Hydrological intelligent display instrument

Abstract

The hydrological intelligent display instrument comprises a host, a touch display screen, a water level gauge and a mounting column, a three-layer intelligent hydrological system is pre-installed in the host, and a sensing and data layer fuses multi-source data and constructs a decision data foundation; the intelligent and calculation layer carries out rapid deduction and quantitative analysis by using a calibrated hydrodynamic model; and the interaction and execution layer provides a multi-scheme comparison view and assists on-site accurate execution through an AR technology. According to the invention, the fundamental transformation from artificial experience research and judgment to man-machine cooperation intelligent rehearsal is realized.

Description

Technical Field

[0001] This invention belongs to the field of hydrological information processing technology, and more specifically, relates to a hydrological intelligent display device. Background Technology

[0002] At the flood control command center, dispatchers need to obtain rainfall forecasts from the meteorological department, collect water level and flow data from scattered hydrological stations, and check the status of gates through the engineering monitoring system. This information is usually presented in the form of paper reports, spreadsheets, or independent system interfaces, and is in a scattered and heterogeneous state. Decision-making relies on senior experts to combine their personal experience to comprehensively analyze the limited information and formulate dispatch plans through meetings and consultations. On-site personnel then operate according to written instructions or drawings.

[0003] The existing technology suffers from fragmented and delayed hydrological information, relying on manual collection and comparison, resulting in slow decision-making. Dispatch plans depend on individual expert experience, making the judgment process highly subjective. The potential consequences of different plans cannot be quickly and accurately quantified and simulated. Furthermore, the process from information gathering, consultation and analysis, and instruction formulation to on-site execution involves numerous steps. In particular, the simulation plan is separated from the physical world; on-site personnel interpret written instructions based on experience, which can easily lead to biases, causing a disconnect between decision-making intent and on-site execution, increasing operational risks and time delays.

[0004] In recent years, although some technologies have introduced single data monitoring or simple hydrological models—such as large data dashboards that only achieve centralized display of information without intelligent fusion and event correlation—offline hydrological model calculations are too time-consuming to support rapid comparison of multiple options, and the issuance and execution of decision-making instructions still rely on traditional methods. Therefore, existing technologies have failed to achieve a fundamental shift from manual judgment to a collaborative pre-simulation mode based on multi-source fusion information and dynamic simulation. Summary of the Invention

[0005] This invention provides a hydrological intelligent display instrument that solves the problem of how to transform basin flood control scheduling decisions from manual judgment relying on scattered information and personal experience to a collaborative pre-simulation mode based on multi-source fusion information and dynamic simulation, so as to significantly shorten the decision-making chain, quantify the consequences of decisions, and improve the timeliness and scientific nature of scheduling plans.

[0006] In view of the above problems, the technical solution proposed by the present invention is as follows:

[0007] This invention provides a hydrological intelligent display instrument, including a main unit, a touch screen, a water level gauge, and a mounting column for installing the touch screen at a hydrological monitoring station. The main unit is pre-installed with an intelligent hydrological system for processing water level gauge monitoring data. The intelligent hydrological system includes a sensing and data layer. The sensing and data layer collects meteorological data, hydrological element data, and engineering monitoring data, and preprocesses and labels the data to build a decision-making data foundation.

[0008] The intelligence and computing layer, based on the data foundation of the perception and data layer, responds to at least two scheduling schemes input by the user, calls the calibrated hydrodynamic model to perform parallel simulations, and generates the pre-simulation results of each scheme;

[0009] The interaction and execution layer is used to provide a synchronous comparison and display of the results of multiple schemes in the pre-drill, and to use AR technology to overlay the instructions of the selected scheme onto the real scene to assist staff in performing precise operations.

[0010] As a preferred technical solution of the present invention, the sensing and data layer includes a data access module for connecting the meteorological forecasting system, the instrument for monitoring hydrological elements, and the water conservancy project monitoring system, including a meteorological data access unit, a hydrological element data acquisition unit, and an engineering monitoring data acquisition unit;

[0011] The data processing module is used to perform spatiotemporal standardization, quality verification, and repair on the raw data from the data access module.

[0012] The monitoring strategy control module has a built-in finite state automaton, which is used to dynamically determine the status of monitoring points based on real-time rainfall intensity and water level change rate rules, and send instructions to the monitoring instruments to switch the data acquisition frequency.

[0013] As a preferred technical solution of the present invention, the data processing module includes a data preprocessing unit, which is used to unify the data to a standard geographic coordinate system and a standard time reference, and to perform threshold verification and spatiotemporal consistency verification based on upstream and downstream hydrological relationships, and to interpolate and repair abnormal or missing data.

[0014] The time association unit automatically creates flood event identifiers based on upstream water conditions from the data preprocessing unit and dynamically tags associated data to the event.

[0015] As a preferred technical solution of the present invention, the intelligence and computing layer includes a simulation and deduction module, which includes a hydrodynamic model and a pre-simulation result generation unit. The module uses a built-in calibrated hydrodynamic model and adopts a parallel computing architecture and an adaptive time step control algorithm to quickly solve the problem and generate pre-simulation results for each scheme.

[0016] The simulation analysis module analyzes the quantification effect of the scheduling scheme based on the pre-simulation results of the simulation module.

[0017] The early warning module is used to generate early warning information based on the analysis results of the simulation and analysis module and the evolution trend of floods in the basin.

[0018] As a preferred technical solution of the present invention, the inference and analysis module includes a flood peak identification unit, which is used to automatically identify flood peaks at upstream and downstream stations and match them using a dynamic time warping algorithm;

[0019] The dynamics calculation unit calculates the propagation speed based on the flood peak pairs matched by the flood peak identification unit, and uses an exponentially weighted moving average algorithm for smoothing to predict the movement trajectory and arrival time of the flood peak.

[0020] As a preferred technical solution of the present invention, the interaction and execution layer includes a decision interaction module, a GIS-based visualization operation interface for displaying simulation and deduction, and a comparison view of the pre-simulation results of multiple scheduling schemes.

[0021] The decision-making interaction module includes a scheme pre-simulation operation comparison unit, which is used to receive the operation of the virtual control by the staff to generate the scheduling scheme and display a comparison view of multiple scheme pre-simulation results from the simulation and deduction module.

[0022] As a preferred technical solution of the present invention, the interaction and execution layer further includes an AR assistance module, which is deployed on the mobile terminal of the staff and is used to visualize and assist the scheduling plan and pre-rehearsal results confirmed by the decision interaction module on site in AR form.

[0023] The AR auxiliary module includes an AR scene simulation unit, which uses image recognition technology to align virtual information with real engineering scenes and overlay scheduling instructions and the simulated future water level line onto the real-time camera screen in the form of a virtual ruler, pointer and water level line.

[0024] The on-site interactive feedback unit is used to receive and transmit on-site personnel's operation confirmation and execution status based on AR information.

[0025] As a preferred technical solution of the present invention, the mobile terminal deployed by the AR auxiliary module is a waterproof tablet computer or AR smart glasses. The AR scene simulation unit built into it integrates a visual inertial odometry and a feature matching algorithm based on a pre-stored engineering model, which is used to realize the superposition of virtual water level lines, command scales and real water conservancy engineering scenes in an unmarked environment in the field.

[0026] As a preferred technical solution of the present invention, the host is equipped with a dedicated data fusion processing unit. This unit integrates a coordinate transformation engine, a rule engine and a real-time stream processing framework, and is used to perform spatiotemporal alignment, quality verification and event tag association on heterogeneous data streams from the meteorological data access unit, the hydrological element data acquisition unit and the engineering monitoring data acquisition unit, so as to generate a real-time decision data pool with a unified spatiotemporal benchmark.

[0027] As a preferred technical solution of the present invention, the parallel computing architecture of the simulation and deduction module is a parallel accelerated computing unit based on a graphics processor, which is used to perform large-scale parallel decomposition and solution of the grid computing task of the hydrodynamic model, so as to realize the rapid pre-simulation of the scheduling scheme.

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] Compared to existing methods that rely on fragmented information and personal experience for manual judgment, this invention constructs a display device with an intelligent hydrological system, consisting of a perception and data layer, an intelligence and computing layer, and an interaction and execution layer. Utilizing hydrodynamic models and parallel computing for rapid simulation, it achieves quantitative pre-simulation and optimization of scheduling schemes, replacing experience-based judgment. Finally, AR technology overlays the simulation results onto the real engineering scenario as virtual water level lines and instruction scales. The synergistic effect of each layer transforms the traditional model into a human-machine collaborative pre-simulation model, significantly shortening the decision-making chain, accurately quantifying the consequences of decisions, and comprehensively improving the timeliness and scientific nature of scheduling.

[0030] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the installation of a hydrological intelligent display instrument disclosed in this invention at a hydrological station;

[0032] Figure 2 This is a block diagram of an intelligent hydrological system for an intelligent hydrological display device disclosed in this invention;

[0033] Figure 3 This is the main interface of the intelligent hydrological system of the intelligent hydrological display instrument disclosed in this invention;

[0034] Figure 4 This is the data viewing interface of an intelligent hydrological system of a hydrological intelligent display instrument disclosed in this invention;

[0035] Figure labeling: 10. Sensing and Data Layer; 11. Data Access Module; 111. Meteorological Data Access Unit; 112. Hydrological Element Data Acquisition Unit; 113. Engineering Monitoring Data Acquisition Unit; 12. Data Processing Module; 121. Data Preprocessing Unit; 122. Time Correlation Unit; 13. Monitoring Strategy Control Module; 20. Intelligence and Computing Layer; 21. Simulation and Deduction Module; 211. Hydrodynamic Model; 212. Pre-simulation Result Generation Unit; 22. Deduction and Analysis Module; 221. Flood Peak Identification Unit; 222. Dynamics Calculation Unit; 23. Early Warning Module; 30. Interaction and Execution Layer; 31. Decision Interaction Module; 32. AR Assistance Module; 321. Scheme Pre-simulation Operation Comparison Unit; 322. AR Scene Simulation Unit; 323. On-site Interactive Feedback Unit. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0038] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0039] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0040] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0041] Example 1

[0042] See attached document Figure 1-2 As shown, the present invention provides a technical solution: a hydrological intelligent display instrument, including a main unit, a touch screen, a water level gauge, and a mounting column for installing the touch screen on a hydrological monitoring station. The main unit and the touch screen are connected to a power source via a power cord. The main unit is an industrial-grade waterproof and dustproof chassis. The touch screen is a high-brightness outdoor waterproof multi-touch screen, connected to the main unit via an armored waterproof HDMI cable and a network cable. The mounting column is a triangular bracket-type column with a flange at the bottom for fixing to the concrete base of the monitoring station. The touch screen is fixed to the column by a wind-resistant U-shaped clamp. The main unit is pre-installed with an intelligent hydrological system for processing water level gauge monitoring data. The intelligent hydrological system includes a sensing and data layer 10. The sensing and data layer 10 collects meteorological data, hydrological element data, and engineering monitoring data, and preprocesses and labels the data to solve the problem of relying on scattered information and build a unified and reliable decision-making data foundation.

[0043] The intelligence and computing layer 20, based on the data foundation of the perception and data layer 10, responds to at least two scheduling schemes input by the user, calls the calibrated hydrodynamic model 211 to perform parallel simulations, generates the pre-simulation results of each scheme, solves the problem of relying on personal experience and the inability to quantify the consequences of decisions, and realizes scientific and quantitative analysis.

[0044] The interaction and execution layer 30 uses animation to demonstrate the simulation process of the intelligence and computing layer 20, and provides a comparison view of the pre-simulation results of multiple scheduling schemes. At the same time, it uses AR technology to visualize and overlay decision instructions onto the on-site engineering scene, solving the problems of long decision-making chains and the disconnect between decision-making and on-site execution, and realizing efficient human-machine collaboration and decision implementation.

[0045] The embodiments of the present invention are also implemented through the following technical solutions.

[0046] In an embodiment of the present invention, the sensing and data layer 10 includes a data access module 11 for connecting a meteorological forecasting system, an instrument for monitoring hydrological elements, and a water conservancy project monitoring system. The data access module 11 includes a meteorological data access unit 111, a hydrological element data acquisition unit 112, and an engineering monitoring data acquisition unit 113.

[0047] The data access module 11 is used to perform the following steps: S101, through the configured protocol adapter (the protocol adapter has a corresponding communication module pre-configured for different data sources, such as the HTTP / JSONAPI adapter for the meteorological forecast system and the ModbusTCP protocol adapter for the water conservancy project monitoring system), establishes connections with the meteorological forecast system, the water conservancy project monitoring system, and instruments such as water level gauges, water temperature gauges, rain gauges, and water quality testers to collect hydrological element data;

[0048] S102 collects raw data such as rainfall forecasts, water level / flow data, gate and pump operating status, and real-time video streams in each hydrological station area in real time.

[0049] S103, the collected raw data is packaged in a unified format and output to the downstream data processing module 12;

[0050] The data processing module 12 is used to perform spatiotemporal standardization, quality verification and repair on the raw data from the data access module 11. The data processing module 12 includes a data preprocessing unit 121, which is used to unify the data to a standard geographic coordinate system and a standard time base, and to perform threshold verification and spatiotemporal consistency verification based on upstream and downstream hydrological relationships (automatically extracted using GIS hydrological analysis tools), and to perform interpolation repair on abnormal or missing data.

[0051] Among them, the upstream and downstream hydrological relationship includes: upstream station set: the numbers of all upstream stations that receive water from it;

[0052] Downstream station: The number of the next station into which the water flows directly;

[0053] River section length: The length of the river channel between the current station and the downstream station;

[0054] Confluence time empirical parameter: range of flood propagation time based on historical data;

[0055] The time association unit 122 automatically creates flood event identifiers based on upstream water conditions from the data preprocessing unit 121, and dynamically marks the associated data to the event, assigning a unified event label, such as associating it with the same flood process;

[0056] The data processing module 12 is used to perform the following steps: S201, for each received data, the data preprocessing unit 121 unifies it to the WGS84 coordinate system and unifies the timestamp to UTC time;

[0057] S202, continue to perform multi-level data quality checks through data preprocessing unit 121, including threshold checks, jump checks based on statistical process control, and spatiotemporal consistency checks based on upstream and downstream relationships. For missing or abnormal data, time series linear interpolation is used for repair.

[0058] The verification and repair process is as follows:

[0059] Threshold verification: Check whether the data value exceeds the reasonable range of physical or historical data. For example, the water level value should be within the range of [station riverbed elevation, historical highest water level * 1.2], and the rainfall intensity value should be >= 0 and < historical maximum minute rainfall intensity.

[0060] Sudden jump verification based on statistical process control: For time series data (such as water level) of a certain station, calculate its moving average over a sliding time window (such as the past 1 hour). ) and moving standard deviation ( ), moving average The calculation formula is: ,in The number of data points within the sliding window. For each data value, the moving standard deviation The calculation formula is: ; sqrt is the square root operation, where Represents the time window within which the first... The water level observation value (or flow rate observation value, rainfall intensity observation value, depending on the data type being verified) at each collection time.

[0061] Set control limits ( If the difference between the current instantaneous data value and the previous value exceeds the control limit, it is judged as a sudden jump;

[0062] Spatiotemporal consistency verification based on upstream and downstream relationships: Utilizing the principles of digital watershed topology and water balance, the digital watershed topology map is automatically generated using high-precision digital elevation model data and GIS hydrological analysis tools (the GIS hydrological analysis tools are lightweight analysis modules built into the system, or can call corresponding functions of existing GIS software; the high-precision digital elevation model data used is provided and imported by the user during system deployment). It is stored in graph structure or relational table format, and its core content includes: unique codes, spatial coordinates (latitude and longitude), elevations, and topological connections between nodes (upstream node set, downstream nodes) and river segment geometric attributes (length, slope), for example, the flow rate of upstream station U at time t. Its transmission time After that, it should be equal to the downstream station D at Flow of time And consider the inflow within the interval;

[0063] The Muskingan channel flow calculation method is used, based on the flow rate at time t at the upstream station. Based on the hydrological parameters calibrated for this river section, the downstream station was calculated. Flow rate estimate at time , serving as a benchmark for spatiotemporal consistency verification;

[0064] Specifically, the Muskingan method uses the following two-time linear formula for calculation:

[0065]

[0066] in, , These represent the current and previous time period traffic of the upstream station, respectively. The coefficient represents the measured flow rate of the downstream station in the previous time period. , , Calculate using the following formula and satisfy... :

[0067]

[0068]

[0069]

[0070] In the formula: K is the storage constant, representing the propagation time of flood waves in the river section, in hours, and its value is greater than 0, usually ranging from 0.5 to 24 hours depending on the characteristics of the river channel; x is the flow-weight factor, reflecting the channel's regulation capacity, a dimensionless number, with a value ranging from 0 to 0.5; Δt is the time step used in the calculation, in hours. To ensure the numerical stability of the calculation and the clarity of the physical meaning of the coefficients ( , , (All are non-negative), and the time step Δt must satisfy the following constraint: 2Kx≤Δt≤2K(1-x);

[0071] The method for determining parameters K and x is as follows: During the system deployment phase, select at least 3 historical flood process data of this basin, take the upstream station flow as input and the downstream station flow as output, and use the least squares method or optimization algorithm (such as SCE-UA algorithm) to fit and calibrate formula (1) so that the Nash efficiency coefficient of the simulated flow process and the measured flow process is the highest. The optimal parameter set (K, x) obtained by calibration is stored in the system along with the digital basin topology.

[0072] During verification, if the measured flow rate of the downstream station at the current moment... Compared with the estimated value If the relative error exceeds a preset threshold (e.g., 30%), the measured data will be marked as suspicious data.

[0073] Repair method: For data points marked as missing or outliers The formula for linear interpolation of time series is:

[0074]

[0075] in, and These are the most recent valid data points before and after the anomaly. For anomalies determined due to upstream / downstream conflicts, the data calculated from upstream data is used first. Replace;

[0076] S203 When an upstream station experiences a flood exceeding the warning level, a flood event ID is automatically created through the time association unit 122, and all related data (hydrological information, engineering information, forecasts) in the subsequent relevant time period are dynamically marked as this event;

[0077] The monitoring strategy control module 13 has a built-in finite state automaton, which is used to dynamically determine the status of the monitoring point according to the real-time rainfall intensity and water level change rate rules, and send instructions to the monitoring instrument to switch the data acquisition frequency.

[0078] The monitoring strategy control module 13 is used to perform the following steps: S301, define three states for each monitoring point: stable period, changing period, and emergency period, as well as the transfer rules;

[0079] S302, periodically (e.g., every 5 minutes) assess station data (e.g., rainfall intensity, water level variability), and if the real-time rainfall intensity I ≥ 20 mm / h or the water level exceeds the warning level, then enter the emergency phase;

[0080] S303, after the status changes, sends a command to the corresponding host via wireless network to switch the data acquisition frequency to a preset value (e.g., switch to 1 minute / time during emergency periods).

[0081] Specifically, the three state definitions and transition rules of the finite state automaton described in step S301 are as follows:

[0082] Stable period: Default state, sampling frequency is (e.g., 5 minutes / session).

[0083] Change period: When the real-time rainfall intensity I satisfies or rate of change of water level This state is entered at that time. The sampling frequency is increased to... (e.g., 2 minutes / session);

[0084] Emergency period: when real-time rainfall intensity or the water level exceeds the warning level. or rate of change of water level When entering this state, the sampling frequency is increased to [value missing]. (e.g., 1 minute / time);

[0085] The evaluation period for state transitions is (e.g., 5 minutes) The condition for returning from the emergency or change period to the stable period is: the triggering condition is not met within N consecutive assessment cycles (e.g., N=6, i.e., 30 minutes).

[0086] In an embodiment of the present invention, the intelligence and computing layer 20 includes a simulation and deduction module 21, which includes a hydrodynamic model 211 and a pre-simulation result generation unit 212. The pre-simulation result generation unit 212 generates the pre-simulation results of each scheme by using the built-in calibrated hydrodynamic model 211, i.e. the adjusted hydrodynamic model 211, and by using a parallel computing architecture and an adaptive time step control algorithm to solve the problem quickly.

[0087] Among them, the hydrodynamic model 211 is constructed based on the one-dimensional Saint-Venant equations, and spatial discretization and numerical algorithm simplification are carried out for fast solution. Specifically, the Preissmann implicit difference scheme of the one-dimensional Saint-Venant equations is used for spatial and temporal discretization, transforming the continuous partial differential equations into a large sparse linear equation system. To meet the requirements of fast solution, the model performs local linearization on the convection term in the momentum equation while ensuring mass conservation. The model solution adopts the chasing method for tridiagonal matrix optimization, and the parallel computing architecture is used to perform parallel decomposition and solution of the block tridiagonal matrix formed by the simultaneous equations of all river segments in the basin, thereby realizing rapid forecasting and extrapolation at the minute level. The calibration steps of the calibrated hydrodynamic model 211 are as follows:

[0088] S400a collects complete data on at least three historical flood events within the basin, including water level and flow process lines at each hydrological station, as well as corresponding spatial distribution data of rainfall. The first 70% of the data for each flood is used as a calibration dataset, and the last 30% is used as a validation dataset.

[0089] S400b sets the initial value range for the key physical parameters of hydrodynamic model 211 (mainly the Manning roughness coefficient n for each river section and floodplain). For example, the n value range for the river channel is 0.025-0.045, and the n value range for the floodplain is 0.05-0.15.

[0090] The S400c uses a Bayesian optimization algorithm with the Nash efficiency coefficient and peak water level error as objective functions. It automatically iterates to optimize the Manning roughness coefficient n. The optimization process is executed offline on the host computer until the objective function converges or the maximum number of iterations (e.g., 1000 times) is reached.

[0091] The Bayesian optimization calibration process includes:

[0092] Input: The vector to be optimized for the Manning roughness coefficient n (one parameter for each river segment and floodplain), and the input of the historical flood calibration dataset (rainfall, upstream boundary).

[0093] Objective function:

[0094] The formula for calculating the Nash efficiency coefficient is as follows: , To observe the flow rate, To simulate traffic, It is the arithmetic mean of all observed flow rates over the entire calibration (or verification) period. , The closer the NSE is to 1, the better the simulation.

[0095] It is the absolute value of the difference between the simulated and observed values ​​of the flood peak water level at the key cross-section. These are weighting coefficients (e.g., all are taken as 0.5);

[0096] Bayesian optimization algorithms construct surrogate models (such as Gaussian processes) of parameters n and objective function F(n). By continuously selecting the value of n that maximizes the expected improvement through simulation and iterative evaluation, they quickly approximate the optimal parameter set that minimizes F(n). ;

[0097] S400d applies the optimal parameter set obtained from calibration to the flood simulation of the validation dataset, calculates the Nash efficiency coefficient during the validation period. If the Nash efficiency coefficient is >0.75 and the peak water level error of the key station is less than 0.1 meters, the model is considered to be qualified for calibration and can be used for real-time simulation.

[0098] The model validation process is as follows:

[0099] Will Applied to an independent validation dataset, the model is run to obtain simulated hydrological hydrographs, and the validation period is calculated. ,like If the value is greater than 0.75, the model is considered to have good generalization ability and can be used for real-time forecasting.

[0100] Among them, the parallel computing architecture is based on the parallel acceleration computing unit of the graphics processor. Specifically, it adopts the CUDA (for NVIDIA GPU) or OpenCL framework to decompose the calculation of hydrodynamic model 211 (fixed structured grid) into several grid units. The hydraulic state (water depth, flow velocity) calculation task of each grid unit is mapped to a GPU thread and executed in parallel.

[0101] The adaptive time step control algorithm uses the largest possible time step to improve computation speed while ensuring numerical stability. The stability condition is determined by the Courant number. ,in, For flow rate, For time step, For spatial grid size, The maximum allowable Courant number is set according to the CFL stability condition to ensure the convergence and stability of numerical calculations, and is usually taken between 0.8 and 1.0.

[0102] The process is as follows: At the beginning of each computation time step, all grid cells are traversed, and the current maximum flow velocity across the entire field is calculated. Calculate the maximum allowable time step that satisfies the stability condition. ;

[0103] The actual time step used Set as Set a percentage (e.g., 80%) and ensure it is between a preset minimum and maximum value (e.g., 1 second to 60 seconds).

[0104] Based on this Δt-progression model calculation, since the flow velocity changes during flood evolution, It will also be dynamically adjusted accordingly;

[0105] The simulation and deduction module 21 is used to perform the following steps: S401, receiving real-time full-basin hydrological element data from the data processing module 12 as the initial field, and receiving future grid rainfall forecasts (provided by the meteorological forecasting system and accessed in real time through the meteorological data access unit 111 of the sensing and data layer 10; the grid refers to the regular spatial units (such as square grids) into which the forecast area is divided, and each grid point provides future rainfall forecast values ​​at different time points, forming a spatially distributed input field) as the upper boundary condition;

[0106] S402, receive user scheduling operation instructions (such as setting the gate opening degree) from the interaction and execution layer 30, and use them as the internal boundary conditions of the hydrodynamic model 211 at a specific moment;

[0107] In this process, the instructions of the interaction and execution layer 30 precede the calculations of the simulation and deduction module 21. Specifically, the user sets a scheduling scheme in the interaction layer (e.g., the gate opening is adjusted to 50% at 10:00). This instruction is sent to the host in real time. When the host starts the simulation and deduction calculation (e.g., starting from the current 9:00), "gate opening 50% at 10:00" is embedded into the model as a preset internal boundary condition.

[0108] S403 calls the calibrated hydrodynamic model 211 and uses parallel computing and an adaptive time step algorithm to numerically integrate and solve the hydrodynamic equations. Finally, it obtains the hydraulic state of all grid cells in the entire computational domain at each future computational time step, including water level, water depth, flow velocity, etc.

[0109] S404 outputs the water level / discharge process line and inundation depth distribution map of key sections for the next 6-72 hours (this time range is the effective duration of the externally input grid rainfall forecast. The system defaults to simulating a flood process of the same length as the forecast. For example, if a 72-hour rainfall forecast is connected, the simulation duration is 72 hours. If only the short term is needed, the results of the first 6 hours can be extracted for analysis).

[0110] The simulation analysis module 22, based on the output simulation results obtained from the hydrodynamic model 211 in step S403, namely the water level / flow process line and inundation distribution for the future period, quantifies the differences in indicators such as water level process line, flood peak, flood peak arrival time, and inundation area at key sections by comparing the no-scheduling scheme (natural evolution) and the scheduling scheme. For example, scheme A reduces the flood peak water level at downstream station X by 0.5 meters and delays the arrival of the flood peak by 2 hours. The simulation analysis module 22 includes a flood peak identification unit 221, which is used to automatically identify the flood peaks at upstream and downstream stations and match them using a dynamic time warping algorithm.

[0111] The dynamics calculation unit 222 calculates the propagation speed based on the flood peak pairs matched by the flood peak identification unit 221, and uses an exponential weighted moving average algorithm for smoothing to predict the movement trajectory and arrival time of the flood peak.

[0112] The early warning module 23 is used to generate early warning information based on the analysis results of the simulation analysis module 22 (referring to the conclusions of the simulation, such as the prediction that the Y station will exceed the guaranteed water level in 5 hours) and the evolution trend of the flood in the basin (referring to the comprehensive judgment of the flood peak propagation speed and intensity change trend based on real-time monitoring and exponential weighted moving average smoothing). The early warning rule is to automatically generate early warning information of the corresponding level when it is predicted that the water level of a certain station will exceed the warning water level or the guaranteed water level, or when it is predicted that the inundation range will involve important residential areas or infrastructure. The early warning level is determined according to the magnitude of the predicted water level exceeding the warning water level. For example, a predicted water level exceeding the warning level by 0-0.5 meters is a blue warning, 0.5-1.0 meters is a yellow warning, 1.0-1.5 meters is an orange warning, and greater than 1.5 meters is a red warning. The specific threshold is configured according to the basin flood control plan.

[0113] The simulation analysis module 22 and the early warning module 23 are used to perform the following steps: S501, for the flow process lines of upstream and downstream stations, the peak is automatically identified by the peak finding algorithm through the peak identification unit 221;

[0114] In step S501, the peak-finding algorithm specifically involves: processing the flow process line... After applying the moving average filter, find the local maxima that satisfy the following conditions. :

[0115] It is greater than M data points before and after it (M represents a period of time centered on the current data point, covering both the front and back). (e.g., a data point set of 3 hours, where the value of M will dynamically change according to the actual collection frequency to ensure a fixed flow rate within a fixed time window);

[0116] Exceeding the set flood peak threshold (For example, 1.5 times the multi-year average flow).

[0117] The time interval between adjacent flood peaks is greater than the set minimum interval. (For example, 12 hours) to exclude minor fluctuations;

[0118] The identified flood peak sequence is denoted as the upstream sequence. and downstream sequences , as input to the dynamic time warping algorithm;

[0119] S502, and through the flood peak identification unit 221, the dynamic time warping algorithm is used to intelligently match the flood peak sequences identified upstream and downstream, and establish the propagation correspondence;

[0120] The dynamic time warping algorithm, when used to match upstream and downstream flood peaks, has the following specific settings:

[0121] Input features: Peak flow values ​​identified in previous steps The flood peak morphology features (such as the slope of the rising segment and the width of the peak) form a feature vector, which serves as the comparison sequence for the dynamic time warping algorithm. The calculation method for the flood peak morphology features is as follows:

[0122] Slope of the rising segment: Take M data points before the flood peak (e.g., M=3, representing the rising segment), and fit a straight line using the least squares method. The slope of this straight line is the slope of the rising segment.

[0123] Peak width: Draw a horizontal line at half the height of the peak flow value. The time difference between the two points where this horizontal line intersects the rising and falling branches of the flow process line is the peak width (also known as half peak width).

[0124] Distance metric: Euclidean distance is used to calculate the distance between the upstream and downstream flood peak feature vectors. Let the upstream flood peak feature vector be... The downstream flood peak feature vector is Then the Euclidean distance ,in, Iterate through all dimensions of the feature vector;

[0125] Path constraints: Introduce monotonicity constraints (forcing the algorithm to maintain time order when searching for the optimal matching path). The use of inclined window constraints (such as the Itakura Parallelogram window) ensures that the matching paths maintain a temporal order (upstream peaks precede downstream peaks) and that the matching does not deviate too far from the diagonal, which conforms to the physical laws of flood wave propagation in the river channel.

[0126] Output: The algorithm outputs the optimal matching path, thereby establishing the upstream flood peak. With downstream flood peak The one-to-one correspondence was established, and the propagation time difference of each pair of matching flood peaks was calculated. ;

[0127] The specific implementation steps of the dynamic time warping algorithm are as follows:

[0128] upstream flood peak sequence and downstream flood peak sequence The feature vectors (flow rate, rise slope, peak width) of each flood peak are arranged in chronological order to form two multidimensional time series. The Euclidean distance between each pair of flood peak feature vectors in the two series is calculated to form a distance matrix. An inclined window constraint is used as a global constraint window, and monotonicity and continuity constraints are applied to ensure that the matching path reflects the physical propagation order of the flood wave from top to bottom. A dynamic programming algorithm is applied to find the path with the minimum cumulative distance under the constraints. The point pairs on this path are the matching upstream and downstream flood peaks.

[0129] The upstream flood peak has been matched using a dynamic time warping algorithm. (time ) and downstream station flood peak (time Given the length L of the river channel between the two stations, then the propagation speed... ;

[0130] S503: Calculate the propagation speed based on the successfully matched flood peak pairs, and use an exponentially weighted moving average algorithm to smooth the speeds of multiple recent flood peaks to obtain the current dynamic propagation speed;

[0131] The formula for smoothing the exponentially weighted moving average is as follows:

[0132]

[0133] in, This is the latest calculated propagation speed. It is the current smoothed speed. It is the smoothed velocity from the previous moment, the initial value for smoothing calculation. The value is set as: the first effectively calculated flood peak propagation velocity V(1), or the n-year average flood peak propagation velocity obtained from historical flood data of this river section. It is a smoothing factor (0 < α < 1, such as 0.3). This algorithm gives greater weight to recent velocities, making the prediction more in line with the current water conditions.

[0134] S504, based on the current dynamic propagation speed, dynamically predicts and visualizes the movement trajectory and estimated arrival time of the flood peak that has not yet reached the downstream area on a GIS map. Specifically, it knows the current location of the flood peak (upstream station) and the smoothed propagation speed of the current river segment. On the GIS map, along the center line of the river, with For velocity, dynamically plot a marker point (trajectory) moving downstream from the upstream station, and calculate the estimated arrival time to the downstream station Z. ,in It represents the remaining river length from the current location of the flood peak to station Z.

[0135] In an embodiment of the present invention, the interaction and execution layer 30 includes a decision interaction module 31, which displays simulation results and compares the results of multiple scheduling schemes through a GIS-based visualization operation interface. The multiple scheme comparison view is generated by the user creating schemes A and B on the system interface. The system submits the instruction sets of the two schemes to the simulation module 21 in parallel. The simulation module 21 calculates and returns two sets of result data (water level process line and inundation map).

[0136] The scheme pre-operation comparison unit 321 of the decision interaction module 31 receives these two sets of data and performs visualization rendering in the same screen area through parallel curves and column flooding range maps to form a comparison view;

[0137] The decision interaction module 31 includes a scheme pre-simulation operation comparison unit 321, which is used to receive the operation of the staff on the virtual control to generate a scheduling scheme and display a comparison view of multiple scheme pre-simulation results from the simulation and deduction module 21.

[0138] The decision interaction module 31 is used to perform the following steps: S601, based on WebGIS technology (used to load electronic maps in the browser environment of the touch screen, and render data such as real-time monitoring points (perception layer), flood peak trajectory (output of the simulation analysis module 22), and inundation range (output of the simulation simulation module 21) on the map in the form of layer overlay, which is the core technical means for the interaction and execution layer 30 to realize the GIS-based visualization operation interface), the flood control plan is displayed on a touch screen, with real-time water and rain conditions provided by the layer rendering technology, flood peak trajectory provided by the simulation analysis module 22, and geographical background information;

[0139] S602 provides virtual controls for projects such as virtual dams and reservoirs, and receives interactive commands from users such as dragging and setting.

[0140] S603 sends user commands to simulation module 21 and receives the simulation results returned by it, and displays them in a multi-view format (process line diagram, cross-section diagram, flooding diagram);

[0141] S604 supports the creation and management of multiple scheduling schemes and provides a synchronous comparative analysis function for the pre-simulation results of multiple schemes.

[0142] In an embodiment of the present invention, the interaction and execution layer 30 further includes an AR assistance module 32, which is deployed on the mobile terminal of the staff and is used to visualize and assist the scheduling plan and pre-rehearsal results confirmed by the decision interaction module 31 on site in AR form.

[0143] The AR auxiliary module 32 includes an AR scene simulation unit 322, which uses image recognition technology to align virtual information with real engineering scenes and overlay dispatch instructions and the simulated future water level line onto the real-time camera screen in the form of a virtual ruler, pointer and water level line.

[0144] The on-site interactive feedback unit 323 is used to receive and transmit on-site personnel's operation confirmation and execution status based on AR information;

[0145] AR-assisted module 32 is used to perform the following steps: S701, using the camera of the mobile terminal, and with the help of image recognition based on natural features, to perform scene registration and pose calculation for the on-site water conservancy facilities (such as gates and water level gauges). Specifically, the mobile terminal camera is pointed at the on-site water conservancy facilities (such as gates), and the ORB algorithm is used to first detect FAST corner points in the image. Then, the BRIEF descriptor is used to generate a binary feature vector for each corner point to describe the features of its surrounding image blocks. The ORB features extracted in the current frame are matched with the ORB feature library of the pre-stored 3D model of the facility to establish the correspondence between 2D image points and 3D model points. Using at least 4 pairs of matching points, the 3D position and pose (i.e., rotation matrix and translation vector) of the camera relative to the real facility are solved by EPnP.

[0146] S702 synchronously obtains scheduling instructions (such as target opening degree) and pre-simulation results (such as future water level) from the decision interaction module 31, and overlays and renders virtual rulers, pointers and water level lines on the real-time image of the camera according to the registered pose;

[0147] S703: On-site personnel operate based on AR overlay information and confirm the execution and status feedback through the terminal, and the information is sent back to the host.

[0148] The rendering of AR overlay information involves calculating the projection position of the virtual information onto the camera image based on the calculated precise pose (rotation matrix and translation vector). For example, to overlay a "future water level" onto a real water level gauge: given the target water level value (e.g., 3.5 meters), combined with a pre-stored 3D model of the water level gauge, the 3D coordinates of the 3.5-meter mark on the water level gauge in the real world can be determined. Through coordinate transformation (Where K is the intrinsic parameter matrix of the mobile terminal camera, which is pre-calibrated and stored in the host using the standard checkerboard calibration method during the system deployment phase. It includes inherent parameters such as focal length and principal point coordinates.) This allows us to obtain... 2D pixel coordinates on the screen .,exist Draw a horizontal, semi-transparent virtual water level line at that location. The principle of superimposing the virtual ruler and pointer is the same.

[0149] In embodiments of the present invention, to address the problem of the unstable application of AR technology in complex hydrological field environments (humid, unmarked), the mobile terminal deployed in the AR auxiliary module 32 is a waterproof tablet or AR smart glasses. Its built-in AR scene simulation unit 322 integrates a visual inertial odometry system and a feature matching algorithm based on a pre-stored engineering model. (The visual inertial odometry system uses the ORB-SLAM3 framework, which estimates the six-degree-of-freedom pose of the mobile terminal in real time by fusing ORB feature points in the camera image with the angular velocity and acceleration data of the inertial measurement unit; the feature matching algorithm based on the pre-stored engineering model specifically employs iterative nearest point...) The algorithm performs precise registration: Starting from the initial pose provided by the visual inertial odometry, the environmental point cloud data collected in real time by the mobile terminal is iteratively aligned with the pre-stored high-precision 3D point cloud model of the engineering facility. By continuously optimizing the rotation and translation matrices, the overall distance error between the two point clouds is minimized, thereby calculating the precise six-degree-of-freedom pose of the mobile terminal relative to the real engineering facility, realizing the high-precision spatial superposition of virtual water level lines, command scales and the physical world. This is used to realize the superposition of virtual water level lines, command scales and real water conservancy engineering scenes in the field unmarked environment, ensuring the reliability and usability of AR auxiliary module 32 in the unmarked environment.

[0150] In the AR scene simulation unit 322, virtual-real fusion is achieved in an unmarked environment through the following steps:

[0151] During the indoor prefabrication stage, the S700a uses 3D scanning to create high-precision 3D point cloud or feature wireframe models for key water conservancy facilities (such as gates and water level gauge calibration columns), and stores them locally on the mobile terminal.

[0152] On-site, the S700b activates the mobile terminal's camera and inertial measurement unit. By analyzing the changes in visual feature points between consecutive video frames and the acceleration and angular velocity data of the inertial measurement unit through visual inertial odometry, it calculates the terminal's six-degree-of-freedom pose changes in real time, achieving preliminary autonomous positioning and mapping.

[0153] The S700c performs feature matching between the real-time camera captured images and the pre-stored 3D model in the S700a (the 3D model is stored in the common .obj or .gltf format). It uses the ORB algorithm to extract feature points in the real-time images and matches them with the expected features projected onto the 3D model under the current estimated pose. Through the iterative nearest point algorithm, it minimizes the distance between the real-time feature point cloud and the pre-stored model point cloud, thereby continuously optimizing the pose of the terminal relative to the real engineering facilities and achieving high-precision spatial alignment between virtual information (such as water level lines and operating scales) and the real scene.

[0154] The S700d, based on the precise pose calculated by the algorithm in the S700c, graphically represents the received scheduling instructions (such as the target water level of 3.5 meters). At the corresponding physical location in the camera's view, it overlays and renders a virtual water level line aligned with the scale of the real water level gauge, as well as a virtual arrow pointing to the target opening scale.

[0155] In an embodiment of the present invention, in order to solve the problems of low efficiency and high latency in real-time fusion processing of multi-source heterogeneous data streams and to support the rapid response capability of the system, a dedicated data fusion processing unit is provided in the host. This unit integrates a coordinate transformation engine, a rule engine and a real-time stream processing framework, and is used to perform spatiotemporal alignment, quality verification and event tag association on heterogeneous data streams from meteorological data access unit 111, hydrological element data acquisition unit 112 and engineering monitoring data acquisition unit 113, and generate a real-time decision data pool with a unified spatiotemporal benchmark, thereby performing efficient spatiotemporal alignment, quality verification and event association on multi-source data streams.

[0156] The dedicated data fusion processing unit is integrated within the host and built on a stream processing framework, such as the Apache Flink DataStream API. Spatiotemporal alignment is achieved by calling a user-defined function through a Map operator to complete the coordinate-time transformation. The quality verification driven by the rule engine is implemented through the ProcessFunction operator, which allows access to the state and processing of each data item, and integrates a configurable verification rule library. Event association and tagging are implemented through the KeyedProcessFunction operator, using the watershed partition as the key. The active flood event context is maintained in the operator's state, thereby dynamically adding event ID tags to data that flow through and belong to the same event spatiotemporal window. Its internal workflow is as follows:

[0157] Spatiotemporal alignment engine: Receives raw data streams from each access unit. For each data item, the engine queries the preset station spatial attribute table based on its own spatial identifier (such as station number) and converts it to latitude and longitude in the WGS84 coordinate system. At the same time, it unifies the timestamps of all data to UTC time and adds a unified data arrival timestamp.

[0158] Quality verification driven by a rules engine: Includes a built-in configurable verification rule library, for example:

[0159] Threshold rule: Check whether the water level value is within the range of historical maximum and minimum values;

[0160] Change rate rule: Check whether the water level change between adjacent data points exceeds the physical maximum possible value (e.g., 1 meter / minute).

[0161] Spatiotemporal consistency rule: Using the flow rate at time t of the upstream station, the flow rate at the downstream station is estimated using river channel calculation methods such as the Muskingan method. If the measured flow rate at a downstream station deviates significantly from this range, it will be marked as abnormal.

[0162] Event association and tagging: The rule engine defines event triggering rules. For example, when the water level of a hydrological station exceeds the warning level, a new flood event is triggered. Once triggered, the system generates a globally unique flood event ID. Subsequently, all relevant data (water level, flow rate, rainfall, and associated gate operations) that are spatially located within the affected basin and whose timestamps are within the event time window will be automatically associated and tagged with the event ID. The affected basin is determined by a pre-loaded basin topology map.

[0163] In an embodiment of the present invention, in order to solve the problem that the hydrodynamic model 211 has a large computational load and that traditional CPU computing is difficult to meet the real-time requirements of scheduling decisions, the parallel computing architecture of the simulation and deduction module 21 is a parallel accelerated computing unit based on a graphics processor, which is used to perform large-scale parallel decomposition and solution of the grid computing task of the hydrodynamic model 211, and to accelerate the grid computing task by utilizing the large-scale parallel capability of the GPU to ensure the rapid pre-simulation of the scheduling scheme.

[0164] Example: Sensing and Data Layer 10 - Taking a flood event as an example:

[0165] Data access: The meteorological forecast adapter unit pulls the grid rainfall forecast for the next 72 hours from the meteorological forecast system; the hydrological element data acquisition unit 112 collects water level and flow data from each hydrological station; and the engineering monitoring unit obtains the gate opening degree and opening / closing status through communication protocols (such as Modbus protocol).

[0166] Data processing: All data is converted to WGS84 coordinates and UTC time. Data preprocessing unit 121 detects a sudden jump in data at a certain station. After spatiotemporal consistency verification, it is determined to be abnormal. The upstream and downstream station data are repaired by using time series linear interpolation. When the water level at upstream station A exceeds the warning level, the intelligent hydrological system creates event ID "FLOOD_20240510_001". Subsequent data exceeding the warning level at stations B and C, as well as related gate operations, are all marked as this event.

[0167] Adaptive acquisition: When the rainfall intensity on the watershed surface is detected to increase sharply to 25 mm / h, the monitoring strategy control module 13 sets the status of the relevant rain gauge and water level stations to the emergency period and immediately issues an instruction to increase the acquisition frequency of these stations from 5 minutes / time to 1 minute / time.

[0168] The specific implementation of the Intelligence and Computing Layer 20 during flood evolution:

[0169] Flood tracking: The system identifies the flood peak sequences of upstream station X and downstream station Y. Through dynamic time warping algorithm, the flood peak Px1 of station X is successfully matched with the flood peak Py1 of station Y. The propagation time is measured to be Δt1=3.5 hours. Combining the propagation speeds of multiple recent flood peaks, the current river propagation speed V=15km / h is obtained after smoothing by exponential weighted moving average. Based on this, the system displays the current location of the flood peak on the electronic map and predicts the time it will reach downstream station Z.

[0170] Dispatch simulation: The decision-maker considers opening the dam for flood diversion. On the interactive interface, the virtual opening degree of the dam is set to 50%. The simulation module 21 is immediately started. Taking the current water situation of the entire basin as the initial condition and future rainfall as the input, the change in the dam opening degree is used as the internal boundary at time t. After the parallel hydrodynamic model 211 is called for rapid calculation (e.g., the calculation time is based on a host equipped with an NVIDIA RTX 4090 graphics card, simulating the flood evolution of about 100,000 computing grids over the next 24 hours), the simulation module 21 outputs the water level process line of the downstream key section H station over the next 24 hours (showing a peak decrease of 0.8 meters) and the possible change in the inundation range.

[0171] Specific implementation of the interaction and execution layer 30

[0172] Decision-making simulation and comparison: In the interactive interface, the staff saw the flood peak trajectory warning provided by the simulation analysis module 22 and formulated two plans: Plan A (gate opening of 50%) and Plan B (gate opening of 30%). The system submitted simulation requests in parallel and displayed the water level process lines of station H under the two plans side by side on the same interface. The staff could intuitively see that Plan A had a better peak reduction effect, but the inundation area was slightly larger. After weighing the options, Plan A was selected and confirmed for issuance.

[0173] On-site AR execution: Operators wearing AR smart glasses arrive at the dam. The glasses' camera identifies the dam body and completes registration through image recognition technology. Subsequently, a clear virtual ruler is superimposed on the corresponding position of the dam column on the glasses lens, with the pointer pointing to "3.5 meters" (i.e., the lifting height corresponding to 50% opening). At the same time, the text prompt "Execution Plan A" is displayed. Based on this visual guidance, the operator operates the hoist to accurately lift the dam to the specified height and confirms the execution via voice command (the voice command recognition function calls the voice recognition API built into the mobile terminal operating system or the integrated offline voice recognition SDK). The on-site image and confirmation information are transmitted back to the touch screen, and the dam status on the interactive interface is updated in real time to "Executed, opening 50%".

[0174] The above are merely preferred embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0175] Those skilled in the art will also understand that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments herein can be implemented as electronic hardware, computer software, or a combination thereof. To clearly illustrate the interchangeability between hardware and software, the various illustrative components, blocks, modules, circuits, and steps described above are generally described in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. Those skilled in the art can implement the described functionality in alternative ways for each specific application; however, such implementation decisions should not be construed as departing from the scope of this disclosure.

[0176] The steps of the methods or algorithms described in conjunction with the embodiments herein can be directly embodied in hardware, software modules executed by a processor, or a combination thereof. The software modules can reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium well known in the art. An exemplary storage medium is connected to the processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside in an ASIC. The ASIC can reside in a user terminal. Alternatively, the processor and storage medium can exist as discrete components in the user terminal.

[0177] For software implementation, the techniques described in this application can be implemented using modules (e.g., procedures, functions, etc.) that perform the functions of this application. This software code can be stored in memory units and executed by a processor. The memory units can be implemented within the processor or outside the processor; in the latter case, they are communicatively coupled to the processor via various means, as is well known in the art.

[0178] The foregoing description includes examples of one or more embodiments. It is certainly impossible to describe all possible combinations of components or methods in order to describe the above embodiments, but those skilled in the art will recognize that further combinations and arrangements of the various embodiments are possible. Therefore, the embodiments described herein are intended to cover all such changes, modifications, and variations that fall within the scope of the appended claims. Furthermore, the term "comprising" as used in the specification or claims is interpreted in a manner similar to the term "including," as interpreted when used as a conjunction in the claims. Additionally, the use of any term "or" in the specification of the claims is intended to mean "non-exclusive or."

Claims

1. A hydrological intelligent display instrument, characterized in that, The system includes a host, a touch screen, a water level gauge, and a mounting column for installing the touch screen at a hydrological monitoring station. The host is pre-installed with an intelligent hydrological system for processing water level gauge monitoring data. The intelligent hydrological system includes a sensing and data layer (10). The sensing and data layer (10) collects meteorological data, hydrological element data, and engineering monitoring data, and preprocesses and labels the data to build a decision data foundation. The intelligence and computing layer (20), based on the data foundation of the perception and data layer (10), responds to at least two scheduling schemes input by the user, calls the calibrated hydrodynamic model (211) to perform parallel simulations, and generates the simulation results of each scheme; The interaction and execution layer (30) is used to provide a synchronous comparison and display of the results of multiple schemes in the pre-drill, and to overlay the instructions of the selected scheme onto the real scene through AR technology to assist the staff in performing precise operations.

2. The intelligent hydrological display instrument according to claim 1, characterized in that, The perception and data layer (10) includes a data access module (11) for connecting the meteorological forecasting system, the instrument for monitoring hydrological elements, and the water conservancy project monitoring system, including a meteorological data access unit (111), a hydrological element data acquisition unit (112), and an engineering monitoring data acquisition unit (113). The data processing module (12) is used to perform spatiotemporal standardization, quality verification and repair on the raw data from the data access module (11); The monitoring strategy control module (13) has a built-in finite state automaton, which is used to dynamically determine the status of the monitoring point according to the real-time rainfall intensity and water level change rate rules, and send instructions to the monitoring instrument to switch the data acquisition frequency.

3. The intelligent hydrological display instrument according to claim 2, characterized in that, The data processing module (12) includes a data preprocessing unit (121) for unifying the data to a standard geographic coordinate system and a standard time reference, performing threshold verification and spatiotemporal consistency verification based on upstream and downstream hydrological relationships, and interpolating and repairing abnormal or missing data. The time association unit (122) automatically creates a flood event identifier based on the upstream water conditions from the data preprocessing unit (121) and dynamically marks the associated data to the event.

4. The intelligent hydrological display instrument according to claim 1, characterized in that, The intelligent and computing layer (20) includes a simulation and deduction module (21), which includes a hydrodynamic model (211) and a pre-simulation result generation unit (212). The simulation and deduction module (21) uses a built-in calibrated hydrodynamic model (211) and adopts a parallel computing architecture and an adaptive time step control algorithm to quickly solve the problem and generate the pre-simulation results of each scheme. The simulation analysis module (22) analyzes the quantification effect of the scheduling scheme based on the simulation results of the simulation module (21); The early warning module (23) is used to generate early warning information based on the analysis results of the simulation analysis module (22) and the evolution trend of the basin flood.

5. A hydrological intelligent display instrument according to claim 4, characterized in that, The inference and analysis module (22) includes a flood peak identification unit (221), which is used to automatically identify flood peaks at upstream and downstream stations and match them using a dynamic time warping algorithm; The dynamics calculation unit (222) calculates the propagation speed based on the flood peak pair matched by the flood peak identification unit (221), and uses an exponentially weighted moving average algorithm for smoothing to predict the movement trajectory and arrival time of the flood peak.

6. The intelligent hydrological display instrument according to claim 1, characterized in that, The interaction and execution layer (30) includes a decision interaction module (31), a GIS-based visualization operation interface for displaying simulations and a comparison view of the pre-simulation results of multiple scheduling schemes; The decision interaction module (31) includes a scheme pre-operation comparison unit (321), which is used to receive the staff’s operation on the virtual control to generate a scheduling scheme and display a comparison view of multiple scheme pre-operation results from the simulation and deduction module (21).

7. A hydrological intelligent display instrument according to claim 1 or 6, characterized in that, The interaction and execution layer (30) also includes an AR assistance module (32), which is deployed on the staff's mobile terminal to visualize and assist the scheduling plan and pre-rehearsal results confirmed by the decision interaction module (31) on-site in AR form. The AR auxiliary module (32) includes an AR scene simulation unit (322), which uses image recognition technology to align virtual information with real engineering scenes and overlay scheduling instructions and the pre-simulated future water level line onto the real-time camera screen in the form of a virtual ruler, pointer and water level line. The on-site interactive feedback unit (323) is used to receive and transmit the operation confirmation and execution status of on-site personnel based on AR information.

8. A hydrological intelligent display instrument according to claim 7, characterized in that, The AR auxiliary module (32) is deployed on a waterproof tablet computer or AR smart glasses. The AR scene simulation unit (322) built into it integrates a visual inertial odometry and a feature matching algorithm based on a pre-stored engineering model, which is used to realize the superposition of virtual water level lines, command scales and real water conservancy engineering scenes in an unmarked environment in the field.

9. A hydrological intelligent display instrument according to claim 1, characterized in that, The host is equipped with a dedicated data fusion processing unit, which integrates a coordinate transformation engine, a rule engine and a real-time stream processing framework. It is used to perform spatiotemporal alignment, quality verification and event tag association on heterogeneous data streams from the meteorological data access unit (111), the hydrological element data acquisition unit (112) and the engineering monitoring data acquisition unit (113), and generate a real-time decision data pool with a unified spatiotemporal benchmark.

10. A hydrological intelligent display instrument according to claim 4, characterized in that, The parallel computing architecture of the simulation and deduction module (21) is a parallel accelerated computing unit based on a graphics processor, which is used to perform large-scale parallel decomposition and solution of the grid computing task of the hydrodynamic model (211) in order to realize the rapid pre-simulation of the scheduling scheme.

Images