A new type of sponge city rainwater reuse system

Abstract

The present invention provides a new type of sponge city rainwater reuse system, which belongs to the technical field of urban rainwater management and resource utilization. The system collects data such as rainfall intensity, soil moisture, water quality and groundwater level in real time through the intelligent integrated monitoring layer, and uploads it to the central intelligent management system; the dynamic infiltration adjustment layer uses nano-level intelligent infiltration materials and adaptive infiltration algorithms to adjust the infiltration rate according to the monitoring data, optimize surface runoff and groundwater recharge; the intelligent water collection and pretreatment layer combines the deep learning water level prediction algorithm to achieve efficient water collection and water quality pretreatment; the water purification and resource layer adopts deep processing technology to ensure that the effluent water quality meets the urban water use standards; the intelligent water resource allocation and reuse layer dynamically optimizes the water resource allocation strategy through integrated learning and demand prediction algorithms to meet the needs of different water use departments. The system realizes the efficient collection, purification and resource utilization of rainwater, and improves the intelligent level of urban water resource management.

Description

Technical Field

[0001] The present invention relates to the technical field of urban rainwater management and resource utilization, and in particular to a novel sponge city rainwater reuse system. Background Art

[0002] In recent years, the concept of sponge city has been proposed and gradually applied to urban rainwater management. It aims to achieve the natural accumulation, infiltration, purification and utilization of rainwater by simulating the natural water cycle process, thereby improving the city's flood control and drainage capabilities and improving the urban ecological environment.

[0003] However, the existing sponge city rainwater treatment system still has some shortcomings. First, in terms of rainwater collection, traditional systems often lack intelligent monitoring and adjustment means, resulting in low rainwater collection efficiency, and unable to dynamically adjust the collection strategy according to real-time data such as rainfall intensity and soil moisture. Secondly, in terms of rainwater purification, existing technologies mostly use a single purification process, which is difficult to meet the purification needs under different water quality conditions, and the purification effect is limited. In addition, in terms of water resource allocation and reuse, traditional systems lack intelligent management and optimization means, and cannot achieve accurate allocation and efficient utilization of water resources, resulting in serious waste of water resources. Summary of the invention

[0004] The purpose of the present invention is to provide a novel sponge city rainwater reuse system, which realizes the efficient collection, purification and resource utilization of rainwater and improves the intelligent level of urban water resource management.

[0005] To achieve the above object, the present invention provides a novel sponge city rainwater reuse system, comprising:

[0006] The intelligent integrated monitoring layer integrates rain gauges, soil moisture sensors, water quality monitoring probes and ultrasonic water level meters to form a monitoring system covering key areas of the city, collecting and uploading data in real time to the central intelligent management system;

[0007] The dynamic permeability adjustment layer uses nano-scale intelligent permeability materials, combined with electromagnetic and pressure sensing control technology, and uses adaptive permeability algorithms to predict and adjust the permeation rate based on the data uploaded by the intelligent integrated monitoring layer received by the central intelligent management system;

[0008] Intelligent water collection and preprocessing layer, using deep learning water level prediction algorithm to predict water level changes in water collection wells;

[0009] The water purification and resource utilization layer is used to deeply process the water provided by the intelligent water collection and pretreatment layer, and the output water quality meets the urban water use standards;

[0010] The intelligent water resource allocation and reuse layer dynamically optimizes the water resource allocation strategy by integrating learning and demand forecasting algorithms, taking into account historical water use patterns, real-time rainfall forecasts, water quality conditions, and water use department needs;

[0011] The central intelligent management and optimization layer is equipped with an adaptive infiltration algorithm, a deep learning water level prediction algorithm, and an integrated learning and demand prediction algorithm. It conducts a comprehensive analysis of the data uploaded by the intelligent integrated monitoring layer, and provides control instructions for the dynamic infiltration regulation layer, the intelligent water collection and pretreatment layer, the water quality purification and resource utilization layer, and the intelligent water resource allocation and reuse layer.

[0012] Preferably, the data collected by the intelligent integrated monitoring layer include: rainfall intensity, soil moisture, water quality and groundwater level.

[0013] Preferably, the water provided by the intelligent water collection and pretreatment layer is deeply treated by sedimentation filtration.

[0014] Preferably, the adaptive penetration algorithm is as follows:

[0015] Set the initial penetration rate based on the penetration performance of the penetration material and historical data and adjustment factor ;

[0016] Calculate the comprehensive factor ;

[0017] ;

[0018] in, , , , All represent weight coefficients; Indicates rainfall intensity; Indicates soil moisture; Indicates water quality index; Indicates the groundwater level;

[0019] Dynamically adjust the penetration rate based on the initial penetration rate and comprehensive factors ;

[0020] ;

[0021] After adjusting the infiltration rate, the system operation effect is monitored in real time, including surface runoff, infiltration efficiency and groundwater level changes. The adjustment coefficient is adjusted according to the feedback data. and weight coefficient , , , Perform dynamic optimization;

[0022] ;

[0023] ;

[0024] in, Represents the optimized adjustment coefficient; Represents the optimized weight coefficient; Represents the weight coefficient The current value of , Both represent learning rate; It represents the change of permeation rate; Indicates the change in comprehensive factors.

[0025] Preferably, the goals of dynamic optimization are as follows:

[0026] ;

[0027] in, Represents comprehensive performance indicators; , , , , All represent weight coefficients; , , , They represent target rainfall intensity, target soil moisture, target groundwater level, and target water quality, respectively; , , , Respectively Rainfall intensity, soil moisture, groundwater level and water quality at any given moment; Indicates the maximum value of the moment;

[0028] The constraints are as follows:

[0029] ;

[0030] ;

[0031] ;

[0032] ;

[0033] ;

[0034] in, , represent the minimum and maximum values ​​of the permeation rate, respectively; , Respectively represent the minimum and maximum values ​​of groundwater level; , Respectively represent the minimum and maximum values ​​of soil moisture; Indicates the maximum permissible pollution concentration; Indicates the maximum change in permeability rate.

[0035] Preferably, the deep learning water level prediction algorithm comprises the following steps:

[0036] Get historical water level data , and normalize it to get the normalized water level data ;

[0037] An improved deep learning model consisting of a bidirectional long short-term memory network and an attention mechanism is constructed to predict future water levels.

[0038] Preferably, the integrated learning and demand forecasting algorithm is as follows:

[0039] Obtain historical water use data, real-time rainfall forecast data, water quality data, and user demand data to form an initial data set;

[0040] Preprocess the initial data set, including data cleaning, normalization, and feature selection;

[0041] Use the random forest algorithm to build a demand forecasting model;

[0042] According to the demand forecast value, the water resource allocation strategy is dynamically adjusted in combination with a multi-objective optimization algorithm. The optimization objectives include maximizing water resource utilization efficiency, maximizing economic benefits, and maximizing user satisfaction.

[0043] ;

[0044] ;

[0045] ;

[0046] in, Indicates the needs of each water-using sector; Indicates that the The amount of water resources in each water-using sector; Indicates The economic benefit function of each water-using sector; Indicates Satisfaction function of each water-using department; Indicates the number of water-using sectors;

[0047] ;

[0048] ;

[0049] in, is a positive coefficient, indicating that when water resources are allocated to When considering a water-using sector, the direct economic benefits brought by each unit of water resources; is a negative coefficient, indicating the effect of diminishing marginal benefits as water allocation increases; is a positive coefficient, indicating that when water resources are allocated to When the number of water users increases, the satisfaction level per unit of water resources increases; Indicates basic needs of each water-using sector; indicates that the marginal satisfaction after exceeding the basic needs decreases, and Less than ;

[0050] The constraints are as follows:

[0051] ;

[0052] ;

[0053] ;

[0054] in, Indicates the total amount of available water resources; , Respectively represent Minimum and maximum demands of each water-using sector;

[0055] The optimized water resource allocation scheme is applied to the sponge city rainwater treatment system to dynamically adjust the infiltration rate and water resource allocation to meet the needs of different water-using departments.

[0056] Preferably, the central intelligent management and optimization layer is also connected to the weather forecast system to obtain future rainfall intensity in real time.

[0057] Preferably, the central intelligent management and optimization layer also has self-learning and self-optimization capabilities. Based on the collected system operation data, it improves the algorithm prediction accuracy and resource allocation efficiency through online learning and model updating.

[0058] Preferably, through online learning and model updating, including:

[0059] Collect system operation data in real time, including rainfall intensity , soil moisture , water quality indicators , groundwater level and water sector needs ;

[0060] Based on the collected system operation data, model parameters are dynamically updated through meta-learning and continuous learning strategies, including:

[0061] Meta-learning optimization:

[0062] ;

[0063] in, Represents task-specific parameters; Represents the global initial parameters; represents the inner loop learning rate; weights representing demand-related losses in the water use sector; represents mean square error; Express The gradient of , , , , They represent the predicted values ​​of rainfall intensity, soil moisture, water quality index, groundwater level and demand of water-using sectors respectively;

[0064] Continuous learning and optimization:

[0065] ;

[0066] in, Represents the updated global parameters; represents the outer loop learning rate; represents the sum of all tasks, Indicates the task index;

[0067] Optimize resource allocation strategy based on updated global parameters and task-specific parameters:

[0068] ;

[0069] in, Indicates the optimized resource allocation plan; represents the cumulative gradient; represents the regularization coefficient; , Respectively represent L1 and L2 regularization coefficients; Indicates that based on task specific parameters resource allocation strategy; Indicates that based on the updated global parameters resource allocation strategy; Represents the resource allocation vector, which is used to represent the resource allocation strategy of the system at different time steps.

[0070] Therefore, the present invention adopts the above-mentioned novel sponge city rainwater reuse system, and the beneficial technical effects are as follows:

[0071] (1) Based on the existing sponge city rainwater treatment system, this invention introduces an intelligent integrated monitoring layer and a central intelligent management and optimization layer. By integrating multiple sensors and advanced algorithms, real-time monitoring and dynamic regulation of rainfall intensity, soil moisture, water quality and groundwater level are achieved. This intelligent improvement not only improves the efficiency of rainwater collection and treatment, but also optimizes the resource allocation strategy, effectively improving the operating performance of the existing system without making large-scale changes to the existing infrastructure.

[0072] (2) The present invention realizes dynamic regulation of infiltration rate and water level change through adaptive infiltration algorithm and deep learning water level prediction algorithm. This dynamic optimization strategy can accurately adjust the operating parameters according to real-time monitoring data on the basis of the existing system, optimize surface runoff and groundwater recharge, effectively alleviate urban waterlogging problems, and improve the utilization efficiency of water resources.

[0073] (3) The present invention introduces an integrated learning and demand prediction algorithm into the existing sponge city rainwater treatment system, and maximizes water resource utilization efficiency, economic benefits and user satisfaction by dynamically optimizing the water resource allocation strategy. This improvement enables the system to better meet the needs of different water-using departments, while promoting the sustainable use of urban water resources without the need to add additional complex hardware facilities.

[0074] (4) The central intelligent management and optimization layer of the present invention has the ability of self-learning and self-optimization, and can update model parameters in real time based on system operation data, thereby improving the algorithm prediction accuracy and resource allocation efficiency. This adaptive capability enables the system to better cope with the complex and changing needs of urban rainwater treatment, further improving the stability and reliability of the existing system without making fundamental changes to the existing system.

[0075] The improvement measures of the present invention have good compatibility and can be seamlessly connected to the existing sponge city rainwater treatment system. At the same time, the system is also scalable and can be further optimized and upgraded according to future urban development needs, providing a more forward-looking solution for urban water resources management. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 This is a flow chart of a novel sponge city rainwater reuse system of the present invention;

[0077] Figure 2 Flowchart of the adaptive infiltration algorithm. DETAILED DESCRIPTION

[0078] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0079] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0080] Embodiment 1

[0081] The present invention provides a novel sponge city rainwater recycling system, comprising:

[0082] The intelligent integrated monitoring layer integrates rain gauges, soil moisture sensors, water quality monitoring probes and ultrasonic water level meters to form an all-round, multi-level monitoring system covering key areas of the city, collecting and uploading data in real time to the central intelligent management system.

[0083] This example takes a typical residential area in a city as an example. The area includes multiple land use types such as residences, parks, schools and commercial facilities. The core component of the system, the intelligent integrated monitoring layer, is equipped with the following equipment:

[0084] Rain gauges: located at 10 key locations across the region, including:

[0085] A park square in the center of the area is used to monitor concentrated rainfall conditions.

[0086] School playground to understand the rainfall intensity in the school area.

[0087] Intersections of major streets to monitor rainfall in areas with heavy traffic.

[0088] At the entrance of the residential area, assess the rainfall conditions in the residential area.

[0089] The central square of the commercial district monitors rainfall intensity in the commercial area.

[0090] Near the lake in the park, it is used to monitor rainfall in natural areas.

[0091] Green belts on both sides of the road, assess rainfall conditions in green areas.

[0092] Community hospital rooftop to understand rainfall conditions on top of the building.

[0093] At the edge of the industrial area, monitor the rainfall intensity in the industrial area.

[0094] Near groundwater monitoring wells, used to compare rainfall effects on surface and groundwater.

[0095] Soil moisture sensors: 50 sensors have been installed in parks, gardens and green belts along roads to monitor soil moisture changes.

[0096] Water quality monitoring probes: Installed in three rainwater collection tanks and two groundwater monitoring wells in the area to detect water quality indicators such as pH, dissolved oxygen, chemical oxygen demand (COD) and ammonia nitrogen content.

[0097] Ultrasonic water level meter: installed at key nodes of rainwater collection tanks and drainage pipes to monitor water level changes.

[0098] The data collected and uploaded by the intelligent integrated monitoring layer include: rainfall intensity, soil moisture, water quality and groundwater level.

[0099] The dynamic infiltration adjustment layer uses nano-scale intelligent infiltration materials and combines electromagnetic and pressure sensing control technology. It uses an adaptive infiltration algorithm to predict and adjust the infiltration rate based on the rainfall intensity, soil moisture, water quality and groundwater level data uploaded by the intelligent integrated monitoring layer received by the central intelligent management system.

[0100] Electromagnetic induction control technology:

[0101] This technology uses changes in the electromagnetic field to adjust the pore structure of nano-scale intelligent permeable materials. By embedding tiny electromagnetic induction elements in the permeable material, the system can dynamically adjust the electromagnetic field strength according to the instructions of the central intelligent management system. When the electromagnetic field strength changes, the pore structure of the permeable material will change accordingly, thereby adjusting the permeation rate. For example, when the rainfall intensity is high, the electromagnetic field is enhanced to increase the pores of the permeable material and accelerate the permeation rate; when the soil moisture is high, the electromagnetic field is weakened, the pores are reduced, the permeation rate is reduced, and the groundwater level is prevented from being too high.

[0102] Pressure sensing control technology:

[0103] Pressure sensing control technology adjusts the penetration rate by monitoring the pressure changes in the permeation layer in real time. High-precision pressure sensors are installed in the permeation layer, which can detect the pressure changes of water in the permeation process in real time. When the pressure sensor detects that the pressure of the permeation layer is too high, the system will automatically adjust the permeation rate to prevent the permeation layer from being blocked or excessively accumulating water. In this way, pressure sensing control technology can effectively ensure the stability and safety of the permeation process.

[0104] like Figure 2 As shown, the adaptive penetration algorithm includes the following steps:

[0105] Set the initial penetration rate based on the penetration performance of the penetration material and historical data and adjustment factor ;

[0106] Calculate the comprehensive factor ;

[0107] ;

[0108] in, , , , All represent weight coefficients; Indicates rainfall intensity; Indicates soil moisture; Indicates water quality index; Indicates the groundwater level;

[0109] Dynamically adjust the penetration rate based on the initial penetration rate and comprehensive factors ;

[0110] ;

[0111] After adjusting the infiltration rate, the system operation effect is monitored in real time, including surface runoff, infiltration efficiency and groundwater level changes. The adjustment coefficient is adjusted according to the feedback data. and weight coefficient , , , Perform dynamic optimization;

[0112] ;

[0113] ;

[0114] in, Represents the optimized adjustment coefficient; Represents the optimized weight coefficient; Represents the weight coefficient The current value of , Both represent learning rate; It represents the change of permeation rate; Indicates the change in comprehensive factors.

[0115] The goals of dynamic optimization are as follows:

[0116] ;

[0117] in, Represents comprehensive performance indicators; , , , , All represent weight coefficients; , , , They represent target rainfall intensity, target soil moisture, target groundwater level, and target water quality, respectively; , , , Respectively Rainfall intensity, soil moisture, groundwater level and water quality at any given moment; Indicates the maximum value of the moment;

[0118] The constraints are as follows:

[0119] ;

[0120] ;

[0121] ;

[0122] ;

[0123] ;

[0124] in, , represent the minimum and maximum values ​​of the permeation rate, respectively; , Respectively represent the minimum and maximum values ​​of groundwater level; , Respectively represent the minimum and maximum values ​​of soil moisture; Indicates the maximum permissible pollution concentration; Indicates the maximum change in permeability rate.

[0125] The intelligent water collection and pretreatment layer includes an automatically cleaning precision filter and an adjustable flow inlet valve, which uses a deep learning water level prediction algorithm to predict water level changes in the water collection well.

[0126] Automatic cleaning fine filter:

[0127] Filtration accuracy: The precision filter uses multiple layers of filtration media, including a stainless steel filter and an activated carbon filter layer. The filter has a pore size of 100 microns, which can effectively remove large particles of impurities (such as leaves, sand, etc.) in rainwater, while the activated carbon filter layer is used to absorb organic matter and odor in the water, further improving water quality.

[0128] Automatic cleaning function: To avoid filter blockage, the system is equipped with an automatic cleaning device. The device uses backwashing technology to clean the filter regularly. The specific working principle is as follows:

[0129] Pressure sensor monitoring: Pressure sensors are installed at the inlet and outlet of the filter to monitor the pressure difference of the filter in real time. When the pressure difference exceeds the set threshold (such as 0.1 MPa), the system automatically starts the backwashing procedure.

[0130] Backwashing process: During backwashing, the water flow direction inside the filter is reversed, and the water flows through the reverse channel to wash away the impurities attached to the filter screen and discharge the system through the sewage outlet. The entire cleaning process does not require manual intervention, and the cleaning time is 30 seconds each time, which does not affect the normal filtering work.

[0131] The precision filter is installed at the entrance of the rainwater collection tank to initially purify the rainwater collected from roofs, roads and green spaces. Through the automatic cleaning function, the filter can operate stably for a long time, ensuring that the rainwater entering the subsequent treatment link meets the water quality requirements.

[0132] Adjustable flow inlet valve:

[0133] Electric regulating valve: The water inlet valve adopts an electric regulating valve, which can accurately adjust the water inlet flow according to the instructions of the central intelligent management system. The adjustment range of the valve is 0-100%, which can achieve stepless adjustment from fully closed to fully open.

[0134] Flow sensor and feedback control: Flow sensors are installed on the water inlet pipe to monitor the water inlet flow in real time. The sensor transmits the flow data to the central intelligent management system, which dynamically adjusts the opening of the water inlet valve according to the water level prediction results and real-time flow data to ensure that the water level in the rainwater collection pool remains within a reasonable range.

[0135] Linkage control: The adjustment of the water inlet valve works in conjunction with the intelligent water level monitor and precision filter. For example, when the water level monitor predicts that the water level is about to exceed the warning value, the system automatically reduces the opening of the water inlet valve; and when the filter starts the backwashing program, the system will appropriately reduce the water flow to ensure the cleaning effect.

[0136] In this embodiment, the inlet valve with adjustable flow is installed on the inlet pipe of the rainwater collection pool. Through linkage with the intelligent water level monitor and the central intelligent management system, the inlet valve can dynamically adjust the inlet flow according to real-time data, optimize the rainwater collection efficiency, and avoid overflow or filter blockage caused by excessive flow.

[0137] The deep learning water level prediction algorithm includes the following steps:

[0138] Get historical water level data , and normalize it to get the normalized water level data ;

[0139] An improved deep learning model consisting of a bidirectional long short-term memory network and an attention mechanism is constructed to predict future water levels.

[0140] The water purification and resource utilization layer is used to deeply process the water provided by the intelligent water collection and pretreatment layer, and the output water quality meets the urban water use standards.

[0141] The water provided by the intelligent water collection and pretreatment layer is deeply treated as sedimentation treatment.

[0142] The treated rainwater is stored in underground tanks and used for green irrigation, road spraying and some non-drinking water needs.

[0143] The intelligent water resource allocation and reuse layer, including intelligent water pumps, flow control valves, and multiple user (water-using department) access points, dynamically optimizes the water resource allocation strategy by integrating learning and demand forecasting algorithms, taking into account historical water use patterns, real-time rainfall forecasts, water quality conditions, and water-using department needs.

[0144] The integrated learning and demand forecasting algorithm includes the following steps:

[0145] Obtain historical water use data, real-time rainfall forecast data, water quality data, user demand data, and crop growth cycle data to form an initial data set;

[0146] Preprocess the initial data set, including data cleaning, normalization, and feature selection;

[0147] The demand forecasting model is constructed using the random forest algorithm. The demand forecasting model is implemented through the following steps:

[0148] a. Perform multiple random sampling with replacement on the preprocessed data set to obtain multiple training sample sets;

[0149] b. For each training sample set, a regression tree is constructed, and all regression trees form a random forest model;

[0150] c. Take a weighted average of the output results of each regression tree to obtain the final demand forecast value;

[0151] According to the demand forecast value, the water resource allocation strategy is dynamically adjusted in combination with a multi-objective optimization algorithm. The optimization objectives include maximizing water resource utilization efficiency, maximizing economic benefits, and maximizing user satisfaction.

[0152] ;

[0153] ;

[0154] ;

[0155] in, Indicates the needs of each water-using sector; Indicates that the The amount of water resources in each water-using sector; Indicates The economic benefit function of each water-using sector; Indicates Satisfaction function of each water-using department; Indicates the number of water-using sectors;

[0156] ;

[0157] ;

[0158] in, is a positive coefficient, indicating that when water resources are allocated to When considering a water-using sector, the direct economic benefits brought by each unit of water resources; is a negative coefficient, indicating the effect of diminishing marginal benefits as water allocation increases; is a positive coefficient, indicating that when water resources are allocated to When the number of water users increases, the satisfaction level per unit of water resources increases; Indicates basic needs of each water-using sector; indicates that the marginal satisfaction after exceeding the basic needs decreases, and Less than ;

[0159] The constraints are as follows:

[0160] ;

[0161] ;

[0162] ;

[0163] in, Indicates the total amount of available water resources; , Respectively represent Minimum and maximum demands of each water-using sector;

[0164] The optimized water resource allocation scheme is applied to the sponge city rainwater treatment system to dynamically adjust the infiltration rate and water resource allocation to meet the needs of different water-using departments.

[0165] The central intelligent management and optimization layer is equipped with an adaptive infiltration algorithm, a deep learning water level prediction algorithm, and an integrated learning and demand prediction algorithm. It conducts a comprehensive analysis of the data uploaded by the intelligent integrated monitoring layer, and provides control instructions for the dynamic infiltration regulation layer, the intelligent water collection and pretreatment layer, the water quality purification and resource utilization layer, and the intelligent water resource allocation and reuse layer.

[0166] The central intelligent management and optimization layer is also connected to the weather forecast system to obtain future rainfall intensity in real time.

[0167] The central intelligent management and optimization layer also has the ability to self-learn and self-optimize. Based on the collected system operation data, it improves the algorithm prediction accuracy and resource allocation efficiency through online learning and model updates, including:

[0168] Collect system operation data in real time, including rainfall intensity , soil moisture , water quality indicators , groundwater level and water sector needs ;

[0169] Based on the collected system operation data, model parameters are dynamically updated through meta-learning and continuous learning strategies, including:

[0170] Meta-learning optimization:

[0171] ;

[0172] in, Represents task-specific parameters; Represents the global initial parameters; represents the inner loop learning rate; weights representing demand-related losses in the water use sector; represents mean square error; Express The gradient of , , , , They represent the predicted values ​​of rainfall intensity, soil moisture, water quality index, groundwater level and demand of water-using sectors respectively;

[0173] Continuous learning and optimization:

[0174] ;

[0175] in, Represents the updated global parameters; represents the outer loop learning rate; represents the sum of all tasks, Indicates the task index;

[0176] Optimize resource allocation strategy based on updated global parameters and task-specific parameters:

[0177] ;

[0178] in, Indicates the optimized resource allocation plan; represents the cumulative gradient; represents the regularization coefficient; , Respectively represent L1 and L2 regularization coefficients; Indicates that based on task specific parameters resource allocation strategy; Indicates that based on the updated global parameters resource allocation strategy; Represents the resource allocation vector, which is used to represent the resource allocation strategy of the system at different time steps.

[0179] It is worth noting that the contents not elaborated in detail in the present invention are all prior art and are well known to those skilled in the art.

[0180] Therefore, the present invention adopts the above-mentioned novel sponge city rainwater reuse system to achieve efficient collection, purification and resource utilization of rainwater, and improve the intelligent level of urban water resource management.

[0181] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A new type of sponge city rainwater reuse system, characterized in that: include: The intelligent integrated monitoring layer integrates rain gauges, soil moisture sensors, water quality monitoring probes and ultrasonic water level meters to form a monitoring system covering key areas of the city, collecting and uploading data in real time to the central intelligent management system; The dynamic permeability adjustment layer uses nano-scale intelligent permeability materials, combined with electromagnetic and pressure sensing control technology, and uses adaptive permeability algorithms to predict and adjust the permeation rate based on the data uploaded by the intelligent integrated monitoring layer received by the central intelligent management system; Intelligent water collection and preprocessing layer, using deep learning water level prediction algorithm to predict water level changes in water collection wells; The water purification and resource utilization layer is used to deeply process the water provided by the intelligent water collection and pretreatment layer, and the output water quality meets the urban water use standards; The intelligent water resource allocation and reuse layer dynamically optimizes the water resource allocation strategy by integrating learning and demand forecasting algorithms, taking into account historical water use patterns, real-time rainfall forecasts, water quality conditions, and water use department needs; The central intelligent management and optimization layer is equipped with an adaptive infiltration algorithm, a deep learning water level prediction algorithm, and an integrated learning and demand prediction algorithm. It conducts a comprehensive analysis of the data uploaded by the intelligent integrated monitoring layer and provides control instructions for the dynamic infiltration adjustment layer, the intelligent water collection and pretreatment layer, the water quality purification and resource utilization layer, and the intelligent water resource allocation and reuse layer. The integrated learning and demand forecasting algorithm is as follows: Obtain historical water use data, real-time rainfall forecast data, water quality data, and user demand data to form an initial data set; Preprocess the initial data set, including data cleaning, normalization, and feature selection; Use the random forest algorithm to build a demand forecasting model; According to the demand forecast value, the water resource allocation strategy is dynamically adjusted in combination with a multi-objective optimization algorithm. The optimization objectives include maximizing water resource utilization efficiency, maximizing economic benefits, and maximizing user satisfaction. ； ； ； in, Indicates the needs of each water-using sector; Indicates that the The amount of water resources in each water-using sector; Indicates The economic benefit function of each water-using sector; Indicates Satisfaction function of each water-using department; Indicates the number of water-using sectors; ； ； in, is a positive coefficient, indicating that when water resources are allocated to When considering a water-using sector, the direct economic benefits brought by each unit of water resources; is a negative coefficient, indicating the effect of diminishing marginal benefits as water allocation increases; is a positive coefficient, indicating that when water resources are allocated to When the number of water users increases, the satisfaction level per unit of water resources increases; Indicates basic needs of each water-using sector; indicates that the marginal satisfaction after exceeding the basic needs decreases, and Less than ; The constraints are as follows: ； ； ； in, Indicates the total amount of available water resources; , Respectively represent Minimum and maximum demands of each water-using sector; The optimized water resource allocation scheme is applied to the sponge city rainwater treatment system to dynamically adjust the infiltration rate and water resource allocation to meet the needs of different water-using departments.

2. According to claim 1, a new type of sponge city rainwater reuse system is characterized in that: The data collected by the intelligent integrated monitoring layer include: rainfall intensity, soil moisture, water quality and groundwater level.

3. According to the novel sponge city rainwater reuse system of claim 1, it is characterized in that: The water provided by the intelligent water collection and pretreatment layer is deeply treated as sedimentation filtration.

4. According to the novel sponge city rainwater reuse system of claim 1, it is characterized in that: The adaptive penetration algorithm is as follows: Set the initial penetration rate based on the penetration performance of the penetration material and historical data and adjustment factor ; Calculate the comprehensive factor ; ； in, , , , All represent weight coefficients; Indicates rainfall intensity; Indicates soil moisture; Indicates water quality indicators; Indicates the groundwater level; Dynamically adjust the penetration rate based on the initial penetration rate and comprehensive factors ; ； After adjusting the infiltration rate, the system operation effect is monitored in real time, including surface runoff, infiltration efficiency and groundwater level changes. The adjustment coefficient is adjusted according to the feedback data. and weight coefficient , , , Perform dynamic optimization; ； ； in, Represents the optimized adjustment coefficient; Represents the optimized weight coefficient; Represents the weight coefficient The current value of , Both represent learning rate; It represents the change of permeation rate; Indicates the change in comprehensive factors.

5. A novel sponge city rainwater reuse system according to claim 4, characterized in that: The goals of dynamic optimization are as follows: ； in, Represents comprehensive performance indicators; , , , , All represent weight coefficients; , , , They represent target rainfall intensity, target soil moisture, target groundwater level, and target water quality, respectively; , , , Respectively Rainfall intensity, soil moisture, groundwater level and water quality at any given moment; Indicates the maximum value of the moment; The constraints are as follows: ； ； ； ； ； in, , represent the minimum and maximum values ​​of the permeation rate, respectively; , Respectively represent the minimum and maximum values ​​of groundwater level; , Respectively represent the minimum and maximum values ​​of soil moisture; Indicates the maximum permissible pollution concentration; Indicates the maximum change in permeability rate.

6. A novel sponge city rainwater reuse system according to claim 5, characterized in that: The deep learning water level prediction algorithm includes the following steps: Get historical water level data , and normalize it to get the normalized water level data ; An improved deep learning model consisting of a bidirectional long short-term memory network and an attention mechanism is constructed to predict future water levels.

7. A novel sponge city rainwater reuse system according to claim 6, characterized in that: The central intelligent management and optimization layer is also connected to the weather forecast system to obtain future rainfall intensity in real time.

8. The novel sponge city rainwater reuse system according to claim 7 is characterized in that: The central intelligent management and optimization layer also has self-learning and self-optimization capabilities. Based on the collected system operation data, it improves the algorithm prediction accuracy and resource allocation efficiency through online learning and model updates.

9. The novel sponge city rainwater reuse system according to claim 8 is characterized in that: Through online learning and model updates, including: Collect system operation data in real time, including rainfall intensity , soil moisture , water quality indicators , groundwater level and water sector needs ; Based on the collected system operation data, model parameters are dynamically updated through meta-learning and continuous learning strategies, including: Meta-learning optimization: ； in, Represents task-specific parameters; Represents the global initial parameters; represents the inner loop learning rate; weights representing demand-related losses in the water use sector; represents mean square error; Express The gradient of , , , , They represent the predicted values ​​of rainfall intensity, soil moisture, water quality index, groundwater level and demand of water-using sectors respectively; Continuous learning and optimization: ； in, Represents the updated global parameters; represents the outer loop learning rate; represents the sum of all tasks, Indicates the task index; Optimize resource allocation strategy based on updated global parameters and task-specific parameters: ； in, Indicates the optimized resource allocation plan; represents the cumulative gradient; represents the regularization coefficient; , Respectively represent L1 and L2 regularization coefficients; Indicates that based on task specific parameters resource allocation strategy; Indicates that based on the updated global parameters resource allocation strategy; Represents the resource allocation vector, which is used to represent the resource allocation strategy of the system at different time steps.

Images