Water resource management method and system based on big data analysis

Abstract

The application discloses a kind of water resource management method and system based on big data analysis, belong to water resource management technical field.The method includes: obtaining the multi-source hydrological monitoring data sequence of target river basin, and executing time sequence semantic alignment processing, generate the standardized data matrix after alignment;According to the standardized data matrix, construct time-varying water level response function, calculate the theoretical water demand urgency of each water intake node;Based on the deviation of theoretical water demand urgency and instantaneous flow reading, generate water resources deployment initial scheme;Obtain the ecological sensitive area distribution map of target river basin, overlay the deployment initial scheme with ecological sensitive area distribution map and carry out flow path deduction, determine abnormal scheduling risk node;Scheduling authority verification is carried out to abnormal scheduling risk node, and the output water resource optimization scheduling instruction sequence.Through the above steps, consistency basis can be obtained from heterogeneous asynchronous monitoring data, and ecological constraints and execution conditions are considered in scheduling.

Description

Technical Field

[0001] This invention relates to the field of water resource management technology, specifically to a water resource management method and system based on big data analysis. Background Technology

[0002] Watershed water resources management requires the integration of multi-source monitoring data, including precipitation, water level, groundwater, and water intake flow, to formulate scheduling schemes that balance supply and demand. Existing management methods typically utilize raw data reported by each monitoring station at its own sampling frequency, performing simple threshold comparisons or fixed-rule scheduling. When data from some rain gauges is missing, that period is often discarded or the average value of the entire period is used to fill the gap. The problem of asynchronous precipitation sequences with river water levels and groundwater depth measurements has not been addressed, leading to misalignments in time and semantics between data from different sources. This misalignment results in lag time estimation biases in the subsequently established precipitation-water level response relationship, causing the estimated water urgency at each water intake node to differ from the actual water shortage situation, directly affecting the accuracy of the scheduling scheme's response to actual drought risks. Furthermore, existing schemes, when generating allocation plans, often focus on numerical matching of water supply and demand gaps, failing to incorporate the spatial distribution of downstream ecologically sensitive areas and their minimum ecological flow constraints into the flow extrapolation process. This can easily lead to insufficient ecological base flow in local river sections after water release. Meanwhile, the generated dispatch instructions are directly issued to control facilities such as sluice gates, lacking an active verification process for the current operating status parameters of the control station and remote control permissions. This results in instructions being unable to be executed due to limitations of on-site equipment or permissions, requiring repeated manual confirmation and adjustment, which reduces dispatch efficiency.

[0003] Eliminating the interference of time-series discrepancies and missing values ​​in multi-source hydrological data on the accuracy of water demand assessment, and simultaneously considering the flow guarantee requirements and engineering control conditions of ecologically sensitive areas in the allocation plan, have become problems that need to be solved in this field. Summary of the Invention

[0004] The purpose of this invention is to provide a water resource management method and system based on big data analysis, in order to solve the problems of deviation in the assessment of water urgency caused by time asynchrony and missing data from multiple sources of hydrological data, and the problems that allocation schemes are prone to causing ecological risks and are difficult to execute automatically.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a water resource management method based on big data analysis, comprising: Acquire multi-source hydrological monitoring data sequences for the target watershed, which include precipitation observation records, river water level changes, groundwater depth dynamics, and instantaneous flow readings at each water intake node; Temporal semantic alignment processing is performed on multi-source hydrological monitoring data sequences to generate aligned standardized data matrices, thereby eliminating inconsistencies in the temporal and spatial dimensions of multi-source data and providing a unified data foundation for subsequent analysis; Based on the standardized data matrix, a time-varying water level response function for the target watershed is constructed, and the theoretical water demand urgency of each water intake node is calculated using the time-varying water level response function, thereby quantifying the degree of water shortage risk of different water intake nodes during the prediction period. Based on the deviation between the theoretical water demand urgency and the instantaneous flow readings of each water intake node, an initial water resource allocation plan for the target watershed is generated, so that the allocation plan can prioritize the water intake node with the most urgent water demand and improve the targeting of water resource allocation. Obtain a map showing the distribution of ecologically sensitive areas in the target watershed. Overlay the initial water resource allocation plan with the map to perform flow path extrapolation to identify abnormal scheduling risk nodes. This will allow for proactive identification of scheduling links that may trigger ecological risks while ensuring water intake needs are met. The system verifies the scheduling permissions of nodes with abnormal scheduling risks, outputs a sequence of water resource optimization scheduling instructions for the target watershed, ensures that the scheduling instructions can be executed within the equipment's capabilities and permissions, and issues alarms or alternative scheduling for nodes that cannot be executed, thereby achieving a dynamic balance between water resource scheduling and ecological protection.

[0006] As a technical solution of the present invention, the process of performing temporal semantic alignment processing on multi-source hydrological monitoring data sequences to generate an aligned standardized data matrix includes: dividing precipitation observation records into independent precipitation time series of multiple rain gauge stations according to spatial distribution; detecting missing values ​​in each independent precipitation time series and spatially interpolating the detected missing periods to generate a complete precipitation time series; decomposing the river water level change process into a baseline sequence of water level during normal water periods and an additional sequence of fluctuations during flood and dry periods; matching the baseline sequence of water level during normal water periods with the complete precipitation time series using timestamps, and correcting the measurement time drift in the dynamics of groundwater depth based on the matching results to obtain time-synchronized precipitation-water level linkage pairs; and resampling the time-synchronized precipitation-water level linkage pairs, the additional sequence of fluctuations during flood and dry periods, and the instantaneous flow readings of each water intake node at a unified sampling frequency to generate an aligned standardized data matrix. Preferably, the spatial interpolation of neighboring stations adopts the inverse distance weighting method, which determines the weighting coefficient of each neighboring station for the precipitation estimate of the missing period based on the reciprocal of the distance between the neighboring rainfall station and the missing station, thereby improving the accuracy of precipitation data interpolation.

[0007] As a further technical solution of the present invention, the process of constructing the time-varying water level response function of the target basin based on the standardized data matrix and calculating the theoretical water demand urgency of each water intake node includes: extracting the average water level value of each river segment in multiple consecutive historical periods from the standardized data matrix to form a historical water level state vector; associating and mapping the historical water level state vector with the cumulative precipitation in the corresponding period to establish a precipitation-water level lag response relationship; fitting the predicted water level change curve of each river segment in the predicted period after the current period based on the precipitation-water level lag response relationship; comparing the water intake elevation of each water intake node with the predicted water level change curve to determine the water level satisfaction duration of each water intake node in the predicted period; marking water intake nodes with a water level satisfaction duration less than the preset water intake duration as nodes of concern, and calculating the theoretical water demand urgency based on the water shortage amount and water shortage duration of the nodes of concern. Preferably, when establishing the precipitation-water level lag response relationship, the Pearson correlation coefficient between the cumulative precipitation and the water level change value with different lag days is calculated, and the lag days corresponding to the largest correlation coefficient are determined as the characteristic response lag parameter of the target watershed, so that the constructed response function is more in line with the actual runoff generation and confluence pattern of the watershed.

[0008] As a preferred embodiment of the present invention, the process of generating an initial water resource allocation plan based on the deviation between the theoretical water demand urgency and the instantaneous flow readings of each water intake node includes: sorting the theoretical water demand urgency of each water intake node from largest to smallest to generate a water demand urgency ranking list; selecting the top few water intake nodes from the water demand urgency ranking list as priority allocation targets; obtaining the current available water volume of the river section where the priority allocation target is located, and calculating the cumulative water demand of the priority allocation target in multiple preset allocation cycles; performing a difference calculation between the current available water volume and the cumulative water demand, and if the current available water volume is less than the cumulative water demand, determining the incremental water release from the upstream reservoir based on the difference; and determining the water release time window and water release flow rate of the upstream reservoir based on the incremental water release, thereby generating an initial water resource allocation plan. When determining the incremental water release from the upstream reservoir based on the difference, the difference is divided by the total water release duration within the water release time window, and the water transfer loss coefficient from the upstream reservoir to the river section where the priority allocation target is located is combined to calculate the water release flow rate, so that the allocated water volume is closer to the effective water volume actually reaching the water intake node.

[0009] As another technical solution of the present invention, the process of overlaying the initial water resource allocation scheme with the distribution map of ecologically sensitive areas and performing flow path extrapolation includes: analyzing the distribution map of ecologically sensitive areas, extracting the boundary coordinates of multiple ecologically sensitive areas and the minimum ecological flow threshold of each ecologically sensitive area; using the water release flow and water release time window in the initial water resource allocation scheme as input conditions, performing water flow evolution calculations along the river topology of the target watershed to obtain the predicted flow process line of each river segment; spatially overlaying the predicted flow process line of each river segment with the boundary coordinates of multiple ecologically sensitive areas to identify the affected river segments located within the ecologically sensitive areas; comparing the predicted flow value of the affected river segment with the minimum ecological flow threshold of the corresponding ecologically sensitive area, and if the predicted flow value is lower than the minimum ecological flow threshold, marking the starting node of the corresponding affected river segment as an abnormal scheduling risk node. In calculating the hydrodynamic evolution, the target watershed is divided into multiple river segment units based on its channel topology. The channel slope, roughness coefficient, and cross-sectional geometric parameters of each segment unit are determined. Using the discharge flow rate as the upstream boundary input, the discharge time window is discretized into multiple time steps in chronological order. Within each time step, the one-dimensional Saint-Venant equations are solved based on the inflow flow rate, channel slope, roughness coefficient, and cross-sectional geometric parameters of the current river segment unit to obtain the outflow flow rate and current water level. The outflow flow rate of the current river segment unit is used as the inflow flow rate of the next river segment unit, and the calculation is recursively performed segment by segment along the flow direction to obtain the flow rate value of each river segment unit at each time step. The flow rate values ​​of each river segment unit at all time steps are arranged in chronological order to generate the predicted flow process line for each river segment. Through refined hydrodynamic evolution simulation, the impact of the scheduling scheme on the flow along the river is accurately reflected.

[0010] The process of verifying scheduling permissions for abnormal scheduling risk nodes and outputting a water resources optimization scheduling instruction sequence includes: obtaining the operating status parameters of the sluice gate control station associated with the abnormal scheduling risk node, including the gate opening limit, current opening value, and remote control permission identifier; determining whether the sluice gate control station has remote scheduling permissions based on the remote control permission identifier; if the sluice gate control station has remote scheduling permissions, calculating the target gate opening adjustment amount for the sluice gate control station based on the risk level of the abnormal scheduling risk node; comparing the target gate opening adjustment amount with the gate opening limit; if it does not exceed the gate opening limit, generating an opening adjustment instruction for the sluice gate control station and incorporating the opening adjustment instruction into the water resources optimization scheduling instruction sequence; if it exceeds the gate opening limit, generating an alarm message and associating the alarm message with the location identifier of the abnormal scheduling risk node for timely intervention by maintenance personnel.

[0011] Preferably, the process of calculating the target gate opening adjustment amount based on the risk level of the abnormal scheduling risk node includes: determining the risk level value of the abnormal scheduling risk node based on the absolute value of the difference between the predicted flow value and the minimum ecological flow threshold; querying a preset risk-adjustment mapping table based on the risk level value to obtain the baseline opening adjustment coefficient of the sluice gate control station; obtaining the measured water level difference between the upstream and downstream of the sluice gate control station, and correcting the baseline opening adjustment coefficient based on the water level difference measurement value to obtain the corrected opening adjustment coefficient; multiplying the corrected opening adjustment coefficient by the current gate water passage capacity parameter of the sluice gate control station to obtain the target gate opening adjustment amount. Through the above quantitative calculation, both the urgency of ecological water replenishment and the actual control capacity of the sluice gate can be considered.

[0012] As a further improvement of the present invention, after generating and incorporating the opening adjustment command, the method further includes: sending the opening adjustment command to the execution terminal of the sluice gate control station and receiving the gate action feedback signal returned by the execution terminal; recording the actual gate opening change value of the sluice gate control station according to the gate action feedback signal; comparing the actual gate opening change value with the gate target opening adjustment amount; if the difference value exceeds the preset deviation threshold, generating a verification command; sending the verification command to the backup control channel of the sluice gate control station to trigger the backup control channel to re-execute the opening adjustment operation, thereby forming a closed-loop verification of the scheduling command execution effect and improving scheduling reliability.

[0013] After outputting the alarm message, the process also includes: querying the set of upstream adjustable reservoirs for the abnormal scheduling risk node based on the alarm message; calculating the adjustable storage capacity between the current storage capacity and dead storage capacity of each candidate reservoir in the upstream adjustable reservoir set; selecting the candidate reservoir with the largest adjustable storage capacity as the alternative scheduling reservoir; obtaining the water conveyance capacity of the water conveyance channel between the alternative scheduling reservoir and the abnormal scheduling risk node, and generating an alternative water release scheduling instruction based on the water conveyance capacity, replacing the original opening adjustment instruction with the alternative water release scheduling instruction. By activating the alternative scheduling path, the flow demand of ecologically sensitive areas can still be guaranteed when the original regulation scheme is not feasible.

[0014] This invention also provides a water resource management system based on big data analysis, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the aforementioned water resource management method. The system outputs an optimized scheduling instruction sequence by performing temporal semantic alignment of multi-source hydrological monitoring data, constructing time-varying water level response functions, calculating water demand urgency, generating allocation schemes, and performing risk extrapolation and authorization verification based on the superposition of ecologically sensitive areas. This improves water resource allocation efficiency while effectively preventing ecological scheduling risks.

[0015] The technical effects and advantages provided by the present invention in the above technical solution are as follows: By spatially dividing precipitation observation records into independent time series and performing spatial interpolation of nearby stations to generate complete series for missing periods, and decomposing the river water level change process into a baseline series during the normal water period and an additional fluctuating series during the flood and dry seasons, the time drift in groundwater depth dynamics is corrected by matching the timestamps of the baseline series during the normal water period with the complete precipitation time series. Furthermore, the precipitation-water level linkage pairs, the additional fluctuating series, and the instantaneous flow readings of each water intake node are resampled at a unified sampling frequency to obtain a standardized data matrix. This processing transforms multi-source data with inconsistent sampling intervals and time bases into a time-synchronized and semantically matched dataset, eliminating systematic biases caused by measurement time drift and gaps. This ensures that the precipitation-water level lag relationship in the constructed time-varying water level response function accurately reflects the actual response characteristics of the watershed, thereby significantly improving the accuracy of calculating the theoretical water demand urgency of each water intake node and making the priority allocation targets generated by the allocation scheme more consistent with their actual water shortage status. After generating the initial water resource allocation plan, a distribution map of ecologically sensitive areas is introduced. Using the discharge flow rate and discharge time window as input, the predicted flow process lines for each river segment are calculated along the river topology. These predicted flow process lines are then overlaid with the boundary coordinates of the ecologically sensitive areas to identify affected river segments located within these areas. Their predicted flow values ​​are compared with the minimum ecological flow threshold to mark abnormal scheduling risk nodes. Subsequently, the remote control permission identifiers of the sluice gate control stations associated with these abnormal scheduling risk nodes are obtained. Only when remote scheduling permission is granted is the target gate opening adjustment calculated based on the risk level. After comparing the gate opening limit, an executable opening adjustment command is generated. For cases exceeding limits or insufficient permission, an alarm is output or an alternative reservoir scheduling is activated. This process allows the scheduling plan to anticipate the impact on ecologically sensitive areas before formal issuance, proactively avoiding river segments with insufficient ecological base flow. It also automatically excludes invalid commands caused by missing permissions or gate limitations. The final output sequence of optimized water resource scheduling commands neither damages ecologically sensitive targets nor lacks the conditions for direct automatic execution. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0017] Figure 1 This is a flowchart of a water resource management method based on big data analysis; Figure 2 This is a flowchart of the time-series semantic alignment processing of multi-source hydrological monitoring data; Figure 3 This is a flowchart for calculating the water demand urgency based on the theory of water intake nodes using precipitation-water level lag response; Figure 4 This is a flowchart for generating the initial water resource allocation plan; Figure 5 This is a flowchart of a water resource optimization and scheduling method based on the identification of ecologically sensitive areas; Figure 6 It is a curve comparing the predicted water level changes at the water intake node with the elevation of the water intake. Figure 7 It is a ranking of the theoretical water demand urgency of water intake nodes and the cumulative expected water demand ratio curve; Figure 8 It is a curve comparing the predicted flow of the affected river section in an ecologically sensitive area with the minimum ecological flow threshold; Figure 9 It is a curve showing the changes in the target opening, actual opening, and opening limit of the gate at the sluice gate control station. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] See Figure 1 This invention provides a water resources management method based on big data analysis, comprising: acquiring a multi-source hydrological monitoring data sequence of a target watershed, the multi-source hydrological monitoring data sequence including precipitation observation records, river water level change processes, groundwater depth dynamics, and instantaneous flow readings of each water intake node; performing time-series semantic alignment processing on the multi-source hydrological monitoring data sequence to generate an aligned standardized data matrix; constructing a time-varying water level response function of the target watershed based on the standardized data matrix, and calculating the theoretical water demand urgency of each water intake node using the time-varying water level response function; generating an initial water resources allocation plan for the target watershed based on the deviation between the theoretical water demand urgency and the instantaneous flow readings of each water intake node; acquiring a distribution map of ecologically sensitive areas of the target watershed, overlaying the initial water resources allocation plan with the distribution map of ecologically sensitive areas, and performing flow path extrapolation to determine abnormal scheduling risk nodes; verifying the scheduling authority of the abnormal scheduling risk nodes, and outputting a water resources optimization scheduling instruction sequence for the target watershed.

[0020] Example 1: In specific implementation, please refer to Figure 2When performing temporal semantic alignment processing on the multi-source hydrological monitoring data sequence to generate an aligned standardized data matrix, the precipitation observation records are divided into independent precipitation time series for multiple rain gauge stations according to their spatial distribution. The precipitation observation records originate from multiple rain gauge stations deployed within the target watershed, with each rain gauge station corresponding to a spatial coordinate. Based on the spatial coordinates of each rain gauge station, the cumulative precipitation observation values ​​of each rain gauge station in the precipitation observation records are extracted in chronological order to form an independent precipitation time series corresponding to each rain gauge station. Each data entry in the independent precipitation time series includes a timestamp and a precipitation value.

[0021] When performing missing value detection on each of the independent precipitation time series, the continuity of timestamps in the independent precipitation time series is scanned. If the interval between two adjacent timestamps exceeds a preset sampling period threshold, the time period within the interval is determined to be a missing period, and the start and end times of the missing period are recorded. At the same time, outlier screening is performed on the precipitation values, and abnormal records that exceed the historical extreme value range are removed. The time periods corresponding to the removed points are also marked as missing periods.

[0022] For the detected missing time periods, spatial interpolation using neighboring stations is performed to generate a complete precipitation time series. The spatial interpolation using neighboring stations employs the inverse distance weighting method. In the inverse distance weighting method, the weighting coefficient of each neighboring station for the precipitation estimate of the missing time period is determined based on the reciprocal of the distance between the neighboring rainfall stations and the missing station. The missing station is denoted as... There are around it Nearby rainfall stations Missing site With neighboring rain gauge stations The spatial distance between them is Nearby rain gauge stations Precipitation observations during the missing period were Then the site is missing. Precipitation estimates for the missing period Calculated using the following formula: ; in, For nearby rain gauge stations The weighting coefficients are calculated as follows: , These are the power parameters of the inverse distance weighting method. The value of is set to 2, based on the fact that the spatial correlation of precipitation within the basin decreases squarely with increasing distance, and the inverse distance square weight can reasonably reflect the spatial structure characteristics of the precipitation field. The unit is kilometers. and The unit is millimeters. All missing time periods are iterated over, and for each missing period, the inverse distance weighting method described above is used to calculate the precipitation estimate. The precipitation estimate is then filled into the corresponding position of the missing time period, thus completing the individual precipitation time series into a complete precipitation time series.

[0023] The river level change process is decomposed into a baseline sequence of water levels during the normal water period and an additional sequence of fluctuations during the flood and dry seasons. During decomposition, a time-series moving average method is used to smooth the river level change process, with the moving window length set to 12 months based on the hydrological and meteorological cycle characteristics of the target basin. The smoothed sequence obtained after moving average processing is used as the baseline sequence of water levels during the normal water period, reflecting the long-term trend and basic seasonal changes of the river level. The original river level change process is subtracted from the baseline sequence of water levels during the normal water period at each time point, and the residual portion constitutes the additional sequence of fluctuations during the flood and dry seasons, reflecting the rapid water level fluctuations caused by factors such as short-term heavy rainfall or drought.

[0024] When matching the baseline water level sequence during the normal water period with the complete precipitation time series using timestamps, the water level values ​​corresponding to each timestamp in the complete precipitation time series are extracted from the baseline water level sequence during the normal water period to form initial precipitation-water level pairs. For cases where timestamps cannot be directly aligned, the water level value of the nearest timestamp in the baseline water level sequence during the normal water period is searched for and matched using the precipitation timestamp as a reference.

[0025] Based on the matching results, the measurement time drift in the groundwater depth dynamics is corrected to obtain a time-synchronized precipitation-water level linkage pair. The cross-correlation coefficient sequence between the baseline water level sequence during the normal water period and the complete precipitation time series is calculated. The lag step corresponding to the maximum value of the cross-correlation coefficient is found; this lag step is the time drift of the groundwater depth monitoring equipment relative to the precipitation-water level observation system. The entire groundwater depth dynamic time series is shifted along the time axis according to the time drift, ensuring that the timestamp of the groundwater depth dynamics is consistent with the timestamp of the precipitation-water level linkage pair. The time-corrected groundwater depth dynamics are then merged with the initial precipitation-water level pair to form a time-synchronized precipitation-water level linkage pair.

[0026] The time-synchronized precipitation-water level linkage pairs, the additional flood and dry season fluctuation sequences, and the instantaneous flow readings of each water intake node are resampled at a unified sampling frequency to generate the aligned standardized data matrix. The unified sampling frequency is set to 1 hour. During resampling, a linear interpolation method is used to recalculate the precipitation, water level, and groundwater depth values ​​in the time-synchronized precipitation-water level linkage pairs, the fluctuation components in the additional flood and dry season fluctuation sequences, and the instantaneous flow readings of each water intake node, forming data rows with the same time index. All resampled variables are arranged in chronological order, with each row corresponding to a unified sampling time and each column corresponding to a data variable type, forming the aligned standardized data matrix. The columns of the standardized data matrix include precipitation observations, river water level values, groundwater depth values, water level fluctuation additional values, and instantaneous flow values ​​of each water intake node.

[0027] Example 2: In specific implementation, please refer to Figure 3 Based on the standardized data matrix, a time-varying water level response function for the target watershed is constructed. When calculating the theoretical water demand urgency for each water intake node using this function, the average water level values ​​of each river segment over multiple consecutive historical periods are extracted from the standardized data matrix to form a historical water level state vector. Each column in the standardized data matrix corresponds to a data variable type, and the river water level values ​​are further divided into multiple river segment water level sequences based on the monitoring section location. For each river segment, using a preset historical observation window as the span, the arithmetic mean of the water level values ​​at all sampling times within the historical observation window is calculated to obtain an average water level value. The historical observation window span is set to 24 hours, and the sliding step is set to 1 hour, thus obtaining a historical water level state vector composed of multiple average water level values ​​over multiple consecutive historical periods. Each element in the historical water level state vector corresponds to the average water level value of a historical period, and the order of the elements is consistent with the chronological order of the periods.

[0028] A correlation mapping is established between historical water level state vectors and corresponding cumulative precipitation over time periods to create a precipitation-water level lag response relationship. The cumulative precipitation corresponding to each time period in the historical water level state vector is obtained; cumulative precipitation refers to the sum of precipitation observations in the standardized data matrix within a certain number of days prior to that time period. Multiple candidate lag days are set, ranging from 1 to 30 days, with a step size of 1 day. For each candidate lag day, the cumulative precipitation sequence for each time period under that lag day is calculated. Correlation analysis is performed between the cumulative precipitation sequence and the historical water level state vector. When establishing the precipitation-water level lag response relationship, the Pearson correlation coefficient between cumulative precipitation and water level changes at different lag days is calculated, and the lag day corresponding to the highest correlation coefficient is determined as the characteristic response lag parameter of the target watershed. The formula for calculating the Pearson correlation coefficient is: ; in, Indicates the number of alternative lag days. The value range is an integer from 1 day to 30 days; Indicates a lag of days. The Pearson correlation coefficient between hourly precipitation accumulation and water level change, with the value of the Pearson correlation coefficient ranging from [-1, 1]. Indicates the time period number. The value ranges from 1 to , Indicates the total number of historical periods; Indicates the first Before the time period The cumulative amount of precipitation within a day, expressed in millimeters; Indicates a lag of days. The average cumulative precipitation for all time periods, in millimeters; Indicates the first The water level change value for each time period is the difference between the average water level value of the current time period and the average water level value of the previous time period, and the unit is meters. This represents the average water level change over all time periods, in meters. When the maximum value is obtained, the corresponding The value is determined as the characteristic response time delay parameter of the target watershed.

[0029] After determining the characteristic response time lag parameters, the predicted water level change curves for each river segment in the prediction period following the current time period are fitted based on the precipitation-water level lag response relationship. The cumulative precipitation within the number of days corresponding to the characteristic response time lag parameters before the current time period is used as input. Combined with the water level change patterns corresponding to the same amount of cumulative precipitation in the same season in historical periods, a linear regression method is used to establish a quantitative mapping expression between precipitation input and water level response. The predicted precipitation in the future prediction period is substituted into the precipitation-water level lag response relationship to sequentially calculate the water level values ​​for multiple future prediction periods. These water level values ​​are then connected in chronological order to obtain the predicted water level change curves.

[0030] The elevation of the intake at each water intake node is compared with the predicted water level change curve to determine the duration for which the water level is satisfied within the predicted period. For each water intake node, the elevation of the structural base plate of its intake is recorded as the intake elevation. On the predicted water level change curve, starting from the prediction start time, the water level value at each prediction time is compared with the intake elevation. If the water level value at a certain prediction time is greater than or equal to the intake elevation, that prediction time is determined to be a valid time when the intake can draw water normally. All consecutive valid times are accumulated, and the total number of valid times is counted. The total number of valid times is multiplied by the prediction time step, and the product is the duration for which the water level is satisfied. The prediction time step is consistent with the unified sampling frequency of the standardized data matrix, and is set to 1 hour.

[0031] Water intake nodes whose water level satisfaction duration is less than the preset water intake duration are marked as nodes of concern. The preset water intake duration is determined based on the designed water supply task of the water intake node, which specifies a minimum number of water intake hours per day. This minimum number of water intake hours is used as the preset water intake duration for that water intake node. When the water level satisfaction duration of a water intake node is less than its preset water intake duration, it indicates that the water intake node cannot meet its water demand through gravity water intake during the predicted period, and the water intake node is marked as a node of concern.

[0032] The theoretical water urgency is calculated based on the water shortage amount and duration of the nodes of interest. The water shortage amount is calculated as the difference between the average daily water demand of the node and the actual available water volume within the water level sufficiency period. The average daily water demand is determined based on the water quota and coverage population or area of ​​the node's service recipients. The water shortage duration refers to the number of consecutive hours during the predicted period when the water level sufficiency period is lower than the preset water intake duration. The theoretical water urgency is obtained by normalizing the product of the water shortage amount and the water shortage duration. The baseline value for normalization is the maximum value of the product of water shortage amount and water shortage duration among all water intake nodes in the target watershed. The theoretical water urgency ranges from 0 to 1; a higher value indicates a higher degree of water urgency for the water intake node.

[0033] See Figure 6In the graph, the horizontal axis represents the prediction time in hours, ranging from 0 to 720 hours (30 days), and the vertical axis represents the water level in meters. The solid line represents the predicted water level change curve, and the dashed line represents the intake elevation, which is constant at 11 meters. The predicted water level curve shows a clear characteristic of phased changes. During the initial stage of the forecast (0-150 hours), the water level gradually rose from about 12.7 meters to a peak of about 13.7 meters, which is above the elevation of the water intake. This indicates that the water level at the water intake is sufficient during this stage to meet normal water intake needs. Subsequently, from 150 hours to 450 hours, the water level gradually decreased, and after about 300 hours, it fell below the intake elevation by 11 meters. During this period, the water level dropped from 13.7 meters to a minimum of about 9.85 meters, which was lower than the intake elevation, indicating that the water level at the intake was insufficient during this period and that water intake difficulties might occur. Between 450 and 720 hours, the water level showed an upward trend, gradually rising from the lowest point to about 12.7 meters, exceeding the elevation of the water intake, indicating that the water level at the water intake had returned to a usable state.

[0034] The gray area in the figure marks the region of difference between the predicted water level curve and the intake elevation, reflecting the duration and magnitude of the water level satisfaction. The intersection of this curve and the intake elevation is the critical time point between water level satisfaction and non-satisfaction. The trend of this curve's change can be used to determine the available water level period and its sustainability at the intake within the predicted timeframe.

[0035] Combining the method of comparing the predicted water level change curve with the water intake elevation in Example 2, this figure accurately reflects the duration of water level sufficiency and the changes in water level at the water intake node during the predicted future period, providing a basis for calculating the duration of water level sufficiency and identifying nodes of concern. Stages where the predicted water level is below the water intake elevation for an extended period indicate a potential risk of water shortage at the water intake node. The theoretical water urgency needs to be calculated by incorporating the amount and duration of water shortage, thereby guiding the formulation of subsequent water resource allocation plans.

[0036] Example 3: In specific implementation, please refer to Figure 4 Based on the deviation between the theoretical water demand urgency and the instantaneous flow readings of each water intake node, when generating the initial water resource allocation plan for the target watershed, the theoretical water demand urgency of each water intake node is sorted from largest to smallest, generating a water demand urgency ranking list. The theoretical water demand urgency is a normalized value calculated by combining the amount of water shortage and the duration of water shortage, with a value range between 0 and 1. The theoretical water demand urgency of all water intake nodes is extracted, and a descending numerical sorting algorithm is used to obtain a list arranged from the largest theoretical water demand urgency to the smallest theoretical water demand urgency. This list is the water demand urgency ranking list, and each item in the list contains a unique identifier for the water intake node and its corresponding theoretical water demand urgency value.

[0037] The top-ranked water intake nodes in the water urgency ranking list are selected as priority allocation targets. The specific method for selecting priority allocation targets is as follows: starting from the first water intake node in the water urgency ranking list, the expected water demand of each water intake node within a preset allocation period is accumulated sequentially. The preset allocation period is set to 24 hours, divided into 24 equal-length time periods, each time period being 1 hour long. The expected water demand is obtained by multiplying the historical average water intake flow rate of the water user corresponding to the water intake node during the same period by the time period length. When the accumulated value reaches a preset percentage threshold of the current available water volume, the selection stops, and all accumulated water intake nodes are identified as priority allocation targets. The preset percentage threshold is set to 0.8, meaning that the selection stops when the accumulated expected water demand reaches 80% of the current available water volume. If the preset percentage threshold is not reached after accumulating all water intake nodes, then all water intake nodes are identified as priority allocation targets.

[0038] Obtain the current available water volume of the river section where the priority allocation target is located. The current available water volume refers to the total amount of water resources that the river section where the priority allocation target is located can be used for supply between the current time and the end time of the preset allocation cycle. The specific method is as follows: the current real-time flow value is collected by the flow monitoring equipment set at the upstream control section of the river section. At the same time, the hourly flow forecast value for the next 24 hours is output by the upstream inflow forecast model. The current real-time flow value and the hourly flow forecast value for the next 24 hours are multiplied by the corresponding time step and then summed to obtain the current available water volume, in cubic meters.

[0039] Calculate the cumulative water demand of priority allocation targets over multiple preset allocation cycles. The number of preset allocation cycles is one, meaning the calculation is performed for the next 24 hours. For each priority allocation target, the expected water demand for each hour within the next 24 hours is summed to obtain the allocation cycle water demand for that single priority allocation target. The allocation cycle water demands of all priority allocation targets are then summed to obtain the cumulative water demand, in cubic meters.

[0040] The difference between the current available water volume and the cumulative water demand is calculated, comparing their magnitudes. If the current available water volume is greater than or equal to the cumulative water demand, it indicates that the river section's own water volume can meet the water needs of the priority allocation targets, and no additional water release from upstream reservoirs is required. In this case, the initial water allocation plan sets the water release increment to zero, the water release flow rate to zero, and the water release time window to an empty set. If the current available water volume is less than the cumulative water demand, the difference is calculated. The difference equals the cumulative water demand minus the current available water volume, and this difference represents the additional water release from upstream reservoirs that needs to be supplemented, expressed in cubic meters.

[0041] When determining the incremental water release from the upstream reservoir based on the difference, the difference is divided by the total release time within the release time window, and the release flow is calculated by combining this with the water transport loss coefficient from the upstream reservoir to the river section where the priority allocation target is located. The release time window is determined as follows: the water propagation time from the upstream reservoir to the river section where the priority allocation target is located is obtained. This propagation time is obtained by dividing the river length by the historical average flow velocity, which is calculated based on the river cross-sectional parameters and the multi-year average flow. The start time of the release time window is obtained by subtracting the water propagation time from the start time of the preset allocation cycle, and the end time of the release time window is obtained by subtracting the water propagation time from the end time of the preset allocation cycle. The duration between the start and end times is the total release time, expressed in hours.

[0042] The water conveyance loss coefficient is determined based on the water conveyance loss characteristics of the river channel between the upstream reservoir and the river section where the priority allocation target is located. Specifically, it is obtained by statistically analyzing the ratio of the actual total amount of water released by the upstream reservoir to the actual amount of water received at the inlet section of the river section where the priority allocation target is located during multiple water releases in the past year, calculating the average value of each ratio, and subtracting the average value from 1 to obtain the water conveyance loss coefficient. The water conveyance loss coefficient represents the proportion of water loss caused by river seepage, evaporation, etc., and its value ranges from 0.05 to 0.30.

[0043] The formula for calculating the water discharge flow rate is as follows: ; in, This indicates the water flow rate, expressed in cubic meters per second. This represents the difference, which is the incremental water release obtained by subtracting the total current available water from the cumulative water demand, and is expressed in cubic meters. This indicates the total water discharge duration within the water discharge time window, in seconds. This represents the water conveyance loss coefficient, which is dimensionless and ranges from 0.05 to 0.30. It is calculated by dividing the difference by the total water release time and the effective water conveyance rate. The product of these two values ​​yields the stable discharge flow that the upstream reservoir needs to maintain after considering water transfer losses.

[0044] After determining the release time window and release flow rate of the upstream reservoir based on the incremental water release, an initial water resource allocation plan is generated. This initial plan records the upstream reservoir's identification information, the start and end times of the release time window, and the change in release flow rate over time. If the release flow rate remains constant within the release time window, the release flow rate will appear as a horizontal straight line. If further adjustments are needed based on the time-specific needs of priority allocation targets, the release time window will be divided into multiple sub-periods, and the corresponding release flow rate will be calculated for each sub-period. The initial water resource allocation plan is generated in the form of a structured data table, containing the reservoir number, release start time, release end time, release flow rate values ​​at each time point, and a corresponding list of priority allocation targets.

[0045] See Figure 7 In the graph, the horizontal axis represents the water intake node number, arranged in descending order according to the theoretical water demand urgency of each node; the left vertical axis corresponds to the theoretical water demand urgency, ranging from 0 to 0.8, represented by a gray bar chart; the right vertical axis corresponds to the cumulative expected water demand percentage, ranging from 0% to 100%, represented by a black dashed line. The graph also includes a reference dashed line (dotted line) indicating a preset threshold of 80%.

[0046] As can be seen from the gray bar chart, the theoretical water urgency of the top-ranked water intake nodes is relatively high, with a maximum value close to 0.78, indicating that these nodes are experiencing severe water shortages and have a high degree of water urgency. As the node number increases, the theoretical water urgency shows a clear decreasing trend, gradually approaching zero, indicating that the risk of water shortage for subsequent nodes is relatively low.

[0047] The black dashed line reflects the trend of the cumulative expected water demand percentage as the water intake node number changes. Observing the curve trend, the cumulative expected water demand percentage of the first approximately 85 water intake nodes rapidly rises to 80%, corresponding to the dotted line position in the graph. At this point, the cumulative expected water demand of the water intake nodes in the theoretical water urgency ranking list reaches the preset threshold of 80% of the current available water volume, meeting the requirement for selecting priority allocation targets in Example 3. Subsequently, the cumulative expected water demand percentage of the remaining approximately 35 water intake nodes slowly approaches 100%.

[0048] Example 4: In specific implementation, please refer to Figure 5The process involves acquiring a distribution map of ecologically sensitive areas in the target watershed, overlaying the initial water resource allocation plan onto this map, and performing flow path simulations to identify abnormal scheduling risk nodes. Then, the ecologically sensitive area distribution map is analyzed to extract the boundary coordinates of multiple ecologically sensitive areas and the minimum ecological flow threshold for each area. The ecologically sensitive area distribution map is stored as a vector geographic information file, containing multiple polygonal patches, each corresponding to an ecologically sensitive area. The attribute table records the name, protection level, and minimum ecological flow threshold of each ecologically sensitive area. The analysis process reads the geometric and attribute data from the vector geographic information file, extracting the latitude and longitude coordinates of the vertices of each polygonal patch to form the boundary coordinate set of the corresponding ecologically sensitive area. Simultaneously, the minimum ecological flow threshold corresponding to that ecologically sensitive area is read. This minimum ecological flow threshold, measured in cubic meters per second, is the minimum river flow required to maintain the basic functions of the ecosystem, determined by the watershed ecological protection authority based on the requirements of water ecological function zoning.

[0049] Using the initial water allocation scheme's discharge flow rate and discharge time window as input conditions, flow evolution calculations are performed along the river topology of the target basin to obtain the predicted flow process lines for each river segment. The river topology of the target basin is represented by a directed acyclic graph (DAG), where nodes represent river confluences, intakes, reservoir discharge outlets, or control sections, and edges represent river segments connecting nodes. Each edge records the river length, riverbed slope, cross-sectional shape parameters, and roughness coefficient. Using the discharge flow rate as the upstream boundary condition, unsteady flow evolution simulations are performed starting from the beginning of the discharge time window, with a uniform calculation step size of 15 minutes. During the unsteady flow evolution simulation, the Muskingan flow evolution algorithm is used to calculate the flow propagation process segment by segment. For each river segment, the downstream outflow process is calculated based on the upstream inflow process, river length, and wave velocity coefficient, while also considering the lateral inflow along the river segment and the intake flow rate. In the initial water resource allocation scheme, the discharge flow of each upstream reservoir is input into the corresponding node in the river topology according to the time series. After the water flow evolution calculation, the flow prediction value of each river segment in the target watershed at each calculation step in the prediction period is obtained. These flow prediction values ​​are arranged in time order to form the predicted flow process line of each river segment.

[0050] By spatially overlaying the predicted flow process lines of each river segment with the boundary coordinates of multiple ecologically sensitive areas, affected river segments located within these areas are identified. Spatial overlay is achieved by superimposing river centerline vector data with the boundary coordinate set of ecologically sensitive areas. The river centerline vector data is generated based on the geographic spatial location of each river segment within the river's topology. For each river centerline, it is determined whether it has a spatial intersection or inclusion relationship with any polygonal patch of an ecologically sensitive area. If all or part of a segment of a river centerline falls within the boundary coordinate range of an ecologically sensitive area, the river segment corresponding to this segment is identified as an affected river segment, and the association between this affected river segment and the corresponding ecologically sensitive area is recorded.

[0051] The predicted flow value of the affected river segment is compared with the minimum ecological flow threshold of the corresponding ecologically sensitive area. If the predicted flow value of the affected river segment at any calculation step within the prediction period is lower than the minimum ecological flow threshold of the corresponding ecologically sensitive area, the starting node of the affected river segment is marked as an abnormal scheduling risk node. During the comparison, the flow value at each time point in the predicted flow process line of the affected river segment is read one by one and compared with the minimum ecological flow threshold. The marking information of the abnormal scheduling risk node includes the node number, the river segment to which it belongs, the name of the corresponding ecologically sensitive area, the time period when the flow is lower than the minimum ecological flow threshold, and the minimum flow deficit value.

[0052] When verifying scheduling permissions for nodes with abnormal scheduling risks and outputting the water resource optimization scheduling instruction sequence for the target watershed, the operational status parameters of the sluice gate control stations associated with the abnormal scheduling risk nodes are obtained. These parameters include the gate opening limit, the current opening value, and the remote control permission identifier. The sluice gate control station associated with an abnormal scheduling risk node refers to the closest sluice gate control station with flow regulation capabilities located upstream of the abnormal scheduling risk node in the river topology. The association is obtained through a pre-established node-control station mapping table. The operational status parameters of the sluice gate control stations are acquired in real-time through a data acquisition and monitoring system. The gate opening limit refers to the maximum opening height of the sluice gate control station's gate within the structural safety allowable range, in meters. The current opening value refers to the current actual opening height of the sluice gate control station's gate, in meters. The remote control permission identifier is a Boolean state variable, with a value of "allowed" or "prohibited," indicating whether the sluice gate control station currently accepts commands from the remote scheduling system.

[0053] Based on the remote control permission identifier, determine whether the sluice gate control station has remote scheduling authority. If the remote control permission identifier is set to "Allow," the sluice gate control station is deemed to have remote scheduling authority. If the remote control permission identifier is set to "Deny," the sluice gate control station is deemed not to have remote scheduling authority. In this case, skip the opening adjustment step for the sluice gate control station, generate a manual intervention prompt, and associate the manual intervention prompt with the location identifier of the abnormal scheduling risk node.

[0054] If the sluice gate control station has remote dispatching authority, the target gate opening adjustment amount is calculated based on the risk level of the abnormal dispatching risk node. When calculating the target gate opening adjustment amount based on the risk level of the abnormal dispatching risk node, the risk level of the abnormal dispatching risk node is determined by the absolute value of the difference between the predicted flow value and the minimum ecological flow threshold. The absolute value of the difference is the maximum difference between the minimum ecological flow threshold and the predicted flow value in the affected river section during the period when the minimum ecological flow threshold is violated, expressed in cubic meters per second. The absolute value of the difference is compared with multiple preset threshold intervals, each interval corresponding to a risk level value. The threshold intervals are divided as follows: risk level 1 when the absolute value of the difference is less than 0.5 cubic meters per second; risk level 2 when the absolute value of the difference is between 0.5 and 2.0 cubic meters per second; and risk level 3 when the absolute value of the difference is greater than 2.0 cubic meters per second.

[0055] Based on the risk level value, a pre-set risk-adjustment mapping table is consulted to obtain the baseline opening adjustment coefficient for the sluice gate control station. The risk-adjustment mapping table is pre-stored in the dispatch system database, recording a one-to-one correspondence between risk level values ​​and baseline opening adjustment coefficients. The baseline opening adjustment coefficient is 0.05 for a risk level value of 1; 0.15 for a risk level value of 2; and 0.30 for a risk level value of 3. The baseline opening adjustment coefficient represents the gate opening adjustment ratio based on the fully open gate state and is dimensionless.

[0056] The water level difference between the upstream and downstream sides of the sluice gate control station is measured, and the baseline opening adjustment coefficient is corrected based on this measurement to obtain the corrected opening adjustment coefficient. The water level difference between the upstream and downstream sides of the sluice gate control station is collected in real time by water level gauges installed at upstream and downstream measuring points. The water level difference is calculated by subtracting the downstream water level from the upstream water level, in meters. The correction method is as follows: the water level difference measurement is compared with the design water level difference reference value, which is the upstream and downstream water level difference corresponding to the design operating conditions of the sluice gate control station, and is taken as 2.0 meters. The corrected opening adjustment coefficient is equal to the baseline opening adjustment coefficient multiplied by a correction factor. The correction factor is obtained by dividing the water level difference measurement by the design water level difference reference value. When this ratio is greater than 1.2, the correction factor is 1.2; when this ratio is less than 0.8, the correction factor is 0.8; otherwise, the correction factor is equal to this ratio.

[0057] The target gate opening adjustment is obtained by multiplying the corrected opening adjustment coefficient by the current gate flow capacity parameter of the sluice gate control station. The current gate flow capacity parameter represents the change in flow rate corresponding to a unit change in gate opening at the current opening state, expressed in cubic meters per second per meter. This parameter is calculated using linear interpolation based on the gate hydraulic characteristic curve and the current opening value. The target gate opening adjustment is equal to the product of the corrected opening adjustment coefficient and the current gate flow capacity parameter, expressed in meters.

[0058] The target gate opening adjustment amount is compared with the gate opening limit. If the target gate opening adjustment amount does not exceed the gate opening limit, an opening adjustment command is generated for the sluice gate control station and incorporated into the water resources optimization scheduling command sequence. The opening adjustment command includes the sluice gate control station number, the target opening value, and the adjustment execution time. The target opening value equals the current opening value plus the target gate opening adjustment amount. If the target opening value exceeds the gate opening limit, the target opening value is limited to within the gate opening limit range. If the target gate opening adjustment amount exceeds the gate opening limit, an alarm message is generated. The alarm message includes the location identifier of the abnormal scheduling risk node, the sluice gate control station number, the specific value exceeding the limit, and a description of the suggested manual handling measures. The alarm message is then associated with the location identifier of the abnormal scheduling risk node and output to the scheduling duty terminal and mobile alarm equipment.

[0059] See Figure 8In the graph, the horizontal axis represents time in minutes, and the vertical axis represents flow rate in cubic meters per second (m³ / s). The solid curve represents the predicted flow rate process line, and the dashed line represents the minimum ecological flow threshold corresponding to ecologically sensitive areas. The predicted flow rate process line exhibits a periodic fluctuation trend, with a maximum peak value of approximately 10.5 m³ / s and a gradually decreasing minimum value, reaching a low of approximately 3.5 m³ / s, showing that the flow rate generally exhibits a gradual decline from high to low within the predicted period.

[0060] The dashed minimum ecological flow threshold remains stable at approximately 5.1 m³ / s, serving as the critical flow standard for maintaining ecological functions. The solid curve crosses this minimum ecological flow threshold at multiple time periods, forming the shaded areas in the figure, representing abnormal scheduling risk intervals where the river section's flow falls below the ecological flow threshold. These risk intervals are mainly distributed at approximately 3600-4200 minutes, 5400-5800 minutes, 6500-7000 minutes, 7800-8400 minutes, and beyond 9500 minutes, reflecting multiple risk events of insufficient ecological flow within the predicted time period.

[0061] The overall flow curve shows a decreasing amplitude over time, and the periods with flow rates below the ecological threshold gradually increase, reflecting the impact and inadequacy of the initial water resource allocation plan's release flow rate and time window on flow maintenance. This figure, through the superposition and comparison of the flow process line and the ecological flow threshold, visually reveals the potential for insufficient flow in ecologically sensitive areas of the river section after the implementation of the water resource allocation plan, providing data support for the identification of subsequent abnormal scheduling risk nodes and the verification of scheduling authority.

[0062] Example 5: In practice, after generating an opening adjustment command for the sluice gate control station and incorporating it into the water resource optimization scheduling command sequence, the opening adjustment command is sent to the execution terminal of the sluice gate control station. The execution terminal of the sluice gate control station is a local control unit installed next to the gate hoist. This local control unit communicates with the scheduling center via industrial Ethernet and receives the opening adjustment command message, which includes the target opening value and adjustment time limit. After parsing the opening adjustment command, the execution terminal drives the hydraulic hoist or winch hoist to perform the gate opening adjustment action. After the action is completed, the execution terminal collects the actual gate opening value obtained by the absolute encoder installed on the gate, and combines it with the action completion timestamp and action execution status code to form a gate action feedback signal, which is returned to the data receiving module of the scheduling center via industrial Ethernet. The dispatch center records the actual gate opening change value of the sluice gate control station based on the gate action feedback signal. This actual gate opening change value is the difference between the actual gate opening value in the gate action feedback signal and the current opening value recorded from the operating status parameters before the opening adjustment command is sent. The absolute value of the actual gate opening change value is compared with the absolute value of the target gate opening adjustment amount. The difference is the absolute value of the actual gate opening change value minus the absolute value of the target gate opening adjustment amount. If the difference exceeds a preset deviation threshold, a verification command is generated. The preset deviation threshold is set to the larger of 5% of the absolute value of the target gate opening adjustment amount and 0.02 meters. The verification command is sent to the backup control channel of the sluice gate control station. The backup control channel is a separate communication link physically isolated from the main control channel and connected to the redundant programmable logic controller of the sluice gate control station. After receiving the verification instruction, the backup control channel drives the redundant programmable logic controller to take over control and re-execute the gate opening adjustment operation until the difference between the actual gate opening change value and the gate target opening adjustment amount is less than or equal to the preset deviation threshold.

[0063] After associating the alarm message with the location identifier of the abnormal scheduling risk node, an upstream tracing query is performed in the pre-constructed watershed topology database based on the location identifier of the abnormal scheduling risk node carried in the alarm message. This retrieves all reservoirs located upstream of the abnormal scheduling risk node that have regulation capabilities, forming a set of upstream adjustable reservoirs. The watershed topology database stores the river network structure of the target watershed using a node-edge model. Nodes represent reservoirs, intakes, and confluences, while edges represent river channels, with the flow direction recorded in the edge attributes. During the query, starting from the abnormal scheduling risk node, all incoming edges are traversed along the reverse flow direction, collecting nodes of type reservoir along the path. For each candidate reservoir in the upstream adjustable reservoir set, its current water storage and dead storage capacity data are obtained. The current water storage is calculated from the real-time reservoir water level-capacity relationship curve of the reservoir management information system, and the dead storage capacity is the storage capacity corresponding to the dead water level specified in the reservoir design documents. Calculate the adjustable storage capacity of each candidate reservoir in the upstream adjustable reservoir set. The adjustable storage capacity is equal to the current storage capacity minus the dead storage capacity. Select the maximum value from the adjustable storage capacity values ​​of each candidate reservoir, and determine the candidate reservoir with the largest adjustable storage capacity as the alternative scheduling reservoir. Obtain the water conveyance capacity of the water conveyance channel between the alternative scheduling reservoir and the abnormal scheduling risk node. The water conveyance capacity of the water conveyance channel is the maximum safe flow rate of the water conveyance channel from the outlet of the alternative scheduling reservoir to the inlet of the river section where the abnormal scheduling risk node is located. This maximum safe flow rate is provided by the water conveyance channel design data and is in cubic meters per second. Based on the water conveyance capacity of the water conveyance channel, generate an alternative water release scheduling instruction. The water release flow rate in the alternative water release scheduling instruction is set to the smaller value between the water conveyance capacity of the water conveyance channel and the required supplementary flow rate of the abnormal scheduling risk node. The required supplementary flow rate is equal to the absolute value of the difference between the predicted flow rate of the river section where the abnormal scheduling risk node is located and the minimum ecological flow threshold. The start time of the alternative water release scheduling instruction is determined by reverse calculation based on the water flow propagation time from the alternative scheduling reservoir to the abnormal scheduling risk node, and the water release duration is consistent with the effective period of the original opening adjustment instruction. The alternative water release scheduling instruction replaces the opening adjustment instruction, that is, the alternative water release scheduling instruction is written into the water resource optimization scheduling instruction sequence, while the opening adjustment instruction is removed from the sequence.

[0064] When using the initial water allocation plan's discharge flow rate and time window as input conditions, and performing flow evolution calculations along the river topology of the target basin to obtain the predicted flow process lines for each river segment, the target basin is divided into multiple river segment units based on its river topology. Each river segment unit is based on an edge in the river topology; edges exceeding a preset maximum calculation length are further subdivided into several sub-segments, ensuring that the length of each river segment unit does not exceed two kilometers. The riverbed slope of each river segment unit is determined; this slope is the ratio of the difference in elevation between the upstream and downstream riverbeds to the unit length. The riverbed elevation is extracted from the cross-sectional measurement data of the river channel. The roughness coefficient of each river segment unit is determined, calibrated according to the riverbed composition and vegetation conditions, referring to the Manning roughness coefficient table in the hydraulics handbook. The natural river roughness coefficient ranges from 0.025 to 0.045. The cross-sectional geometric parameters of each river segment are determined, including cross-sectional shape identifiers, bottom width, slope coefficients, and left and right bank heights. These parameters are extracted from the large-section river channel measurement results. Using the discharge flow rate as the upstream boundary input, the discharge time window is discretized into multiple time steps in chronological order. The determination of the time steps must satisfy the numerical stability condition, i.e., the Coulomb number must be less than 1.0. The time steps are then selected. Sixty seconds, space step Let be the length of each river segment unit. Within each time step, based on the inflow rate, channel slope, roughness coefficient, and cross-sectional geometric parameters of the current river segment unit, the one-dimensional Saint-Venant equations are solved to obtain the outflow rate and current water level of the current river segment unit. The one-dimensional Saint-Venant equations consist of continuity equations and dynamic equations, and are discretized and solved simultaneously using the Pressman four-point eccentric implicit difference scheme. One of the discrete equations is as follows: ; in, Indicates the first Each cross-section in the time layer The flow rate is expressed in cubic meters per second. Indicates the first Each cross-section in the time layer The flow rate is expressed in cubic meters per second. Indicates the first Each cross-section in the time layer The water level is expressed in meters. Indicates the first Each cross-section in the time layer The water level is expressed in meters. The spatial step size is the length of the current river segment unit, expressed in meters. The time step is set to sixty seconds, based on satisfying the Courrant condition and ensuring computational stability. Indicates the first Each cross-section in the time layer The width of the water surface at the cross-section is measured in meters. This width of the water surface at the cross-section is determined by the first... The cross-sectional geometric parameters of each section and The water level values ​​of the time layers are calculated using rectangular or trapezoidal formulas. The Pressman four-point eccentric implicit difference scheme establishes a nonlinear system of equations for the continuity and dynamic equations, and the Newton-Raphson method is used iteratively to solve for the flow rates at each cross-section within each time step. and water level During the calculation, the upstream boundary condition of the first river segment unit is the time-series value of the discharge flow, while the downstream boundary condition of the last river segment unit adopts the water level-discharge relationship curve or the assumption of steady uniform flow. After the solution is completed, the outflow flow and current water level of the downstream section of the current river segment unit are obtained. The outflow flow of the downstream section of the current river segment unit is used as the inflow flow of the next river segment unit, and the calculation is recursively performed segment by segment along the flow direction to obtain the flow value of each river segment unit at each time step. The flow values ​​of each river segment unit at all time steps are arranged in chronological order to generate the predicted flow process line of each river segment.

[0065] See Figure 9 The horizontal axis of the graph represents time in minutes, covering a time range of approximately 0 to 1450 minutes; the vertical axis represents gate opening in meters, ranging from 0.8 meters to 1.5 meters. The legend includes three curves: "Target gate opening (m)" is represented by a solid line, "Actual gate opening (m)" is represented by a dashed line, and "Gate opening limit (m)" is represented by a dashed line, with the limit fixed at 1.5 meters.

[0066] The solid line "target gate opening" in the figure shows a stepped change, which can be divided into five stages: the initial stage (0 to about 200 minutes) in which the target gate opening is stable at 1.0 meter, then suddenly increases to about 1.3 meters and remains there for about 500 minutes; the middle stage (about 500 to 700 minutes) in which the target opening rapidly decreases to 1.0 meter and remains there, and then maintains a low opening state in the range of 700 to 1000 minutes; the final stage (about 1000 to 1300 minutes) in which the target opening rises back to about 1.0 meter; and the final stage (about 1300 to 1450 minutes) in which it slightly increases to about 1.1 meters and then falls back to 1.0 meter.

[0067] The dotted line representing the "actual gate opening" closely tracks the target opening curve, exhibiting a certain degree of delay and smooth transition. The actual opening shows smooth fluctuations at points where the target opening suddenly increases or decreases, with the maximum deviation not exceeding approximately 0.02 meters. The actual opening consistently does not exceed the 1.5-meter dotted line representing the gate opening limit, indicating that the operating command did not exceed the safe opening limit.

[0068] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A water resource management method based on big data analysis, characterized in that, The method includes: Acquire multi-source hydrological monitoring data sequences for the target watershed, which include precipitation observation records, river water level change processes, groundwater depth dynamics, and instantaneous flow readings at each water intake node; Perform temporal semantic alignment processing on the multi-source hydrological monitoring data sequence to generate an aligned and standardized data matrix; The time-varying water level response function of the target watershed is constructed based on the standardized data matrix, and the theoretical water demand urgency of each water intake node is calculated using the time-varying water level response function. Based on the deviation between the theoretical water demand urgency and the instantaneous flow readings at each water intake node, an initial water resource allocation plan for the target watershed is generated. Obtain the distribution map of ecologically sensitive areas in the target watershed, overlay the initial water resource allocation plan with the distribution map of ecologically sensitive areas, and then perform flow path extrapolation to determine abnormal scheduling risk nodes; The scheduling authority of the abnormal scheduling risk node is verified, and the water resource optimization scheduling instruction sequence of the target watershed is output.

2. The water resource management method based on big data analysis according to claim 1, characterized in that, Perform temporal semantic alignment processing on the multi-source hydrological monitoring data sequence to generate an aligned and standardized data matrix, including: The precipitation observation records are divided into independent precipitation time series of multiple rain gauge stations according to their spatial distribution; Missing values ​​are detected for each of the independent precipitation time series, and spatial interpolation of neighboring stations is performed on the detected missing periods to generate a complete precipitation time series; The river water level change process is decomposed into a baseline sequence of water levels during normal water periods and an additional sequence of fluctuations during flood and dry periods. The baseline water level sequence during the normal water period is timestamped with the complete precipitation time series, and the measurement time drift in the dynamics of groundwater depth is corrected according to the matching results to obtain a time-synchronized precipitation-water level linkage pair. The time-synchronized precipitation-water level linkage pair, the additional sequence of flood and dry season fluctuations, and the instantaneous flow readings of each water intake node are resampled at a unified sampling frequency to generate the aligned standardized data matrix.

3. The water resource management method based on big data analysis according to claim 2, characterized in that, The spatial interpolation of neighboring stations adopts the inverse distance weighting method, which determines the weighting coefficient of each neighboring station for the precipitation estimate of the missing period based on the reciprocal of the distance between the neighboring rainfall station and the missing station.

4. The water resource management method based on big data analysis according to claim 2, characterized in that, Based on the standardized data matrix, a time-varying water level response function for the target watershed is constructed, and the theoretical water demand urgency of each water intake node is calculated using the time-varying water level response function, including: The average water level values ​​of each river section over multiple consecutive historical periods are extracted from the standardized data matrix to form a historical water level state vector. The historical water level state vector is correlated and mapped with the cumulative precipitation in the corresponding period to establish a precipitation-water level lag response relationship; Based on the precipitation-water level lag response relationship, the predicted water level change curves for each river segment within the predicted time period following the current time period are fitted. The elevation of the water intake at each water intake node is compared with the predicted water level change curve to determine the duration for which the water level at each water intake node meets the predicted time period. Water intake nodes whose water level meets the requirement for a duration less than the preset water intake duration are marked as nodes of concern, and the theoretical water urgency is calculated based on the water shortage amount and duration of the nodes of concern.

5. A water resource management method based on big data analysis according to claim 4, characterized in that, When establishing the precipitation-water level lag response relationship, the Pearson correlation coefficient between the cumulative precipitation and the water level change value with different lag days is calculated, and the lag day corresponding to the largest correlation coefficient in the Pearson correlation coefficient is determined as the characteristic response lag parameter of the target watershed.

6. A water resource management method based on big data analysis according to claim 3, characterized in that, Based on the deviation between the theoretical water demand urgency and the instantaneous flow readings at each water intake node, an initial water resource allocation plan for the target watershed is generated, including: The theoretical water urgency of each water intake node is sorted from largest to smallest to generate a water urgency sorting list; Select the top few water intake nodes from the water urgency ranking list as priority allocation targets; Obtain the current available water volume of the river section where the priority allocation target is located, and calculate the cumulative water demand of the priority allocation target in multiple preset allocation cycles; The difference between the current available water volume and the cumulative water demand is calculated. If the current available water volume is less than the cumulative water demand, the incremental water release from the upstream reservoir is determined based on the difference. Based on the incremental water release, the release time window and release flow rate of the upstream reservoir are determined, and the initial water resource allocation plan is generated.

7. A water resource management method based on big data analysis according to claim 6, characterized in that, When determining the incremental water release from the upstream reservoir based on the difference, the difference is divided by the total water release duration within the water release time window, and the water release flow rate is calculated by combining the water transport loss coefficient from the upstream reservoir to the river section where the priority allocation target is located.

8. A water resource management method based on big data analysis according to claim 6, characterized in that, Obtain a distribution map of ecologically sensitive areas in the target watershed. Overlay the initial water resource allocation plan with the ecologically sensitive area distribution map to perform flow path extrapolation, thereby identifying abnormal scheduling risk nodes, including: The distribution map of the ecologically sensitive areas is analyzed to extract the boundary coordinates of multiple ecologically sensitive areas and the minimum ecological flow threshold for each ecologically sensitive area. Using the initial water allocation scheme and the water release flow rate and time window as input conditions, the water flow evolution calculation is performed along the river topology of the target watershed to obtain the predicted flow process line of each river segment. By spatially overlaying the predicted flow process lines of each river segment with the boundary coordinates of the multiple ecologically sensitive areas, the affected river segments located within the ecologically sensitive areas are identified. The predicted flow value of the affected river segment is compared with the minimum ecological flow threshold of the corresponding ecologically sensitive area. If the predicted flow value is lower than the minimum ecological flow threshold, the starting node of the corresponding affected river segment is marked as the abnormal scheduling risk node.

9. A water resource management method based on big data analysis according to claim 8, characterized in that, The scheduling authority of the abnormal scheduling risk node is verified, and the water resource optimization scheduling instruction sequence of the target watershed is output, including: Obtain the operating status parameters of the sluice gate control station associated with the abnormal scheduling risk node. The operating status parameters include the gate opening limit, the current opening value, and the remote control permission identifier. Determine whether the sluice gate control station has remote dispatch authority based on the remote control permission identifier; If the sluice gate control station has the remote scheduling authority, then the target gate opening adjustment amount of the sluice gate control station is calculated according to the risk level of the abnormal scheduling risk node. The target gate opening adjustment amount is compared with the gate opening limit. If the target gate opening adjustment amount does not exceed the gate opening limit, an opening adjustment command is generated for the sluice gate control station, and the opening adjustment command is incorporated into the water resource optimization scheduling command sequence. If the gate target opening adjustment exceeds the gate opening limit, an alarm message is generated and the alarm message is associated with the location identifier of the abnormal scheduling risk node and output.

10. A water resource management system based on big data analysis, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the water resource management method based on big data analysis as described in any one of claims 1 to 9.

Images