Reservoir water rain condition monitoring system and monitoring method

Abstract

The invention provides a reservoir water rain condition monitoring system and a monitoring method, the system constructs a dual monitoring system of physical observation reference and electronic automatic acquisition, a customized stainless steel water gauge and a multi-element acquisition device are deployed on key sections of a reservoir dam and a drainage basin, and a front-end edge calculation module is utilized to invert the flow in real time. And after the data is transmitted to the cloud platform, intelligent management and control are carried out through the four-pre functional module. The system displays the watershed situation based on a GIS graph, integrates an online flow measurement algorithm to generate a water and rain condition process line in real time, and realizes visual supervision by combining video monitoring and OSD data superposition. The core multi-target optimization scheduling model comprehensively considers flood control safety and power generation benefits, generates an optimal scheduling scheme through simulation deduction, and provides a closed-loop control interface. The problems of low monitoring precision, scheduling decision lag and the like in the prior art are solved, and the automatic management level and the water resource utilization rate of the reservoir group are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of water conservancy and hydropower engineering management and information technology, and in particular to a reservoir water and rainfall monitoring system and monitoring method. Background Technology

[0002] Currently, in the management of reservoirs in mountainous river basins, most reservoirs are located in remote mountainous areas, are numerous and widely distributed, and are greatly affected by harsh climates and complex geographical environments. For a long time, there have been problems such as the lack of upstream and downstream hydrological monitoring stations and a scarcity of reservoir scheduling data. Existing monitoring methods often rely on the most primitive manual patrols and simple manual observation and calculation methods. This approach not only increases the input of manpower and resources and is labor-intensive, but is also highly susceptible to data distortion due to human factors, failing to meet the needs of modern, refined reservoir scheduling.

[0003] In the prior art, although there have been some studies on reservoir monitoring systems, for example, prior art 1 (CN115388949A) discloses an intelligent reservoir monitoring and management system, which proposes to use a digital twin platform to view water and rainfall information and realize the "four predictions" function, focusing on data acquisition and three-dimensional visualization of front-end equipment; Prior art 2 (CN116299780A) discloses a water and rainfall monitoring system and its monitoring method, which mainly uses a rainfall monitoring module and a water volume monitoring module to perform basic data statistics, storage and threshold alarm, focusing on preventing the risk of untimely flood control; Existing technology 3 (CN107622113B) discloses an intelligent management system for reservoir safety, which constructs a hardware architecture based on a cloud service platform and realizes centralized data processing and GIS application services. However, most of the above-mentioned existing technologies focus on data collection and transmission and general information display, lacking a mechanism for "dual verification of physical benchmarks and electronic monitoring" in the specific environment of complex mountainous areas. They do not fully solve the problem that electronic sensors are prone to zero-point drift in harsh environments and are difficult to calibrate remotely. At the same time, they lack a comprehensive solution that deeply integrates on-site customized stainless steel water gauges, island-shore station construction, and online flow measurement algorithms. As a result, the accuracy of flood forecasting and scheduling under extreme weather conditions is still insufficient, and it cannot effectively support the multi-objective optimization scheduling between maximizing the power generation benefits of cascade power stations and flood control safety. Summary of the Invention

[0004] The main objective of this invention is to provide a reservoir water and rainfall monitoring system and method, which solves the technical problems of data distortion and lack of scientific basis for scheduling decisions caused by the lack of upstream and downstream monitoring stations and reliance on manual observation in existing mountain reservoirs, which in turn leads to large water loss and flood control safety hazards.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a reservoir water and rainfall monitoring system, with island-shore monitoring stations configured in the reservoir area and upstream and downstream basins, the monitoring stations integrating front-end acquisition equipment, physical observation reference facilities and communication modules; A communication network is established to connect with monitoring stations and create data transmission channels. The software platform, deployed on the server side, receives data packets sent by monitoring stations through a communication network, and has built-in data verification module, online flow measurement algorithm module, and multi-objective optimization scheduling model; The monitoring stations are equipped with water and rainfall monitoring devices and online flow monitoring systems that support basic data calculations on the hardware side, while the software platform supports advanced simulation and command issuance on the server side.

[0006] In the preferred scheme, the island-shore monitoring station adopts an independent station building structure; The physical observation benchmark facility is a vertical stainless steel water gauge, which is custom-made based on the dam elevation data. The stainless steel water gauge is fixedly installed on the bedrock or concrete base of the dam and the monitoring section, serving as a physical benchmark for manual comparison and sensor calibration. The front-end data acquisition equipment includes a telemetry terminal, a tipping bucket rain gauge, a float-type water level gauge, a Doppler current meter, and a high-definition monitoring camera. The telemetry terminal has an embedded hydraulic calculation unit. The software platform adopts a containerized microservice architecture and integrates IoT communication middleware, time-series database, hydraulics calculation service daemon, and simulation engine.

[0007] In the preferred scheme, the multi-objective optimization scheduling model is configured to receive real-time water and rainfall data and meteorological forecast data, and output digital scheduling instruction packages; The online flow measurement algorithm module includes the water level-cross-sectional area relationship function and hydraulic parameter correction logic.

[0008] A method for monitoring water and rainfall conditions in reservoirs, S1, constructing island-shore monitoring stations at key nodes in the reservoir and watershed, and installing customized stainless steel water gauges on the observation sections to establish a physical water level observation benchmark; S2. Deploy front-end data acquisition devices within the monitoring site. After the front-end data acquisition devices are powered on and initialized, they establish a long-term connection channel with the server through the built-in communication protocol stack. S3. The front-end acquisition device synchronously acquires rainfall data, water level data, and flow velocity data at a preset frequency, and calls the online flow measurement algorithm module to calculate and encapsulate the acquired raw signals to generate monitoring data packets; S4. The server-side software platform receives monitoring data packets, parses and stores them in the standard hydrological database, and simultaneously calls the verification module to perform linear regression analysis and parameter correction on the automatically collected water level data using the manually observed stainless steel water gauge data. S5. Based on the corrected water and rainfall data and online flow measurement data, the software platform drives the multi-objective optimization scheduling model to perform simulation and generate the optimal scheduling scheme and issue it for execution.

[0009] In the preferred embodiment, the deployment and communication connection between the front-end acquisition device and the server in step S2 specifically includes: S21. Configure high-performance data processing servers and application server clusters on the server side, deploy Linux operating system and Docker containerized runtime environment, and build a load-balanced high-availability service architecture. S22. Deploy IoT communication middleware in a container environment, configure TCP / UDP listening ports and message parsing engine, define a private communication protocol stack, and the data frame structure of the communication protocol stack includes frame header identifier, device unique ID, function control code, encrypted data field and CRC cyclic redundancy check code. S23. Configure the server's public IP address or domain name in the embedded system of the front-end acquisition device, and set the reconnection trigger threshold and heartbeat packet sending interval. S24. The front-end acquisition device initiates a connection request to the server. The communication middleware responds with a handshake signal and establishes a TCP long connection. Subsequently, the front-end acquisition device periodically sends heartbeat packets containing the device's operating status. The communication middleware maintains a dedicated data transmission channel through a heartbeat keep-alive mechanism. If no heartbeat packet is received within a preset time, the device is determined to be offline and a reconnection waiting logic is triggered.

[0010] In the preferred embodiment, the calculation steps in step S3, which call the online flow measurement algorithm module, include: S31. Constructing the water level-cross-sectional area relationship function The least squares method was used to perform polynomial fitting on the measured cross-sectional topographic data, and the generated function formula is as follows: ; in, The cross-sectional area of ​​the water passage. This is the real-time water level value. These are the fitting coefficients. It is the highest power of the polynomial; The function formula and fitting coefficients are compiled into a binary calculation configuration file and written to the storage unit of the front-end acquisition device; S32. Obtain real-time water level values ​​using a float-type water level gauge. Real-time flow velocity values ​​at representative points are obtained using a flow meter. ; S33, The computing unit of the front-end acquisition device reads the configuration file and displays the real-time water level value. Substitute the values ​​into the function formula to calculate the current cross-sectional area of ​​the water flow. ; S34. The computer deployment steps for calling the hydraulic parameter model are as follows: A hydraulics calculation service daemon is deployed in the embedded Linux system of the front-end acquisition device. The daemon loads a pre-compiled dynamic link library of hydraulics algorithms. The dynamic link library internally encapsulates the Manning formula or the logarithmic velocity distribution formula, and reads the Manning roughness coefficient of the river channel from the configuration file at runtime. Hydraulic radius and water surface gradient ; The main program transmits the real-time flow rate value through an inter-process communication mechanism. The auxiliary parameters are passed to the dynamic link library, the library function performs iterative calculations, and outputs the average flow velocity of the cross section. ; S35, through formula Complete real-time traffic Perform multiplication operations and upload the calculation results in a package.

[0011] In the preferred embodiment, the data processing and verification deployment of the software platform in step S4 includes: S41. Deploy a relational database and a time-series database on the database server. The relational database stores basic site information, and the time-series database stores high-frequency monitoring data. S42. Develop a data cleaning service module. The data cleaning service module resides in memory and scans the data entering the database in real time. S43. Execute outlier removal logic. When outlier data is detected within a short period of time... When the jump amplitude exceeds the preset physical limit threshold, the data is automatically marked as a suspected anomaly and a video surveillance capture command is triggered. S44. Execute the dynamic comparison logic, periodically input the manual observation values ​​of the stainless steel water gauge, and calculate the deviation between the manual observation values ​​and the automatic values ​​of the float-type water level gauge. ,like If the value remains above the calibration threshold, sensor zero-point drift correction parameters are automatically generated and sent to the front-end acquisition device via the communication network for parameter updates.

[0012] In the preferred solution, the software platform deploys a "four-pre" functional module, and the module's operation steps include: S51. Forecast: Integrate gridded rainfall forecast data released by meteorological departments and combine it with watershed runoff generation and confluence models to calculate the inflow process in the future period; S52. Early Warning: Set graded early warning thresholds. The system monitors water level and rainfall in real time. Once the threshold is reached, an alarm message is immediately pushed through the SMS gateway and mobile application. S53. Pre-simulation: Using the simulation engine, load the reservoir scheduling rules, input the predicted inflow, and simulate the reservoir water level changes and downstream discharge process under different scheduling schemes. S54. Contingency Plan: Based on the results of the pre-rehearsal, intelligently recommend scheduling schemes and generate a scheduling execution plan table that includes the number of gate openings, opening height, and discharge duration.

[0013] In the preferred scheme, the specific steps for generating the optimal scheduling scheme in step S5 include: S91. Construct a multi-objective optimization scheduling model and define objective functions in the software platform. The objective functions include the objective of minimizing flood disaster losses and the objective of maximizing power generation benefits. S92. Set a set of constraints. The set of constraints is integrated into the rule engine of the software platform. The set of constraints includes the maximum design water level of the dam, the safe discharge of the downstream river channel, the maximum opening speed of the gate, and the power output limit of the unit. S93. The software platform calls the simulation engine and, based on the current real-time water and rainfall data and the future forecast inflow, uses a genetic algorithm or particle swarm algorithm to generate a set of multiple candidate scheduling schemes. Each candidate scheduling scheme contains different flood discharge time and flood discharge flow sequences. S94. Perform parallel simulation calculations for each candidate scheduling scheme, predict the reservoir water level change curve and downstream flooding risk after each scheme is implemented, and automatically eliminate infeasible schemes that violate the constraint set.

[0014] In the preferred solution, the steps for decision-making and execution to achieve the optimal solution include: S101. A multi-dimensional decision matrix algorithm is used to score the selected set of feasible solutions. The scoring weights are dynamically adjusted according to the current flood control level. During the high-risk period of flood control, the weight of the disaster loss target is automatically increased, and during the non-flood season, the weight of the power generation benefit target is automatically increased. S102. Select the scheme with the highest score as the optimal scheduling scheme. The system automatically generates a digital scheduling instruction package corresponding to the scheme. The digital scheduling instruction package includes the gate action sequence, target opening degree and execution timestamp. S103. The digital dispatch instruction package is pushed to the decision-maker's client for visualization and confirmation. After confirmation, it is sent to the reservoir's local control unit for execution through the Internet of Things communication middleware, thereby realizing closed-loop control of the dispatch scheme.

[0015] This invention provides a reservoir hydrological and rainfall monitoring system and method. First, the system constructs island-shore monitoring stations at the reservoir dam and key watershed nodes, and innovatively introduces a customized stainless steel water gauge as a physical observation benchmark. This enables dynamic comparison and verification of automatic monitoring data and manual physical observation data, effectively solving the data drift problem caused by long-term operation of electronic sensors in the field environment, and significantly improving the authenticity and reliability of basic hydrological data.

[0016] Secondly, the system integrates an online flow measurement algorithm with an embedded hydraulic model. Through polynomial fitting and dynamic link library calculations, it can accurately retrieve flow data in real time. Combined with the "four predictions" (forecast, early warning, rehearsal, and contingency plan) functional modules, it realizes the transformation from single water level monitoring to comprehensive hydrological perception. Furthermore, this invention constructs a software platform including IoT communication middleware and a multi-objective optimization scheduling model. Based on accurate real-time data and weather forecasts, it can automatically generate optimal scheduling schemes that balance flood control safety and power generation benefits, reducing water wastage losses caused by inaccurate forecasting and significantly improving the water resource utilization rate of the cascade reservoir group. Finally, the system achieves visualized remote supervision through video monitoring and data character overlay technology. It can promptly detect and warn of "gradual" hidden dangers and faults, realizing the digitalization, networking, and intelligence of reservoir operation and management, reducing operation and maintenance costs, and improving the overall safety management level. Attached Figure Description

[0017] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a diagram of the reservoir water and rainfall monitoring system of the present invention; Figure 2 This is a flowchart of the reservoir water and rainfall monitoring method of the present invention; Figure 3 This is a schematic diagram of the island-shore monitoring station of the present invention; Attached labels: 1. Surveillance camera; 2. Multi-angle camera; 3. Stainless steel water gauge; 4. Detection well; 5. Island-shore monitoring station. Detailed Implementation

[0019] Example 1 like Figure 1-3 As shown, a reservoir water and rainfall monitoring system includes 5 island-shore monitoring stations configured in the reservoir area and upstream and downstream basins. The monitoring stations integrate front-end acquisition equipment, physical observation reference facilities and communication modules. A communication network is established to connect with monitoring stations and create data transmission channels. The software platform, deployed on the server side, receives data packets sent by monitoring stations through a communication network, and has built-in data verification module, online flow measurement algorithm module, and multi-objective optimization scheduling model; The monitoring stations are equipped with water and rainfall monitoring devices and online flow monitoring systems that support basic data calculations on the hardware side, while the software platform supports advanced simulation and command issuance on the server side.

[0020] In the preferred scheme, the five island-shore monitoring stations adopt an independent station building structure; The physical observation benchmark facility is a vertical stainless steel water gauge 3, which is customized and manufactured according to the dam elevation data. The stainless steel water gauge is fixedly installed on the bedrock or concrete base of the dam and the monitoring section, serving as a physical benchmark for manual comparison and sensor calibration. The front-end data acquisition equipment includes a telemetry terminal, a tipping bucket rain gauge, a float level gauge, a Doppler current meter, and a high-definition monitoring camera 1 and a multi-angle camera 2. The telemetry terminal has a hydraulic calculation unit embedded inside. The software platform adopts a containerized microservice architecture and integrates IoT communication middleware, time-series database, hydraulics calculation service daemon, and simulation engine.

[0021] In the preferred scheme, the multi-objective optimization scheduling model is configured to receive real-time water and rainfall data and meteorological forecast data, and output digital scheduling instruction packages; The online flow measurement algorithm module includes the water level-cross-sectional area relationship function and hydraulic parameter correction logic.

[0022] This embodiment details a reservoir water and rainfall monitoring system and its operating mechanism, aiming to provide sufficient disclosure and support for the technical features in the claims, and to ensure that the technical solution is feasible in practical applications and produces the expected beneficial technical effects.

[0023] This invention proposes a reservoir hydrological and rainfall monitoring system. By deploying island-shore monitoring stations at key control sections in the reservoir area and upstream and downstream basins, a comprehensive sensing network is constructed. The monitoring stations adopt an independent island-shore structure, which adapts to the complex and varied slope terrain of mountain reservoirs. This ensures that the monitoring equipment is close to the water body to obtain accurate data, while the robust civil engineering foundation ensures the safety of the station building under flood impact. The stations integrate front-end acquisition equipment, physical observation benchmark facilities, and communication modules, forming a complete on-site sensing and transmission system. The communication network, acting as a bridge between the physical and digital worlds, is responsible for establishing stable data transmission channels with each monitoring station, ensuring that on-site data can be transmitted back to the server in real time and without loss. The software platform deployed on the server is the core brain of the entire system, containing a data verification module, an online flow measurement algorithm module, and a multi-objective optimization scheduling model. This architecture employs a collaborative model of edge computing and cloud computing. The monitoring stations' hydrological and meteorological monitoring devices and online flow monitoring systems support basic data computation on the hardware side, such as filtering and encapsulating raw signals. Meanwhile, the software platform leverages the powerful computing capabilities of the servers to support advanced simulations and command issuance. This layered processing mechanism not only reduces the concurrent load on the servers but also ensures that the front-end devices retain independent data processing and storage capabilities even when the communication network is unstable, significantly improving the system's robustness and data continuity.

[0024] In a preferred embodiment, the physical observation reference facility is specifically designed as a vertical stainless steel water gauge, custom-manufactured based on dam elevation data. This stainless steel water gauge is not a standard, universal component; rather, it is calibrated to the absolute elevation benchmark of a specific reservoir dam and is permanently fixed to the hard bedrock or reinforced concrete base of the dam and monitoring section. The advantage of this physical reference is that it provides the electronic sensors with an absolute "true value" reference that is unaffected by power failures or signal drift. During long-term operation, float-type or radar level gauges may experience zero-point drift. By periodically comparing the readings from the manually observed stainless steel water gauge with the sensor readings, on-site calibration of the electronic monitoring equipment can be achieved. This solves the technical problem of decreased data reliability under harsh environments using a single electronic monitoring method, ensuring the absolute accuracy of the water level data.

[0025] The front-end data acquisition equipment underwent deep hardware optimization, including a telemetry terminal, a tipping bucket rain gauge, a float-type water level gauge, a Doppler current meter, and a high-definition monitoring camera. The telemetry terminal incorporates a dedicated hydraulic calculation unit, distinguishing it from traditional pure transmission devices by enabling it to execute complex hydraulic formulas. The tipping bucket rain gauge accurately captures rainfall events, the float-type water level gauge provides high-precision water level data, the Doppler current meter measures water velocity using the Doppler effect, and the high-definition monitoring camera provides a visualized real-time view. This multi-sensor fusion configuration, combined with the edge computing capabilities of the telemetry terminal, enables the simultaneous acquisition and initial fusion of four-dimensional data: rainfall, water level, flow velocity, and video, laying a solid data foundation for subsequent accurate flow inversion.

[0026] The software platform adopts an advanced containerized microservice architecture, integrating IoT communication middleware, a time-series database, a hydraulics calculation service daemon, and a simulation engine. The benefits of this microservice architecture lie in the independent decoupling of various functional modules, such as data reception, algorithm calculation, and alarm push. Upgrades or failures of a single module will not affect the operation of the entire system, greatly improving system stability and maintainability. The IoT communication middleware can handle massive concurrent connections and heterogeneous protocol parsing, shielding the differences between underlying devices. The time-series database is specifically designed to store high-frequency monitoring data with timestamps, significantly improving the efficiency of writing and querying historical data, and meeting the needs of long-term backtracking analysis of hydrological data. The hydraulics calculation service daemon, as a resident background service, can quickly respond to calculation requests and call dynamic link libraries to calculate real-time flow rate and volume.

[0027] Addressing the core technical challenge of online flow measurement, the online flow measurement algorithm module incorporates and applies the water level-cross-sectional area relationship function and hydraulic parameter correction logic. Specifically, the system first uses measured river cross-sectional topographic data and employs the least squares method for polynomial fitting to construct a mathematical mapping relationship between water level and cross-sectional area. This relationship function is expressed as: ; In the above formula, Represents the cross-sectional area of ​​the water passage. Represents the real-time water level value. Representing the Order of fit coefficients, This represents the highest degree of the polynomial. Using this formula, the system can adjust the water level based on real-time changes. The current water flow area is calculated in milliseconds.

[0028] Furthermore, the system combines the representative point flow velocities collected by the front-end Doppler current meter. The average flow velocity of the cross section is calculated by correcting the logic using hydraulic parameters. The correction logic incorporates parameters such as Manning roughness coefficient, hydraulic radius, and water surface gradient, using a velocity distribution model to correct the single-point velocity to the average velocity across the entire cross-section. Ultimately, the real-time flow rate... The calculation formula is as follows: ; This method, which uses real-time water level and flow velocity to dynamically invert flow rate, overcomes the shortcomings of traditional water level-flow relationship curves, which are prone to failure during riverbed deformation or flood seasons. It enables automatic flow rate measurement across the entire range, significantly improving the timeliness and accuracy of hydrological measurements.

[0029] A multi-objective optimization scheduling model is key to achieving intelligent reservoir scheduling. It is configured to receive real-time hydrological and meteorological data and output digital scheduling instruction packages. This model constructs an objective function encompassing flood control safety and power generation benefits. Under constraints such as dam safety water level and downstream river safety discharge, it utilizes genetic algorithms or particle swarm optimization to perform optimization calculations. The final generated digital scheduling instruction package not only includes macro-level scheduling strategies but also quantifies specific parameters such as the number of gates to be opened, the target opening degree, and the execution timestamp. The beneficial effect of this technical solution is that it transforms complex scheduling decisions into standardized instructions that machines can directly read and execute, achieving closed-loop control from "data monitoring" to "decision execution." This effectively avoids the arbitrariness and lag of manual experience-based scheduling, maximizing water resource utilization efficiency while ensuring reservoir flood control safety.

[0030] Example 2 Further explanation in conjunction with Example 1, such as Figure 1 As shown, a method for monitoring water and rainfall conditions in a reservoir includes: S1. Constructing island-shore monitoring stations at key nodes in the reservoir and watershed, and installing customized stainless steel water gauges on the observation sections to establish a physical water level observation benchmark; S2. Deploy front-end data acquisition devices within the monitoring site. After the front-end data acquisition devices are powered on and initialized, they establish a long-term connection channel with the server through the built-in communication protocol stack. S3. The front-end acquisition device synchronously acquires rainfall data, water level data, and flow velocity data at a preset frequency, and calls the online flow measurement algorithm module to calculate and encapsulate the acquired raw signals to generate monitoring data packets; S4. The server-side software platform receives monitoring data packets, parses and stores them in the standard hydrological database, and simultaneously calls the verification module to perform linear regression analysis and parameter correction on the automatically collected water level data using the manually observed stainless steel water gauge data. S5. Based on the corrected hydrological and rainfall data and online flow measurement data, the software platform drives a multi-objective optimization scheduling model to perform simulation and generate the optimal scheduling scheme, which is then issued for execution. This embodiment discloses a specific implementation process of a reservoir hydrological and rainfall monitoring method. This method achieves closed-loop control of reservoir operation and management through five closely linked steps: physical benchmark construction, intelligent data acquisition and transmission, edge computing, cloud verification, and simulation scheduling. The following details each step.

[0031] In step S1, island-shore monitoring stations are constructed at key nodes in the reservoir and watershed, and customized stainless steel water gauges are installed at the observation sections to establish a physical water level observation benchmark. Specifically, the island-shore monitoring stations are located on stable bank slopes in front of the reservoir dam or at the watershed control section, with independent station foundations constructed using reinforced concrete to ensure structural stability under flood erosion. The physical observation benchmark facility uses a vertical stainless steel water gauge customized based on the dam's absolute elevation data. The gauge's scale is strictly consistent with the dam's elevation system and is permanently fixed to the bedrock of the dam body or the monitoring section. The beneficial effects of this step are that by constructing stable island-shore stations, the problems of difficult equipment installation and susceptibility to damage in complex mountainous terrain are solved; at the same time, the introduction of customized stainless steel water gauges as a physical truth benchmark provides an unalterable physical reference for subsequent electronic sensor calibration, effectively eliminating systematic errors caused by geological subsidence or equipment aging during long-term operation.

[0032] In step S2, front-end data acquisition devices are deployed within the monitoring site. After power-on initialization, these devices establish a long-term connection with the server via their built-in communication protocol stack. The front-end data acquisition devices include telemetry terminals and various sensors. Upon power-on, the devices first execute a self-test procedure; if confirmed to be in normal condition, they read the server address from the configuration file. Using their built-in IoT communication protocol stack, the devices initiate a handshake request to the server. After establishing a TCP or UDP connection, a heartbeat keep-alive mechanism is activated, whereby the devices send extremely small data packets to the server at preset time intervals to maintain link activity. The advantage of this long-term connection technology is that it avoids the network resources and time delays consumed by frequent connection establishment and disconnection, ensuring that alarm data can be uploaded in milliseconds during emergencies such as sudden floods, significantly improving the system's real-time response capability and communication stability.

[0033] In step S3, the front-end acquisition device synchronously collects rainfall data, water level data, and flow velocity data at a preset frequency, and calls the online flow measurement algorithm module to calculate and encapsulate the collected raw signals to generate a monitoring data packet. This step embodies the concept of edge computing. The front-end device triggers the sampling actions of the rain gauge, water level gauge, and flow velocity meter simultaneously based on a unified clock signal, ensuring strict alignment of multi-source data in the time dimension. Subsequently, the computing unit inside the telemetry terminal calls the online flow measurement algorithm. This algorithm first calculates the water flow area based on a pre-mapped model of the relationship between water level and cross-sectional area written into the device. Let the real-time collected water level value be... Then the cross-sectional area of ​​the water passage The calculation formula is as follows: ; In the above formula, Representing the Order of fit coefficients, Representing the highest degree of the polynomial, this formula accurately describes the cross-sectional morphology of irregular river channels through polynomial fitting.

[0034] Next, the equipment uses a Doppler current meter to measure the flow velocity at a representative point. The average flow velocity of the cross section was calculated by combining hydraulic correction factors. Then calculate the real-time traffic. The calculation formula is as follows: ; After completing the calculation, the device encapsulates the raw data and calculation results into a monitoring data packet with a checksum. The advantage of this step is that it directly outputs flow data through front-end edge computing, reducing the computational burden on the server side. Even in the event of a network outage, the front-end device can independently complete flow accumulation and storage, ensuring the continuity and integrity of hydrological data.

[0035] In step S4, the server-side software platform receives monitoring data packets, parses and stores them in the standard hydrological database. Simultaneously, it calls the verification module to perform linear regression analysis and parameter correction on the automatically collected water level data using manually observed stainless steel water gauge data. The software platform's data receiving middleware performs protocol parsing and validity verification on the uploaded data packets, and writes them to the time-series database after successful verification. The verification module is a crucial step in ensuring data accuracy; it periodically receives manually observed stainless steel water gauge values ​​entered by management personnel. And extract the water level value automatically collected by the sensor at the same time. The module uses linear regression analysis to compare the two sets of data. If a systematic bias is found, a correction equation is established: ; In the above formula, This is the corrected final water level value. This is the slope correction factor. This is the intercept offset. The system dynamically updates this based on historical comparison data. and The data is then distributed to the front-end equipment. The beneficial effect of this technology is that it realizes a dual "physical-electronic" verification mechanism, effectively solving the technical problem that electronic water level gauges are prone to zero-point drift in harsh environments and are difficult to detect, thus greatly improving the confidence of the monitoring data.

[0036] In step S5, the software platform drives a multi-objective optimization scheduling model to perform simulations based on the corrected hydrological and rainfall data and online flow measurement data, generating an optimal scheduling scheme and issuing it for execution. The software platform inputs verified, accurate real-time data and meteorological forecast data into the simulation engine. The multi-objective optimization scheduling model constructs a mathematical model that includes flood control safety objectives and power generation benefit objectives. Under the premise of meeting the dam's safe water level constraints and the downstream river channel's safe discharge constraints, an intelligent optimization algorithm searches for the optimal solution in the solution space. The scheduling scheme finally generated by the system is not a vague suggestion, but a specific digital instruction, including the specific number of gates to be opened, the target opening degree, and the start time of the execution action. The beneficial effect of this step is that it transforms experience-based manual scheduling into data- and model-based intelligent scheduling. While ensuring the safety of the reservoir during the flood season, it reduces unnecessary water wastage through refined control, significantly improving the power generation efficiency of cascade power stations and achieving optimal allocation of water resources.

[0037] In the preferred embodiment, the deployment and communication connection between the front-end acquisition device and the server in step S2 specifically includes: S21. Configure high-performance data processing servers and application server clusters on the server side, deploy Linux operating system and Docker containerized runtime environment, and build a load-balanced high-availability service architecture. S22. Deploy IoT communication middleware in a container environment, configure TCP / UDP listening ports and message parsing engine, define a private communication protocol stack, and the data frame structure of the communication protocol stack includes frame header identifier, device unique ID, function control code, encrypted data field and CRC cyclic redundancy check code. S23. Configure the server's public IP address or domain name in the embedded system of the front-end acquisition device, and set the reconnection trigger threshold and heartbeat packet sending interval. S24. The front-end acquisition device initiates a connection request to the server. The communication middleware responds with a handshake signal and establishes a TCP long connection. Subsequently, the front-end acquisition device periodically sends heartbeat packets containing the device's operating status. The communication middleware maintains a dedicated data transmission channel through a heartbeat keep-alive mechanism. If no heartbeat packet is received within a preset time, the device is determined to be offline and a reconnection waiting logic is triggered.

[0038] In step S21, a high-performance data processing server and application server cluster are configured on the server side, and a Linux operating system and Docker containerized runtime environment are deployed to build a load-balanced, highly available service architecture. Specifically, a high-performance server with multi-core processors and large-capacity memory is selected as the hardware foundation, and a Linux operating system is installed to provide stable underlying support. Based on this, Docker container technology is introduced to encapsulate different services such as data reception, processing, storage, and web applications in independent containers. Simultaneously, a load balancer is built using Nginx or HAProxy to distribute concurrent requests from front-end devices to different nodes in the application server cluster. The beneficial effects of implementing this step are that containerized deployment decouples the application from the underlying environment, greatly simplifying the system migration and expansion process; load balancing and cluster architecture avoid single points of failure, meaning that when a server fails, traffic automatically switches to other normal nodes, ensuring the reservoir monitoring system operates 24 / 7 under high concurrency access during the flood season.

[0039] In step S22, an IoT communication middleware is deployed in a container environment, configuring TCP and UDP listening ports and a packet parsing engine, and defining a private communication protocol stack. The data frame structure of the communication protocol stack includes a frame header identifier, a unique device ID, a function control code, an encrypted data field, and a CRC cyclic redundancy check code. Specifically, the IoT communication middleware, as the core component for data access, listens to a designated network port to receive data packets from the telemetry terminal. The system defines a dedicated private binary communication protocol stack, where the frame header identifier is used by the receiving end to identify the start position of the data packet and prevent packet fragmentation; the unique device ID is used to accurately identify the data source among a large number of devices; the function control code is used to distinguish service types such as reported data, heartbeat packets, and configuration commands; the encrypted data field uses AES or national cryptographic algorithms to encrypt and encapsulate core monitoring data; and the CRC cyclic redundancy check code is used to detect whether bit flips or packet loss occur during data transmission. The benefits of implementing this step are that the private protocol stack is more compact and efficient than the general text protocol, reducing network traffic consumption; the encryption and verification mechanisms effectively prevent data from being maliciously tampered with or from being corrupted by network noise, ensuring the security and integrity of hydrological data.

[0040] In step S23, the public IP address or domain name of the server is configured in the embedded system of the front-end acquisition device, and the network reconnection trigger threshold and heartbeat packet sending interval are set. During implementation, technicians access the configuration interface of the telemetry terminal via serial port or Bluetooth interface and write the fixed public IP address of the server or the domain name resolved by dynamic DNS into the non-volatile memory. Simultaneously, the network reconnection trigger threshold is set, meaning that when the number of consecutive network unavailability detections reaches a set value, the device automatically restarts the communication module; the heartbeat packet sending interval is set, meaning that when there is no business data transmission, the device sends a status packet to the server every fixed number of seconds. The beneficial effect of implementing this step is that, through preset network parameters and self-healing logic, the front-end device is endowed with network adaptability in unattended environments, enabling it to adapt to the unstable communication signal quality in mountainous areas and reducing the frequency of manual on-site maintenance.

[0041] In step S24, the front-end acquisition device initiates a connection request to the server. The communication middleware responds with a handshake signal and establishes a TCP long connection. Subsequently, the front-end acquisition device periodically sends heartbeat packets containing the device's operating status. The communication middleware maintains a dedicated data transmission channel through a heartbeat keep-alive mechanism. If no heartbeat packet is received within a preset time, the device is determined to be offline and reconnection waiting logic is triggered. In specific operation, the front-end device actively initiates a three-way handshake with the server to establish a full-duplex TCP long connection channel. After the connection is established, the channel remains open. The device periodically sends heartbeat packets at set intervals, which contain status information such as voltage and signal strength. The communication middleware on the server side monitors these heartbeat signals in real time. If no heartbeat packet is received from a device within a preset timeout period, the connection resource is immediately released, the device status is updated to offline, and the system waits for the device's next reconnection request. The benefits of implementing this step are that the TCP long connection mechanism avoids the network overhead and latency caused by frequent connection establishment and disconnection, and realizes millisecond-level real-time data transmission; the heartbeat keep-alive mechanism can quickly detect link anomalies, ensuring the system's accurate control over the online status of devices, and providing a solid communication guarantee for real-time flood early warning.

[0042] In the preferred embodiment, the calculation steps in step S3, which call the online flow measurement algorithm module, include: S31. Constructing the water level-cross-sectional area relationship function The least squares method was used to perform polynomial fitting on the measured cross-sectional topographic data, and the generated function formula is as follows: ; in, The cross-sectional area of ​​the water passage. This is the real-time water level value. These are the fitting coefficients. It is the highest power of the polynomial; The function formula and fitting coefficients are compiled into a binary calculation configuration file and written to the storage unit of the front-end acquisition device; S32. Obtain real-time water level values ​​using a float-type water level gauge. Real-time flow velocity values ​​at representative points are obtained using a flow meter. ; S33, The computing unit of the front-end acquisition device reads the configuration file and displays the real-time water level value. Substitute the values ​​into the function formula to calculate the current cross-sectional area of ​​the water flow. ; S34. The computer deployment steps for calling the hydraulic parameter model are as follows: A hydraulics calculation service daemon is deployed in the embedded Linux system of the front-end acquisition device. The daemon loads a pre-compiled dynamic link library of hydraulics algorithms. The dynamic link library internally encapsulates the Manning formula or the logarithmic velocity distribution formula, and reads the Manning roughness coefficient of the river channel from the configuration file at runtime. Hydraulic radius and water surface gradient ; The main program transmits the real-time flow rate value through an inter-process communication mechanism. The auxiliary parameters are passed to the dynamic link library, the library function performs iterative calculations, and outputs the average flow velocity of the cross section. ; S35, through formula Complete real-time traffic Perform multiplication operations and upload the calculation results in a package.

[0043] This embodiment details the specific calculation process and computer implementation steps of calling the online flow measurement algorithm module in step S3. This process achieves real-time and accurate inversion of traffic data through mathematical modeling, edge computing, and modular software deployment.

[0044] In step S31, the water level-cross-sectional area relationship function is constructed. This forms the mathematical foundation for automated flow measurement. In practice, technicians first conduct detailed mapping of the riverbed topography at the monitoring section, obtaining a series of discrete points corresponding to water levels and areas. Subsequently, the least squares method is used to perform polynomial fitting on these measured data. This algorithm finds the best function match for the data by minimizing the sum of squared errors, thereby generating a function formula that accurately describes the irregular riverbed cross-section. ; In the above formula, Represents the cross-sectional area of ​​the water passage. Represents the real-time water level value. to These are the fitting coefficients for each order. The value is the highest power of the polynomial, typically chosen based on the complexity of the cross-section. To improve the reading and calculation efficiency of the embedded system, this function formula and all fitting coefficients are not stored in text format, but are compiled into a binary calculation configuration file and written to the non-volatile storage unit of the front-end acquisition device. The advantages of using binary storage are that it reduces the overhead of the embedded processor parsing ASCII codes, significantly improves data loading speed, and also increases the security of the configuration file, preventing parameters from being accidentally tampered with.

[0045] In steps S32 and S33, the system enters the real-time calculation phase. The front-end equipment obtains the real-time water level value through a high-precision float-type water level gauge. Real-time flow velocity values ​​at representative points are obtained using Doppler current meters or radar current meters installed underwater. The computing unit of the front-end acquisition device directly reads the aforementioned binary configuration file and processes the real-time water level value. Substituting the independent variable into the polynomial function formula, the cross-sectional area of ​​the water passage at the current moment is quickly calculated using the hardware floating-point arithmetic unit. This process is completed entirely locally on the front end, enabling real-time data processing at the edge.

[0046] In step S34, the computer deployment step of calling the hydraulic parameter model demonstrates the advanced nature of the software architecture. A dedicated hydraulic calculation service daemon is deployed in the embedded Linux operating system of the front-end acquisition device. This daemon runs automatically at system startup and resides in memory, responsible for loading pre-compiled hydraulic algorithm dynamic link libraries. The dynamic link libraries encapsulate core hydraulic algorithms such as the Manning formula or the logarithmic velocity distribution formula. This design, which encapsulates the algorithm logic in independent library files, allows algorithm upgrades to be performed without reflashing the entire system firmware; only the library files need to be updated. During program execution, the Manning roughness coefficient of the river channel is first read from the configuration file. Hydraulic radius and water surface gradient Environmental parameters, etc. Then, the main program uses inter-process communication mechanisms to transmit the real-time collected representative point flow rate values. The aforementioned auxiliary parameters are passed to the dynamic link library. The library function performs iterative calculations based on the velocity distribution model, eliminating the representativeness error of single-point measurements, and outputting the cross-sectional average velocity that represents the flow characteristics of the entire cross-section. .

[0047] In step S35, the system performs the final flow synthesis calculation. The calculation unit uses the multiplication formula: ; Multiply the average flow velocity across the cross-section by the cross-sectional area to obtain the real-time flow rate. Ultimately, the device will process the traffic data. The data, along with the original water level and flow velocity data, is packaged and uploaded to the server via a communication module. The beneficial effect of this implementation method is that, through a rigorous mathematical model and efficient computer software architecture, it enables the scientific calculation of flow velocity from a single point to the average flow velocity across a cross section. This solves the flow measurement error problem caused by irregular river cross sections or uneven flow velocity distribution in traditional methods, greatly improving the accuracy and practicality of hydrological monitoring data.

[0048] In the preferred embodiment, the data processing and verification deployment of the software platform in step S4 includes: S41. Deploy a relational database and a time-series database on the database server. The relational database stores basic site information, and the time-series database stores high-frequency monitoring data. S42. Develop a data cleaning service module. The data cleaning service module resides in memory and scans the data entering the database in real time. S43. Execute outlier removal logic. When outlier data is detected within a short period of time... When the jump amplitude exceeds the preset physical limit threshold, the data is automatically marked as a suspected anomaly and a video surveillance capture command is triggered. S44. Execute the dynamic comparison logic, periodically input the manual observation values ​​of the stainless steel water gauge, and calculate the deviation between the manual observation values ​​and the automatic values ​​of the float-type water level gauge. ,like If the value remains above the calibration threshold, sensor zero-point drift correction parameters are automatically generated and sent to the front-end acquisition device via the communication network for parameter updates.

[0049] This embodiment details the data processing and verification mechanism of the server-side software platform. This part ensures the accuracy of the data entering the database and the long-term stability of the monitoring system through four core steps: dual-mode database architecture, memory-level data cleaning, physical limit logic judgment, and remote dynamic calibration.

[0050] In step S41, a relational database and a time-series database are deployed on the database server. The relational database stores basic site information, while the time-series database stores high-frequency monitoring data. In practice, the software platform backend adopts a hybrid storage architecture. The relational database (such as MySQL or PostgreSQL) stores static attribute data of the monitoring sites, including site name, geographic latitude and longitude coordinates, equipment model, zero-point elevation of the stainless steel water gauge, and management personnel information. This data is highly structured and has a low frequency of change, making it suitable for relational database management. Simultaneously, the time-series database (such as InfluxDB or TDengine) is specifically used to store high-frequency monitoring data streams uploaded by front-end devices, including rainfall, water level, flow velocity, and flow rate data with precise millisecond-level timestamps. The advantage of this dual-mode database architecture is that it fully leverages the high write throughput and high compression ratio advantages of time-series databases when processing massive continuous time-series data, solving the IO bottleneck problem caused by traditional relational databases when facing tens of thousands of concurrent writes per second, and significantly improving the query efficiency of historical hydrological data and the overall system response speed.

[0051] In step S42, a data cleaning service module is developed. This module resides in memory and scans the incoming data in real time. This module runs as an independent service process in the server's random access memory (RAM), utilizing the high-speed read / write capabilities of memory to construct a data buffer. When the communication middleware receives and parses a data packet, the data first enters the memory buffer, and the cleaning service module uses a sliding window algorithm to scan the data stream in real time. The advantage of this step is that by placing the data cleaning process before database writing and using in-memory computation instead of disk I / O operations, millisecond-level interception of abnormal data is achieved, preventing dirty data from polluting the database and ensuring the purity of the input data for subsequent simulation models.

[0052] In step S43, outlier removal logic is executed when water level data within a short period of time is detected. When the fluctuation range exceeds a preset physical limit threshold, the data is automatically marked as potentially abnormal, and a video surveillance capture command is triggered. The specific logic is as follows: the system calculates the current water level value. Compared with the previous effective water level value The absolute value of the difference, if this absolute value is greater than the set physical limit threshold. If so, the data is considered abnormal. Physical limit threshold. The settings are based on the dam's discharge capacity and the maximum possible confluence velocity of the watershed, representing the physical upper limit of water level changes in nature. Once an anomaly is triggered, the system not only tags the data point as abnormal in the database but also immediately calls the video monitoring interface to control the on-site camera to capture images or record short videos of the corresponding water level observation section. The beneficial effect of this technology is that it establishes a linkage mechanism of "data anomaly-driven visual verification," which can not only automatically filter out false jump data caused by sensor circuit interference or floating object impacts but also provide managers with on-site visual evidence of anomalies, greatly reducing the false alarm rate and improving troubleshooting efficiency.

[0053] In step S44, the dynamic comparison logic is executed, the manual observation values ​​of the stainless steel water gauge are periodically entered, and the deviation between the manual observation values ​​and the automatic values ​​of the float-type water level gauge is calculated. ,like If the water level remains above the calibration threshold, sensor zero-point drift correction parameters are automatically generated and transmitted to the front-end data acquisition equipment via the communication network for parameter updates. In practice, managers periodically (e.g., weekly or monthly) input visually estimated water level values ​​from the stainless steel water gauge using a mobile app. The system automatically retrieves the most recent sensor data collected around the time of entry. And calculate the absolute value of the deviation between the two. The system will perform statistical analysis on the deviations from the most recent comparisons. If it finds... Exceeding the set calibration threshold multiple times in a row If the deviation direction is consistent, it is determined that the sensor has experienced zero-point drift. At this point, the system automatically calculates new correction parameters and sends them to the front-end acquisition device through the long-term connection established in step S2. Upon receiving the instruction, the front-end device updates the calculated parameters in its local storage unit. The beneficial effect of implementing this step is that it replaces high-cost manual on-site hardware calibration with low-cost software calibration, solving the industry pain point of decreased accuracy of electronic sensors after long-term operation. It achieves remote, full-lifecycle accuracy maintenance of the equipment without interrupting monitoring operations, ensuring the consistency between monitoring data and the true physical value.

[0054] In the preferred solution, the software platform deploys a "four-pre" functional module, and the module's operation steps include: S51. Forecast: Integrate gridded rainfall forecast data released by meteorological departments and combine it with watershed runoff generation and confluence models to calculate the inflow process in the future period; S52. Early Warning: Set graded early warning thresholds. The system monitors water level and rainfall in real time. Once the threshold is reached, an alarm message is immediately pushed through the SMS gateway and mobile application. S53. Pre-simulation: Using the simulation engine, load the reservoir scheduling rules, input the predicted inflow, and simulate the reservoir water level changes and downstream discharge process under different scheduling schemes. S54. Contingency Plan: Based on the results of the pre-rehearsal, intelligently recommend scheduling schemes and generate a scheduling execution plan table that includes the number of gate openings, opening height, and discharge duration.

[0055] This embodiment describes in detail the specific operation steps and technical implementation details of the "Four Predictions" functional module in the software platform. This module achieves intelligent control of the entire chain from meteorological sources to engineering scheduling through the organic combination of four dimensions: forecasting, early warning, rehearsal, and contingency plan.

[0056] In step S51, the forecasting function is executed. The system integrates gridded rainfall forecast data released by the meteorological department and, combined with the watershed runoff generation and concentration model, calculates the inflow process for the future period. Specifically, the software platform periodically retrieves high-resolution gridded numerical weather forecast data released by the meteorological department through a standard API interface. This data includes rainfall distribution for the next 72 hours or longer. The system is equipped with distributed hydrological models or watershed runoff generation and concentration models such as the Xin'anjiang model. These models divide the watershed into several calculation units, convert the gridded rainfall data into surface runoff based on underlying surface conditions and soil moisture content, and, through river runoff calculations, predict the future inflow process curve. The beneficial effect of implementing this step is that it extends the perspective of reservoir scheduling from current measured data to future forecast data, significantly prolonging the forecast period for flood control scheduling. This allows managers to take advantage of the time difference between rainfall and flood peaks to release water from reservoirs in advance and stagger peak flows, thus realizing a shift from passive defense to proactive response.

[0057] In step S52, the early warning function is executed. The system sets tiered early warning thresholds and monitors water levels and rainfall in real time. Once a threshold is reached, an alarm message is immediately pushed through the SMS gateway and mobile application. A real-time monitoring service runs in the system background, scanning the latest monitoring data at millisecond intervals and comparing it with the preset four-level early warning thresholds in the database. The early warning thresholds include the flood control limit water level, warning water level, guaranteed water level, and short-duration heavy rainfall intensity threshold. When the real-time water level... Or the cumulative rainfall exceeds a certain threshold level. When the system triggers an event response mechanism, it automatically generates an alarm message and simultaneously sends the alarm information to the mobile terminals of all relevant responsible persons through the integrated SMS gateway and mobile APP message push interface. The beneficial effects of implementing this step are that automated machine monitoring replaces manual screen watching, eliminating the risk of missed reports due to human negligence. At the same time, the multi-channel real-time push mechanism ensures that emergency information can be obtained by management as soon as possible, even during special periods such as nighttime or holidays, guaranteeing the timeliness of emergency response.

[0058] In step S53, a pre-simulation function is executed. The simulation engine loads the reservoir scheduling rules, inputs the predicted inflow, and simulates reservoir water level changes and downstream discharge processes under different scheduling schemes. The simulation engine is a digital twin component built based on the water balance equation and hydraulic discharge formula. The engine loads the approved reservoir scheduling procedures and historical flood scheduling experience, combined with the predicted inflow calculated in step S51. Multiple scheduling schemes are simulated in parallel in a virtual environment. For example, the reservoir water level is simulated for several hours or even days under different strategies, such as keeping the current gate opening unchanged, increasing the discharge, or decreasing the discharge. The change curve and the flow evolution process in the downstream river channel. The beneficial effect of implementing this step is that it provides a low-cost, zero-risk trial-and-error mechanism, allowing decision-makers to intuitively see the possible consequences of various scheduling schemes before actually operating the gates. This effectively identifies and avoids the risks of dam overflow or downstream man-made floods caused by improper scheduling, thus achieving scientific decision-making in flood control scheduling.

[0059] In step S54, the contingency plan function is executed. Based on the pre-simulation results, an intelligent scheduling scheme is recommended, generating a scheduling execution plan table that includes the number of gates to be opened, their opening height, and the discharge duration. The system automatically selects the optimal scheme that ensures dam safety while also considering downstream interests based on the comprehensive scores of each scheme in the pre-simulation results, and translates it into engineering language. The generated scheduling execution plan table details the specific operation instructions, including the gate numbers to be operated and the target opening height. Adjust the start time of the action. and duration The beneficial effect of implementing this step is that it transforms complex decision-making logic into standardized execution work orders, directly guiding on-site operators to perform standardized operations. This avoids safety accidents caused by unclear verbal instructions or human error in emergency situations, and significantly improves the level of refined management and execution efficiency of reservoir scheduling.

[0060] In the preferred scheme, the specific steps for generating the optimal scheduling scheme in step S5 include: S91. Construct a multi-objective optimization scheduling model and define objective functions in the software platform. The objective functions include the objective of minimizing flood disaster losses and the objective of maximizing power generation benefits. S92. Set a set of constraints. The set of constraints is integrated into the rule engine of the software platform. The set of constraints includes the maximum design water level of the dam, the safe discharge of the downstream river channel, the maximum opening speed of the gate, and the power output limit of the unit. S93. The software platform calls the simulation engine and, based on the current real-time water and rainfall data and the future forecast inflow, uses a genetic algorithm or particle swarm algorithm to generate a set of multiple candidate scheduling schemes. Each candidate scheduling scheme contains different flood discharge time and flood discharge flow sequences. S94. Perform parallel simulation calculations for each candidate scheduling scheme, predict the reservoir water level change curve and downstream flooding risk after each scheme is implemented, and automatically eliminate infeasible schemes that violate the constraint set.

[0061] This embodiment details the specific implementation logic and mathematical model construction process for generating the optimal scheduling scheme in step S5. This process, through multi-objective optimization, constraint conditions, and parallel simulation screening, realizes the leap from experience-based decision-making to digital intelligent decision-making in reservoir scheduling.

[0062] In step S91, a multi-objective optimization scheduling model is constructed, and objective functions are defined in the software platform. These objective functions include minimizing flood disaster losses and maximizing power generation benefits. Specifically, the software platform's algorithm engine first establishes the objective function for minimizing flood disaster losses. This function aims to reduce the risk of downstream flooding caused by the discharge, and its mathematical expression is: ; In the above formula, Represents the total number of time periods in the scheduling cycle. represent Total flow of the downstream river during the period This represents the safe discharge threshold for downstream river channels. This formula ensures flood control safety by penalizing the portion of the flow exceeding the safe discharge limit.

[0063] At the same time, establish an objective function to maximize power generation efficiency. The aim is to increase power generation by utilizing water head under safe conditions, and its mathematical expression is: ; In the above formula, This is the overall output coefficient of the unit. for The flow rate through the turbine during the power generation period. for The unit's clean water head during a given time period.

[0064] The final integrated objective function is constructed using a weighted method, i.e. ,in and These are weighting coefficients that are dynamically adjusted according to the flood season and non-flood season. The beneficial effect of implementing this step is that it unifies the two originally contradictory demands in reservoir scheduling—flood control safety and economic benefits—into a computable mathematical model, enabling scheduling decisions to maximize the economic value of every cubic meter of water while ensuring the safety of life and property downstream.

[0065] In step S92, a set of constraints is set. This set of constraints is integrated into the rule engine of the software platform. The set of constraints includes the dam's design maximum water level constraint, the downstream river channel's safe discharge constraint, the gate's maximum opening speed constraint, and the unit's output limit constraint. These constraints constitute the feasible domain boundary of the scheduling scheme. Specifically, they include: Water level constraints: ,in For the reservoir water level, The highest permitted flood control water level; Leakage constraint: ,in Maximum flow rate variation is limited to prevent scouring or ship grounding caused by sudden changes in downstream water level; Unit constraints: This ensures that the generator set operates within a stable operating range.

[0066] The beneficial effect of implementing this step is that by digitizing the rigid regulations that were originally scattered in paper procedures into mathematical inequalities through the rule engine, it ensures that any solution generated by the machine is physically executable and legally compliant, thus avoiding the risk of irregular scheduling due to human negligence.

[0067] In step S93, the software platform invokes the simulation engine to generate multiple candidate scheduling schemes based on current real-time water and rainfall data and future forecasted inflows, using either a genetic algorithm or a particle swarm optimization algorithm. The simulation engine reads real-time inflows and the future inflow sequence reported by weather forecasts. Using this as the boundary condition, the system employs a genetic algorithm as the solver, initializing a population containing dozens of different discharge strategies. Each individual represents a candidate scheduling scheme, comprising a set of time-varying gate opening sequences and discharge flow sequences. The algorithm iteratively searches the solution space through crossover, mutation, and selection operations to find a solution that satisfies the comprehensive objective function. The optimal solution set. The beneficial effect of implementing this step is that, by leveraging the powerful global optimization capability of evolutionary algorithms, it is possible to traverse tens of thousands of scheduling combinations in a very short time and discover the optimal scheduling strategy that is difficult to detect by human experience. Especially in complex working conditions such as multiple peak floods, it can provide the best timing for operation that balances peak shifting and peak reduction.

[0068] In step S94, parallel simulation calculations are performed on each candidate scheduling scheme to predict the reservoir water level change curve and downstream inundation risk after each scheme is implemented, and infeasible schemes that violate the constraint set are automatically eliminated. The software platform utilizes the multi-core computing power of the server cluster to start multiple parallel simulation threads. Each thread loads one candidate scheduling scheme and combines the reservoir capacity curve with the water balance equation: ; Deducing the future reservoir water level changes after the implementation of this plan. The system then compares the simulation results with the set of constraints in step S92. Any scheme that results in water level exceeding limits, excessive discharge fluctuations, or unit output exceeding limits is marked as infeasible and automatically eliminated. The beneficial effect of implementing this step is that it establishes a robust logical firewall, ensuring that the final solution submitted to decision-makers is not only the theoretically optimal solution but also a safe solution that has undergone rigorous risk verification, thus realizing the principle of "calculation before execution" in scheduling decisions.

[0069] In the preferred solution, the steps for decision-making and execution to achieve the optimal solution include: S101. A multi-dimensional decision matrix algorithm is used to score the selected set of feasible solutions. The scoring weights are dynamically adjusted according to the current flood control level. During the high-risk period of flood control, the weight of the disaster loss target is automatically increased, and during the non-flood season, the weight of the power generation benefit target is automatically increased. S102. Select the scheme with the highest score as the optimal scheduling scheme. The system automatically generates a digital scheduling instruction package corresponding to the scheme. The digital scheduling instruction package includes the gate action sequence, target opening degree and execution timestamp. S103. The digital dispatch instruction package is pushed to the decision-maker's client for visualization and confirmation. After confirmation, it is sent to the reservoir's local control unit for execution through the Internet of Things communication middleware, thereby realizing closed-loop control of the dispatch scheme.

[0070] This embodiment details the specific operational process for achieving optimal solution decision-making and execution. This process solves the technical problems of difficulty in quantifying qualitative decisions and delays in scheduling command execution in traditional reservoir scheduling by using multi-dimensional quantitative scoring, standardized instruction encapsulation, and human-machine collaborative closed-loop control.

[0071] In step S101, a multi-dimensional decision matrix algorithm is used to score the selected set of feasible solutions. The scoring weights are dynamically adjusted according to the current flood control level. During periods of high flood risk, the weight of the disaster loss target is automatically increased, while during non-flood seasons, the weight of the power generation benefit target is automatically increased. In specific implementation, the software platform first normalizes each solution in the set of feasible solutions to eliminate the influence of data with different dimensions. Let the i-th solution in the set be... The flood disaster loss value for each scenario is Power generation benefit value Normalized disaster loss score The formula for handling extremely small indicators is used, meaning the smaller the value, the higher the score: ; in and These represent the maximum and minimum disaster losses in the scheme set, respectively.

[0072] Meanwhile, the normalized power generation efficiency score The formula for handling extremely large indicators is used, meaning the larger the value, the higher the score: ; Subsequently, the system automatically matches weighting coefficients based on the current water conditions. Let the disaster loss weight corresponding to the current flood control level be... The power generation benefit weight is And satisfy During the flood season when water levels are high, the system will automatically... Adjust to a larger value, such as the range of 0.8 to 0.9; automatically increase during dry seasons or when the water level is low. The final value. Overall score of each option The calculation formula is: ; The beneficial effects of implementing this step are that the standardization and quantification of scheduling decisions are achieved through mathematical modeling, and the introduction of dynamic weights enables the scheduling strategy to flexibly adapt to the core requirements of different operating cycles of the reservoir. This avoids the safety hazards caused by blindly pursuing economic benefits during the flood season, and also prevents the waste of water resources caused by excessive conservatism during the non-flood season.

[0073] In step S102, the scheme with the highest score is selected as the optimal scheduling scheme. The system automatically generates a digital scheduling instruction package corresponding to this scheme. The digital scheduling instruction package includes the gate action sequence, target opening degree, and execution timestamp. The system then performs a comprehensive evaluation of all schemes. The schemes are sorted in descending order, and the top-ranked scheme is selected. Subsequently, the scheduling engine transforms the abstract flow process line in this scheme into specific electromechanical control parameters. The system constructs a structured JSON or XML data packet, which internally defines an array of action sequences. For example, for the sequence numbered... The gate is defined as either opening or closing, and a target opening value is set. And specify the absolute timestamp for executing the action. The beneficial effect of implementing this step is that it eliminates the ambiguity and errors that may arise from relying on telephone or paper-based communication in traditional scheduling, and solidifies complex scheduling intentions into standardized machine-readable instructions, laying a data foundation for subsequent automated execution.

[0074] In step S103, the digital dispatch instruction package is pushed to the decision-maker's client for visualization and confirmation. After confirmation, it is sent to the reservoir's on-site control unit via the IoT communication middleware for execution, achieving closed-loop control of the dispatch plan. The software platform uses real-time communication technologies such as WebSocket to push the generated dispatch plan to the flood control command screen or the decision-maker's mobile terminal. The interface intuitively displays the predicted water level curve, gate action simulation animation, and expected discharge flow. The decision-maker confirms the confirmation via biometric identification or password verification using an electronic signature. After the confirmation instruction is issued, the IoT communication middleware sends the digital dispatch instruction package to the on-site control unit at the reservoir via an encrypted channel. The on-site control unit parses the instruction and drives the hydraulic gate hoist to perform the corresponding gate operation, while simultaneously feeding back data from the gate displacement sensor to the platform in real time, forming a control closed loop. The beneficial effect of implementing this step is that it establishes a dual guarantee mechanism of "intelligent recommendation plus human review," which not only leverages the advantages of artificial intelligence in complex calculations but also retains the final decision-making power of human experts at critical moments. At the same time, it achieves seamless connection from cloud decision-making to edge execution, greatly improving the response speed and execution accuracy of reservoir scheduling.

[0075] The above embodiments are merely preferred technical solutions of the present invention and should not be considered as limitations on the present invention. The scope of protection of the present invention should be defined as the technical solutions described in the claims, including equivalent substitutions of the technical features described in the claims. That is, equivalent substitutions and improvements within this scope are also within the scope of protection of the present invention.

Claims

1. A reservoir water and rainfall monitoring system, characterized in that: Island-shore monitoring stations are deployed in reservoir areas and upstream and downstream basins. The monitoring stations integrate front-end data acquisition equipment, physical observation benchmark facilities and communication modules. A communication network is established to connect with monitoring stations and create data transmission channels. The software platform, deployed on the server side, receives data packets sent by monitoring stations through a communication network, and has built-in data verification module, online flow measurement algorithm module, and multi-objective optimization scheduling model; The monitoring stations are equipped with water and rainfall monitoring devices and online flow monitoring systems that support basic data calculations on the hardware side, while the software platform supports advanced simulation and command issuance on the server side.

2. The reservoir water and rainfall monitoring system according to claim 1, characterized in that: The island-shore monitoring stations adopt an independent station building structure; The physical observation benchmark facility is a vertical stainless steel water gauge, which is custom-made based on the dam elevation data. The stainless steel water gauge is fixedly installed on the bedrock or concrete base of the dam and the monitoring section, serving as a physical benchmark for manual comparison and sensor calibration. The front-end data acquisition equipment includes a telemetry terminal, a tipping bucket rain gauge, a float-type water level gauge, a Doppler current meter, and a high-definition monitoring camera. The telemetry terminal has an embedded hydraulic calculation unit. The software platform adopts a containerized microservice architecture and integrates IoT communication middleware, time-series database, hydraulics calculation service daemon, and simulation engine.

3. The reservoir water and rainfall monitoring system according to claim 1, characterized in that: The multi-objective optimization scheduling model is configured to receive real-time water and rainfall data and meteorological forecast data, and output digital scheduling instruction packages; The online flow measurement algorithm module includes the water level-cross-sectional area relationship function and hydraulic parameter correction logic.

4. A method for monitoring reservoir water and rainfall conditions, characterized by: The method is implemented based on the system architecture of claims 1-3 and includes the following steps: S1. Construct island-shore monitoring stations at key nodes in reservoirs and watersheds, and install customized stainless steel water gauges on the observation sections to establish physical water level observation benchmarks. S2. Deploy front-end data acquisition devices within the monitoring site. After the front-end data acquisition devices are powered on and initialized, they establish a long-term connection channel with the server through the built-in communication protocol stack. S3. The front-end acquisition device synchronously acquires rainfall data, water level data, and flow velocity data at a preset frequency, and calls the online flow measurement algorithm module to calculate and encapsulate the acquired raw signals to generate monitoring data packets; S4. The server-side software platform receives monitoring data packets, parses and stores them in the standard hydrological database, and simultaneously calls the verification module to perform linear regression analysis and parameter correction on the automatically collected water level data using the manually observed stainless steel water gauge data. S5. Based on the corrected water and rainfall data and online flow measurement data, the software platform drives the multi-objective optimization scheduling model to perform simulation and generate the optimal scheduling scheme and issue it for execution.

5. The reservoir water and rainfall monitoring method according to claim 4, characterized in that: Step S2, specifically the deployment and communication connection between the front-end acquisition device and the server, includes: S21. Configure high-performance data processing servers and application server clusters on the server side, deploy Linux operating system and Docker containerized runtime environment, and build a load-balanced high-availability service architecture. S22. Deploy IoT communication middleware in a container environment, configure TCP / UDP listening ports and message parsing engine, define a private communication protocol stack, and the data frame structure of the communication protocol stack includes frame header identifier, device unique ID, function control code, encrypted data field and CRC cyclic redundancy check code. S23. Configure the server's public IP address or domain name in the embedded system of the front-end acquisition device, and set the reconnection trigger threshold and heartbeat packet sending interval. S24. The front-end acquisition device initiates a connection request to the server. The communication middleware responds with a handshake signal and establishes a TCP long connection. Subsequently, the front-end acquisition device periodically sends heartbeat packets containing the device's operating status. The communication middleware maintains a dedicated data transmission channel through a heartbeat keep-alive mechanism. If no heartbeat packet is received within a preset time, the device is determined to be offline and a reconnection waiting logic is triggered.

6. The reservoir water and rainfall monitoring method according to claim 4, characterized in that: The calculation steps in step S3, which call the online flow measurement algorithm module, include: S31. Constructing the water level-cross-sectional area relationship function The least squares method was used to perform polynomial fitting on the measured cross-sectional topographic data, and the generated function formula is as follows: ； in, The cross-sectional area of ​​the water passage. This is the real-time water level value. These are the fitting coefficients. It is the highest power of the polynomial; The function formula and fitting coefficients are compiled into a binary calculation configuration file and written to the storage unit of the front-end acquisition device; S32. Obtain real-time water level values ​​using a float-type water level gauge. Real-time flow velocity values ​​at representative points are obtained using a flow meter. ; S33, The computing unit of the front-end acquisition device reads the configuration file and displays the real-time water level value. Substitute the values ​​into the function formula to calculate the current cross-sectional area of ​​the water flow. ; S34. The computer deployment steps for calling the hydraulic parameter model are as follows: A hydraulics calculation service daemon is deployed in the embedded Linux system of the front-end acquisition device. The daemon loads a pre-compiled dynamic link library of hydraulics algorithms. The dynamic link library internally encapsulates the Manning formula or the logarithmic velocity distribution formula, and reads the Manning roughness coefficient of the river channel from the configuration file at runtime. Hydraulic radius and water surface gradient ; The main program transmits the real-time flow rate value through an inter-process communication mechanism. The auxiliary parameters are passed to the dynamic link library, the library function performs iterative calculations, and outputs the average flow velocity of the cross section. ; S35, through formula Complete real-time traffic Perform multiplication operations and upload the calculation results in a package.

7. The reservoir water and rainfall monitoring method according to claim 4, characterized in that: Step S4, data processing and verification deployment of the software platform, includes: S41. Deploy a relational database and a time-series database on the database server. The relational database stores basic site information, and the time-series database stores high-frequency monitoring data. S42. Develop a data cleaning service module. The data cleaning service module resides in memory and scans the data entering the database in real time. S43. Execute outlier removal logic. When outlier data is detected within a short period of time... When the jump amplitude exceeds the preset physical limit threshold, the data is automatically marked as a suspected anomaly and a video surveillance capture command is triggered. S44. Execute the dynamic comparison logic, periodically input the manual observation values ​​of the stainless steel water gauge, and calculate the deviation between the manual observation values ​​and the automatic values ​​of the float-type water level gauge. ,like If the value remains above the calibration threshold, sensor zero-point drift correction parameters are automatically generated and sent to the front-end acquisition device via the communication network for parameter updates.

8. The reservoir water and rainfall monitoring method according to claim 4, characterized in that: The software platform deploys a "four-prevention" functional module, and the module's operation steps include: S51. Forecast: Integrate gridded rainfall forecast data released by meteorological departments and combine it with watershed runoff generation and confluence models to calculate the inflow process in the future period; S52. Early Warning: Set graded early warning thresholds. The system monitors water level and rainfall in real time. Once the threshold is reached, an alarm message is immediately pushed through the SMS gateway and mobile application. S53. Pre-simulation: Using the simulation engine, load the reservoir scheduling rules, input the predicted inflow, and simulate the reservoir water level changes and downstream discharge process under different scheduling schemes. S54. Contingency Plan: Based on the results of the pre-rehearsal, intelligently recommend scheduling schemes and generate a scheduling execution plan table that includes the number of gate openings, opening height, and discharge duration.

9. The reservoir water and rainfall monitoring method according to claim 8, characterized in that: The specific steps in step S5 for generating the optimal scheduling scheme include: S91. Construct a multi-objective optimization scheduling model and define objective functions in the software platform. The objective functions include the objective of minimizing flood disaster losses and the objective of maximizing power generation benefits. S92. Set a set of constraints. The set of constraints is integrated into the rule engine of the software platform. The set of constraints includes the maximum design water level of the dam, the safe discharge of the downstream river channel, the maximum opening speed of the gate, and the power output limit of the unit. S93. The software platform calls the simulation engine and, based on the current real-time water and rainfall data and the future forecast inflow, uses a genetic algorithm or particle swarm algorithm to generate a set of multiple candidate scheduling schemes. Each candidate scheduling scheme contains different flood discharge time and flood discharge flow sequences. S94. Perform parallel simulation calculations for each candidate scheduling scheme, predict the reservoir water level change curve and downstream flooding risk after each scheme is implemented, and automatically eliminate infeasible schemes that violate the constraint set.

10. The reservoir water and rainfall monitoring method according to claim 9, characterized in that: The steps to achieve optimal decision-making and execution include: S101. A multi-dimensional decision matrix algorithm is used to score the selected set of feasible solutions. The scoring weights are dynamically adjusted according to the current flood control level. During the high-risk period of flood control, the weight of the disaster loss target is automatically increased, and during the non-flood season, the weight of the power generation benefit target is automatically increased. S102. Select the scheme with the highest score as the optimal scheduling scheme. The system automatically generates a digital scheduling instruction package corresponding to the scheme. The digital scheduling instruction package includes the gate action sequence, target opening degree and execution timestamp. S103. The digital dispatch instruction package is pushed to the decision-maker's client for visualization and confirmation. After confirmation, it is sent to the reservoir's local control unit for execution through the Internet of Things communication middleware, thereby realizing closed-loop control of the dispatch scheme.

Images