High-altitude ballastless track construction period prediction modeling method based on hybrid simulation

Abstract

This invention provides a modeling method for predicting the construction period of high-altitude ballastless track based on hybrid simulation, belonging to the field of computer simulation technology. The method includes: S1: defining agent attributes and interaction rules; S2: constructing an adaptive mesh simulation environment; S3: constructing an efficiency dynamic evolution model based on SD; S4: constructing a Bayesian prior probability network based on historical data to introduce perturbations; S5: introducing PERT construction speed distribution to obtain the probability distribution for construction period prediction; S6: building a coupled simulation model and solving for the construction period distribution. This invention achieves unified modeling of construction efficiency evolution, the impact of random perturbations, and spatial interaction constraints through multi-model coupling, dynamic optimization, and probabilistic analysis, improving the accuracy and engineering applicability of construction period prediction, and providing technical support for the planning and risk management of high-altitude railway tunnel construction.

Description

Technical Field

[0001] This invention relates to the field of computer simulation technology, and in particular to a method for predicting the construction period of high-altitude ballastless track based on hybrid simulation. Background Technology

[0002] As railway infrastructure extends into the western plateau regions, the proportion of tunnel engineering has increased significantly. Ballastless track, as a key structure within tunnels, faces challenges such as confined space, complex human-machine collaboration, and difficulties in material transportation. The low air pressure and lack of oxygen in high-altitude environments can lead to a decline in workers' physical functions. The cumulative fatigue from continuous work causes labor productivity to exhibit non-linear time-varying characteristics. Furthermore, the linear organization and cyclical operation characteristics of tunnel construction can easily lead to the propagation of local disturbances along the work chain, causing cumulative delays.

[0003] Existing methods for project schedule prediction and management commonly include the Critical Path Method (CPM), Discrete Event Simulation (DES), System Dynamics (SD), and Multi-Agent Modeling (ABM). Some studies have attempted hybrid modeling, but for the specific scenario of ballastless track construction in high-altitude tunnels, there are still significant shortcomings: a single method cannot take into account both micro-level interactions and macro-level efficiency evolution; many methods use deterministic schedules or simple uniformly distributed stochastic simulations, which do not match engineering reality; they ignore or simplify the causal transmission of low-frequency, high-impact interference events; and fixed-grid modeling makes it difficult to balance the accuracy and efficiency of long-distance tunnel simulation, resulting in large prediction biases and insufficient applicability. Summary of the Invention

[0004] This invention provides a modeling method for predicting the construction period of ballastless track in high-altitude tunnels based on hybrid simulation. Addressing the problems of existing technologies in predicting the construction period of ballastless track in high-altitude tunnels, such as "disconnect between micro-level interaction and macro-level efficiency," "distortion of random distribution models," "lack of causal mapping for sudden disturbances," and "low efficiency of long-distance simulation," this invention achieves unified modeling of construction efficiency evolution, the impact of random disturbances, and spatial interaction constraints through multi-model coupling, dynamic optimization, and probabilistic analysis. This improves the accuracy and engineering applicability of the construction period prediction, providing technical support for the formulation of construction plans and risk management in high-altitude railway tunnels.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A modeling method for predicting the construction period of high-altitude ballastless track based on hybrid simulation includes: S1: Define the heterogeneous agent types, attribute states, and interaction rules in the construction system, and output the rule system used to drive the agent's behavior and interaction; S2: Divide the longitudinal spatial region of the tunnel according to the construction progress, construct an adaptive mesh simulation environment based on the Gaussian function, and output a mesh space that adapts to the construction scenario; S3: Analyze the causal effects of altitude and fatigue on construction efficiency, build a dynamic efficiency evolution model, and output real-time updated dynamic work efficiency. S4: Identify low-frequency, high-impact interference events based on historical engineering data, construct a Bayesian prior probability network, and output the mapping rules and triggering mechanisms between disturbance events and construction procedures. S5: Set the optimistic, most likely, and pessimistic values ​​for construction speed, and use the program review and approval technique combined with Monte Carlo sampling to output the benchmark construction speed for each type of work and each construction cycle. S6: Build a multi-agent modeling and system dynamics coupled simulation framework, use the grid space as the operation carrier, drive the operation of the agents according to the rule system, correct the benchmark construction speed with dynamic operation efficiency, apply process delays according to mapping rules and triggering mechanisms, and output the project duration probability distribution and related quantitative indicators after iterative simulation.

[0006] In this specification, in step S1, the heterogeneous intelligent agents include worker intelligent agents, concrete mixer truck intelligent agents, and gantry crane intelligent agents; among them, the worker intelligent agent is responsible for the execution and connection of construction procedures, triggering state transitions based on the progress of the preceding procedures, and its attributes include the procedure status; the concrete mixer truck intelligent agent is responsible for the transportation of concrete from the mixing plant to the work surface, and its attributes include vehicle capacity, cumulative mileage, failure probability threshold, and current travel speed; the gantry crane intelligent agent is responsible for the coordination of material hoisting and concrete pouring, and its attributes include hoisting capacity, failure rate, and current operating status; the rule system covers the behavioral interaction logic between the three types of intelligent agents and between the intelligent agents and the tunnel environment.

[0007] In this specification, in step S2, the longitudinal spatial region of the tunnel is divided into a completed construction area, a construction area, and a construction-to-be-constructed area; the Gaussian function is used to construct the grid density curve, so that the construction area is allocated a high-resolution fine-grained grid, and the completed construction area and the construction-to-be-constructed area far away from the construction area are allocated a low-resolution coarse-grained grid; as the construction cycle progresses, the grid state function of the entire line is updated in real time to reconstruct the grid density distribution, ensuring that computing resources are concentrated in the currently active construction area.

[0008] In this specification, in step S3, the causal loops corresponding to the causal influence include an altitude-hypoxia-efficiency negative feedback loop and a continuous operation-fatigue-efficiency negative feedback loop; the evolution function of the efficiency dynamic evolution model integrates the altitude attenuation effect and the fatigue-recovery law, wherein the altitude attenuation effect is constructed for hypoxic conditions in high-altitude environments, and the fatigue-recovery law is constructed based on Jaber fatigue theory; the dynamic operation efficiency is used to correct the benchmark construction speed in step S6.

[0009] In this specification, in step S4, the parent node of the Bayesian prior probability network is the identified low-frequency, high-impact interference event, and the child nodes are the construction procedures affected by the interference; the directed edges in the network represent the interference propagation path, and the conditional probability table on the edge describes the delay probability and degree of the corresponding procedure when the interference occurs; the triggering mechanism is to randomly sample and determine whether the interference has occurred based on the prior probability in each time step or each procedure cycle in step S6, and if it has occurred, extend the remaining operation time of the corresponding procedure.

[0010] In this specification, in step S5, the optimistic value of the construction speed is the construction speed under ideal conditions, the most likely value is the construction speed under normal progress, and the pessimistic value is the construction speed under unfavorable conditions; in the single simulation initialization phase, the Monte Carlo method extracts a benchmark construction speed sample for each type of work and each construction cycle, and this sample serves as the initial speed basis for the agent's operation in step S6.

[0011] In this specification, step S6, the time step loop of the coupled simulation includes the following operations: (1) Based on the benchmark construction speed output in step S5, extract the initial speed of the agent operation in the current simulation cycle; (2) Based on the mapping rules and triggering mechanism of step S4, detect and determine whether to apply interference delay; (3) Using the dynamic work efficiency output in step S3, calculate the efficiency correction coefficient and update the agent's work speed; (4) Adjust the resolution of the adaptive mesh based on the Gaussian function in step S2 to suit the current construction area; (5) In the grid space of step S2, the agent performs tasks and interacts according to the rule system of step S1, detects and avoids positional conflicts, and updates the agent's position, task completion and fatigue accumulation.

[0012] In this specification, the behavioral interaction logic of the intelligent agents includes: the interaction logic between the concrete mixer truck intelligent agent and the tunnel environment, covering three transportation stages: from the mixing plant to the inclined shaft entrance, queuing judgment at the inclined shaft entrance, and from the inclined shaft entrance to the working face; the interaction logic between the worker intelligent agent and the gantry crane intelligent agent, covering the coordination and waiting logic in the concrete pouring process; and the interaction logic between the concrete mixer truck intelligent agent and the gantry crane intelligent agent, covering the material handover and waiting logic at the working face.

[0013] In this specification, in step S4, the low-frequency high-impact interference events are identified and obtained through historical engineering data or on-site measurement records, including equipment failure, material interruption, quality rework, power outage in the tunnel, local traffic congestion, etc.; the prior probability of occurrence and average delay time of various interference events are determined by statistical analysis of historical engineering data or on-site measurement records.

[0014] In this specification, in step S6, the number of iterative simulations is no less than N, where N is determined based on simulation accuracy and computational resource consumption requirements; the quantitative indicators include average construction period, confidence interval, most optimistic construction period, most pessimistic construction period, and the variance of construction period fluctuation for each process; the output results also include a histogram of construction period frequency distribution and a cumulative probability curve, which are used for optimization of construction organization and risk management of construction period for high-altitude ballastless track.

[0015] In summary, the present invention has at least the following beneficial effects: To improve the reliability of construction period prediction for high-altitude ballastless track, coupled modeling is used to realistically reflect the dynamic changes in construction efficiency over time and environmental conditions, breaking through the limitations of traditional fixed efficiency or static correction.

[0016] This enables a probabilistic characterization of construction period uncertainty, freeing us from the constraints of a single deterministic construction period or unreasonable random distribution assumptions. It provides quantitative support for construction period guarantee rate analysis and risk reservation, thereby reducing decision-making risks.

[0017] Enhance the ability to express and interpret low-frequency, high-impact construction disturbances, clarify the source, action path, and degree of impact of disturbance events, and improve the adaptability of prediction results to actual construction scenarios.

[0018] While ensuring the simulation accuracy of key areas, we optimize the allocation of computing resources, significantly reduce the computational overhead of long-distance tunnel simulation, and improve simulation efficiency. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the high-altitude ballastless track construction period prediction modeling method based on hybrid simulation involved in this invention.

[0021] Figure 2 This is a schematic diagram of the adaptive mesh simulation environment involved in this invention.

[0022] Figure 3 This is a schematic diagram of the causal loop of the system dynamics involved in this invention.

[0023] Figure 4 This is a schematic diagram of the prior probability network that takes into account perturbations involved in this invention.

[0024] Figure 5This is a schematic diagram of the project schedule prediction process based on PERT velocity distribution involved in this invention.

[0025] Figure 6 This is a schematic diagram of the coupling model framework involved in this invention.

[0026] Figure 7 This is a schematic diagram of the project duration prediction distribution and cumulative frequency curve involved in this invention.

[0027] Figure 8 This is a schematic diagram illustrating the simulation details of the processes involved in this invention. Detailed Implementation

[0028] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the embodiments of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0029] The following disclosure provides many different implementations or examples for carrying out different structures of the embodiments of the present invention. To simplify the disclosure of the embodiments of the present invention, specific examples of components and arrangements are described below. Of course, these are merely examples and are not intended to limit the embodiments of the present invention. Furthermore, reference numerals and / or reference letters may be repeated in different examples of the embodiments of the present invention; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various implementations and / or arrangements discussed.

[0030] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0031] like Figure 1 As shown, this embodiment provides a method for predicting the construction period of high-altitude ballastless track based on hybrid simulation, including: S1: Define the heterogeneous agent types, attribute states, and interaction rules in the construction system, and output the rule system used to drive the agent's behavior and interaction; S2: Divide the longitudinal spatial region of the tunnel according to the construction progress, construct an adaptive mesh simulation environment based on the Gaussian function, and output a mesh space that adapts to the construction scenario; S3: Analyze the causal effects of altitude and fatigue on construction efficiency, build a dynamic efficiency evolution model, and output real-time updated dynamic work efficiency. S4: Identify low-frequency, high-impact interference events based on historical engineering data, construct a Bayesian prior probability network, and output the mapping rules and triggering mechanisms between disturbance events and construction procedures. S5: Set the optimistic, most likely, and pessimistic values ​​for construction speed, and use the program review and approval technique combined with Monte Carlo sampling to output the benchmark construction speed for each type of work and each construction cycle. S6: Build a multi-agent modeling and system dynamics coupled simulation framework, use the grid space as the operation carrier, drive the operation of the agents according to the rule system, correct the benchmark construction speed with dynamic operation efficiency, apply process delays according to mapping rules and triggering mechanisms, and output the project duration probability distribution and related quantitative indicators after iterative simulation.

[0032] In some embodiments, in step S1, the heterogeneous intelligent agents include a worker intelligent agent, a concrete mixer truck intelligent agent, and a gantry crane intelligent agent; wherein, the worker intelligent agent is responsible for the execution and connection of construction procedures, triggering state transitions based on the progress of the preceding procedures, and its attributes include the procedure status; the concrete mixer truck intelligent agent is responsible for the transportation of concrete from the mixing plant to the work surface, and its attributes include vehicle capacity, cumulative mileage, failure probability threshold, and current travel speed; the gantry crane intelligent agent is responsible for the coordination of material hoisting and concrete pouring, and its attributes include hoisting capacity, failure rate, and current operating status; the rule system covers the behavioral interaction logic between the three types of intelligent agents and between the intelligent agents and the tunnel environment.

[0033] In some embodiments, in step S2, the longitudinal spatial region of the tunnel is divided into a completed construction area, a construction area, and a construction-to-be-constructed area; the Gaussian function is used to construct a grid density curve, so that the construction area is allocated a high-resolution fine-grained grid, and the completed construction area and the construction-to-be-constructed area far away from the construction area are allocated a low-resolution coarse-grained grid; as the construction cycle progresses, the grid state function of the entire line is updated in real time to reconstruct the grid density distribution, ensuring that computing resources are concentrated in the currently active construction area.

[0034] In some embodiments, in step S3, the causal loop corresponding to the causal influence includes an altitude-hypoxia-efficiency negative feedback loop and a continuous operation-fatigue-efficiency negative feedback loop; the evolution function of the efficiency dynamic evolution model integrates the altitude attenuation effect and the fatigue-recovery law, wherein the altitude attenuation effect is constructed for hypoxic conditions in high-altitude environments, the fatigue-recovery law is constructed based on Jaber fatigue theory, and the dynamic operation efficiency is used to correct the benchmark construction speed in step S6.

[0035] In some embodiments, in step S4, the parent node of the Bayesian prior probability network is the identified low-frequency, high-impact interference event, and the child node is the construction process affected by the interference; the directed edges in the network represent the interference propagation path, and the conditional probability table on the edge describes the delay probability and degree of the corresponding process when the interference occurs; the triggering mechanism is to determine whether the interference has occurred based on prior probability random sampling in each time step or each process cycle in step S6, and if it has occurred, extend the remaining operation time of the corresponding process.

[0036] In some embodiments, in step S5, the optimistic value of the construction speed is the construction speed under ideal conditions, the most likely value is the construction speed under normal progress, and the pessimistic value is the construction speed under unfavorable conditions; in the single simulation initialization phase, the Monte Carlo method extracts a benchmark construction speed sample for each type of work and each construction cycle, and this sample serves as the initial speed basis for the agent's operation in step S6.

[0037] In some embodiments, step S6, the time-step loop of the coupled simulation includes the following operations: (1) Based on the benchmark construction speed output in step S5, extract the initial speed of the agent operation in the current simulation cycle; (2) Based on the mapping rules and triggering mechanism of step S4, detect and determine whether to apply interference delay; (3) Using the dynamic work efficiency output in step S3, calculate the efficiency correction coefficient and update the agent's work speed; (4) Adjust the resolution of the adaptive mesh based on the Gaussian function in step S2 to suit the current construction area; (5) In the grid space of step S2, the agent performs tasks and interacts according to the rule system of step S1, detects and avoids positional conflicts, and updates the agent's position, task completion and fatigue accumulation.

[0038] In some embodiments, the behavioral interaction logic of the intelligent agent includes: the interaction logic between the concrete mixer truck intelligent agent and the tunnel environment, covering three transportation stages: from the mixing plant to the entrance of the inclined shaft, queuing judgment at the entrance of the inclined shaft, and from the entrance of the inclined shaft to the working face; the interaction logic between the worker intelligent agent and the gantry crane intelligent agent, covering the coordination and waiting logic in the concrete pouring process; and the interaction logic between the concrete mixer truck intelligent agent and the gantry crane intelligent agent, covering the material handover and waiting logic at the working face.

[0039] In some embodiments, in step S4, the low-frequency high-impact interference events are identified and obtained through historical engineering data or on-site measurement records, including equipment failure, material interruption, quality rework, power outage in the tunnel, local traffic congestion, etc.; the prior probability of occurrence and average delay time of various interference events are determined by statistically analyzing historical engineering data or on-site measurement records.

[0040] In some embodiments, in step S6, the number of iterative simulations is not less than N, where N is determined based on simulation accuracy and computational resource consumption requirements; the quantitative indicators include average construction period, confidence interval, most optimistic construction period, most pessimistic construction period, and the variance of construction period fluctuation for each process; the output results also include a histogram of construction period frequency distribution and a cumulative probability curve, which are used for optimization of construction organization and risk management of construction period for high-altitude ballastless track.

[0041] In some embodiments, the worker agent state transition threshold θ in step S1 is determined by statistical analysis of historical construction data. The threshold range for different processes is 0.4-1.0 (i.e., the subsequent process is triggered when the preceding process is 40%-100% completed), for example, 0.4 for upper layer steel bar installation and 1.0 for concrete pouring process; for example, θ for steel bar installation process is set to 0.85 and θ for formwork installation process is set to 0.8; the fault probability threshold for concrete mixer truck agent is set to 0.05-0.15. When the cumulative mileage of the mixer truck reaches the fault trigger mileage, if the fault probability exceeds the threshold, it is determined to be a fault state.

[0042] In some embodiments, the Gaussian function in step S2 is specifically in the form of: ,in The minimum grid density (values ​​range from 0.05 to 0.1 grids / m) is used. This represents the highest grid density (0.5-1.0 grids / m). The coordinates of the center of the current construction area are: The width parameter (values ​​30-80m); the grid state update trigger condition is to update the state of the environmental grid after a fixed time step.

[0043] In some embodiments, the altitude attenuation function in step S3 is specifically: ,in The altitude attenuation coefficient (values ​​range from 0.0001 to 0.0003 / m). The baseline elevation is 2000m. Construction speed at the baseline altitude; fatigue accumulation rate of the fatigue-recovery model. and recovery rate Classified by process type, for labor-intensive processes (such as concrete pouring) The value ranges from 0.02 to 0.04, indicating a recovery rate. Values ​​ranging from 0.01 to 0.02 are used for technology-intensive processes (such as track panel fine-tuning). The value ranges from 0.01 to 0.02, representing the recovery rate. The value ranges from 0.005 to 0.01.

[0044] In some embodiments, in the Bayesian prior probability network of step S4, the prior probability of the interference event is obtained through historical data statistics. For example, the prior probability of equipment failure is 0.02-0.05, the prior probability of power outage in the tunnel is 0.01-0.03, and the prior probability of quality rework is 0.03-0.06. In the conditional probability table, the degree of delay is divided into three levels: mild (0.5-2h), moderate (2-8h), and severe (8-24h). The probability of delay for each level of delay corresponding to different interference events is clear. For example, the probability of mild delay of the concrete pouring process due to loader failure is 60%, moderate is 30%, and severe is 10%.

[0045] In some embodiments, the three parameters of the PERT distribution in step S5 are determined according to the process type, including the optimistic value. Most likely value pessimistic value The value of satisfies For example, the track panel installation process , , Concrete pouring process , , The sample size for Monte Carlo sampling is determined by the number of jobs. Each job is sampled 50-100 times per construction cycle. The average value of the sampling results after removing extreme values ​​(5% above and below) is taken as the baseline construction speed.

[0046] In some embodiments, the PERT distribution adds an "environmental adaptation adjustment": at high altitudes, the pessimistic value... Adjusted for altitude, the correction formula is as follows: This makes the pessimistic speed more consistent with the extreme conditions at high altitudes; at the same time, the sampling process introduces "historical data weighting", with the construction speed data within the last 3 months having a weight of 1.5-2 times that of ordinary data in the sampling, improving the timeliness and accuracy of the sampling results.

[0047] In some embodiments, the time step of step S6 adopts a dynamic adjustment strategy. The basic time step is 0.5-2h. When the interaction frequency of intelligent agents in the construction area is high, it is automatically shortened to 0.5h. When the interaction frequency is low, it is extended to 2h. The convergence condition of the iterative simulation is to stop when the number of iterations reaches 10,000. The variance of the project duration distribution after convergence is controlled within 2-5 days.

[0048] In some embodiments, the coupled simulation framework adds a "multi-scenario comparison module": it supports the simultaneous input of 3-5 different resource configuration schemes (such as different numbers of gantry cranes, different numbers of construction team members), and outputs the schedule probability distribution, cost input and risk level of each scheme through parallel simulation. The risk level is quantified by the schedule overdue probability (under a 90% guarantee rate). This module can directly provide a multi-scheme decision basis for construction organization optimization, breaking through the limitations of traditional single-scheme simulation.

[0049] The core content of this invention includes: defining three heterogeneous intelligent agents—workers, concrete mixer trucks, and gantry cranes—and clarifying their attribute states and micro-interaction rules; constructing an adaptive grid simulation environment based on Gaussian functions, dynamically allocating grid resolution according to the construction area; building a system dynamics (SD) efficiency evolution model to quantify the nonlinear impact of altitude hypoxia and continuous operation fatigue on work efficiency; introducing low-frequency, high-impact construction disturbances through a Bayesian prior probability network to establish a causal mapping between disturbances and work processes; using the Program Evaluation and Review Technique (PERT) distribution combined with Monte Carlo sampling to describe the randomness of construction speed; and building a coupled simulation framework of multi-agent modeling (ABM) and system dynamics (SD), outputting the probability distribution of the construction period and related quantitative indicators through large-scale iterative calculations, providing a basis for optimizing construction organization.

[0050] Step S1: Define agent attributes and interaction rules The relevant symbols and definitions involved in the S1.1 modeling process are explained in Table 1: Table 1. Symbol Definition Table ; S1.2 Define heterogeneous agent types and their attribute states.

[0051] Based on the characteristics of ballastless track construction technology, the core elements of the construction system are abstracted into three types of intelligent agents, and their state variables are defined: S1.2.1 Worker Intelligent Agent: Responsible for the specific execution and connection of each construction process, and determines whether to trigger state transition based on the completion progress of the preceding process. It is the core entity for promoting the construction progress.

[0052] (1.1) (1.2) In the formula C k ( t The function indicates the current construction progress. Used to determine whether a subsequent construction process has begun; The threshold for state transition is determined by experience at the construction site for different processes. For construction workers The status of the process at any given moment; This corresponds to the process sequence number.

[0053] S1.2.2 Concrete Mixer Truck Intelligent Agent: Responsible for transporting concrete from the mixing plant to the work site. Its attributes include: vehicle capacity, current load, cumulative mileage, failure probability threshold, and current speed. The truck's failure rate and failure mileage as a function of mileage are expressed by the following formula: β, η These are the shape parameter and the scale parameter, respectively, which are calibrated based on the vehicle's historical operating data through maximum likelihood estimation.

[0054] (1.3) (1.4) In the formula Let represent a random variable that follows a continuous uniform distribution on the interval (0,1); S1.2.3 Gantry Crane Agent: Responsible for coordinating material hoisting and concrete pouring within the work area. Its attributes include: hoisting capacity, movement range, failure rate, and current operating status. The failure rate is a constant c, expressed by the following formula.

[0055] (1.5) S1.3 Construct rule-based micro-interaction logic.

[0056] Establish behavioral interaction rules between agents and between agents and the environment to drive simulation and deduction.

[0057] S1.3.1 Interaction between concrete mixer trucks and the tunnel environment: Concrete mixer truck transportation includes three stages: from the mixing plant to the inclined shaft entrance, determining whether to queue at the inclined shaft entrance, and from the inclined shaft entrance to the working face. The total transportation time is shown below: (1.6) The queuing logic for judging concrete mixer trucks is as follows, where... v n,j ( t Let be the speed of the nth tanker truck at time t. (1.7) S1.3.2 Interaction between construction personnel and gantry crane: The gantry crane is responsible for transporting concrete during the concrete pouring process. In this process, concrete is transferred from a concrete mixer truck to the gantry crane, which then transfers it to the construction area for pouring. The time the gantry crane waits for construction personnel at the work surface during the pouring process is expressed by the following formula: (1.8) S1.3.3 Interaction between concrete mixer truck and gantry crane: The waiting time of the mixer truck at the working face is expressed by the following formula: (1.9) (1.10) Step S2: Construct a simulation environment with an adaptive mesh S2.1 Zoning Definition of Tunnel Spatial Environment. Based on the dynamic process of construction progress, for example, during cycle j, the corresponding construction area is [ The tunnel's longitudinal space is divided into three characteristic regions, and the tunnel environment is divided as follows: Figure 2 As shown.

[0058] (1.11) 1. Completed construction area: Located behind the work site, personnel and equipment have been removed, and the interaction density is extremely low; 2. Area under construction: The section where the current construction cycle takes place, which is also a high-frequency area for human-machine interaction and material flow; 3. Area to be constructed: Located in front of the work area, there is currently no actual construction work except for some vehicle traffic.

[0059] S2.2 Dynamic generation of mesh density based on Gaussian function.

[0060] To balance the simulation accuracy of microscopic interactions with the efficiency of large-scale computation, a non-uniform mesh generation strategy is adopted. Specifically: 1. Use high-resolution, fine-grained meshes in the "construction area" to accurately capture the interactions and location updates of workers and equipment in confined spaces; 2. Use low-resolution coarse-grained grids in the "completed construction area" and the "to-be-constructed area" which is far from the "construction area" to reduce unnecessary computational overhead.

[0061] A grid density curve is constructed using the Gaussian distribution function to gradually transition the "area to be constructed" to the minimum grid density along the construction direction, ensuring the stability of the simulation values ​​and small cumulative error. The density function is shown in the following formula.

[0062] (1.12) In the formula The minimum density function is defined, corresponding to the lowest resolution coarse-grained network; μ The mean value of the density function is determined based on the location of the working surface. σ The variance is determined based on the planned construction schedule for each cycle.

[0063] S2.3 Dynamic update of grid state.

[0064] During the simulation process, as the process cycle is completed, the start and end coordinates of the work activity area […]. The system moves forward longitudinally along the tunnel. The system updates the grid state function across the entire line in real time. Dynamically reconfigure the grid density distribution to ensure that high computing resources are always allocated to the currently active construction areas.

[0065] Step S3: Construct an efficiency dynamic evolution model based on SD S3.1 Causal loop analysis of efficiency influencing factors.

[0066] By introducing the system dynamics (SD) method, the nonlinear effects of high-altitude environment and continuous operation on work efficiency are quantified, and a system is constructed as follows: Figure 3 Feedback loop shown: S3.1.1 Altitude-Hypoxia-Efficiency Negative Feedback: As tunnel altitude increases, air oxygen content decreases, workers' blood oxygen saturation decreases, physiological functions decline, and construction efficiency decreases.

[0067] S3.1.2 Continuous operation - fatigue - efficiency negative feedback: Increased continuous operation time -> physical energy consumption and fatigue accumulation -> decreased action accuracy and reaction speed -> reduced construction efficiency.

[0068] S3.2 Establish the efficiency evolution function of multi-factor coupling.

[0069] The above causal relationship is transformed into a mathematical model and embedded into the simulation time step: S3.2.1 Modeling the Altitude Attenuation Function: Setting a Baseline Altitude When the construction site is at an altitude When using an exponential function Describe the nonlinear decay trend of efficiency with increasing altitude.

[0070] (1.13) S3.2.2 Fatigue-Recovery Modeling: Based on Jaber fatigue theory, fatigue state variables are introduced. This variable accumulates with increasing continuous working time and recovers with rest. And it is defined as follows: This describes the trend of efficiency changes as fatigue accumulates.

[0071] (1.14) (1.15) In the formula Indicates the time of a certain process The fatigue level after normalization is set to [0,1], where 1 represents extreme fatigue; To extend the construction time step; These represent the cumulative fatigue rate and the fatigue recovery rate, respectively, and are calibrated by regression analysis of different processes. ak , b k , c k The coefficients are fatigue function parameters, fitted based on field data. L The target length of the construction cycle. n i For process i The number of construction workers allocated.

[0072] (1.16) S3.2.3 Dynamic Efficiency Output: Combining the above factors, a standard efficiency function is constructed for the actual construction speed at each time step. This function is updated in real time during each simulation step and fed back to the ABM module.

[0073] (1.17) Step S4: Construct a Bayesian prior probability network based on historical data and introduce perturbations. S4.1 Identification and statistics of typical interference events.

[0074] Based on historical engineering data or on-site measurement records, identify "unplanned disruption events" that occur infrequently but have a significant impact on the project schedule. Calculate the prior probability and average delay duration for each event.

[0075] S4.2 Construct a Bayesian prior probability network, such as Figure 4 As shown.

[0076] By utilizing the topology of Bayesian networks, a mapping relationship between interfering factors and specific construction procedures is established: S4.2.1 Node Definition: Interference events are defined as parent nodes in the network, and affected processes are defined as child nodes.

[0077] S4.2.2 Edges and Weights: Directed edges represent the transmission path of interference, and the conditional probability table on the edge describes the probability and degree of delay in the corresponding process when interference occurs.

[0078] Random disturbances triggered during the S4.3 simulation process.

[0079] In each time step or cycle of the ABM process, random sampling is performed based on the prior probabilities of the Bayesian network. If a disruptive event is determined to have occurred, the remaining operation time of the current process is directly increased according to the consequences defined by the network, thereby explicitly simulating the impact of the unexpected event on the project schedule.

[0080] Step S5: Introduce the PERT construction speed distribution to obtain the probability distribution of project duration prediction. S5.1 PERT distribution modeling of construction speed.

[0081] Abandoning the traditional uniform distribution assumption, the PERT distribution is adopted to describe the randomness of the construction workers' work speed, so as to reflect the characteristics of the central limit theorem of engineering data and the long-tail risk under adverse conditions: three key parameters are set: the optimistic speed b under ideal conditions, the most likely speed m under normal progress, and the pessimistic speed a under adverse conditions.

[0082] S5.2 Monte Carlo random sampling.

[0083] In the initialization phase of a single simulation experiment, the Monte Carlo method is used to extract specific construction speed samples from the PERT distribution for each type of work and each cycle as benchmark speeds. Multiple simulations at different speeds are then performed to obtain different project duration predictions. Figure 5 As shown.

[0084] Step S6: Build a coupled simulation model and solve for the project duration distribution. S6.1 Coupled Model Framework, such as Figure 6 As shown.

[0085] Construct a unified time-step advancement mechanism to organically integrate the above models: S6.1.1 Initialization: Input tunnel and environmental parameters; resource configuration; PERT distribution parameters, etc.

[0086] S6.1.2 Time step loop. In each loop: The construction team speed was randomly sampled based on the PERT distribution and used for agent simulation calculations. The Bayesian network trigger conditions are detected and determined, and interference delays are applied in the corresponding processes. Within each time step t in the loop: The system dynamics calculates the efficiency correction coefficient based on the current altitude of the agent and the continuous working time, and returns the basic parameters of the agent for correction. Adaptive meshing divides the mesh based on the density function for different mesh environments at the current time step; The agent performs movement or tasks in the grid environment based on the corrected speed and specific task rules; agents interact with each other; detect positional conflicts and perform avoidance, and update agent positions, task completion rate, and agent fatigue accumulation.

[0087] S6.2 Large-scale iteration and probability distribution output.

[0088] Execute N independent full-process simulation experiments and record the total project duration data for each experiment; Statistical analysis is performed on N project duration samples to generate a frequency distribution histogram and cumulative probability curve for the project duration. It outputs the average construction period, confidence interval (such as the construction period under a 90% guarantee rate), the most optimistic / most pessimistic construction period prediction, and the variance of the construction period fluctuation for each process, providing a quantitative basis for construction organization optimization.

[0089] This solution has at least the following technical advantages: 1. Improving the reliability of construction period prediction for high-altitude ballastless track: This invention constructs a hybrid simulation framework that couples multi-agent modeling with system dynamics. It unifies the modeling of the interaction behaviors of construction personnel, loading / unloading, and transportation equipment with the efficiency evolution process caused by high-altitude hypoxia and continuous operation fatigue, enabling dynamic updates of construction efficiency over time and environmental conditions. Compared to traditional methods using fixed efficiency or static correction coefficients, this invention more realistically reflects the efficiency changes during actual construction.

[0090] 2. Achieving a probabilistic characterization of construction period uncertainty and reducing the risk of deterministic decision-making: This invention introduces a stochastic modeling method for construction speed based on the PERT distribution, combined with Monte Carlo simulation, breaking through the limitations of traditional assumptions of a single deterministic construction period or uniform distribution. It can output a complete probability distribution and cumulative probability curve of the construction period. This distribution can reasonably reflect the right-skewed and long-tailed characteristics of the construction period, providing a quantitative basis for the analysis of the project schedule guarantee rate and risk reservation.

[0091] 3. Enhanced expressiveness and engineering interpretability of low-frequency, high-impact construction disturbances: By constructing a construction disturbance model based on a Bayesian prior probability network, low-frequency, high-impact events such as equipment failure, logistical delays, and quality rework are explicitly mapped to specific construction procedures, and corresponding delay effects are triggered according to their prior probabilities during simulation. Compared to existing technologies that simplify disturbances to random noise, this invention can clearly identify the source of disturbances and their action paths, improving the adaptability of prediction results to actual construction scenarios.

[0092] 4. Significantly improve simulation calculation efficiency while ensuring prediction accuracy: This invention proposes an adaptive mesh generation strategy based on Gaussian function, which uses high-resolution mesh in active construction areas and low-resolution mesh in non-critical areas to achieve dynamic optimization of computing resources.

[0093] In summary, this invention enables unified modeling of construction efficiency evolution, random disturbance effects, and spatial interaction constraints under complex high-altitude tunnel construction conditions, significantly improving the accuracy and engineering applicability of ballastless track construction period prediction, and has good promotional value and application prospects.

[0094] In one specific embodiment, the construction period prediction for ballastless track construction in tunnels operating in high-altitude, highly disruptive environments is as follows: (I) Implementation Environment and Basic Data This embodiment uses the construction of ballastless track at the entrance of a double-track high-altitude railway tunnel as an example. The construction area is 4.7km long with an average altitude of 3500m, and the track slab adopts a CRTS double-block ballastless track structure.

[0095] At the implementation level, this embodiment uses Python to develop a simulation program and performs large-scale Monte Carlo iterative calculations on a 64-bit Windows 11 workstation environment equipped with an Intel Core i7-12700KF CPU and an NVIDIA GeForce RTX 3060 GPU.

[0096] (II) Resource allocation and process parameter setting As shown in Tables 2 and 3, this embodiment initializes the work teams and mechanical equipment based on the actual on-site construction logs and quotas.

[0097] Table 2. Personnel Configuration and Parameters for the Work Area Involved in the Case Study ; Table 3. Equipment Configuration and Parameters in the Work Areas Involved in the Case Study ; (III) Model parameter calibration and coupling logic Based on the principles of system dynamics and historical data regression, this embodiment calibrates the key parameters affecting efficiency as follows: Altitude and fatigue correction: Set the altitude impact threshold H0=2000m. When the altitude exceeds this value, nonlinear efficiency decay is triggered. Set the fatigue accumulation rate λ=0.03 to calculate the efficiency loss under continuous operation.

[0098] Equipment failure model: The failure mileage of concrete mixer trucks follows a Weibull distribution. The shape parameter β≈1.8 and the scale parameter are set as follows: =1600km; the failure rate of the gantry crane foundation is set at 0.1%.

[0099] Adaptive mesh parameters: Set the length of a single construction cycle to 100m, and the width parameter σ of the Gaussian mesh density function to 50m.

[0100] (iv) Definition of interference terms and setting of prior probabilities This embodiment introduces unplanned disturbances through a Bayesian prior probability network. Based on historical engineering records, 16 typical disturbance events are identified, including site clearing, power outages, equipment failures, and traffic congestion. Each disturbance item is calibrated based on historical data and associated with relevant processes. During the simulation, the system randomly triggers the above events according to prior probabilities and applies corresponding delays. The calibration data and impact propagation paths are as follows: Figure 4 As shown.

[0101] (V) Hybrid Simulation Solution Process After completing the above input preparations, this embodiment starts the ABM-SD coupled simulation process, as follows: Figure 6 As shown: Time-step progression: Within each time step, the agent performs movement and tasks based on the current grid environment; Dynamic feedback: The SD module calculates the current worker's fatigue level and the efficiency correction coefficient based on the altitude in real time, and feeds it back to ABM to update the work speed for the next moment; Interference detection: The status of Bayesian network nodes is detected in parallel. If an interference is triggered, the remaining operation time of the corresponding process is dynamically extended.

[0102] To ensure the stability of the statistical results, this embodiment sets the number of Monte Carlo simulations N=10,000.

[0103] (vi) Operational Results and Performance Analysis After 10,000 iterations of calculation, the project duration prediction results and performance indicators output by this embodiment are as follows: Project duration distribution characteristics: The output results follow a slightly right-skewed approximate normal distribution, such as... Figure 7 As shown, the predicted average construction period was 47.80 days, mainly distributed in the range of 46-50 days. The actual project completion time was 48.63 days, located at the 84th percentile of the cumulative probability of the predicted distribution, with a relative prediction error of only 2%, verifying the accuracy of the model.

[0104] Long-tail risk identification: Simulation results reveal that each process has significant variance in schedule variation, such as... Figure 8 As shown, the long tail on the right side of the overall project duration distribution quantifies the delay risk under extreme interference combinations, providing a quantitative basis for the project team to reserve a buffer period.

[0105] The embodiments described above are for illustrative purposes only and are not intended to limit the invention. Therefore, any changes in numerical values ​​or substitutions of equivalent elements should still fall within the scope of this invention.

[0106] The above detailed description will enable those skilled in the art to understand that the present invention can indeed achieve the aforementioned objectives and has complied with the provisions of the Patent Law.

[0107] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention. The above descriptions are merely preferred embodiments of the invention and are not intended to limit the invention. It should be noted that any modifications, equivalent substitutions, and improvements made within the spirit and principles of the invention should be included within the scope of protection of the invention.

[0108] It should be noted that the above description of the process is for illustrative purposes only and does not limit the scope of this specification. Those skilled in the art can make various modifications and changes to the process under the guidance of this specification. However, these modifications and changes remain within the scope of this specification.

[0109] The basic concepts have been described above. Obviously, for those skilled in the art who have read this application, the above disclosure is merely illustrative and does not constitute a limitation of this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, and therefore, such modifications, improvements, and corrections still fall within the spirit and scope of the exemplary embodiments of this application.

[0110] Furthermore, this application uses specific terms to describe its embodiments. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic related to at least one embodiment of this application. Therefore, it should be emphasized and noted that "an embodiment," "one embodiment," or "an alternative embodiment" mentioned twice or more in different positions in this specification do not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of this application can be appropriately combined.

[0111] Furthermore, those skilled in the art will understand that aspects of this application can be described and illustrated through several patentable types or situations, including any new and useful combination of processes, machines, products, or substances, or any new and useful improvements thereof. Therefore, aspects of this application can be implemented entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or a combination of hardware and software. All of the above hardware or software can be referred to as a “unit,” “module,” or “system.” Furthermore, aspects of this application can take the form of a computer program product embodied in one or more computer-readable media, wherein computer-readable program code is contained therein.

[0112] The computer program code required for the operation of each part of this application can be written in any one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, and Python; general programming languages ​​such as C; Visual Basic, Fortran2103, Perl, COBOL2102, PHP, and ABAP; dynamic programming languages ​​such as Python, Ruby, and Groovy; or other programming languages. This program code can run entirely on the user's computer, or as a standalone software package on the user's computer, or partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer via any network, such as a local area network (LAN) or wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as Software as a Service (SaaS).

[0113] Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this application are not intended to limit the order of the processes and methods of this application. Although some currently considered useful embodiments of the invention have been discussed in the foregoing disclosure by way of various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the substance and scope of the embodiments of this application. For example, although the implementation of the various components described above can be embodied in a hardware device, it can also be implemented as a purely software solution, such as an installation on an existing server or mobile device.

[0114] Similarly, it should be noted that, in order to simplify the description of the present application and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of the embodiments of the present application sometimes combines multiple features into a single embodiment, drawing, or description thereof. However, this approach of the present application should not be construed as reflecting an intention that the claimed subject matter requires more features than expressly recited in each claim. Rather, the subject of the invention should possess fewer features than in any single embodiment described above.

Claims

1. A high-altitude ballastless track construction period prediction modeling method based on hybrid simulation, characterized in that, include: S1: Define the heterogeneous agent types, attribute states, and interaction rules in the construction system, and output the rule system used to drive the agent's behavior and interaction; S2: Divide the longitudinal spatial region of the tunnel according to the construction progress, construct an adaptive mesh simulation environment based on the Gaussian function, and output a mesh space that adapts to the construction scenario; S3: Analyze the causal effects of altitude and fatigue on construction efficiency, build a dynamic efficiency evolution model, and output real-time updated dynamic work efficiency. S4: Identify low-frequency, high-impact interference events based on historical engineering data, construct a Bayesian prior probability network, and output the mapping rules and triggering mechanisms between disturbance events and construction procedures. S5: Set the optimistic, most likely, and pessimistic values ​​for construction speed, and use the program review and approval technique combined with Monte Carlo sampling to output the benchmark construction speed for each type of work and each construction cycle. S6: Build a multi-agent modeling and system dynamics coupled simulation framework, use the grid space as the operation carrier, drive the operation of the agent according to the rule system, correct the benchmark construction speed with dynamic operation efficiency, apply the process delay according to the mapping rules and triggering mechanism, and output the construction period probability distribution and related quantitative indicators after iterative simulation. In step S1, the heterogeneous intelligent agents include worker intelligent agents, concrete mixer truck intelligent agents, and gantry crane intelligent agents. The worker intelligent agent is responsible for the execution and connection of construction procedures, triggering state transitions based on the progress of preceding procedures, and its attributes include procedure status. The concrete mixer truck intelligent agent is responsible for transporting concrete from the mixing plant to the work surface, and its attributes include vehicle capacity, cumulative mileage, fault probability threshold, and current speed. The gantry crane intelligent agent is responsible for coordinating material hoisting and concrete pouring, and its attributes include hoisting capacity, fault rate, and current operating status. The rule system covers the behavioral interaction logic between the three types of intelligent agents and between the intelligent agents and the tunnel environment. In step S3, the causal loops corresponding to the causal influence include an altitude-hypoxia-efficiency negative feedback loop and a continuous operation-fatigue-efficiency negative feedback loop; the evolution function of the efficiency dynamic evolution model integrates the altitude decay effect and the fatigue-recovery law, wherein the altitude decay effect is constructed for hypoxia conditions in high-altitude environments, and the fatigue-recovery law is constructed based on Jaber fatigue theory.

2. The method for predicting the construction period of high-altitude ballastless track based on hybrid simulation according to claim 1, characterized in that, In step S2, the longitudinal spatial region of the tunnel is divided into a completed construction area, a construction area, and a construction-to-be-constructed area. The Gaussian function is used to construct the grid density curve, so that the construction area is allocated a high-resolution fine-grained grid, while the completed construction area and the construction-to-be-constructed area far from the construction area are allocated a low-resolution coarse-grained grid. As the construction cycle progresses, the grid state function of the entire line is updated in real time to reconstruct the grid density distribution, ensuring that computing resources are concentrated in the currently active construction area.

3. The high-altitude ballastless track construction period prediction modeling method based on hybrid simulation according to claim 1, characterized in that, In step S4, the parent node of the Bayesian prior probability network is the identified low-frequency, high-impact interference event, and the child nodes are the construction procedures affected by the interference. The directed edges in the network represent the interference propagation path, and the conditional probability table on the edge describes the delay probability and degree of the corresponding procedure when the interference occurs. The triggering mechanism is to randomly sample and determine whether the interference has occurred based on the prior probability in each time step or each procedure cycle in step S6. If it has occurred, the remaining operation time of the corresponding procedure is extended.

4. The high-altitude ballastless track construction period prediction modeling method based on hybrid simulation according to claim 1, characterized in that, In step S5, the optimistic value of the construction speed is the construction speed under ideal conditions, the most likely value is the construction speed under normal progress, and the pessimistic value is the construction speed under unfavorable conditions. In the single simulation initialization phase, the Monte Carlo method extracts a benchmark construction speed sample for each type of work and each construction cycle. This sample serves as the initial speed basis for the agent's operation in step S6.

5. The method for predicting the construction period of high-altitude ballastless track based on hybrid simulation according to claim 1, characterized in that, In step S6, the time-step loop of the coupled simulation includes the following operations: (1) Based on the benchmark construction speed output in step S5, extract the initial speed of the agent operation in the current simulation cycle; (2) Based on the mapping rules and triggering mechanism of step S4, detect and determine whether to apply interference delay; (3) Using the dynamic work efficiency output in step S3, calculate the efficiency correction coefficient and update the agent's work speed; (4) Adjust the resolution of the adaptive mesh based on the Gaussian function in step S2 to suit the current construction area; (5) In the grid space of step S2, the agent performs tasks and interacts according to the rule system of step S1, detects and avoids positional conflicts, and updates the agent's position, task completion and fatigue accumulation.

6. The method for predicting the construction period of high-altitude ballastless track based on hybrid simulation according to claim 1, characterized in that, The behavioral interaction logic of the intelligent agents includes: the interaction logic between the concrete mixer truck intelligent agent and the tunnel environment, covering three transportation stages: from the mixing plant to the entrance of the inclined shaft, queuing judgment at the entrance of the inclined shaft, and from the entrance of the inclined shaft to the working face; the interaction logic between the worker intelligent agent and the gantry crane intelligent agent, covering the coordination and waiting logic in the concrete pouring process; and the interaction logic between the concrete mixer truck intelligent agent and the gantry crane intelligent agent, covering the material handover and waiting logic at the working face.

7. The method for predicting the construction period of high-altitude ballastless track based on hybrid simulation according to claim 1, characterized in that, In step S4, the low-frequency high-impact interference events are identified and obtained through historical engineering data or on-site measurement records, including equipment failure, material interruption, quality rework, power outage in the tunnel, and local traffic congestion; the prior probability of occurrence and average delay time of various interference events are determined by statistical analysis of historical engineering data or on-site measurement records.

8. The method for predicting the construction period of high-altitude ballastless track based on hybrid simulation according to claim 1, characterized in that, In step S6, the number of iterative simulations is no less than N, where N is determined based on simulation accuracy and computational resource consumption requirements. The quantitative indicators include average construction period, confidence interval, most optimistic construction period, most pessimistic construction period, and the variance of construction period fluctuation for each process. The output results also include a histogram of construction period frequency distribution and a cumulative probability curve. The output results are used for optimization of construction organization and risk management of construction period for high-altitude ballastless track.

Images