A method for optimizing urban public space based on crowd behavior simulation

Abstract

The application discloses a kind of urban public space optimization methods based on crowd behavior simulation, comprising: obtaining city wind field parameter, city air pollutant concentration data and the geometric characteristics of building;Adopt the method of computational fluid dynamics simulation to calculate turbulent diffusion coefficient;According to real-time weather data API, calculate atmospheric stability by real-time weather data;Determine the concentration distribution of pollutants at different times and spatial positions by pollutant Lagrangian particle tracking method;Build crowd behavior pattern classification model, determine the crowd behavior pattern in different public space;Determine the optimal green belt and ventilation corridor configuration and the optimal public space ventilation optimization scheme;Evaluate the influence of different configurations on ventilation effect and pollution suppression.The present application provides more accurate and efficient optimization scheme for urban public space design by comprehensively considering meteorological, building distribution, crowd behavior and air flow and other multidimensional factors, which significantly improves the quality of life of citizens and the comfort of public space.

Description

Technical Field

[0001] This invention relates to the field of smart city planning, and in particular to a method for optimizing urban public spaces based on crowd behavior simulation. Background Technology

[0002] With the acceleration of global urbanization, urban population density is increasing daily, and urban air quality problems are becoming increasingly serious. High population density and heavy traffic activities lead to the accumulation of air pollutants, which has a huge negative impact on human health and the environment. Currently, most urban air quality optimization solutions focus on simple measures such as increasing greenery and improving ventilation. While these methods can alleviate air pollution to some extent, they often overlook the complexity of airflow and pollutant diffusion, and lack accurate prediction and control of actual environmental dynamics. For example, traditional green belt design and ventilation corridor layout are based solely on experience, lacking comprehensive analysis of wind fields, air pollutant concentrations, building distribution, and population activities. Therefore, they cannot achieve optimal airflow and pollutant control in complex urban environments. In addition, many existing ventilation optimization solutions typically use static calculation models, ignoring dynamic factors such as meteorological conditions, wind speed and direction, and real-time changes in pollutant concentrations. In most cases, ventilation models only consider the wind field characteristics of local areas, but fail to account for the impact of population density on airflow. The presence and movement of people influence airflow paths and pollutant dispersion processes, especially in high-density areas. The effectiveness of airflow and pollutant dispersion is closely related to population activity patterns, a point often overlooked by traditional airflow models. Furthermore, traditional airflow and pollutant dispersion models rely heavily on static data, lacking adaptability to dynamic meteorological environments and failing to provide effective air quality management strategies in rapidly changing urban environments. Current urban spatial design often neglects the interaction between population behavior patterns and airflow, leading to poor airflow or pollutant accumulation, impacting the comfort and air quality of public spaces. Particularly in busy areas of large cities, there is a lack of comprehensive airflow and population behavior analysis models, failing to scientifically optimize the layout of public spaces, resulting in inefficient airflow and ineffective pollutant dispersion. Therefore, current urban air quality optimization technologies face several challenges, including a lack of real-time response to dynamic meteorological data, neglect of the impact of population density and behavior patterns on airflow, and limitations in ventilation and pollution control schemes. These problems have rendered existing technologies less than satisfactory in addressing urban air pollution and optimizing public space design, creating an urgent need for new, more comprehensive, and more precise technological approaches to tackle the complex issues of urban air quality and public space optimization. Summary of the Invention

[0003] This invention addresses the problems existing in the prior art by providing a method for optimizing urban public spaces based on crowd behavior simulation, mainly comprising:

[0004] By using a meteorological station monitoring network, urban wind field parameters and urban air pollutant concentration data are obtained, and UAV lidar technology is used to obtain the geometric features of buildings.

[0005] Based on historical meteorological data from urban monitoring points, the turbulent kinetic energy at the monitoring points is predicted, and the turbulent dissipation rate is calculated. The three-dimensional velocity field of air flow within the urban area is obtained by computational fluid dynamics simulation, and the turbulent diffusion coefficient of air flow in different urban areas is calculated.

[0006] Real-time meteorological data is obtained from the real-time meteorological data API, atmospheric stability is calculated, and the stability classification result is determined based on the preset stability classification standard.

[0007] Large eddy simulation was used to determine the changing trends of airflow under different weather conditions, and the pollutant Lagrange particle tracking method was used to simulate the diffusion process of pollutants in the air, and to determine the concentration distribution of pollutants at different times and spatial locations.

[0008] Acquire public space layout data and real-time pedestrian density data within urban areas, construct a crowd behavior pattern classification model, determine crowd behavior patterns in different public spaces, and predict pedestrian density in each public space within a preset time period in the future.

[0009] Based on data on pollutant concentration, pedestrian density, air velocity, and wind direction in green areas, the optimal configuration of green belts and ventilation corridors is determined. Combining turbulent diffusion and pollutant concentration gradients, the optimal ventilation optimization scheme for public spaces is determined.

[0010] Based on the optimized layout of green belts and ventilation corridors, building distribution and wind field parameters, combined with air flow simulation and pollutant diffusion trajectory, the impact of different configurations on ventilation effect and pollution suppression is evaluated, and the ventilation path of public space, green belt distribution and building spacing configuration are optimized.

[0011] Furthermore, the process of acquiring urban wind field parameters and urban air pollutant concentration data through a meteorological station monitoring network, and using UAV lidar technology to acquire the geometric features of buildings, includes:

[0012] Based on the meteorological station monitoring network, wind speed, humidity, and temperature sensors were used to acquire urban wind field parameters, record the timestamp of each data point, and upload them to the urban air monitoring database. The wind field parameters included wind direction, wind speed, humidity, and temperature. Fixed and mobile PM2.5, NO2, and CO sensors were deployed to acquire urban air pollutant concentration data, and the monitoring point locations and monitoring periods were recorded. The sliding window method was used to process the monitoring data by time period, obtaining hourly, daily, and monthly pollutant concentration variation curves. By comparing with the wind field parameters, spatiotemporal interpolation methods were used to fill in the missing pollutant concentration data, resulting in a pollutant concentration distribution map with spatiotemporal continuity. Unmanned aerial vehicle (UAV) lidar technology was used to scan urban building structures, acquiring the geometric features of buildings, including building height, density, shape, and ground features.

[0013] Furthermore, based on historical meteorological data from urban monitoring points, the turbulent kinetic energy of the monitoring points is predicted, and the turbulent dissipation rate is calculated. A three-dimensional velocity field of airflow within the urban area is obtained using computational fluid dynamics simulation, and the turbulent diffusion coefficient of airflow in different urban areas is calculated, including:

[0014] Air viscosity is determined using the Sutherland formula based on air temperature and pressure. A turbulent kinetic energy prediction model is constructed by training a recurrent neural network using historical meteorological data from urban monitoring points and corresponding turbulent kinetic energy. Meteorological data includes wind field parameters and air pressure. Turbulent kinetic energy at monitoring points is determined using the real-time meteorological data prediction model. Based on the turbulent kinetic energy, the standard turbulent dissipation rate formula is applied. Calculate the turbulent dissipation rate ε, where k is the turbulent kinetic energy, l is the characteristic length, is the size of the building, and C μ These are model constants, obtained by fitting historical data. Based on the turbulence model parameters, computational fluid dynamics simulation is used to simulate the acquired wind field parameters, air pollutant concentration data, and building geometric characteristics to obtain the three-dimensional velocity field of airflow throughout the city. According to the air viscosity, turbulent kinetic energy, and turbulent dissipation rate of different urban areas, the standard k-ε turbulence model is used to calculate the turbulent diffusion effect, obtaining the turbulent diffusion coefficient of airflow in different urban areas. Based on the emission data of urban pollution sources, the emission rate of each pollution source is obtained. Combined with the pollutant diffusion coefficient and meteorological data, the Reynolds-averaged Navier-Stokes equations are used to simulate the changes in pollutant concentration to determine the distribution of pollutant concentration within the city.

[0015] Furthermore, the step of obtaining real-time meteorological data based on the real-time meteorological data API, calculating atmospheric stability, and determining the stability classification result based on a preset stability classification standard includes:

[0016] Real-time meteorological data is obtained using the real-time meteorological data API, and meteorological data at different time points is obtained by setting the request time interval; meteorological data at different altitudes of the atmosphere is extracted by calling the radar system interface and setting the monitoring altitude range; atmospheric stability is calculated using the Monin-Obukhov length formula based on ground temperature, wind speed, and atmospheric boundary layer height data; and the stability classification result is determined based on the atmospheric stability calculation results and a preset stability classification standard, including stable, neutral, and unstable.

[0017] Furthermore, the method employs large eddy simulation to determine the changing trends of airflow under different weather conditions, and uses the pollutant Lagrange particle tracking method to simulate the diffusion process of pollutants in the air, determining the concentration distribution of pollutants at different times and spatial locations, including:

[0018] Based on wind field parameters, historical meteorological data, real-time meteorological data, and atmospheric stability, the large eddy simulation method is used to calculate the airflow response under different weather conditions, obtaining the airflow variation trend under different weather conditions. Based on the coordinates of the pollution source, wind speed, pollutant diffusion coefficient, large eddy simulation results, and real-time meteorological data, the pollutant Lagrange particle tracking method is used to simulate the diffusion and movement process of pollutants in the air, obtaining the flow trajectory of pollutants. Combining the fluid dynamics simulation and large eddy simulation results, the diffusion path and concentration variation map of pollutants are generated, determining the concentration distribution of pollutants at different times and spatial locations.

[0019] Furthermore, the acquisition of public space layout data and real-time pedestrian density data within urban areas, the construction of a crowd behavior pattern classification model, the determination of crowd behavior patterns in different public spaces, and the prediction of pedestrian density in each public space within a preset future time period include:

[0020] Based on the urban planning database, GIS data analysis methods were used to obtain the location coordinates and spatial geometry of urban green spaces, parks, squares, and open streets. GIS software was used for data processing to divide the city's public spaces into functional zones, identify the location and boundaries of different public spaces, and determine public space layout data, including the location, area, shape, and interrelationships of each public space. Real-time video surveillance data from the city's video surveillance system was obtained to determine real-time pedestrian density data within the urban area. A convolutional neural network algorithm was used to identify pedestrian trajectories, dwell time, and walking speed. The pedestrian density data was combined with the public space layout data to obtain differences in pedestrian distribution across different public space types. Based on pedestrian trajectories, dwell time, and walking speed, a decision tree algorithm was used to train a model to construct a crowd behavior pattern classification model, determining crowd behavior patterns in different public spaces, including but not limited to gathering, loitering, and movement patterns. Based on historical pedestrian data, crowd behavior models, and meteorological data, a long short-term memory network algorithm was used to train a model to construct a crowd flow trend prediction model, predicting the pedestrian density of each public space within a preset future time period.

[0021] Furthermore, based on pollutant concentration, pedestrian density, air velocity, and wind direction data in green areas, the optimal configuration of green belts and ventilation corridors is determined. Combining turbulent diffusion and pollutant concentration gradients, the optimal public space ventilation optimization scheme is determined, including:

[0022] Based on pollutant concentration, pedestrian density, and air velocity data, an optimization model was constructed using a non-dominated sorting genetic algorithm. Multiple objective functions were set, including minimizing pollutant concentration, maximizing airflow efficiency, and optimizing the public space layout. The non-dominated sorting method was used to evaluate the merits of each candidate scheme, resulting in the optimal public space layout scheme, encompassing both pollutant concentration minimization and optimal airflow configuration. Based on green space distribution, wind direction changes, and air quality index data, a grid search method was employed. By setting the grid range, the location and width of green belts and ventilation corridors were gradually adjusted. After each adjustment, fluid dynamics simulations were used to determine airflow, pollutant concentration distribution, and air quality index. Each layout scheme was evaluated to assess its effectiveness in airflow and pollution diffusion, yielding the optimal configuration of green belts and ventilation corridors. Finally, based on the turbulent diffusion coefficient and pollutant concentration gradient, the convection-diffusion equation was used... Determine the diffusion rate J of the pollutant, where K t It is the turbulent diffusion coefficient. It is the pollutant concentration gradient; based on data of building spacing, wind speed, and air pollution diffusion rate, a multi-objective linear programming model is used to set multiple optimization objectives, including minimizing pollutant concentration and maximizing air flow efficiency. Constraints are established, including building spacing, wind speed range, and ventilation outlet location. The linear programming method is used to optimize the ventilation design of public spaces and determine the public space ventilation optimization scheme; based on the ventilation effect of different public space ventilation optimization schemes, the optimal public space ventilation optimization scheme is determined, including the optimal location of ventilation outlets, optimization of ventilation paths, and optimal configuration of air flow.

[0023] Furthermore, based on the optimized layout of green belts and ventilation corridors, building distribution, and wind field parameters, combined with airflow simulation and pollutant diffusion trajectories, the impact of different configurations on ventilation effects and pollution suppression is evaluated, and the ventilation paths, green belt distribution, and building spacing configurations in public spaces are optimized, including:

[0024] Based on the adjusted green belt and ventilation corridor layout using the optimal configuration, we obtain data on building distribution and geometric characteristics of green belts. Combining wind field parameters and air pollutant concentration data, we use fluid dynamics simulation to determine the airflow velocity field, pollutant diffusion trajectory, and air quality index, thus identifying airflow and pollutant concentration variations under different green belt and ventilation corridor configurations. Based on these variations, we evaluate the airflow optimization effect and pollutant diffusion suppression effect of different green belt configurations, and optimize the green belt and ventilation corridor design. Based on the adjusted public space ventilation optimization scheme and pedestrian density data, we evaluate the airflow effect and crowd activity patterns in public spaces, and optimize the spatial configuration and layout of public spaces. Based on the evaluation results of the green belt and ventilation corridor layout, and the evaluation results of the public space ventilation optimization scheme, we determine the optimal configuration of the public space layout, including ventilation paths, green belt layout, ventilation opening locations, green space distribution, and building spacing for each public space area.

[0025] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0026] This invention provides a method for optimizing urban public spaces based on crowd behavior simulation. By acquiring real-time wind field parameters, air pollutant concentrations, and building geometric characteristics, this invention comprehensively analyzes airflow characteristics, predicts airflow trends under different weather conditions, and combines historical meteorological data to predict turbulent kinetic energy and turbulent diffusion coefficients, simulating the dynamic diffusion process of pollutants. This invention uses real-time meteorological data to calculate atmospheric stability, dynamically predicts airflow trends using large eddy simulation, and simulates pollutant diffusion processes using the Lagrange particle tracking method, tracking pollutant diffusion in real time. This invention combines crowd behavior pattern data to dynamically adjust the layout of public spaces, ensuring good airflow and pollutant diffusion in high-density areas. Through atmospheric stability calculations and pollutant diffusion simulation, this invention can dynamically assess pollutant concentration distribution, optimize the configuration of green belts and ventilation corridors, and achieve efficient dilution and diffusion of pollutants. Combining crowd density data and behavior pattern classification, this invention optimizes the ventilation paths and spatial layout of public spaces, improving air quality and enhancing the comfort of crowd activities. This invention provides a more precise and efficient optimization solution for urban public space design by comprehensively considering multiple factors such as meteorology, building distribution, crowd behavior and air flow, promoting the sustainable development of the urban environment and significantly improving the quality of life of citizens and the comfort of public spaces. Attached Figure Description

[0027] Figure 1 This is a flowchart of an urban public space optimization method based on crowd behavior simulation according to the present invention;

[0028] Figure 2 This is a schematic diagram of an urban public space optimization method based on crowd behavior simulation according to the present invention;

[0029] Figure 3 This is another schematic diagram of an urban public space optimization method based on crowd behavior simulation according to the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0031] like Figure 1-3 This embodiment of a method for optimizing urban public spaces based on crowd behavior simulation may specifically include:

[0032] Step S101: Obtain urban wind field parameters and urban air pollutant concentration data through the meteorological station monitoring network, and use UAV lidar technology to obtain the geometric features of buildings.

[0033] Based on the meteorological station monitoring network, wind speed, humidity, and temperature sensors were used to acquire urban wind field parameters. The timestamps of each data point were recorded and uploaded to the urban air monitoring database. Wind field parameters included wind direction, wind speed, humidity, and temperature. Fixed and mobile PM2.5, NO2, and CO sensors were deployed to acquire urban air pollutant concentration data. The locations of monitoring points and monitoring periods were recorded, and the data was processed by time period using a sliding window method to obtain hourly, daily, and monthly pollutant concentration variation curves. By comparing these data with wind field parameters, spatiotemporal interpolation methods were used to fill in missing pollutant concentration data, resulting in a spatiotemporally continuous pollutant concentration distribution map. UAV lidar technology was used to scan urban building structures, acquiring the geometric features of buildings, including height, density, shape, and ground features.

[0034] For example, a set of meteorological sensors and pollutant monitoring equipment, such as wind speed sensors, humidity sensors, and temperature sensors, were deployed in the central area of ​​a city, and monitoring began at 8:00 AM on November 1, 2020. These sensors acquired real-time data including wind direction, wind speed, humidity, and temperature. For instance, at that time, the wind speed was 5 m / s, the wind direction was northerly, the humidity was 60%, and the temperature was 18°C. Simultaneously, PM2.5, NO2, and CO sensors were installed at multiple monitoring points throughout the city, including in the south, east, and west of the urban area. These sensors recorded the concentration of air pollutants hourly. At the southern monitoring point, the PM2.5 sensor measured a PM2.5 concentration of 35 micrograms per cubic meter, an NO2 concentration of 25 micrograms per cubic meter, and a CO concentration of 0.3 milligrams per cubic meter at 8:00 AM on November 1, 2020. These data were recorded with timestamps during the monitoring process and uploaded to the city's air monitoring database. The sliding window method was used to process hourly data into daily and monthly pollutant concentration variation curves. For example, the sliding window method yielded the PM2.5 concentration variation curve for the entire day of November 1, 2020, at this monitoring point, showing that the PM2.5 concentration remained roughly between 30-40 micrograms per cubic meter throughout the day. However, due to missing data at some monitoring points, complete pollutant concentration data could not be provided. To address this issue, a spatiotemporal interpolation method was used to interpolate the missing data based on wind field parameters from surrounding monitoring points, such as wind speed and direction, thereby generating a complete pollutant concentration distribution map. Furthermore, UAV LiDAR technology was used to scan the city's building structures, obtaining the geometric characteristics of each building. For instance, one building was 120 meters high, had a density of 30%, and was a multi-story structure. The surrounding ground features were a smooth urban plaza, suitable for data collection.

[0035] Step S102: Based on historical meteorological data of urban monitoring points, predict the turbulent kinetic energy of the monitoring points and calculate the turbulent dissipation rate. Use computational fluid dynamics simulation to obtain the three-dimensional velocity field of air flow within the urban area and calculate the turbulent diffusion coefficient of air flow in different urban areas.

[0036] Air viscosity is determined using the Sutherland formula based on air temperature and pressure. A turbulent kinetic energy prediction model is constructed by training a recurrent neural network using historical meteorological data from urban monitoring stations, along with corresponding turbulent kinetic energy. Meteorological data includes wind field parameters and air pressure. Turbulent kinetic energy at the monitoring stations is determined using the real-time meteorological data prediction model. Based on the turbulent kinetic energy, the standard turbulent dissipation rate formula is applied. Calculate the turbulent dissipation rate ε, where k is the turbulent kinetic energy, l is the characteristic length, is the size of the building, and C μ These are model constants, obtained through fitting historical data. Based on the turbulence model parameters, computational fluid dynamics simulations are used to simulate the acquired wind field parameters, air pollutant concentration data, and building geometry to obtain the three-dimensional velocity field of airflow throughout the city. According to the air viscosity, turbulent kinetic energy, and turbulent dissipation rate of different urban areas, the standard k-ε turbulence model is used to calculate the turbulent diffusion effect, obtaining the turbulent diffusion coefficient of airflow in different urban areas. Based on urban pollution source emission data, the emission rate of each pollution source is obtained. Combined with the pollutant diffusion coefficient and meteorological data, the Reynolds-averaged Navier-Stokes equations are used to simulate the changes in pollutant concentration, determining the pollutant concentration distribution within the city.

[0037] For example, in a city's air quality monitoring system, meteorological stations determine air viscosity by recording real-time temperature and air pressure data. Assuming that on November 2, 2020, the temperature was 20℃ and the air pressure was 101325 Pa, the Sutherland formula is used to calculate the air viscosity. According to this formula, the air viscosity is approximately 1.81 × 10⁻⁵ Pa·s. In the city's historical meteorological data, the turbulent kinetic energy k at the monitoring point is recorded as 0.5 m² / s². Combining the wind field parameters at the monitoring point, such as wind speed of 5.2 m / s and wind direction of southeast, as well as the air pressure data, a recurrent neural network is used to model the turbulent kinetic energy, constructing a turbulent kinetic energy prediction model. After training, the turbulent kinetic energy prediction model can predict turbulent kinetic energy based on real-time meteorological data. In a new time period, the turbulent kinetic energy prediction model predicts the turbulent kinetic energy at the monitoring point as 0.6 m² / s². Based on this turbulent kinetic energy and known building dimensions, such as a 50-meter-high building near the monitoring point, the standard turbulent dissipation rate formula is used. Calculate the turbulent dissipation rate ε, where k is the turbulent kinetic energy, l is the characteristic length, typically the size of a building (e.g., 50 meters), and C.μ The model constant, 0.09, was obtained by fitting historical data. Substituting the data, the turbulent dissipation rate ε was calculated to be 0.015 W / kg. Computational fluid dynamics simulations were performed using these turbulence model parameters. The simulation included acquired wind field parameters, pollutant concentrations (e.g., PM2.5 concentration of 35 μg / m3), and building geometry. The standard k-ε turbulence model was used to calculate the turbulent diffusion coefficient for various urban areas. In the city center, the turbulent diffusion coefficient was 0.12 m² / s, while in the suburbs it was 0.08 m² / s. By combining urban pollution source emission data, such as the NO2 emission rate of 10 g / s in an industrial area, the Reynolds-averaged Navier-Stokes equations were used to simulate the changes in pollutant concentrations. The simulation results show that the PM2.5 concentration is higher around the industrial area, approximately 45 μg / m3, while it is lower in the urban suburbs, only 25 μg / m3.

[0038] Step S103: Obtain real-time meteorological data according to the real-time meteorological data API, calculate atmospheric stability, and determine the stability classification result based on the preset stability classification standard.

[0039] Real-time meteorological data is acquired using a real-time meteorological data API, and meteorological data at different time points is obtained by setting request intervals. By calling the radar system interface and setting the monitoring altitude range, meteorological data at different atmospheric altitudes are extracted. Atmospheric stability is calculated using the Monin-Obukhov length formula based on ground temperature, wind speed, and atmospheric boundary layer height data. Based on the atmospheric stability calculation results and a preset stability classification standard, the stability classification is determined, including stable, neutral, and unstable.

[0040] Exemplarily, on November 3, 2020, meteorological data of a certain city was obtained using a real-time meteorological data API. The request time interval was set to once per hour through the API, and data such as the temperature, wind speed, and wind direction of the city were recorded. At 9 am, the real-time data showed that the temperature was 22°C, the wind speed was 3.5 m / s, the wind direction was southwest, and the air pressure was 101320 Pa. Next, the radar system interface was called, the monitoring altitude range was set from 0 to 500 meters, and meteorological data at different altitudes in the atmosphere was extracted. At an altitude of 100 meters, the temperature was 19°C and the wind speed was 4.1 m / s. At an altitude of 300 meters, the temperature dropped to 16°C and the wind speed was 5.2 m / s. Based on the ground temperature of 22°C, wind speed of 3.5 m / s, and atmospheric boundary layer height of 1000 meters, the Monin-Obukhov length formula was used to calculate the atmospheric stability. According to these input values, the calculated Monin-Obukhov length was 50 meters, which indicates that the atmosphere is in a relatively unstable state. Because according to the classification of the Monin-Obukhov length, when its value is small, when the Monin-Obukhov length > 200 meters, the atmosphere is usually in a stable state. In a stable situation, the vertical mixing of air currents is weak, the air is relatively calm, and the vertical diffusion of heat and pollutants is restricted, usually manifested as lower convective activities. When 100 meters < Monin-Obukhov length < 200 meters, the atmospheric stability is neutral. In this case, the vertical mixing ability of the atmosphere is moderate, and the diffusion of heat and pollutants is relatively balanced, not particularly strong and not completely restricted. When the Monin-Obukhov length < 100 meters, the atmosphere is usually in an unstable state. The vertical mixing of the atmosphere is strong, the vertical diffusion of heat and pollutants is obvious, convection is likely to occur, and the air flow is intense. In this case, the pollutants diffuse quickly, usually manifested as strong convective activities. The current Monin-Obukhov length is 50 meters, indicating that the atmosphere is relatively unstable and prone to convection. Therefore, the stability state of the atmosphere in this city is unstable. This unstable state usually means that the vertical mixing of air currents is strong, which is suitable for the diffusion of pollutants, but may also lead to drastic changes in local meteorological conditions. Under this condition, the concentration of pollutants may be affected by changes in wind speed and temperature, the diffusion effect is strong, and the air quality may change rapidly.

[0041] Step S104, use the large eddy simulation method to determine the change trend of air flow under different weather conditions, and use the Lagrangian particle tracking method for pollutants to simulate the diffusion process of pollutants in the air, and determine the concentration distribution of pollutants at different time and space positions.

[0042] Based on wind field parameters, historical meteorological data, real-time meteorological data, and atmospheric stability, the Large Eddy Simulation (LES) method is used to calculate the airflow response under different weather conditions, obtaining the changing trends of airflow under different weather conditions. Based on the coordinates of the pollution source, wind speed, pollutant diffusion coefficient, LES results, and real-time meteorological data, the pollutant Lagrange particle tracking method is used to simulate the diffusion and movement of pollutants in the air, obtaining the flow trajectory of pollutants. Combining the results of fluid dynamics simulation and LES, a diffusion path and concentration variation map of pollutants are generated, determining the concentration distribution of pollutants at different times and spatial locations.

[0043] For example, in a city's air quality monitoring system, meteorological data is collected regularly. On November 4, 2020, real-time wind field parameters and historical meteorological data for the city were acquired. At 8:00 AM, the real-time data showed a wind speed of 4 m / s, a southeasterly wind direction, a temperature of 20°C, and an air pressure of 101,300 Pa. Based on these data and historical meteorological data, atmospheric stability was calculated, resulting in a Monin-Obukhov length of 40 meters. This indicates that the atmosphere is unstable and prone to vertical convection. Based on atmospheric stability and wind field parameters, large eddy simulation (LES) was used to simulate the airflow in the city. Under these unstable weather conditions, the simulation results show that the airflow exhibits relatively intense turbulent activity, with wind speed increasing with altitude. Specifically, at a height of 100 meters above the ground, the wind speed increases from 4 m / s to 5.5 m / s, demonstrating strong vertical airflow fluctuations. A pollutant diffusion model was applied to the pollution source in this city. Assuming a pollution source in an industrial area in the north of the city, with an emission rate of 10 g / s, primarily emitting NO2, a pollutant diffusion coefficient of 0.1 m² / s and real-time meteorological data was used to simulate the airborne diffusion process of pollutants using the pollutant Lagrange particle tracking method. According to the large eddy simulation results, the diffusion of pollutants is influenced by the wind field. The flow trajectory shows that pollutants diffuse rapidly in areas with high wind speeds and move southeastward. Over time, the concentration of pollutants gradually diffuses from the source area to the surrounding areas. During real-time monitoring, it was identified that pollutants had begun to diffuse into the southern part of the city at 10:00 AM, reaching a concentration of 30 μg / m³, while the concentration near the source was 120 μg / m³. The simulation yielded a path and concentration variation map of the pollutant diffusion, showing the changes in pollutant concentration at different times and spatial locations. For example, at 11:00 AM, the pollutants had spread to an area approximately 2 kilometers from the pollution source, and the concentration in this area reached 50 μg / m³. Combining fluid dynamics and large eddy simulation results, the analysis reveals that in urban centers, pollutant diffusion is complex due to the dense distribution of buildings and the influence of wind speed. Strong fluctuations in airflow cause the diffusion paths of pollutants to exhibit a certain curved shape, while in open areas, pollutant diffusion is more uniform. Based on the fluid dynamics and large eddy simulation results, diffusion paths and concentration variation maps of pollutants are plotted, demonstrating the pollutant concentration distribution at different times and spatial locations.

[0044] Step S105: Obtain public space layout data and real-time pedestrian density data within the urban area, construct a crowd behavior pattern classification model, determine crowd behavior patterns in different public spaces, and predict pedestrian density in each public space within a preset time period in the future.

[0045] Based on the urban planning database, GIS data analysis methods were used to obtain the location coordinates and spatial geometry of urban green spaces, parks, squares, and open streets. GIS software was used for data processing to divide the city's public spaces into functional zones, identify the location and boundaries of different public spaces, and determine public space layout data, including the location, area, shape, and interrelationships of each public space. Real-time video surveillance data was obtained from the city's installed video surveillance system to determine real-time pedestrian density data within the urban area. A convolutional neural network algorithm was used to identify pedestrian trajectories, dwell time, and walking speed. Combining pedestrian density data with public space layout data revealed differences in pedestrian distribution across different public space types. Based on pedestrian trajectories, dwell time, and walking speed, a decision tree algorithm was used for model training to construct a crowd behavior pattern classification model, determining crowd behavior patterns in different public spaces, including but not limited to gathering, loitering, and movement patterns. Based on historical pedestrian data, crowd behavior models, and meteorological data, a long short-term memory network algorithm was used for model training to construct a crowd flow trend prediction model, predicting the pedestrian density of each public space within a preset future time period.

[0046] For example, on November 5, 2020, urban planning databases and GIS software were used to obtain public space layout data for a certain city. Through GIS analysis, the location coordinates and spatial geometry of green spaces, parks, squares, and open streets within the city were extracted. Specifically, in the city center, there is a park with an area of ​​15,000 square meters, surrounded by open streets and squares. Through GIS data processing, these public spaces were functionally zoned, determining the location, area, and shape of different public spaces. The relationships between them were also analyzed. For instance, the park connects with the surrounding open streets, forming a highly mobile public area, while the nearby square is a more static area, mainly used for leisure and social activities. Simultaneously, the city's video surveillance system collected real-time pedestrian density data. Through analysis of the surveillance video, a convolutional neural network algorithm was used to identify pedestrian trajectories, dwell time, and walking speed. At 10:00 AM on November 5, 2020, surveillance data showed approximately 500 people in the city center park area, with an average dwell time of 20 minutes and a walking speed of approximately 1.2 m / s. At the same time, there were approximately 300 people in the plaza area, with a dwell time of 30 minutes and a walking speed of 1.0 m / s. Based on this data, the differences in pedestrian flow distribution across different types of public spaces were analyzed. It was found that pedestrian flow in the park area was more dynamic, while pedestrian flow in the plaza area was more concentrated, with longer dwell times. Combining pedestrian trajectories, dwell times, and walking speeds, a decision tree algorithm was used for training to construct a crowd behavior pattern classification model. This model can identify crowd behavior patterns in different public spaces. For the park area, the model identified more flow patterns, i.e., people moving quickly through the space, while in the plaza area, it identified lingering and gathering patterns, indicating that people tend to stay and gather more. Combining historical pedestrian flow data, crowd behavior models, and meteorological data (e.g., a temperature of 23°C and a wind speed of 2 m / s on that day), a long short-term memory network algorithm was used to train the model, constructing a crowd flow trend prediction model. This model predicted that within the next 4 hours, i.e., until 2:00 PM on November 5, 2020, the pedestrian density in the city center park would increase to 700 people, while the density in the square would increase to 450 people. The model also showed that due to the warm weather and low wind speed, the crowd would mainly concentrate in the open areas of the park, while the crowd flow in the square would be smaller, mainly remaining in a stagnant state.

[0047] Step S106: Based on pollutant concentration, pedestrian density, air velocity, and green space wind direction data, determine the optimal configuration of green belts and ventilation corridors, and combine turbulent diffusion and pollutant concentration gradient to determine the optimal ventilation optimization scheme for public spaces.

[0048] Based on pollutant concentration, pedestrian density, and air velocity data, an optimization model was constructed using a non-dominated sorting genetic algorithm. Multiple objective functions were set, including minimizing pollutant concentration, maximizing airflow efficiency, and optimizing the public space layout. The non-dominated sorting method was used to evaluate the merits of each candidate scheme, resulting in the optimal public space layout scheme, encompassing both pollutant concentration minimization and optimal airflow configuration. Based on green space distribution, wind direction changes, and air quality index data, a grid search method was employed. By setting the grid range, the location and width of green belts and ventilation corridors were progressively adjusted. After each adjustment, fluid dynamics simulations were used to determine airflow, pollutant concentration distribution, and air quality index. Each layout scheme was evaluated to assess its effectiveness in airflow and pollution diffusion, yielding the optimal configuration of green belts and ventilation corridors. Based on the turbulent diffusion coefficient and pollutant concentration gradient, the convection-diffusion equation was used... Determine the diffusion rate J of the pollutant, where K t It is the turbulent diffusion coefficient. This involves the pollutant concentration gradient. Based on data on building spacing, wind speed, and air pollution diffusion rate, a multi-objective linear programming model is used to set multiple optimization objectives, including minimizing pollutant concentration and maximizing airflow efficiency. Constraints are established, including building spacing, wind speed range, and vent location. Linear programming methods are then used to optimize the ventilation design of public spaces, determining the optimal ventilation scheme. Based on the ventilation effects of different public space ventilation optimization schemes, the optimal public space ventilation optimization scheme is determined, including the best location of vents, optimization of ventilation paths, and optimal configuration of airflow.

[0049] For example, in an air quality optimization case study of a city center park, the monitoring system obtained the following data: PM2.5 concentration of 55 μg / m3, NO2 concentration of 45 μg / m3, wind speed of 3.5 m / s, wind direction of southeast, pedestrian density of 200 people (maximum density of 400 people), and temperature of 22℃. The goal is to construct an optimization model using a non-dominated sorting genetic algorithm to minimize pollutant concentrations, maximize airflow efficiency, and optimize the layout of public spaces. The genetic algorithm is used to optimize the park's layout, aiming to find a layout that effectively reduces PM2.5 and NO2 concentrations and improves airflow efficiency. Based on different candidate schemes, the genetic algorithm evaluates the merits of each scheme. The optimal scheme is obtained by adjusting the green belt, building spacing, ventilation opening location, and green space distribution, ultimately achieving the best configuration. In this scheme, the green belt in the center of the park is set at 5000 square meters, and the ventilation openings are located at the north and south ends of the park to maximize the use of natural wind direction to enhance airflow and reduce pollutant accumulation. Based on data on green space distribution, wind direction changes, and air quality index (AQI), a grid search method was used to further optimize the layout of green belts and ventilation corridors. By adjusting the width and location of the ventilation corridors and using fluid dynamics simulations to evaluate the effects of each adjustment, an optimized configuration of green belts and ventilation corridors was derived. In the simulation, the adjusted airflow path showed that the wind speed reached 4.2 m / s within the park, pollutant concentration decreased by 15% at the southern end of the park, and the AQI improved by 10%. Based on this, the convection-diffusion equation was used to calculate the pollutant diffusion rate J. If the turbulent diffusion coefficient is 0.12 m² / s and the pollutant concentration gradient is 0.5 μg / m³ / m, then the pollutant diffusion rate is... This indicates that the diffusion of pollutants was effectively controlled in the optimized configuration. A multi-objective linear programming model was used to optimize building spacing, wind speed, and ventilation opening locations. During the optimization process, several constraints were set, such as a building spacing of no less than 10 meters, a wind speed range of 2.5-5 m / s, and the layout of ventilation openings to avoid obstructing airflow as much as possible. Through linear programming optimization, the optimal ventilation scheme for public spaces was found, including the optimization of ventilation opening locations, ventilation paths, and the optimal configuration of airflow.

[0050] Step S107: Based on the optimized layout of green belts and ventilation corridors, building distribution and wind field parameters, combined with air flow simulation and pollutant diffusion trajectory, evaluate the impact of different configurations on ventilation effect and pollution suppression, and optimize the ventilation path of public space, green belt distribution and building spacing configuration.

[0051] Based on the adjusted green belt and ventilation corridor layout using the optimal configuration, data on building distribution and the geometric characteristics of green belts are obtained. Combined with wind field parameters and air pollutant concentration data, fluid dynamics simulations are used to determine the airflow velocity field, pollutant diffusion trajectories, and air quality index, thus identifying airflow and pollutant concentration variations under different green belt and ventilation corridor configurations. Based on these variations, the optimization effect of airflow and the suppression effect of pollutant diffusion under different green belt configurations are evaluated, and the design of green belts and ventilation corridors is optimized. Based on the adjusted public space ventilation optimization scheme and pedestrian density data, the airflow effect and crowd activity patterns in public spaces are evaluated, and the spatial configuration and layout of public spaces are optimized. Based on the evaluation results of the green belt and ventilation corridor layout, and the evaluation results of the public space ventilation optimization scheme, the optimal configuration of the public space layout is determined, including ventilation paths, green belt layout, ventilation opening locations, green space distribution, and building spacing for each public space area.

[0052] For example, in an airflow and pollutant control optimization project for a city center park, a new design scheme was obtained by optimizing the layout of green belts and ventilation corridors. This optimization scheme aims to achieve optimal airflow and pollutant control based on multiple factors such as wind field parameters, air pollutant concentration, building distribution, and pedestrian density. The new green belt layout is set with a width of 30 meters and a length of 500 meters, arranged around the park's main flow paths. Furthermore, the ventilation corridors are designed as two areas, each 50 meters wide and 200 meters deep, located at the east and west ends of the park, respectively, to guide and accelerate airflow. Based on this layout and combined with wind field data, the wind speed near the ventilation corridors reached 4 m / s, while the wind speed in the central area of ​​the park was 3.5 m / s, with a southeasterly wind direction. Using fluid dynamics simulations, the airflow velocity field and pollutant diffusion trajectories under different green belt and ventilation corridor configurations were analyzed. Simulation results show that the optimized layout effectively reduced PM2.5 concentration in the southern area of ​​the park by approximately 20%, specifically decreasing it from 55 μg / m³ to 42 μg / m³. Simultaneously, the overall air quality index of the park improved by 15%, reaching a healthier level. The pollutant diffusion path also exhibited a more uniform distribution, and the accumulation of pollutants was significantly suppressed through the guidance of the ventilation corridors. Based on this optimized configuration, the changes in airflow and pollutant concentration under different green belt and ventilation corridor layouts were evaluated. After optimization, the green belts not only effectively increased airflow but also effectively reduced the accumulation of high-concentration pollutants through the blocking effect of vegetation on airflow. Further simulations revealed that this configuration could reduce pollutant concentration in the park by 25% while maintaining relatively stable airflow, ensuring a clean air environment. Based on the ventilation optimization plan for public spaces and pedestrian density data, the airflow effect and crowd activity patterns within the park were further evaluated. If the maximum pedestrian density in the park is 400 people, under the optimized layout, the crowd activity pattern is primarily a flow pattern, meaning pedestrians mainly move along the green belts and ventilation corridors. Due to smooth airflow, crowd activity no longer exhibits excessive concentration or stagnation, improving the comfort of the public space. Combining the evaluation results of the green belt and ventilation corridor layout, the optimal configuration of the public space layout was determined. In this configuration, the park's green belts are evenly distributed on both sides of the main flow paths, with appropriate width and density. Ventilation openings are located in the middle and at the ends of the ventilation corridors to ensure unobstructed airflow. The building spacing is set at 20 meters to ensure smooth airflow without affecting crowd activities. This optimal solution not only significantly improves the park's air quality but also optimizes crowd mobility and comfort, achieving the dual optimization goals of airflow and public space design.

[0053] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the concept of this application. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for optimizing urban public spaces based on crowd behavior simulation, characterized in that, The method includes: By using a meteorological station monitoring network, urban wind field parameters and urban air pollutant concentration data are obtained, and UAV lidar technology is used to obtain the geometric features of buildings. Based on historical meteorological data from urban monitoring points, the turbulent kinetic energy at the monitoring points is predicted, and the turbulent dissipation rate is calculated. The three-dimensional velocity field of air flow within the urban area is obtained by computational fluid dynamics simulation, and the turbulent diffusion coefficient of air flow in different urban areas is calculated. Real-time meteorological data is obtained from the real-time meteorological data API, atmospheric stability is calculated, and the stability classification result is determined based on the preset stability classification standard. Large eddy simulation was used to determine the changing trends of airflow under different weather conditions, and the pollutant Lagrange particle tracking method was used to simulate the diffusion process of pollutants in the air, and to determine the concentration distribution of pollutants at different times and spatial locations. Acquire public space layout data and real-time pedestrian density data within urban areas, construct a crowd behavior pattern classification model, determine crowd behavior patterns in different public spaces, and predict pedestrian density in each public space within a preset time period in the future. Based on data on pollutant concentration, pedestrian density, air velocity, and wind direction in green areas, the optimal configuration of green belts and ventilation corridors is determined. Combining turbulent diffusion and pollutant concentration gradients, the optimal ventilation optimization scheme for public spaces is determined. Based on the optimized layout of green belts and ventilation corridors, building distribution and wind field parameters, combined with air flow simulation and pollutant diffusion trajectory, the impact of different configurations on ventilation effect and pollution suppression is evaluated, and the ventilation path of public space, green belt distribution and building spacing configuration are optimized.

2. The method according to claim 1, wherein, The process involves acquiring urban wind field parameters and urban air pollutant concentration data through a meteorological station monitoring network, and using UAV lidar technology to obtain the geometric features of buildings, including: Based on the meteorological station monitoring network, wind speed, humidity, and temperature sensors were used to acquire urban wind field parameters, record the timestamp of each data point, and upload them to the urban air monitoring database. The wind field parameters included wind direction, wind speed, humidity, and temperature. Fixed and mobile PM2.5, NO2, and CO sensors were deployed to acquire urban air pollutant concentration data, and the monitoring point locations and monitoring periods were recorded. The sliding window method was used to process the monitoring data by time period, obtaining hourly, daily, and monthly pollutant concentration variation curves. By comparing with the wind field parameters, spatiotemporal interpolation methods were used to fill in the missing pollutant concentration data, resulting in a pollutant concentration distribution map with spatiotemporal continuity. Unmanned aerial vehicle (UAV) lidar technology was used to scan urban building structures, acquiring the geometric features of buildings, including building height, density, shape, and ground features.

3. The method according to claim 1, wherein, Based on historical meteorological data from urban monitoring points, the turbulent kinetic energy at these points is predicted, and the turbulent dissipation rate is calculated. Computational fluid dynamics simulation is used to obtain the three-dimensional velocity field of airflow within the urban area, and the turbulent diffusion coefficient of airflow in different urban areas is calculated, including: Air viscosity is determined using the Sutherland formula based on air temperature and pressure. A turbulent kinetic energy prediction model is constructed by training a recurrent neural network using historical meteorological data from urban monitoring points and corresponding turbulent kinetic energy. Meteorological data includes wind field parameters and air pressure. Turbulent kinetic energy at monitoring points is determined using the real-time meteorological data prediction model. Based on the turbulent kinetic energy, the standard turbulent dissipation rate formula is applied. Calculate the turbulent dissipation rate ε, where k is the turbulent kinetic energy, l is the characteristic length, is the size of the building, and C μ These are model constants, obtained by fitting historical data. Based on the turbulence model parameters, computational fluid dynamics simulation is used to simulate the acquired wind field parameters, air pollutant concentration data, and building geometric characteristics to obtain the three-dimensional velocity field of airflow throughout the city. According to the air viscosity, turbulent kinetic energy, and turbulent dissipation rate of different urban areas, the standard k-ε turbulence model is used to calculate the turbulent diffusion effect, obtaining the turbulent diffusion coefficient of airflow in different urban areas. Based on the emission data of urban pollution sources, the emission rate of each pollution source is obtained. Combined with the pollutant diffusion coefficient and meteorological data, the Reynolds-averaged Navier-Stokes equations are used to simulate the changes in pollutant concentration to determine the distribution of pollutant concentration within the city.

4. The method according to claim 1, wherein, The process of obtaining real-time meteorological data from a real-time meteorological data API, calculating atmospheric stability, and determining the stability classification result based on a preset stability classification standard includes: Real-time meteorological data is obtained using the real-time meteorological data API, and meteorological data at different time points is obtained by setting the request time interval; meteorological data at different altitudes of the atmosphere is extracted by calling the radar system interface and setting the monitoring altitude range; atmospheric stability is calculated using the Monin-Obukhov length formula based on ground temperature, wind speed, and atmospheric boundary layer height data; and the stability classification result is determined based on the atmospheric stability calculation results and a preset stability classification standard, including stable, neutral, and unstable.

5. The method according to claim 1, wherein, The method employs large eddy simulation to determine the changing trends of airflow under different weather conditions, and uses the pollutant Lagrange particle tracking method to simulate the diffusion process of pollutants in the air, determining the concentration distribution of pollutants at different times and spatial locations, including: Based on wind field parameters, historical meteorological data, real-time meteorological data, and atmospheric stability, the large eddy simulation method is used to calculate the airflow response under different weather conditions, obtaining the airflow variation trend under different weather conditions. Based on the coordinates of the pollution source, wind speed, pollutant diffusion coefficient, large eddy simulation results, and real-time meteorological data, the pollutant Lagrange particle tracking method is used to simulate the diffusion and movement process of pollutants in the air, obtaining the flow trajectory of pollutants. Combining the fluid dynamics simulation and large eddy simulation results, the diffusion path and concentration variation map of pollutants are generated, determining the concentration distribution of pollutants at different times and spatial locations.

6. The method according to claim 1, wherein, The process of acquiring public space layout data and real-time pedestrian density data within urban areas, constructing a crowd behavior pattern classification model, determining crowd behavior patterns in different public spaces, and predicting pedestrian density in each public space within a preset future time period includes: Based on the urban planning database, GIS data analysis methods were used to obtain the location coordinates and spatial geometry of urban green spaces, parks, squares, and open streets. GIS software was used for data processing to divide the city's public spaces into functional zones, identify the location and boundaries of different public spaces, and determine public space layout data, including the location, area, shape, and interrelationships of each public space. Real-time video surveillance data from the city's video surveillance system was obtained to determine real-time pedestrian density data within the urban area. A convolutional neural network algorithm was used to identify pedestrian trajectories, dwell time, and walking speed. The pedestrian density data was combined with the public space layout data to obtain differences in pedestrian distribution across different public space types. Based on pedestrian trajectories, dwell time, and walking speed, a decision tree algorithm was used to train a model to construct a crowd behavior pattern classification model, determining crowd behavior patterns in different public spaces, including but not limited to gathering, loitering, and movement patterns. Based on historical pedestrian data, crowd behavior models, and meteorological data, a long short-term memory network algorithm was used to train a model to construct a crowd flow trend prediction model, predicting the pedestrian density of each public space within a preset future time period.

7. The method according to claim 1, wherein, The optimal configuration of green belts and ventilation corridors is determined based on data such as pollutant concentration, pedestrian density, air velocity, and wind direction in green areas. Combining turbulent diffusion and pollutant concentration gradients, the optimal ventilation optimization scheme for public spaces is determined, including: Based on pollutant concentration, pedestrian density, and air velocity data, an optimization model was constructed using a non-dominated sorting genetic algorithm. Multiple objective functions were set, including minimizing pollutant concentration, maximizing airflow efficiency, and optimizing the public space layout. The non-dominated sorting method was used to evaluate the merits of each candidate scheme, resulting in the optimal public space layout scheme, encompassing both pollutant concentration minimization and optimal airflow configuration. Based on green space distribution, wind direction changes, and air quality index data, a grid search method was employed. By setting the grid range, the location and width of green belts and ventilation corridors were gradually adjusted. After each adjustment, fluid dynamics simulations were used to determine airflow, pollutant concentration distribution, and air quality index. Each layout scheme was evaluated to assess its effectiveness in airflow and pollution diffusion, yielding the optimal configuration of green belts and ventilation corridors. Finally, based on the turbulent diffusion coefficient and pollutant concentration gradient, the convection-diffusion equation was used... Determine the diffusion rate J of the pollutant, where K t It is the turbulent diffusion coefficient. It is the pollutant concentration gradient; based on data of building spacing, wind speed, and air pollution diffusion rate, a multi-objective linear programming model is used to set multiple optimization objectives, including minimizing pollutant concentration and maximizing air flow efficiency. Constraints are established, including building spacing, wind speed range, and ventilation outlet location. The linear programming method is used to optimize the ventilation design of public spaces and determine the public space ventilation optimization scheme; based on the ventilation effect of different public space ventilation optimization schemes, the optimal public space ventilation optimization scheme is determined, including the optimal location of ventilation outlets, optimization of ventilation paths, and optimal configuration of air flow.

8. The method according to claim 1, wherein, Based on the optimized layout of green belts and ventilation corridors, building distribution, and wind field parameters, combined with airflow simulation and pollutant diffusion trajectories, the impact of different configurations on ventilation effects and pollution suppression is evaluated. The optimization of ventilation paths in public spaces, green belt distribution, and building spacing configurations includes: Based on the adjusted green belt and ventilation corridor layout using the optimal configuration, we obtain data on building distribution and geometric characteristics of green belts. Combining wind field parameters and air pollutant concentration data, we use fluid dynamics simulation to determine the airflow velocity field, pollutant diffusion trajectory, and air quality index, thus identifying airflow and pollutant concentration variations under different green belt and ventilation corridor configurations. Based on these variations, we evaluate the airflow optimization effect and pollutant diffusion suppression effect of different green belt configurations, and optimize the green belt and ventilation corridor design. Based on the adjusted public space ventilation optimization scheme and pedestrian density data, we evaluate the airflow effect and crowd activity patterns in public spaces, and optimize the spatial configuration and layout of public spaces. Based on the evaluation results of the green belt and ventilation corridor layout, and the evaluation results of the public space ventilation optimization scheme, we determine the optimal configuration of the public space layout, including ventilation paths, green belt layout, ventilation opening locations, green space distribution, and building spacing for each public space area.

Images