Method for constructing twin management and control system of urban drainage unit based on Al

Abstract

The invention provides a method for constructing an urban drainage unit twin system based on an AI data analysis technology. The drainage unit twin system is composed of an online monitoring device, a network communication system and a remote monitoring platform. The NBIOT or OneNET technology is adopted to carry out networking on the online monitoring device, so that online accurate data acquisition and real-time transmission of a drainage unit pipe network catchment point, a drainage port and a rainfall station intelligent terminal carrying lightweight model are realized; and the remote monitoring platform is combined with an unreal engine technology to truly reflect a 3D real scene of the drainage unit. The intelligent monitoring system has a plurality of data monitoring functions such as rainfall, underground liquid level, flow velocity and flow, water quality, pavement ponding and well lid abnormity alarm monitoring, and provides a comprehensive data source for a drainage unit twin system. The AI technology is used for comprehensively mastering the condition of a drainage system, and visual monitoring of pavement ponding and waterlogging in the urban drainage unit and intelligent monitoring of a drainage pipe network are achieved.

Description

Technical Field

[0001] The present invention relates to the field of urban flood control and drainage information technology, and in particular to a method for constructing an urban drainage unit twin control system based on AI. Background Art

[0002] In the existing technology, urban drainage units, as small units of the drainage network system, play a key role in urban infrastructure. In the drainage system, the main function of the drainage unit is to achieve rainwater and sewage separation, improve the sewage collection and treatment rate, and enhance the city's flood control and drainage capabilities. At the same time, it is also convenient for the management and maintenance of the drainage system, which is of great significance to protecting the urban environment and preventing flood disasters.

[0003] Although some intelligent monitoring terminals and information management systems have been introduced in recent years for the digital management of drainage networks divided into drainage units, there are still certain limitations.

[0004] In the urban drainage system, the drainage network of 1 square kilometer is usually distributed in three layers, from the source drainage to the main pipe drainage outlet, which is the branch pipe layer, the main pipe layer and the main pipe layer in turn. The drainage network is designed according to the buried depth and design flow rate of the gravity flow drainage network of different pipe diameters. The number of rainwater manhole covers, sewage manhole covers, rainwater inlets, water collection wells, drainage outlets and pipes in the urban drainage unit will vary significantly due to factors such as the area, topography, and function of the drainage unit. According to the "Outdoor Drainage Design Standard", the distance between rainwater inlets should be 25-50 meters, and generally arranged at about 30 meters. Based on this spacing, the number of rainwater inlets in a drainage unit of 1 square kilometer is about 300-400, and the number of water collection wells is about 50-100. The number of drainage outlets depends on the size of the drainage unit, drainage needs, and connection with external water bodies. Small drainage units may have only 1-2 drainage outlets, large and complex drainage units may have 3-5 or even more, and drainage units of 1 square kilometer usually have 2-4 drainage outlets.

[0005] Smart monitoring products for drainage outlets such as drainage networks, rivers, and lakes on the market often focus on water level monitoring and rainfall detection data, combined with drainage unit topography (GIS) and other factors to develop a certain type of drainage management system, and its structure and function are relatively simple. At present, drainage twin system products with complete solutions and built using AI technology are still very scarce.

[0006] In addition, the online monitoring devices of the existing drainage network intelligent monitoring system work in outdoor environments for a long time and need to rely on numerous transmission nodes to connect a large number of monitoring devices. There are problems such as limited monitoring point distribution, limited monitoring capabilities, insufficient pre-control capabilities, and high networking costs. The data monitoring effect is unsatisfactory. This is not only not conducive to the timely handling of waterlogging hazards and fault tracking by the drainage management system, but also greatly affects the daily maintenance of the drainage network system. Summary of the Invention

[0007] The object of the present invention is to solve the technical problems in the prior art that the monitoring data types of urban drainage pipe network systems are single, the accuracy is unstable, the communication network is congested, and the overall monitoring ability is limited. A single relational model cannot meet the factors such as the terrain CIS of drainage units, pipe network status, environment, etc. in the traditional drainage system model, and there are problems such as incomplete results and low accuracy due to the conditional environment and instability of monitoring data. By using new AI technology, it is possible to comprehensively grasp the dynamic conditions of drainage outlets and pipe networks in the drainage system, such as the causes of poor drainage and blockage. Intuitively monitor the road surface waterlogging and waterlogging in the urban drainage unit, intelligent monitoring of the drainage pipe network, and a drainage twin system and method constructed by 3D technology to improve the data monitoring range and monitoring efficiency, reduce the system layout cost and data desensitization, improve data security and the management and maintenance level of the drainage pipe network, and further enhance the urban emergency flood control ability.

[0008] The present invention discloses an urban drainage unit twin control system based on Al, which is characterized by including:

[0009] An on-line monitoring device, which is based on a hybrid architecture of edge computing and cloud computing and is configured to perform on-line monitoring, and includes a sensing module;

[0010] A network communication system, which is configured for network communication transmission, and connects the on-line monitoring device and a remote monitoring platform;

[0011] A remote monitoring platform, which is configured to dynamically present the monitoring data of the drainage pipe network by using AI technology and perform trend analysis;

[0012] Among them, the sensing module includes:

[0013] A liquid level acquisition device, which is configured to acquire the liquid level height in the well, and is arranged in the well at the pipe intersection, pipe drainage outlet and / or pipe diameter change point in the drainage pipe network within the drainage unit, where the drainage unit is the pipe network that flows to the same drainage outlet in the urban drainage unit drainage pipe network system;

[0014] A water accumulation acquisition device, which is configured to measure the road surface water volume, and is arranged in the historical waterlogging area of the road within the drainage unit;

[0015] A flow velocity and flow rate acquisition device, which is configured to measure the water level flow velocity and flow rate of the pipe network, and is arranged at the drainage outlet or the bottom water collection point of the main pipe network;

[0016] A water quality acquisition device, which is configured to measure the water quality of the pipe network drainage, and is arranged at the drainage outlet;

[0017] A rainfall acquisition device, which is configured to measure the real-time rainfall, and is arranged at the central position of the area within the drainage unit; and

[0018] A manhole cover alarm device is used to detect abnormal state data of the manhole cover, including abnormal opening, tilting, damage, and whether it is lost, and is arranged on the manhole cover.

[0019] In an embodiment of the present invention, the network communication system adopts the narrowband Internet of Things protocol, and through the narrowband Internet of Things base station and the network communication module of the online monitoring device, the collected monitoring data in the drainage unit is transmitted to the remote monitoring platform in real time.

[0020] In another embodiment of the present invention, the remote monitoring platform includes:

[0021] A data center is used to store various tools and key data such as the geographical information of the drainage unit, ground elevation, pipe well depth, pipe network topology map, and traditional drainage models;

[0022] A monitoring terminal adopts AI technology and is used for map positioning and query, real-time / historical data query, trend analysis, and gives an alarm prompt for abnormal state data;

[0023] Among them, the monitoring terminal includes a C / S client, a Web client, and a wireless terminal, and the wireless terminal includes a smart phone, a tablet, and a PDA.

[0024] In another embodiment of the present invention, the remote monitoring platform is based on a B / S + C / S hybrid architecture to support different types of clients.

[0025] The present invention also provides a method for constructing a twin control system for urban drainage units based on Al, which is characterized by including:

[0026] Deploy an online monitoring device based on a hybrid architecture of edge computing and cloud computing, and build and deploy a remote monitoring platform based on a B / S and C / S hybrid architecture;

[0027] Use the online monitoring device to dynamically collect various monitoring data in the drainage unit, including the underground liquid level height, flow velocity and flow rate, water quality, road surface water accumulation, real-time rainfall, and manhole cover abnormal state data in the drainage pipe network of the drainage unit;

[0028] Through the network communication module of the online monitoring device, the real-time monitoring data is sent to the corresponding narrowband Internet of Things base station through the communication protocol;

[0029] Transmit the real-time monitoring data or the edge computing lightweight model to the data center server of the remote monitoring platform through the network communication system;

[0030] Receive the real-time monitoring data and the data processed by the edge computing lightweight model through the data center server of the remote monitoring platform, and store them in the corresponding database; and

[0031] The corresponding monitoring data is read from the database according to the demand through the monitoring terminal, and is dynamically updated and presented in real time in the software interface.

[0032] In one embodiment of the present invention, the following steps are also included:

[0033] Collect basic data such as geographic information data, engineering design data and meteorological data of urban drainage systems, remove erroneous, duplicated and incomplete data, and standardize and normalize the data for subsequent analysis and modeling;

[0034] Based on the basic principles of hydraulics, a water flow calculation model of the drainage unit is established to simulate the flow state of water in the pipe network, and the model parameters are calibrated through measured data to make the model simulation results consistent with the actual situation, and then verified with another set of measured data;

[0035] Fuse data of different modes, apply multimodal fusion technology in deep learning, conduct joint analysis and modeling of multi-source data, and explore the complex relationships between data;

[0036] Integrate quantitative models and multimodal models to achieve data interaction and feedback, so that they complement and work together to fully describe the hydraulic characteristics and operating status of the drainage system; and

[0037] Evaluation indicators are used to evaluate the performance of the integrated model, analyze the advantages and disadvantages of the model, and make optimization adjustments.

[0038] In another embodiment of the present invention, the step of dynamically updating and presenting in real time in the software interface includes:

[0039] Collect drainage system CAD drawings, BIM models, point cloud data, etc., use 3DMAX, Blender and other tools to create or process models, and use lightweight tools to optimize models to reduce the number of faces;

[0040] Import the lightweight model into the corresponding engine, and perform coordinate positioning and scaling adaptation in the engine according to the actual size and position of the drainage system;

[0041] Set up realistic materials for the drainage system model in the engine;

[0042] Add light sources in Unreal Engine and Unity to simulate real light and shadow reflection effects;

[0043] Implement user interaction with the drainage system through scripting in Unreal Engine and Unity, and combine real-time data to achieve dynamic effects of the drainage system; and

[0044] Integrate the models and functions of each part in the main scene, and optimize the performance in the engine.

[0045] In another embodiment of the present invention, the process of building and deploying an online monitoring device for the edge computing and cloud computing hybrid architecture includes:

[0046] Determine the functions and performance requirements of the drainage digital twin system, determine the types of data that need to be processed in real time and analyzed remotely, and plan the edge node and cloud resources according to the processing requirements and budget.

[0047] Place edge computing devices at the drainage facility site or at a certain distance from it, install the operating system, edge computing software, and data caching tools, and configure the network connection to enable communication with sensors and on-site devices.

[0048] Select a public cloud, private cloud, or hybrid cloud, create virtual machines, storage, and databases in the cloud, deploy software and services such as big data processing, data analysis, and digital twin modeling, and set security access control and data backup policies.

[0049] Formulate data transmission protocols and rules, develop data processing and analysis programs, enable the edge node to preprocess the collected data and upload key data to the cloud, and the cloud issues analysis results and control instructions to the edge node.

[0050] Conduct functional and performance tests on the overall architecture, check the accuracy of data transmission, etc., and optimize and adjust the resource configuration and data processing process according to the test results.

[0051] In another embodiment of the present invention, the process of building and deploying a remote monitoring platform for the B / S and C / S hybrid architecture includes:

[0052] Determine the usage scenarios of the B / S part and the C / S part;

[0053] Develop a C / S client software for the operating system platform to achieve data collection and real-time monitoring, install and deploy the client software to relevant management terminal devices, and configure the communication parameters with the server.

[0054] Develop a B / S-side application using Web technology to view the status of the drainage system, deploy the Web application to the Web server, and ensure the connection with the database and backend services.

[0055] Design a database to store the data of the drainage system, select a suitable database management system, configure the database server, and set user permissions and data backup and recovery policies for access by the B / S and C / S sides; and

[0056] Integrate the B / S and C / S sides with the database and other backend services, conduct functional and compatibility tests, fix problems, and optimize performance.

[0057] In another embodiment of the present invention, in the step of fusing data of different modalities, the following formula is used to process multi-source data:

[0058] Prediction equation: and

[0059] Update equation:

[0060] where represents the predicted value of the system state at time k based on the information available at time k - 1;

[0061] represents the optimized estimated value of the system state at time k - 1;

[0062] F k , the state transition matrix, which describes how the system state transfers from time k - 1 to time k;

[0063] B k , the control matrix, which determines how external adjustments affect the system state;

[0064] U k , the control vector, which represents the external control input applied to the system at time k;

[0065] combines the measurement value Z at time k k to obtain the optimal estimated value of the system state at time k;

[0066] K k , the Kalman gain matrix, which weighs the relative weights of the predicted value and the measurement value Z k when updating the state estimate;

[0067] Z k , the measurement vector, which is the actual measurement data obtained from various sensors at time k;

[0068] H k , the observation matrix, which describes the relationship between the system state and the actual measurement value.

[0069] In another embodiment of the present invention, in the step of using an added light source to simulate the real light and shadow reflection effect, the following equation is used for rendering:

[0070] L O (P, W O ) = L e (P, W O ) + ∫ Ω f r (P, W i, , W O )L i (P, W i )(n · Wi )dW i

[0071] The meanings of the parameters are as follows:

[0072] L O (P,W O ):Indicates the exit luminance of the light at point p in the direction of,W O ;

[0073] L e (P,W O ):Represents the luminance of the light emitted by itself at point p in the direction of,W O ;

[0074] ∫ Ω : It is an integral symbol, indicating integration over all directions on the hemispherical surface Ω;

[0075] f r (P,W i, ,W O ):Is called the bidirectional reflectance distribution function (BRDF), which describes the proportion of the light incident from the direction W at point p and reflected in the direction W; i ; O ;

[0076] L i (P,W i ):Is the luminance of the light incident at point p in the direction of W; i ;

[0077] n·W i : Where n is the surface normal direction at point p, and W is the incident light direction. The dot product of the two represents the cosine of the angle between the incident light and the surface normal, which is used to calculate the effective projected area of the incident light on the surface. i ;

[0078] The present invention has the following beneficial effects:

[0079] (1) The present invention is equipped with a lightweight model, integrating the monitoring functions of various types of data such as the water level height, flow velocity, water quality, road surface water accumulation, manhole cover status, and real-time rainfall in the drainage unit. It can comprehensively, integrally, and accurately obtain the measured data in the urban drainage unit in real time, so as to accurately evaluate the drainage capacity of the drainage pipe network in each drainage unit. Through real-time monitoring, problems such as urban waterlogging areas, pipe network blockages, road surface water accumulation, and abnormal water quality exceeding the standard at the drainage outlets of rainwater pipes can be detected in a timely manner. When local or global cleaning or maintenance is required for these problems, it can provide detailed and comprehensive data basis, effectively improving the management level of the drainage unit. In addition, it can also provide complete data support for the drainage unit twin system equipped with a quantization model and a multi-modal model, helping with data simulation and prediction.

[0080] (2) The present invention adopts an NB-IoT module, transmits data through the OneNet platform, and is paired with a sensing module with data acquisition and preliminary processing capabilities to replace the traditional wireless communication module, achieving efficient data transmission. The device has strong wall penetration ability and extremely low power consumption. In addition, it supports multiple communication methods such as 4G, 5G, LoRa, etc., which can be flexibly selected according to the actual application scenario to meet different requirements for distance, power consumption, and data transmission rate. This can not only significantly extend the service life of online monitoring devices working outdoors for a long time, but also overcome the disadvantages of short transmission distance and the need to set up too many base station nodes of traditional wireless communication modules. Within the drainage unit range, only a few base stations for communication need to be set up to connect all online intelligent terminal monitoring devices equipped with lightweight models to the remote monitoring center, effectively reducing the networking cost of the monitoring devices.

[0081] (3) The present invention combines lightweight models with virtual engine technology, AMRT3D, Cesium, and Unity, and they cooperate with each other in the drainage unit twin system to render 3D simulation images of real scenarios such as waterlogging, road surface water accumulation, pipe network blockage, and abnormal water quality exceeding the standard at the drainage outlet of the rainwater pipe. This helps to intuitively and accurately predict the causes of problems, quickly formulate decision-making treatment plans, and reasonably allocate resources such as people, vehicles, and materials. It is possible to clean the road surface rainwater inlets in advance, dredge blocked pipes, trace the pollution sources of abnormal water quality exceeding the standard, and command the rescue personnel to quickly carry out rescue and drainage work. At the same time, it provides improvement plans for facility maintenance and quality improvement and transformation, improving the overall response efficiency. Description of the Drawings

[0082] Figure 1 Shows the schematic diagram of the urban drainage unit twin control system based on Al in an embodiment of the present invention;

[0083] Figure 2 Shows the method flow chart of the urban drainage unit twin control system based on Al in an embodiment of the present invention;

[0084] Figure 3 Shows the flow chart of building a drainage unit twin system based on a quantization model and a multi-modal model in an embodiment of the present invention;

[0085] Figure 4 Shows the flow chart of rendering 3D simulation real scenarios in an embodiment of the present invention; and

[0086] Figure 5 Shows the deployment flow chart based on a hybrid architecture in an embodiment of the present invention. Detailed Embodiments

[0087] To describe in detail the technical content, structural features, achieved objectives and effects of the technical solution, the following will be described in detail in combination with specific embodiments and with reference to the accompanying drawings.

[0088] Figure 1 It shows the schematic diagram of the twin control system for urban drainage units constructed based on Al in an embodiment of the present invention.

[0089] As Figure 1 shown, the twin control system for drainage units in the present invention mainly consists of an on-line monitoring device 100, a network communication system 200 and a remote monitoring platform 300.

[0090] The on-line monitoring device 100 includes a control module 120, a network communication module 130 and a sensing module 110. Based on a hybrid architecture of edge computing and cloud computing, it has the monitoring function of various types of data, and realizes rainfall monitoring, underground well liquid level monitoring, flow velocity and flow rate monitoring, water quality monitoring, road surface water accumulation monitoring and manhole cover abnormal alarm monitoring. This system can provide comprehensive and comprehensive data support for the twin system of urban drainage units.

[0091] In an embodiment of the present invention, the network communication module 130 uses narrowband Internet of Things to interact with the network communication system 200 for data.

[0092] In order to perform accurate data monitoring, in an embodiment of the present invention, the sensing module 110 includes: a liquid level acquisition device 111, a water accumulation acquisition device 112, a rainfall acquisition device 113, a flow velocity and flow rate device 114, a water quality device 115 and a manhole cover alarm device 116. Among them, the liquid level acquisition device 111 is used to collect the underground well liquid level height and is arranged underground at the pipe intersection, pipe drainage outlet and pipe diameter change point in the drainage pipe network within the drainage unit; the water accumulation acquisition device 112 is used to measure the road surface water accumulation and is arranged in the historical waterlogging area of the road within the drainage unit; the flow velocity and flow rate acquisition device 113 is used to measure the water level flow velocity and flow rate of the pipe network and is arranged at the drainage outlet or the bottom water collection point of the main pipe network; the water quality acquisition device 114 is used to measure the water quality of the pipe network drainage and is arranged at the drainage outlet; the rainfall acquisition device 115 is used to measure the real-time rainfall and is arranged at the central position of the area within the drainage unit; the manhole cover alarm device 116 is used to detect the abnormal state data of the manhole cover, including abnormal opening, tilting, damage and whether it is lost, and is arranged on the manhole cover.

[0093] The network communication system 200 uses NBIOT or OneNET technology to network the on-line monitoring devices, and can realize accurate on-line acquisition and real-time transmission of data by intelligent terminals carrying lightweight models at the water collection points, drainage outlets and rain gauges of the drainage unit pipe network. The real-time monitoring data is collected through the sensing module 110 of the on-line monitoring device 100 and transmitted to the network communication system 200 via the network communication module 130.

[0094] The remote monitoring platform 300 consists of a data center 310 and a monitoring terminal 320 equipped with a lightweight model. It includes various tools and key data such as the geographical information (GIS) of the drainage unit, ground elevation, well depth, pipe network topology map, and traditional drainage models. By means of a quantization model and a multi-modal model, a drainage unit twin system is built and deployed based on a hybrid B / S + C / S architecture. These two architectures complement each other and can support different types of clients. At the same time, the remote monitoring platform is combined with Unreal Engine technology, Unity, Cesium, and AMRT3D engine technology, and can truly and intuitively present a 3D scene based on elements such as the drainage unit GIS, ground elevation, well depth, and pipe network topology map.

[0095] Figure 2 The flowchart of the method for building a twin control system for urban drainage units based on Al in an embodiment of the present invention is shown.

[0096] Figure 3 The flowchart of building a drainage unit twin system based on a quantization model and a multi-modal model in an embodiment of the present invention is shown.

[0097] Combined Figure 2 、 Figure 3 It can be seen that the present invention builds a drainage unit twin system based on a quantization model and a multi-modal model, so as to predict the future state changes of the system and perform real-time comparison monitoring on the system according to the intelligent monitoring of the drainage unit pipe network.

[0098] In an embodiment of the present invention, the method for building a twin control system for urban drainage units based on Al is as follows:

[0099] Device deployment 11, deploy the online monitoring device 100 based on a hybrid architecture of edge computing and cloud computing, and build and deploy the remote monitoring platform 300 based on a hybrid B / S and C / S architecture.

[0100] Monitoring data collection 12, use the online monitoring device 100 equipped with a lightweight model to dynamically collect various monitoring data of the drainage in the drainage unit. These data include the underground liquid level height, flow velocity and flow rate, water quality, road surface water accumulation, real-time rainfall, and manhole cover abnormal state data of the drainage pipe network in the drainage unit.

[0101] Data sending 13, with the help of the network communication module 130 of the online monitoring device 100, send the real-time monitoring data to the corresponding NBIOT base station according to the communication protocol.

[0102] Data transmission 14, through the corresponding NBIOT module, OneNET platform or 5G base station, transmit the real-time monitoring data or the edge computing lightweight model to the data center 310 of the remote monitoring platform 300.

[0103] Data storage 15. The data center 310 of the remote monitoring platform 300 receives the real-time monitoring data and the data processed by the edge computing lightweight model, and stores them in the corresponding database.

[0104] Data presentation 16. The monitoring terminal 320 reads the corresponding monitoring data from the database according to its own needs and presents it in real-time dynamic update on the software interface. In the embodiment of the present invention, the data presentation methods are rich and diverse, covering map positioning and query, real-time data display, historical data query, trend chart analysis, alarm prompt, and can also display various data, analysis results, and three-dimensional visualization scenarios of the drainage twin system in browsers on the large screen end, desktop end, and mobile phone end.

[0105] The process of building the drainage unit twin system based on the quantization model and multi-modal model is as follows:

[0106] S1: Data collection and preprocessing 101: Collect basic data such as geographical information data, engineering design data, and meteorological data of the urban drainage system, remove incorrect, duplicate, and incomplete data, and perform standardization and normalization processing on the data for subsequent analysis and modeling.

[0107] S2: Quantization model construction 102: Based on the basic principles of hydraulics, such as the continuity equation and the energy equation, establish a calculation model for the water flow movement of the drainage unit to simulate the flow state of the water flow in the pipe network, calibrate the model parameters with the measured data to make the model simulation results consistent with the actual situation, and then verify with another set of measured data to ensure the accuracy and reliability of the model.

[0108] When building the drainage unit twin system, the quantization model needs to accurately apply relevant formulas to ensure the scientificity and reliability of the model.

[0109] In an embodiment of the present invention, a calculation model for the water flow movement of the drainage unit is established through the Manning formula.

[0110] Manning formula:

[0111] Among them, V is the flow rate, representing the volume of fluid passing through a certain cross-section of flow per unit time.

[0112] n is the Manning roughness coefficient, which is related to factors such as the roughness of the channel wall surface. For example, the n value of a smooth concrete channel is smaller, and the n value of a rough rock channel is larger. For example, for a plastic pipe with a smooth inner wall, its n value is usually between 0.009 and 0.011; while for a masonry channel with a relatively rough surface, the n value is approximately in the range of 0.020 to 0.025. In practical applications, accurately selecting the n value is crucial for the accuracy of flow rate calculation.

[0113] R is the hydraulic radius, R: Where X is the wetted perimeter, i.e., the perimeter length of the water-passing section in contact with the solid boundary. A is the water-passing cross-sectional area, which refers to the cross-sectional area perpendicular to the water flow direction. In one embodiment of the present invention, when a circular pipe is full of flow, for a circular pipe, its water-passing cross-sectional area is a circle. According to the area formula of a circle, A=πr 2 , where r is the radius, and the pipe diameter d = 2r, that is Substituting this into the area formula we get In a circular pipe, the wetted circumference is the circumference of the circle. According to the circumference formula of the circle, C = 2πr = πd, where X represents the wetted circumference, so X = πd. Hydraulic radius formula: Bundle Substituting X=πd into the equation, we get

[0114] S is the hydraulic gradient, which is generally equal to the water surface slope and approximately equal to the channel bottom slope. It is dimensionless and represents the head loss per unit length of flow. In one embodiment of the present invention, for an open channel, under the ideal condition that local head loss is ignored and the water flow is uniform and stable, the hydraulic gradient is approximately equal to the channel bottom slope. For example, for a straight trapezoidal open channel, if its bottom slope is i, then the hydraulic gradient S can be approximately taken as i. For pipelines, the hydraulic gradient is mainly related to the head loss along the way. The greater the head loss along the way, the greater the hydraulic gradient. Specifically, it can be calculated and analyzed by the Darcy-Weisbach formula, etc.

[0115] The continuity equation is Among them, A generally represents the cross-sectional area of ​​the water flow, Q is usually the flow rate, q represents the source and sink term, which indicates the increase or decrease of the flow rate per unit length (such as the inflow or outflow of tributaries, etc.), t is the time, and x is the spatial coordinate along the pipe section. This formula describes that in the pipe section, the sum of the rate of change of the cross-sectional area of ​​the water flow with time and the rate of change of the flow rate along the length of the pipe section is equal to the source and sink term, which is used to strictly describe the law of conservation of water mass in the pipe section.

[0116] In one embodiment of the present invention, when the fluid is an incompressible fluid and there is no source (such as an additional water injection point) or sink (such as an area where the water flow disappears for no reason) in the pipe section, the formula is: If the side inflow q is considered, then The conversion between these two formulas mainly depends on whether the side inflow q is considered. When the side inflow is not considered, that is, q = 0 It is simplified to When the side inflow q needs to be considered, Adding the term q to the basis of

[0117] This equation indicates the rate of change of the fluid volume within the pipe segment per unit time (which is a partial derivative expression related to the change of the cross-sectional area of the water flow with time), and is equal to the difference in the flow rates flowing into and out of the pipe segment (the partial derivative of the flow rate Q with respect to the spatial coordinates), plus the lateral inflow q per unit length of the pipe segment

[0118] Among them, the sources of the lateral inflow q are diverse and their properties are also different. It is mainly divided into rainwater inflow and groundwater infiltration. The inflow of rainwater usually has obvious intermittency and randomness. During the rainfall process, surface runoff will flow into the drainage pipeline through facilities such as rainwater inlets and inspection wells, and the magnitude of its inflow is closely related to factors such as rainfall amount, rainfall intensity, and catchment area. The infiltration of groundwater is relatively stable and mainly depends on the difference between the groundwater level and the pipeline internal water level as well as the sealing performance of the pipeline. When there are cracks, loose joints, etc. in the pipeline, if the groundwater level is higher than the internal water level of the pipe, groundwater will infiltrate into the pipe under the action of the pressure difference

[0119] The energy equation is used to deeply analyze the energy relationship of the water flow and is a form of expression of the Bernoulli equation

[0120] The meanings of the parameters are as follows

[0121] Total head H: It is the sum of the elevation head z, the pressure head and the velocity head , and its physical meaning is the total energy possessed by a unit weight of fluid

[0122] Elevation head z: It represents the height of the water flow relative to a certain reference plane, reflecting the gravitational potential energy possessed by a unit weight of fluid

[0123] Pressure head where p is the pressure, ρ is the density of water, and g is the acceleration due to gravity. The pressure head reflects the energy possessed by a unit weight of fluid due to the action of pressure

[0124] Velocity head where v is the flow velocity, and the velocity head represents the kinetic energy possessed by a unit weight of fluid due to the flow velocity

[0125] Based on the law of conservation of energy, in the case of ideal energy loss-free, the total head of the water flow along the flow path is conserved. However, in actual engineering, due to factors such as friction between the water flow and the pipe wall or channel wall and turbulence within the water flow, energy loss will inevitably occur. When applying in practice, it is usually necessary to introduce energy loss coefficients (such as the friction head loss coefficient, local head loss coefficient, etc.) to correct the energy equation to more accurately reflect the actual energy change of the water flow

[0126] S3: Multi-modal Model Construction 103: Integrate data of different modalities such as geographical information data, sensor monitoring data, video image data, etc., and use multi-modal fusion technology in deep learning to jointly analyze and model multi-source data, and explore the complex relationships between data.

[0127] S4: Model Integration and Coupling 104: Integrate the quantization model and the multi-modal model to achieve data interaction and feedback, so that they complement and cooperate with each other to comprehensively describe the hydraulic characteristics and operating status of the drainage system.

[0128] S5: Model Evaluation and Optimization 105: Use a variety of evaluation indicators to evaluate the performance of the integrated model, analyze the advantages and disadvantages of the model, and perform optimization and adjustment.

[0129] When using a multi-modal model to construct a drainage unit twin system, specific functions can be achieved through different formulas.

[0130] In an embodiment of the present invention, when the multi-modal model processes image data (such as satellite images of drainage network distribution, on-site monitoring images of drainage facilities, etc.), key features such as pipeline contours and facility locations are extracted from the original images through a convolution formula, providing a basis for subsequent analysis and model construction.

[0131] The two-dimensional convolution formula is (I*K)(X,Y) = ∑ m ∑ n I(x-m,y-n)K(m,n)

[0132] Where, in this embodiment, I represents the input image (which can be regarded as a two-dimensional matrix), and I(x,y) represents the pixel value of the input image at coordinates (x,y). K is the convolution kernel (also a two-dimensional matrix), and K(m,n) is the value of the convolution kernel at coordinates (m,n). (I*K)(x,y) represents the result value at coordinates (x,y) after the convolution operation. The two summation symbols ∑ m Σ n Indicate traversing all elements of the convolution kernel. For each element K(m,n) in the convolution kernel, multiply it by the pixel value at the corresponding position I(x-m,y-n) in the input image, and then sum all the products to obtain the value of the convolution result at (x,y).

[0133] In an embodiment of the present invention, in the drainage unit twin system, it is necessary to integrate data from multiple sensors (such as flow sensors, pressure sensors, water quality sensors, etc.) to comprehensively and accurately reflect the operating status of the drainage system. The Kalman filter formula can not only accurately estimate the current state through effective processing of multi-source data, but also predict the future state changes of the system, providing strong support for the real-time monitoring and control of the system.

[0134] Prediction equation:

[0135] where This is a vector representing the predicted value of the system state at time k based on the information available at time k - 1. In the application context of a drainage system, it may comprehensively include the prediction of relevant state information such as the water flow rate, pressure, and water quality parameters in the pipeline. For example, if a certain section of the drainage pipeline is of concern, a certain element in it may be the predicted value of the water flow rate in this pipeline at time k.

[0136] is also a vector, representing the optimized estimated value of the system state at time k - 1. This value is the most accurate state estimate based on all available information at time k - 1 and before.

[0137] F k : Known as the state transition matrix, it describes how the system state transfers from time k - 1 to time k. In a drainage system, this matrix is determined according to the physical characteristics and dynamic laws of the system, for example, reflecting the relationship between the flow rate and pressure changes of water flow caused by factors such as pipeline characteristics and slope at different time periods.

[0138] B k : Is the control matrix, which relates the control input U k to the system state. In a drainage system, U k may represent some external controllable factors, such as the adjustment of the pumping rate of a pumping station, etc., and U k determines how these adjustments affect the system state.

[0139] U k : Control vector, representing the external control input applied to the system at time k.

[0140] Update equation:

[0141] where the parameter meanings are as follows:

[0142] In vector form, it is the optimal estimated value of the system state at time k after combining the measurement value Z k . Compared with , it incorporates the latest measurement information and is thus more accurate.

[0143] K k : Kalman gain matrix, which weighs the relative weights of the predicted value and the measurement value Z k when updating the state estimate. The calculation of the Kalman gain matrix comprehensively considers the noise characteristics of the system as well as the uncertainties of prediction and measurement.

[0144] Z k : The measurement vector, which is the actual measurement data obtained from various sensors (such as flow sensors, pressure sensors, water quality sensors, etc.) at time k. These data reflect part of the state information of the system at time k, but due to measurement errors in the sensors, they need to be combined with the predicted values for optimization.

[0145] H k : The observation matrix, which maps the state variables of the system to the measurement space, that is, describes the relationship between the system state and the actual measurement values. In the drainage system, H k will be determined according to factors such as the measurement principle and installation location of the sensors. For example, it maps the actual state parameters of the water flow in the pipeline to the flow value corresponding to the electrical signal measured by the flow sensor.

[0146] Through the collaborative work of the above prediction equation and update equation, the Kalman filter can effectively process multi-source sensor data in the drainage unit twin system, and achieve accurate estimation and prediction of the system state.

[0147] Figure 4 shows the flowchart of rendering a 3D simulation real scene in an embodiment of the present invention.

[0148] Combined with Figure 4 the shown process, the steps and methods of rendering a 3D simulation real scene based on the cooperation of the lightweight model and the Unreal Engine technology in the drainage unit twin system are as follows:

[0149] Step1: Data collection and processing 161: Collect CAD drawings, BIM models, point cloud data, etc. of the drainage system, and use tools such as 3DMAX and Blender for model creation or processing. And use lightweight tools to optimize the model and reduce the number of model faces.

[0150] Step2: Import and adaptation 162: Import the lightweight model into the corresponding engine. In an embodiment of the present invention, in the Unreal Engine, models in formats such as FBX can be directly imported; in AMRT3D, import according to its supported formats; in Cesium, convert the model to formats such as 3DTiles and then load; in Unity, multiple common formats are supported for import. According to the actual size and position of the drainage system, perform coordinate positioning and scaling and other adaptation operations on the model in each engine.

[0151] Step3: Material and texture setting 163: Set realistic materials for the drainage system model in each engine. For example, set metal or concrete materials for the pipes, and set transparent flowing materials for the water bodies. Add texture maps, such as rust textures for the pipes and brick textures for the ground, to enhance the realism.

[0152] Step 4: Lighting and Environment Settings 164: In Unreal Engine and Unity, various light sources are added, such as directional lights to simulate sunlight and point lights to illuminate specific areas. In Cesium, real lighting effects are set according to geographical information and time. Ambient light, reflection probes, etc. are set to simulate real light and shadow reflection effects.

[0153] Step 5: Interaction and Dynamic Effect Settings 165: In Unreal Engine and Unity, user interaction with the drainage system is achieved through script writing, such as clicking to view device information. In AMRT3D, similar functions are achieved using its interaction interface. Combining real-time data, dynamic effects of the drainage system are implemented in each engine, such as water flow and valve opening / closing.

[0154] Step 6: Integration and Optimization 166: Integrate and layout each part of the model and functions in the main scene. Perform performance optimization in each engine, such as setting LOD (Level of Detail), occlusion culling, etc., to improve rendering efficiency.

[0155] Step 7: Lightweight Model Unreal Engine Technology.

[0156] In an embodiment of the present invention, the present invention reduces the number of model parameters, computational amount, etc. according to the pruning operation function, and improves the model operation efficiency while maintaining a certain accuracy, reduces the demand for hardware resources, and is convenient for deployment and application in resource-constrained devices or scenarios.

[0157] In model pruning, assume the original model parameters are W, and the parameters after pruning are W 1 , which can be expressed as W 1 = prune(W), where prune is the pruning operation function. The function is to screen and process the original model parameters W, removing those parameters that have less impact on the model performance, so as to obtain the refined parameters W 1 .

[0158] In an embodiment of the present invention, knowledge distillation is another important technology in lightweight model construction. In this process, the output of the teacher model is denoted as Yt, the output of the student model is Ys, and the true label is denoted as Ytrue.

[0159] At this time, the loss function L used to train the student model consists of two parts, and the specific formula is as follows: L = L CE (Ys, Ytrue)+λ·KL(Ys, Yt)

[0160] Where: L CEDenotes the cross - entropy loss, which measures the difference between the output Ys of the student model and the true label Ytrue. The cross - entropy loss can guide the student model to learn in the direction of correct classification, prompting its output to be as close as possible to the true label.

[0161] KL represents the KL divergence, which is used to measure the difference in probability distributions between the output Ys of the student model and the output Yt of the teacher model. By minimizing the KL divergence, the student model can learn the knowledge contained in the teacher model, especially the information hidden in the output probability distribution.

[0162] λ is a balancing coefficient, whose role is to adjust the relative weights of the cross - entropy loss L CE and the KL divergence KL in the total loss function L. An appropriate value of λ can enable the student model to learn the true label information while effectively acquiring knowledge from the teacher model, thus ensuring the performance of the model under the premise of light - weight.

[0163] In an embodiment of the present invention, when using Unreal Engine technology for model rendering, i.e., in step 164, the rendering equation:

[0164] L O (P,W O )=L e (P,W O )+∫ Ω f r (P,W i, ,W O )L i (P,W i )(n·W i )dW i

[0165] The meanings of the parameters are as follows:

[0166] L O (P,W O ):Represents the outgoing luminance at point p in the direction of,W O .

[0167] L e (P,W O ):Represents the self - emitted luminance at point p in the direction of,W O .

[0168] ∫ Ω : Is an integral symbol, indicating integration over all directions on the hemispherical surface Ω.

[0169] f r (P,W i, ,W O):Called the Bidirectional Reflectance Distribution Function (BRDF), it describes, at point p, the proportion of light incident from direction W i that is reflected in direction W O .

[0170] L i (P, W i ):Is the luminance of the light incident at point p in direction W i .

[0171] n·W i : Where n is the surface normal direction at point p, and W i is the incident light direction. The dot product of the two represents the cosine of the angle between the incident light and the surface normal, and is used to calculate the effective projected area of the incident light on the surface.

[0172] In an embodiment of the present invention, the frustum culling formula is used in the rendering optimization of AMRT3D. Assuming that the frustum is represented by the plane equation Ax + By + Cz + D = 0, and the model vertex coordinates are (X0, Y0, Z0). By judging the value of A X0 + B Y0 + C Z0 + D to determine whether the vertex is inside the frustum. When A X0 + B Y0 + C Z0 + D meets specific conditions (such as greater than or equal to 0 or less than or equal to 0, specifically depending on the definition of the frustum), the vertex is inside the frustum; otherwise, it is outside the frustum.

[0173] In an embodiment of the present invention, Cesium is used to create virtual globes and various 3D geospatial applications. With its powerful functions, it can conveniently load and intuitively visualize a variety of geospatial data.

[0174] In the actual process of geospatial analysis and visualization, coordinate system conversion is often required. Taking the conversion of longitude and latitude coordinates to Cartesian coordinates (x, y, z) as an example, this conversion process is realized based on specific mathematical formulas. The conversion formulas are as follows:

[0175]

[0176] In the above formula system: λ represents longitude, which is used to measure the angle of a point on the Earth's surface east or west of the prime meridian (0° longitude), and its range is usually -180° to 180°, which determines the position of the point in the east-west direction on the Earth.

[0177] Latitude, denoted as , is the angle between a point on the Earth's surface and the Earth's equatorial plane. Its range is from -90° to 90°. Positive values indicate north latitude, and negative values indicate south latitude. It is used to determine the position of a point in the north-south direction on the Earth.

[0178] R represents the radius of the Earth, which is an important parameter. Since the Earth is not a standard sphere, in practical applications, the value of the Earth radius R may vary according to specific Earth models (such as the WGS84 model, etc.). For example, in an embodiment of the present invention, using the WGS84 model, the average radius of the Earth is approximately 6,371,008.7714 meters. This parameter plays a crucial role in coordinate transformation, determining the scale relationship of each dimension when converting from latitude and longitude coordinates to Cartesian coordinates.

[0179] Through the synergistic effect of these parameters and formulas, the accurate conversion from latitude and longitude coordinates to Cartesian coordinates is achieved, laying a foundation for subsequent processing, analysis, and visualization of geographical data in three-dimensional space.

[0180] In an embodiment of the present invention, in the physical simulation section of the Unity engine, Newton's second law plays a crucial role. The expression of this law is: F = ma.

[0181] Where F represents the force acting on an object, with the unit of Newton (N); m represents the mass of the object, usually with the unit of kilogram; a is the acceleration of the object, with the unit of meters per second squared. In Unity, this formula is used to accurately calculate the force on an object and then determine its motion state, providing an important theoretical basis for simulating physical phenomena in the real world.

[0182] Figure 5 Shows the deployment flowchart of the hybrid architecture in an embodiment of the present invention.

[0183] The application of the hybrid architecture provided by the present invention in the field of drainage digital twins is a breakthrough and innovation for the traditional architecture, which can flexibly configure resources and functions according to the characteristics and requirements of the drainage system, improving the adaptability and scalability of the system.

[0184] Such as Figure 5 shown, an online monitoring device is built and deployed through the edge computing + cloud computing hybrid architecture, and a monitoring terminal is built and deployed through the B / S + C / S hybrid architecture.

[0185] The deployment process of the edge computing + cloud computing hybrid architecture includes:

[0186] Requirement analysis and planning 111-1: clarify the functions and performance requirements of the drainage digital twin system, determine the types of data that need to be processed in real time and analyzed remotely, and plan the edge nodes and cloud resources according to the processing requirements and budget.

[0187] Edge Node Deployment 111-2: Install an operating system, edge computing software, and data caching tools on edge computing devices such as small servers or industrial computers at or a certain distance from the drainage facility site. Configure the network connection to enable communication with sensors and on-site devices. The distance is generally 1-10m.

[0188] Cloud Platform Setup 111-3: Select a public cloud, private cloud, or hybrid cloud. Create resources such as virtual machines, storage, and databases in the cloud. Deploy software and services such as big data processing, data analysis, and digital twin modeling. Set up security access controls and data backup policies.

[0189] Data Interaction and Collaboration 111-4: Develop data transmission protocols and rules, and data processing and analysis programs. Have the edge node preprocess the collected data, upload key data to the cloud, and have the cloud send analysis results and control instructions to the edge node.

[0190] System Testing and Optimization 111-5: Conduct functional and performance tests on the overall architecture, check data transmission accuracy, etc. Optimize and adjust resource configurations and data processing flows based on test results.

[0191] The deployment process of the B / S + C / S hybrid architecture includes:

[0192] Architecture Design 112-1: Determine the functional modules of the B / S and C / S parts. The B / S is used for general information display and simple operations, and the C / S is used for complex professional functions and scenarios with high-performance requirements.

[0193] C / S-side Development and Deployment 112-2: Develop C / S client software for platforms such as Windows to achieve functions such as data collection and real-time monitoring. Install and deploy the client software on relevant management terminal devices and configure communication parameters with the server.

[0194] B / S-side Development and Deployment 112-3: Develop B / S-side applications using web technologies to achieve functions such as viewing the status of the drainage system. Deploy the web application to a web server to ensure connection with the database and other backend services.

[0195] Database Design and Management 112-4: Design a database to store drainage system data, select a suitable database management system, configure the database server, and set user permissions and data backup and recovery policies for access by the B / S and C / S sides.

[0196] System Integration and Testing 112-5: Integrate the B / S and C / S sides with the database and other backend services, conduct functional, compatibility, etc. tests, and fix problems and optimize performance.

[0197] Although the above-described embodiments have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concepts. Therefore, the above are only the embodiments of the present invention and do not limit the patent protection scope of the present invention. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, is similarly included in the patent protection scope of the present invention.

Claims

1. A twin control system for urban drainage units based on AI, characterized in that: include: An online monitoring device, which is based on a hybrid architecture of edge computing and cloud computing and is configured to perform online monitoring, comprises a sensor module; A network communication system, configured for network communication transmission, connecting the online monitoring device and the remote monitoring platform; The remote monitoring platform is configured to use AI technology to dynamically present real-time monitoring data of the drainage network and conduct trend analysis; Wherein, the sensor module comprises: A liquid level acquisition device, configured to acquire the underground liquid level height, is arranged underground at a pipe intersection, a pipe outlet and / or a pipe diameter change in a drainage pipe network in a drainage unit, wherein the drainage unit is a pipe network flowing to the same outlet in a drainage pipe network system of an urban drainage unit; a water collection device configured to measure the amount of water on the road surface and arranged in a historical waterlogging area of ​​the road within the drainage unit; The flow rate and flow collection device is configured to measure the water level, flow rate and flow of the pipe network, and is arranged at the drainage outlet or the water confluence point at the bottom of the main network pipe network; A water quality collection device is configured to measure the water quality of the drainage from the pipe network and is arranged at the drainage outlet; A rainfall collection device, configured to measure real-time rainfall, arranged at a central location within the drainage unit; and The manhole cover alarm device is used to detect abnormal status data of the manhole cover, including abnormal opening, abnormal tilt, abnormal damage and whether it is lost, and is arranged on the manhole cover.

2. The urban drainage unit twin control system based on Al as claimed in claim 1 is characterized in that: The network communication system adopts the narrowband Internet of Things protocol, and transmits the collected monitoring data in the drainage unit to the remote monitoring platform in real time through the narrowband Internet of Things base station and the network communication module of the online monitoring device.

3. The urban drainage unit twin control system based on Al as claimed in claim 1, characterized in that: The remote monitoring platform includes: Data center, used to store various tools and key data such as drainage unit geographic information, ground elevation, pipe well depth, pipe network topology and traditional drainage models; The monitoring terminal uses AI technology for map positioning and query, real-time / historical data query, trend analysis, and alarm prompts for abnormal status data; The monitoring terminal includes a C / S client, a Web client and a wireless terminal, wherein the wireless terminal includes a smart phone, a tablet and a PDA.

4. The urban drainage unit twin control system based on Al as claimed in claim 3 is characterized in that: The remote monitoring platform is based on a B / S+C / S hybrid architecture to support different types of clients.

5. A method for constructing a twin control system of urban drainage units based on AI, characterized in that: include: Deploy online monitoring devices based on the hybrid architecture of edge computing and cloud computing, and build and deploy remote monitoring platforms based on the hybrid architecture of B / S and C / S; Use online monitoring devices to dynamically collect various monitoring data in the drainage unit, including the underground liquid level height, flow rate, water quality, road surface water volume, real-time rainfall and abnormal status data of the drainage network in the drainage unit; Through the network communication module of the online monitoring device, the real-time monitoring data is sent to the corresponding narrowband Internet of Things base station via the communication protocol; Transmit real-time monitoring data or edge computing lightweight models to the data center server of the remote monitoring platform through the network communication system; Receive real-time monitoring data and data processed by the edge computing lightweight model through the data center server of the remote monitoring platform, and store them in the corresponding database; and The corresponding monitoring data is read from the database according to the demand through the monitoring terminal, and is dynamically updated and presented in real time in the software interface.

6. The method for constructing a twin control system of urban drainage units based on Al as claimed in claim 5, characterized in that: The following steps are also included: Collect basic data such as geographic information data, engineering design data and meteorological data of urban drainage systems, remove erroneous, duplicated and incomplete data, and standardize and normalize the data for subsequent analysis and modeling; Based on the basic principles of hydraulics, a water flow calculation model of the drainage unit is established to simulate the flow state of water in the pipe network, and the model parameters are calibrated through measured data to make the model simulation results consistent with the actual situation, and then verified with another set of measured data; Fuse data of different modes, apply multimodal fusion technology in deep learning, conduct joint analysis and modeling of multi-source data, and explore the complex relationships between data; Integrate quantitative models and multimodal models to achieve data interaction and feedback, so that they complement and work together to fully describe the hydraulic characteristics and operating status of the drainage system; as well as Evaluation indicators are used to evaluate the performance of the integrated model, analyze the advantages and disadvantages of the model, and make optimization adjustments.

7. The method for constructing a twin control system of urban drainage units based on Al as claimed in claim 5, characterized in that: The step of dynamically updating and presenting in real time in the software interface comprises: Collect drainage system CAD drawings, BIM models, point cloud data, etc., use 3DMAX, Blender and other tools to create or process models, and use lightweight tools to optimize models to reduce the number of faces; Import the lightweight model into the corresponding engine, and perform coordinate positioning and scaling adaptation in the engine according to the actual size and position of the drainage system; Set up realistic materials for the drainage system model in the engine; Add light sources in Unreal Engine and Unity to simulate real light and shadow reflection effects; Implement user interaction with the drainage system through scripting in Unreal Engine and Unity, and combine real-time data to achieve dynamic effects of the drainage system; and Integrate the models and functions of each part in the main scene, and optimize the performance in the engine.

8. The method for constructing a twin control system of urban drainage units based on Al as claimed in claim 5, characterized in that: The process of building and deploying an online monitoring device using a hybrid architecture of edge computing and cloud computing includes: Determine the functional and performance requirements of the drainage twin system, determine the types of data that need to be processed in real time and analyzed remotely, and plan edge nodes and cloud resources based on processing requirements and budget; Place edge computing devices at or a distance from the drainage facility, install operating systems, edge computing software, and data caching tools, and configure network connections to communicate with sensors and field devices; Choose a public cloud, private cloud, or hybrid cloud, create virtual machines, storage, and databases in the cloud, deploy software and services such as big data processing, data analysis, and digital twin modeling, and set up security access control and data backup policies; Formulate data transmission protocols and rules, develop data processing and analysis programs, enable edge nodes to pre-process collected data and upload key data to the cloud, and the cloud will issue analysis results and control instructions to edge nodes; Conduct functional and performance tests on the overall architecture, check data transmission accuracy, etc., and optimize and adjust resource allocation and data processing procedures based on test results.

9. The method for constructing a twin control system of urban drainage units based on Al as claimed in claim 5, characterized in that: The process of building and deploying a remote monitoring platform with a hybrid B / S and C / S architecture includes: Determine the usage scenarios of the B / S and C / S parts; Develop C / S client software for the operating system platform to realize data collection and real-time monitoring, install and deploy the client software to relevant management terminal devices and configure communication parameters with the server; Use Web technology to develop B / S end applications to view the status of the drainage system, deploy the Web application to the Web server and ensure connection with the database and back-end services; Design a database to store drainage system data, select a suitable database management system, configure the database server, set user permissions and data backup and recovery strategies for B / S and C / S access; and Integrate B / S and C / S with database and other backend services, perform functional and compatibility testing, fix issues and optimize performance.

10. The method for constructing a twin control system of urban drainage units based on Al as claimed in claim 6, characterized in that: In the step of fusing data of different modalities, the following formula is used to process multi-source data: Prediction equation: as well as Update equation: in It represents the predicted value of the system state at time k based on the information at time k-1; Represents the estimated value of the system state after optimization at time k-1; F k , the state transfer matrix, which describes how the system state is transferred from time k-1 to time k; B k , the control matrix, determines how external adjustments affect the state of the system; U k , the control vector, represents the external control input applied to the system at time k; Combined with the measurement value Z at time k k After that, the optimal estimate of the system state at time k; K k , the Kalman gain matrix, which weighs the predicted value and the measured value Z k Relative weights when updating state estimates; Z k , measurement vector, which is the actual measurement data obtained from various sensors at time k; H k ,The observation matrix describes the relationship between the system state and the actual measured values.

11. The method for constructing a twin control system of urban drainage units based on Al as claimed in claim 7, characterized in that: In the step of adding light sources to simulate the real light and shadow reflection effect, the following equation is used for rendering: L O (P,W O )=L e (P,W O )+∫ Ω f r (P,W i, ,W O )L i (P,W i )(n W i )dW i The meaning of each parameter is as follows: L O (P,W O ): indicates that at point p, along direction, W O The brightness of the emitted light; L e (P,W O ): represents the direction at point p, W O The brightness of the light emitted by itself; ∫ Ω : is an integral symbol, indicating integration in all directions on the hemisphere Ω; f r (P,W i, ,W O ): called the bidirectional reflectance distribution function (BRDF), which describes the reflection at point p from the direction W i The incident light, along the direction W O The proportion of reflection; L i (P,W i ): is at point p, along direction W i The brightness of the incident light; n.W i : where n is the surface normal direction at point p, and W i is the direction of the incident light, and the dot product of the two represents the cosine of the angle between the incident light and the surface normal, which is used to calculate the effective projection area of ​​the incident light on the surface.

Images