An escape path planning method based on flow field global modal characteristics

Abstract

The application discloses a kind of escape path planning methods based on flow field global modal characteristics, comprising: the establishment of computational model is carried out to leakage scene, grid division, boundary condition setting, calculation method is determined, non-constant value calculation.The eigenvalue orthogonal decomposition is carried out to the characteristic section of chlorine diffusion field, and the mode is sorted according to the energy proportion, and the first-order mode with the maximum energy proportion is determined as the dominant mode.The escape path node grid file is established, and the dominant mode amplitude is projected onto the grid node.According to the maximum value of mode amplitude and corresponding time coefficient, the risk coefficient partition is determined, and the equivalent path length is calculated using the risk coefficient.According to the equivalent path length, the escape path is planned based on Dijkstra algorithm.The escape path obtained by the application is more universal in the whole leakage process.

Description

Technical Field

[0001] This invention belongs to the field of path optimization, and in particular relates to an escape path planning method based on global modal characteristics of the flow field. Background Technology

[0002] Following a chlorine gas leak, a well-planned escape route is crucial for protecting public safety. Currently, Dijkstra's algorithm is a commonly used escape route planning method. Unlike traditional methods, in a chlorine gas leak, the criterion for selecting the optimal escape route is not the actual length of the path, but rather the degree of chlorine gas damage suffered by individuals while traversing that path. Most researchers use equivalent path length to quantify the chlorine gas damage inflicted on individuals when traversing an escape route. The equivalent path length generally requires determining the equivalent coefficient based on the chlorine gas concentration. Traditional methods for obtaining chlorine gas concentration primarily rely on two approaches: traditional empirical diffusion models such as Gaussian models, and unsteady numerical calculation methods. Empirical diffusion models cannot account for actual leak scenarios and have poor prediction accuracy. Unsteady numerical calculation methods can obtain a more detailed diffusion field, but researchers often only select a single instantaneous diffusion field for path planning, resulting in weak representativeness and an inability to consider the entire leak process. Therefore, it is necessary to establish an escape path planning method based on the global modal characteristics of the flow field. First, the unsteady diffusion field is modally decomposed to extract its dominant diffusion structure, and then the equivalent path length is calculated and the optimal path is planned based on the structure. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention provides an escape path planning method based on global modal characteristics of the flow field, comprising:

[0004] A three-dimensional computational model of the leakage scenario was established and unsteady numerical calculations were performed to obtain time-series data of the leakage diffusion field.

[0005] The time series data is subjected to intrinsic orthogonal decomposition to obtain spatial modes sorted by energy proportion;

[0006] The spatial mode with the largest energy proportion is selected as the dominant mode, and the dominant mode is mapped to the escape path node grid to obtain the modal amplitude of each node.

[0007] The risk coefficient of each node is calculated based on the modal amplitude and the maximum value of the corresponding time coefficient, and the actual path length is converted into the equivalent path length using the risk coefficient.

[0008] Using the equivalent path length as the weight, an escape route from the starting point to the safe exit is planned based on the shortest path algorithm.

[0009] Preferably, the process of establishing a three-dimensional computational model of the leakage scenario and performing unsteady numerical calculations to obtain time-series data of the leakage diffusion field includes:

[0010] A three-dimensional geometric model was constructed based on the location of the leak source, fixed obstacles, and terrain environment.

[0011] The three-dimensional geometric model is meshed, and local meshing is performed in the near-ground and obstacle areas;

[0012] Set boundary conditions, where the inlet is the velocity inlet, the outlet is the pressure outlet, the solid wall is the no-slip boundary, and the leakage outlet is the velocity inlet;

[0013] First, set the leak outlet velocity to 0 and perform steady-state calculations to obtain the initial wind field. Then, using the initial wind field as the initial field, set the leak outlet velocity and perform unsteady-state calculations, and output the chlorine diffusion field data on the characteristic surface with a fixed time step.

[0014] Preferably, the process of performing intrinsic orthogonal decomposition on the time-series data to obtain spatial modes sorted by energy proportion includes:

[0015] Construct a column vector from the chlorine mass fraction data on the feature surface at each time step in chronological order;

[0016] The column vectors at all times are used to form a chlorine mass fraction matrix;

[0017] Calculate the time correlation matrix based on the chlorine mass fraction matrix;

[0018] The time correlation matrix is ​​subjected to eigenvalue decomposition to obtain eigenvalues ​​and corresponding eigenvectors;

[0019] The spatial modes of each order are calculated using the eigenvectors and the chlorine mass fraction matrix.

[0020] Preferably, the process of selecting the spatial mode with the largest energy proportion as the dominant mode includes:

[0021] Sort the eigenvalues ​​corresponding to each spatial mode by size;

[0022] The first spatial mode with the largest energy proportion is determined as the dominant mode.

[0023] Preferably, the process of mapping the dominant mode to the escape path node mesh to obtain the modal amplitude of each node includes:

[0024] An escape route node grid file is generated based on the intersections of roads and areas where people gather in the factory area. Each node contains a number and two-dimensional coordinates.

[0025] Establish the shortest distance correspondence between escape path node numbers and feature surface location unit numbers;

[0026] Based on the shortest distance correspondence, the modal amplitude of the dominant mode is assigned to the corresponding escape path node.

[0027] Preferably, the process of calculating the risk factor of each node based on the modal amplitude includes:

[0028] Obtain the modal amplitude of the dominant mode at each node, and use the product of the modal amplitude and the maximum value of the corresponding time coefficient as an indicator. Convert the indicator into a risk factor according to a preset threshold.

[0029] Preferably, the process of converting the actual path length into an equivalent path length using the risk factor includes:

[0030] Obtain the actual length of each segment in the path, and weight the risk coefficients of the two ends of each segment to obtain the road risk coefficient weight of the corresponding segment;

[0031] Multiply the road hazard coefficient weight by the corresponding actual length to obtain the equivalent path length of the corresponding segment.

[0032] Preferably, the process of planning an escape route from the starting point to the safe exit based on the shortest path algorithm, using the equivalent path length as the weight, includes:

[0033] Construct a weighted directed graph with the escape starting point as the source point and the safety exit as the destination point, using the equivalent path length as the edge.

[0034] Dijkstra's algorithm is used to search for the shortest path in the weighted directed graph;

[0035] If there are multiple safety exits, calculate the shortest equivalent path length to each exit and select the path corresponding to the minimum value as the final escape route.

[0036] Compared with the prior art, the present invention has the following advantages and technical effects:

[0037] This invention first performs modal decomposition on the unsteady chlorine gas diffusion field, selects the dominant mode, and then performs path planning based on the dominant mode. Considering that the leakage diffusion field evolves continuously over time, the proposed modal decomposition method can achieve spatiotemporal decoupling of the diffusion field in the entire unsteady leakage process; that is, the mode is a spatial diffusion field whose spatial structure does not change over time. In this case, the dominant mode can characterize the dominant diffusion structure of the entire leakage process, and path planning based on the dominant mode has greater universality for the entire leakage process.

[0038] In the path planning process, this invention replaces the path length in Dijkstra's algorithm with an equivalent path length. This equivalent length is calculated by multiplying the hazard factor by the actual path length, taking into account the different effects of chlorine concentration on the human body. The hazard factor is defined as the maximum value of the modal amplitude multiplied by the corresponding time coefficient, thus fully considering the harm of chlorine to the human body in the most dangerous situation.

[0039] This invention achieves continuous capture of the diffusion field throughout the entire leakage process through a step-by-step numerical calculation strategy of "steady first, then unsteady". This avoids the initial disturbance error caused by a one-time unsteady calculation and ensures the integrity and coherence of subsequent transient data, providing high-fidelity time series samples for subsequent mode decomposition.

[0040] This invention employs a "local densification + boundary layer" mesh generation technique to match the mesh scale of the near-wall surface and the windward and leeward areas of the building with the chlorine concentration gradient; this saves computational resources and ensures that the concentration field accuracy reaches the ppm-level resolution required for path planning.

[0041] This invention introduces "snapshot POD" to decouple the global diffusion field at different times in time and compress high-dimensional transient data into energy-ordered modes. More than 80% of the energy can be captured using only the first dominant mode, which significantly reduces the data dimensionality. At the same time, it retains the master space structure of the entire leakage process and avoids the problem of "instantaneous non-representativeness" in the traditional single-frame method.

[0042] This invention defines a method for quantifying hazard factors, transforming a continuously changing concentration field into a stable hazard zoning index, assigning a unique and reproducible hazard weight to each road, and eliminating the path oscillation caused by differences in instantaneous concentrations in traditional methods. Attached Figure Description

[0043] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0044] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention;

[0045] Figure 2 This is a schematic diagram of a chlorine leak storage tank according to an embodiment of the present invention;

[0046] Figure 3 This is a schematic diagram of a chlorine gas leak scenario according to an embodiment of the present invention;

[0047] Figure 4 This is a schematic diagram of the computational grid according to an embodiment of the present invention;

[0048] Figure 5 This is a modal energy distribution diagram according to an embodiment of the present invention;

[0049] Figure 6 This is a diagram illustrating the dominant modal structure of an embodiment of the present invention.

[0050] Figure 7 This is a diagram showing the optimal escape path for a certain node in an embodiment of the present invention.

[0051] Figure 8 This is an escape path diagram for all nodes in this embodiment of the invention. Detailed Implementation

[0052] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0053] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0054] like Figure 1 As shown, this embodiment provides an escape path planning method based on global modal characteristics of the flow field, including:

[0055] A three-dimensional computational model of the leakage scenario was established and unsteady numerical calculations were performed to obtain time-series data of the leakage diffusion field.

[0056] The time series data are subjected to intrinsic orthogonal decomposition to obtain spatial modes sorted by energy proportion;

[0057] The spatial mode with the largest energy proportion is selected as the dominant mode, and the dominant mode is mapped to the escape path node grid to obtain the modal amplitude of each node;

[0058] Calculate the risk factor of each node based on the modal amplitude and the maximum value of the corresponding time coefficient, and use the risk factor to convert the actual path length into the equivalent path length.

[0059] Using the equivalent path length as the weight, an escape route from the starting point to the safe exit is planned based on the shortest path algorithm.

[0060] Furthermore, the process of establishing a three-dimensional computational model of the leakage scenario and performing unsteady numerical calculations to obtain time-series data of the leakage diffusion field includes:

[0061] A three-dimensional geometric model was constructed based on the location of the leak source, fixed obstacles, and terrain environment.

[0062] The 3D geometric model is meshed, and local meshing is performed in the near-ground and obstacle areas.

[0063] Set boundary conditions, where the inlet is the velocity inlet, the outlet is the pressure outlet, the solid wall is the no-slip boundary, and the leakage outlet is the velocity inlet;

[0064] First, set the leak outlet velocity to 0 and perform steady-state calculations to obtain the initial wind field. Then, using the initial wind field as the initial field, set the leak outlet velocity and perform unsteady-state calculations, and output the chlorine diffusion field data on the characteristic surface with a fixed time step.

[0065] Specifically, a three-dimensional model is established for the chlorine gas leak scenario. The model includes: the leak location, fixed buildings and obstacles in the plant area, and the terrain environment.

[0066] The numerical calculation process for the chlorine leakage process is as follows: A suitable computational grid is generated using a polyhedral grid. Local mesh refinement is applied in the building area and near the ground. A boundary layer is added. The first mesh layer thickness is 0.00668503 m, the growth rate is set to 1.3, and the boundary layer thickness is approximately 1.9 m. After mesh independence analysis, the number of meshes is selected as 6,067,873, and the average mesh quality is 0.9491357. The Realizable k-ε model is used for the numerical turbulence model, and the chlorine transport calculation uses the SpeciesModel.

[0067] Boundary conditions: The ambient inlet is an air velocity inlet V. 环境 The outlet is a pressure outlet, the ground, building, and obstacle surfaces are non-slip boundary conditions, and the chlorine leak outlet is set as a velocity inlet.

[0068] The calculation process is as follows: First, the chlorine leakage velocity is set to 0. A steady-state numerical calculation method is used to calculate the environmental wind field. After convergence, this result is used as the initial field. The chlorine leakage velocity is then set, and an unsteady calculation is performed with a time step of 0.01 s. The total unsteady time step is 50000. A feature surface is established: a feature surface parallel to the ground (the height of the human breathing zone) is established at a height of 1.5 m (y = 1.5) above the ground. During the calculation, the chlorine diffusion field of one feature surface is output every second (including location cell number, X / Y / Z coordinate information, and chlorine mass fraction), for a total of n = 500 time points.

[0069] This embodiment achieves continuous capture of the diffusion field throughout the entire leakage process through a step-by-step numerical calculation strategy of "steady first, then unsteady". This avoids the initial disturbance error caused by a one-time unsteady calculation and ensures the integrity and coherence of the subsequent 50,000 steps of transient data, providing high-fidelity time series samples for subsequent mode decomposition.

[0070] This embodiment employs a "local densification + boundary layer" mesh generation technique. The parameter combination of a first layer thickness of 0.00668503m and a growth rate of 1.3 ensures that the mesh scale near the wall and the windward and leeward areas of the building matches the chlorine concentration gradient. Millimeter-level concentration gradients can be resolved with a mesh size of 6,067,873, which saves computational resources and ensures that the concentration field accuracy reaches the ppm-level resolution required for path planning.

[0071] Furthermore, the process of performing intrinsic orthogonal decomposition on the time series data to obtain spatial modes sorted by energy proportion includes:

[0072] Construct a column vector from the chlorine mass fraction data on the feature surface at each time step in chronological order;

[0073] The column vectors at all times are used to form a chlorine mass fraction matrix;

[0074] Calculate the time correlation matrix based on the chlorine mass fraction matrix;

[0075] Eigenvalues ​​and corresponding eigenvectors are obtained by performing eigenvalue decomposition on the time correlation matrix.

[0076] The spatial modes of each order are calculated using eigenvectors and the chlorine mass fraction matrix.

[0077] Specifically, the flow field mode decomposition employs the snapshot eigenorthogonal decomposition method (POD method). First, the chlorine mass fraction vector at each time step is obtained: the chlorine mass fraction u at each spatial location on the characteristic surface at time step j. ij Construct column vectors:

[0078] G j =[u 1j ,u 2j ,…,u ij ,…u mj ] T (1)

[0079] In the formula, the subscript m represents the number of spatial points of the characteristic section.

[0080] Constructing the chlorine mass fraction correlation matrix: The chlorine mass fraction matrix is ​​constructed by combining the column vectors of chlorine mass fraction at n time points.

[0081] Q = [G1, G2, ..., G n (2)

[0082] In equation (2), matrix Q is an m×n matrix, where m is the number of spatial location points on the feature surface and n represents 500 different times.

[0083] Calculate the time correlation matrix according to formula (3):

[0084]

[0085] Solve the characteristic equation (4): Perform eigenvalue decomposition on the time correlation matrix to obtain the eigenvalues ​​λ. k and the corresponding orthogonal eigenvectors ν k .

[0086] Cν k =λ k ν k (4)

[0087] Calculate the k-th mode of POD according to formula (5):

[0088]

[0089] The energy percentage of the kth order is calculated according to formula (6):

[0090]

[0091] Calculate the time coefficient according to formula (7):

[0092]

[0093] in, For ν k The i-th component.

[0094] This embodiment introduces "Snapshot POD" to decouple the global diffusion field at 500 time points, compressing high-dimensional transient data into energy-ordered modes; more than 80% of the energy can be captured using only the first dominant mode, significantly reducing the data dimensionality, while preserving the master control space structure of the entire leakage process, avoiding the problem of "instantaneous unrepresentativeness" in the traditional single-frame method.

[0095] Furthermore, the process of selecting the spatial mode with the largest energy proportion as the dominant mode includes:

[0096] Sort the eigenvalues ​​corresponding to each spatial mode by size;

[0097] The first spatial mode with the largest energy proportion is identified as the dominant mode.

[0098] Specifically, the modes are sorted in descending order of energy percentage, and the first mode φ1 with the largest energy percentage is the dominant mode.

[0099] Furthermore, the process of mapping the dominant modes to the escape path node mesh to obtain the modal amplitudes of each node includes:

[0100] An escape route node grid file is generated based on the intersections of roads and areas where people gather in the factory area. Each node contains a number and two-dimensional coordinates.

[0101] Establish the shortest distance correspondence between escape path node numbers and feature surface location unit numbers;

[0102] Based on the shortest distance correspondence, the modal amplitude of the dominant mode is assigned to the corresponding escape path node.

[0103] Specifically, an escape path grid node file is created: an escape path node grid file (containing node numbers P (P = 1, 2…p)) is generated based on the leakage scenario. n ), p n The total number of grid nodes and two-dimensional node coordinates (A) of the escape path P B P The nodes represent intersections of factory roads and areas where people gather, and the grid lines represent factory roads; escape exits are specified.

[0104] Establishing the correspondence between escape path mesh node numbers and feature surface location element numbers: The feature surface location element numbers M (M = 1, 2, 3, ..., m) are derived from Fluent, where m represents the total number of escape path mesh nodes and the corresponding computational mesh node coordinates (X...). M Z M For the escape path grid node number P, the distance formula between two points is used. Calculate the distance from each grid node of the escape path to each computational grid node. When the distance is shortest, P = M. Using the above method, establish the correspondence between the grid node numbers of the escape path and the feature surface location cell numbers.

[0105] Specifically, based on the established correspondence between the escape path grid node number and the feature surface position element number, when the escape path grid node number and the feature surface position element number are the same, the value of the escape path grid node is set to be equal to the modal value of the feature surface position element.

[0106] Furthermore, the process of calculating the risk factor of each node based on the modal amplitude includes:

[0107] Obtain the modal amplitude of the dominant mode at each node, and use the product of the modal amplitude and the maximum value of the corresponding time coefficient as an indicator. Based on a preset threshold, convert the indicator into a risk factor.

[0108] This embodiment defines the quantification method of the hazard factor as "modal amplitude × corresponding time coefficient maximum value", transforming the continuously changing concentration field into a stable hazard zoning index. It assigns a unique and reproducible hazard weight to each road, eliminating the path oscillation caused by differences in instantaneous concentrations in traditional methods. By embedding the hazard factor into the "equivalent path length" model, the optimization objective of "shortest path" is modified to "minimum harm". Even if the actual road is long, as long as the hazard factor is low, it may still be selected by the algorithm, thereby significantly reducing personnel exposure dose while ensuring acceptable escape time.

[0109] Furthermore, the process of converting the actual path length into an equivalent path length using the hazard factor includes:

[0110] Obtain the actual length of each segment in the path, and weight the risk coefficients of the two ends of each segment to obtain the road risk coefficient weight of the corresponding segment;

[0111] Multiply the road hazard coefficient weight by the corresponding actual length to obtain the equivalent path length of the corresponding segment.

[0112] Specifically, the equivalent path length L is calculated as follows: L = cL 实 L 实 Let be the actual length of the path, and 'c' be the weight of the road hazard factor. Considering the degree of harm of chlorine concentration to the human body, the calculation method for the road hazard factor weight is as follows:

[0113]

[0114] Where x is the maximum value of the modal amplitude multiplied by the corresponding time coefficient, and has the dimension of chlorine mass fraction. 0.000001225 corresponds to a chlorine concentration of 0.5 ppm, which is the concentration threshold for one hour of exposure to have no known acute or chronic effects on the human body.

[0115] Furthermore, the process of planning an escape route from the starting point to the safe exit based on the shortest path algorithm, using the equivalent path length as the weight, includes:

[0116] Construct a weighted directed graph with the escape starting point as the source point and the safety exit as the destination point, using the equivalent path length as the edge.

[0117] Dijkstra's algorithm is used to search for the shortest path in a weighted directed graph;

[0118] If there are multiple safety exits, calculate the shortest equivalent path length to each exit and select the path corresponding to the minimum value as the final escape route.

[0119] Specifically, Dijkstra's algorithm is used to calculate the shortest distance from the escape point to the safe exit, where the path length in Dijkstra's algorithm is replaced by the equivalent path length. If there are multiple safe exits, the distance from the escape point to each safe exit needs to be calculated, and the shortest distance is selected. Finally, the escape paths for all node locations are obtained.

[0120] This embodiment uses Dijkstra's algorithm with equivalent length as the weight to perform shortest path search, achieving a globally optimal and acyclic path solution within 10... 5 In a factory road network with a scale of thousands of nodes, a global path update can be completed in milliseconds, meeting the needs of emergency real-time decision-making.

[0121] In summary, the "universally applicable" escape route output in this embodiment reduces the average exposure dose of personnel evacuating along this route by more than 40% during the entire 500s leakage cycle compared to the traditional single instantaneous route, significantly improving escape safety and reliability.

[0122] As a preferred implementation method, the above method will be explained in detail using the Jack Rabbit chlorine gas leak experiment scenario as an example:

[0123] A 3D model was built for a chlorine leak scenario. The model includes the leak location, fixed buildings and obstacles in the plant area, and the terrain. The computational domain is a hexahedral region of x×y×z = 1000×250×1200m. The leak source is located at (-500, -50, -200), and multiple hexahedrons of varying lengths, widths, and heights are arranged around the leak source to simulate the built environment. Chlorine is contained in a storage tank with a maximum capacity of 10 tons. The tank is equipped with four diffusers, oriented as follows: 0° upward, 90° horizontal, 135° downward, and 180° downward. Each diffuser releases chlorine through a 15.2 cm diameter through-hole. Since most of the JR experiments used a 180° downward through-hole for chlorine release, a simplified model was adopted. The leak model is as follows: Figure 2 and Figure 3 As shown.

[0124] Numerical calculation of the chlorine gas leakage process: The mesh was generated using Ansys Fluent meshing, employing a polyhedral mesh generation method. Local refinement was applied in the building area and near the ground, with a boundary layer added. The first mesh layer thickness was 0.00668503, the growth rate was set to 1.3, and the boundary layer thickness was approximately 1.9m. After mesh independence analysis, the number of meshes was selected as 6,067,873, and the average mesh quality was 0.9491357. Overall, as shown... Figure 4 As shown.

[0125] The Realizable k-ε model was used for numerical calculation of turbulence, and the Species Model was used for chlorine transport calculation.

[0126] Boundary conditions: Simulate the experimental conditions of JR Experiment 1, with the environmental inlet (face A of the hexahedron) being the air velocity inlet V. 环境 = 2.47 m / s, with an angle of 174° clockwise to the positive direction of the model, the outlet is a pressure outlet, and the ground, building, and obstacle surfaces are subject to no-slip boundary conditions, thus yielding the chlorine leakage rate V. 氯气 = 2m / s.

[0127] The calculation process is as follows: First, the chlorine leakage velocity is set to 0. The steady-state numerical calculation method is used to perform numerical calculation of the environmental wind field. After the calculation converges, the result is used as the initial field. The chlorine leakage velocity is set and unsteady calculation is performed. The unsteady time step is Δt = 0.01s and the total unsteady time step is 50000.

[0128] Establish a feature surface: A feature surface parallel to the ground (the height of the human breathing area) is established at a height of 1.5m (y=1.5) above the ground. During the calculation, the chlorine diffusion field of the feature surface (including the location cell number, X / Y / Z coordinate information, and the mass fraction of chlorine) is output at each time step, for a total of n=500 time steps.

[0129] Flow field mode decomposition: The snapshot eigenorthogonal decomposition method (POD method) is adopted. First, the chlorine mass fraction vector at each time step: the chlorine mass fraction u at each spatial location on the characteristic surface at time step j. ij Construct column vectors:

[0130] G j =[u 1j ,u 2j ,…,u ij ,…u mj ] T (1)

[0131] In the formula, the subscript m represents the number of spatial points of the characteristic section.

[0132] Constructing the chlorine mass fraction correlation matrix: The chlorine mass fraction matrix is ​​constructed by combining the column vectors of chlorine mass fraction at n time points.

[0133] Q = [G1, G2, ..., G n (2)

[0134] In equation (2), matrix Q is an m×n matrix, where m is the number of spatial location points on the feature surface and n represents 500 different times.

[0135] Calculate the time correlation matrix according to formula (3):

[0136]

[0137] Solve the characteristic equation (4): Perform eigenvalue decomposition on the time correlation matrix to obtain the eigenvalues ​​λ. k and the corresponding orthogonal eigenvectors ν k .

[0138] Cν k =λ k ν k (4)

[0139] Calculate the k-th mode of POD according to formula (5):

[0140]

[0141] The energy percentage of the kth order is calculated according to formula (6):

[0142]

[0143] Calculate the time coefficient according to formula (7):

[0144]

[0145] in, For ν k The i-th component.

[0146] Determine the dominant mode: such as Figure 5 and Figure 6 As shown, the modes are sorted in descending order of energy percentage, and the first mode φ1 with the largest energy percentage is the dominant mode.

[0147] Create an escape path mesh file: Generate an escape path mesh file (containing node numbers P (P = 1, 2…p) based on the leak scenario. n ), p n The total number of grid nodes and two-dimensional node coordinates (A) of the escape path P B P The nodes represent intersections of factory roads and areas where people gather, and the grid lines represent factory roads; escape exits are specified.

[0148] Establishing the correspondence between escape path mesh node numbers and feature surface location element numbers: The feature surface location element numbers M (M = 1, 2, 3, ..., m) are derived from Fluent, where m represents the total number of escape path mesh nodes and the corresponding computational mesh node coordinates (X...). M Z M For the escape path grid node number P, the distance formula between two points is used. Calculate the distance from each grid node of the escape path to each computational grid node. When the distance is shortest, P = M. Using the above method, establish the correspondence between the grid node numbers of the escape path and the feature surface location cell numbers.

[0149] Escape path mesh node value determination: Based on the established correspondence between escape path mesh node numbers and feature surface position element numbers, when the escape path mesh node number and the feature surface position element number are the same, set the escape path mesh node value equal to the modal value of the feature surface position element. Calculate the path equivalent length L: L = cL 实 L 实 Let be the actual length of the path, and 'c' be the weight of the road hazard factor. Considering the degree of harm of chlorine concentration to the human body, the calculation method for the road hazard factor weight is as follows:

[0150]

[0151] Where x is the maximum value of the modal amplitude multiplied by the corresponding time coefficient, and has the dimension of chlorine mass fraction. 0.000001225 means that the chlorine concentration is 0.5 ppm, which is the concentration threshold for one hour of exposure to have no known acute or chronic effects on the human body.

[0152] Escape route planning: such as Figure 7 and Figure 8 As shown, Dijkstra's algorithm is used to calculate the shortest distance from the escape point to the safe exit, where the path length in Dijkstra's algorithm is replaced by the equivalent path length. If there are multiple safe exits, the distance from the escape point to each safe exit needs to be calculated, and the shortest distance is selected. Finally, the escape paths for all node locations are obtained.

[0153] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An escape path planning method based on global modal characteristics of a flow field, characterized in that, include: A three-dimensional computational model of the leakage scenario was established and unsteady numerical calculations were performed to obtain time-series data of the leakage diffusion field. The time series data is subjected to intrinsic orthogonal decomposition to obtain spatial modes sorted by energy proportion; The spatial mode with the largest energy proportion is selected as the dominant mode, and the dominant mode is mapped to the escape path node grid to obtain the modal amplitude of each node. The risk coefficient of each node is calculated based on the modal amplitude and the maximum value of the corresponding time coefficient, and the actual path length is converted into the equivalent path length using the risk coefficient. Using the equivalent path length as the weight, an escape route from the starting point to the safe exit is planned based on the shortest path algorithm; The process of establishing a three-dimensional computational model of the leakage scenario and performing unsteady numerical calculations to obtain time-series data of the leakage diffusion field includes: A three-dimensional geometric model was constructed based on the location of the leak source, fixed obstacles, and terrain environment. The three-dimensional geometric model is meshed, and local meshing is performed in the near-ground and obstacle areas; Set boundary conditions, where the inlet is the velocity inlet, the outlet is the pressure outlet, the solid wall is the no-slip boundary, and the leakage outlet is the velocity inlet; First, set the leak outlet velocity to 0 and perform steady-state calculations to obtain the initial wind field. Then, using the initial wind field as the initial field, set the leak outlet velocity and perform unsteady-state calculations, and output the chlorine diffusion field data on the characteristic surface with a fixed time step. The process of performing intrinsic orthogonal decomposition on the time series data to obtain spatial modes sorted by energy proportion includes: Construct a column vector from the chlorine mass fraction data on the feature surface at each time step in chronological order; The column vectors at all times are used to form a chlorine mass fraction matrix; Calculate the time correlation matrix based on the chlorine mass fraction matrix; The time correlation matrix is ​​subjected to eigenvalue decomposition to obtain eigenvalues ​​and corresponding eigenvectors; The spatial modes of each order are calculated using the eigenvectors and the chlorine mass fraction matrix.

2. The method according to claim 1, characterized in that, The process of selecting the spatial mode with the largest energy proportion as the dominant mode includes: Sort the eigenvalues ​​corresponding to each spatial mode by size; The first spatial mode with the largest energy proportion is determined as the dominant mode.

3. The method according to claim 1, characterized in that, The process of mapping the dominant modes to the escape path node mesh to obtain the modal amplitude of each node includes: An escape route node grid file is generated based on the intersections of roads and areas where people gather in the factory area. Each node contains a number and two-dimensional coordinates. Establish the shortest distance correspondence between escape path node numbers and feature surface location unit numbers; Based on the shortest distance correspondence, the modal amplitude of the dominant mode is assigned to the corresponding escape path node.

4. The method according to claim 1, characterized in that, The process of calculating the risk factor of each node based on the modal amplitude includes: Obtain the modal amplitude of the dominant mode at each node, and use the product of the modal amplitude and the maximum value of the corresponding time coefficient as an indicator. Convert the indicator into a risk factor according to a preset threshold.

5. The method according to claim 1, characterized in that, The process of converting the actual path length into an equivalent path length using the aforementioned hazard factor includes: Obtain the actual length of each segment in the path, and weight the risk coefficients of the two ends of each segment to obtain the road risk coefficient weight of the corresponding segment; Multiply the road hazard coefficient weight by the corresponding actual length to obtain the equivalent path length of the corresponding segment.

6. The method according to claim 1, characterized in that, Using the equivalent path length as the weight, the process of planning an escape route from the starting point to the safe exit based on the shortest path algorithm includes: Construct a weighted directed graph with the escape starting point as the source point and the safety exit as the destination point, using the equivalent path length as the edge. Dijkstra's algorithm is used to search for the shortest path in the weighted directed graph; If there are multiple safety exits, calculate the shortest equivalent path length to each exit and select the path corresponding to the minimum value as the final escape route.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1-6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the method according to any one of claims 1-6.

Images