Ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis
By combining CFD simulation and the A-Star algorithm, building a three-dimensional urban model and introducing satellite remote sensing data for verification, the problems of height discrepancy and scientificity in ventilation corridor identification were resolved, and accurate ventilation corridor identification and planning were achieved.
Patent Information
- Application Number
- CN202510829807.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-20
- Publication Date
- 2025-10-03
AI Technical Summary
Existing technologies fail to effectively consider the vertical differentiation characteristics of different heights in ventilation corridor identification, resulting in insufficient applicability of identification results in three-dimensional space and a lack of result verification, affecting scientificity and reliability.
Combining CFD simulation with the A-Star algorithm, a three-dimensional urban model was constructed, and wind field simulations at multiple height levels were performed. Satellite remote sensing data was introduced for verification, and path similarity quantification and hierarchical clustering methods were used to divide the ventilation corridor height intervals and screen the corridors with the highest ventilation efficiency.
It achieves accurate identification of ventilation corridors, improves the reliability and scientific nature of the identification results, and provides quantifiable technical support for urban wind environment planning.
Smart Images
Figure CN120744360A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of urban planning and relates to urban ventilation corridor identification by integrating numerical simulation technology and control algorithm, and specifically is a ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis. Background Art
[0002] As cities continue to develop, their size and population density continue to increase. Overly concentrated construction patterns alter the underlying surface properties. Large amounts of impervious surfaces and densely packed buildings exacerbate the urban heat island effect, leading to significant increases in urban temperatures and hindering the accumulation and diffusion of pollutants, posing significant risks to human health. Ventilation corridors, by guiding fresh air infiltration, can improve air exchange efficiency within and outside the city, reduce heat island intensity, and accelerate pollutant diffusion, helping to improve urban wind and heat conditions and promote sustainable urban environmental development.
[0003] Ventilation corridor identification, as an upstream step in ventilation corridor construction, plays a crucial role in subsequent ventilation corridor planning, management, and construction. Accurate ventilation corridor identification provides a key basis for optimizing urban spatial layout and formulating strategies to improve wind and heat environments. Currently, three main technical approaches exist for ventilation corridor identification. The first is to identify potential corridors between wind sources and wind sinks based on satellite data. The second is to analyze urban ventilation resistance based on factors such as urban topography and urban construction, and to identify ventilation corridors using a minimum-cost path algorithm. The third is to use numerical simulation tools such as WRF and CFD to simulate wind velocity flows within cities and subsequently identify ventilation corridors. While these ventilation corridor methods have been widely used, they still have shortcomings. Existing analysis methods ignore urban wind fields, the three-dimensional characteristics of cities, and air flow characteristics such as local air circulation. They also fail to account for differences between ventilation corridors at different altitudes. This results in inaccurate ventilation corridor identification results and provides limited guidance for ventilation corridor planning, construction, and management.
[0004] Patent CN117436174A proposes a ventilation corridor construction method based on CFD and circuit theory. The urban ventilation corridor identification method proposed in this patent includes the following steps: (1) obtaining and preprocessing meteorological, topographic, and building data; (2) simulating the urban wind field based on CFD simulation; (3) inverting the surface temperature by combining remote sensing data; (4) analyzing ventilation suitability through multi-factor evaluation; (5) simulating the wind environment current distribution using circuit theory; and (6) constructing a "wind source-corridor-repair zone" ventilation system. However, this method has obvious shortcomings: it only simulates a single height, ignoring the differences in wind field characteristics at different heights, and fails to verify the simulation results.
[0005] In summary, the technical problems that need to be solved in previous studies are: first, the vertical differentiation characteristics of ventilation corridors at different heights are not taken into account, and the simulation is only based on a single height, resulting in insufficient applicability of the identification results in three-dimensional space; second, there is a lack of verification of the ventilation corridor identification results, which affects its scientificity and reliability. Summary of the Invention
[0006] The purpose of the present invention is to provide a ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis. By combining CFD simulation and A-Star algorithm, accurate identification of ventilation corridors is achieved. In the identification process, the three-dimensional structure of urban buildings and mountains and the fluid characteristics of wind are comprehensively considered to improve the identification accuracy; the vertical differentiation law of ventilation corridors at different heights is revealed; satellite remote sensing data is introduced to verify the identification results, ensure their reliability and practicality, and provide a scientific basis for urban ventilation corridor planning.
[0007] To achieve the above object, the technical solution adopted by the present invention is: a ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis, which includes the following steps:
[0008] Step 1: Build a 3D city model based on multi-source geospatial data: aggregate building planes and assign heights, slice the mountains, and construct a simplified 3D city model.
[0009] Step 2: Computational fluid dynamics is used to simulate wind fields at different altitudes, and the A-Star path search algorithm is integrated to enumerate potential ventilation corridors at different altitudes.
[0010] Step 3: Verify the reliability of corridor enumeration results through satellite-derived surface temperature and vegetation index: Based on high-resolution satellite data, compare the surface temperature and vegetation index data inside and outside the potential ventilation corridor to verify the reliability of the corridor enumeration results;
[0011] Step 4: Use path similarity measurement and hierarchical clustering method to divide the ventilation corridor height intervals;
[0012] Step 5: Screen typical height layer ventilation corridors based on ventilation efficiency index: At the typical characteristic height surface, identify higher efficiency ventilation corridors based on ventilation corridor efficiency index as ventilation corridor identification results.
[0013] Furthermore, in step 1, when processing buildings: the average width of the city's main roads is used to aggregate the building planes, and the average building height is used to assign heights to the aggregated building patches; when processing mountains: the height of the mountain slices is determined according to the total height difference of the urban terrain and the number of slices; and after splicing the model, it is converted into a format that can be recognized by the CFD simulation software.
[0014] Furthermore, the mountain slice height is calculated using the following formula:
[0015]
[0016] Among them, H i is the height of the i-th mountain slice, n is the number of slices, H M The overall height of the mountain.
[0017] Furthermore, in step 2, the wind field distribution at different altitudes is obtained through CFD simulation, exported and projected into a wind speed grid plane after grayscale conversion, and then integrated with the A-Star path search algorithm to enumerate potential ventilation corridors.
[0018] Furthermore, CFD simulation requires simulation parameter setting, solution, and post-processing steps; the A-Star algorithm is implemented on the Python platform, and it is necessary to set 12 pairs of ventilation corridor starting and ending points on the boundary line of the identification range according to the dominant wind direction, execute the algorithm, and project and export the ventilation corridor enumeration results.
[0019] Furthermore, in step 3, the Landsat8 TIRS sensor bands ST_B10, SR_B4, and SR_B5 are downloaded, and then 1-fold, 2-fold, and 3-fold mutually exclusive buffer zones of the ventilation corridor are established. The differences in surface temperature and vegetation index values within the buffer zones are compared, and then the Mann-WhitneyU is used to verify that the surface temperature data inside the ventilation corridor enumeration results is smaller than that outside, and the vegetation index data is larger than that outside, so as to verify the reliability of the ventilation corridor enumeration results.
[0020] Furthermore, step 4 includes calculating the bidirectional Hausdorff distance of the ventilation corridor, constructing a similarity matrix of ventilation corridors at different heights, and classifying the height intervals of the ventilation corridors using a hierarchical clustering method.
[0021] Furthermore, in step 4, the corridors generated by A-Star are first discretized into point sets at certain intervals, and the Hausdorff distances between any two paths are calculated to construct a similarity matrix. The similarity calculation formula is:
[0022]
[0023] Where, Simlarity(A,B) is the similarity between ventilation corridors A and B at different heights; L in L is the intersection length of the two ventilation corridor buffer zones, in meters; A ,L B are the lengths of ventilation corridors A and B, respectively, in meters;
[0024] Based on the similarity matrix, a hierarchical clustering algorithm was used to group the height surfaces. The number of clusters was optimized by the silhouette coefficient, and clusters with sample size ≤ 3 were merged. Finally, the ventilation corridor was divided into a small number of height intervals.
[0025] Furthermore, in step 5, the average wind speed in the ventilation corridor is calculated by counting the average grid value of the ventilation corridor, and the resistance-free wind speed is calculated according to the gradient wind correction formula, and the ventilation efficiency index of the ventilation corridor is calculated.
[0026] Furthermore, the ventilation corridor efficiency calculation formula is:
[0027]
[0028] Where VEI refers to the ventilation efficiency index of the corridor, is the average wind speed in the ventilation corridor at the Z plane height, in m / s; V z is the Z-plane height ventilation corridor no-resistance wind speed, dynamically calculated according to the gradient wind correction formula, The surface roughness coefficient must be selected according to actual conditions; is the effective length ratio of the ventilation corridor.
[0029] The beneficial effects of the present invention are as follows:
[0030] This paper systematically addresses the core challenge of insufficient vertical differentiation analysis in the three-dimensional identification of urban ventilation corridors by innovatively integrating computational fluid dynamics (CFD) simulation with the A-Star heuristic search algorithm, achieving accurate identification of ventilation corridors. Its effects are as follows: First, CFD multi-level wind field simulation based on a three-dimensional urban model is combined with A-Star path enumeration to achieve a refined mapping from macroscopic wind fields to corridor paths, overcoming the limitations of traditional single-height surface simulation results. Second, a spatial statistical validation mechanism using satellite-derived Land Surface Temperature (LST) and National Natural Distress Vegetation Index (NDVI) is introduced to significantly improve the reliability of corridor identification results through stratified sampling and significance testing. Furthermore, the Hausdorff distance is used to quantify path morphological similarity, combined with a hierarchical clustering algorithm to intelligently group 21 height layers, objectively revealing the vertical distribution patterns and characteristic height ranges of ventilation corridors. Finally, critical sections are selected based on the clustering results, and typical corridors are dynamically screened using the Ventilation Efficiency Index (VEI), forming a scientific and practical urban wind environment optimization solution. This framework significantly improves the accuracy and efficiency of urban ventilation corridor identification through the collaboration of multidisciplinary methods and full-process data-driven, providing quantifiable and reusable technical support for mitigating the heat island effect and optimizing urban wind environment planning. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] Figure 1 This is a flow chart of a ventilation corridor identification system based on CFD simulation and A-Star algorithm synthesis;
[0032] Figure 2 This is a schematic diagram of the building plane aggregation method;
[0033] Figure 3 is a schematic diagram of the wind speed grid;
[0034] Figure 4 Schematic diagram of potential paths enumerated by the A-Star algorithm;
[0035] Figure 5 This is a schematic diagram of the ventilation corridor similarity matrix;
[0036] Figure 6 Schematic diagram of hierarchical clustering results;
[0037] Figure 7 Schematic diagram of ventilation corridor identification results. DETAILED DESCRIPTION
[0038] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
[0039] like Figure 1 This embodiment provides a ventilation corridor identification method based on CFD simulation and A-Star algorithm. By integrating computational fluid dynamics (CFD) simulation and A-Star path search algorithm, it achieves refined identification of ventilation corridors at different height levels, providing a scientific basis for urban three-dimensional wind environment planning. The specific steps are as follows:
[0040] 1. Data Acquisition
[0041] The patent requires the following data: (1) Geospatial data, including digital elevation models (DEMs), building vector data, and traffic network data. (2) The target area's dominant wind direction in summer or winter, the annual average wind speed at a height of 10 meters, and the surface roughness index in the incoming flow direction. (3) High-resolution satellite inversion data, including land surface temperature (LST) and normalized difference vegetation index (NDVI).
[0042] 2. Construction of 3D urban model
[0043] A refined 3D urban model was constructed to facilitate subsequent CFD wind field simulation. The building model was simplified using the GIS platform. The building planes were aggregated to generate new patches based on the average width of the main roads, and the patch heights were assigned based on the average building heights. The building plane aggregation method is shown in Figure 2 .
[0044] The mountain model is simplified through the GIS platform. The buildings are sliced according to the overall terrain height of the city and the mountain slice height is assigned. The height is calculated using the following formula:
[0045]
[0046] Among them, H iis the height of the i-th mountain slice, n is the number of slices, H M The overall height of the mountain.
[0047] In ArcScene, the aggregated building model and the sliced mountain model are combined to form a three-dimensional city model.
[0048] 3. CFD wind field simulation
[0049] Computational fluid dynamics (CFD) is used to simulate urban wind farms. The following steps are required:
[0050] (1) Simulation parameter setting. Parameters such as inlet wind direction, inlet wind speed, roughness index of incoming flow, computational domain size, minimum computational grid, turbulence model, number of iterations, and global convergence criteria are set based on relevant data, specifications, and cases. The inlet wind direction, inlet wind speed, and roughness index of incoming flow refer to the actual meteorological data of the simulation area. The computational domain size should be based on the length, width, and height of the urban model, with the inlet expanding by 3 times the maximum building height and the outlet expanding by 5 times the maximum building height to meet the space requirements for turbulent development. The minimum computational grid should ensure that each building patch has at least 10 computational grids. The steady-state RNGk-ε model should be selected as the turbulence model. The number of iterations should be at least 1000. The global convergence criteria should be less than or equal to 1.000E-4%.
[0051] (2) Solving. To improve simulation efficiency, a parallel solver is used.
[0052] (3) Post-processing. Output the wind speed plane at 0-100m height (step length is 5, a total of 21 height planes) as a grayscale image, and project the image into a wind speed grid with geospatial coordinates in GIS. The wind speed grid is as follows: Figure 3 shown.
[0053] 4. A-Star corridor path enumeration
[0054] Based on the wind speed grid data from CFD simulation, the A-Star algorithm is used to enumerate potential ventilation corridors. The A-Star algorithm with the introduction of an evaluation function can effectively overcome the short-sightedness when searching for ventilation corridors and improve the reliability of the simulation. The algorithm evaluation function is:
[0055] f(n)=g(n)+h(n)
[0056] Among them, f(n) is the overall cost, g(n) is the actual cost from the starting point to the current point, and h(n) is the estimated cost from the current point to the target point.
[0057] The starting and ending points are set as 12 pairs of points generated uniformly by percentage along the boundary line of the study area. The upwind starting point and downwind ending point are determined according to the prevailing wind direction. Enumeration is performed for the 0-100m altitude layer (5m step length, 21 altitude planes in total), and 144 potential paths are generated for each altitude plane. The enumerated ventilation corridors are as follows: Figure 4 shown.
[0058] 5. Verification of corridor enumeration results
[0059] The reliability of corridor enumeration results was verified by inverting Landsat 8TIRS sensor data, including Land Surface Temperature (LST) and Normalized Difference Vegetation Index (NDVI). Taking the 0-meter elevation corridor as an example, sampling points were generated at 50-meter intervals. The average distance from each point to obstacles was calculated as the corridor width. Three mutually exclusive buffer zones (1, 2, and 3 times the corridor width) were established. The variation in data within and outside the corridor was calculated. If the Mann-Whitney U test was valid, the reliability of the enumeration results at different corridor elevations was confirmed.
[0060] 6. Similarity Analysis and Hierarchical Clustering
[0061] In order to clarify the vertical differentiation characteristics of ventilation corridors, it is necessary to calculate the similarity of ventilation corridors at different heights, construct a similarity matrix, and perform hierarchical clustering. The specific steps are as follows:
[0062] The Hausdorff distance is introduced to quantify the morphological similarity of corridor paths at different height layers and construct a similarity matrix. First, the corridors generated by A-Star are discretized into point sets at 5-meter intervals, and the Hausdorff distance between each two paths is calculated to construct a similarity matrix. The similarity matrix is as follows: Figure 5 , the similarity calculation formula is:
[0063]
[0064] Where, Simlarity(A,B) is the similarity between ventilation corridors A and B at different heights; L in L is the intersection length of the two ventilation corridor buffer zones, in meters; A ,L B are the lengths of ventilation corridors A and B respectively, in meters.
[0065] Based on the similarity matrix, a hierarchical clustering algorithm was used to group the 21 height surfaces. The number of clusters was optimized by the silhouette coefficient, and clusters with sample size ≤ 3 (continuous height interval ≤ 15 meters) were merged. Finally, the ventilation corridor was divided into a small number of height intervals. The hierarchical clustering results are shown in Figure 2. Figure 6 .
[0066] 7. Ventilation corridor identification
[0067] In order to take into account the vertical differentiation characteristics of ventilation corridors at different heights and achieve accurate identification of ventilation corridors, it is necessary to select a typical height surface and consider the potential path with the highest ventilation corridor efficiency as the urban ventilation corridor. The specific steps are as follows:
[0068] Based on the hierarchical clustering results, ventilation corridors were screened at critical points in different height intervals. The screening index was the ventilation corridor efficiency (Ventilation Efficiency Index, VEI), which was calculated as follows:
[0069]
[0070] Where VEI refers to the ventilation efficiency index of the corridor, is the average wind speed in the ventilation corridor at the Z plane height, in m / s; V z is the Z-plane height ventilation corridor no-resistance wind speed, dynamically calculated according to the gradient wind correction formula, The surface roughness coefficient must be selected according to actual conditions; is the effective length ratio of the ventilation corridor.
[0071] The final ventilation corridor identification results are as follows: Figure 7 shown.
[0072] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the above embodiments do not limit the scope of protection of the present invention in any form. All technical solutions obtained by equivalent substitution, etc., fall within the scope of protection of the present invention. Parts not covered by the present invention are the same as the existing technology or can be implemented using existing technology.
Claims
1. A ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis, characterized in that: The steps include: Step 1: Build a 3D city model based on multi-source geospatial data: aggregate building planes and assign heights, slice the mountains, and construct a simplified 3D city model. Step 2: Computational fluid dynamics is used to simulate wind fields at different altitudes, and the A-Star path search algorithm is integrated to enumerate potential ventilation corridors at different altitudes. Step 3: Verify the reliability of corridor enumeration results through satellite-derived surface temperature and vegetation index: Based on high-resolution satellite data, compare the surface temperature and vegetation index data inside and outside the potential ventilation corridor to verify the reliability of the corridor enumeration results; Step 4: Use path similarity measurement and hierarchical clustering method to divide the ventilation corridor height intervals; Step 5: Screen typical height layer ventilation corridors based on ventilation efficiency index: At the typical characteristic height surface, identify higher efficiency ventilation corridors based on ventilation corridor efficiency index as ventilation corridor identification results.
2. The ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis according to claim 1 is characterized in that: In step 1, when processing buildings, the average width of urban main roads is used to aggregate building planes, and the average building height is used to assign heights to the aggregated building patches; when processing mountains, the height of mountain slices is determined according to the total height difference of the urban terrain and the number of slices; and after splicing the model, it is converted into a format that can be recognized by CFD simulation software.
3. The ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis according to claim 2 is characterized in that: The mountain slice height is calculated using the following formula: Among them, H i is the height of the i-th mountain slice, n is the number of slices, H M The overall height of the mountain.
4. The ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis according to claim 1 is characterized in that: In step 2, the wind field distribution at different altitudes is obtained through CFD simulation, which is then exported and projected into a wind speed grid plane after grayscale conversion, and then the A-Star path search algorithm is integrated to enumerate potential ventilation corridors.
5. The ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis according to claim 4 is characterized in that: CFD simulation requires simulation parameter setting, solution, and post-processing steps; the A-Star algorithm is implemented on the Python platform. It is necessary to set 12 pairs of ventilation corridor starting and ending points on the boundary line of the identification range according to the dominant wind direction, execute the algorithm, and project and export the ventilation corridor enumeration results.
6. The ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis according to claim 1 is characterized in that: In step 3, the Landsat8 TIRS sensor bands ST_B10, SR_B4, and SR_B5 are downloaded, and then 1-fold, 2-fold, and 3-fold mutually exclusive buffer zones of the ventilation corridor are established. The differences in surface temperature and vegetation index values within the buffer zones are compared, and the Mann-Whitney U method is used to verify that the surface temperature data inside the ventilation corridor enumeration results is smaller than that outside, and the vegetation index data is larger than that outside, so as to verify the reliability of the ventilation corridor enumeration results.
7. The ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis according to claim 1 is characterized in that: The step 4 includes calculating the bidirectional Hausdorff distance of the ventilation corridor, constructing a similarity matrix of ventilation corridors at different heights, and classifying the height intervals of the ventilation corridors using a hierarchical clustering method.
8. The ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis according to claim 1 or 7, characterized in that: In step 4, the corridors generated by A-Star are first discretized into point sets at certain intervals, the Hausdorff distances between any two paths are calculated, and a similarity matrix is constructed. The similarity calculation formula is: Where, Simlarity(A,B) is the similarity between ventilation corridors A and B at different heights; L in L is the intersection length of the two ventilation corridor buffer zones, in meters; A ,L B are the lengths of ventilation corridors A and B, respectively, in meters; Based on the similarity matrix, a hierarchical clustering algorithm was used to group the height surfaces. The number of clusters was optimized by the silhouette coefficient, and clusters with sample size ≤ 3 were merged. Finally, the ventilation corridor was divided into a small number of height intervals.
9. The ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis according to claim 1 is characterized in that: In step 5, the average wind speed of the ventilation corridor is calculated by counting the average grid value of the ventilation corridor, and the resistance-free wind speed is calculated according to the gradient wind correction formula, and the ventilation efficiency index of the ventilation corridor is calculated.
10. The ventilation corridor identification method based on CFD simulation and A-Star algorithm synthesis according to claim 9 is characterized in that: The calculation formula for ventilation corridor efficiency is: Where VEI refers to the ventilation efficiency index of the corridor, is the average wind speed in the ventilation corridor at the Z plane height, in m / s; V z is the Z-plane height ventilation corridor no-resistance wind speed, dynamically calculated according to the gradient wind correction formula, The surface roughness coefficient must be selected according to actual conditions; is the effective length ratio of the ventilation corridor.