Real-time optimization scheduling method based on basin water regimen change trend analysis model

Abstract

The invention relates to the technical field of water conservancy model scheduling, in particular to a real-time optimization scheduling method based on a watershed water regimen change trend analysis model. The method comprises the following steps: continuously monitoring a water level change value in a drainage basin; when the water level change value exceeds a preset first threshold value, judging a water regimen change state, and recording the water regimen change state as initial water regimen feature information; acquiring gradient data, slope direction data, wetland area data and vegetation coverage data in a water regimen change state; and performing topographic feature correction on a preset watershed water regimen change trend analysis model based on the slope data and the slope direction data. By monitoring the water level change and combining the terrain and ecological characteristic data to dynamically correct the water regimen change trend analysis model, accurate prediction and real-time optimization scheduling of the watershed water regimen are achieved, and therefore the timeliness of watershed water resource management is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of water conservancy model scheduling, and in particular to a real-time optimization scheduling method based on a watershed water regime change trend analysis model. BACKGROUND

[0002] The monitoring sites in the watershed are unevenly distributed, resulting in missing water regime data in some areas, especially remote mountainous areas and small watersheds; for example, the rainstorm center area in the mountainous area lacks enough rain monitoring points, making it impossible to accurately capture the rapid response of local rainstorms to the flow of rivers. In addition, most existing monitoring equipment can only provide data at fixed time intervals, making it difficult to capture the instantaneous changes in water regime; and the existing watershed water regime change trend analysis model has a slow response speed when facing sudden water regime changes (such as sudden rainstorms, dam breaks, etc.), because the model needs a certain amount of time to process and analyze new data and update the prediction results, resulting in a lag in scheduling decisions; the complex terrain in the watershed (such as mountainous areas, gorges, and plain transition zones) has a significant impact on water regime changes, but existing models often have difficulty accurately simulating water regime changes when dealing with these complex terrains; specifically, the flow speed and direction in mountainous areas are greatly affected by the terrain, and existing models cannot accurately capture this change, resulting in a decrease in prediction accuracy. SUMMARY

[0003] Therefore, it is necessary to provide a real-time optimization scheduling method based on a watershed water regime change trend analysis model to solve at least one of the above technical problems.

[0004] To achieve the above-mentioned purpose, a real-time optimization scheduling method based on a watershed water regime change trend analysis model, the method comprising the following steps:

[0005] Step S1: continuously monitor the water level change value in the watershed; when the water level change value exceeds a preset first threshold value, determine that the water regime is in a change state, and record it as initial water regime characteristic information; in the water regime change state, obtain slope data, slope direction data, wetland area data, and vegetation coverage data;

[0006] Step S2: correct the preset watershed water regime change trend analysis model based on the slope data and the slope direction data; adjust the watershed water regime change trend analysis model that has been subjected to the terrain feature correction processing based on the wetland area data and the vegetation coverage data; adjust the watershed water regime change trend analysis model that has been subjected to the ecological feature adjustment to an optimal prediction accuracy state and store the optimal prediction model parameters;

[0007] Step S3: If the duration of the hydrological change is greater than the preset duration of the hydrological change, a dispatch warning signal is generated and a hydrological change reminder is triggered; when the water level change value is detected to decrease to zero and continues until the preset hydrological stabilization time, the basin hydrological change trend analysis model is restored to the initial prediction state.

[0008] Step S4: When the monitored water level change value increases from zero to the first threshold again, real-time hydrological characteristic information is obtained; when the real-time hydrological characteristic information matches the initial hydrological characteristic information, the basin hydrological change trend analysis model is adjusted according to the optimal prediction model parameters to perform real-time optimized scheduling operation of basin hydrological changes.

[0009] Preferably, the continuous monitoring of water level changes within the watershed in step S1 includes:

[0010] Multiple water level monitoring nodes are set up within the basin. The monitoring nodes are evenly distributed along the direction of water flow in the basin, and the distance between adjacent nodes is 1 / 10 of the width of the basin.

[0011] Each monitoring node uses a time-division multiplexing method to collect water level data. The first collection lasts for 5 seconds, the second for 10 seconds, and the third for 15 seconds, with each collection lasting 5 seconds longer than the previous one.

[0012] The water level data collected each time is stored in the order of collection;

[0013] Differential processing is performed on two consecutive water level data collections to obtain the water level difference change.

[0014] The arithmetic mean of the water level data in the water level difference change is obtained to obtain the water level arithmetic mean data.

[0015] The arithmetic mean water level data is segmented according to the direction of water flow in the basin to obtain the water level change gradient for each segment.

[0016] The water level change values ​​within the basin are divided according to the water level change gradient of each segment.

[0017] Preferably, in step S1, acquiring slope data, aspect data, wetland area data, and vegetation cover data under changing water conditions includes:

[0018] Multiple topographic sections are selected within the watershed, and multiple slope monitoring points are arranged along the direction of water flow in each section.

[0019] At each monitoring point, a tilt sensor is used to measure the tilt angle of the ground. The tilt angle at each monitoring point is used to determine the change in cross-sectional slope and obtain slope data.

[0020] Multiple terrain nodes were selected within the watershed, and multiple slope aspect monitoring points were set up at each node, with the monitoring points distributed in different directions.

[0021] At each monitoring point, a level was used to measure the slope orientation angle, and the angle data was converted into slope aspect data.

[0022] Preferably, in step S1, acquiring slope data, aspect data, wetland area data, and vegetation cover data under changing water conditions further includes:

[0023] Multiple wetland sample areas were selected within the watershed, and multiple area monitoring points were set up in each area, with the monitoring points distributed in different locations of the wetland;

[0024] At each monitoring point, a laser rangefinder was used to measure the wetland boundary and calculate the wetland area data;

[0025] Multiple vegetation quadrat areas were selected within the watershed, and multiple vegetation monitoring points were set up in each area. The monitoring points were distributed in areas with different vegetation types.

[0026] At each monitoring point, optical sensors were used to measure the reflectance spectrum of the vegetation;

[0027] Vegetation cover data are determined by the amount of reflectance spectrum.

[0028] Of particular importance, determining vegetation cover data through the amount of reflectance spectrum includes:

[0029] The reflectance spectral data is segmented according to wavelength intervals, and the number of bands with reflected light intensity higher than a preset threshold within each wavelength interval is counted.

[0030] For each monitoring point, the number of bands is summed to obtain the total number of reflectance spectra for that monitoring point;

[0031] For each monitoring point, the total number of reflectance spectra is compared with the preset vegetation coverage reference value, and the total number of reflectance spectra is adjusted according to the comparison results;

[0032] The sum of the adjusted reflectance spectra is linearly fitted to the vegetation coverage reference value to obtain the vegetation coverage estimate.

[0033] The vegetation coverage data is obtained by averaging the estimated vegetation coverage values ​​of all monitoring points.

[0034] Preferably, step S2, which involves correcting the terrain features of the preset watershed hydrological trend analysis model based on slope and aspect data, includes:

[0035] Calculate the slope change rate for each monitoring point;

[0036] Calculate the rate of change of slope aspect for each monitoring point;

[0037] The slope change rate and the aspect change rate are vector-synthesized to obtain the topographic change vector.

[0038] The watershed is divided into regions with topographic change characteristics based on the magnitude and direction of the topographic change vector.

[0039] Within the region of terrain change characteristics, the magnitude and direction of the average terrain change vector are calculated and used as the terrain change characteristic index of the region.

[0040] Calculate the topographic correlation coefficient between topographic change characteristic indicators and hydrological change data;

[0041] Topographic correlation coefficients with an absolute value greater than 0.5 were selected as topographic feature correction factors.

[0042] The terrain feature correction factor is input into the preset watershed hydrological change trend analysis model to correct the terrain features.

[0043] Of particular importance is the vector synthesis of the slope change rate and the aspect change rate, including;

[0044] Multiple topographic monitoring points were selected within the watershed, and the slope change rate and aspect change rate of each monitoring point were measured using a topographic measuring instrument.

[0045] For each monitoring point, the measured slope change rate and aspect change rate are recorded as two independent values.

[0046] Multiply the slope change rate by the preset slope weighting coefficient to obtain the slope weighted value;

[0047] Multiply the slope aspect change rate by the preset slope aspect weighting coefficient to obtain the slope aspect weighted value;

[0048] The slope weighted value and the slope aspect weighted value are assigned to the corresponding vector directions, where the slope weighted value corresponds to the vertical direction and the slope aspect weighted value corresponds to the horizontal direction.

[0049] In a two-dimensional coordinate system, with the monitoring point as the center, the slope weighted value and the aspect weighted value are used as the vertical and horizontal coordinates, respectively, to draw two vectors;

[0050] In a two-dimensional coordinate system, connecting the endpoints of two vectors forms a new composite vector;

[0051] Calculate the length and direction of the composite vector, defining the length of the composite vector as the terrain change intensity and the direction of the composite vector as the terrain change direction;

[0052] The intensity and direction of terrain change at each monitoring point are recorded as terrain change vectors.

[0053] Preferably, step S2, which involves adjusting the ecological characteristics of the watershed hydrological trend analysis model after topographic feature correction based on wetland area data and vegetation cover data, includes:

[0054] Time series analysis was performed on the wetland area data of each monitoring point, and the linear regression slope of the wetland area in multiple consecutive time windows was calculated to determine the trend characteristics of the wetland area.

[0055] Time series analysis was performed on the vegetation coverage data of each monitoring point, and the average value of vegetation coverage over multiple consecutive time windows was calculated to determine the characteristics of vegetation coverage fluctuation.

[0056] By linking the trend characteristics of wetland area and the fluctuation characteristics of vegetation cover point by point, ecological characteristic data of each monitoring point are obtained.

[0057] Cluster analysis was performed on the ecological characteristic data of all monitoring points, and the watershed was divided into multiple ecological characteristic regions;

[0058] Within each ecological characteristic region, the standard deviation of the comprehensive ecological characteristic vector is calculated and used as the ecological characteristic index of that region.

[0059] Calculate the ecological correlation coefficient between ecological characteristic indicators and hydrological change data;

[0060] Ecological characteristic indicators with an absolute value of ecological correlation coefficient greater than 0.6 were selected as ecological characteristic adjustment factors.

[0061] Ecological characteristic adjustment factors are input into the watershed water situation change trend analysis model after topographic feature correction to adjust ecological characteristics.

[0062] Preferably, step S2, adjusting the watershed hydrological change trend analysis model after ecological feature adjustment to the optimal prediction accuracy state and storing the optimal prediction model parameters, includes:

[0063] Using a pre-set historical hydrological dataset, the accuracy of the watershed hydrological change trend analysis model after ecological feature adjustment was tested multiple times, and the model mean and variance of the multiple accuracy tests were calculated.

[0064] Select the set of model parameters with the highest average value and the smallest variance, and store the selected model parameters in the preset storage unit.

[0065] Preferably, in step S3, if the duration of the water situation change exceeds the preset water situation change duration, generating a dispatch warning signal and triggering a water situation change alert includes:

[0066] The duration of the water condition change detection exceeds the preset water condition change duration;

[0067] If the duration of the water condition change exceeds the preset water condition change duration, the early warning signal generation module is activated to generate a dispatch early warning signal;

[0068] At the same time, the reminder trigger module is activated to send reminders of changes in water conditions to dispatchers or equipment within the basin;

[0069] The status information of dispatch early warning signals and water condition change reminders are recorded in the system log.

[0070] Preferably, in step S4, when the monitored water level change value increases from zero to the first threshold again, the acquisition of real-time water condition characteristic information includes:

[0071] The water level changes are monitored in real time using a pre-set water level monitoring device, and the initial moment when the water level starts to increase from zero is recorded.

[0072] When the water level change value increases from zero, a timer is started to record the time interval between the water level change value increasing from zero to the first threshold.

[0073] When the water level change reaches the first threshold, the high-precision water level sensor is activated to collect the water level change value at a preset sampling frequency.

[0074] During the data collection process, the timestamp of each sampling point is recorded, and the water level change value is associated with the corresponding time information.

[0075] Time series analysis was performed on the collected water level change values ​​to extract the fluctuation characteristics of the water level changes;

[0076] The water level change data is segmented based on the fluctuation characteristics, dividing the data into multiple time intervals;

[0077] Interpolate the water level change data within each time interval to fill the data gaps between sampling points and generate real-time water condition feature information.

[0078] Preferably, in step S4, when the real-time hydrological characteristic information matches the initial hydrological characteristic information, the watershed hydrological change trend analysis model is adjusted according to the optimal prediction model parameters to perform real-time optimized scheduling operations for watershed hydrological changes, including:

[0079] Compare the water level changes between real-time hydrological information and initial hydrological information, and calculate the difference between the two.

[0080] If the difference value is less than the preset matching threshold, then the real-time hydrological feature information is determined to match the initial hydrological feature information;

[0081] Select the optimal prediction model parameters that match the current hydrological characteristics from the preset optimal prediction model parameter library;

[0082] The selected optimal prediction model parameters are substituted into the watershed hydrological change trend analysis model to update the model's prediction parameters.

[0083] The updated watershed hydrological change trend analysis model was used to recalculate the watershed hydrological change trend.

[0084] Based on the recalculated hydrological change trends, real-time dispatch instructions are generated.

[0085] Real-time dispatch instructions are sent to dispatch equipment within the basin to execute real-time optimized dispatch operations based on changes in the basin's water conditions.

[0086] The beneficial effects of the present invention are as follows:

[0087] By continuously monitoring water level changes within the watershed, when the water level change exceeds a preset first threshold, the system can promptly determine the water situation and record initial water characteristic information, including slope data, aspect data, wetland area data, and vegetation cover data. Based on the slope and aspect data, the watershed water situation trend analysis model is corrected for topographic features to ensure that the model accurately reflects the impact of the watershed's topography on water situation changes. Then, based on wetland area and vegetation cover data, the model after topographic feature correction is further adjusted for ecological characteristics, improving its adaptability to the watershed's ecological features. Through these corrections and adjustments, the watershed water situation trend analysis model is adjusted to its optimal prediction accuracy, and the optimal prediction model parameters are stored, thus providing a high-precision model foundation for subsequent water situation predictions.

[0088] When the duration of a change in water conditions exceeds the preset duration, a timely dispatching early warning signal can be generated and a water condition change alert can be triggered, providing timely early warning information for watershed dispatching and management. When the monitored water level change value decreases to zero and remains so until the preset water condition stabilization time, the watershed water condition change trend analysis model is restored to its initial prediction state, ensuring the applicability of the model in different water condition stages. When the water level change value increases from zero to the first threshold again, real-time water condition characteristic information is acquired and matched with the initial water condition characteristic information. If the match is successful, the watershed water condition change trend analysis model is adjusted according to the optimal prediction model parameters, and then real-time optimized dispatching operations for watershed water condition changes are executed. This process realizes the organic combination of real-time monitoring, accurate prediction, and optimized dispatching of watershed water condition changes, improving the scientific nature and timeliness of watershed dispatching, and effectively enhancing the efficiency of watershed water resources management and the ability to respond to water condition changes. Attached Figure Description

[0089] Fig. 1This is a flowchart illustrating the steps of a real-time optimization scheduling method based on a watershed hydrological trend analysis model.

[0090] Fig. 2 This is a schematic diagram of a watershed hydrological monitoring scenario.

[0091] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0092] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0093] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0094] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0095] To achieve the above objectives, please refer to Figs. 1-2 A real-time optimization scheduling method based on a watershed hydrological trend analysis model, the method comprising the following steps:

[0096] Step S1: Continuously monitor the water level change value within the watershed; when the water level change value exceeds the preset first threshold, it is determined to be a water situation change state and recorded as the initial water situation characteristic information; under the water situation change state, acquire slope data, slope aspect data, wetland area data and vegetation coverage data;

[0097] In this embodiment of the invention, a high-precision water level sensor network is first used to monitor the water level in the basin in real time. Water level sensors are evenly distributed at key locations in the basin, including river cross-sections, lake perimeters, and areas prone to flooding. The monitoring frequency is set to collect water level data every 5 minutes to ensure timely detection of even minor changes in water level. The sensors transmit the collected water level data to a central data processing system via a wireless communication module. The central data processing system analyzes the received water level data in real time, calculating the water level change value, i.e., the difference between the current water level and the previous water level. When the water level change value exceeds a preset first threshold, which is pre-set to 0.5 meters per hour based on historical hydrological data and basin characteristics, the system determines that the basin has entered a hydrological change state and records the current water level, rate of change, and timestamp as initial hydrological characteristic information. After determining a hydrological change state, the system automatically triggers a terrain data acquisition module to obtain slope and aspect data within the basin. Slope data is measured using multi-point tilt sensors installed within the watershed. These sensors employ high-precision accelerometers with a measurement accuracy of 0.1 degrees, accurately reflecting slope changes. Aspect data is extracted by a Geographic Information System (GIS) platform combining satellite remote sensing imagery and Digital Elevation Model (DEM) data. The GIS platform divides the watershed into 10m x 10m grids, extracts the aspect information for each grid, and stores it in a database. Simultaneously, the system utilizes a wetland area monitoring module, employing satellite remote sensing technology to acquire wetland area data within the watershed. The satellite remote sensing imagery uses optical images with a resolution better than 1 meter, and image processing algorithms are used to identify and segment wetlands. The algorithm classifies wetlands based on their spectral and textural characteristics, setting thresholds to calculate the wetland area, which is then stored in the data system in square meters. Furthermore, vegetation cover data is collected using a multispectral camera mounted on a UAV. The UAV flies at low altitude over the watershed along a pre-set flight path, and the multispectral camera captures images in red and near-infrared bands to obtain spectral reflectance information of the vegetation. After the image data is transmitted to the ground processing system, it is processed using the Normalized Difference Vegetation Index (NDVI) algorithm to calculate the vegetation coverage. The result is stored as a percentage and associated with the corresponding geographic location information.

[0098] Step S2: Based on slope and aspect data, perform terrain feature correction on the preset watershed hydrological trend analysis model; based on wetland area and vegetation coverage data, perform ecological feature adjustment on the watershed hydrological trend analysis model after terrain feature correction; adjust the watershed hydrological trend analysis model after ecological feature adjustment to the optimal prediction accuracy state and store the optimal prediction model parameters.

[0099] In this embodiment of the invention, a pre-defined watershed hydrological trend analysis model is corrected for topographic features. Using the slope and aspect data obtained in step S1, the topographic parameters in the model are adjusted through a Geographic Information System (GIS) platform. The GIS platform corrects the topographic slope factor in the model based on the slope data, dividing the slope data into multiple intervals with an accuracy of 0.1 degrees, each interval corresponding to a correction coefficient. The correction coefficients are pre-calculated based on historical hydrological data and topographic features; for example, the correction coefficient for slopes between 0° and 5° is 0.9, the correction coefficient for slopes between 5° and 10° is 0.95, and so on. Simultaneously, the water flow direction parameter in the model is corrected based on the aspect data, dividing the aspect data into 36 directional intervals with an accuracy of 10 degrees, each directional interval corresponding to a water flow direction correction value. The correction value is pre-set based on the degree of influence of topography on water flow direction. After topographic feature correction, the model can more accurately reflect the influence of watershed topography on hydrological changes. Subsequently, ecological feature adjustments were made to the watershed hydrological trend analysis model after topographic correction, based on wetland area and vegetation cover data. Wetland area data was input into the model in square meters, and the model adjusted the buffering capacity parameter for hydrological changes according to the proportion of wetland area to the total watershed area. For example, for every 1% increase in the wetland area proportion, the buffering capacity parameter in the model increased by 0.05. Vegetation cover data was input into the model as a percentage, and the model adjusted the model according to the impact of vegetation cover on soil and water conservation and runoff regulation. Specifically, for every 10% increase in vegetation cover, the soil and water conservation parameter increased by 0.1, and the runoff regulation parameter increased by 0.08. Through ecological feature adjustments, the model can more accurately reflect the impact of watershed ecological characteristics on hydrological changes. Finally, the watershed hydrological trend analysis model after ecological feature adjustments was adjusted to its optimal prediction accuracy. By comparing the model's prediction results with actual hydrological monitoring data, the root mean square error (RMSE) was used as an evaluation index to optimize and adjust the model parameters. The optimization process employs a genetic algorithm with a population size of 100, a crossover probability of 0.8, a mutation probability of 0.05, and 1000 iterations. During optimization, the model parameters are dynamically adjusted based on the RMSE (Recovery Mean Squared Error) until the RMSE reaches its minimum, at which point the model is in its optimal prediction accuracy state. These optimal model parameters are then stored in the database as the best prediction model parameters, providing accurate model support for subsequent real-time optimization scheduling.

[0100] Step S3: If the duration of the hydrological change is greater than the preset duration of the hydrological change, a dispatch warning signal is generated and a hydrological change reminder is triggered; when the water level change value is detected to decrease to zero and continues until the preset hydrological stabilization time, the basin hydrological change trend analysis model is restored to the initial prediction state.

[0101] In this embodiment of the invention, a central data processing system monitors and evaluates the duration of changes in water conditions. The system records the start time of changes in water conditions in real time and continuously tracks water level changes. The preset duration of water condition changes is set to 3 hours, a parameter predetermined based on historical water data and basin characteristics. When the duration of a water condition change exceeds 3 hours, the system automatically generates a dispatch warning signal. The dispatch warning signal is transmitted to the dispatch control center via a data communication interface, simultaneously triggering a water condition change alert. The alert is sent to a designated manager's mobile phone via SMS, containing key information such as the start time, duration, and current water level change. The SMS sending module uses a standard short message service protocol to ensure timely information delivery. Simultaneously, the system continuously monitors water level changes. When the water level change decreases to zero, the system starts a timer to record the water level stabilization time. The preset water level stabilization time is set to 2 hours, a parameter predetermined based on the typical time range for basin water condition recovery. When the water level change remains zero for 2 hours, the system determines that the basin water condition has stabilized. At this point, the system restores the watershed hydrological trend analysis model to its initial prediction state via the data interface. The restoration process includes clearing the dynamic parameter adjustment records related to the current hydrological changes in the model, reloading the initial model parameters, and adjusting the model's prediction accuracy to the initial set value. The initial model parameters are retrieved from the database to ensure that the model, after being restored to its initial state, accurately reflects the watershed's conventional hydrological characteristics, providing fundamental support for subsequent monitoring and scheduling.

[0102] Step S4: When the monitored water level change value increases from zero to the first threshold again, real-time hydrological characteristic information is obtained; when the real-time hydrological characteristic information matches the initial hydrological characteristic information, the basin hydrological change trend analysis model is adjusted according to the optimal prediction model parameters to perform real-time optimized scheduling operation of basin hydrological changes.

[0103] In this embodiment of the invention, a water level monitoring system continuously monitors water level changes within the watershed. When the monitored water level change increases from zero and reaches a preset first threshold (0.5 m / h), the system automatically triggers a real-time hydrological feature information acquisition module. This module, through a deployed sensor network including water level sensors, slope sensors, tilt sensors, and multispectral cameras, simultaneously collects current water level, slope, aspect, wetland area, and vegetation cover data as real-time hydrological feature information. Water level data is in millimeters, slope and aspect data are in degrees, wetland area data is in square meters, and vegetation cover data is recorded as a percentage. Subsequently, the system compares the real-time hydrological feature information with the initial hydrological feature information stored in the database. The comparison process is implemented using a feature matching algorithm, which sets matching thresholds for features such as water level, slope, aspect, wetland area, and vegetation cover. For example, the matching threshold for water level change is ±0.1 m / h, the matching threshold for slope and aspect is ±1 degree, the matching threshold for wetland area is ±10 square meters, and the matching threshold for vegetation coverage is ±5%. A successful match is determined when all parameters of the real-time hydrological characteristic information and the corresponding parameters of the initial hydrological characteristic information are within the set matching threshold ranges. After a successful match, the system retrieves the optimal prediction model parameters from the database. These parameters were optimized and stored in step S2. The system adjusts the watershed hydrological change trend analysis model based on the optimal prediction model parameters. The adjustment process includes updating the topographic parameters, ecological parameters, and hydrological change prediction coefficients in the model. Specifically, the topographic parameters are fine-tuned based on slope and aspect data, the ecological parameters are corrected based on wetland area and vegetation coverage data, and the hydrological change prediction coefficients are optimized based on historical matching data. The adjusted model can more accurately reflect the current watershed hydrological change trend. Finally, the system uses the adjusted watershed hydrological change trend analysis model to perform real-time optimized scheduling operations. The scheduling operations are implemented through an automated control system, which precisely controls water conservancy facilities (such as reservoir gates and pumping stations) within the basin based on model-predicted hydrological trends. For example, based on the model's predicted flood peak arrival time and flow rate, the opening and closing degree of reservoir gates is adjusted in advance to achieve reasonable flood storage and discharge. Scheduling instructions are transmitted to the control terminals of each water conservancy facility via industrial communication protocols, ensuring the timeliness and accuracy of the scheduling operations.

[0104] Preferably, the continuous monitoring of water level changes within the watershed in step S1 includes:

[0105] Multiple water level monitoring nodes are set up within the basin. The monitoring nodes are evenly distributed along the direction of water flow in the basin, and the distance between adjacent nodes is 1 / 10 of the width of the basin.

[0106] Each monitoring node uses a time-division multiplexing method to collect water level data. The first collection lasts for 5 seconds, the second for 10 seconds, and the third for 15 seconds, with each collection lasting 5 seconds longer than the previous one.

[0107] The water level data collected each time is stored in the order of collection;

[0108] Differential processing is performed on two consecutive water level data collections to obtain the water level difference change.

[0109] The arithmetic mean of the water level data in the water level difference change is obtained to obtain the water level arithmetic mean data.

[0110] The arithmetic mean water level data is segmented according to the direction of water flow in the basin to obtain the water level change gradient for each segment.

[0111] The water level change values ​​within the basin are divided according to the water level change gradient of each segment.

[0112] In this embodiment of the invention, multiple water level monitoring nodes are set up within the watershed. These nodes are evenly distributed along the direction of water flow, and the distance between adjacent nodes is set to 1 / 10 of the watershed width. For example, if the watershed width is 1000 meters, the distance between adjacent nodes is 100 meters. Each monitoring node is equipped with a high-precision water level sensor, which transmits the collected water level data to the central data processing system via a wireless communication module. Each monitoring node uses a time-division multiplexing method to collect water level data. The initial collection duration is set to 5 seconds, the second to 10 seconds, and the third to 15 seconds, with each subsequent collection duration increasing by 5 seconds. Specifically, at each monitoring node, the water level sensor first starts and continuously collects water level data for 5 seconds, then stops collecting and enters a brief sleep state; then it starts collecting again for 10 seconds, stops again, and enters sleep mode; finally, it starts collecting for 15 seconds and then stops. The collected data is transmitted in real time to the central data processing system via the wireless communication module and stored according to the collection order. The storage format includes a timestamp, monitoring node number, and corresponding water level value. The central data processing system performs differential processing on the received water level data. For each monitoring node, the difference between two adjacent water level data acquisitions is calculated to obtain the water level differential change. For example, if the water level in the first acquisition is 1.2 meters and the water level in the second acquisition is 1.3 meters, then the water level differential change is 0.1 meters. The system stores the water level differential changes of all monitoring nodes in a database, with the data format being the monitoring node number, acquisition sequence number, and water level differential change. Subsequently, the system takes the arithmetic mean of the water level data in the water level differential change data. For each monitoring node, the arithmetic mean of the water level differential changes corresponding to all its acquisition durations is calculated to obtain the water level arithmetic mean data. For example, if the water level differential changes of a certain monitoring node are 0.1 meters, 0.2 meters, and 0.15 meters, then its water level arithmetic mean is 0.15 meters. The system segments the water level arithmetic mean data according to the direction of water flow in the basin, based on the distribution location of the monitoring nodes. Assuming there are 10 monitoring nodes in the watershed, the watershed is divided into 9 segments, each containing the area between any two adjacent monitoring nodes. The system calculates the difference between the arithmetic mean water levels of all monitoring nodes within each segment to obtain the water level change gradient for each segment. For example, if the arithmetic mean water levels of the two monitoring nodes in the first segment are 0.1 meters and 0.2 meters respectively, then the water level change gradient for that segment is 0.1 meters / 100 meters (assuming a node spacing of 100 meters). Finally, the water level change values ​​within the watershed are classified according to the water level change gradient of each segment. The system sets thresholds for the water level change gradient; for example, areas with a water level change gradient less than 0.05 meters / 100 meters are classified as low water level change areas, areas with a water level change gradient between 0.05 meters / 100 meters and 0.1 meters / 100 meters are classified as medium water level change areas, and areas with a water level change gradient greater than 0.1 meters / 100 meters are classified as high water level change areas.The system stores the division results in a database, providing a basis for subsequent hydrological analysis and scheduling.

[0113] Preferably, in step S1, acquiring slope data, aspect data, wetland area data, and vegetation cover data under changing water conditions includes:

[0114] Multiple topographic sections are selected within the watershed, and multiple slope monitoring points are arranged along the direction of water flow in each section.

[0115] At each monitoring point, a tilt sensor is used to measure the tilt angle of the ground. The tilt angle at each monitoring point is used to determine the change in cross-sectional slope and obtain slope data.

[0116] Multiple terrain nodes were selected within the watershed, and multiple slope aspect monitoring points were set up at each node, with the monitoring points distributed in different directions.

[0117] At each monitoring point, a level was used to measure the slope orientation angle, and the angle data was converted into slope aspect data.

[0118] In the embodiments of this invention, please refer to Fig. 2Within the watershed, several representative topographic cross-sections are selected. Multiple slope monitoring points are evenly distributed along the slope monitoring area 102 of each cross-section, for example, 10 slope monitoring points are arranged at each topographic cross-section, with a spacing of 10 meters between the points. A high-precision tilt sensor is installed at each monitoring point to measure the ground tilt angle. The tilt sensor uses a dual-axis accelerometer with a measurement accuracy of 0.01 degrees, enabling precise measurement of the ground tilt angle. The sensor transmits the measured tilt angle data to the central data processing system in real time via its built-in wireless communication module. The system uses the received tilt angle data to determine the cross-sectional slope change in the slope aspect monitoring area 103 and calculates the slope value for each monitoring point. The slope value is calculated using the following formula: Slope (degrees) = arctan(tilt angle), and the calculation result is stored as slope data in the format of cross-section number, monitoring point number, and corresponding slope value. Multiple topographic nodes are selected within the watershed, with multiple slope aspect monitoring points arranged at each node, distributed in different directions to comprehensively cover the slope aspect characteristics of the terrain. For example, six aspect monitoring points are set up at each terrain node, pointing in the directions of north, northeast, east, southeast, south, southwest, west, and northwest. At each monitoring point, a level assembly is used to measure the slope orientation angle. The level assembly includes an electronic level and a total station; the electronic level has a measurement accuracy of 0.01 mm, and the total station has a measurement accuracy of 1 second. The electronic level measures the horizontal deviation of the slope, and the total station measures the azimuth angle of the slope. The system converts the measured horizontal deviation and azimuth angle data into aspect data. The conversion process is: aspect angle (degrees) = azimuth angle - 180 degrees (if the azimuth angle is greater than 180 degrees, then aspect angle = azimuth angle - 360 degrees). The aspect data is stored in the central data processing system in the form of monitoring point number and aspect angle, providing basic data support for subsequent terrain feature analysis.

[0119] Preferably, in step S1, acquiring slope data, aspect data, wetland area data, and vegetation cover data under changing water conditions further includes:

[0120] Multiple wetland sample areas were selected within the watershed, and multiple area monitoring points were set up in each area, with the monitoring points distributed in different locations of the wetland;

[0121] At each monitoring point, a laser rangefinder was used to measure the wetland boundary and calculate the wetland area data;

[0122] Multiple vegetation quadrat areas were selected within the watershed, and multiple vegetation monitoring points were set up in each area. The monitoring points were distributed in areas with different vegetation types.

[0123] At each monitoring point, optical sensors were used to measure the reflectance spectrum of the vegetation;

[0124] Vegetation cover data are determined by the amount of reflectance spectrum.

[0125] In the embodiments of this invention, please refer to Fig. 2 Within the watershed, 104 wetland area monitoring areas were selected, with multiple area monitoring points set up in each area, evenly distributed across different locations within the wetland. For example, nine area monitoring points were set up in each wetland sample area, forming a 3×3 grid layout. A high-precision laser rangefinder was installed at each monitoring point to measure the wetland boundary. The laser rangefinder had a measurement accuracy of 1 mm and a measurement range of 0.1 m to 1000 m. The laser rangefinder measured the straight-line distance from the emission point to the wetland boundary by emitting a laser beam to the wetland boundary and receiving the reflected light signal. The laser rangefinder at each monitoring point measured the distance to the wetland boundary in different directions, obtaining distance data in at least three directions for subsequent geometric calculations. The system calculated the wetland area using geometric calculation methods based on the measured distance data. The specific calculation method is as follows: first, the coordinates of the polygon vertices of the wetland boundary are determined based on the measured distance data; then, the wetland area is calculated using the polygon area formula. The calculation formula is:

[0126]

[0127] in( () represents the coordinates of the polygon's vertices. This represents the number of vertices. The calculated wetland area data is stored in the central data processing system in square meters. The data format is wetland sample area number, monitoring point number, and corresponding wetland area value. Please refer to [link to relevant documentation]. Fig. 2 Within the watershed, 101 vegetation cover monitoring areas were selected, with multiple vegetation monitoring points set up in each area, distributed across different vegetation types. For example, each vegetation quadrat area had 12 monitoring points, covering different vegetation types such as herbaceous plants, shrubs, and trees. At each monitoring point, an optical sensor was installed to measure the reflectance spectrum of the vegetation. The optical sensor used a multispectral camera, measuring wavelengths including red light (640-680 nm), near-infrared light (760-900 nm), and green light (520-560 nm), with a measurement accuracy of 0.1 reflectance units. The multispectral camera acquired vegetation reflectance spectral data by measuring the intensity of reflected light from the vegetation in different wavelengths. The system calculated vegetation cover based on the reflectance spectral data. The specific calculation method was as follows: first, based on the reflectance spectral characteristics of the vegetation, the Normalized Difference Vegetation Index (NDVI) was used to calculate the NDVI value for each monitoring point; then, based on the empirical relationship between the NDVI value and vegetation cover, the NDVI value was converted into vegetation cover data.

[0128] Of particular importance, determining vegetation cover data through the amount of reflectance spectrum includes:

[0129] The reflectance spectral data is segmented according to wavelength intervals, and the number of bands with reflected light intensity higher than a preset threshold within each wavelength interval is counted.

[0130] For each monitoring point, the number of bands is summed to obtain the total number of reflectance spectra for that monitoring point;

[0131] For each monitoring point, the total number of reflectance spectra is compared with the preset vegetation coverage reference value, and the total number of reflectance spectra is adjusted according to the comparison results;

[0132] The sum of the adjusted reflectance spectra is linearly fitted to the vegetation coverage reference value to obtain the vegetation coverage estimate.

[0133] The vegetation coverage data is obtained by averaging the estimated vegetation coverage values ​​of all monitoring points.

[0134] In this embodiment of the invention, when processing the reflectance spectral data of vegetation monitoring points, the reflectance spectral data is first segmented according to wavelength intervals. The wavelength range of the reflectance spectral data is set to 400 nm to 900 nm, with a wavelength interval of 10 nm. Specifically, the reflectance spectral data is divided into multiple bands, for example, 400 nm to 410 nm as the first band, 410 nm to 420 nm as the second band, and so on, up to 900 nm. For each band, the number of bands with reflected light intensity higher than a preset threshold is counted. The preset reflected light intensity threshold is preset to 0.2 reflectance units based on vegetation type and environmental conditions. If the reflected light intensity of a certain band is higher than 0.2 reflectance units, then that band is counted as a valid band. For each monitoring point, the number of bands with reflected light intensity higher than the threshold in all bands is summed to obtain the total number of reflectance spectra for that monitoring point. For example, if the reflected light intensity at a monitoring point is higher than the threshold in both the 400-410 nm and 410-420 nm wavelength bands, while the reflected light intensity in other bands is lower than the threshold, then the sum of the reflected light spectra at that monitoring point is 2. Subsequently, the sum of the reflected light spectra at each monitoring point is compared with a preset vegetation cover reference value. The vegetation cover reference value is preset based on historical data and field surveys; for example, for herbaceous vegetation areas, the reference value is set to 50. If the sum of the reflected light spectra is higher than the reference value, it is adjusted proportionally by multiplying the sum of the reflected light spectra by the ratio of the reference value to the sum of the reflected light spectra. If the sum of the reflected light spectra is lower than the reference value, it remains unchanged. The adjusted sum of the reflected light spectra is then linearly fitted to the vegetation cover reference value to obtain an estimated vegetation cover value. The linear fitting method involves multiplying the adjusted sum of the reflected light spectra by a fitting coefficient and adding a constant term. The fitting coefficient and constant term are pre-calculated based on historical data. For example, for a certain vegetation type, the fitting coefficient is 0.5, and the constant term is 10. Finally, the estimated vegetation cover values ​​of all monitoring points are averaged to obtain the vegetation cover data for that vegetation quadrat area. The averaging method is to add the estimated vegetation cover values ​​of all monitoring points and then divide by the total number of monitoring points. The calculated vegetation cover data is stored in the central data processing system as a percentage, with the data format being the vegetation quadrat area number and the corresponding vegetation cover data.

[0135] Preferably, step S2, which involves correcting the terrain features of the preset watershed hydrological trend analysis model based on slope and aspect data, includes:

[0136] Calculate the slope change rate for each monitoring point;

[0137] Calculate the rate of change of slope aspect for each monitoring point;

[0138] The slope change rate and the aspect change rate are vector-synthesized to obtain the topographic change vector.

[0139] The watershed is divided into regions with topographic change characteristics based on the magnitude and direction of the topographic change vector.

[0140] Within the region of terrain change characteristics, the magnitude and direction of the average terrain change vector are calculated and used as the terrain change characteristic index of the region.

[0141] Calculate the topographic correlation coefficient between topographic change characteristic indicators and hydrological change data;

[0142] Topographic correlation coefficients with an absolute value greater than 0.5 were selected as topographic feature correction factors.

[0143] The terrain feature correction factor is input into the preset watershed hydrological change trend analysis model to correct the terrain features.

[0144] In this embodiment of the invention, within the watershed, multiple slope and aspect monitoring points are deployed, and the slope change rate and aspect change rate for each monitoring point are calculated. The slope change rate is calculated by the ratio of the difference in slope data between adjacent monitoring points to the distance between the monitoring points. The distance between monitoring points is determined based on the topographic cross-section layout, for example, 10 meters. Assuming a monitoring point has a slope of 10 degrees and its adjacent monitoring point has a slope of 15 degrees, the slope change rate for that monitoring point is (15 degrees - 10 degrees) / 10 meters = 0.5 degrees / meter. Similarly, the aspect change rate is calculated by the ratio of the difference in aspect data between adjacent monitoring points to the distance between the monitoring points. For example, if a monitoring point has an aspect of 30 degrees east of north and its adjacent monitoring point has an aspect of 40 degrees east of north, the aspect change rate for that monitoring point is (40 degrees - 30 degrees) / 10 meters = 1 degree / meter. Subsequently, the slope change rate and aspect change rate for each monitoring point are vector-synthesized to obtain the topographic change vector. In vector synthesis, the slope change rate is considered as the ordinate component of the vector, and the aspect change rate as the abscissa component. The magnitude and direction of the synthesized vector are calculated using geometric methods. For example, if the slope change rate at a monitoring point is 0.5 degrees / meter and the aspect change rate is 1 degree / meter, then the magnitude of its topographic change vector is √(0.5² + 1²) degrees / meter, and the direction is calculated using the arctangent function, which is arctan(1 / 0.5). Based on the magnitude and direction of the topographic change vector, the watershed is divided into different topographic change characteristic regions. A threshold for the magnitude of the topographic change vector is set; for example, regions with a vector magnitude less than 1 degree / meter are classified as low topographic change regions, regions with a vector magnitude between 1 and 3 degrees / meter are classified as medium topographic change regions, and regions with a vector magnitude greater than 3 degrees / meter are classified as high topographic change regions. Simultaneously, based on the direction of the topographic change vector, the regions are further subdivided; for example, regions with a vector direction between 0 and 45 degrees are classified as northeast-facing topographic change regions. Within each topographic change characteristic region, the magnitude and direction of the average topographic change vector are calculated. The specific operation involves vector averaging of the topographic change vectors from all monitoring points within the region. This involves averaging the abscissa and ordinate components of each vector separately, and then using geometric methods to calculate the magnitude and direction of the composite vector, which serves as the topographic change characteristic index for the region. The correlation coefficient between the topographic change characteristic index and the hydrological change data is calculated. Hydrological change data includes water level change rates, and the correlation coefficient between the two is calculated using correlation analysis methods. For example, using the Pearson correlation coefficient formula, with the topographic change characteristic index as the independent variable and the hydrological change data as the dependent variable, the correlation coefficient is calculated. Topographic correlation coefficients with an absolute value greater than 0.5 are selected as topographic feature correction factors. These topographic feature correction factors are then input into a pre-defined watershed hydrological change trend analysis model to correct the topographic correlation parameters in the model.The correction process involves adjusting the prediction parameters related to topographic changes in the model based on the magnitude and direction of the correction factor for topographic features, so that the model can more accurately reflect the impact of topographic changes on watershed hydrological changes.

[0145] Of particular importance is the vector synthesis of the slope change rate and the aspect change rate, including;

[0146] Multiple topographic monitoring points were selected within the watershed, and the slope change rate and aspect change rate of each monitoring point were measured using a topographic measuring instrument.

[0147] For each monitoring point, the measured slope change rate and aspect change rate are recorded as two independent values.

[0148] Multiply the slope change rate by the preset slope weighting coefficient to obtain the slope weighted value;

[0149] Multiply the slope aspect change rate by the preset slope aspect weighting coefficient to obtain the slope aspect weighted value;

[0150] The slope weighted value and the slope aspect weighted value are assigned to the corresponding vector directions, where the slope weighted value corresponds to the vertical direction and the slope aspect weighted value corresponds to the horizontal direction.

[0151] In a two-dimensional coordinate system, with the monitoring point as the center, the slope weighted value and the aspect weighted value are used as the vertical and horizontal coordinates, respectively, to draw two vectors;

[0152] In a two-dimensional coordinate system, connecting the endpoints of two vectors forms a new composite vector;

[0153] Calculate the length and direction of the composite vector, defining the length of the composite vector as the terrain change intensity and the direction of the composite vector as the terrain change direction;

[0154] The intensity and direction of terrain change at each monitoring point are recorded as terrain change vectors.

[0155] In this embodiment of the invention, multiple topographic monitoring points are selected within the watershed, and a high-precision topographic surveying instrument is used to measure the slope change rate and aspect change rate at each monitoring point. The topographic surveying instrument employs a total station or a 3D laser scanner, achieving a measurement accuracy of 0.01 degrees. For each monitoring point, the instrument first determines the slope change rate, which is the difference in slope between adjacent monitoring points divided by the horizontal distance between the two points. For example, if the horizontal distance between adjacent monitoring points is 10 meters and the slope difference is 2 degrees, then the slope change rate is 0.2 degrees / meter. Simultaneously, the instrument determines the aspect change rate, which is the difference in aspect between adjacent monitoring points divided by the horizontal distance between the two points. For example, if the aspect difference is 3 degrees, then the aspect change rate is 0.3 degrees / meter. The measured slope change rate and aspect change rate are recorded as two independent values ​​and stored in the data acquisition system. Preset slope weighting coefficients and aspect weighting coefficients are pre-set based on topographic features and watershed characteristics. For example, the slope weighting coefficient is set to 1.5, and the aspect weighting coefficient is set to 1.0. For each monitoring point, the measured slope change rate is multiplied by a slope weighting coefficient to obtain a slope-weighted value. For example, if the slope change rate at a monitoring point is 0.2 degrees / meter, the slope weighting value is 0.2 degrees / meter × 1.5 = 0.3. Simultaneously, the aspect change rate is multiplied by an aspect weighting coefficient to obtain an aspect-weighted value. For example, if the aspect change rate is 0.3 degrees / meter, the aspect weighting value is 0.3 degrees / meter × 1.0 = 0.3. The slope weighting value is assigned to the vertical direction, and the aspect weighting value to the horizontal direction. In a two-dimensional coordinate system, with the monitoring point as the center, two vectors are plotted, with the slope weighting value as the ordinate and the aspect weighting value as the abscissa. For example, if the slope weighting value and aspect weighting value of a monitoring point are both 0.3, then in a two-dimensional coordinate system, starting from the monitoring point, draw a vector of length 0.3 upwards (representing the slope weighting value) and another vector of length 0.3 to the right (representing the aspect weighting value). Connect the endpoints of the two vectors in the two-dimensional coordinate system to form a new composite vector. Calculate the length and direction of the composite vector. The length of the composite vector is calculated using the Pythagorean theorem, that is, the length is equal to the square root of the sum of the squares of the slope weighting value and the aspect weighting value. For example, the length of the composite vector is √(0.3² + 0.3²) = 0.424. The direction of the composite vector is calculated using the arctangent function, that is, the direction is arctan(slope weighting value / aspect weighting value). For example, the direction is arctan(0.3 / 0.3) = 45 degrees. Define the length of the composite vector as the terrain change intensity and the direction of the composite vector as the terrain change direction. The intensity and direction of terrain change at each monitoring point are recorded as terrain change vectors and stored in the central data processing system. The data format is monitoring point number, terrain change intensity, and terrain change direction.

[0156] Preferably, step S2, which involves adjusting the ecological characteristics of the watershed hydrological trend analysis model after topographic feature correction based on wetland area data and vegetation cover data, includes:

[0157] Time series analysis was performed on the wetland area data of each monitoring point, and the linear regression slope of the wetland area in multiple consecutive time windows was calculated to determine the trend characteristics of the wetland area.

[0158] Time series analysis was performed on the vegetation coverage data of each monitoring point, and the average value of vegetation coverage over multiple consecutive time windows was calculated to determine the characteristics of vegetation coverage fluctuation.

[0159] By linking the trend characteristics of wetland area and the fluctuation characteristics of vegetation cover point by point, ecological characteristic data of each monitoring point are obtained.

[0160] Cluster analysis was performed on the ecological characteristic data of all monitoring points, and the watershed was divided into multiple ecological characteristic regions;

[0161] Within each ecological characteristic region, the standard deviation of the comprehensive ecological characteristic vector is calculated and used as the ecological characteristic index of that region.

[0162] Calculate the ecological correlation coefficient between ecological characteristic indicators and hydrological change data;

[0163] Ecological characteristic indicators with an absolute value of ecological correlation coefficient greater than 0.6 were selected as ecological characteristic adjustment factors.

[0164] Ecological characteristic adjustment factors are input into the watershed water situation change trend analysis model after topographic feature correction to adjust ecological characteristics.

[0165] In this embodiment of the invention, time series analysis is performed on wetland area data at each monitoring point within the watershed. Specifically, the wetland area data is arranged chronologically, and multiple consecutive time windows are selected for analysis. Each time window is set to a length of 30 days, with a sliding step of 10 days. For example, the first time window is from day 1 to day 30, the second time window is from day 11 to day 40, and so on. Within each time window, linear regression analysis is performed on the wetland area data, and the linear regression slope is calculated to determine the trend characteristics of the wetland area. The linear regression slope reflects the changing trend of the wetland area within that time window; a positive slope indicates an increase in wetland area, and a negative slope indicates a decrease. Time series analysis is also performed on vegetation cover data at each monitoring point. The average vegetation cover is calculated within multiple consecutive time windows to determine the fluctuation characteristics of vegetation cover. The length and sliding step of each time window are the same as those for the wetland area analysis, both being 30 days and 10 days. Within each time window, the sum of all vegetation cover data within that window is divided by the number of data points to obtain the average vegetation cover value for that time window, thus reflecting the fluctuation of vegetation cover within that time window. The wetland area trend characteristics (linear regression slope) and vegetation cover fluctuation characteristics (average value) of each monitoring point are correlated point-by-point to form the ecological characteristic data for each monitoring point. The ecological characteristic data is stored in the data processing system in the form of monitoring point number, wetland area trend characteristic value, and vegetation cover fluctuation characteristic value. Cluster analysis is performed on the ecological characteristic data of all monitoring points to divide the watershed into multiple ecological characteristic regions. The K-means clustering algorithm is used for cluster analysis, with a pre-set number of clusters of 5, meaning the watershed is divided into 5 ecological characteristic regions. During the clustering process, the wetland area trend characteristic value and vegetation cover fluctuation characteristic value are used as the feature dimensions of the clusters. By calculating the distance between each monitoring point and the cluster center, the monitoring points are assigned to different ecological characteristic regions. Within each ecological characteristic region, the standard deviation of the comprehensive ecological characteristic vector is calculated as the ecological characteristic index for that region. The comprehensive ecological characteristic vector consists of the trend characteristic values ​​of wetland area and the fluctuation characteristic values ​​of vegetation cover at all monitoring points within the region. Specifically, the mean of the trend characteristic values ​​of wetland area and the fluctuation characteristic values ​​of vegetation cover at each monitoring point within the region is first calculated. Then, the difference between the characteristic value and the mean for each monitoring point is calculated. Finally, the standard deviation of these differences is calculated to obtain the ecological characteristic index of the region. The ecological correlation coefficient between the ecological characteristic index and the hydrological change data is calculated. The hydrological change data includes water level change rate, flow change rate, etc. The correlation coefficient between the ecological characteristic index and the hydrological change data is calculated using correlation analysis. The correlation coefficient is calculated using the Pearson correlation coefficient formula, with the ecological characteristic index as the independent variable and the hydrological change data as the dependent variable. Ecological characteristic indicators with an absolute value of ecological correlation coefficient greater than 0.6 are selected as ecological characteristic adjustment factors.These ecological characteristic adjustment factors are input into the watershed hydrological change trend analysis model after topographic feature correction, and the ecologically related parameters in the model are adjusted. The adjustment process involves adjusting the prediction parameters related to ecological characteristics in the model according to the magnitude and direction of the ecological characteristic adjustment factors, so that the model can more accurately reflect the impact of ecological characteristics on watershed hydrological changes.

[0166] Preferably, step S2, adjusting the watershed hydrological change trend analysis model after ecological feature adjustment to the optimal prediction accuracy state and storing the optimal prediction model parameters, includes:

[0167] Using a pre-set historical hydrological dataset, the accuracy of the watershed hydrological change trend analysis model after ecological feature adjustment was tested multiple times, and the model mean and variance of the multiple accuracy tests were calculated.

[0168] Select the set of model parameters with the highest average value and the smallest variance, and store the selected model parameters in the preset storage unit.

[0169] In this embodiment of the invention, a preset historical hydrological dataset is used to conduct multiple accuracy tests on a watershed hydrological change trend analysis model adjusted for ecological characteristics. The historical hydrological dataset includes data on water level changes, flow changes, and rainfall over the past few years, with a data interval of one hour, covering the entire watershed. The accuracy test is conducted 10 times, with data from different time periods in the dataset selected as test samples for each test to ensure sample diversity and representativeness. In each accuracy test, the test samples are input into the model, and the model outputs predicted hydrological change trends, including predicted water level changes, flow changes, and other indicators. Simultaneously, the model's predicted results are compared with the actual hydrological data in the test samples to calculate the model's accuracy index. The accuracy index uses two statistical methods: root mean square error (RMSE) and coefficient of determination (R²). RMSE measures the deviation between predicted and actual values, while R² measures the model's goodness of fit to the actual data. Specifically, for each test sample, the sum of squares of the differences between predicted and actual values ​​is calculated, then the average is taken and the square root is taken to obtain the RMSE; simultaneously, the R² value is calculated to evaluate the model's interpretability of the data. After completing 10 accuracy tests, calculate the model mean and model variance for each test. The model mean is the arithmetic mean of the accuracy metrics (RMSE and R²) for each test, and the model variance is the variance of the accuracy metrics for each test. Specifically, sum the RMSE values ​​from all 10 tests and divide by 10 to obtain the average RMSE; sum the R² values ​​from all 10 tests and divide by 10 to obtain the average R². Simultaneously, calculate the sum of the squares of the differences between each RMSE value and the average RMSE, and then divide by 10 to obtain the variance of the RMSE; similarly, calculate the variance of the R². From the results of the 10 tests, select the set of model parameters with the highest average and lowest variance. The specific criterion is: select the set of parameters with the lowest average RMSE and lowest variance across all tests, while ensuring that the corresponding R² average is the highest and the variance is the lowest. This set of parameters guarantees high prediction accuracy and stability for the model across different test samples. Store the selected model parameters in a preset storage unit. The storage unit uses non-volatile storage media, such as solid-state drives or flash memory chips, to ensure that model parameters are not lost after power failure or system restart. The storage process is completed through data write instructions, storing the model parameters in binary format to the specified address of the storage unit. After storage is completed, the system verifies the storage unit to ensure the integrity and accuracy of the model parameters.

[0170] Preferably, in step S3, if the duration of the water situation change exceeds the preset water situation change duration, generating a dispatch warning signal and triggering a water situation change alert includes:

[0171] The duration of the water condition change detection exceeds the preset water condition change duration;

[0172] If the duration of the water condition change exceeds the preset water condition change duration, the early warning signal generation module is activated to generate a dispatch early warning signal;

[0173] At the same time, the reminder trigger module is activated to send reminders of changes in water conditions to dispatchers or equipment within the basin;

[0174] The status information of dispatch early warning signals and water condition change reminders are recorded in the system log.

[0175] In this embodiment of the invention, the watershed hydrological monitoring system detects in real time whether the duration of a hydrological change exceeds a preset duration. The duration of the hydrological change is calculated by a central data processing system, which records the start time of the hydrological change and continuously tracks its duration. The preset duration is 3 hours, a parameter pre-set based on historical hydrological data and watershed characteristics and stored in the system parameter configuration file. When the duration of the hydrological change exceeds the preset 3 hours, the system automatically activates the warning signal generation module. This module uses digital signal processing technology to generate a dispatch warning signal conforming to a preset format. This signal contains key information such as the start time, duration, and current water level change of the hydrological change, generated in binary code, and transmitted to the signal output interface via the system's internal communication bus. Simultaneously, the reminder trigger module is activated to send hydrological change reminders to dispatchers or equipment within the watershed. The reminder trigger module implements the reminder function through an SMS gateway and device communication interface. For dispatchers, the system sends alerts via SMS to a pre-defined list of mobile phone numbers through an SMS gateway. The SMS content includes key information about the water situation and the alert level. For equipment, the system sends alert signals in data frames through the equipment communication interface. These data frames contain the equipment identifier, alert type, and water situation status information, ensuring that the equipment can receive and respond to the alert signals. The status information of dispatch warning signals and water situation alerts is recorded in the system log. The system log module uses timestamp technology to record the time, content, and target of each warning signal generation and alert transmission. Log information is stored in text format in the system log file, with the file path and format pre-defined in the system configuration file. Log records include the warning signal generation time, content summary, alert transmission time, list of receiving mobile phone numbers, and equipment identifier, for subsequent querying and auditing.

[0176] Preferably, in step S4, when the monitored water level change value increases from zero to the first threshold again, the acquisition of real-time water condition characteristic information includes:

[0177] The water level changes are monitored in real time using a pre-set water level monitoring device, and the initial moment when the water level starts to increase from zero is recorded.

[0178] When the water level change value increases from zero, a timer is started to record the time interval between the water level change value increasing from zero to the first threshold.

[0179] When the water level change reaches the first threshold, the high-precision water level sensor is activated to collect the water level change value at a preset sampling frequency.

[0180] During the data collection process, the timestamp of each sampling point is recorded, and the water level change value is associated with the corresponding time information.

[0181] Time series analysis was performed on the collected water level change values ​​to extract the fluctuation characteristics of the water level changes;

[0182] The water level change data is segmented based on the fluctuation characteristics, dividing the data into multiple time intervals;

[0183] Interpolate the water level change data within each time interval to fill the data gaps between sampling points and generate real-time water condition feature information.

[0184] In this embodiment of the invention, a preset water level monitoring device monitors water level changes in real time. This device employs a high-precision water level sensor with a measurement accuracy at the millimeter level and a monitoring frequency of once per second. The system records the initial moment when the water level begins to increase from zero. Specifically, when the water level change is detected to increase from zero, a timestamp is recorded at that moment, accurate to the millisecond level. When the water level change begins to increase from zero, the system starts a timer with an accuracy of 1 millisecond. The timer records the time interval from the water level change increasing from zero to a first threshold, which is preset to 50 millimeters. The timer stops timing when the water level change reaches the first threshold and records the time interval, stored in the system data recording module in milliseconds. When the water level change reaches the first threshold, the system activates a high-precision water level sensor with a measurement accuracy of 0.1 millimeters and a sampling frequency preset to 10 times per second. The sensor collects water level change values ​​at the preset sampling frequency and associates each collected water level change value with a corresponding timestamp, provided by the system's internal clock, accurate to the millisecond level. The collected water level change values ​​and timestamps are stored as data pairs in the system data buffer. Time series analysis is performed on the collected water level change values, using Fourier transform to extract the fluctuation characteristics of the water level changes. Specifically, the collected water level change value sequence is subjected to Fourier transform to calculate its spectrum, extracting the main frequency components and corresponding amplitudes, thereby determining the fluctuation characteristics of the water level changes, including the fluctuation frequency and amplitude. Based on the fluctuation characteristics, the water level change value data is segmented into multiple time intervals. Specifically, the time series is divided into multiple time intervals of equal length according to the fluctuation frequency, with the length of each time interval being an integer multiple of the fluctuation period. For example, if the fluctuation frequency is 0.1 Hz, the fluctuation period is 10 seconds, and the time interval length can be set to 30 seconds. Interpolation is performed on the water level change value data within each time interval to fill data gaps between sampling points. A linear interpolation method is used to calculate the water level change values ​​corresponding to time points between adjacent sampling points based on the known water level change values ​​and timestamps of the sampling points. For example, if the timestamps of two sampling points are 1 second and 2 seconds, and the water level changes are 10 mm and 15 mm, respectively, then the interpolated water level change at 1.5 seconds is 12.5 mm. The time series data generated by interpolation is used as real-time hydrological characteristic information and stored in the system's data processing module for subsequent hydrological analysis and scheduling decisions.

[0185] Preferably, in step S4, when the real-time hydrological characteristic information matches the initial hydrological characteristic information, the watershed hydrological change trend analysis model is adjusted according to the optimal prediction model parameters to perform real-time optimized scheduling operations for watershed hydrological changes, including:

[0186] Compare the water level changes between real-time hydrological information and initial hydrological information, and calculate the difference between the two.

[0187] If the difference value is less than the preset matching threshold, then the real-time hydrological feature information is determined to match the initial hydrological feature information;

[0188] Select the optimal prediction model parameters that match the current hydrological characteristics from the preset optimal prediction model parameter library;

[0189] The selected optimal prediction model parameters are substituted into the watershed hydrological change trend analysis model to update the model's prediction parameters.

[0190] The updated watershed hydrological change trend analysis model was used to recalculate the watershed hydrological change trend.

[0191] Based on the recalculated hydrological change trends, real-time dispatch instructions are generated.

[0192] Real-time dispatch instructions are sent to dispatch equipment within the basin to execute real-time optimized dispatch operations based on changes in the basin's water conditions.

[0193] In this embodiment of the invention, water level sensors deployed at key sections of the watershed collect current water level data in real time, which is recorded as the real-time water level value in the real-time hydrological feature information. At the same time, the initial water level data of the section within the historical scheduling cycle is extracted from the database and recorded as the initial water level value in the initial hydrological feature information. The real-time water level value is subtracted from the initial water level value to obtain the water level change value, which is recorded as the water level change amount. A matching threshold of 0.05 meters is set. If the absolute value of the water level change amount is less than 0.05 meters, it is determined that the real-time hydrological feature information matches the initial hydrological feature information. The optimal prediction model parameters that match the current hydrological characteristics are retrieved from a pre-defined optimal prediction model parameter library. This parameter library stores model parameter sets corresponding to different hydrological characteristics, each set including rainfall-runoff coefficient, evaporation coefficient, soil permeability coefficient, channel roughness coefficient, and confluence time coefficient. Based on the rainfall intensity, soil moisture, and channel flow in the current hydrological characteristics, the corresponding rainfall-runoff coefficient, evaporation coefficient, soil permeability coefficient, channel roughness coefficient, and confluence time coefficient parameter values ​​are selected from the parameter library using a parameter matching algorithm, and these are recorded as the optimal prediction model parameter set. The parameters from the optimal prediction model parameter set are substituted into the watershed hydrological change trend analysis model to replace the original model parameters, completing the model parameter update operation. The updated model parameters include updated rainfall-runoff coefficient, evaporation coefficient, soil permeability coefficient, channel roughness coefficient, and confluence time coefficient. Using the updated watershed hydrological trend analysis model, input the current rainfall intensity, upstream inflow, downstream water level, soil moisture, and evaporation. The model outputs a sequence of watershed water level changes over the next 24 hours, denoted as the water level change trend sequence. This sequence includes hourly water level predictions, with the time index ranging from one to twenty-four. Based on the predicted water level values ​​in the water level change trend sequence and combined with the scheduling strategies in the basin scheduling rule base, real-time scheduling instructions are generated. These instructions include gate opening degree, pump station start / stop status, flood diversion gate opening time, and flood storage / detention area activation flag. The gate opening degree is calculated based on the difference between the predicted water level and the target water level, using the formula: gate opening degree equals adjustment coefficient multiplied by the target water level difference. The adjustment coefficient ranges from 0.1 to 0.5. The pump station start / stop status is determined based on whether the predicted water level exceeds the warning level. If the predicted water level is greater than the warning level, the pump station is set to start; otherwise, it is set to stop. The flood diversion gate opening time is determined based on the time when the predicted water level reaches the flood diversion level. The flood storage / detention area activation flag is set as a Boolean value based on whether the predicted water level exceeds the flood storage / detention area activation level.The generated real-time dispatch instructions are sent to the target dispatch equipment, including gate controllers, pump station control systems, flood diversion gate control terminals, and flood storage and detention area control units, through the basin dispatch communication network. After receiving the instructions, the dispatch equipment performs corresponding operations, including adjusting the gate opening to a specified value, starting or stopping the pump station, opening the flood diversion gate at a specified time, and activating the flood storage and detention area; thus completing the real-time optimized dispatch operation for changes in the basin's water conditions.

[0194] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0195] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A real-time optimization scheduling method based on a watershed hydrological trend analysis model, characterized in that, Includes the following steps: Step S1: Continuously monitor the water level change value within the watershed; when the water level change value exceeds the preset first threshold, it is determined to be a water situation change state and recorded as the initial water situation characteristic information; under the water situation change state, acquire slope data, slope aspect data, wetland area data and vegetation coverage data; Step S2: Based on slope and aspect data, perform terrain feature correction on the preset watershed hydrological trend analysis model; based on wetland area and vegetation coverage data, perform ecological feature adjustment on the watershed hydrological trend analysis model after terrain feature correction; adjust the watershed hydrological trend analysis model after ecological feature adjustment to the optimal prediction accuracy state and store the optimal prediction model parameters. Step S3: If the duration of the hydrological change is greater than the preset duration of the hydrological change, a dispatch warning signal is generated and a hydrological change reminder is triggered; when the water level change value is detected to decrease to zero and continues until the preset hydrological stabilization time, the basin hydrological change trend analysis model is restored to the initial prediction state. Step S4: When the water level change value is detected to increase from zero to the first threshold again, real-time water condition characteristic information is obtained; When real-time hydrological characteristics match initial hydrological characteristics, the basin hydrological change trend analysis model is adjusted according to the optimal prediction model parameters to perform real-time optimized scheduling operations for basin hydrological changes.

2. The real-time optimization scheduling method based on the watershed hydrological change trend analysis model according to claim 1, characterized in that, Step S1 involves continuously monitoring water level changes within the watershed, including: Multiple water level monitoring nodes are set up within the basin. The monitoring nodes are evenly distributed along the direction of water flow in the basin, and the distance between adjacent nodes is 1 / 10 of the width of the basin. Each monitoring node uses a time-division multiplexing method to collect water level data. The first collection lasts for 5 seconds, the second for 10 seconds, and the third for 15 seconds, with each collection lasting 5 seconds longer than the previous one. The water level data collected each time is stored in the order of collection; Differential processing is performed on two consecutive water level data collections to obtain the water level difference change. The arithmetic mean of the water level data in the water level difference change is obtained to obtain the water level arithmetic mean data. The arithmetic mean water level data is segmented according to the direction of water flow in the basin to obtain the water level change gradient for each segment. The water level change values ​​within the basin are divided according to the water level change gradient of each segment.

3. The real-time optimization scheduling method based on the watershed hydrological change trend analysis model according to claim 1, characterized in that, Step S1, under the condition of water situation change, includes obtaining slope data, aspect data, wetland area data, and vegetation cover data, including: Multiple topographic sections are selected within the watershed, and multiple slope monitoring points are arranged along the direction of water flow in each section. At each monitoring point, a tilt sensor is used to measure the tilt angle of the ground. The tilt angle at each monitoring point is used to determine the change in cross-sectional slope and obtain slope data. Multiple terrain nodes were selected within the watershed, and multiple slope aspect monitoring points were set up at each node, with the monitoring points distributed in different directions. At each monitoring point, a level was used to measure the slope orientation angle, and the angle data was converted into slope aspect data.

4. The real-time optimization scheduling method based on the watershed hydrological change trend analysis model according to claim 3, characterized in that, Step S1, under the condition of water situation change, also includes obtaining slope data, aspect data, wetland area data, and vegetation cover data, which further includes: Multiple wetland sample areas were selected within the watershed, and multiple area monitoring points were set up in each area, with the monitoring points distributed in different locations of the wetland; At each monitoring point, a laser rangefinder was used to measure the wetland boundary and calculate the wetland area data; Multiple vegetation quadrat areas were selected within the watershed, and multiple vegetation monitoring points were set up in each area. The monitoring points were distributed in areas with different vegetation types. At each monitoring point, optical sensors were used to measure the reflectance spectrum of the vegetation; Vegetation cover data are determined by the amount of reflectance spectrum.

5. The real-time optimization scheduling method based on the watershed hydrological change trend analysis model according to claim 1, characterized in that, Step S2 involves correcting the terrain features of the pre-set watershed hydrological trend analysis model based on slope and aspect data, including: Calculate the slope change rate for each monitoring point; Calculate the rate of change of slope aspect for each monitoring point; The slope change rate and the aspect change rate are vector-synthesized to obtain the topographic change vector. The watershed is divided into regions with topographic change characteristics based on the magnitude and direction of the topographic change vector. Within the region of terrain change characteristics, the magnitude and direction of the average terrain change vector are calculated and used as the terrain change characteristic index of the region. Calculate the topographic correlation coefficient between topographic change characteristic indicators and hydrological change data; Topographic correlation coefficients with an absolute value greater than 0.5 were selected as topographic feature correction factors. The terrain feature correction factor is input into the preset watershed hydrological change trend analysis model to correct the terrain features.

6. The real-time optimization scheduling method based on the watershed hydrological change trend analysis model according to claim 1, characterized in that, Step S2 involves adjusting the ecological characteristics of the watershed hydrological trend analysis model after topographic feature correction based on wetland area data and vegetation cover data. This includes: Time series analysis was performed on the wetland area data of each monitoring point, and the linear regression slope of the wetland area in multiple consecutive time windows was calculated to determine the trend characteristics of the wetland area. Time series analysis was performed on the vegetation coverage data of each monitoring point, and the average value of vegetation coverage over multiple consecutive time windows was calculated to determine the characteristics of vegetation coverage fluctuation. By linking the trend characteristics of wetland area and the fluctuation characteristics of vegetation cover point by point, ecological characteristic data of each monitoring point are obtained. Cluster analysis was performed on the ecological characteristic data of all monitoring points, and the watershed was divided into multiple ecological characteristic regions; Within each ecological characteristic region, the standard deviation of the comprehensive ecological characteristic vector is calculated and used as the ecological characteristic index of that region. Calculate the ecological correlation coefficient between ecological characteristic indicators and hydrological change data; Ecological characteristic indicators with an absolute value of ecological correlation coefficient greater than 0.6 were selected as ecological characteristic adjustment factors. Ecological characteristic adjustment factors are input into the watershed water situation change trend analysis model after topographic feature correction to adjust ecological characteristics.

7. The real-time optimization scheduling method based on the watershed hydrological change trend analysis model according to claim 1, characterized in that, Step S2 involves adjusting the watershed hydrological trend analysis model, after ecological feature adjustment, to its optimal prediction accuracy and storing the optimal prediction model parameters, including: Using a pre-set historical hydrological dataset, the accuracy of the watershed hydrological change trend analysis model after ecological feature adjustment was tested multiple times, and the model mean and variance of the multiple accuracy tests were calculated. Select the set of model parameters with the highest average value and the smallest variance, and store the selected model parameters in the preset storage unit.

8. The real-time optimization scheduling method based on the watershed hydrological change trend analysis model according to claim 1, characterized in that, If the duration of the water situation change in step S3 is greater than the preset duration of the water situation change, then a dispatching early warning signal is generated and a water situation change alert is triggered, including: The duration of the water condition change detection exceeds the preset water condition change duration; If the duration of the water condition change exceeds the preset water condition change duration, the early warning signal generation module is activated to generate a dispatch early warning signal; At the same time, the reminder trigger module is activated to send reminders of changes in water conditions to dispatchers or equipment within the basin; The status information of dispatch early warning signals and water condition change reminders are recorded in the system log.

9. The real-time optimization scheduling method based on the watershed hydrological change trend analysis model according to claim 1, characterized in that, In step S4, when the monitored water level change value increases from zero to the first threshold again, the real-time water situation characteristic information obtained includes: The water level changes are monitored in real time using a pre-set water level monitoring device, and the initial moment when the water level starts to increase from zero is recorded. When the water level change value increases from zero, a timer is started to record the time interval between the water level change value increasing from zero to the first threshold. When the water level change reaches the first threshold, the high-precision water level sensor is activated to collect the water level change value at a preset sampling frequency. During the data collection process, the timestamp of each sampling point is recorded, and the water level change value is associated with the corresponding time information. Time series analysis was performed on the collected water level change values ​​to extract the fluctuation characteristics of the water level changes; The water level change data is segmented based on the fluctuation characteristics, dividing the data into multiple time intervals; Interpolate the water level change data within each time interval to fill the data gaps between sampling points and generate real-time water condition feature information.

10. The real-time optimization scheduling method based on the watershed hydrological change trend analysis model according to claim 1, characterized in that, In step S4, when the real-time hydrological characteristic information matches the initial hydrological characteristic information, the basin hydrological change trend analysis model is adjusted according to the optimal prediction model parameters to perform real-time optimized scheduling operations for basin hydrological changes, including: Compare the water level changes between real-time hydrological information and initial hydrological information, and calculate the difference between the two. If the difference value is less than the preset matching threshold, then the real-time hydrological feature information is determined to match the initial hydrological feature information; Select the optimal prediction model parameters that match the current hydrological characteristics from the preset optimal prediction model parameter library; The selected optimal prediction model parameters are substituted into the watershed hydrological change trend analysis model to update the model's prediction parameters. The updated watershed hydrological change trend analysis model was used to recalculate the watershed hydrological change trend. Based on the recalculated hydrological change trends, real-time dispatch instructions are generated. Real-time dispatch instructions are sent to dispatch equipment within the basin to execute real-time optimized dispatch operations based on changes in the basin's water conditions.

Images