Ventilation simulation calculation method for uninclined shaft tunnel by combining side pilot tunnel with double main tunnels

Abstract

The invention relates to the crossing field of tunnel engineering and computational fluid mechanics, and discloses a non-inclined vertical shaft tunnel side pilot tunnel combined double-main-tunnel ventilation simulation calculation method which comprises the following steps: step 1, decomposing a three-dimensional model of a tunnel system into a first type region and a second type region; step 2, for the second type of region, establishing a mathematical agent model representing a mapping relationship between input and output boundary states of the second type of region by acquiring high-fidelity simulation data of the second type of region under various boundary conditions; and step 3, aiming at a specific ventilation simulation task, based on a decomposition result in the step 1, constructing and solving a global system equation set. A tunnel system is decomposed into a linear flow area and a nonlinear node area, and a global system equation set with mixed fidelity is finally constructed and solved, so that the method avoids carrying out three-dimensional computational fluid dynamics simulation which is huge in time consumption on the whole tunnel system, and unification of high efficiency and high precision of analog computation is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of intersection of tunnel engineering and computational fluid dynamics, and in particular to a ventilation simulation calculation method for a side pilot tunnel combined with dual main tunnels in a non-inclined vertical shaft tunnel. Background Art

[0002] As a key component of modern transportation infrastructure, tunnels are crucial for operational safety and efficiency. Especially for long, complex tunnels, such as dual-main tunnels with no inclined shafts and supplemented by side guide tunnels, the design and operational control of internal ventilation systems are crucial for ensuring both daily operational conditions and the safe evacuation of personnel in emergencies such as fires. Therefore, accurate and efficient numerical simulation of tunnel ventilation systems to predict airflow organization, temperature, and smoke distribution under various operating conditions is of great engineering significance. Currently, numerical simulation methods for tunnel ventilation primarily rely on computational fluid dynamics (CFD). Three-dimensional CFD simulation, by solving the Navier-Stokes equations, provides the highest-fidelity description of the detailed flow field within tunnels and is widely recognized as the most accurate analytical tool. However, its application is severely limited by inherent limitations. Due to the large scale of tunnel structures and the highly complex flow, a complete 3D CFD simulation often requires enormous computing resources and lengthy computation cycles, making it difficult to apply to scenarios requiring rapid emergency decision support or ventilation scheme design involving extensive parameter optimization.

[0003] To overcome the efficiency bottleneck of three-dimensional CFD simulation, the engineering field also widely uses one-dimensional pipe network models for simplified calculations. This method abstracts the tunnel system into a network consisting of pipe segments and nodes, and the calculation speed is extremely fast. However, its simulation accuracy is heavily dependent on the empirical selection of local resistance coefficients in complex connection areas (such as branch and confluence points). For areas with significant nonlinear characteristics such as the connection between the side guide tunnel and the main tunnel involved in the present invention, a simple one-dimensional model cannot accurately capture the complex three-dimensional flow details such as separation and vortexes. This will lead to large deviations in the prediction of the system's air volume distribution and pollutant diffusion, thereby affecting the reliability of the ventilation system design and the effectiveness of the emergency control strategy.

[0004] Therefore, existing tunnel ventilation simulation technologies generally face an irreconcilable contradiction between computational accuracy and efficiency. On the one hand, high-precision three-dimensional simulation methods lack practicality due to their high time cost; on the other hand, highly efficient one-dimensional simplified models lack accuracy and are unreliable, especially in fire conditions, which place extremely stringent safety requirements. This technological status quo not only limits the ability to quickly and accurately evaluate the performance of complex tunnel ventilation systems, but also hinders the development and application of intelligent optimal control strategies based on precise simulations. As a result, current emergency control of tunnel ventilation systems relies heavily on preset plans or manual experience, lacking the ability to adaptively optimize for unexpected situations. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a ventilation simulation calculation method for a side guide tunnel of a non-slanted vertical shaft tunnel combined with dual main tunnels, which solves the contradiction between the low efficiency of high-precision three-dimensional simulation calculation and the insufficient calculation accuracy of the one-dimensional pipe network model in the existing tunnel ventilation simulation technology, as well as the resulting difficulty in quickly and reliably simulating sudden accidents and formulating optimal control strategies.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a ventilation simulation calculation method for a non-slanted vertical shaft tunnel side guide tunnel combined with dual main tunnels, comprising the following steps: Step 1: Decompose the three-dimensional model of the tunnel system into a first type of region and a second type of region, wherein the first type of region is a linear flow region with uniform flow field, and the second type of region is a nonlinear node region with complex three-dimensional flow; Step 2: For the second type of region, by obtaining high-fidelity simulation data under various boundary conditions, a mathematical proxy model is established to characterize the mapping relationship between its input and output boundary states; Step 3: For specific ventilation simulation tasks, based on the decomposition results of step 1, construct and solve the global system equations, where the physical behavior of the first type of area is described by a one-dimensional pipe network model, and the physical behavior of the second type of area is described by the mathematical proxy model established in step 2.

[0007] Preferably, the step of performing topological decomposition of the tunnel system specifically includes abstracting each of the nonlinear node areas into a node of a network topology graph, and abstracting each of the linear flow areas connecting the nonlinear node areas into an edge connecting the nodes, thereby generating a network topology graph that represents the physical connection relationship of the tunnel system.

[0008] Preferably, the step of drawing the network topology diagram also includes assigning physical property parameters required by the corresponding linear flow area in the one-dimensional pipe network model to each edge of the network topology diagram, and the physical property parameters include the pipe section length, hydraulic diameter and friction coefficient along the linear flow area.

[0009] Preferably, the steps of the mathematical proxy model are specifically as follows: performing multiple three-dimensional computational fluid dynamics simulations to obtain the high-fidelity simulation data, and using a machine learning algorithm for training based on the high-fidelity simulation data to obtain the mathematical proxy model.

[0010] Preferably, the mathematical agent model outputs a prediction result for characterizing its output boundary state according to its input boundary conditions, and also outputs a prediction confidence index for characterizing the credibility of the prediction result.

[0011] Preferably, the step of constructing and solving the global system equations further includes an adaptive correction step, which is triggered based on a prediction confidence indicator output by the mathematical agent model.

[0012] Preferably, the adaptive correction step is specifically as follows: in the process of solving the global system equations, the prediction confidence index is monitored in real time, and when the prediction confidence index is lower than a preset threshold, a three-dimensional computational fluid dynamics simulation is performed online for the corresponding nonlinear node area to obtain a high-precision online simulation result, and the mathematical agent model is updated online using the online simulation result.

[0013] Preferably, the method further comprises: constructing a topological state graph, wherein the nodes of the topological state graph represent the macroscopic states of the tunnel system, and the edges of the topological state graph represent control operations that can cause transitions between the macroscopic states.

[0014] Preferably, the method also includes: for a preset control target, for at least part of the edges in the topological state graph, constructing and solving the global system equations to determine the system evolution result caused by executing the control operation represented by the edge, and calculating the weight of the edge based on the system evolution result.

[0015] Preferably, the method further includes: on the topological state graph, based on the weight of the edge, using a graph search algorithm to search for a path with the optimal cumulative weight from the starting state node to the target state node, and outputting the control operation sequence corresponding to the optimal path.

[0016] The present invention provides a ventilation simulation calculation method for a non-slanted vertical shaft tunnel with a side guide tunnel and dual main tunnels. It has the following beneficial effects: 1. This method decomposes the tunnel system into linear flow regions and nonlinear node regions, describing them using a one-dimensional pipe network model and a pre-trained mathematical proxy model. Ultimately, a mixed-fidelity global system equation set is constructed and solved. This approach significantly improves the computational efficiency of full-system ventilation simulation while accurately capturing the physical behavior of key complex regions. This approach avoids the time-consuming and expensive three-dimensional computational fluid dynamics simulation of the entire tunnel system, achieving both high efficiency and high precision in simulation computation.

[0017] 2. This invention achieves adaptability and high reliability in simulation calculations by introducing an adaptive correction step based on the prediction confidence index of a mathematical surrogate model during the solution of the global system equations. This method monitors the predictive reliability of the surrogate model in real time under unknown operating conditions. When its reliability is insufficient, it uses high-precision simulation on demand and online for targeted data supplementation and model updates. This ensures that the entire simulation framework can output stable and reliable calculation results even when facing rare or extreme conditions not covered by the training data.

[0018] 3. This invention constructs a topological state map that characterizes the system's macroscopic state and control operations. Leveraging the aforementioned efficient and precise simulation computing capabilities, this method dynamically assigns cost weights to the control operations in the map based on physical evolution results. Finally, it combines this with a graph search algorithm to deduce optimal control paths, elevating traditional simulation tools into intelligent decision support systems. This method automatically and rapidly generates a quantitatively evaluated optimal control operation sequence for specific control objectives, achieving intelligent and optimized ventilation control strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the present specification. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] Please see the attached Figure 1An embodiment of the present invention provides a ventilation simulation calculation method for a side guide tunnel combined with dual main tunnels in a non-slanted vertical shaft tunnel. The method decomposes the tunnel system into a linear flow area and a nonlinear node area, and describes them respectively using a one-dimensional pipe network model and a pre-trained mathematical proxy model. Finally, a mixed-fidelity global system equation group is constructed and solved. While ensuring the accurate capture of the physical behavior of key complex areas, the computational efficiency of the ventilation simulation of the entire system is greatly improved.

[0022] like Figure 1 As shown, the ventilation simulation calculation method for the side guide tunnel of a non-slanted vertical shaft tunnel combined with dual main tunnels may include the following steps: Step 1, decomposing the three-dimensional model of the tunnel system into a first type of area and a second type of area, wherein the first type of area is a linear flow area with uniform flow field, and the second type of area is a nonlinear node area with complex three-dimensional flow; Step 2: For the second type of region, by obtaining high-fidelity simulation data under various boundary conditions, a mathematical proxy model is established to characterize the mapping relationship between its input and output boundary states; Step 3: For specific ventilation simulation tasks, based on the decomposition results of step 1, construct and solve the global system equations. The physical behavior of the first type of area is described by the one-dimensional pipe network model, and the physical behavior of the second type of area is described by the mathematical proxy model established in step 2.

[0023] The method begins with a detailed topological decomposition of the tunnel system consisting of a non-slanted vertical shaft tunnel with side guides and dual main tunnels. This step serves as a foundation for the subsequent construction of a hybrid-fidelity simulation model that balances computational accuracy and efficiency. In tunnel engineering, the complexity of fluid motion varies significantly across different regions, making it difficult to balance accuracy and cost using a unified computational model. Therefore, this step begins by identifying and classifying the tunnel system based on its physical characteristics.

[0024] Specifically, this segmentation process is based on an analysis of the complete 3D geometry of the tunnel system and its inherent fluid dynamics. This analysis is based on factors such as geometric curvature, cross-sectional change rates, the presence of diverging and converging flows, and the presence of interference from external power sources (such as fans). Based on these factors, the entire tunnel system is physically divided into two distinct regions: The first type of area is the linear flow area. The typical characteristics of this type of area are that the internal flow field is relatively fully and uniformly developed, the flow direction is basically parallel to the tunnel axis, and the three-dimensional effect is not significant. In this type of area, the energy loss of the fluid is mainly dominated by the wall friction along the way, and its physical behavior can be described accurately enough by one-dimensional macroscopic parameters, such as the average flow velocity and average pressure of the section. In the specific scenario of the present invention, the linear flow area usually refers to a long straight tunnel section away from complex structures such as intersections and connections, such as the middle straight section of a double main tunnel and the straight section of a side guide tunnel.

[0025] The second type of area is the nonlinear node area. This type of area is in sharp contrast to the linear flow area, and there are strong and non-negligible complex three-dimensional flow phenomena inside it. These phenomena include but are not limited to the separation and reattachment of airflow, the formation of secondary flow, strong vortex motion and significant energy and momentum exchange. These complex physical processes make the simple one-dimensional model completely invalid. In the tunnel structure targeted by the present invention, the nonlinear node area clearly includes the connection between the side guide tunnel and the main tunnel, the T-shaped or cross-shaped connection between the transverse connecting channel and the main tunnel, which are the core research objects, and the affected section where the jet fan group is located to meet the ventilation needs. Under accident conditions such as fire, the fire source itself and the thermal buoyancy-dominated area near it should also be regarded as a dynamically formed nonlinear node area.

[0026] After completing the physical division of the tunnel system, to facilitate subsequent system-level numerical solutions, this step also includes a key topological abstraction process. This process converts the aforementioned physical division results into a structured mathematical expression.

[0027] Specifically, the topology abstraction step abstracts each identified nonlinear node region into a node in a network topology graph. These nodes represent "passages" in the graph where physical behavior is complex and requires focused analysis.

[0028] Accordingly, each linear flow region connecting two nonlinear node regions is abstracted as an edge connecting the two corresponding nodes in the graph. These edges represent the "channels" for macroscopic fluid transport in the graph.

[0029] Through this abstraction process, a complex, continuous three-dimensional tunnel entity is transformed and simplified into a discretized network topology consisting of nodes and edges. Preferably, this network topology can be expressed mathematically as where is the node set consisting of all nonlinear node regions, and is the edge set consisting of all linear flow regions. This transformation greatly simplifies the complexity of the problem, making it possible to subsequently construct and solve a global set of equations.

[0030] It should be further explained that the generated network topology not only contains the connection relationship, but also must carry the physical information required for subsequent calculations. Therefore, this step also includes assigning physical attribute parameters to the network topology.

[0031] Specifically, for each edge in the network topology, that is, each linear flow area, a set of physical property parameters required in the one-dimensional pipe network model must be calculated and assigned. These parameters are the direct basis for describing the physical behavior of the area in subsequent steps. In this embodiment, the physical property parameters include at least: Pipe section length: the actual length of the linear flow area along its center line.

[0032] Hydraulic diameter: An equivalent diameter calculated based on the cross-sectional geometry of the linear flow area, used for calculating one-dimensional flow losses.

[0033] Friction coefficient along the tunnel: a dimensionless coefficient that characterizes the frictional resistance encountered by the fluid when flowing through the inner wall of the linear flow zone. Its value is related to the roughness of the tunnel wall and the Reynolds number.

[0034] The core task of step 2 of the present method is to construct a computationally lightweight mathematical proxy model for each type of nonlinear node region identified in step 1, accurately replicating its complex physical behavior. This step is key to achieving both high efficiency and high precision in the present method. Directly and frequently calling a three-dimensional computational fluid dynamics (CFD) model in system-level simulations is impractical. This mathematical proxy model serves as a high-fidelity "digital twin" of the CFD model, integrated into subsequent calculations.

[0035] The specific implementation of this step can be further decomposed into the process of acquiring high-fidelity simulation data and the process of data-based model training.

[0036] First, high-fidelity simulation data is acquired. For a specific type of nonlinear node region, such as a three-port side-guide-main-hole T-junction, it is necessary to systematically explore the intrinsic mapping relationship between its input and output boundary states.

[0037] The first step in the process is to define parameters. The fluid state parameters of all physical ports in the nonlinear node area, such as static pressure , flow rate ,temperature and the concentration of key components , defined as a boundary state vector. Some of the parameters are used as input vectors , and the other part is used as the output vector Next, in order to efficiently explore the entire working parameter space of the nonlinear node region, preferably, a Design of Experiments (DoE) method can be adopted, such as Latin Hypercube Sampling (LHS), to generate a set of representative and evenly covered input boundary condition sample points within a reasonable range of its input parameters. For each generated input sample point ,This method will perform a complete and high-precision three-dimensional computational fluid dynamics simulation. By solving the Navier-Stokes equations, the stable three-dimensional flow field distribution inside the nonlinear node area under the input conditions can be obtained, and the corresponding output boundary state vector can be extracted from it. After completing the calculation of all sample points, a high-fidelity simulation database consisting of a large number of "input-output" data pairs can be obtained, which can fully characterize the physical characteristics of the nonlinear node area. .

[0038] Then, the process of model training is based on the acquired data. The goal of this process is to use machine learning algorithms to obtain high-fidelity data points from discrete Learn and fit a continuous mathematical function with strong generalization ability, namely the mathematical agent model , making .

[0039] Preferably, in this embodiment, the Gaussian Process Regression (GPR) algorithm can be used to construct the mathematical proxy model. GPR is chosen due to its unique, non-parametric modeling capabilities based on Bayesian theory. It not only provides predictions of output results but also quantitatively assesses the uncertainty of the predictions, which is crucial for the subsequent adaptive corrections of the present invention.

[0040] Specifically, when given a new input boundary condition outside the database When , the trained Gaussian process regression model can simultaneously output two types of information: First, it is used to characterize the prediction result of its output boundary state, that is, the predicted mean The prediction result is the real physical output The best estimate of can be directly used in subsequent system-level simulation calculations.

[0041] The second is the prediction confidence index used to characterize the credibility of the prediction result, namely the prediction variance This indicator quantifies the model's confidence in the accuracy of the prediction. When the input is close to the sample point in the training database, the prediction variance is small, indicating that the prediction result is highly reliable. On the contrary, when the input is close to the sample point in the training database, the prediction variance is small, indicating that the prediction result is highly reliable. When we are far away from all known data points and in the model’s “cognitive blind spot”, the prediction variance will increase significantly, thus sending a warning signal that the reliability of the current prediction result is low.

[0042] Through the above process, this step equips each type of nonlinear nodal region with a dedicated, fully trained mathematical proxy model. These models provide prediction accuracy similar to that of 3D CFD simulations at a very low computational cost, and creatively add the ability to self-diagnose the reliability of their predictions. This provides the core technical support for building a robust and efficient adaptive mixed-fidelity solution framework in Step 3.

[0043] Step 3 of the method is the core computational step in performing the specific ventilation simulation task. Rather than using a single-scale model, this step constructs and solves an innovative, multi-fidelity coupled global system equations based on the topological decomposition results of Step 1 and the mathematical model library constructed in Step 2. This design aims to organically combine the computational efficiency of a one-dimensional model with the physical accuracy of a three-dimensional model.

[0044] First, the global system equations are constructed. For a given ventilation simulation task, such as smoke diffusion analysis under a specific fire scenario, this method automatically assembles a large set of nonlinear equations describing the steady-state or transient behavior of the entire tunnel system based on the network topology generated in step 1, which carries the physical properties. The "mixed fidelity" characteristic of this system of equations is reflected in the fact that different mathematical descriptions are used for different regions: For each edge in the network topology, that is, each tunnel segment divided into a linear flow region, its internal physical behavior is described by the classic one-dimensional pipe network model. Specifically, the relationship between its pressure drop and flow rate can be characterized by the Darcy-Weisbach equation: ; in, is the pressure difference at both ends of the pipe section, and the pipe section length in the formula , hydraulic diameter and friction coefficient along the way The physical property parameters such as , are exactly what are assigned to the edge in step 1. For each node in the network topology, that is, each complex connection identified as a nonlinear node area, its physical behavior is determined by the mathematical agent model that is tailored for it and pre-trained in step 2. The model is described in the form of a function The nonlinear exchange relationship between physical quantities such as pressure, flow, energy and component concentration among all ports of the node is given, thus replacing the calculation results that originally had to be obtained through expensive three-dimensional CFD simulation. In addition, in order to ensure the physical conservation of the entire system, the constraint equations of mass conservation, momentum conservation and energy conservation must be established and satisfied at the interface between each nonlinear node area and the linear flow area connected to it. Combining all the above equations for edges, nodes and interfaces forms a global system equation group that fully describes the behavior of the entire tunnel system. .in, is a global state vector that contains all unknown quantities to be solved (e.g., pressure at each node, flow rate at each pipe section, etc.) Preferably, a robust numerical iterative method such as the Newton-Raphson method can be used to solve the large nonlinear system of equations.

[0045] However, a key innovation of this step lies not only in the aforementioned mixed-fidelity modeling but also in its inclusion of an intelligent adaptive correction step. This step is designed to address the potential loss of accuracy in mathematical surrogate models for unknown operating conditions outside the training database, thereby ensuring the robustness and reliability of the entire simulation process.

[0046] The operating mechanism of this adaptive correction step is deeply coupled with the iterative solution process of the global system equations, and its triggering is entirely based on the prediction confidence index output by the mathematical agent model itself in step 2. The specific implementation process is as follows: First, in each iteration of solving the global system equations, when the system calls the mathematical agent model of any nonlinear node area To calculate its output When the system not only obtains its prediction results , and also synchronously obtains its confidence index for this prediction, such as the prediction variance given by the Gaussian process regression model . This is a real-time monitoring process.

[0047] Next, the system compares the prediction confidence index with a preset threshold that can be defined by the user. When the prediction confidence index of a nonlinear node area is found to be lower than the preset threshold, for example threshold, which constitutes a clear trigger signal. This signal indicates that the current input working condition This is a weak area in the cognitive ability of the agent model, and the reliability of its prediction results is not enough to support subsequent calculations.

[0048] Once the correction step is triggered, the system will immediately execute an online update process. This process will first suspend the iterative solution of the main program. Then, for the nonlinear node area that issued the alarm, the current input boundary conditions that lead to low confidence prediction are updated. As input, a complete 3D computational fluid dynamics simulation is performed online and on demand. This simulation will obtain a high-precision online simulation result for the current unknown working conditions. . Subsequently, this newly acquired high-fidelity data point It will be used as a new sample for online updating or retraining of the mathematical agent model, thereby improving its cognitive level and prediction accuracy in this unknown area.

[0049] After the mathematical agent model completes the online update, the iterative solution process of the main program will resume from the pause point and continue calculations using this "targeted enhanced" and more reliable model until the entire global system equations converge.

[0050] Through this closed-loop adaptive correction mechanism of monitoring, triggering, online simulation and model updating, the method of the present invention ensures the stability of the calculation process and the accuracy of the final result. It can intelligently identify and compensate for the cognitive shortcomings of the proxy model, thereby maximizing the advantages of mixed-fidelity simulation.

[0051] First, this step requires constructing a topological state graph, which differs from the physical network topology graph in step 1 and has a higher dimensionality. This graph is a decision graph, whose nodes represent the macro-state of the entire tunnel system. A macro-state is defined by two pieces of information: the state of physical events (e.g., the specific location of the fire source in the tunnel, combustion power, etc.), and the state of controllable devices (e.g., the on / off / adjustment status of jet fans, axial fans, and dampers). The edges of this graph represent atomic control operations that can cause transitions between macro-states, such as "turn on jet fan unit 1" or "adjust the speed of the smoke exhaust fan in main tunnel 2 to 80%."

[0052] Given a predefined control objective (for example, in a fire scenario, reducing smoke concentration in an evacuation corridor to below a safe threshold within a specified timeframe with minimal energy consumption), this method transforms this complex decision-making problem into a shortest path search problem within the topological state graph. The core of this process lies in dynamically and physically meaningfully weighting the edges in the graph.

[0053] Specifically, for a specific edge in the topological state graph to be evaluated (i.e., a potential control action), the system uses the system state represented by the starting node of that edge as the initial condition and fully invokes and executes the mixed-fidelity ventilation simulation method described in step three. By constructing and solving the global system equations, the system can efficiently predict the next stable state to which the tunnel system will evolve after executing the control action, thereby obtaining the system evolution result caused by the action.

[0054] The system then quantitatively evaluates the system's evolutionary results based on a pre-set cost function (which integrates various control objectives such as personnel safety, smoke exhaust efficiency, control time, and equipment energy consumption) and calculates a scalar value. This value is then assigned to the edge's weight.

[0055] After assigning weights to the relevant edges in the topological state graph, this method preferably adopts an efficient graph search algorithm such as A* (A-star) to search for a path with the optimal cumulative weight from the starting node to the target node on the dynamically weighted graph, with the initial state of the accident as the starting node and the state that meets the control target as the target node.

[0056] This optimal path clearly indicates a series of sequential, cost-minimizing control steps. Ultimately, the method outputs the corresponding control operation sequence for this optimal path, providing clear, reliable, and quantitatively evaluated intelligent decision support for tunnel operators.

[0057] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A ventilation simulation calculation method for a non-slanted vertical shaft tunnel with a side guide tunnel and dual main tunnels, characterized in that: The following steps are involved: Step 1: Decompose the three-dimensional model of the tunnel system into a first type of region and a second type of region, wherein the first type of region is a linear flow region with uniform flow field, and the second type of region is a nonlinear node region with complex three-dimensional flow; Step 2: For the second type of region, by obtaining high-fidelity simulation data under various boundary conditions, a mathematical proxy model is established to characterize the mapping relationship between its input and output boundary states; Step 3: For specific ventilation simulation tasks, based on the decomposition results of step 1, construct and solve the global system equations, where the physical behavior of the first type of area is described by a one-dimensional pipe network model, and the physical behavior of the second type of area is described by the mathematical proxy model established in step 2.

2. A ventilation simulation calculation method for a non-slanted vertical shaft tunnel side guide tunnel combined with dual main tunnels according to claim 1, characterized in that: The step of topological decomposition of the tunnel system specifically includes abstracting each of the nonlinear node areas into a node of a network topology graph, and abstracting each of the linear flow areas connecting the nonlinear node areas into an edge connecting the nodes, thereby generating a network topology graph representing the physical connection relationship of the tunnel system.

3. The method for simulating ventilation of a non-slanted vertical shaft tunnel with a side guide tunnel and dual main tunnels according to claim 2 is characterized in that: The step of drawing the network topology diagram also includes assigning physical property parameters required by the corresponding linear flow area in the one-dimensional pipe network model to each edge of the network topology diagram, and the physical property parameters include the pipe section length, hydraulic diameter and friction coefficient along the linear flow area.

4. The method for simulating ventilation of a non-slanted vertical shaft tunnel with a side guide tunnel and dual main tunnels according to claim 1 is characterized in that: The steps of the mathematical proxy model are specifically as follows: performing multiple three-dimensional computational fluid dynamics simulations to obtain the high-fidelity simulation data, and using a machine learning algorithm to train based on the high-fidelity simulation data to obtain the mathematical proxy model.

5. The method for simulating ventilation of a non-slanted vertical shaft tunnel with a side guide tunnel and dual main tunnels according to claim 1 is characterized in that: The mathematical agent model outputs a prediction result for characterizing its output boundary state according to its input boundary conditions, and also outputs a prediction confidence index for characterizing the credibility of the prediction result.

6. The method for simulating ventilation of a non-slanted vertical shaft tunnel with a side guide tunnel and dual main tunnels according to claim 1 is characterized in that: The step of constructing and solving the global system equations further includes an adaptive correction step, which is triggered based on the prediction confidence index output by the mathematical agent model.

7. A ventilation simulation calculation method for a non-slanted vertical shaft tunnel side guide tunnel combined with dual main tunnels according to claim 6, characterized in that: The adaptive correction step specifically includes: monitoring the prediction confidence index in real time during the process of solving the global system equations, and when the prediction confidence index is lower than a preset threshold, performing a three-dimensional computational fluid dynamics simulation online for the corresponding nonlinear node area to obtain a high-precision online simulation result, and using the online simulation result to update the mathematical agent model online.

8. The method for simulating ventilation of a non-slanted vertical shaft tunnel with a side guide tunnel and dual main tunnels according to claim 1 is characterized in that: The method further includes: constructing a topological state graph, wherein the nodes of the topological state graph represent macroscopic states of the tunnel system, and the edges of the topological state graph represent control operations that can cause transitions between the macroscopic states.

9. The method for simulating ventilation of a non-slanted vertical shaft tunnel with a side guide tunnel and dual main tunnels according to claim 1 is characterized in that: The method also includes: for a preset control target, for at least part of the edges in the topological state graph, constructing and solving the global system equations to determine the system evolution result caused by executing the control operation represented by the edge, and calculating the weight of the edge based on the system evolution result.

10. A ventilation simulation calculation method for a non-slanted vertical shaft tunnel side guide tunnel combined with dual main tunnels according to claim 1, characterized in that: The method further includes: searching for a path with the optimal cumulative weight from the starting state node to the target state node on the topological state graph based on the weight of the edge using a graph search algorithm, and outputting a control operation sequence corresponding to the optimal path.

Images