Intelligent water supply decision generation method and system

Abstract

The invention discloses an intelligent water supply decision generation method and system. The method belongs to the technical field of municipal water conservancy project decision. The method mainly comprises the steps that multi-source heterogeneous data and GIS spatial data are collected in real time, and data quality verification and standardization processing are executed; establishing an analysis system comprising a water supply safety evaluation model, a pipe network leakage detection model and a water quality pollution prediction model, and generating a comprehensive evaluation result by each model output index; performing spatial superposition on real-time monitoring data and a GIS map to draw a water affair map, and generating a pipe network pressure pressurization path and a water quality pollution traceability path through topological relation analysis; and a regulation and control instruction is dynamically generated based on a visual analysis result, and early warning information is pushed. According to the method, the GIS map and the real-time multi-source data stream are deeply fused, a multi-layer linkage analysis system is constructed, and the cross-system data utilization rate is improved. And integrating multi-source heterogeneous data by using a water affair graph, and checking monitoring data in real time.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of municipal water conservancy engineering decision-making, in particular to a smart water supply decision-making generation method and system. BACKGROUND

[0002] Traditional water management has significant technical bottlenecks in the fields of pipe network pressure regulation and water quality pollution tracing: on the one hand, pipe network pressure optimization mainly relies on manual experience to develop pressure schemes, lacking multi-physical field coupling analysis of pipe segment topological parameters (length, pipe diameter), dynamic pressure data and pipe pressure threshold. This experience-driven mode leads to uneven pressure distribution, which easily causes local overpressure or low pressure area, increasing the risk of pipe explosion; the mismatch between pump station energy consumption strategy and actual pipe network carrying capacity causes energy waste and increases the risk of pipe explosion. In particular, in large-scale complex pipe networks, existing pressure optimization algorithms (such as genetic algorithm) face slow convergence speed, insufficient constraint condition processing and other problems in large-scale complex pipe networks, making it difficult to generate optimal paths that meet safety constraints and economic efficiency in real time; on the other hand, linear water quality monitoring systems have the phenomenon of "data island", and water quality pollution tracing relies mainly on single-point sensor detection and manual investigation, lacking sufficient extraction accuracy of characteristic parameters such as concentration gradient and diffusion rate, making it difficult to accurately locate the pollution source in the early stage of diffusion, resulting in the expansion of pollution range and the lag of disposal. At the same time, the existing system lacks the ability to integrate GIS geographic information and real-time monitoring data, and the spatial topological analysis is limited to static layer display, which cannot dynamically superimpose multi-dimensional data flow such as pressure field and water quality field, resulting in a lack of spatial correlation and timeliness in decision support. In addition, the pipe network pressure model and water quality model in traditional technology are isolated from each other, and a data-driven analysis framework across physical fields cannot be constructed, making it difficult to deal with coupled problems such as water quality pollution caused by pipe network leakage and pressure fluctuation aggravating pipe aging. The above defects lead to the problems of low decision-making accuracy, slow response speed and high operation and maintenance cost in water management systems. Therefore, the present application proposes a smart water supply decision-making generation method and system. SUMMARY

[0003] Based on the above problems existing in the prior art, the present application solves the problem of providing a smart water supply decision-making generation method and system, which solves the problems of data island, low decision-making accuracy, slow response speed and high operation and maintenance cost in water management systems.

[0004] The technical solution adopted by the present application to solve its technical problems is: a smart water supply decision-making generation method, comprising the following steps:

[0005] S1: Real-time collection of pipe network pressure data, water quality monitoring data, user water consumption data and GIS spatial data, and execution of data quality verification and standardization processing;

[0006] S2: Establish an analysis system containing a water supply safety assessment model, a pipe network leakage detection model, and a water quality pollution prediction model. The output indicators of each model are fused by weighting to generate a comprehensive evaluation result;

[0007] S3: Spatially superimpose real-time monitoring data and GIS maps to draw a water map. Through topological relationship analysis, generate pipe network pressure boosting paths and water quality pollution tracing paths;

[0008] S4: Dynamically generate control instructions based on visual analysis results, including pump station pressure adjustment instructions and cross-system early warning information push.

[0009] Further, the data quality verification uses a multi-dimensional evaluation formula:

[0010] ;

[0011] Where: Q is the data quality comprehensive score, which is controlled by threshold Q0. If the quality score is lower than Q0, it is determined as abnormal data; n is the number of data sources; is the weight of the ith type of data; V is the number of valid data; T is the total data volume; is the consistency factor; C is the cross-system data consistency score.

[0012] Further, the water supply safety assessment model uses a dynamic balance algorithm:

[0013] ;

[0014] Where: is the pressure anomaly index; is the real-time water supply; is the water demand predicted by the LSTM network; is the variance of water supply; is the temperature influence coefficient; T is the pipe network water temperature.

[0015] Further, the pipe network leakage detection model is:

[0016] ;

[0017] Where: L is the leakage index, which quantifies the severity of pipe network leakage in a specific time period; are the pressures at the user pipe network inlet and outlet, respectively; is the pipe network material attenuation coefficient; t1, t2 are the start and end times of leakage detection.

[0018] Furthermore, a comprehensive water resources map is created by spatially overlaying real-time monitoring data with a GIS map. Topological relationship analysis is used to generate pipeline pressure boosting paths and water pollution source tracing paths. This comprehensive water resources map, created by spatially overlaying real-time monitoring data with a GIS map, includes the following elements:

[0019] Basic geographical layer: includes administrative divisions and water system distribution;

[0020] Facilities layer: Includes the topological relationships of water plants, pumping stations, and pipe networks;

[0021] Monitoring layer: Integrates real-time data from pressure, flow, and water quality monitoring points;

[0022] User layer: Mark the water usage characteristics of key users;

[0023] Protection zone layer: Delineate water source protection areas and ecological red lines.

[0024] Furthermore, in step S3, the spatial topology analysis includes pipeline pressure boosting paths and water pollution source tracing.

[0025] Furthermore, the pipeline pressure boosting path is generated based on the Dijkstra algorithm, and its calculation model is as follows:

[0026] ;

[0027] Where: Cost is the comprehensive cost of the route, which represents the comprehensive cost of pressurization and regulation of the pipeline route. The smaller the value, the better the route. is the pipe segment length, specifically representing the actual physical length of the i-th pipe segment; The diameter is the diameter of the i-th segment of the pipe, which reflects the water conveyance capacity. The current pressure is represented by the real-time water pressure monitoring value of the i-th segment of the pipeline. This is the pressure threshold, specifically representing the maximum safe pressure that the pipeline can withstand.

[0028] Furthermore, the water pollution source tracing path uses the backpropagation algorithm to locate the pollution source, and its calculation model is as follows: ;

[0029] Where: S(x, y) are the coordinates of the pollution source, specifically the coordinates of the source location of the pollutant diffusion; For concentration gradient, specifically the rate of change of pollutant concentration field in space, pointing in the direction of the fastest concentration increase; It is the partial derivative of time with respect to space, specifically the rate at which time changes with spatial location during the diffusion of pollutants; t represents the pollutant concentration field, specifically the pollutant concentration value at coordinates (x, y); t represents the diffusion time, specifically the time it takes for the pollutant to diffuse from the source to the detection point.

[0030] The wisdom water supply decision generation system is a system for implementing a wisdom water supply decision generation method, comprising a data acquisition and exchange layer, a data middle platform layer, an intelligent analysis layer, and a visualization and regulation layer.

[0031] Further, the data acquisition and exchange layer comprises:

[0032] A multi-source adaptation module: supports data acquisition of relational databases MySQL / Oracle, time series databases InfluxDB, GIS spatial databases PostGIS, and API interfaces, and simultaneously has an ETL work flow engine built-in to realize data cleaning, splitting, and merging;

[0033] A real-time exchange module: uses Apache Kafka to build a distributed message queue, and the real-time exchange module is based on Kafka partition horizontal expansion, and under a 32-node cluster, the measured throughput is 1.2M per second (1KB per data), the end-to-end delay is <50ms, and a full / incremental data synchronization strategy is provided;

[0034] A disaster recovery module: based on database log replication technology to realize millisecond-level synchronization of master and standby databases, and supports off-site disaster recovery;

[0035] The data middle platform layer comprises:

[0036] A subject data warehouse: divides data subjects according to water supply safety, pipe network loss, and water pollution, and the storage structure comprises fact tables and dimension tables;

[0037] A data governance engine, comprising:

[0038] Metadata management: tracks data sources and conversion processes through lineage analysis;

[0039] Quality monitoring: defines quality rules;

[0040] Security control: realizes field-level data permission control based on the RBAC model;

[0041] The intelligent analysis layer comprises: a prediction model library, which comprises water supply demand prediction based on an ARIMA time series model, water quality anomaly detection using an isolated forest algorithm, and pipe network life assessment using a Weibull distribution model;

[0042] A real-time computing engine: uses Flink to realize stream computing to calculate a pipe network pressure balance index;

[0043] The visualization and regulation layer comprises: a water affair one map module: integrates an OpenLayers geographic engine to support 3D pipe network rendering and section analysis, heat map superposition, and buffer zone analysis functions;

[0044] The wisdom big screen module preinstalls emergency command, daily monitoring, leader board template, and supports multi-chart linkage;

[0045] The open platform module provides RESTful API and SDK, and supports third-party system to call data service and early warning interface.

[0046] The beneficial effects of the present application are: the present application constructs a pipe network pressure optimization model based on the Dijkstra algorithm, generates an optimal pressurization path considering safety and economy by comprehensively calculating the dynamic relationship of pipe length, pipe diameter, real-time pressure and pressure threshold, and reduces the energy consumption of the pump station and improves the pressure balance compared with the traditional method. The present application applies the back propagation algorithm to analyze the gradient characteristics of the pollutant concentration field, and reversely traces the diffusion path combined with the topological relationship of the pipe network, so that the positioning accuracy of the pollution source is improved, the response time is shortened, and the pollution diffusion is effectively curbed. The present application deeply integrates GIS map and real-time data flow, constructs a multi-layer linkage analysis system of pipe network pressure field, water quality field and user water field, realizes the minute-level decision of complex scenes such as pipe burst influence range prediction and pollution diffusion simulation through spatial topological calculation, and improves the utilization rate of cross-system data. The present application integrates multi-source heterogeneous data through water affair map, and realizes real-time monitoring of the data. The present application early warns the water quality risk caused by pipe network leakage through the pressure-water quality coupling model, assists in formulating preventive maintenance strategy, reduces the comprehensive operation and maintenance cost of the pipe network, improves the standard rate of water supply safety, provides high-precision, full-factor and real-time decision support for smart water affair construction. The present application realizes one-stop data integration and data management. In the global planning aspect: globally design a big data center, design a standard model, and unify data index caliber; in the data fusion aspect: connect any data source, automatically reconstruct metadata and theme data, and provide unified data service for application; in the asset management aspect: globally control and intelligently manage data assets, efficiently govern data, track data usage and generated value; in the intelligent analysis aspect: analyze the managed data through machine learning algorithm, and sublimate traditional statistical analysis to the level of data mining and BI analysis; in the data sharing aspect: provide the data obtained or calculated by the data platform to third-party application through API, file, database and other ways, unify the standard, and avoid waste of computing power. In addition to the purposes, features and advantages described above, the present application has other purposes, features and advantages. The present application will be further described below with reference to the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0047] The drawings accompanying the specification of the present application form a part thereof, serve to provide further understanding of the present application, and together with the description of the present application, serve to explain the present application, and do not constitute improper limitations on the present application. In the drawings:

[0048] Fig. 1This is a flowchart of a smart water supply decision generation method proposed in this invention;

[0049] Fig. 2 This is a diagram of the architecture of a smart water supply decision generation system proposed in this invention;

[0050] Fig. 3 The present invention proposes a system thermal map interface. Detailed Implementation

[0051] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0052] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0053] Example 1:

[0054] like Figs. 1-3 As shown, this application provides a smart water supply decision generation method, which includes the following steps:

[0055] S1: Real-time collection of water supply network pressure data, water quality monitoring data, user water consumption data and GIS spatial data, and execution of data quality verification and standardization processing;

[0056] Water supply network pressure data, water quality monitoring data, user water consumption data, and GIS spatial data are multi-source heterogeneous data, which are collected in real time through a memory-level data exchange module; among them, water quality monitoring data is obtained through water quality monitoring terminals, and user water consumption data is obtained through smart water meters; the data undergoes standardized cleaning processing.

[0057] Data standardization and cleaning processes include handling missing values, handling outliers, data standardization and normalization, and data format unification.

[0058] Handling missing values ​​involves checking for null values, placeholders, or implicit missing values ​​(such as "unknown" in text) in the data. For detected missing values, the missing rows or columns are directly deleted. Missing values ​​are filled with the mean, median, mode, fixed values, or predictive models (such as regression or KNN). A new binary column is added to mark missing values ​​(e.g., is_missing=1).

[0059] Outlier handling specifically involves: Statistical methods: Z-score (>3 or <-3 is considered outlier), IQR (interquartile range) method (for values ​​exceeding 1.5×IQR). Visualization: Boxplots, scatter plots. Deletion: Directly remove outlier records. Correction: Replace with a cutoff value (e.g., 99th percentile) or discretize by binning.

[0060] Data standardization and normalization specifically involve eliminating dimensional differences to make the data suitable for model calculations. Common methods include Z-score standardization and Min-Max normalization.

[0061] The specific manifestations of data format standardization are: formatting dates / times to YYYY-MM-DD or timestamps; and standardizing text by converting case (e.g., converting all to lowercase) and removing spaces and special characters (e.g., #, !).

[0062] The data quality verification employs a multi-dimensional evaluation formula:

[0063] ;

[0064] Where: Q is the overall data quality score, which is controlled by a threshold Q0. Data with a quality score lower than Q0 is considered abnormal; n is the number of data sources. V represents the weight of the i-th data category; V represents the number of valid data entries; T represents the total data volume. C is the consistency factor; C is the cross-system data consistency score.

[0065] S2: Establish an analysis system that includes a water supply safety assessment model, a pipeline leakage detection model, and a water pollution prediction model. The output indicators of each model are weighted and fused to generate a comprehensive assessment result.

[0066] Specifically, the water security assessment model employs a dynamic equilibrium algorithm:

[0067] ;

[0068] in: This is an abnormal stress index; This refers to the real-time water supply volume. To predict water demand using an LSTM network; The variance of the water supply; is the temperature influence coefficient; T is the water temperature in the pipe network.

[0069] The pipeline leakage detection model is as follows:

[0070] ;

[0071] Where: L is the leakage index, which quantifies the severity of leakage in the pipe network within a specific time period; These are the pressures at the user's pipeline inlet and outlet, respectively. t1 represents the attenuation coefficient of the pipeline material; t1 and t2 represent the start and end times of the leakage detection.

[0072] The water pollution prediction model adopts a spatiotemporal diffusion equation:

[0073] ;

[0074] in: D represents the pollutant concentration at position x, y at time t; x D y and are the diffusion coefficients in the x and y directions, respectively, and are inversely proportional to the water flow velocity; v x ,v y λ represents the velocity field component of the pipe network water flow; λ represents the pollutant decay rate; S(x0,y0) is the pollution source term function, which takes the value of 1 when (x,y)=(x0,y0), and 0 otherwise.

[0075] The location of pollution sources is achieved by solving the inverse problem of the spatiotemporal diffusion equation, and the specific steps include: (a) Constructing the cost function: ; Where: C obs C represents the actual measured value at the monitoring point. sim These are simulated values; (b) The gradient is calculated using the adjoint method; (c) Iteratively update the pollution source coordinates (x0, y0) using the conjugate gradient method until J < $.

[0076] S3: A comprehensive water resources map is created by spatially overlaying real-time monitoring data with a GIS map. This spatial overlay generates a dynamic topology network, providing graph structure input for the Dijkstra's algorithm and backpropagation algorithm. Topological relationships are then used to analyze the pressure boosting paths of the water supply network and the source tracing paths of water pollution. Specifically, this comprehensive water resources map integrates all elements involved in water enterprise management onto a single GIS map to achieve comprehensive monitoring and management of water resources. It includes the following elements:

[0077] Basic geographic layer: includes administrative divisions and water system distribution, specifically showing the location and boundaries of water bodies such as water sources, rivers, lakes, and reservoirs, as well as the administrative divisions to which they belong;

[0078] Facilities layer: This includes the topological relationships of water plants, pumping stations, and pipe networks. Specifically, it includes the location, capacity, and spatial layout of water supply facilities such as water treatment plants, pumping stations, water supply networks, and water tanks.

[0079] Monitoring layer: Integrates real-time data from pressure, flow, and water quality monitoring points. Specifically, it indicates the location of water quality monitoring points and is used to monitor the water quality status and changes in the water body.

[0080] User layer: Mark the water usage characteristics of key users, specifically: mark important water-using units, such as industrial enterprises, farmland, urban residential areas, etc., as well as their water consumption and water usage characteristics;

[0081] Protection zone layer: Delineating water source protection areas and ecological red lines, specifically manifested as marking environmental protection areas, ecological protection areas, and water source protection areas to strengthen the protection and management of water resources.

[0082] Specifically, the pipeline pressure optimization calculation model based on the Dijkstra algorithm to generate the optimal pressurization path is as follows:

[0083] ;

[0084] Where: Cost is the comprehensive cost of the route, which represents the comprehensive cost of pressurization and regulation of the pipeline route. The smaller the value, the better the route. is the pipe segment length, specifically representing the actual physical length of the i-th pipe segment; The diameter is the diameter of the i-th segment of the pipe, which reflects the water conveyance capacity. The current pressure is represented by the real-time water pressure monitoring value of the i-th segment of the pipeline. This is the pressure threshold, specifically representing the maximum safe pressure that the pipeline can withstand.

[0085] The water pollution source tracing path uses the backpropagation algorithm to locate the pollution source, and its calculation model is as follows: Where: (x, y) are the coordinates of the pollution source, specifically the coordinates of the source location of the pollutant diffusion; For concentration gradient, specifically the rate of change of pollutant concentration field in space, pointing in the direction of the fastest concentration increase; It is the partial derivative of time with respect to space, specifically the rate at which time changes with spatial location during the diffusion of pollutants; t represents the pollutant concentration field, specifically the pollutant concentration value at coordinates (x, y); t represents the diffusion time, specifically the time it takes for the pollutant to diffuse from the source to the detection point.

[0086] The method for drawing a single water resources map is as follows:

[0087] A virtual node is generated for each physical monitoring device with multi-source heterogeneous data, and the node is bound to the device ID and GPS coordinates;

[0088] Specifically, a monitoring point encapsulation engine is used to encapsulate virtual nodes;

[0089] For example, a pressure monitoring node declares a data slot `pressure_value` and automatically subscribes to the MQTT topic ` / sensor / pressure / {device_id}`.

[0090] Flow rate monitoring node: Declare data slot flow_speed and automatically connect to Kafka topic water_flow;

[0091] Metadata injection unit: Adds the field {device_id: "P-982", gps: [121.34,31.19]} to each data packet. Automatically associates each monitoring point with its corresponding coordinates on the GIS map to achieve spatial linkage; specifically, a spatial linkage gateway is used to achieve spatial linkage, which includes:

[0092] Coordinate Matcher: Dynamically binds the device's GPS coordinates to map tile coordinates (error ≤ 3 meters);

[0093] Data Pipeline Generator: Automatically creates data channels to all monitoring points when users drag and drop pressure nodes;

[0094] Heatmap renderer: Unclicked areas display average pressure / flow rate in gradient colors (blue = normal, red = abnormal).

[0095] When a user clicks on a location on the map, the system automatically retrieves monitoring nodes within a 100-meter radius and displays pressure and flow rate data in real time.

[0096] Execute when user clicks on a map location:

[0097] Real-time retrieval: Based on R tree spatial index, it returns all monitoring devices within a 100-meter radius within 50ms;

[0098] Dynamic Information Panel: Overlays dashboard controls at clicked locations, displaying pressure values ​​(numerical + dashboard) and flow rate values ​​(curve graph) in parallel.

[0099] Abnormal warning: When the pressure is <1.2MPa and the flow rate is >2m³ / s, a flashing animation will be triggered and an alarm sound will be played.

[0100] Distributed storage with more than 100 nodes processes data streams from 100,000+ monitoring points with a latency of ≤50ms;

[0101] Distributed memory processing monitoring data streams include:

[0102] Data stream partitioning: Distributed storage based on grid blocks (1km × 1km) of device GPS coordinates;

[0103] Real-time calculation: For monitoring points within 100 meters of the clicked location, calculate the standard deviation of pressure / flow rate (if >0.5, a yellow warning will be issued);

[0104] Memory release: Data in inactive areas is automatically unloaded after 30 seconds, releasing ≥40% of memory resources.

[0105] In this embodiment, the scale of the distributed memory processing monitoring data stream is expanded to more than 100 nodes, which can meet the high-concurrency and high-throughput writing from the data source, and the data writing volume of a single node can reach more than 500M / S.

[0106] The creation of a unified water resources map encapsulates each monitoring data point within a node. Users can configure the map in a WYSIWYG manner by clicking on a location, requiring no programming skills. Simply click on a location to view all monitoring data for that location, or drag and drop nodes to view data, or drag nodes to the work area to generate heatmaps. Heatmaps provide users with a visual understanding of the entire water resources system, facilitating monitoring, scheduling, and coordination of key elements. Furthermore, the map supports decision-makers in spatial analysis and simulation, optimizing water supply network layout, improving water use efficiency, and protecting water sources, thereby achieving sustainable management and rational utilization of water resources. The unified water resources map connects and integrates various sensors, instruments, and data sources, such as water level, water quality, and water supply network pressure. This allows users to monitor and track water resources status on a centralized platform.

[0107] S4: The smart screen dynamically controls and integrates with the open platform. It generates decision commands through the visualization engine and performs the following operations: triggering pressurized pump station control commands and pushing early warning information to third-party systems through the open platform API.

[0108] Example 2:

[0109] A smart water supply decision generation system includes a data acquisition and exchange layer, a data middleware layer, an intelligent analysis layer, and a visualization and control layer.

[0110] The data acquisition and exchange layer includes:

[0111] Multi-source adaptation module: Supports data acquisition from relational databases MySQL / Oracle, time-series databases InfluxDB, GIS spatial databases PostGIS, and API interfaces. It also has a built-in ETL workflow engine to achieve data cleaning, splitting, and merging.

[0112] Real-time exchange module: Utilizes Apache Kafka to build a distributed message queue, supporting a throughput of millions of data points per second and providing full / incremental data synchronization strategies;

[0113] Disaster recovery module: Based on database log replication technology (such as MySQLBinlog), it achieves millisecond-level synchronization between primary and standby databases and supports off-site disaster recovery; Database high availability disaster recovery can realize real-time synchronization of table data in the primary database to the off-site disaster recovery center through the platform's built-in database log collection module.

[0114] The data platform layer includes:

[0115] The thematic data warehouse is divided into data themes based on water supply security, pipeline losses, and water pollution. The storage structure includes fact tables (pressure values, flow values, and water quality parameters) and dimension tables (time dimension and spatial dimension).

[0116] Data governance engines, which include:

[0117] Metadata management: Tracking data sources and transformation processes through lineage analysis;

[0118] Quality monitoring: Define quality rules (such as non-empty checks, range checks, and consistency checks);

[0119] Security Management: Implement field-level data access control based on the RBAC model;

[0120] The intelligent analysis layer includes:

[0121] The predictive model library includes water demand forecasting based on the ARIMA time series model, water quality anomaly detection using the isolated forest algorithm, and pipeline life assessment using the Weibull distribution model.

[0122] The real-time computing engine uses Flink to implement streaming computing and calculates the pipeline pressure balance index.

[0123] The visualization and control layer includes:

[0124] The Water Affairs One Map module integrates the OpenLayers geographic engine and supports 3D pipeline rendering and cross-sectional analysis, heat map overlay (pressure distribution, water pollution), and buffer zone analysis (pollution source impact range calculation).

[0125] The smart screen module has pre-set templates for emergency command, daily monitoring, and leadership dashboards, and supports multi-chart linkage (clicking on a map area will display trend charts and statistical tables in a linked manner).

[0126] The open platform module provides RESTful APIs and SDKs, supporting third-party systems to call data services and alert interfaces.

[0127] Example 3:

[0128] A smart water supply decision generation system, based on Embodiment 2, also provides target planning services, as detailed below:

[0129] ①Utilize integrated reporting software to provide the data reports required by various business systems, and display these reports through a visualization and control layer;

[0130] ②Utilize the data middle platform layer to integrate data scattered across various business systems based on the divided data themes, providing full-process analytical data and early warning monitoring models to achieve cross-system thematic data analysis and early warning;

[0131] ③ The system also includes a decision support layer, which analyzes the core operational indicators of the enterprise based on the predictions and calculations of the intelligent analysis layer, provides effective data support for decision support, and generates decision-making solutions.

[0132] The intelligent water supply decision generation system, based on the data acquisition and exchange layer, enables the acquisition of data from various databases and data structures; it has pluggable interfaces for template-based data access; the data acquisition and exchange layer, through the processing of structured and unstructured data, enables efficient transmission and centralized management of data from data warehouses, big data platforms, and various business systems according to a unified transmission and exchange strategy.

[0133] The multi-source adaptation module of the data acquisition and exchange layer adapts to multiple data sources, enabling it to acquire data from different data sources such as structured data, semi-structured data, and unstructured data. It can also achieve remote data acquisition and transmission across networks to the target source database.

[0134] The real-time exchange module of the data acquisition and exchange layer enables real-time data acquisition, loading, and external exchange, supports real-time batch data exchange between systems, and improves the timeliness of data analysis and use.

[0135] The data acquisition and exchange layer, based on the multi-source adaptation module, performs data cleaning to verify the quality of key business data and meet transcoding requirements; it also performs data splitting and merging to achieve incremental identification of large batches of data files and database data, and full-volume merging and push functions, thereby improving data exchange efficiency; and it performs data quality verification to verify the quality of loaded data files or database data, including but not limited to the accuracy of data format, verification of non-empty data tables, and identification of abnormal data.

[0136] The data platform layer adopts a distributed memory storage structure with a built-in distributed cache cluster. The cluster size can be expanded to more than 100 nodes. It can perform real-time data cleaning and processing on the incoming data stream. The cluster size can be expanded according to the amount of data to be processed. It can meet the high-concurrency and high-throughput writing from the data source, and the single-node write data volume can reach more than 500M / S.

[0137] The data acquisition and exchange layer includes multiple data disaster recovery methods.

[0138] Supports multiple backup and restore methods

[0139] It offers backup methods such as full backup, incremental backup, and log backup, and provides multiple recovery methods such as point-in-time recovery and automatic disaster recovery.

[0140] In practical applications, the platform needs to support diverse business applications. To ensure the platform's flexibility and scalability, and to handle different business data, the data processing models and algorithms need to be separated to adapt to varying business requirements. In specific implementation, this is done based on the defined goals of big data processing and the selection of appropriate data processing models.

[0141] Specifically, the platform manages various data analysis models, loads sample data, creates and schedules tasks, and generates intermediate or final results, which are then provided to different application systems or users for access and querying. The system will be encapsulated using internationally standardized enterprise-level service interfaces to meet diverse needs. Through a visual workflow, the platform provides visual monitoring of the MR (Mapping Reference) execution status of each analysis model and enables the evaluation and optimization of these models.

[0142] In summary, the main methods of data governance in this invention are as follows:

[0143] Unified data standards: Standardized and unified management of data classification, definitions, models, and other rules.

[0144] Metadata management: Starting with the establishment of enterprise-level data models and indicator systems, business classifications, business rules, and data cubes are incorporated into metadata management.

[0145] Data quality management: Establish a cross-disciplinary, end-to-end data quality management system to ensure the accuracy, standardization, completeness, and consistency of data information.

[0146] Data lifecycle management: Enables multi-level management of the data lifecycle and rationally allocates data usage frequency and resource consumption.

[0147] Data security management: Comprehensive security management of data assets, transmission, environment, access control, and personnel permissions throughout the entire data management process.

[0148] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for generating intelligent water supply decisions, characterized in that, Includes the following steps: S1: Real-time collection of water supply network pressure data, water quality monitoring data, user water consumption data and GIS spatial data, and execution of data quality verification and standardization processing; S2: Establish an analysis system that includes a water supply safety assessment model, a pipeline leakage detection model, and a water pollution prediction model. The output indicators of each model are weighted and fused to generate a comprehensive assessment result. S3: Spatially overlay real-time monitoring data with GIS maps to create a single water affairs map, and generate pipeline pressure boosting paths and water pollution source tracing paths through topological relationship analysis; S4: Dynamically generate control commands based on visualization analysis results, including pump station pressure regulation commands and cross-system early warning information push.

2. The intelligent water supply decision generation method according to claim 1, characterized in that, The data quality verification employs a multi-dimensional evaluation formula: ； Where: Q is the overall data quality score, which is controlled by a threshold Q0. Data with a quality score lower than Q0 is considered abnormal; n is the number of data sources. V represents the weight of the i-th data category; V represents the number of valid data entries; T represents the total data volume. C is the consistency factor; C is the cross-system data consistency score.

3. The intelligent water supply decision generation method according to claim 1, characterized in that, The water supply safety assessment model employs a dynamic equilibrium algorithm. ； in: This is a stress anomaly index; This refers to the real-time water supply volume. To predict water demand using an LSTM network; The variance of the water supply; is the temperature influence coefficient; T is the water temperature in the pipe network.

4. The intelligent water supply decision generation method according to claim 1, characterized in that, The pipeline leakage detection model is as follows: ； Where: L is the leakage index, which quantifies the severity of leakage in the pipe network within a specific time period; These are the pressures at the user's pipeline inlet and outlet, respectively. t1 represents the attenuation coefficient of the pipeline material; t1 and t2 represent the start and end times of the leakage detection.

5. The intelligent water supply decision generation method according to claim 1, characterized in that, A comprehensive water resources map is created by spatially overlaying real-time monitoring data with a GIS map. Topological relationship analysis is used to generate pipeline pressure boosting paths and water pollution source tracing paths. This comprehensive water resources map includes the following elements: Basic geographical layer: includes administrative divisions and water system distribution; Facilities layer: Includes the topological relationships of water plants, pumping stations, and pipe networks; Monitoring layer: Integrates real-time data from pressure, flow, and water quality monitoring points; User layer: Mark the water usage characteristics of key users; Protection zone layer: Delineate water source protection areas and ecological red lines.

6. The intelligent water supply decision generation method according to claim 5, characterized in that, In step S3, the spatial topology analysis includes pipeline pressure boosting paths and water pollution source tracing.

7. The intelligent water supply decision generation method according to claim 6, characterized in that, The pipeline pressure boosting path is generated based on the Dijkstra algorithm, and its calculation model is as follows: ； Where: Cost is the comprehensive cost of the route, which represents the comprehensive cost of pressurization and regulation of the pipeline route. The smaller the value, the better the route. is the pipe segment length, specifically representing the actual physical length of the i-th pipe segment; The diameter is the diameter of the i-th segment of the pipe, which reflects the water conveyance capacity. The current pressure is represented by the real-time water pressure monitoring value of the i-th segment of the pipeline. This is the pressure threshold, specifically representing the maximum safe pressure that the pipeline can withstand.

8. The intelligent water supply decision generation method according to claim 6, characterized in that, The water pollution source tracing path uses the backpropagation algorithm to locate the pollution source. Its calculation model is as follows: ; Where: S(x, y) are the coordinates of the pollution source, specifically the coordinates of the source location of the pollutant diffusion; For concentration gradient, specifically the rate of change of pollutant concentration field in space, pointing in the direction of the fastest concentration increase; It is the partial derivative of time with respect to space, specifically the rate at which time changes with spatial location during the diffusion of pollutants; t represents the pollutant concentration field, specifically the pollutant concentration value at coordinates (x, y); t represents the diffusion time, specifically the time it takes for the pollutant to diffuse from the source to the detection point.

9. A smart water supply decision generation system, which is a system for implementing the smart water supply decision generation method according to any one of claims 1-8, characterized in that, It includes a data acquisition and exchange layer, a data middle platform layer, an intelligent analysis layer, and a visualization and control layer.

10. The intelligent water supply decision generation system according to claim 9, characterized in that, The data acquisition and exchange layer includes: Multi-source adaptation module: Supports data acquisition from relational databases MySQL / Oracle, time-series databases InfluxDB, GIS spatial databases PostGIS, and API interfaces. It also has a built-in ETL workflow engine to achieve data cleaning, splitting, and merging. Real-time exchange module: Utilizes Apache Kafka to build a distributed message queue, supporting a throughput of millions of data points per second and providing full / incremental data synchronization strategies; Disaster recovery module: Based on database log replication technology, it achieves millisecond-level synchronization between primary and standby databases and supports disaster recovery in different locations; The data middle platform layer includes: Thematic data warehouses are divided into data themes based on water supply security, pipeline losses, and water pollution. The storage structure includes fact tables and dimension tables. Data governance engines, which include: Metadata management: Tracking data sources and transformation processes through lineage analysis; Quality monitoring: defining quality rules; Security Management: Implement field-level data access control based on the RBAC model; The intelligent analysis layer includes: a prediction model library, which includes water demand prediction based on the ARIMA time series model, water quality anomaly detection using the isolated forest algorithm, and pipeline life assessment using the Weibull distribution model. The real-time computing engine uses Flink to implement streaming computing and calculates the pipeline pressure balance index. The visualization and control layer includes: a water affairs map module, which integrates the OpenLayers geographic engine and supports 3D pipeline rendering and cross-sectional analysis, heat map overlay, and buffer analysis functions; The smart screen module has pre-set templates for emergency command, daily monitoring, and leadership dashboards, and supports multi-chart linkage. The open platform module provides RESTful APIs and SDKs, supporting third-party systems to call data services and alert interfaces.

Images