) in the geological exploration report; plane curve table, vertical curve table, and tunnel left and right line pile-by-pile coordinate table in the route construction design drawing; tunnel main structure material and size, lining type material and size in the tunnel construction drawing design; excavation method and construction technology in the special construction plan; progress information and resource information in the construction organization design, where the progress information includes total construction period, node construction period, key processes, construction technology excavation method, support type, and process arrangement, and the resource information includes labor plan, construction material and measure material plan, and mechanical equipment allocation information, etc. The mechanical equipment allocation information includes the types, locations, and working states of mechanical equipment.

[0017] Furthermore, the three-dimensional geometric model of the tunnel established based on BIM technology in S1 includes the three-dimensional route model, cross-sectional model, structure model, and equipment and facility model of the tunnel; the parametric model includes the material, size, and model of components. Among them, the three-dimensional route and cross-sectional models of the tunnel include the longitudinal alignment, transverse section, and curve radius of the tunnel; the tunnel structure model includes soil layer structure, portal, open cut tunnel, lining support, ancillary structures, and auxiliary adits; the equipment and facility model includes ventilation facilities, lighting equipment, and drainage facilities.

[0018] Furthermore, the specific steps for establishing the three-dimensional geometric model of the tunnel based on BIM technology in S1 are as follows:

[0019] (1) Based on the route design data and geological information, establish the three-dimensional route model of the tunnel using civil 3D.

[0020] (2) Import the 3D route model and the pile-by-pile coordinate points into Revit in.dwg file format respectively, and import them into Dynamo in.xlsx file format;

[0021] (3) Establish a conventional model core library according to the components included in the tunnel main structure and lining types: portal structure library, open cut tunnel structure library, rock tunnel lining type library, soil tunnel lining type library; The library is the presentation of 3D geometric models and parametric models, and the lining type library includes components such as bolts, steel arches, steel mesh, and concrete components involved in primary support, secondary lining, and invert support;

[0022] (4) Use the Dynamo node FamilyInstance.ByPoint to place the components in the library along the 3D curve points, and intercept the curve through the mileage stake number to place the set of lining type library components corresponding to the mileage;

[0023] (5) Export the tunnel 3D model in.nwc format, import it into Navisworks, and add the progress information in the construction organization design to the Timeliner toolbar to preliminarily simulate the construction process.

[0024] Even more preferably, the sub-project division of the tunnel project and the model components affecting the tunnel construction simulation main structure in the present invention are shown in Table 1 below. The division of the BIM model components in the present invention refers to the model fineness LOD. Through the optimization of the component division, the workload of modeling can be effectively reduced, and at the same time, the reliability of the subsequent resource allocation and cost management optimization plan can be effectively guaranteed, and the selected components are more representative.

[0025] Table 1 Sub-project division of tunnel project and model components affecting the tunnel construction simulation main structure

[0026]

[0027]

[0028] According to the construction management method described in the first aspect of the present invention, in S2, the construction tasks are decomposed, the construction sections, construction processes, and construction units are divided, and the time parameters of construction activities are calculated, specifically including:

[0029] (1) Divide the construction sections according to the surrounding rock grades, and divide the construction surfaces according to the construction section lengths;

[0030] (2) Determine the construction processes that affect the construction period of each construction section, further decompose the construction processes into construction activities with similar workloads and number them; generate a network diagram based on the dependency relationship of the tunnel construction processes to clarify the coordination relationship between the processes; determine the reasonable time interval and distance interval between the key processes (determined according to the specific regulations of the construction requirements). In this invention, the sub-projects of the tunnel project mainly include: portal and open cut tunnel project, tunnel body excavation project, support and lining project, ventilation and drainage project, and auxiliary facilities project.

[0031] (3) Resource requirement estimation and project quota determination

[0032] According to the characteristics of each construction unit, determine the types of resources required for each process, including the types of construction workers, construction machinery and construction materials; according to the project quantity Q i of the construction unit and the daily work quota S i of a construction team or construction machinery in the construction process, calculate the number R i of construction teams or construction machinery required for each process, and the number of work shifts N i per day in the construction activity; Resource scheduling includes the scheduling of human, material and mechanical resources.

[0033] (4) Calculate the duration of the construction activity, combine the intermittent time and interval distance between the construction activities, and integrate the construction organization information into a linear engineering construction progress plan. Dynamically track the actual consumption of materials, equipment and labor costs during the construction process.

[0034] The construction organization information includes: work unit, construction process, mileage parameter, time parameter and construction rate. The intermittent time includes the concrete curing time and the monitoring and measurement time; the interval distance is the interval distance between the construction processes.

[0035] Furthermore, the formula for calculating the duration is as follows:

[0036]

[0037] Among them, T i represents the duration of the i-th construction activity; Q i represents the workload (project quantity) of the i-th construction activity; S i represents the daily work quota of a work team or construction machinery in the i-th construction activity; R i represents the number of work teams or construction machinery in the i-th construction activity; N i represents the number of work shifts per day in the i-th construction activity; α i represents the effective utilization rate of the resources invested in the i-th construction activity.

[0038] When calculating the duration of construction activities, the present invention fully considers the effective utilization rate of resources invested in construction activities, thereby facilitating the improvement of the accuracy of calculation results.

[0039] According to the construction management method described in the first aspect of the present invention, in S3, based on the quantities of work derived from the tunnel BIM model, calculate the cost of materials, machinery, and labor for each construction unit, and summarize the total project cost. Specifically, import the quantity of work file into pricing software (such as Glodon costing software), apply quotas to form comprehensive unit prices, list the sum of unit prices of five items including labor cost, material cost, machinery cost, management fee, and profit, and generate a pricing file.

[0040] The total project cost (i.e., the total cost) includes direct costs, indirect costs, and rewards and penalties. Among them, the direct cost is the sum of the direct costs of each construction activity i, including: labor cost, material cost, machinery cost, safety and civilized construction cost, and reserved unforeseen expenses; the indirect cost includes management fees; the rewards and penalties include rewards for early project completion and penalties for project delays. Among them, the labor cost includes the salaries of construction workers, technicians, and managers; the material cost includes the costs of concrete, steel bars, linings, reinforcement materials, and waterproof materials; the machinery cost includes the quantity of various construction machinery, rental or purchase costs, and operation and maintenance costs.

[0041] According to the construction management method described in the first aspect of the present invention, in S4, use the genetic algorithm to optimize the construction period and resource allocation under multiple constraints, and give the optimal construction schedule and resource allocation plan with the lowest cost and a reasonable construction period. Specifically, it includes:

[0042] (1) Collect the project parameters and genetic parameters generated in S2 and S3. The project parameters include the logical relationship of processes, the planned construction time T p and the optimized time T, the standard cost c iN and the optimized cost c iM ; the genetic parameters include the length L of the individual coding string, the population size P, the iteration frequency G, the crossover probability Pc, and the mutation probability Pm. Among them: the number of construction units is the length of the string.

[0043] (2) Use real number coding to solve the genetic algorithm

[0044] The number of genes contained in each chromosome is the number of operations, and the attributes of each gene include the work name, the duration t i the earliest start time ES h and the immediate predecessor work h.

[0045] (3) Randomly generate an initial population that meets the constraints and basic assumptions as the initial solution to the problem, select an appropriate population size (which can be determined based on the size of the project), calculate the fitness value through the objective function, and evaluate the fitness.

[0046] More preferably, the objective function is:

[0047]

[0048] The constraints are:

[0049]

[0050] Among them, C is the total cost of the project, n is the number of processes, k is the indirect cost coefficient, T is the optimized construction time, e is the reward and penalty coefficient, and T p Indicates the planned construction time; ES i is the earliest start time of process i; ES1 is the earliest start time of the process without a preceding process; ES h is the start time of the immediately preceding process h; ES k , T K c is the earliest start time and duration of process k without a subsequent process; i and t i represents the direct cost and duration of process i; c iN and c iM represents the standard cost and optimized cost of process i, t iN and t iM represents the standard duration and optimized duration of process i.

[0051] The present invention provides an objective function with the lowest cost, optimizes the objective function, and introduces a penalty coefficient in cost calculation, thereby further ensuring the accurate calculation of construction cost parameters and time parameters in the design phase, and improving construction efficiency on the basis of reducing construction costs as much as possible.

[0052] (4) Selection, crossover and mutation gene operations to generate new individuals and form a better population

[0053] The Monte Carlo selection method and uniform crossover method are used to randomly select a construction process plan with a construction mode change, repeatedly check the fitness value, and output the optimal solution after iterative convergence to obtain the optimized construction schedule and resource allocation plan {T 0 ,C 0}.

[0054] According to the construction management method described in the first aspect of the present invention, in S5, based on the construction history ledger data and the BIM model, the future process duration and construction progress are predicted by the LSTM time series prediction system, and the in-depth excavation and spatio-temporal fusion prediction of engineering features are realized through the collaborative work of multiple modules. The prediction system specifically includes the following functional modules: multi-source data fusion processing module, dynamic feature fusion module, LSTM dual-channel deep learning model module, and multi-task output module.

[0055] (1) Multi-source data fusion processing module

[0056] 1) Standardization processing of historical construction data

[0057] Construct a structured time series database, including the following core fields: spatial identifier: mileage stake number (format: KXX+XXX); time identifier: construction date and cycle number (timestamp accuracy: minute level); dynamic construction parameters: surrounding rock grade (I-VI categories, One-hot encoding), excavation method (full face / bench method, etc., represented by embedded vectors), cycle footage (m, range 0.5-3.0, accuracy ±0.05m), single cycle duration (h, range 4-12, accuracy ±0.1h), mechanical parameters (excavation speed m / h, construction rate m / h).

[0058] 2) BIM model feature extraction technology

[0059] Extract three-dimensional engineering features through the IFC standard interface: geological parameters: such as rock mass elastic modulus (GPa, measurement point spacing ≤5m), joint density (number of joints / m 3 , three-dimensional point cloud statistics), groundwater status (discrete classification: dry / seepage / gushing water); structural parameters: such as designed cross-sectional area (m 2 , directly parsed from the BIM model), support type (bolt length / m, shotcrete thickness / cm).

[0060] 3) Spatio-temporal data feature alignment mechanism

[0061] Spatial alignment: Establish a mapping table of the stake number coordinate system to associate BIM parameters with construction data;

[0062] Time alignment: Take the excavation cycle as the basic unit to construct a unified time series index.

[0063] (2) LSTM dual-channel deep learning model module

[0064] Input tensor dimension: [batch_size, time_steps, features];

[0065] Time step setting: According to the construction cycle, take T = 10 historical cycles as the input;

[0066] The specific structure of the LSTM dual-channel deep learning model module is as follows:

[0067] a) Temporal feature processing channel

[0068] Input: The standardized historical construction sequence (including features such as footage, duration, and mechanical parameters, with a time window T = 10 cycles);

[0069] Core module: 3-layer stacked LSTM with 256 hidden units, used to capture the temporal dependencies of the construction progress; the dilated convolutional layer enhances the long-period pattern recognition ability.

[0070] b) Spatial feature processing channel

[0071] Input: Static engineering feature parameters extracted from BIM (such as cross-sectional area, support type, etc.); the current surrounding rock grade (One-hot encoding) and construction method type (embedded vector);

[0072] Core module: The graph convolutional network processes the BIM topological structure, the convolutional layer extracts local correlation features, and the bidirectional LSTM captures the global interaction between parameters.

[0073] (3) Dynamic feature fusion module

[0074] Introduce an adaptive weight allocation mechanism to achieve dual-channel feature fusion. Dynamically calculate the feature weights through the attention gating unit, and the weighted fusion formula is:

[0075] H = α·H 时序 +(1 - α)·H 空间

[0076] where the weight coefficient α is dynamically generated by the feature importance; H 时序 is the temporal feature; H 条件 is the spatial feature.

[0077] (4) Multi-task output module

[0078] Main task output: Predict the footage speed (unit: m / cycle) for the next 3 construction cycles (y1 ∈ R 3 );

[0079] Output dimension:

[0080] Auxiliary task output: Classification of process duration (y2 ∈ {0,1} k ), using a multi-task learning framework; Auxiliary task linked to the objective function: y3 = MLP(h t ) ∈ R 2 , (output [ΔT pred ,ΔCpred ) where ΔT pred is the predicted construction period deviation, and ΔC pred is the predicted cost deviation, forming a closed-loop feedback with the construction period optimization amount and cost optimization amount of S4. Among them, h t refers to the hidden state vector generated by the temporal feature channel and the spatial feature channel in the feature fusion layer.

[0081] Furthermore, the LSTM time series prediction system predicts the future process duration and construction progress, and its process is specifically divided into the following steps:

[0082] a. Receive the surrounding rock grade encoding (One-hot) of the current pile number and the construction method type embedding vector;

[0083] b. Extract the temporal features of the most recent T cycles and perform Z-score normalization;

[0084] c. Call the three-dimensional structure parameters of the corresponding section from the BIM model;

[0085] d. Perform multi-feature fusion and parallel reasoning, and output the prediction result.

[0086] According to the construction management method described in the first aspect of the present invention, the operation of step S6 is specifically as follows:

[0087] Import the construction progress plan file and three-dimensional model file into the BIM construction management platform to complete the association of model, progress, and cost information; obtain the dynamic data of the construction site, transmit the collected actual construction data to the BIM model in real time for updating, and compare it with the design parameters to track the construction progress in real time; bind the design result and the prediction result to the BIM model to generate a 4D progress simulation animation with time dimension and cost dimension to achieve comprehensive information management; when the predicted total construction period exceeds the set proportion of the designed construction period, use the genetic algorithm for optimization and adjustment to automatically generate construction method optimization suggestions.

[0088] Specifically, the ahead / behind areas can be marked with different colors (for example, red: behind > 10%, green: normal, blue: ahead). When the predicted total construction period exceeds the set proportion of the designed construction period (for example, 5%, which can be set manually according to the situation), use the genetic algorithm for optimization and adjustment to automatically generate construction method optimization suggestions.

[0089] The second aspect of the present invention also provides a full-cycle management system for tunnel construction organization optimization based on BIM, including:

[0090] A data collection module for extracting and collecting basic information and parameters related to construction from engineering documents;

[0091] A tunnel model construction module, which is used to establish a three-dimensional geometric model and a parametric model of the tunnel based on BIM technology;

[0092] A construction task decomposition and construction activity time parameter calculation module, which is used to decompose construction tasks, divide construction sections, construction processes, and construction units, and calculate construction activity time parameters;

[0093] A construction cost calculation module, which is used to calculate the cost of materials, machinery, and labor for each construction unit according to the engineering quantities derived from the tunnel BIM model, and summarize the total project cost;

[0094] A construction schedule and resource allocation plan optimization module, which is used to optimize the construction period and resource allocation under multiple constraints by using the genetic algorithm to obtain the optimal construction schedule and resource allocation plan;

[0095] A future process duration and construction progress prediction module, which is used to predict the future process duration and construction progress based on the LSTM time series prediction model according to the construction historical ledger data and the BIM model; and

[0096] A BIM model update module, which is used to integrate the dynamic data of the construction site and update the BIM model in real time for dynamic tracking and adjustment of the construction process.

[0097] Adopting the technical solution provided by the present invention, compared with the prior art, the following beneficial effects can be obtained:

[0098] (1) By constructing a three-dimensional model and a parametric model through BIM technology, the present invention can intuitively display the spatial structure and construction parameters of tunnel construction, improving the accuracy of construction organization and cost management; at the same time, using the genetic algorithm to intelligently optimize the construction period and resource allocation, the optimal construction plan can be quickly obtained under multiple constraints, effectively improving the construction efficiency and reducing the project cost.

[0099] (2) Further, the present invention optimizes the design by dividing the BIM model components, so as to balance the relationship between the modeling workload and the subsequent resource allocation and cost management optimization effect; when calculating the duration of construction activities, the present invention also fully considers the effective utilization rate of the resources invested in the construction activities, which is beneficial to improving the accuracy and reliability of the calculation results.

[0100] (3) When optimizing the construction schedule and resource allocation plan by using the genetic algorithm, the present invention gives the objective function with the lowest cost, and adds a reward and punishment coefficient to the cost calculation. Thus, the calculation of the construction cost parameters and time parameters in the design stage is completed, which is beneficial to further improving the construction efficiency on the basis of minimizing the construction cost as much as possible.

[0101] (4) The present invention combines BIM technology with the dynamic data of the construction site, and dynamically adjusts the construction plan through visual management, ensuring the safety of the project and the rationality of decision-making, and significantly improving the scientificity and efficiency of the whole process management of tunnel construction. Description of the Drawings

[0102] Figure 1 Schematic diagram of the method flow of the present invention. Detailed Implementation Modes

[0103] To further understand the content of the present invention, the present invention will be described in detail below in combination with specific embodiments.

[0104] Combined with Figure 1 , the full-cycle management method for optimizing tunnel construction organization based on BIM in the embodiment of the present invention includes:

[0105] S1. Based on BIM technology, establish a three-dimensional geometric model and a parametric model of the tunnel, and extract basic information and parameters related to construction from engineering data and the BIM model.

[0106] The three-dimensional geometric model includes a three-dimensional route model, a cross-sectional model, a structural model, and an equipment and facility model of the tunnel; the parametric model includes the material, size, and model of components. Among them, the three-dimensional route and cross-sectional models of the tunnel include the longitudinal alignment, transverse section, and curve radius of the tunnel; the tunnel structural model includes the soil layer structure, portal, open cut tunnel, lining support, auxiliary structures, and auxiliary tunnels; the equipment and facility model includes ventilation facilities, lighting equipment, and drainage facilities.

[0107] In the embodiment of the present invention, the sources of engineering data collection for extracting basic information parameters include geological exploration reports, design and construction drawings, special construction plans, and construction organization designs. The basic information extracted from the above engineering data includes the geological information, progress information, and resource information of the tunnel. Among them, the geological information of the tunnel includes rock layer distribution and geological joints; the progress information includes the total construction period, node construction period, and key processes; the resource information includes the required amount of labor, the required amount of construction materials and measure materials, and the mechanical equipment allocation information. Among them, the mechanical equipment allocation information includes the type, location, and working status of mechanical equipment. The required amount of materials is calculated according to sub-projects and summarized separately; the types of construction materials include: steel sections, steel bars, concrete, anchor rods, pipe roofs, advanced small ducts, anchor rods, geotextiles, and waterproof coiled materials.

[0108] According to the embodiment of the present invention, the establishment of the three-dimensional geometric model of the tunnel based on BIM in S1 specifically includes the following steps:

[0109] (1) Based on the route design data and geological information, establish a three-dimensional route model of the tunnel based on civil 3D;

[0110] (2) Import the 3D route model and the coordinates of each pile point into Revit in the form of.dwg files and into Dynamo in the form of.xlsx files;

[0111] (3) Establish a conventional model core library according to the components included in the main tunnel structure and lining types: portal structure library, open cut tunnel structure library, rock tunnel lining type library, soil tunnel lining type library; the library is the presentation of the 3D geometric model and parametric model;

[0112] (4) Use the Dynamo node FamilyInstance.ByPoint to place the components in the library along the 3D curve points, and intercept the curve by mileage to place the set of lining type library components corresponding to the mileage;

[0113] (5) Export the tunnel 3D model in the.nwc format, import it into Navisworks, add the progress information in the construction organization design to the Timeliner toolbar, and preliminarily simulate the construction process.

[0114] S2. Decompose the construction tasks, divide the construction sections, construction processes, and construction units, and calculate the time parameters of the construction activities.

[0115] In the embodiment of the present invention, step S2 specifically includes the following steps:

[0116] (1) Divide the construction sections according to the surrounding rock grades, divide the construction surfaces according to the lengths of the construction sections, and determine the sub-projects and corresponding construction processes that affect the construction periods of each construction section.

[0117] Among them, the number of processes is n, and the tunnel engineering sub-projects include: portal project, open cut tunnel project, tunnel body excavation project, lining project, waterproofing and drainage project, and auxiliary facilities project.

[0118] (2) Determine the construction processes that affect the construction periods of each construction section, further decompose the construction processes into construction units with similar workloads and number them; on the premise of ensuring the total project duration in calendar days, select the key processes from the construction schedule as the main control points for the construction period. Determine the reasonable time intervals and distance intervals between the key processes; generate a network plan diagram according to the dependence relationship of the tunnel construction processes. Coordinate the construction processes through the network plan to minimize the total float of the construction process as much as possible. The above interval time includes the concrete curing time and the monitoring and measurement time; the interval distance refers to the physical safety distance between the construction processes.

[0119] (3) Resource requirement estimation and engineering quota calculation: According to the characteristics of each construction unit, determine the types of resources required for each process, including the types of construction workers, construction machinery and equipment, and construction materials; according to the engineering quantity Q of the construction unit i, calculate the quantity R of construction teams or construction machinery required for each process i ; The daily work quota of a construction team or construction machinery in the construction process is S i , and the number of work shifts per day in the construction activity is N i ; Resource scheduling includes the scheduling of human, material, and mechanical resources. Table 2 below shows the time required for each meter of construction process and the consumption of labor, materials, and machinery in the single-side drift method, that is, the time parameters and resource parameters in the project.

[0120] Table 2 shows the time required for each meter of construction process and the consumption of labor, materials, and machinery in the single-side drift method

[0121]

[0122]

[0123] (4) Calculate the time parameters of the construction unit, determine the intermittent time and interval distance between various construction activities, and organize the construction progress plan for the linear project.

[0124] Dynamically track the actual consumption of materials, equipment, and labor costs during the construction process, compare with the estimated consumption, and calculate the effective utilization rate α of the resources invested in the construction process i . The calculation of the time parameters of the construction unit includes: the planned construction time T p ; the planned time t of work i iN ; the earliest start time ES of process i i ; the start time ES of the preceding process h h ; the earliest start time ES of process k without subsequent processes k .

[0125] The method for calculating the duration according to the workload, construction quota, and resource effective utilization rate of the construction process is:

[0126]

[0127] T i represents the duration of construction activity i; Q i represents the workload of construction activity i; S i represents the daily work quota of a work team or construction machinery in construction activity i; R i represents the number of work teams or construction machinery in construction activity i; N i represents the number of work shifts per day in construction activity i; α i represents the effective utilization rate of the resources invested in construction activity i.

[0128] S3. Calculate the total project cost and the cost parameter of materials, machinery, and labor for the construction unit.

[0129] Specifically, import the partial and sub - item engineering quantity files in step S1 into the BIM software, apply the quota data in step S2, group the comprehensive unit prices, and calculate the total project cost. The total project cost C includes direct costs, indirect costs, and rewards and penalties. Among them, the direct cost c i is the sum of the direct costs of each construction process i, including: labor costs, material costs, machinery costs, safety and civilized construction costs, and reserved unforeseen expenses; indirect costs include management fees. The rewards and penalties are the rewards for early project completion and the penalties for project delays. Among them, labor costs include the salaries of project personnel; material costs include construction material costs and turnover material costs. The construction material costs are the costs of steel, steel bars, concrete, anchor rods, pipe roofs, advanced small ducts, anchor rods, geotextiles, and waterproof coiled materials; machinery costs include machinery usage fees; machinery rental fees; operation and maintenance costs.

[0130] S4. Use the genetic algorithm to optimize the construction period and resource allocation under multiple constraints, and give the optimal construction schedule and resource allocation plan with the lowest cost and reasonable construction period, specifically including:

[0131] (1) Determine variables and parameters

[0132] Collect the project parameters and genetic parameters generated in steps S2 and S3. The project parameters include the logical relationship between processes, the planned construction time T p and the optimized time T, the standard cost c iN and the optimized cost c iM . The genetic parameters include the individual coding string length L, the population size P, the iteration frequency G, the crossover probability Pc, and the mutation probability Pm. Among them: the number of construction units is the string length.

[0133] (2) Parameter coding

[0134] To facilitate the solution of the genetic algorithm, transform the actual problem into a problem represented by parameters. Due to the large number of partial and sub - item projects in tunnel engineering, the chromosome capacity is relatively large. To simplify the coding and decoding steps, the embodiment of the present invention adopts real - number coding. The attributes of these genes include the work name, time, the earliest start time of the work, and the previous work. The number of genes contained in each chromosome is the number of operations. The attributes of each gene include the work name, the duration t i the earliest start time ES i and the immediate predecessor work h.

[0135] (3) Initialize the population

[0136] In this embodiment, a random initial population that meets the constraint conditions and basic assumptions is generated as the initial solution to the problem.

[0137] (4) Fitness evaluation

[0138] Calculate the fitness value according to the objective function and evaluate the fitness. The fitness reflects the degree to which an individual approaches the objective function, and the smaller the fitness value, the better.

[0139] The objective function of the embodiment of the present invention is preferably as follows:

[0140]

[0141] The constraint conditions are:

[0142]

[0143] C is the total project cost, n represents the number of processes, k represents the indirect cost coefficient, T represents the optimized construction time, e represents the reward and penalty coefficient, and T p represents the planned construction time; ES i is the earliest start time of process i; ES1 is the earliest start time of the process without a preceding process; ES h is the start time of the preceding process h; ES k and T K are respectively the earliest start time and the duration of process k without a succeeding process; c i and t i represent the direct cost and the duration of process i; c iN and c iM represent the standard cost and the optimized cost of process i, and t iN and t iM represent the standard duration and the optimized duration of process i.

[0144] (5) Gene operation

[0145] By performing gene operations such as selection, crossover, and mutation, simulate the reproductive function of organisms to generate new individuals and form a better population. The Monte Carlo selection method and the uniform crossover method are used to randomly select a construction process plan with construction mode changes, repeatedly check the fitness value, and output the optimal solution after iterative convergence to obtain the optimized construction schedule and resource allocation plan.

[0146] S5. Based on the LSTM time series prediction system, predict the future process duration and construction progress according to the construction historical ledger data and the BIM model.

[0147] Step S5 is specifically to input the surrounding rock grade and excavation method of the current mileage section, the excavation historical construction ledger data (mileage stake number, cycle footage, single cycle duration), and predict the tunneling speed.

[0148] The loss function for model training is: L = α·MAE(y1) + β·CrossEntropy(y2), where α = 0.7 and β = 0.3;

[0149] Optimization strategy: Use the AdamW optimizer with an initial learning rate of 3e-4 and cosine annealing scheduling;

[0150] Regularization: Dropout rate of 0.2 and L2 weight decay of 1e-5;

[0151] The real-time prediction process is specifically divided into the following steps:

[0152] a. Receive the surrounding rock grade encoding (One-hot) and construction method type embedding vector of the current station number;

[0153] b. Extract the time series features of the nearest T cycles and perform Z-score standardization;

[0154] c. Call the three-dimensional structure parameters of the corresponding section from the BIM model;

[0155] d. Perform multi-feature fusion and parallel inference and output the prediction result;

[0156] e. When the construction period deviation ΔT > 5%, automatically call the S4 module to generate a re-optimization plan and update the progress parameters of the BIM model.

[0157] S6. Integrate the dynamic data of the construction site, update the BIM model in real time, and dynamically track and adjust the construction plan to achieve visual management.

[0158] Step S6 is specifically as follows: Import the construction progress plan file and three-dimensional model file in the design stage into the BIM construction management platform to complete the association of model, progress, and cost information; Provide a three-dimensional visual display of the entire construction process; Obtain the dynamic data on the construction site during the construction stage, compare it with the predicted future construction progress, conduct deviation analysis, and transmit the collected data to the BIM model in real time for updating to achieve visual management. Real-time query the progress and cost details of any station number or construction area, including geological information, construction progress, cost distribution, and resource scheduling. Dynamically track the cost consumption of materials, equipment, and labor, analyze the cumulative deviation in real time, use the genetic algorithm for optimization and adjustment to automatically generate construction method optimization suggestions, and use the genetic algorithm to generate an adjusted resource scheduling plan. The dynamic data of the construction site includes the real-time construction progress; the cost consumption of materials, equipment, and labor. Dynamically track the real-time construction progress, dynamically adjust the construction plan, and generate an adjusted construction plan.

[0159] In summary, the present invention focuses on the optimization of construction progress and resource allocation, runs through the entire process of construction management in the design stage and the construction stage, designs and optimizes the construction activity content reasonably and scientifically, can form a detailed construction plan to guide the actual construction, collects the ledger information in real time during the construction stage, and uses the LSTM algorithm model to predict the usage of labor, materials, and machinery and the construction time information of the subsequent construction processes to guide the construction.

Claims

1. A full-cycle management method for optimizing tunnel construction organization based on BIM, characterized in that, Including: Extracting basic construction-related information and parameters from engineering data, and establishing a 3D geometric model and parametric model of the tunnel based on BIM technology; Decomposing the construction tasks, dividing the construction sections, construction processes, and construction units, and calculating the time parameters of construction activities; Calculating the cost of materials, machinery, and labor for each construction unit based on the quantities of work derived from the tunnel BIM model, and summarizing the total project cost; Using a genetic algorithm to optimize the construction period and resource allocation under multiple constraints to obtain the optimal construction schedule and resource allocation plan; Based on the construction history ledger data and BIM model, predicting the time consumption of future processes and construction progress using an LSTM time series prediction system; Integrating the dynamic data at the construction site, updating the BIM model in real time, and dynamically tracking and adjusting the construction process.

2. The management method according to claim 1, characterized in that The sources of the engineering data include geological exploration reports, design and construction drawings, special construction plans, and construction organization designs. The basic construction-related information and parameters extracted from the engineering data include, but are not limited to, the tunnel geological information in the geological exploration report; the horizontal curve table, vertical curve table, and tunnel left and right line pile-by-pile coordinate table in the route construction design drawing; the tunnel main structure material and size, lining type material and size in the tunnel construction drawing design; the excavation method and construction technology in the special construction plan; the progress information and resource information in the construction organization design, where the progress information includes the total construction period, node construction period, key processes, construction technology excavation method, support type, and process arrangement, and the resource information includes: labor plan, construction material and measure material plan, and mechanical equipment allocation information. And / or the 3D geometric model of the tunnel established based on BIM technology includes the 3D route model, cross-section model, structure model, and equipment and facility model of the tunnel; the parametric model includes the material, size, and model of the components; And / or when establishing the BIM model of the tunnel, the model components related to the tunnel main structure include portal structure, open cut tunnel lining, inverted arch, advanced anchor rod, casing arch, system anchor rod, locking foot anchor rod steel arch frame, concrete pouring, steel arch frame, concrete pouring, inverted arch reinforcement mesh, concrete pouring, geotextile and waterproof board, secondary lining reinforcement mesh, concrete pouring, central water channel, and left and right side cable troughs and covers.

3. The management method according to claim 1, wherein The establishment of the 3D geometric model and parametric model of the tunnel based on BIM technology specifically includes: (1) Based on the route design data and geological information, establishing a 3D route model of the tunnel using civil 3D; (2) Importing the 3D route model and pile-by-pile coordinate points into Revit in.dwg file format and into dynamo in.xlsx file format respectively; (3) Establishing a conventional model core library according to the components included in the tunnel main structure and lining type: portal structure library, open cut tunnel structure library, rock tunnel lining type library, soil tunnel lining type library; (4) Using the dynamo node FamilyInstance.ByPoint to place the components in the library along the 3D curve points, and intercepting the curve by the mileage stake number to place the set of lining type library components corresponding to the mileage; (5) Export the 3D tunnel model in the.nwc format, import it into Navisworks, add the progress information in the construction organization design to the Timeliner toolbar, and preliminarily simulate the construction process.

4. The management method according to any one of claims 1-3, characterized in that The work breakdown of the construction tasks, division of construction sections, construction processes, and construction units, and calculation of the time parameters of construction activities specifically include: (1) Divide the construction sections according to the surrounding rock grade, and divide the construction surfaces according to the length of the construction sections; (2) Determine the construction processes that affect the construction period of each construction section, further decompose the construction processes into construction activities with similar workloads and number them; generate a network diagram based on the dependency relationship of the tunnel construction processes to clarify the coordination relationship between the processes; determine the reasonable time interval and distance interval between the key processes; (3) Estimation of resource requirements and determination of engineering quotas According to the characteristics of each construction unit, determine the types of resources required for each process, including the types of construction workers, construction machinery and equipment, and construction materials; according to the engineering quantity Q of the construction unit i and the daily work quota S of a construction team or construction machinery in the construction process i , calculate the number R of construction teams or construction machinery and equipment required for each process i , and the number of work shifts N per day during the construction activities i ; (4) Calculate the duration of construction activities, combine the intermittent time and interval distance between various construction activities, and integrate the construction organization information into a linear engineering construction progress plan.

5. The management method according to claim 4, characterized in that, The calculation of the duration of the construction activities is as follows: Among them, T i represents the duration of the i-th construction activity; Q i represents the workload of the i-th construction activity; S i represents the daily work quota of a work team or construction machinery in the i-th construction activity; R i represents the number of work teams or construction machinery in the i-th construction activity; N i represents the number of daily work shifts in the i-th construction activity; α i represents the effective utilization rate of resources invested in the i-th construction activity.

6. The management method according to any one of claims 1-3, characterized in that, Use the genetic algorithm to optimize the construction period and resource allocation under multiple constraints to obtain the optimal construction progress plan and resource allocation plan, specifically including: (1)Collect the project parameters and genetic parameters generated in S2 and S3. The project parameters include the process logical relationship, the planned construction time T p and the optimized time T, the standard cost c iN and the optimized cost c iM ; the genetic parameters include the individual coding string length L, the population size P, the iteration frequency G, the crossover probability Pc, and the mutation probability Pm; (2) Solve the genetic algorithm using real number coding The number of genes contained in each chromosome is the number of operations, and the attributes of each gene include the work name, duration, earliest start time, and preceding work; (3) Randomly generate an initial population that meets the constraints and basic assumptions as the initial solution to the problem, select an appropriate population size, calculate the fitness value through the objective function, and evaluate the fitness; the expression of the objective function is as follows: The constraints are: Among them, C is the total project cost, n represents the number of processes, k represents the indirect cost coefficient, T represents the optimized construction time, e represents the reward and penalty coefficient, and T p represents the planned construction time; ES i is the earliest start time of process i; ES1 is the earliest start time of the process without a preceding process; ES h is the start time of the preceding process h; ES k and T K are respectively the earliest start time and the duration of process k without a succeeding process; c i and t i represent the direct cost and the duration of process i; c iN and c iM represent the standard cost and the optimized cost of process i, and t iN and t iM represent the standard duration and the optimized duration of process i; (4) Use the Monte Carlo selection method and the uniform crossover method to randomly select a construction process plan with changing construction modes, repeatedly check the fitness value, and output the optimal solution after iterative convergence to obtain the optimized construction progress plan and resource allocation plan.

7. The management method according to any one of claims 1-3, characterized in that, The prediction system based on the LSTM time series specifically includes a multi-source data fusion processing module, an LSTM dual-channel deep learning model module, and a multi-task output module. The technical implementation and prediction process of each module are as follows: Receive the surrounding rock grade code and construction method type embedding vector of the current pile number; Extract the time series features of the last T cycles and perform Z-score standardization; Call the 3D structure parameters of the corresponding section from the BIM model; Execute multi-feature fusion and parallel inference to output the prediction result.

8. The management method according to claim 7, characterized in that, The processing process of the multi-source data fusion processing module includes: (1) Standardization processing of historical construction data Construct a structured database with fields including: mileage pile number, construction date, surrounding rock grade, excavation method, cycle footage, single cycle duration, and mechanical parameters; (2) Feature extraction of the BIM model Extract the 3D model structure parameters through the IFC standard interface, including geological parameters and structural parameters; (3) Feature alignment of spatio-temporal data Spatial alignment: Establish a mapping table of the pile number coordinate system to associate the BIM parameters with the construction data; Temporal alignment: Take the excavation cycle as the basic unit to construct a unified time series index; And / or the specific structure of the LSTM dual-channel deep learning model module is as follows: (1) Temporal feature processing channel Input: Standardized historical construction sequences; Core module: 3-layer stacked LSTM with 256 hidden units to capture the temporal dependencies of construction progress; (2) Spatial feature processing channel Input: Static parameters extracted by BIM, current surrounding rock grade, and construction method type; Core module: Graph convolutional network processes the BIM topological structure, convolutional layers extract local correlation features, and bidirectional LSTM captures the global interaction between parameters; And / or the dynamic feature fusion module introduces an adaptive weight allocation mechanism to achieve dual-channel feature fusion. The feature weights are dynamically calculated through an attention gating unit. The weighted fusion formula: H = α·H 时序 +(1 - α)·H 空间 where the weight coefficient α is dynamically generated by feature importance; and / or the main task output of the multi-task output module: prediction of the footage advance rate for the next 3 construction cycles, y1 ∈ R 3 ; Its auxiliary task output: process time classification, y2 ∈ {0, 1} k , using a multi-task learning framework; auxiliary task linked to the objective function: y3 = MLP(h t ) ∈ R 2 , the output is [ΔT pred , ΔC pred , where ΔT pred is the predicted construction period deviation, and ΔC pred is the predicted cost deviation. When the construction period deviation amount of ΔT pred > 5%, the S4 module is automatically called to generate a re-optimization plan, and the BIM model progress parameters are updated, forming a closed-loop feedback with the construction period optimization amount and cost optimization amount of S4.

9. The management method according to claim 8, wherein, Integrating the dynamic data of the construction site, updating the BIM model in real time, and dynamically tracking and adjusting the construction process, including: Importing the construction progress plan file and 3D model file into the BIM construction management platform to complete the association of model, progress, and cost information; Obtaining the dynamic data of the construction site, transmitting the collected actual construction data to the BIM model in real time for updating, and comparing it with the design parameters to track the construction progress in real time; Binding the design results and prediction results to the BIM model to generate a 4D progress simulation animation with time dimension and cost dimension to achieve comprehensive information management; When the predicted total construction period exceeds a set proportion of the design construction period, use the genetic algorithm for optimization and adjustment to automatically generate construction method optimization suggestions.

10. A full-cycle management system for optimizing tunnel construction organization based on BIM, characterized in that, Including: Data collection module, used to extract and collect basic information and parameters related to construction from engineering documents; Tunnel model construction module, used to establish a 3D geometric model and parametric model of the tunnel based on BIM technology; Construction task decomposition and construction activity time parameter calculation module, used to decompose construction tasks, divide construction sections, construction processes, and construction units, and calculate construction activity time parameters; Construction cost calculation module, used to calculate the cost of materials, machinery, and labor for each construction unit based on the engineering quantities exported from the tunnel BIM model, and summarize the total project cost; Construction progress plan and resource allocation plan optimization module, used to optimize the construction period and resource allocation under multiple constraints using the genetic algorithm to obtain the optimal construction progress plan and resource allocation plan; Future process duration and construction progress prediction module, used to predict the future process duration and construction progress based on the construction historical ledger data and BIM model using the LSTM time series prediction model; BIM model update module, used to integrate the dynamic data of the construction site and update the BIM model in real time for dynamically tracking and adjusting the construction process.

Images