Intelligent water conservancy monitoring method and system based on Internet of Things

Abstract

The invention provides an intelligent water conservancy monitoring method and system based on the Internet of Things. The method comprises the steps of obtaining heterogeneous water conservancy data of various sources; preprocessing and standardizing the heterogeneous water conservancy data from various sources; constructing a water conservancy domain knowledge graph, and performing multi-dimensional data fusion to generate comprehensive water conservancy situation data; based on the comprehensive water conservancy situation data, constructing a water conservancy digital twinborn body, and realizing real-time driving and hydrological and hydraulic process simulation of the water conservancy digital twinborn body; and multi-modal deep analysis is performed on the comprehensive water conservancy situation data and the water conservancy digital twin, and interpretable intelligent decision support information is generated by using an artificial intelligence technology. According to the method, the problems of scattered data sources, high isomerism, relatively low intelligent degree and insufficient information presentation modes of a water conservancy management system are solved by acquiring and preprocessing the multi-source heterogeneous water conservancy data, fusing and cognizing the water conservancy data, twinning and simulating the water conservancy system and performing analysis and decision intelligent support.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of artificial intelligence and water conservancy, and specifically relates to a smart water conservancy monitoring method and system based on the Internet of Things. BACKGROUND

[0002] As the foundation support for national economic and social development, the importance of water conservancy is self-evident. Traditional water conservancy monitoring and management methods face many challenges in dealing with increasingly complex water regime changes, severe water supply and demand contradictions, and frequent drought and flood disasters. At present, although the informatization construction of the water conservancy industry has made some progress, there are still many deficiencies in data acquisition, information integration, intelligent analysis and decision support.

[0003] Specifically, the existing water conservancy monitoring system often has the problems of scattered data sources and strong heterogeneity. There is a lack of unified interface standards and data protocols between different types of sensors, monitoring equipment and historical data platforms, which leads to difficulties in data integration and makes it difficult to form a comprehensive water situation awareness. These scattered data are difficult to effectively align in time and space and deeply integrate, so that the data value cannot be fully tapped, which in turn restricts the macroscopic grasp and fine management of the overall operation of the water conservancy system.

[0004] In addition, the traditional monitoring means have relatively low intelligence in terms of prediction and early warning, anomaly identification and auxiliary decision-making. Most systems still stay at the stage of simple data display and threshold alarm, lacking deep mining of complex hydrological laws and intelligent judgment ability for sudden events. For example, for complex hydrodynamics and water environment processes such as flood evolution and water pollution diffusion, the existing models and analysis methods often have limited accuracy and delayed response, making it difficult to provide timely and accurate prediction results. At the same time, the decision-making process still highly depends on manual experience, lacking scientific and quantitative intelligent recommendation mechanism, especially in dealing with extreme weather events and sudden water conservancy hazards, the decision-making efficiency and accuracy are difficult to be effectively guaranteed.

[0005] Furthermore, the existing water conservancy management system also has room for improvement in terms of visualization and human-computer interaction. Although some systems can provide chart and simple two-dimensional map display, they lack high-fidelity three-dimensional modeling and real-time digital twin capabilities for physical water conservancy facilities. This makes it difficult for users to intuitively understand the dynamic changes and complex spatial relationships of the water conservancy system, and also limits the possibility of scenario deduction and simulation analysis in a virtual environment. The shortcomings of this information presentation method directly affect the cognitive efficiency and decision-making efficiency of water conservancy management personnel on complex water regime and working conditions. SUMMARY

[0006] Therefore, the application provides a smart water conservancy monitoring method and system based on Internet of Things, which solves the problems of scattered data sources, strong heterogeneity, relatively low intelligence and insufficient information presentation mode of the water conservancy management system by acquiring and preprocessing multi-source heterogeneous water conservancy data, fusing and cognizing water conservancy data, twinning and simulating water conservancy system, and analyzing and decision-making intelligent support.

[0007] The first aspect of the application provides a smart water conservancy monitoring method based on Internet of Things, which comprises steps S1 to S5.

[0008] Step S1, acquiring heterogeneous water conservancy data from multiple sources of Internet of Things sensors, remote sensing devices, video monitoring devices and weather stations.

[0009] Step S2, preprocessing and standardizing the heterogeneous water conservancy data from multiple sources; Step S3, based on the preprocessed and standardized heterogeneous water conservancy data from multiple sources, constructing a water conservancy knowledge graph, and performing multi-dimensional data fusion to generate comprehensive water conservancy situation data; Step S4, based on the comprehensive water conservancy situation data, constructing a water conservancy digital twin, and realizing real-time driving of the water conservancy digital twin and hydrological and hydraulic process simulation; Step S5, multi-modal deep analysis of the comprehensive water conservancy situation data and the water conservancy digital twin, and generating interpretable intelligent decision support information using artificial intelligence technology.

[0010] In one specific embodiment of the application, step S2 includes protocol adaptation, format conversion, data denoising, missing value filling and outlier detection of the heterogeneous water conservancy data from multiple sources.

[0011] In one specific embodiment of the application, the multi-dimensional data fusion in step S3 includes: based on the water conservancy knowledge graph, the standardized heterogeneous water conservancy data from multiple sources is semantically associated with the pre-defined water conservancy entities in the knowledge graph to give the data context background information; for the same monitoring indicators from multiple data sources, the reliability weights of each data source are dynamically determined according to the historical data performance of each data source, and the values of the monitoring indicators are weighted and fused to generate a more highly confident comprehensive indicator value.

[0012] In one specific embodiment of the application, the hydrological and hydraulic process simulation in step S4 includes: embedding a hydrodynamic or water quality transport model in the water conservancy digital twin, and using the comprehensive water conservancy situation data as the real-time boundary conditions of the model to dynamically simulate and predict the processes of flood evolution, reservoir storage and discharge or pollutant diffusion.

[0013] In an embodiment of the present application, the artificial intelligence technology used in step S5 includes: using a time series prediction model to predict water level and flow; and / or using an image recognition model to analyze video monitoring data to identify dam safety hazards or unusual floating objects on the water surface.

[0014] In an embodiment of the present application, the smart water conservancy monitoring method based on the Internet of Things further includes step S6.

[0015] Step S6, according to the intelligent decision support information, performing automatic inspection or intelligent emergency response to cope with water conservancy abnormal events.

[0016] In an embodiment of the present application, the automatic inspection performed in step S6 includes: planning the inspection path and task of a drone or an underwater robot based on the intelligent decision support information, and remotely controlling it to perform autonomous inspection of water conservancy engineering facilities. The intelligent emergency response performed in step S6 includes: when a risk of flood or water pollution is predicted, automatically generating and executing optimal control instructions for reservoirs, gates or pumping stations based on the intelligent decision support information.

[0017] In an embodiment of the present application, the smart water conservancy monitoring method based on the Internet of Things further includes step S7.

[0018] Step S7, energy management and network security protection are performed on the method of step S1 to step S6, and based on scenario simulation and risk assessment, system resilience is improved, and system resilience specifically includes: Simulate multiple failure scenarios of extreme weather or equipment failure in the digital twin, evaluate the response capability and weak links of the system, and optimize the emergency plan and resource allocation strategy according to the evaluation results.

[0019] In an embodiment of the present application, simulating multiple failure scenarios of extreme weather or equipment failure in the digital twin, evaluating the response capability and weak links of the system, and optimizing the emergency plan and resource allocation strategy according to the evaluation results, includes: Use the water conservancy digital twin as a virtual test field to actively simulate one or more pre-set extreme working conditions or key equipment failure scenarios for system-level stress testing; Under the simulated scenario, quantitatively evaluate the performance of the whole link from data perception to decision response to accurately identify its potential weak links; According to the evaluation results, the artificial intelligence model in the analysis and decision support module, or the emergency response strategy in the intelligent execution and response module is iteratively optimized and reinforced.

[0020] The second aspect of the application provides a smart water conservancy monitoring system based on the Internet of Things, which comprises an Internet of Things sensing module, a data processing module, a data fusion and knowledge graph module, a digital twin module, a deep analysis and decision-making module, and an automatic inspection and emergency response module.

[0021] The Internet of Things sensing module is configured to acquire heterogeneous water conservancy data from various sources of Internet of Things sensors, remote sensing devices, video monitoring devices and weather stations, and to perform preprocessing and standardization operations on the heterogeneous water conservancy data from various sources.

[0022] The data processing module is configured to preprocess and standardize the heterogeneous water conservancy data from various sources.

[0023] The data fusion and knowledge graph module is configured to construct a water conservancy knowledge graph based on the preprocessed and standardized data, and to perform multi-dimensional data fusion to generate comprehensive water conservancy situation data.

[0024] The digital twin module is configured to construct a water conservancy digital twin based on the comprehensive water conservancy situation data, and to realize real-time driving of the water conservancy digital twin and simulation of hydrological and hydraulic processes.

[0025] The deep analysis and decision-making module is configured to perform multi-modal deep analysis on the comprehensive water conservancy situation data and the water conservancy digital twin, and to generate interpretable intelligent decision support information using artificial intelligence technology.

[0026] The automatic inspection and emergency response module is configured to perform automatic inspection and / or intelligent emergency response according to the intelligent decision support information to deal with water conservancy abnormal events.

[0027] The third aspect of the application provides a computer device comprising a processor and a memory. The processor is configured to execute the smart water conservancy monitoring method based on the Internet of Things of the first aspect of the application. The memory is configured to store executable instructions of the processor.

[0028] The fourth aspect of the application provides a computer readable storage medium having computer executable instructions stored thereon. The executable instructions are executed by the processor to implement the smart water conservancy monitoring method based on the Internet of Things of the first aspect of the application.

[0029] The fifth aspect of the application provides a computer program product comprising computer programs / instructions, which are executed by the processor to implement the smart water conservancy monitoring method based on the Internet of Things of the first aspect of the application.

[0030] The beneficial effects of the technical scheme of the present application are: in order to realize comprehensive and stereoscopic perception of the state of the water conservancy system, firstly, a variety of sources of heterogeneous water conservancy data are widely obtained, which overcomes the limitations of traditional monitoring methods relying on a single data source and information being one-sided. In view of the diversity of the above-mentioned data sources leading to significant differences in communication protocols, data formats, timestamp accuracy and data structure, that is, "heterogeneity", a series of preprocessing and standardization operations are performed on the heterogeneous water conservancy data from multiple sources to eliminate data barriers and realize the availability and credibility of the data. By performing knowledge graph construction and multi-dimensional data fusion, the knowledge graph aims to formally describe various entities, concepts and their complex relationships in the water conservancy system in a structured manner, which overcomes the defect that traditional databases can only store data but cannot effectively express the logic and association behind the data, and also lays a solid and reliable data and cognitive foundation for step S4 to build a high-fidelity water conservancy digital twin that can truly reflect the operation rules of the physical world. In order to ensure the synchronization of virtual and reality, the integrated water conservancy situation data fused in step S3 is used to drive the water conservancy digital twin in real time, and through this real-time driving mechanism, the water conservancy digital twin becomes a dynamic virtual copy that coexists and evolves synchronously with the physical water conservancy system. In order to realize deeper causal analysis, the integrated water conservancy situation data output by step S3 and the water conservancy digital twin and its simulation results output by step S4 are connected, aiming to convert massive data and complex simulation information into intuitive, interpretable and guiding decision support information through multi-modal deep analysis, so as to provide forward-looking and scientific decision-making basis for water conservancy management. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 Fig. 1 shows a flowchart of a smart water conservancy monitoring method based on the Internet of Things according to an embodiment of the present application.

[0032] Figure 2 Fig. 2 shows a structural diagram of a smart water conservancy monitoring system based on the Internet of Things according to an embodiment of the present application. DETAILED DESCRIPTION

[0033] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0034] Please refer to the drawings Figure 1In at least one embodiment of the present application, a smart water conservancy monitoring method based on Internet of Things is provided, which takes a smart water conservancy monitoring system based on Internet of Things (which can be referred to as system for short) provided by an embodiment of the present application as an execution subject. The smart water conservancy monitoring method based on Internet of Things includes the following steps S1 to S6.

[0035] Step S1, acquiring heterogeneous water conservancy data from multiple sources of Internet of Things sensors, remote sensing devices, video monitoring devices and weather stations.

[0036] Step S2, pre-processing and standardizing the heterogeneous water conservancy data from multiple sources.

[0037] In this embodiment, steps S1 and S2 are to acquire and pre-process heterogeneous water conservancy data from multiple sources, and the core purpose is to provide high-quality, standardized data basis for subsequent deep fusion, twin modeling and intelligent analysis. Steps S1 and S2 are the starting point of the whole smart water conservancy monitoring and management method, and the execution quality directly affects the accuracy and reliability of all subsequent steps.

[0038] In this embodiment, in order to realize comprehensive and three-dimensional perception of the state of the water conservancy system, the system first widely acquires heterogeneous water conservancy data from multiple sources. This overcomes the limitations of traditional monitoring methods that rely on a single data source and one-sided information. Specifically, the data sources cover multiple dimensions: through the deployment of sensing devices such as Internet of Things sensors at key locations such as reservoirs, rivers, and gate stations, high-frequency time series data such as water level, flow rate, flow speed, and water quality parameters are acquired in real time; by calling remote sensing devices carried by satellites or drones, macro spatial information such as large-area water area, land cover, and soil moisture is acquired; through video monitoring devices distributed in key areas, visual image data such as water surface floating objects, dam conditions, and water activities is acquired. In addition, the system also accesses meteorological data such as rainfall, wind speed, and air temperature provided by weather stations, and integrates engineering design data, hydrological records from historical archives, and hidden danger information reported through manual patrol.

[0039] In this embodiment, due to the diversity of the above-mentioned data sources, there are significant differences in communication protocols, data formats, timestamp accuracy, and data structures, i.e. heterogeneity, so a series of preprocessing and standardization operations must be performed on the original data to eliminate data barriers and achieve data usability and credibility.

[0040] Preferably, the system first performs protocol adaptation and format conversion. For various transmission protocols such as MQTT, CoAP, HTTP adopted by different devices and platforms, the system configures corresponding adapters or gateways to realize stable access to various data. After access, the system uniformly parses and converts data of different sources such as JSON, XML, binary code stream, and text file into an internal standardized structured data format. The standardized format includes a unified timestamp, a unique device identifier, a geographic spatial coordinate, and a standardized monitoring indicator field, laying a foundation for subsequent unified processing.

[0041] In this embodiment, to improve the quality of the original monitoring data, the system then performs a series of data cleaning operations.

[0042] For random noise or high-frequency disturbance that may be introduced by the sensor in the collection or data in the transmission process, preferably, the system uses a filtering algorithm for data denoising. For example, moving average filtering, exponential smoothing, or wavelet transform-based filtering methods can be used to smooth the time series data, while retaining the true trend of the data.

[0043] For the problem of missing monitoring data caused by network interruption, device offline, or storage error, the system performs missing value filling. Preferably, different interpolation strategies are used according to the pattern and length of data missing. For single or short-term missing values, linear interpolation, polynomial interpolation, or spline interpolation using data correlation at the front and rear time points can be used for estimation and filling. For long-term continuous data missing, the system will mark the data segment as low confidence to avoid false filling misleading subsequent analysis.

[0044] In this embodiment, to identify and eliminate obviously erroneous abnormal data points, the system also performs outlier detection. These outliers may be caused by sensor transient failure or extreme environmental interference, and their values are far from the normal physical change range. Preferably, the system uses statistical methods such as setting thresholds using the standard deviation or interquartile range of the data to identify isolated outliers. In addition, the system can also combine a rule base based on domain knowledge, for example, to determine whether the water level change rate exceeds the maximum value physically possible, to identify and process abnormal data that do not conform to the water conservancy operation rules.

[0045] By performing the above acquisition and preprocessing steps, the present application can convert the chaotic, original, and heterogeneous water conservancy data of various sources into a clean, complete, and unified format and reliable standardized data set, providing a solid guarantee for the deep data fusion and knowledge graph construction of step S3.

[0046] In one specific implementation of the present application, step S3 and step S4 jointly constitute a data cognition and twin stage. The core goal of this stage is to first realize deep understanding and semantic association of data through knowledge graph technology (i.e., data cognition), and then build a dynamic virtual copy (i.e., twin) that accurately maps and synchronously runs with the physical world on this basis.

[0047] Step S3, based on pre-processing and standardization of multi-source heterogeneous water conservancy data, constructs a water conservancy field knowledge graph, and performs multi-dimensional data fusion to generate comprehensive water conservancy situation data.

[0048] In at least one embodiment of the present application, multi-dimensional data fusion specifically includes: based on the water conservancy field knowledge graph, the standardized multi-source heterogeneous water conservancy data is semantically associated with the pre-defined water conservancy entities in the knowledge graph to give data context background information; for the same monitoring index from multiple data sources, the reliability weight of each data source is dynamically determined according to the historical data performance of each data source, and the numerical value of the monitoring index is weighted and fused to generate a more highly confident comprehensive index value.

[0049] In this embodiment, step S3 is to fuse and cognize water conservancy data. This step takes the standardized data set output by step S2, and its core purpose is to extract comprehensive water conservancy information with deep semantics and internal relations from massive and isolated data points, realize the transition from data to cognition, and provide a high-quality cognitive basis for subsequent twin simulation and intelligent decision-making.

[0050] In one specific implementation of the present application, the execution of this step is not simply data superposition, but through the construction of a water conservancy field knowledge graph and the implementation of a multi-dimensional data fusion strategy, the depth understanding and characterization of the water conservancy system state are realized.

[0051] First, the system starts to construct a water conservancy field knowledge graph. This knowledge graph aims to formally describe various entities, concepts and their complex relationships in the water conservancy system in a structured manner. This overcomes the defect that traditional databases can only store data but cannot effectively express the implicit logic and association behind the data.

[0052] Specifically, the construction of the knowledge graph includes entity definition, relation extraction, and attribute annotation. Entities cover various water conservancy elements in the physical world, such as reservoirs, dams, gates, pumping stations, river sections, monitoring stations, and sensor devices. They also include management and conceptual entities, such as administrative divisions, management units, and emergency plans. Relations define the interactions and logical connections between these entities, such as a sensor deployed at a dam, a gate controlling the flow of a river section, and the flood discharge of an upstream reservoir affecting the water level of a downstream river. Properties are used to describe the specific characteristics of entities, such as the total storage capacity of a reservoir, the design standard of a dam, and the measurement type of a sensor. The sources of this knowledge include, but are not limited to, engineering design drawings, historical hydrological data, industry standards and specifications, and prior knowledge of water conservancy experts. Through this knowledge graph, discrete data points are given a clear semantic context.

[0053] In this embodiment, based on the constructed semantic framework of the water conservancy knowledge graph, the system further performs multi-dimensional data fusion to generate comprehensive water conservancy situation data with rich semantics and strong correlation. The fusion here is multi-level.

[0054] Preferably, the system first performs semantic fusion based on the knowledge graph. It associates and maps the standardized data after preprocessing in step S1 with the corresponding entities in the knowledge graph. For example, a water level data point with geographic coordinates is no longer just a numerical value, but is explicitly identified as the water level at the center of a certain reservoir area, and is logically connected with the reservoir's capacity, historical water level, and associated flood discharge gates. Similarly, a video stream that identifies cracks in a dam will be automatically associated with the corresponding dam entity in the knowledge graph and its design data and historical inspection records. This way, data can utilize its background knowledge when fused, greatly enhancing the dimension and depth of information.

[0055] In addition, to improve the accuracy and robustness of key monitoring indicators, the application also adopts a dynamic weighting data fusion method. In actual scenarios, multiple different sensors or data sources may monitor the same physical quantity. At this time, simply taking the average value may lead to distorted results due to errors from some devices. Therefore, the system introduces a dynamic reliability evaluation mechanism for each data source.

[0056] The mechanism continuously analyzes the historical performance of each data source, such as the stability of the data, the degree of deviation from the group mean, the continuity of data reporting, etc., and dynamically assigns a weight to each data source accordingly. A stable and high-quality data source will have a higher weight; otherwise, a source that frequently appears abnormal or deviates recently will be given a lower weight. In each round of data fusion calculation, the final comprehensive monitoring value is obtained by weighted sum of the data of all related sources. This adaptive weighting strategy ensures that the fused result can approximate the true physical value to the maximum extent and effectively suppresses the negative impact of individual data source failure or interference.

[0057] By performing the above knowledge graph construction and multi-dimensional data fusion, the final output of this step is no longer a scattered set of data points, but structured, semantic, and high-confidence comprehensive water situation data. These data not only contain accurate monitoring values, but also contain the logical relationship between water elements, which lays a solid and reliable data and cognitive foundation for step S4 to build a high-fidelity water digital twin that truly reflects the operating rules of the physical world.

[0058] Step S4, based on the comprehensive water situation data, constructs a water digital twin and realizes real-time driving of the water digital twin and hydrological and hydraulic process simulation.

[0059] In this embodiment, step S4 is a twin and simulation water system, which is based on the comprehensive water situation data generated in step S3, aims to construct a virtual mapping body highly corresponding to the physical water system, i.e. water digital twin, and endows it with dynamic simulation and deduction ability. This step is the key bridge connecting comprehensive perception and intelligent decision-making, and its core lies in realizing the transition from static cognition to dynamic simulation.

[0060] In one specific implementation of the present application, the construction of the water digital twin is a multi-level and multi-dimensional process, aiming to realize accurate reproduction of physical entities. First, the system performs high-fidelity geometric modeling. Preferably, this process comprehensively utilizes building information model (BIM) data, geographic information system (GIS) data, laser point cloud scanning data, and high-resolution remote sensing images to finely reconstruct the three-dimensional geometry of reservoir dams, river embankments, gate stations, and other water conservancy engineering facilities and their surrounding topography. This ensures that the water digital twin is consistent with the physical world in terms of spatial form, size, and position.

[0061] In this embodiment, only the geometric twin is not enough to reflect the system dynamics. Therefore, the system further embeds physical properties, operating mechanisms and management rules into the geometric model. Specifically, the system gives each entity in the twin its physical properties, such as the concrete material and strength of the dam, the roughness coefficient of the river channel, the permeability of the soil, etc. At the same time, the system also constructs a mechanism model of the facility to describe its operating logic, such as the relationship curve between the opening of the gate and the discharge flow, the functional relationship between the power of the water pump and the pumping efficiency, etc. In addition, the management rules such as the flood season limit water level of the reservoir and flood control scheduling scheme are also integrated into the digital water twin, making it not only similar in shape, but also similar in spirit.

[0062] In this embodiment, to ensure the synchronization of virtual and real, the system uses the integrated water situation data fused in step S3 to drive the digital water twin constructed in real time. This is a continuous data injection and state updating process. The system will obtain the data of water level, flow, rainfall, equipment working condition, etc. in real time as the driving source, and update the state of the corresponding elements in the digital water twin in real time. For example, when the sensor in the physical world detects that the water level of the reservoir rises, the water surface of the reservoir model in the digital water twin will also rise synchronously; when the physical gate is opened, the twin gate model will also perform virtual opening with the same opening degree and rate, and render its water flow changes in real time. Through this real-time driving mechanism, the digital water twin becomes a dynamic virtual copy that coexists and evolves synchronously with the physical water system.

[0063] In this embodiment, the digital water twin is not only used for state display, but also a simulation environment with strong prediction ability. To achieve this purpose, the system deeply integrates professional hydrology and hydraulics models in the digital twin environment.

[0064] Preferably, to simulate the evolution process of flood in the river network, the system embeds a hydrodynamic model based on the basic principles of fluid mechanics. For example, one-dimensional Saint-Venant equations can be used to describe unsteady flow in the main and branch rivers, or two-dimensional shallow water equations can be used to simulate the flood inundation process in a wide area. These models take the fused real-time monitoring data (such as upstream inflow, inter-zone rainfall runoff) as their dynamic boundary conditions and initial conditions for solving, so as to simulate and predict key indicators such as water level, flow velocity and flood arrival time in the future period.

[0065] In addition, to deal with sudden water pollution incidents, the system also integrates a water quality transport and diffusion model. Based on the convection-diffusion equation, this model can couple the calculation results of the hydrodynamic model (i.e. the velocity field and direction of water flow), and combine with the discharge information of the pollution source, to simulate and predict the diffusion path, concentration distribution and spatio-temporal evolution law of pollutants in the water body.

[0066] By integrating and driving these professional models by real-time data in the digital twin of water conservancy, the application realizes the scientific deduction of the future state of the water conservancy system. The final output of this step is a dynamically evolving and predictive digital twin system. The core output is a high-fidelity twin model and simulation results based on the model, which provides intuitive and quantitative scenario data for the deep analysis and intelligent decision-making of step S5, making hypothesis analysis and scenario deduction possible.

[0067] Step S5, multi-modal deep analysis of comprehensive water situation data and water digital twin, and generation of interpretable intelligent decision support information using artificial intelligence technology.

[0068] In this embodiment, step S5 is analysis and decision-making intelligent support. This step is the core link of the application to realize wisdom. It takes the comprehensive water situation data output in step S3 and the water digital twin and its simulation results output in step S4, aiming to transform massive data and complex simulation information into intuitive, interpretable and guiding decision support information through multi-modal deep analysis, so as to provide forward-looking and scientific decision-making basis for water management.

[0069] It should be noted that the interpretable intelligent decision support information is obtained by introducing an explainable analysis algorithm to attribute analyze the output results of the artificial intelligence technology, and presenting the key influence characteristics in a visual form, so as to improve the credibility of the decision-making information.

[0070] In one specific implementation of the application, this step first performs deep analysis on the fused comprehensive water situation data to realize accurate prediction of the future state of the water conservancy system and deep insight into the current state.

[0071] Preferably, for the prediction of key parameters such as water level, flow rate and water quality, the system uses an artificial intelligence model that can capture long-term dependencies in time series data. For example, a deep learning model such as long short-term memory network (LSTM) or its variant gated recurrent unit (GRU) can be used. These models take historical, high-quality time series data after fusion processing as input, learn the complex nonlinear variation law and periodic characteristics of hydrological processes through their internal memory units and gating mechanisms, and then output the predicted values of key parameters in the future period. This prediction capability overcomes the defects of traditional models such as response lag and limited accuracy, and provides valuable advance for flood warning and water resource scheduling.

[0072] Preferably, the system adopts a multi-modal analysis strategy for the identification of water-related abnormal events. For video monitoring data or remote sensing images, an image recognition model such as a convolutional neural network (CNN) can be used. After being trained with a large number of labeled samples, this model can automatically detect abnormal conditions such as dam surface cracks, leaks, large-scale floating objects on the water surface, or illegal buildings within the river management area from real-time video streams or images, achieving all-weather and high-efficiency automated visual patrol.

[0073] In this embodiment, to achieve deeper causal analysis, such as tracing the source of pollution when a water pollution event occurs, the system combines artificial intelligence technology with the water-related knowledge graph constructed in step S3. Preferably, a graph neural network (GNN) model can be used. This model uses knowledge graph information such as water system connectivity, sewage outlet distribution, and upstream and downstream relationships as its reasoning foundation. When an abnormal water quality is detected downstream, the graph neural network can simulate the reverse propagation process of pollutants in the water network structure, combine the characteristics of each node (such as sewage outlet type and enterprise information), calculate and output the contribution probability of each potential pollution source, and thus provide scientific tracing guidance for environmental law enforcement and emergency disposal.

[0074] In this embodiment, one key technical feature of the present application is that the intelligent decision support information generated is interpretable. To avoid the artificial intelligence model becoming a black box that decision makers cannot understand, the system introduces an explainability analysis algorithm. This algorithm can perform attribution analysis on the decision-making process of the model while giving the prediction or identification results, quantifying the contribution of each input feature to the final output result. For example, when issuing a flood risk warning, the system not only gives the warning conclusion, but also shows users through visual forms such as feature importance charts how factors such as upstream water inflow, local rainfall, or downstream backwater level influenced the generation of this warning to what extent. This interpretability significantly enhances users' trust and acceptance of artificial intelligence decision results.

[0075] In this embodiment, the system comprehensively analyzes and judges the results of the above-mentioned analyses and predictions, and combines the simulation and deduction capabilities of the water-related digital twin in step S4 to generate intelligent decision support information. For example, when predicting that a flood may occur in the future, the system will automatically call the water-related digital twin to simulate the flood evolution process and inundation range under different gate scheduling schemes, and recommend one or more optimal scheduling plans that balance flood control safety and benefit maximization to decision makers by comparing the simulation results of different schemes.

[0076] By performing the above-mentioned analysis and decision intelligence support steps, the present application can distill raw data and simulation results into high-value decision intelligence, providing clear, reliable, and credible instructions and basis for the automated execution and intelligent response in step S6.

[0077] Step S6, according to the intelligent decision support information, performing automated inspection or intelligent emergency response to deal with water conservancy abnormal events.

[0078] In this embodiment, step S6 is to execute and respond to intelligent operations, which is the execution end of the monitoring and management closed-loop control realized by the present application. This step directly receives the intelligent decision support information generated in step S5, and its core purpose is to convert the high-level strategy obtained through analysis and judgment into accurate and timely control instructions for physical world water conservancy facilities and inspection equipment, realizing the landing from wisdom decision to intelligent execution.

[0079] In a specific implementation of the present application, the execution of this step embodies two main operation modes: automated inspection and intelligent emergency response. These two modes can work independently or cooperatively to meet the needs of daily management and emergencies, thereby changing the traditional passive response mode relying on manual operation into an active and efficient automated or man-machine cooperative operation mode.

[0080] For automated inspection operation, when the analysis result of step S5 indicates that there is a potential risk in a certain area, or according to the preset inspection plan, the system will automatically trigger the automated inspection process. This process aims to use unmanned equipment to conduct detailed exploration of areas that are difficult for humans to reach or need high-frequency inspection.

[0081] Preferably, this process first generates specific inspection tasks according to the decision information. For example, if step S5 identifies image features of suspected cracks in a certain section of the embankment, the system will automatically plan an optimal unmanned aerial vehicle or remotely operated vehicle (ROV) inspection path centered on this location. This path planning not only considers the shortest distance between two points, but also takes into account various constraints such as topography, flight or navigation safety areas, equipment endurance, and angles that need to be observed in detail. Subsequently, the system sends a detailed task instruction set containing path points, flight attitude, gimbal angle, camera zoom, and other parameters to the designated unmanned equipment through a wireless communication link. The equipment receives the instructions and autonomously executes the inspection task, and real-time returns the collected high-definition video, infrared image, or three-dimensional point cloud data. These returned data will be used as new data sources and re-enter step S1, forming a dynamic closed loop of discovery, exploration, and re-analysis to achieve continuous tracking and confirmation of hidden dangers.

[0082] For intelligent emergency response operation, when the prediction and analysis module of step S5 determines that an emergency water conservancy event is about to occur or is occurring, such as a predicted flood peak is about to arrive at a certain time or an upstream water pollution has been confirmed, the system will start the intelligent emergency response mechanism. This mechanism aims to convert the optimal scheduling scheme or emergency plan recommended by S5 into actual control actions for water conservancy engineering facilities.

[0083] Preferably, to ensure the safety and reliability of critical infrastructure control, the system supports multiple execution modes. In the fully automatic mode, for high-confidence early warnings and mature dispatch plans, the system can directly parse the optimal control strategy (such as the staged opening sequence of reservoir gates in the next few hours) into instructions conforming to the industrial control protocol, and issue them to the remote control unit (RTU) or programmable logic controller (PLC) of the relevant water conservancy facilities through a secure isolation gateway, realizing unmanned and automated control of devices such as gates and pump stations.

[0084] In the semi-automatic mode of human-machine collaboration, the system will first present the generated control instruction sequence to the management personnel, along with the decision explanation information provided in step S5 and the expected effects deduced in step S4. After the management personnel confirm the plan, they can execute it with one click. This mode balances the computational efficiency of artificial intelligence and the final decision-making power of human experts. All operations, whether automatically executed or manually confirmed, their execution status and actual effects (such as the actual discharge flow after gate opening, the actual change of downstream water level) will be monitored in real time and used as first-hand feedback data to input the system for the next step of rolling optimization.

[0085] By performing the above automated inspection and intelligent emergency response, the application effectively applies the results of decision analysis to the physical world, significantly improving the response speed and control accuracy of water conservancy systems to various events, and providing valuable practical data and feedback basis for the overall optimization and resilience enhancement of the system in step S7.

[0086] Step S7, energy management and network security protection for the method of steps S1 to S6, and based on scenario simulation and risk assessment, to enhance system resilience, which specifically includes: Using water conservancy digital twin as a virtual test field, actively simulating one or more pre-set extreme working conditions or key device failure scenarios, and conducting system-level stress testing; Under simulated scenarios, quantitatively evaluate the performance of the system from data perception to decision response, to accurately identify potential weak links; According to the evaluation results, iteratively optimize and reinforce the artificial intelligence models in the analysis and decision support module, or the emergency response strategies in the intelligent execution and response module; In this embodiment, step S7 is optimization and resilience management, which is a guarantee for the long-term, stable, safe and efficient operation of the whole smart water monitoring and management method of the application. This step is not an isolated terminal operation, but a continuous management and self-improvement mechanism throughout the whole process of system operation. Its core is to improve the energy efficiency, network security, and resilience and adaptability of the whole methodological framework in the face of unknown disturbances through proactive system-level optimization and risk management.

[0087] In a specific implementation of the application, this step first focuses on the energy management of the system. Since the method relies on a large number of Internet of Things sensors, unmanned aerial vehicles and other terminal devices deployed in the wild, the sustainability of energy is directly related to the continuity and integrity of data acquisition. Therefore, the system conducts fine energy state monitoring and management of all energy-consuming devices. Preferably, the system establishes power consumption models for different types of devices and intelligently schedules the operation mode of the devices based on the business demand prediction output in step S5. For example, during the normal water period, the system can automatically reduce the sampling frequency of non-core area sensors or put them into periodic sleep mode to save power; and when the flood is predicted to come, it automatically wakes up the related devices and switches to high-frequency working mode to ensure the complete acquisition of key data. For mobile inspection devices such as unmanned aerial vehicles, the system optimizes their flight path and charging strategy when performing task planning in step S6 to achieve the optimal balance between inspection coverage and energy consumption.

[0088] In this embodiment, to protect the safety of the water conservancy, a key infrastructure information system, this step implements a comprehensive network security protection strategy. This is to defend against external attacks and internal misoperations and to ensure the confidentiality, integrity and availability of data. Preferably, at the data transmission level, all data collected from step S1 is transmitted using an encryption protocol as soon as it leaves the terminal device to prevent eavesdropping or tampering during communication. At the data storage and access level, the system builds a rigorous role-based access control system, assigns the minimum necessary permissions related to their business to management personnel with different responsibilities, and prevents unauthorized operations. At the same time, the system records all key operation logs, including each control instruction issued in step S6, for security auditing and post-tracing. The system also deploys an anomaly detection mechanism for its own running state to continuously monitor network traffic and service invocation behavior to identify potential intrusions or abnormal activities.

[0089] In this embodiment, one of the key innovations of the application is to improve the resilience and adaptability of the whole system through this step. Resilience is the ability of the system to maintain its core functions and quickly recover after being subjected to predictable or unpredictable disturbances. To achieve this purpose, the system makes full use of the water conservancy digital twin built in step S4 as a virtual test field to conduct systematic scenario simulation and risk assessment.

[0090] Specifically, instead of passively waiting for failures to occur, the system actively simulates various extreme or failure scenarios in the digital twin environment. These scenarios include: partial key sensor networks being massively interrupted in step S1 due to natural disasters; core servers failing, causing the analysis capability in step S5 to decrease; or the remote control link in step S6 being interrupted due to strong electromagnetic interference. In these simulated stress tests, the system comprehensively evaluates the performance of the method of the present application, for example, how much the accuracy of the fusion and prediction model will decrease in the case of data missing, and whether the effectiveness of the emergency decision will be significantly affected.

[0091] Through analysis of the simulation results described above, the system can identify weak links in the entire methodological system. Based on these identified potential risk points, the system can automatically or assist managers to optimize and reinforce. For example, if the evaluation results show that there is a single point failure risk in the sensor network of a certain area, the system will suggest adding redundant backup devices in that area. If the simulation shows that the existing emergency plan is not effective under a certain extreme flood combination, the system will drive the water conservancy digital twin to perform a new round of deduction to iteratively generate a more robust emergency dispatch strategy and update the plan library in step S6.

[0092] Finally, this step will form a continuous closed-loop feedback with the actual execution effect of step E and the experience obtained in various scenario simulations. These information will be used to update the knowledge graph of step B, calibrate the water conservancy digital twin of step C, and retrain the artificial intelligence model of step D. Through this cycle of self-optimization and resilience improvement, the present application ensures that it not only can handle current business, but also can continuously learn and evolve to adapt to future more complex and uncertain challenges, ensuring the long-term vitality and reliability of the entire smart water conservancy monitoring and management system.

[0093] At least one embodiment of the present application also provides a smart water conservancy monitoring system based on Internet of Things, which comprises an Internet of Things perception module, a data processing module, a data fusion and knowledge graph module, a digital twin module, a deep analysis and decision module, and an automatic inspection and emergency response module.

[0094] The Internet of Things perception module is used to acquire heterogeneous water conservancy data from multiple sources of Internet of Things sensors, remote sensing devices, video monitoring devices and weather stations, and to perform preprocessing and standardization operations on the heterogeneous water conservancy data from multiple sources.

[0095] In this embodiment, the Internet of Things perception module constitutes the sensory network of the entire smart water conservancy monitoring system, and its core responsibility is to widely and continuously capture raw water conservancy data from the physical world, which is the data source for all subsequent analysis and decision-making.

[0096] The Internet of Things sensing module is not a single device, but a heterogeneous and distributed collection system. The module builds a comprehensive sensing network by integrating sensing devices deployed throughout the water conservancy system. It integrates various data acquisition interfaces and adapters to accommodate data from different sources. In at least one embodiment of the present application, the Internet of Things sensing module includes a field sensor interface, a visual and remote sensing data access unit, a third party system interface, and a manual data reporting interface.

[0097] Field sensor interface: used to directly connect various Internet of Things sensors deployed at key locations such as reservoirs, dams, river channels, pumping stations, and gateways. These sensing devices cover devices that measure water level, flow rate, flow velocity, water pressure, rainfall, soil moisture, water turbidity, pH value, dissolved oxygen, and other parameters. The module ensures stable connection and data collection with these front-end devices by supporting multiple industrial communication protocols (such as Modbus, RS-485) and wireless transmission technologies (such as LoRa, NB-IoT, 4G / 5G).

[0098] Visual and remote sensing data access unit: This unit is responsible for receiving visual data from high-definition video monitoring networks, unmanned aerial vehicle aerial photography, satellite remote sensing images, etc. It can process real-time video streams and periodic remote sensing image files, providing intuitive unstructured information such as macro water area changes, key project construction conditions, and water surface floating objects for the system.

[0099] Third party system interface: Through standard API (Application Programming Interface) or data middleware, the module also actively pulls or receives pushed data from external information systems (such as weather forecasting systems, hydrological information networks, and land and resources databases), obtaining key background information such as rainfall forecasts, typhoon paths, and historical hydrological sequences.

[0100] Manual data reporting interface: provides mobile applications or web forms for inspection personnel to report hidden danger information discovered through manual inspection that cannot be automatically monitored by sensors, such as small cracks in embankments, abnormal sounds of equipment, or water-related activities within the management scope, achieving human-machine collaborative data collection.

[0101] The purpose of the Internet of Things sensing module is to ensure the comprehensiveness and multidimensionality of data collection, providing the most original and extensive data foundation for building a panoramic digital portrait of the water conservancy system.

[0102] Data processing module for pre-processing and standardizing heterogeneous water conservancy data from multiple sources.

[0103] In this embodiment, the data processing module plays the role of a system data purifier and interpreter. It receives raw, mixed, and format-variant data from the IoT sensing module and transforms it into high-quality, standardized data that subsequent modules can understand and use. The IoT sensing module and the data processing module together form the data acquisition and processing part of the system, which is the cornerstone of the entire intelligent system.

[0104] The core functions of this data processing module revolve around data preprocessing and standardization, including: Data cleaning unit: This unit is responsible for improving data quality. For common signal noise and glitches in the sensor acquisition process, it optimally uses Kalman filtering, wavelet denoising, and other algorithms for smoothing. For sporadic data point missing due to network delay or device failure, it can use linear interpolation, spline interpolation, and other methods for filling. At the same time, it also has an embedded outlier detection algorithm, which automatically identifies and marks those obviously deviating from the normal range of error data by setting statistical thresholds or using domain knowledge rules (such as water level change rate cannot exceed physical limit).

[0105] Data standardization and conversion unit: This unit is responsible for solving the heterogeneity problem of data. It uniformly parses and converts data of different sources and protocols (such as JSON, XML, binary code stream, etc.) into the system's internally predefined standard data structure. This structured operation not only unifies the data format, but also standardizes the key fields, such as unifying all time information into time stamp format with time zone (UTC) and all spatial information into standard geographic coordinate system (such as WGS84), ensuring consistency of data in time and space dimensions.

[0106] Time and space alignment unit: Since the sampling frequency and reporting time of different data sources may differ, this unit is responsible for aligning all data streams on the time axis. Preferably, it uses a unified time reference to resample or interpolate different frequency data to the same time granularity, laying the foundation for subsequent multi-data fusion analysis.

[0107] Through the processing of this data processing module, the originally chaotic raw data is transformed into a clean, complete, and uniformly formatted data set, providing reliable raw materials for subsequent in-depth analysis.

[0108] The data fusion and knowledge graph module is used to construct a water conservancy field knowledge graph based on preprocessed and standardized data, and to perform multi-dimensional data fusion to generate comprehensive water conservancy situation data.

[0109] In this embodiment, the data fusion and knowledge graph module is the core of the leap from data to cognition, which plays the role of the system's knowledge brain and is responsible for deep mining the internal relationship between data. This module, together with the subsequent digital twin module, constitutes the core function of data cognition and twin of the system.

[0110] The data fusion and knowledge graph module is mainly composed of two interrelated water conservancy knowledge graph management sub-modules and multi-dimensional data fusion sub-modules.

[0111] Water conservancy knowledge graph management sub-module: This sub-module is responsible for building and maintaining a structured knowledge base describing the water conservancy field. It stores and manages physical entities in the water conservancy system (such as reservoirs, dams, and stations), conceptual entities (such as emergency plans and dispatching rules), and their complex relationships (such as location, control, and influence) in the form of a graph. The construction of the knowledge graph absorbs engineering design data, industry standards, and the experience of water conservancy experts, providing rich background semantics for understanding data.

[0112] Multi-dimensional data fusion sub-module: This sub-module performs deep fusion on pre-processed data based on the knowledge graph. Its fusion process is multi-level: Semantic fusion: It associates standardized data points with entities in the knowledge graph, making data no longer isolated values. For example, a water level data will be associated with a reservoir entity, thereby automatically linking to all related knowledge of the reservoir's capacity, design standards, and downstream river channels.

[0113] Spatio-temporal fusion: It integrates various data from different locations and times under a unified spatio-temporal reference, forming a situation slice that can comprehensively reflect the comprehensive state of a water conservancy system in a certain region and at a certain time.

[0114] Feature fusion: To improve the accuracy of key indicators, when multiple sensors monitor the same indicator, this sub-module dynamically evaluates the reliability of each data source and assigns different weights for weighted fusion, outputting a more reliable comprehensive result than any single source.

[0115] The final output of this data fusion and knowledge graph module is comprehensive water conservancy situation data, which is a rich semantic, highly correlated, and high-confidence information set, providing high-quality cognitive input for the construction of water conservancy digital twins and intelligent decision-making.

[0116] Digital twin module, for constructing water conservancy digital twins based on comprehensive water conservancy situation data, and realizing real-time driving of water conservancy digital twins and hydrological and hydraulic process simulation.

[0117] In this embodiment, the digital twin module is a virtual mirror connecting the physical world and the information world, which builds a simulation environment that evolves synchronously with the real water system and has the ability to deduce.

[0118] According to the functional division, the digital twin module includes a high-fidelity modeling unit, a real-time driving and state synchronization unit, and a hydrological and hydraulic simulation engine unit.

[0119] High-fidelity modeling unit: This unit is responsible for building a three-dimensional visual digital model of the water system. It comprehensively utilizes BIM, GIS, oblique photography, and laser point cloud technologies, not only accurately reproducing the geometric shape, spatial position, and topological relationship of facilities such as reservoirs, river channels, embankments, and gates and pumps, but also endowing them with real physical properties (such as material, roughness) and operating mechanisms (such as the relationship between gate opening and flow).

[0120] Real-time driving and state synchronization unit: This unit is the key to ensuring synchronization between the virtual and real worlds. It uses the comprehensive water situation data output by the data fusion and knowledge graph module as the driving source to continuously update the state of each element in the twin body. The rising water level, gate opening, and water quality changes in the physical world are all mapped to the virtual model in real time and synchronously, making it a dynamic, visual, and living system.

[0121] Hydrological and hydraulic simulation engine unit: This is the core of the twin module that goes beyond simple visualization. This unit deeply integrates professional hydrodynamic, water quality transport and diffusion mechanism models in the twin environment. Using real-time monitoring data as the boundary condition of the model, this engine can dynamically simulate possible future events, such as simulating the flood evolution process, inundation range under different flood discharge schemes, or predicting the diffusion path and concentration changes of pollutants in sudden pollution events.

[0122] Through this digital twin module, the system ultimately integrates the high-fidelity twin model and simulation results on a panoramic intelligent visualization platform. Managers can not only see the current state of the water system through this platform, but also predict its future changes, and can use the interactive simulation functions provided by the platform to conduct risk-free hypothesis-deduction analysis in the virtual environment, such as manually adjusting gate opening or setting different rainfall scenarios, to intuitively observe and evaluate the execution effect of different strategies.

[0123] Deep analysis and decision-making module for multi-modal deep analysis of comprehensive water situation data and water digital twin, and generation of interpretable intelligent decision support information using artificial intelligence technology.

[0124] In this embodiment, the deep analysis and decision-making module is the intelligent decision-making hub of the entire system, responsible for extracting meaningful insights and solutions from massive data and simulation results.

[0125] The deep analysis and decision module internally integrates an algorithm library and an analysis engine. According to the functional division, the deep analysis and decision module includes a predictive analysis unit, an abnormal event identification unit, a causal tracing and correlation analysis unit, an interpretable decision generation unit, and an optimization scheme recommendation unit.

[0126] Predictive analysis unit: This unit is built-in with various artificial intelligence prediction models. Preferably, time series models such as long short-term memory networks (LSTM) are used to learn from historical and real-time monitoring data to achieve accurate prediction of key parameters such as water level, flow, etc. in the future hours or even days, providing valuable time for flood prevention warning and water resource scheduling.

[0127] Abnormal event identification unit: This unit uses multi-modal analysis technology to detect anomalies in various types of data. For example, convolutional neural networks (CNN) are used to analyze real-time video streams to automatically identify diseases such as seepage and cracks in embankments, or abnormal floating objects on the water surface; by analyzing the mutation of time series data, the operation of the device is identified.

[0128] Causal tracing and correlation analysis unit: When an abnormal event is detected (such as water pollution), this unit uses techniques such as graph neural networks (GNN) to combine water system connectivity and pollution outlet information in the knowledge graph for intelligent tracing analysis, quickly locking potential pollution sources.

[0129] Interpretable decision generation unit: To avoid black box decisions, this unit introduces explainable AI technology. When outputting any prediction, warning or recommendation scheme, it will provide decision basis simultaneously, in the form of heat map, feature importance ranking, etc. to show users which factors led to the current result, thereby enhancing the credibility of the decision.

[0130] Optimization scheme recommendation unit: This unit combines prediction results and the simulation and deduction capabilities of water conservancy digital twins to automatically generate and evaluate various response schemes using reinforcement learning or optimization algorithms. For example, before a flood is forecast, it can simulate various reservoir scheduling schemes and recommend a comprehensive optimal scheduling strategy in terms of flood safety, power generation efficiency, ecological water demand, etc.

[0131] The deep analysis and decision module turns raw data into high-value decision intelligence, providing core support for scientific, forward-looking and intelligent water management.

[0132] Automatic inspection and emergency response module, for performing automatic inspection and / or intelligent emergency response according to intelligent decision support information to deal with water abnormal events.

[0133] In this embodiment, the automatic inspection and emergency response module is an intelligent executor that connects decision-making and action, responsible for accurately and efficiently executing the instructions generated by the upper module in the physical world.

[0134] The automatic inspection and emergency response module realizes the transition from passive response to active intervention. According to the function, the automatic inspection and emergency response module includes an automatic inspection task management unit, an intelligent emergency response execution unit, and an execution feedback and recording unit.

[0135] Automatic inspection task management unit: When the analysis module identifies potential hazards or according to the preset inspection plan, this unit will automatically generate unmanned inspection tasks. It can intelligently plan the optimal inspection path for devices such as drones and underwater robots, and generate detailed task instructions (including waypoints, flight height, shooting angle, etc.). During task execution, this unit is responsible for remotely monitoring device status and receiving high-definition data transmission, forming a closed loop of discovery, thorough investigation, and reconfirmation.

[0136] Intelligent emergency response execution unit: After receiving high-confidence early warnings and decision-making schemes, this unit is responsible for converting them into actual control of water conservancy facilities. It supports multiple execution modes: Automatic execution mode: For emergency and mature scenarios, dispatch instructions (such as timing and quantitative opening of gates) can be parsed into industrial control protocols and directly issued to on-site PLC or RTU through a secure channel, achieving millisecond-level automated response.

[0137] Human-machine collaborative mode: The recommended dispatch scheme and pre-play results are presented to management personnel in a clear and visual manner, and after final review and confirmation by human, one-key execution is performed. This mode ensures human supervision of critical decisions, balancing efficiency and safety.

[0138] Execution feedback and recording unit: All executed operations (whether automatic or manual), their status, process, and actual effects will be accurately recorded by this unit and input as feedback data into the system for subsequent effect evaluation and model optimization.

[0139] The automatic inspection and emergency response module turns intelligent decision-making into smart action, significantly improving the efficiency and accuracy of water conservancy system management.

[0140] In at least one embodiment of the present application, the Internet of Things-based smart water conservancy monitoring system further includes a system operation and resilience module.

[0141] The system operation and resilience module is used for energy management, network security protection of the Internet of Things-based smart water conservancy monitoring method, and based on scenario simulation and risk assessment, the system resilience is improved.

[0142] In this embodiment, the system operation and resilience module is the health manager and immune system of the entire system. Its goal is to ensure the long-term, safe and stable operation of the entire method system, and continuously evolve to adapt to new challenges.

[0143] The function of the system operation and resilience module is continuous throughout the system operation. According to the functional division, the system operation and resilience module includes a system energy and resource management unit, a network and data security protection unit, and a system resilience evaluation and enhancement unit.

[0144] System energy and resource management unit: This unit monitors and optimizes the energy consumption of all hardware devices in the system, especially the sensors deployed in the field. It can intelligently adjust the sampling frequency and working mode of the devices according to the business tide characteristics, to maximize energy saving under the premise of ensuring data quality. At the same time, it also performs elastic scaling scheduling on the computing and storage resources of the server to ensure the performance stability of the system under high concurrency access or large data processing.

[0145] Network and data security protection unit: The network and data security protection unit builds a three-dimensional security line for the system. It is responsible for the encrypted transmission of data throughout the link to prevent eavesdropping and tampering; establishes a strict access permission control system based on roles to prevent data leakage and unauthorized operations; and records detailed operation logs for security audit. In addition, it also deploys an intrusion detection system to actively defend against network attacks.

[0146] System resilience evaluation and enhancement unit: This is the most forward-looking function of this module. It uses the digital twin module as a sandbox to actively simulate various extreme failure scenarios, such as interruption of key sensor networks, core server downtime, and encounter with super-standard floods, etc. By evaluating the system's performance under these pressures, it identifies the weak links in the system. Based on the evaluation results, the system can propose reinforcement suggestions (such as adding redundant devices), or drive model iteration to optimize emergency plans, so that the system's resistance and recovery capabilities are improved before the real disaster occurs.

[0147] At least one embodiment of the present application also provides a computer device, which includes a processor and a memory. The processor is used to execute the method for monitoring intelligent water conservancy based on Internet of Things provided by any of the above embodiments of the present application. The memory is used to store executable instructions of the processor, such as application programs. The number of processors can be one or more. The application programs stored in the memory can include one or more than one module corresponding to a set of instructions. In addition, the processor is configured to execute the instructions to perform the above-mentioned method for monitoring intelligent water conservancy based on Internet of Things.

[0148] The computer device can further include a power supply component configured to manage power supply of the computer device, a wired or wireless network interface configured to connect the computer device to a network, and an input / output (I / O) interface. The computer device can operate based on an operating system stored in the memory, such as Windows ServerTM, Mac OSXTM, UnixTM, LinuxTM, FreeBSDTM, or the like.

[0149] The computer readable storage medium stores computer executable instructions. When the computer executable instructions are executed by the processor, the computer readable storage medium enables the computer device to perform the above-mentioned method for monitoring smart water conservancy based on Internet of Things.

[0150] The non-transitory computer readable storage medium stores computer executable instructions. When the computer executable instructions are executed by the processor, the non-transitory computer readable storage medium enables the computer device to perform the above-mentioned method for monitoring smart water conservancy based on Internet of Things. The method for monitoring smart water conservancy based on Internet of Things is executed by the agent program.

[0151] Those skilled in the art can realize that the algorithm steps of each example described in combination with the embodiments disclosed in the present application can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized by hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0152] The computer program product includes computer program / instructions. When the computer program / instructions are executed by the processor, the computer program product enables the computer device to perform the above-mentioned method for monitoring smart water conservancy based on Internet of Things.

[0153] If the above functions are realized in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the parts that contribute to the prior art or parts of the technical solutions can be embodied in the form of a computer program product stored in a storage medium, including a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the Internet of Things-based smart water conservancy monitoring method of the embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a magnetic disk or an optical disk, and various program code storage media.

[0154] As shown in the present application and claims, unless the context clearly indicates otherwise, "one", "a", and / or "the" do not refer to the singular, but can also include the plural. Generally, the term "comprising" only indicates the inclusion of the steps and elements explicitly identified, and these steps and elements do not constitute an exclusive list, and the method or device can also include other steps or elements.

[0155] The above only describes the preferred embodiments of the present application and does not limit the present application. Any modification, equivalent replacement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A smart water conservancy monitoring method based on Internet of Things, characterized in that, The method comprises the following steps: Step S1, acquiring heterogeneous water conservancy data from multiple sources including Internet of Things sensors, remote sensing devices, video monitoring devices, and weather stations; Step S2, preprocessing and standardizing the heterogeneous water conservancy data from multiple sources; Step S3, based on the preprocessed and standardized heterogeneous water conservancy data from multiple sources, constructing a water conservancy domain knowledge graph, and performing multi-dimensional data fusion to generate comprehensive water conservancy situation data; Step S4, based on the comprehensive water conservancy situation data, constructing a water conservancy digital twin, and realizing real-time driving of the water conservancy digital twin and hydrological and hydraulic process simulation; Step S5, performing multi-modal deep analysis on the comprehensive water conservancy situation data and the water conservancy digital twin, and generating interpretable intelligent decision support information using artificial intelligence technology. 2.The smart water conservancy monitoring method based on the Internet of Things according to claim 1, characterized in that, Step S2 includes protocol adaptation, format conversion, data denoising, missing value filling, and outlier detection of the heterogeneous water conservancy data from multiple sources. 3.The wisdom water conservancy monitoring method based on the Internet of Things according to claim 1, characterized in that, In step S3, multi-dimensional data fusion includes: based on the water conservancy domain knowledge graph, the standardized heterogeneous water conservancy data from multiple sources is semantically associated with the pre-defined water conservancy entities in the knowledge graph to provide data with contextual background information; for the same monitoring indicators from multiple data sources, the reliability weights of each data source are dynamically determined according to the historical data performance of each data source, and the values of the monitoring indicators are weighted and fused to generate a more reliable comprehensive indicator value. 4.The wisdom water conservancy monitoring method based on the Internet of Things according to claim 1, characterized in that, In step S4, the hydrological and hydraulic process simulation includes: embedding hydrodynamic or water quality transport models in the water conservancy digital twin, and using the comprehensive water conservancy situation data as the real-time boundary conditions of the models to dynamically simulate and predict the processes of flood evolution, reservoir storage and discharge, or pollutant diffusion. 5.The wisdom water conservancy monitoring method based on the Internet of Things according to claim 1, characterized in that, In step S5, the artificial intelligence technology includes: using a time series prediction model to predict water level and flow; and / or using an image recognition model to analyze video monitoring data to identify dam safety hazards or abnormal floating objects on the water surface.

6. The Internet of Things-based intelligent water conservancy monitoring method according to any one of claims 1 to 5, characterized in that, Further comprising: Step S6, based on the intelligent decision support information, performing automated inspection or intelligent emergency response to deal with water conservancy abnormal events.

7. The smart water conservancy monitoring method based on the Internet of Things according to claim 6, characterized in that, In step S6, performing automated inspection includes: planning the inspection path and task of a drone or an underwater robot based on the intelligent decision support information, and remotely controlling it to perform autonomous inspection of water conservancy engineering facilities. 8.The wisdom water conservancy monitoring method based on the Internet of Things according to claim 6, characterized in that, Further comprising: Step S7, energy management and network security protection for the method of steps S1 to S6, and based on scenario simulation and risk assessment, improving system resilience, which specifically includes: Simulating multiple failure scenarios of extreme weather or equipment failure in the digital twin to evaluate the response capability and weak links of the system, and optimizing emergency plans and resource allocation strategies according to the evaluation results. 9.The smart water conservancy monitoring method based on the Internet of Things according to claim 8, characterized in that, Simulating multiple failure scenarios of extreme weather or equipment failure in the digital twin to evaluate the response capability and weak links of the system, and optimizing emergency plans and resource allocation strategies according to the evaluation results, including: Using the water conservancy digital twin as a virtual test field to actively simulate one or more pre-set extreme working conditions or key equipment failure scenarios for system-level stress testing; In the simulation scenario, the quantitative evaluation system evaluates the performance of the whole link from data perception to decision response to accurately identify potential weak links; According to the evaluation results, the artificial intelligence models in the analysis and decision support module or the emergency response strategies in the intelligent execution and response module are iteratively optimized and reinforced.

10. A smart water conservancy monitoring system based on Internet of Things, characterized in that, It includes: The Internet of Things perception module is used to obtain heterogeneous water conservancy data from multiple sources such as Internet of Things sensors, remote sensing devices, video monitoring devices and weather stations, and to preprocess and standardize the heterogeneous water conservancy data from multiple sources; The data processing module is used to preprocess and standardize the heterogeneous water conservancy data from multiple sources; The data fusion and knowledge graph module is used to construct a water conservancy knowledge graph based on preprocessed and standardized data, and to perform multi-dimensional data fusion to generate comprehensive water conservancy situation data; The digital twin module is used to construct a water conservancy digital twin based on comprehensive water conservancy situation data, and to realize real-time driving of the water conservancy digital twin and simulation of hydrological and hydraulic processes; The deep analysis and decision module is used to perform multi-modal deep analysis on comprehensive water conservancy situation data and water conservancy digital twins, and to generate interpretable intelligent decision support information using artificial intelligence technology; The automatic inspection and emergency response module is used to perform automatic inspection and / or intelligent emergency response based on intelligent decision support information to deal with water conservancy abnormal events.

Images