Simulation forecasting method for runoff pollutant transportation process in urban area

Abstract

The invention discloses an urban area runoff pollutant transportation process simulation forecasting method, which is characterized in that a building area runoff production calculation method under the influence of a wind field is provided, and the building area runoff production calculation method is coupled with an SWMM rainfall flood model; secondly, constructing a two-dimensional hydrodynamic model based on an FVCOM and a Lagrange particle model, coupling the two-dimensional hydrodynamic model with the SWMM rainfall flood model, and driving the two-dimensional hydrodynamic model by taking the overflow flow and the pollutant concentration simulated by the SWMM rainfall flood model as boundary conditions; simulating a transportation process of pollutants along with evolution of surface ponding by establishing a'concentration-quantity 'conversion relation between the pollutants and particles at grid units and nodes, and obtaining a time sequence pollutant concentration data set; and training and establishing a pollutant concentration prediction model based on the obtained data, and then predicting the pollutant concentration of each prediction point after specific time in an actual rainfall event. According to the method, the runoff pollutant concentration can be predicted, and the urban non-point source pollution risk under the rainstorm weather is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of urban stormwater management, and in particular to a method for simulating and forecasting the transport process of runoff pollutants in urban areas. Background Art

[0002] Driven by urbanization and global climate change, urban flooding and the accompanying runoff pollution caused by extreme rainstorms are becoming increasingly serious, threatening urban water quality and public health. Currently, the drainage system in most Chinese cities remains a traditional combined sewer system. Non-point source pollution from rainfall runoff and point source pollution from pipe overflows have become a major constraint on water quality. During heavy rain events, pollutants such as tire wear, pesticides, heavy metals, and suspended solids are carried and driven by rainfall runoff into urban lakes, rivers, and wetlands, exacerbating environmental and ecological risks. Therefore, developing a timely and accurate forecasting method for the transport of runoff pollutants is crucial for maintaining and improving urban water quality and reducing the risk of non-point source pollution.

[0003] Rainfall runoff and pollutant transport patterns in highly urbanized areas are influenced by numerous factors, including topography and meteorological and hydrological factors. Studies have shown that major Chinese cities experience significant synchronicity between rain and wind. In stormy environments, raindrops are affected by wind resistance, gaining horizontal velocity and resulting in oblique rainfall, which significantly impacts rainfall runoff processes in urban areas. Complex, uneven urban terrain disrupts local wind patterns, affecting the spatial distribution of rainfall. Numerous studies have shown that spatial variability in rainfall can translate into significant flow variations. Compared to natural watersheds, buildings are a significant factor influencing runoff generation and runoff in urban areas. Due to the shielding effect of buildings, during oblique rainfall, rainfall that would normally fall on the leeward surface of buildings instead falls on the building's side walls. This results in runoff generation in urban built-up areas characterized by a combination of building side walls, roofs, and the ground surface. The runoff coefficient from building side walls is typically no less than that from the ground (permeable pavement). Runoff from building side walls typically enters the underground drainage network directly through surrounding diffused water systems. The combined effects of these factors result in an increase in the runoff coefficient from built-up areas in windy conditions compared to calm conditions. Furthermore, wind and rainfall are generally non-constant, causing the slope of rainfall to exhibit time-varying characteristics. Changes in rainfall slope can alter the distribution of rainwater between building walls and the ground surface, thereby changing the runoff characteristics of built-up areas and causing the runoff coefficient in urban built-up areas to exhibit significant time-varying characteristics.

[0004] Suspended particulate matter (SS) in stormwater runoff is one of the primary pollutants in urban surface runoff and serves as a carrier for most pollutants in stormwater runoff, such as COD, TN, TP, and heavy metals. The concentration and form of SS influence the characteristics of other pollutants in stormwater runoff. Numerous hydrological, hydrodynamic, and water quality models have been developed to simulate urban flooding and the transport of runoff pollutants, particularly SS. For example, the open-source SWMM model developed by the US Environmental Protection Agency is widely used to calculate urban stormwater runoff and the movement of pollutants within drainage networks. However, it cannot directly simulate the hydrodynamics of overland flow and the associated pollutant transport processes. Two-dimensional hydrodynamic models address these limitations, but they cannot capture the hydrodynamics of underground pipe networks. Consequently, a single model cannot fully describe urban hydrodynamics and water pollution processes. Therefore, researchers worldwide have employed commercial models or independently developed numerical models to investigate the spatiotemporal distribution and concentration variations of runoff pollutants at the watershed and city scales. A coupled one- and two-dimensional hydrodynamic and pollutant model is needed to better describe the hydrodynamic and water quality exchanges between the drainage system and the two-dimensional urban surface, the guiding and blocking effects of topography and buildings on water flow, and the resulting changes in pollutant migration patterns. However, existing models and methods lack consideration of wind fields, a key factor in calculating rainfall runoff and pollutant transport.

[0005] The transport of particulate pollutants is essentially a Lagrangian process, and the transport of pollutants through pipe networks and on the surface can be simulated by coupling a one- and two-dimensional hydrodynamic model with a Lagrangian particle model. Furthermore, even with the use of currently advanced GPU processors for parallel computing, the computational efficiency of physics-based hydrodynamic and water quality numerical models still struggles to meet the timeliness requirements of early warnings in real-world scenarios. Therefore, with the rapid development of artificial intelligence technology, the ability to combine hydrodynamic and water quality numerical models with deep learning algorithms to simulate and predict pollutant transport processes is a technology that urgently needs to be implemented and improved. Summary of the Invention

[0006] In order to overcome the above-mentioned defects, the technical problem to be solved by the present invention is: how to provide a method for simulating and forecasting the runoff pollutant transport process in urban areas, which can quickly predict and evaluate the concentration of pollutants in urban rainfall runoff during rainfall, so that it has good timeliness and prediction accuracy, and better provide technical support for improving the quality of urban water environment and reducing the risk of non-point source pollution.

[0007] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0008] A method for simulating and predicting the transport process of runoff pollutants in urban areas is characterized by establishing a SWMM (Stormwater Flood Management Model) stormwater model for a target urban area to simulate the rainfall runoff and pipe network hydraulic and water quality processes in the drainage area of ​​the target area; then constructing a two-dimensional hydrodynamic model based on FVCOM and Lagrangian particle model, and coupling it with the established SWMM stormwater model; using the overflow flow and pollutant concentration obtained by SWMM stormwater model simulation as boundary conditions to drive the two-dimensional hydrodynamic model; by establishing a "concentration-quantity" conversion relationship between pollutants and particles at grid cells and nodes, simulating the transport process of pollutants as surface water evolves, and obtaining a time series pollutant concentration data set; and establishing a pollutant concentration prediction model based on the obtained data for training, and then predicting the pollutant concentration after a specific time at each prediction point in an actual rainfall event.

[0009] The present invention specifically comprises the following steps:

[0010] S1. Establish a SWMM (Stormwater Management Model) stormwater model for the target urban area to simulate rainfall runoff and pipe network hydraulic and water quality processes in the drainage area of ​​the target area;

[0011] S2. Based on the topographic data of the target area, a two-dimensional hydrodynamic model based on FVCOM and the Lagrangian particle model was constructed. This model was then coupled with the SWMM stormwater model. The overflow flow and pollutant concentrations simulated by the SWMM stormwater model were used as boundary conditions to drive the two-dimensional hydrodynamic model. By establishing a "concentration-mass" conversion relationship between pollutants at grid cells and nodes, the transport process of pollutants as surface water evolves was simulated to obtain a time-series pollutant concentration dataset.

[0012] S3. Filtering rainfall events that have caused waterlogging from the historical rainfall events recorded in the target area's hydrological and meteorological database, obtaining corresponding rainfall information data, and obtaining a set of rainfall information data that may have caused waterlogging. The rainfall information data includes various rainfall-related data that directly affect surface water collection, including but not limited to rainfall amount, rainfall duration, and rainfall type.

[0013] S4. Based on the rainfall information data set obtained in S3, randomly interweave and combine the rainfall information required for simulating several single rainfall events. These are then input into the rainfall flood model established in S2 to simulate a single rainfall process, thereby obtaining a time-series pollutant concentration dataset. The time-series pollutant concentration dataset includes rainfall information data corresponding to each set time point during the rainfall process, as well as pollutant concentration data for the target grid.

[0014] S5. Using the rainfall information data at each time point as the input factor and the target grid pollutant concentration data after a specific time point (i.e., the time of evaluation and prediction) as the output factor, a training set and a test set are selected from the time series pollutant concentration dataset. A long-short-term memory artificial neural network model is used for training and testing to form a pollutant concentration prediction model.

[0015] S6. In an actual rainfall event, the rainfall information data at the current time point is used as an input factor and input into the pollutant concentration prediction model obtained in S5 to predict the pollutant concentration data after a specific time at each prediction point.

[0016] Furthermore, in step S1, in the established SWMM stormwater model, a method for calculating runoff generation on the building sidewalls under the influence of the wind field is coupled to the calculation of rainfall runoff generation and convergence.

[0017] The calculation method for building side wall runoff under the influence of the wind field is as follows: For building side wall runoff caused by inclined rainfall, under the influence of wind, the runoff flow rate of the building side wall driven by the horizontal component of rainfall can be calculated by the following formula:

[0018]

[0019] r h =r w tanθ

[0020] Where Q h is the flow rate of the building side wall, unit is m 3 / s;r w and r h are the vertical and horizontal components of the inclined rainfall intensity, in mm / h; f wall is the infiltration rate of the vertical surface of the building side wall, in mm / h, and is determined by referring to the relevant provisions in the Technical Specification for Rainwater Control and Utilization Engineering in Buildings and Residential Areas (GB50400-2016); b and l are the width and length of the building, respectively, in meters; n is the number of buildings in the building area; is the angle between the wind speed direction and the normal of the long side of the building, and its value range is [0,π / 2]; θ is the rainfall inclination angle in degrees; h is the net height of the building in meters; It is the effective “rain-receiving width” of the building side wall in the direction of wind speed.

[0021] In this way, the runoff generation in the built-up area under the influence of the wind field is taken into account and coupled in the established SWMM stormwater model, making the simulated rainfall surface runoff process in urban areas more realistic and accurate, and better improving the reliability of the prediction.

[0022] Furthermore, for the part of the surface on the leeward side of the building where there is no rainfall runoff due to the obstruction of adjacent buildings, the area of ​​this part of the area is deducted when the subcatchment area parameter of the SWMM stormwater model is selected.

[0023] Furthermore, when calculating the sidewall runoff and obstructed area, the coordinate system of the building and terrain data is aligned, and the total height of the building is equal to the net height of the building itself plus the terrain value of the terrain grid closest to the geometric center of the building bottom surface. The area of ​​the obstructed area between the buildings is further calculated based on this. The calculation process includes the following steps:

[0024] If the distance between two adjacent buildings is l d Greater than or equal to (h+zz a )tanθ, then the area S of the blocked area is:

[0025]

[0026] Where z and z0 are the terrain heights of the building, in meters. At this time, the windward side of the building on the right, that is, its left wall, can be fully affected by rainfall, and the flow rate is Q h ;

[0027] On the contrary, if the distance between two adjacent buildings is l d is less than (h+z-z0)tanθ, then the area S of the blocked area is:

[0028] At this time, the windward side of the building on the right is partially blocked, and the rain-receiving area of ​​the side wall needs to be calculated. The rain-receiving height h′0 of its side wall is:

[0029]

[0030] Its rain-receiving area S′0 is:

[0031]

[0032] Then the flow rate generated by the side wall of the right building is:

[0033] Q′ h =r h nS′0-f wall n(b+l)h′0.

[0034] In this way, similar analytical methods can be used to derive runoff calculation formulas for other situations where the height difference between adjacent buildings and terrain exists. All buildings are numbered according to the terrain grid number corresponding to the geometric center coordinates of their bases, and their long side orientations (due north is 0°) and facade geometric dimensions are counted to facilitate calculation using the program.

[0035] Furthermore, when calculating the area S of the blocked region, the average value of S calculated during the forecast period is used.

[0036] This is because S will change with wind speed and direction. For example, an increase in wind speed will increase the rainfall inclination angle, which may cause S to increase. Vice versa, the impact of wind direction changes is more complicated. In the SWMM stormwater model, the subcatchment area is a static parameter. Therefore, using the average value of S calculated during the analysis period when deducting it can make the final calculation result more accurate.

[0037] Furthermore, in S1, for buildings with drainage ditches around their exterior walls that enter the catchment network, after the flow generation calculation of the building side walls, the flow enters the associated node of the corresponding catchment area in the SWMM stormwater model in the form of node inflow; for buildings with green spaces or open spaces around their exterior walls, coupling is achieved by adding their flow sequence to the boundary condition file of the FVCOM model for calculation.

[0038] This allows for better coupling of calculated building sidewall runoff results into the model for different scenarios. Specifically, if building sidewall runoff enters the pipe network through drainage ditches arranged along the building's exterior walls, it must be fed into the SWMM model as a node inflow for pipe drainage hydraulic calculations. If it is dispersed into green spaces or open spaces surrounding the building's exterior walls and does not enter the pipe network, it becomes surface runoff and serves as a boundary condition in the two-dimensional hydrodynamic model to simulate the surface runoff process. This allows for effective simulation of various scenarios.

[0039] Furthermore, in S2, the overflow volume and pollutant concentration at the overflow node obtained by SWMM stormwater model simulation are used as boundary conditions of the two-dimensional hydrodynamic model. The pollutant concentration at the overflow node is converted into a certain number of particles by the following formula:

[0040]

[0041] Where, is the overflow node i at time j n The number of particles at , which is used as input data for the Lagrangian particle model; is the overflow node i at time j n Pollutant concentration at the site (mg / L), calculated by the SWMM stormwater model; m p is the mass of a single particle (kg); is the overflow node i at time j n The volume of water accumulated on the corresponding grid unit (m 3 );

[0042] During the transport of pollutants on the surface along with the evolution of accumulated water, their concentration can be converted from the number of particles to concentration using the following formula:

[0043]

[0044] Where, is the grid cell i at time j c Pollutant concentration at the site (mg / L); is the grid cell i at time j c The number of particles at is the grid cell i at time j c Volume of water on the 3 ).

[0045] In the SWMM stormwater model, the concentration of runoff pollutants is determined during the simulation process. This is determined by determining the surface land use type, accumulation and erosion calculation parameters, and other simulation factors based on actual conditions. However, the traditional SWMM model can only simulate the changes in pollutants within the pipe, but cannot simulate their transport with water on the surface. Therefore, one of the innovations of this solution is to couple SWMM, FVCOM, and the Lagrangian particle model. SWMM first simulates the overflow flow and pollutant concentration at the overflow manhole node. The node overflow flow serves as a boundary condition to drive FVCOM to simulate the flow of the overflow flow on the surface. The node overflow pollutant concentration is converted to the number of particles at the node. The hydrodynamic flow field provided by FVCOM drives the particle transport on the surface, thereby simulating the surface transport process of pollutant particles, making it possible to predict the pollution situation after a specific time at each prediction point.

[0046] Compared with the prior art, the present invention has at least the following beneficial effects:

[0047] 1. In the process of calculating runoff generated by rainfall in urban areas, the runoff generated by building side walls under the influence of wind fields is considered. Compared with the traditional calculation method that ignores the influence of wind fields, the physical mechanism of the present invention is more sound, so it can better improve the accuracy and reliability of simulation calculations.

[0048] 2. Since traditional hydrological models, such as the SWMM stormwater model, cannot simulate the transport process of runoff pollutants on the surface as the accumulated water evolves, the present invention proposes a coupled simulation method for the runoff pollutant transport process based on SWMM-FVCOM and the Lagrangian particle model. On the one hand, it can more accurately depict the dynamic change process of pollutant concentration on the surface as the accumulated water flows. On the other hand, by describing and visualizing the transport process of pollutants in the form of particles, it can more vividly and intuitively display their movement process on the surface. It can be applied to the field of dynamic monitoring and visualization of pollutants and has good practical application value.

[0049] 3. Based on the fusion of physical mechanisms and data-driven approaches, a mechanistic model is constructed to ensure the physical rationality of the simulation process. An LSTM network is used to capture complex nonlinear relationships, forming a comprehensive modeling framework combining physical mechanisms with temporal features. Tests have shown that this hybrid model achieves 25% higher prediction accuracy than a purely data-driven model.

[0050] 4. In terms of prediction efficiency, through model coupling and parallel computing optimization, the prediction time is shortened from hours in traditional mechanism models to minutes, meeting the needs of real-time prediction.

[0051] 5. Wide applicability: The present invention is not only applicable to mountainous cities, but can also be applied to different types of urban areas such as plain cities and coastal cities, providing a universal technical solution to solve the problem of urban waterlogging in my country. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. The drawings described below are only some embodiments of the present invention.

[0053] Figure 1 It is a schematic diagram of the principle of the present invention when calculating the flow generation of the building side wall.

[0054] Figure 2 It is a schematic flow diagram of the method of the present invention. DETAILED DESCRIPTION

[0055] In the following description, specific details are provided to provide a clearer understanding of the present invention. However, it will be apparent to those skilled in the art that the present invention may be practiced without one or more of these details. In other instances, some technical features well known in the art are not described to avoid confusion with the present invention.

[0056] See also Figure 1-2This embodiment discloses a method for simulating and predicting the transport process of runoff pollutants in urban areas. The method is characterized in that a SWMM (Stormwater Flood Management Model) stormwater model is established for a target urban area to simulate the rainfall runoff and pipe network hydraulic and water quality processes in the drainage area of ​​the target area; then a two-dimensional hydrodynamic model based on FVCOM and Lagrangian particle model is constructed and coupled with the established SWMM stormwater model. The overflow flow and pollutant concentration obtained by SWMM stormwater model simulation are used as boundary conditions to drive the two-dimensional hydrodynamic model. By establishing a "concentration-quantity" conversion relationship between pollutants and particles at grid units and nodes, the transport process of pollutants as the surface water evolves is simulated to obtain a time series pollutant concentration data set; and a pollutant concentration prediction model is established based on the obtained data for training, and then the pollutant concentration after a specific time at each prediction point is predicted in an actual rainfall event.

[0057] The specific implementation includes the following steps, see Figure 2 :

[0058] S1. Collect and process data on the topography, pipeline network, and land use types of the target urban area, and establish a SWMM (Stormwater Management Model) stormwater model for the target urban area (the specific modeling process itself is common knowledge in the field and will not be described in detail here). This model is used to simulate rainfall runoff and pipeline network hydraulic and water quality processes in the drainage area of ​​the target area.

[0059] S2. Based on the topographic data of the target area, a two-dimensional hydrodynamic model based on FVCOM and the Lagrangian particle model was constructed. This model was then coupled with the SWMM stormwater model. The overflow flow and pollutant concentrations simulated by the SWMM stormwater model were used as boundary conditions to drive the two-dimensional hydrodynamic model. By establishing a "concentration-mass" conversion relationship between pollutants at grid cells and nodes, the transport process of pollutants as surface water evolves was simulated to obtain a time-series pollutant concentration dataset.

[0060] S3. Filtering rainfall events that have caused waterlogging from the historical rainfall events recorded in the target area's hydrological and meteorological database, obtaining corresponding rainfall information data, and obtaining a set of rainfall information data that may have caused waterlogging. The rainfall information data includes various rainfall-related data that directly affect surface water collection, including but not limited to rainfall amount, rainfall duration, and rainfall type.

[0061] S4. Based on the rainfall information data set obtained in S3, randomly interweave and combine the rainfall information required for simulating several single rainfall events. These are then input into the rainfall flood model established in S2 to simulate a single rainfall process, thereby obtaining a time-series pollutant concentration dataset. The time-series pollutant concentration dataset includes rainfall information data corresponding to each set time point during the rainfall process, as well as pollutant concentration data for the target grid.

[0062] S5. Using the rainfall information data at each time point as the input factor and the target grid pollutant concentration data after a specific time point (i.e., the time of evaluation and prediction) as the output factor, a training set and a test set are selected from the time series pollutant concentration dataset. A long-short-term memory artificial neural network model is used for training and testing (the establishment, training, and testing processes of the long-short-term memory artificial neural network model are conventional techniques in the art and are not described in detail here), thereby forming a pollutant concentration prediction model.

[0063] S6. In an actual rainfall event, the rainfall information data at the current time point is used as an input factor and input into the pollutant concentration prediction model obtained in S5 to predict the pollutant concentration data after a specific time at each prediction point.

[0064] During implementation, in step S1, in the established SWMM stormwater model, the calculation method of building side wall runoff under the influence of wind field is coupled to the calculation of rainfall runoff.

[0065] The calculation method for the flow generated by the building side wall under the influence of the wind field is as follows (see Figure 1 ): For the runoff generated on the building side wall due to oblique rainfall, under the influence of wind, the runoff flow on the building side wall driven by the horizontal component of rainfall can be calculated by the following formula:

[0066]

[0067] r h =r w tanθ

[0068] Where Q h is the flow rate of the building side wall, unit is m 3 / s;r w and r h are the vertical and horizontal components of the inclined rainfall intensity, in mm / h; f wall is the infiltration rate of the vertical surface of the building side wall, in mm / h, and is determined by referring to the relevant provisions in the Technical Specification for Rainwater Control and Utilization Engineering in Buildings and Residential Areas (GB50400-2016); b and l are the width and length of the building, respectively, in meters; n is the number of buildings in the building area; is the angle between the wind speed direction and the normal of the long side of the building, and its value range is [0,π / 2]; θ is the rainfall inclination angle in degrees; h is the net height of the building in meters; It is the effective “rain-receiving width” of the building side wall in the direction of wind speed.

[0069] In this way, the runoff generation in the built-up area under the influence of the wind field is taken into account and coupled in the established SWMM stormwater model, making the simulated rainfall surface runoff process in urban areas more realistic and accurate, and better improving the reliability of the prediction.

[0070] During implementation, if there is a part of the surface on the leeward side of the building where there is no rainfall runoff due to the obstruction of adjacent buildings, the area of ​​this part of the surface will be deducted when the subcatchment area parameter of the SWMM stormwater model is selected.

[0071] During implementation, when calculating the sidewall runoff and blocked area, the coordinate system of the building and terrain data is aligned. The total height of the building is equal to the net height of the building itself plus the terrain value of the terrain grid closest to the geometric center of the building bottom surface. The area of ​​the blocked area between buildings is further calculated based on this. The calculation process includes the following steps:

[0072] If the distance between two adjacent buildings is l d is greater than or equal to (h+z-z0)tanθ, then the area S of the occluded region is:

[0073]

[0074] Where z and z0 are the terrain heights of the building, in meters. At this time, the windward side of the building on the right, that is, its left wall, can be fully affected by rainfall, and the flow rate is Q h ;

[0075] On the contrary, if the distance between two adjacent buildings is l d is less than (h+z-z0)tanθ, then the area S of the blocked area is:

[0076]

[0077] At this time, the windward side of the building on the right is partially blocked, and the rain-receiving area of ​​the side wall needs to be calculated. The rain-receiving height h′0 of its side wall is:

[0078]

[0079] Its rain-receiving area S′0 is:

[0080]

[0081] Then the flow rate generated by the side wall of the right building is:

[0082] Q′ h =r h nS′0-f wall n(b+l)h′0.

[0083] In this way, similar analytical methods can be used to derive runoff calculation formulas for other situations where the height difference between adjacent buildings and terrain exists. All buildings are numbered according to the terrain grid number corresponding to the geometric center coordinates of their bases, and their long side orientations (due north is 0°) and facade geometric dimensions are counted to facilitate calculation using the program.

[0084] During implementation, when calculating the area S of the obscured region, the average value of S calculated during the forecast period is used.

[0085] This is because S will change with wind speed and direction. For example, an increase in wind speed will increase the rainfall inclination angle, which may cause S to increase. Vice versa, the impact of wind direction changes is more complicated. In the SWMM stormwater model, the subcatchment area is a static parameter. Therefore, using the average value of S calculated during the analysis period when deducting it can make the final calculation result more accurate.

[0086] During implementation, in S1, for buildings with drainage ditches arranged around their exterior walls entering the catchment network, after the flow generation calculation of the building side walls, the flow enters the associated nodes of the corresponding catchment area in the SWMM stormwater model in the form of node inflow; for buildings with green spaces or open spaces around their exterior walls, coupling is achieved by adding their flow sequences to the boundary condition file of the FVCOM model for calculation.

[0087] This allows for better coupling of calculated building sidewall runoff results into the model for different scenarios. Specifically, if building sidewall runoff enters the pipe network through drainage ditches arranged along the building's exterior walls, it must be fed into the SWMM model via node inflow for pipe drainage hydraulic calculations. If it is dispersed into green spaces or open spaces surrounding the building's exterior walls and does not enter the pipe network, it becomes surface runoff and serves as the boundary condition of the two-dimensional hydrodynamic model (prepared according to the FVCOM model groundwater input file format; see the model manual for specific format instructions) to simulate the surface runoff convergence process. This allows for effective simulation of various scenarios.

[0088] In addition, during implementation, in S1, the land use type data required to construct the SWMM stormwater model to simulate the accumulation and washout of pollutants in storm runoff can be obtained free of charge through the StarCloud Data Service Platform (https: / / data-starcloud.pcl.ac.cn / ).

[0089] During implementation, in S2, the overflow volume at the overflow node obtained by SWMM stormwater model simulation can be prepared according to the groundwater input file format of the FVCOM model as the boundary condition driving the FVCOM model. For specific format instructions, please refer to the model manual, and the groundwater module should be turned on in the model operation management file (_run.nml).

[0090] During implementation, in S2, when compiling the FVCOM model, you can uncomment 'FlAG_31=-DLAG_PARTICLE' in the make.inc file in the source code folder FVCOM_source to couple the Lagrangian particle model.

[0091] During implementation, in S2, the overflow volume and pollutant concentration at the overflow node obtained by SWMM stormwater model simulation are used as the boundary conditions of the two-dimensional hydrodynamic model. The input file of the Lagrangian particle model needs to be prepared in accordance with the file format specified in the model manual. The number of pollutant particles at the overflow node required in the preparation process is converted from the pollutant concentration at the overflow node obtained by SWMM stormwater model simulation according to the following formula:

[0092]

[0093] Where, is the overflow node i at time j n The number of particles at , which is used as input data for the Lagrangian particle model; is the overflow node i at time j n Pollutant concentration at the site (mg / L), calculated by the SWMM stormwater model; m p is the mass of a single particle (kg); is the overflow node i at time j n The volume of water accumulated on the corresponding grid unit (m 3 ).

[0094] During implementation, in S2, the Lagrangian particle model is used to simulate the transport of pollutants on the surface as the water accumulates. The number of particles in the grid cells of the two-dimensional hydrodynamic model can be converted into concentration using the following formula:

[0095]

[0096] Where, is the grid cell i at time j c Pollutant concentration at the site (mg / L); is the grid cell i at time j c The number of particles at is the grid cell i at time j c Volume of water on the 3 ).

[0097] For implementation, the model information in this application is: software version: FVCOM (https: / / www.fvcom.org / ). Time step: 0.05 s, subject to the CFL stability condition. Manning's coefficient and infiltration parameters are adjusted according to actual conditions.

Claims

1. A method for simulating and predicting the transport of runoff pollutants in urban areas, characterized in that: A SWMM stormwater model was established for the target urban area to simulate the rainfall runoff and pipe network hydraulic and water quality processes in the drainage area of ​​the target area. A two-dimensional hydrodynamic model based on FVCOM and Lagrangian particle model was then constructed and coupled with the established SWMM stormwater model. The overflow flow and pollutant concentrations simulated by the SWMM stormwater model were used as boundary conditions to drive the two-dimensional hydrodynamic model. By establishing a "concentration-number" conversion relationship between pollutants and particles at grid cells and nodes, the transport process of pollutants as surface water evolved was simulated to obtain a time-series pollutant concentration dataset. A pollutant concentration prediction model was established based on the obtained data, and the pollutant concentrations at each prediction point after a specific time were predicted in actual rainfall events.

2. The method for simulating and predicting the transport of runoff pollutants in urban areas according to claim 1, characterized in that: The specific steps include: S1. Develop a SWMM stormwater model for the target urban area to simulate the runoff and water quality processes of the drainage network in the target area. S2. Based on the topographic data of the target area, a two-dimensional hydrodynamic model based on FVCOM and the Lagrangian particle model was constructed. This model was then coupled with the SWMM stormwater model. The overflow flow and pollutant concentrations simulated by the SWMM stormwater model were used as boundary conditions to drive the two-dimensional hydrodynamic model. By establishing a "concentration-mass" conversion relationship between pollutants at grid cells and nodes, the transport process of pollutants as they evolve with surface water was simulated, resulting in a time-series pollutant concentration dataset. S3. Filtering rainfall events that have caused waterlogging from the historical rainfall events recorded in the target area's hydrological and meteorological database, obtaining corresponding rainfall information data, and obtaining a set of rainfall information data that may have caused waterlogging. The rainfall information data includes various rainfall-related data that directly affect surface water collection, including but not limited to rainfall amount, rainfall duration, and rainfall type. S4. Based on the rainfall information data set obtained in S3, randomly interweave and combine the rainfall information required for simulating several single rainfall events. These are then input into the rainfall flood model established in S2 to simulate a single rainfall process, thereby obtaining a time-series pollutant concentration dataset. The time-series pollutant concentration dataset includes rainfall information data corresponding to each set time point during the rainfall process, as well as pollutant concentration data for the target grid. S5. Using the rainfall information data at each time point as the input factor and the pollutant concentration data of the target grid at a specific time after the time point as the output factor, a training set and a test set are selected from the time series pollutant concentration dataset. The training and testing are performed using a long-short-term memory artificial neural network model to form a pollutant concentration prediction model. S6. In an actual rainfall event, the rainfall information data at the current time point is used as an input factor and input into the pollutant concentration prediction model obtained in S5 to predict the pollutant concentration data after a specific time at each prediction point.

3. The method for simulating and predicting the transport of runoff pollutants in urban areas according to claim 2, wherein: In step S1, in the established SWMM stormwater model, the calculation method of building sidewall runoff under the influence of wind field is coupled to the calculation of rainfall runoff. The calculation method for building side wall runoff under the influence of the wind field is as follows: For building side wall runoff caused by inclined rainfall, under the influence of wind, the runoff flow rate of the building side wall driven by the horizontal component of rainfall can be calculated by the following formula: r h =r w tanθ Where Q h is the flow rate of the building side wall, unit is m 3 / s;r w and r h are the vertical and horizontal components of the inclined rainfall intensity, in mm / h; f wall is the infiltration rate of the vertical surface of the building side wall, in mm / h; b and l are the width and length of the building, in m; n is the number of buildings in the building area; is the angle between the wind speed direction and the normal of the long side of the building, and its value range is [0,π / 2]; θ is the rainfall inclination angle in degrees; h is the net height of the building in meters; It is the effective "rain-receiving width" of the building's side wall in the direction of wind speed.

4. The method for simulating and predicting the transport of runoff pollutants in urban areas according to claim 3, wherein: For the part of the surface on the leeward side of the building where there is no rainfall runoff due to the obstruction of adjacent buildings, the area of ​​this part of the area will be deducted when the subcatchment area parameter value of the SWMM stormwater model is selected.

5. The method for simulating and predicting the transport of runoff pollutants in urban areas according to claim 4, characterized in that: When calculating the sidewall runoff and blocked area, the coordinate system of the building and terrain data is aligned. The total height of the building is equal to the net height of the building itself plus the terrain value of the terrain grid closest to the geometric center of the building bottom surface. The area of ​​the blocked area between buildings is further calculated based on this. The calculation process includes the following steps: If the distance between two adjacent buildings is l d is greater than or equal to (h+z-z0)tanθ, then the area S of the occluded region is: Where z and z0 are the terrain heights of the building, in meters; At this time, the windward side of the right building, that is, its left wall, can be fully affected by rainfall, and the flow rate is Q h ; On the contrary, if the distance between two adjacent buildings is l d is less than (h+z-z0)tanθ, then the area S of the blocked area is: At this time, the windward side of the building on the right is partially blocked, and the rain-receiving area of ​​the side wall needs to be calculated. The rain-receiving height h′0 of its side wall is: Its rain-receiving area S′0 is: Then the flow rate generated by the side wall of the right building is: Q′ h =r h nS′0-f wall n(b+l)h′0。 6. The method for simulating and predicting the transport of runoff pollutants in urban areas according to claim 5, characterized in that: When calculating the area S of the obscured region, the average value of S calculated during the forecast period is used.

7. The method for simulating and predicting the transport of runoff pollutants in urban areas according to claim 3, wherein: In S1, for buildings with drainage ditches around their exterior walls entering the catchment network, the flow generated by the building side walls is calculated and then enters the associated node of the corresponding catchment area in the SWMM stormwater model in the form of node inflow; for buildings with green spaces or open spaces around their exterior walls, coupling is achieved by adding their flow sequence to the boundary condition file of the FVCOM model for calculation.

8. The method for simulating and predicting the transport of runoff pollutants in urban areas according to claim 3, wherein: In S2, the overflow volume and pollutant concentration at the overflow node obtained by SWMM stormwater model simulation are used as boundary conditions of the two-dimensional hydrodynamic model. The pollutant concentration at the overflow node is converted into a certain number of particles by the following formula: Where, is the overflow node i at time j n The number of particles at , which is used as input data for the Lagrangian particle model; is the overflow node i at time j n Pollutant concentration at the site (mg / L), calculated by the SWMM stormwater model; m p is the mass of a single particle (kg); is the overflow node i at time j n The volume of water accumulated on the corresponding grid unit (m 3 ); During the transport of pollutants on the surface along with the evolution of accumulated water, their concentration can be converted from the number of particles to concentration using the following formula: Where, is the grid cell i at time j c Pollutant concentration at the site (mg / L); is the grid cell i at time j c The number of particles at is the grid cell i at time j c Volume of water on the 3 ).

Images