An all-region intelligent management and control platform for irrigation districts based on digital twins

Abstract

The present application relates to the technical field of data processing, in particular to a kind of based on digital twin's irrigation area global wisdom management and control platform, system includes: disturbance structure construction module constructs three-dimensional disturbance characteristic matrix, rhythm division module executes amplitude contrast and generates irrigation area time series disturbance partition map, evapotranspiration layer generation module constructs evapotranspiration difference heat map layer, fluctuation mapping module generates flow evaporation correlation graph, and regulation and control graph generation module constructs wisdom management and control graph.The present application, by collecting multidimensional data to construct disturbance matrix, accurately identify fluctuation rhythm characteristics to divide high and low disturbance area, utilize plot evapotranspiration record to invert high dynamic area heat distribution, with evapotranspiration intensity path and canal system water flow node space connectivity matching, establish flow evaporation correlation mechanism, according to the cross-path evapotranspiration density ranking and regulation and control element position comparison result, automatically screen key control valve, realize water resource dynamic multidimensional response and high-precision space-time configuration.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a smart management and control platform for the entire irrigation district based on digital twins. Background Technology

[0002] The data processing platform technology field primarily involves the collection, management, analysis, and visualization of large amounts of structured or unstructured data. The data types covered include real-time streaming data, historical database information, and heterogeneous data sources. Its core aspects include data modeling, information fusion, database system design, data retrieval methods, interactive queries, visualization, and decision support. This technology field is widely applied in scenarios such as urban governance, industrial monitoring, smart agriculture, and energy dispatching. By constructing a unified system platform, it efficiently processes and organically integrates distributed data, supporting intelligent decision-making mechanisms and system-level management capabilities. Among them, the traditional intelligent management and control platform for the entire irrigation area refers to the technical system built in farmland irrigation areas to achieve comprehensive management goals such as water resource allocation, scheduling and water use monitoring. It mainly addresses technical issues such as untimely water quantity measurement and control, delayed scheduling information and low irrigation efficiency. The traditional intelligent management and control platform for the entire irrigation area usually uses on-site sensing devices such as water level gauges and flow meters to collect data through remote monitoring terminals and transmit it to the management backend. It implements manual control or semi-automatic scheduling operations in combination with preset time periods or experience models. It realizes irrigation area allocation, water source scheduling and irrigation cycle determination through predefined rules. The management logic is based on static zoning or single water source supply models and lacks dynamic response capabilities and system-level digital modeling.

[0003] Traditional irrigation district management models typically rely on on-site sensing devices to collect data and transmit it to the backend. Manual and semi-automatic scheduling is then implemented in conjunction with preset time periods or experience models. The management logic is mostly based on static zoning or single water source supply models, lacking the ability to dynamically respond to water resources across the entire area and system-level digital modeling. This results in untimely water quantity monitoring and control, as well as serious lag in scheduling information. It is difficult to cope with complex and ever-changing irrigation demands, leading to low efficiency in water resource allocation, a lack of scientific basis for setting irrigation cycles, and an inability to achieve refined water resource allocation. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide a smart management and control platform for the entire irrigation district based on a digital twin. On one hand, a smart management and control platform for the entire irrigation district based on a digital twin is provided, the system comprising:

[0005] The disturbance structure construction module collects the water level time series from the headwater level sensor, the inflow and outflow times series from the boundary flow meter, and the pressure difference change images from the upstream and downstream pressure sensors. It calculates the vertical fluctuation of water level, the boundary frequency change value, and the number of water potential reversals, and combines them to generate a three-dimensional disturbance feature matrix.

[0006] The rhythm segmentation module reads the disturbance time series in the three-dimensional disturbance feature matrix, performs amplitude comparison on the number and spacing of fluctuation peaks, divides the disturbance into dense and sparse areas, numbers and identifies high-disturbance areas and low-disturbance areas, and generates a time-series disturbance zoning map of the irrigation area.

[0007] The evapotranspiration layer generation module obtains the irrigation area time-series disturbance zoning map, extracts the evapotranspiration record sequence, calculates the evapotranspiration difference curve for each pair of plots, filters areas with frequent difference reversals, and constructs an evapotranspiration difference thermal layer.

[0008] The fluctuation mapping module extracts the path number with the maximum evaporation intensity from the evaporation difference thermal layer, obtains the corresponding water flow node number in the canal system map, compares the connectivity of the numbers, filters the connectivity combinations, and generates a flow-evaporation correlation map.

[0009] The regulation map generation module reads the numbered paths in the flow-evapotranspiration correlation map, extracts the evapotranspiration density ranking, compares the positions of regulation elements in the canal system records, filters the control valve numbers of the fields to be controlled, and constructs a smart management and control map of the entire irrigation area.

[0010] As a further embodiment of the present invention, the three-dimensional disturbance feature matrix includes a water level vertical fluctuation index, a boundary frequency change index, and a water potential reversal frequency index; the irrigation district temporal disturbance zoning map includes a high-disturbance area number, a low-disturbance area number, and a disturbance interval type label; the evapotranspiration difference thermal layer includes an evapotranspiration difference amplitude distribution, an evapotranspiration difference reversal frequency, and evapotranspiration dynamic distribution characteristics; the flow-evapotranspiration correlation map includes evapotranspiration path plot numbers, water flow change node numbers, and spatial connectivity matching relationships; and the irrigation district-wide intelligent management and control map includes an evapotranspiration density ranking sequence, control element location information, and a set of field control valve numbers.

[0011] As a further aspect of the present invention, the dense disturbance region refers to the region in the disturbance time series that has a large number of fluctuation peaks and small spacing, while the sparse disturbance region is the region that has a small number of fluctuation peaks and large spacing.

[0012] As a further aspect of the present invention, the frequently reversing difference region refers to the spatiotemporal region in the evaporation difference curve where positive and negative changes are frequent and the direction switching rate is high.

[0013] As a further aspect of the present invention, the disturbance structure construction module includes:

[0014] The data receiving submodule collects the water level time series from the headwater level sensor, the inflow and outflow sequence from the boundary flow meter, and the pressure difference trend image from the upstream and downstream pressure sensors. It then performs time alignment and missing data imputation to generate the boundary disturbance input dataset.

[0015] The disturbance extraction submodule calculates the adjacent differences in water level time series based on the boundary disturbance input dataset to obtain the vertical fluctuation of water level, counts the inflow and outflow frequency per unit time to obtain the boundary frequency change value, extracts the slope change symbol sequence count in the pressure difference image to obtain the number of water potential reversals, and generates a disturbance response index set.

[0016] The feature matrix generation submodule calls the vertical fluctuation of water level, the change of boundary frequency and the number of water potential reversals in the disturbance response index set, and combines them in dimensional order to fill the axes and generate a three-dimensional disturbance feature matrix.

[0017] As a further aspect of the present invention, the rhythm segmentation module includes:

[0018] The time series extraction submodule reads the numerical sequence of the perturbation index along the time axis in the three-dimensional perturbation feature matrix, establishes a unified index structure according to the time step, integrates them into a time series data group, and generates a time series set of perturbation index.

[0019] The disturbance interval identification submodule extracts the location of local extreme points of continuous fluctuations in each sequence based on the disturbance index time series set, calculates the amplitude difference and time interval between adjacent extreme points, counts the fluctuation density according to a preset time window, assigns high-density windows as high-disturbance areas and low-density windows as low-disturbance areas, and generates a disturbance area classification sequence.

[0020] The time-series partitioning output submodule calls the perturbation area classification sequence, maps it to a two-dimensional region structure according to the time index, constructs a corresponding numbered matrix layer, and generates a time-series perturbation partitioning map of the irrigation area.

[0021] As a further aspect of the present invention, the evaporation layer generation module includes:

[0022] The area extraction submodule obtains the plot numbers marked as high-disturbance areas in the irrigation district time-series disturbance zoning map, retrieves the continuous evapotranspiration record sequence of the corresponding plot, aligns the sequence according to the time index and establishes a mapping relationship to generate an evapotranspiration sequence mapping set.

[0023] The difference curve generation submodule performs time synchronization difference operation on any two evaporation record sequences based on the evaporation sequence mapping set, calculates the evaporation difference between adjacent time steps, detects the number of reversals of the difference sign in continuous segments and filters segments according to the reversal density, and generates evaporation difference reversal interval groups.

[0024] The heat map output submodule calls the evapotranspiration difference inversion interval group, establishes a two-dimensional raster structure according to the segment position and inversion density, assigns numerical weights to the raster and forms a thermal distribution layer, and generates an evapotranspiration difference thermal layer.

[0025] As a further aspect of the present invention, the wave mapping module includes:

[0026] The path extraction submodule obtains the evapotranspiration intensity value of the grid area in the evapotranspiration difference thermal layer, traverses the continuous grid path along the direction of the maximum intensity value, extracts the plot number sequence on the path and stores it in order, and generates the evapotranspiration main path number group.

[0027] The node association extraction submodule calls the main evapotranspiration path number group, finds the water flow change record position corresponding to the path according to the preset canal system map, extracts the canal system control node identifier corresponding to the number, records the coordinate information, and generates a path association node number set.

[0028] The connectivity matching output submodule, based on the main evapotranspiration path number group and the path-associated node number set, calls the two-dimensional coordinates of the plots and nodes respectively, performs Euclidean distance determination and filters number pairs with a positional distance less than a threshold, establishes a mapping relationship between the numbers, and generates a flow-evapotranspiration correlation map.

[0029] As a further aspect of the present invention, the regulation spectrum generation module includes:

[0030] The path number parsing submodule reads the number combination in the flow-evapotranspiration correlation map, extracts the set of plot numbers corresponding to the intersecting paths, retrieves the evapotranspiration density sequence of the path segment in the evapotranspiration difference thermal layer, sorts them in descending order of density value, and generates a path evapotranspiration density ranking table.

[0031] The control element screening submodule locates the plot number corresponding to the top-ranked path segment according to the path evapotranspiration density ranking table, calls the control element coordinate set recorded in the canal system record, compares the position of each path segment with the position of the control element according to the spatial overlap principle, filters the element numbers with spatial overlap relationship, and generates the valve number set of the control zone.

[0032] The map output submodule calls the flow-evaporation correlation map and the set of valve numbers for the control zone. Based on the number mapping relationship, it constructs a graphical data structure including path structure, control unit location and control number to generate a smart management and control map of the entire irrigation area.

[0033] As a further aspect of the present invention, the spatial overlap principle means that when the geographical location range of the path segment coincides with and intersects with the coordinate area of ​​the control element, it is considered that there is a spatial relationship.

[0034] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0035] By collecting multi-dimensional sensor data to construct a three-dimensional disturbance feature matrix, the system accurately identifies the fluid fluctuation patterns and rhythmic characteristics within the irrigation area. Based on temporal changes, it divides high and low disturbance regions, utilizes plot evapotranspiration records to invert the thermal distribution of high-dynamic areas, matches the spatial connectivity of evapotranspiration intensity paths with canal flow nodes, establishes a flow-evapotranspiration correlation mechanism, and automatically selects key control valves based on the ranking of evapotranspiration density along cross paths and the comparison of control element locations. This constructs a comprehensive control map, realizing a transformation from static single scheduling to dynamic multi-dimensional response, significantly improving the accuracy of water resource spatiotemporal allocation and the systematization level of irrigation scheduling. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a schematic diagram of the process of this invention;

[0038] Figure 2 This is a flowchart of the disturbance structure construction module in this invention;

[0039] Figure 3 This is a flowchart of the rhythm segmentation module in this invention;

[0040] Figure 4 This is a flowchart of the evaporation layer generation module in this invention;

[0041] Figure 5 This is a flowchart of the wave mapping module in this invention;

[0042] Figure 6 This is a flowchart of the control spectrum generation module in this invention. Detailed Implementation

[0043] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0044] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0045] This invention provides a smart management and control platform for the entire irrigation district based on digital twins, such as... Figure 1 The diagram shown illustrates a comprehensive intelligent management and control platform system for irrigation districts based on digital twins. This system includes:

[0046] The disturbance structure construction module collects water level time series from water level sensors at the head of the canal, obtains inflow and outflow frequency series from flow meters at the boundary, and extracts pressure difference change trend images from pressure sensors at upstream and downstream nodes. It calculates the vertical fluctuation of water level, boundary frequency change value and water potential reversal number respectively, and combines the three results according to the dimension and position to construct a three-dimensional disturbance feature matrix.

[0047] The rhythm segmentation module reads the time axis change sequence corresponding to the disturbance index in the three-dimensional disturbance feature matrix, performs amplitude comparison operation on the number of fluctuation peaks and interval spacing, divides the disturbance dense distribution area and sparse distribution area according to the time window, identifies the high disturbance area and low disturbance area based on the interval type number, and generates the irrigation area time series disturbance zoning map.

[0048] The evapotranspiration layer generation module obtains the continuous evapotranspiration record sequence of the corresponding plots based on the plot numbers identified as high-disturbance areas in the irrigation district temporal disturbance zoning map. It calculates the evapotranspiration difference curve from the record sequence between each pair of plots, filters out curve areas with frequent inversion of the difference, and uses them as features of high evapotranspiration dynamic areas to construct an evapotranspiration difference thermal layer.

[0049] The fluctuation mapping module extracts the plot number of the path with the maximum evapotranspiration intensity from the evapotranspiration difference thermal layer, calls the water flow change node number associated with the path in the canal system map, performs a position comparison operation on the spatial connectivity between the plot number and the canal system node number in the path, filters the mutually connected number combinations, and generates a flow-evapotranspiration correlation map.

[0050] The control map generation module reads the cross paths of each numbered combination in the flow-evapotranspiration correlation map, extracts the evapotranspiration density ranking on each path in the thermal layer, compares the location of the control element corresponding to the path from the canal system record, and selects the field zoning control valve numbers that need to be controlled to construct a smart control map of the entire irrigation area.

[0051] The three-dimensional disturbance feature matrix includes indicators of vertical water level fluctuation, boundary frequency change, and water potential reversal frequency. The irrigation district temporal disturbance zoning map includes high-disturbance area numbers, low-disturbance area numbers, and disturbance interval type labels. The evapotranspiration difference thermal layer includes evapotranspiration difference amplitude distribution, evapotranspiration difference reversal frequency, and evapotranspiration dynamic distribution characteristics. The flow-evapotranspiration correlation map includes evapotranspiration path plot numbers, water flow change node numbers, and spatial connectivity matching relationships. The irrigation district-wide intelligent management and control map includes evapotranspiration density ranking sequence, control element location information, and field control valve number set.

[0052] Specifically, such as Figure 2 As shown, the disturbance structure building module includes:

[0053] The data receiving submodule collects the water level time series from the headwater level sensor, the inflow and outflow sequence from the boundary flow meter, and the pressure difference trend image from the upstream and downstream pressure sensors. It then performs time alignment and missing data imputation to generate the boundary disturbance input dataset.

[0054] First, a two-way communication link is established with the on-site sensing equipment at the canal head via an industrial-grade IoT gateway. This link, based on the Modbus RTU communication protocol, polls at a 100-millisecond cycle to read the analog signals from the headwater level sensor and the pulse signals from the boundary flow meter. For the 4-20 mA current signal uploaded by the water level sensor, the module internally uses a high-precision analog-to-digital converter to parse it into the corresponding water level value, accurate to 0.001 meters, and stores it in the original buffer queue in chronological order, combined with the received system timestamp. For the boundary flow meter, the module accumulates the pulse count values ​​of its inflow and outflow directions in real time, marking them as forward inflow count and reverse outflow count, forming a discrete sequence of inflow and outflow counts. Simultaneously, the module calls the image generation unit integrated into a high-speed industrial camera or pressure transmitter to capture pressure difference trend images of the cross-sections where the upstream and downstream pressure sensors are located at 1-minute intervals. These images are bitmap formats with a resolution of 1920 x 1080 pixels, recording the pressure fluctuation pattern of the fluid passing through the boundary cross-section. Subsequently, the module initiates a time alignment process, setting a standard time axis with a time step of 5 minutes, and performs resampling operations on the three types of heterogeneous data. For water level data, linear interpolation is used to fill data gaps at non-standard time points; for flow meter entry / exit times, the cumulative frequency within the standard time step is calculated; for differential pressure images, the image frame closest to the standard time point is selected as the representative data for that moment. After alignment, the module executes missing data imputation logic, traversing the entire time series. If the number of consecutive missing points is less than or equal to 3, the arithmetic mean of the points before and after the missing segment is calculated for imputation; if the number of consecutive missing points is greater than 3, the historical database for the same period is called, and the weighted average of the data from the same time period over the past 7 days is calculated as the imputation value, with the weighting coefficient decreasing linearly with increasing time distance. Finally, the module encapsulates the cleaned, aligned, and imputed water level time series, boundary flow meter entry / exit times series, and differential pressure trend image series with a unified time index to construct a boundary disturbance input dataset.

[0055] Assuming the current time is 08:00 on January 2nd, and the standard time step is set to 10 minutes, the module collected the original water level value at 08:02 as 3.52 meters and at 08:12 as 3.58 meters. For the standard time node of 08:10, due to the lack of direct observations, the module performs linear interpolation. The specific mathematical calculation steps are as follows: Step 1: Calculate the time difference between the target time and the previous observation time. That is, subtract the previous time 08:02 from the target time 08:10, resulting in a time difference of 8 minutes. Step 2: Calculate the total time interval between the two observations. That is, subtract 08:02 from 08:12, resulting in a total interval of 10 minutes. Step 3: Calculate the time interpolation ratio. Divide the time difference of 8 minutes obtained in Step 1 by the total interval of 10 minutes obtained in Step 2, resulting in a ratio of 0.8. Step 4: Calculate the difference between the two water level observations. The difference between the preceding and following water levels is calculated as follows: Subtracting the preceding water level from the following water level (3.58 meters) yields a water level difference of 0.06 meters. Step 5: Calculate the interpolation increment. Multiply the water level difference of 0.06 meters obtained in Step 4 by the interpolation ratio of 0.8 obtained in Step 3, resulting in 0.048 meters. Step 6: Calculate the final aligned water level value. Add the interpolation increment of 0.048 meters obtained in Step 5 to the preceding water level of 3.52 meters, resulting in 3.568 meters. Similarly, for the number of flow meter inflows and outflows, if the cumulative positive pulses from 08:00 to 08:10 are 50 and the negative pulses are 2, then the sequence of inflows and outflows during this period is directly recorded as (50, 2). Through this specific numerical calculation and logical processing, the strict synchronization and integrity of the input dataset in the time dimension are ensured.

[0056] Table 1. Monitoring Point Data Acquisition and Alignment Processing

[0057]

[0058] The disturbance extraction submodule calculates the adjacent differences in water level time series based on the boundary disturbance input dataset to obtain the vertical fluctuation of water level, counts the inflow and outflow frequency per unit time to obtain the boundary frequency change value, extracts the slope change symbol sequence count in the pressure difference image to obtain the number of water potential reversals, and generates a disturbance response index set.

[0059] The module reads the boundary disturbance input dataset and first performs a first-order difference operation on the water level time series. It iterates through each discrete point on the time axis, subtracting the previous water level value from the current value to obtain the absolute height difference between the two points. This height difference represents the instantaneous fluctuation amplitude of the water body in the vertical direction, thus generating a water level vertical fluctuation sequence. Next, the module processes the inflow and outflow frequency sequence of the boundary flow meter, setting a unit statistical time window of 1 hour. Within this window, the inflow and outflow frequencies are summed and statistically analyzed. Then, the absolute value of the difference between the total inflow and outflow frequencies is calculated and divided by the unit time window length to obtain the boundary frequency change value, reflecting the activity level of boundary water flow exchange. For the pressure difference trend image, the module first uses the Canny edge detection algorithm to extract the pressure curve contour in the image and converts the image pixel coordinates into pressure-time coordinate system data. Subsequently, the module performs piecewise linear fitting on the extracted pressure curve and calculates the slope value of each small line segment. The module sets up a slope sign change judgment logic. When the product of the slopes of two consecutive line segments is less than 0, it is judged as a water potential reversal event. The module counts the total number of reversal events occurring within a unit time step (e.g., 10 minutes) to quantify the degree of turbulence in the internal pressure of the water flow and generate a sequence of water potential reversal times. Finally, the module encapsulates the vertical water level fluctuation, boundary frequency change value, and water potential reversal number into tuples using the same time index to generate a disturbance response index set. Specific examples are explained below: Calculation steps for vertical water level fluctuation: Step 1: Obtain the water level value at time point t as 3.568 meters and the water level value at time point t-1 as 3.520 meters. Step 2: Perform a subtraction operation: subtract 3.520 meters from 3.568 meters to obtain a difference of 0.048 meters. This value is the vertical water level fluctuation at that moment. Calculation steps for boundary frequency change value: Step 1: Set the time window to 1 hour, and count the total inflow frequency of 300 times and the total outflow frequency of 20 times within this window. Step 2: Calculate the absolute value of the frequency difference, i.e., 300 minus 20, resulting in 280. Step 3: Calculate the frequency change value. Divide the result of Step 2, 280, by the time window of 1 hour to obtain the boundary frequency change value of 280 times / hour. Calculation steps for the number of water potential reversals: Step 1: Extract the pressure curve slope sequence within 10 minutes: 0.5, 0.2, -0.1, -0.3, 0.4. Step 2: Perform adjacent slope product judgment. Calculate 0.2 multiplied by -0.1, the result is -0.02 (less than 0), judged as the 1st reversal; calculate -0.3 multiplied by 0.4, the result is -0.12 (less than 0), judged as the 2nd reversal. Step 3: Count the total number of reversals, the result is 2. Finally, integrate the above calculation results, the data item in the disturbance response index set at this moment is (0.048, 280, 2). This computational logic can transform multi-source heterogeneous physical quantities into numerical indicators with a unified dimension.

[0060] The feature matrix generation submodule calls the vertical fluctuation of water level, the change of boundary frequency and the number of water potential reversals in the disturbance response index set, and combines and fills the axes in dimensional order to generate a three-dimensional disturbance feature matrix.

[0061] First, a three-dimensional data container is defined, with its three axes set as the time dimension axis, the spatial station dimension axis, and the feature attribute dimension axis, respectively. The module receives a set of disturbance response indicators from multiple upstream monitoring stations and first performs standardization processing on the data of each indicator. For three types of data with inconsistent dimensions—vertical water level fluctuation, boundary frequency change, and water potential reversal number—the module uses the range standardization method (Min-MaxNormalization) to find the maximum and minimum values ​​of each indicator in the historical period. The current real-time data is subtracted from the minimum value and then divided by the difference between the maximum and minimum values, thus mapping all indicator values ​​to a closed interval of 0 to 1. After standardization, the module fills the processed indicator data into the corresponding positions of the spatial station dimension axis according to the geographic number of the monitoring station; it fills it into the corresponding scale of the time dimension axis according to the data collection timestamp; and it fills it into the corresponding level of the feature attribute dimension axis according to the type of indicator (i.e., fluctuation, frequency value, and reversal number). The module continuously adds new data slices along the time axis as the time step progresses, and finally constructs a dynamically updated three-dimensional disturbance feature matrix in memory. The specific calculation example is as follows: Assume that the data being processed is from monitoring station ID_01, whose vertical water level fluctuation in the disturbance response index set is 0.048 meters. The module retrieves historical records, setting the historical minimum water level fluctuation to 0 meters and the historical maximum to 0.5 meters. The standardized calculation steps for the vertical water level fluctuation are as follows: Step 1: Calculate the numerator, i.e., the current value 0.048 minus the historical minimum value 0, resulting in 0.048. Step 2: Calculate the denominator (range), i.e., the historical maximum value 0.5 minus the historical minimum value 0, resulting in 0.5. Step 3: Perform a division operation, dividing the result of Step 1, 0.048, by the result of Step 2, 0.5, resulting in 0.096. Similarly, for the boundary frequency change value, if the real-time value is 280 times / hour and the historical range is 0 to 1000 times / hour, calculate the difference between 280 and 0, then divide by the difference between 1000 and 0, resulting in 0.28. Regarding the number of water potential reversals, if the real-time value is 2 times and the historical range is 0 to 10 times, calculate the difference between 2 and 0, then divide by the difference between 10 and 0, resulting in 0.2. At this point, for time point t and station ID_01, the module fills the vector [0.096, 0.28, 0.2] at the coordinate position (t, ID_01, :) in the feature matrix. By constructing this three-dimensional feature matrix, the system can comprehensively express the perturbation state of the entire irrigation area in the spatiotemporal dimension in tensor form.

[0062] Specifically, such as Figure 3 As shown, the rhythm segmentation module includes:

[0063] The time series extraction submodule reads the numerical sequence of the perturbation index along the time axis in the three-dimensional perturbation feature matrix, establishes a unified index structure according to the time step, integrates them into a time series data group, and generates a time series set of perturbation index;

[0064] This module directly accesses the 3D perturbation feature matrix in memory, locks the feature attribute dimension axis, and extracts all data from the water level vertical fluctuation layer, boundary frequency change value layer, and water potential reversal number layer through slicing operations. For each feature layer, the module performs dimensionality reduction processing on the spatial station dimension axis or retains the data as multi-channel input. Here, the module chooses to arrange the same indicator data from different stations according to spatial adjacency, integrating them into a multi-dimensional time series vector. The module establishes a unified time index structure, which uses year, month, day, hour, minute, and second as keys, strictly corresponding to the time dimension axis scale in the feature matrix. The module traverses each time step on the time axis, concatenating the extracted feature indicator values ​​in a predefined order (e.g., water level first, then frequency, then reversal number) to form a comprehensive feature vector for that time step. Subsequently, the module concatenates the comprehensive feature vectors of consecutive time steps in time sequence to generate a continuous perturbation indicator time series set. A specific example is illustrated below: Assume the 3D feature matrix contains data from two stations. At time step t1, the standardized feature vector of station 1 is [0.1, 0.2, 0.3], and the standardized feature vector of station 2 is [0.4, 0.5, 0.6]. The steps for concatenating the comprehensive feature vector are as follows: Step 1: Extract the feature components of station 1 [0.1, 0.2, 0.3]. Step 2: Extract the feature components of station 2 [0.4, 0.5, 0.6]. Step 3: Perform a vector concatenation operation, appending the vector from step 2 directly to the end of the vector from step 1 to form a new vector [0.1, 0.2, 0.3, 0.4, 0.5, 0.6]. Step 4: Bind this comprehensive feature vector of length 6 to time index t1. If the time step is 10 minutes and data from the past hour needs to be extracted, the module will perform this extraction and concatenation operation 6 times in sequence, ultimately generating a sequence set containing 6 comprehensive feature vectors. For example, the extracted time series set of disturbance indicators can be represented as a matrix with dimensions of 6 rows (time steps) multiplied by 6 columns (number of features). This process transforms complex spatial tensor data into a standardized sequence format suitable for time series analysis algorithms.

[0065] The disturbance interval identification submodule extracts the location of local extreme points of continuous fluctuations in each sequence based on the time series set of disturbance indicators, calculates the amplitude difference and time interval between adjacent extreme points, counts the fluctuation density according to the preset time window, assigns high-density windows as high-disturbance areas and low-density windows as low-disturbance areas, and generates a disturbance area classification sequence.

[0066] The module receives a time series set of disturbance indicators and employs a sliding window algorithm, setting the window width to 6 time steps (i.e., 1 hour). Within each window, the module first identifies local extreme points in the sequence. The judgment logic is: if the indicator value at a certain time point is strictly greater than the values ​​at its immediate preceding and following time points, it is marked as a maximum point; otherwise, it is marked as a minimum point. The module records the time position and numerical amplitude of all extreme points. Subsequently, the module calculates the amplitude difference (fluctuation amplitude) and time interval (fluctuation period) between adjacent extreme points. The module introduces a fluctuation density indicator, which is calculated as: the sum of the absolute values ​​of the amplitude differences of all adjacent extreme points within the window, divided by the time length of the window. The module sets a disturbance density benchmark threshold, which is obtained through statistical analysis of historical normal water period data, taking 1.5 times the historical average fluctuation density as the dividing line. If the fluctuation density calculated for the current window is greater than the threshold, the time period corresponding to the window is marked as a "high disturbance area"; if it is less than or equal to the threshold, it is marked as a "low disturbance area". The module performs this classification operation on all time windows and concatenates the classification results in chronological order to generate a perturbation region classification sequence. A specific example is illustrated below: As shown in Table 2, the module selects a 60-minute time series for analysis. Assume extreme points are detected at the 10th, 20th, and 40th minutes, with amplitudes of 0.8, 0.2, and 0.9 respectively. The fluctuation density determination calculation steps are as follows: Step 1: Calculate the amplitude difference between adjacent extreme points in the first segment. Subtract 0.2 from the extreme point value of 0.8, and take the absolute value to obtain 0.6. Step 2: Calculate the amplitude difference between adjacent extreme points in the second segment. Subtract 0.9 from the extreme point value of 0.2, and take the absolute value to obtain 0.7. Step 3: Calculate the cumulative sum of amplitude differences within the window. Add the result of Step 1 (0.6) to the result of Step 2 (0.7), obtaining a total of 1.3. Step 4: Calculate the fluctuation density. Divide the total of 1.3 by the window length of 1 hour, resulting in 1.3. Step 5: Calculate the baseline threshold. Assuming the historical average fluctuation density is 0.5, multiply it by a coefficient of 1.5 to obtain a threshold of 0.75. Step 6: Perform a comparison and judgment. Compare the current fluctuation density of 1.3 with the threshold of 0.75. Since 1.3 is greater than 0.75, this 1-hour window is determined to be a high-disturbance area. This calculation logic achieves accurate capture and classification of unstable water flow states by quantifying the density of fluctuations.

[0067] Table 2. Disturbance Interval Identification and Classification Calculation Table

[0068]

[0069] The time-series partitioning output submodule calls the perturbation area classification sequence, maps it to a two-dimensional region structure according to the time index, constructs the corresponding number matrix layer, and generates a time-series perturbation partitioning map of the irrigation area;

[0070] The module reads the perturbation zone classification sequence, which consists of a series of time-labeled "high perturbation" or "low perturbation" status markers. It establishes a two-dimensional region structure, where the horizontal axis corresponds to the time axis and the vertical axis corresponds to the spatial region numbers of the irrigation district. The module iterates through the classification sequence, mapping the classification result of each time period to the corresponding spatial region. For time periods marked as "high perturbation zones," the module fills the corresponding two-dimensional grid with a specific color code (e.g., red code value 255, 0, 0); for "low perturbation zones," it fills with another color code (e.g., green code value 0, 255, 0). The module stacks these filled grids according to region numbers to construct a multi-layered matrix structure. Each layer represents the perturbation state distribution of a specific irrigation district on a continuous time axis. Finally, the module renders this matrix structure as a visual graphic file or serializes it into a JSON data packet, generating a temporal perturbation zoning map of the irrigation district. A specific example is illustrated below: Assume the irrigation district is divided into region A and region B. The classification sequence shows that region A experiences high perturbation from 08:00 to 09:00 and low perturbation from 09:00 to 10:00; region B experiences low perturbation from 08:00 to 09:00 and high perturbation from 09:00 to 10:00. The two-dimensional matrix assignment calculation steps are as follows: Step 1: Process the row vectors of region A. Assign a value of 1 (representing high perturbation) at the index positions corresponding to 08:00 to 09:00; assign a value of 0 (representing low perturbation) at the index positions corresponding to 09:00 to 10:00. Step 2: Process the row vectors of region B. Assign a value of 0 (representing high perturbation) at the index positions corresponding to 08:00 to 09:00; assign a value of 1 (representing low perturbation) at the index positions corresponding to 09:00 to 10:00. Step 3: Construct the final matrix. Combine the vectors from steps 1 and 2 to form a matrix data structure [[1, 0], [0, 1]]. This process transforms the abstract time series classification results into an intuitive spatial distribution map, clearly showing the perturbation status of different regions at different times.

[0071] Specifically, such as Figure 4 As shown, the evaporation layer generation module includes:

[0072] The submodule for extracting the irrigated area obtains the plot numbers marked as high-disturbance areas in the irrigation district time-series disturbance zoning map, retrieves the continuous evapotranspiration record sequence of the corresponding plot, aligns the sequence according to the time index and establishes a mapping relationship, and generates an evapotranspiration sequence mapping set.

[0073] The module parses the input irrigation district temporal perturbation zoning map, and filters out all grid cells with values ​​marked as high perturbation (e.g., a value of 1) by traversing the state matrix in the map. The module extracts the corresponding plot numbers and time periods for these grid cells. Then, using these plot numbers as primary keys, the module initiates a query request to the irrigation district meteorological and soil monitoring database to retrieve continuous evapotranspiration records for the corresponding plots within the corresponding time periods. The evapotranspiration records contain real-time evapotranspiration rate data (in millimeters per hour). After acquiring the data, the module cuts and aligns the retrieved evapotranspiration sequences according to the start and end times recorded in the zoning map, ensuring that all extracted sequences have the same start point and time step. The module establishes a hash mapping table to map plot numbers to their corresponding aligned evapotranspiration data sequences, generating an evapotranspiration sequence mapping set. A specific example is as follows: Assume that the zoning map shows plot ID_101 in a high-perturbation state from 10:00 to 12:00. After reading this information, the module queries the evapotranspiration data of ID_101 during that time period, obtaining the sequence data [0.5, 0.6, 0.55, 0.7] (one data point every 30 minutes). If plot ID_102 is also marked as high disturbance at the same time, its sequence is obtained as [0.4, 0.45, 0.4, 0.5]. Mapping set construction steps: Step 1: Create key-value pair 1. Bind the key "ID_101" to the value list [0.5, 0.6, 0.55, 0.7]. Step 2: Create key-value pair 2. Bind the key "ID_102" to the value list [0.4, 0.45, 0.4, 0.5]. Step 3: Store the key-value pairs from Steps 1 and 2 into a hash table, forming a mapping set {ID_101: [0.5, 0.6, 0.55, 0.7], ID_102: [0.4, 0.45, 0.4, 0.5]}. This process accurately identifies water consumption behavior data in the affected area, providing direct data objects for analyzing the specific impact of the disturbance on crop evapotranspiration.

[0074] The difference curve generation submodule is based on the evapotranspiration sequence mapping set. It performs time synchronization difference operation on any two evapotranspiration record sequences, calculates the evapotranspiration difference between adjacent time steps, detects the number of reversals of the difference sign in continuous segments and filters segments according to the reversal density, and generates evapotranspiration difference reversal interval groups.

[0075] Based on the evapotranspiration sequence mapping set, the module uses a full permutation combination method to select any two plots of evapotranspiration record sequences for pairwise analysis. For each pair of sequences, the module performs time-synchronized differential operations, that is, at each identical time step, the evapotranspiration value of the first sequence is subtracted from the evapotranspiration value of the second sequence to generate an evapotranspiration difference curve. The module then scans this difference curve, detecting changes in the sign of the difference. If the difference is positive at a certain time step, and becomes negative (or vice versa) at the next time step, a "sign inversion" is determined. The module counts the total number of inversions in the entire difference curve and calculates the inversion density, which is the total number of inversions divided by the total time length of the sequences. The module sets an inversion density screening threshold (e.g., once per hour), retaining sequence pairs and their corresponding difference intervals whose inversion density is greater than this threshold. These high-frequency inverted difference intervals indicate significant asynchrony or disorder in the evapotranspiration behavior between the two plots, and the module encapsulates them to generate evapotranspiration difference inversion interval groups. The specific example is explained below: Select plot A sequence [0.5, 0.6, 0.55, 0.7] and plot B sequence [0.4, 0.7, 0.5, 0.8]. The steps for calculating the difference and inversion density are as follows: Step 1: Calculate the difference sequence. The first point is 0.5 minus 0.4, which equals 0.1 (positive); the second point is 0.6 minus 0.7, which equals -0.1 (negative); the third point is 0.55 minus 0.5, which equals 0.05 (positive); the fourth point is 0.7 minus 0.8, which equals -0.1 (negative). The resulting difference sequence is [0.1, -0.1, 0.05, -0.1]. Step 2: Count the number of sign inversions. The first inversion is from 0.1 to -0.1; the second inversion is from -0.1 to 0.05; the third inversion is from 0.05 to -0.1. A total of 3 inversions are recorded. Step 3: Calculate the inversion density. Assuming the four points cover a two-hour duration, dividing the number of reversals (3) by the duration of two hours yields a result of 1.5 reversals per hour. Step 4: Perform threshold filtering. Compare the calculated result of 1.5 with the preset threshold of 1.0. Since 1.5 is greater than 1.0, the interval is deemed valid and included in the reversal interval group. This logic effectively reveals the differences and instabilities in the evapotranspiration response of different plots under disturbance through frequent changes in the sign of the difference.

[0076] The heat map output submodule calls the evapotranspiration difference inversion interval group, establishes a two-dimensional raster structure according to the segment position and inversion density, assigns numerical weights to the raster and forms a thermal distribution layer, and generates an evapotranspiration difference thermal layer.

[0077] The module invokes the evapotranspiration difference inversion interval grouping. First, it constructs a two-dimensional rasterized structure consistent with the geographic scale of the irrigation area, dividing it into several small square grids (e.g., 10 meters by 10 meters). For each pair of plots within the interval group, the module identifies their spatial location and connects these two locations in the two-dimensional raster to form a virtual line. The module uses the calculated inversion density value as a weight to accumulate the values ​​of the raster cells covered by this line. Specifically, if a raster is located on the line connecting plot A and plot B, and the inversion density between A and B is 1.5, then the value of that raster increases by 1.5. The module iterates through all selected interval groups, completing the accumulation of all weights. Next, the module normalizes the entire raster matrix, mapping the values ​​to a color scale range of 0 to 255; higher values ​​indicate more drastic fluctuations in evapotranspiration differences in the area. Finally, the module generates the raster matrix with weighted values ​​as an image-formatted evapotranspiration difference thermal layer. The specific example is explained below: Assume that grid G ​​(5, 5) is located on the line connecting plots A and B, and simultaneously on the line connecting plots C and D. The inversion density of A and B is 1.5, and the inversion density of C and D is 2.0. The steps for accumulating and normalizing thermal values ​​are as follows: Step 1: Initialize the value of grid G ​​(5, 5) to 0. Step 2: Accumulate the weight of the first connection. Add the initial value of 0 to the density value of AB (1.5), resulting in 1.5. Step 3: Accumulate the weight of the second connection. Add the result of Step 2 (1.5) to the density value of CD (2.0), resulting in 3.5. Step 4: Assume the maximum accumulated value found in the entire graph is 10.0, and perform a normalization operation. Divide the current grid value of 3.5 by the maximum value of 10.0 to obtain a scale of 0.35. Step 5: Calculate the color level value. Multiply the scale of 0.35 by the upper limit of the color level of 255, resulting in 89.25, which is rounded down to 89. The raster will be rendered as the corresponding color level 89 in the heatmap. Through this overlay calculation, the module transforms discrete sequence differences into a continuous spatial thermal distribution, visually presenting the core area within the irrigation district most severely affected by disturbances.

[0078] Specifically, such as Figure 5 As shown, the wave mapping module includes:

[0079] The path extraction submodule obtains the evapotranspiration intensity value of the raster area in the evapotranspiration difference thermal layer, traverses the continuous raster path along the direction of the maximum intensity value, extracts the plot number sequence on the path and stores it in order, and generates the main evapotranspiration path number group.

[0080] The module reads raster data from the evapotranspiration difference thermal layer. First, it scans the entire map to find the raster point with the highest pixel grayscale value (i.e., evapotranspiration intensity value) as the starting node. Then, using the reverse logic of gradient descent (i.e., gradient ascent or steepest ascent), it searches its 8-neighborhood (up, down, left, right, and diagonal directions) from the starting node, finding the neighboring raster with the highest value as the next path point. The module repeats this process, continuously extending towards higher values ​​until the value of the adjacent raster is lower than the current raster value by a certain percentage (e.g., 90%) or reaches the map boundary, thus forming a continuous high-intensity path. The module records the coordinates of all raster traversed by the path and maps these coordinates back to the irrigation area plot number layer, extracting the plot number sequence covered by the path and storing them sequentially to generate the main evapotranspiration path number group. A specific example is as follows: Assume the maximum value point in the thermal map is located at coordinates (50, 50), with a value of 200. Path extension determination calculation steps: Step 1: Retrieve the neighboring raster values ​​of the starting point (50, 50). Step 1: The value of (50, 51) is 195, and the value of (51, 50) is 180. All other neighboring values ​​are below 180. Step 2: Compare neighboring values. Since 195 is greater than 180, the module selects (50, 51) as the next node in the path. Step 3: Search the neighborhood of the new node (50, 51). The value of (50, 52) is 190, which is the maximum value in its neighborhood. Step 4: Execute the termination condition judgment. Calculate the ratio: divide the new value 190 by the current value 195, and the result is approximately 0.97. Since 0.97 is greater than the set termination ratio of 0.90, the path continues to extend. Step 5: Repeat the above steps to finally form the path coordinate sequence [(50, 50), (50, 51), (50, 52)...]. The module compares the plot number layer and finds that coordinates (50, 50) to (50, 52) all belong to plot ID_205, and the subsequent coordinates belong to plot ID_206. Therefore, the module generates path number groups [ID_205, ID_206]. This process utilizes pathfinding algorithms from image processing to automatically identify the propagation chain with the most intense evapotranspiration fluctuations.

[0081] The node association extraction submodule calls the evapotranspiration main path number group, finds the water flow change record position corresponding to the path according to the preset canal system map, extracts the canal system control node identifier corresponding to the number, records the coordinate information, and generates a path association node number set.

[0082] The module retrieves the parcel number from the main evapotranspiration path number group and performs a correlation query based on a pre-set canal system map database. The canal system map database stores the topology of the irrigation district, including canal routes, parcel locations, and the distribution information of control nodes (such as gates and pumping stations). For each parcel on a path, the module retrieves the upstream canal branch that directly supplies water and traces upwards along that canal branch until the nearest primary control node is found. The module extracts the unique identifier (such as valve number) and its GPS coordinates for that control node. The module integrates the parcel number, the corresponding control node identifier, and the node coordinates into a single correlation record. After traversing all parcels along the path, the module removes duplicate node records and generates a set of path-related node numbers. A specific example is as follows: For parcel ID_205, the module queries the canal system map and finds that it is supplied by "East Main Canal - Branch Canal 03". Correlation tracing steps: Step 1: Locate the water supply inlet of parcel ID_205, identified as "Branch Canal 03". Step 2: Traverse the topological nodes upstream along "Branch Canal 03" until a node of type "Control Gate" is encountered. Step 3: Read the node's attributes, obtaining the ID "Gate_E03" and coordinates (X_gate, Y_gate). Step 4: Repeat steps 1 to 3 for the next plot ID_206, finding the result is still "Gate_E03". Step 5: Perform deduplication. Compare the node IDs obtained in the two steps; since they are identical, only one record is retained. If the path extends to plot ID_301, which is supplied with water by "Branch Canal 04," the control node is Gate_E04. The final set contains {Gate_E03, Gate_E04}. This step achieves a logical reverse inference from the "affected plot" to the "control source," establishing a physical spatial tracing relationship.

[0083] The connectivity matching output submodule, based on the main evapotranspiration path number group and the path-related node number set, calls the two-dimensional coordinates of the plots and nodes respectively, performs Euclidean distance determination and filters number pairs with a position distance less than a threshold, establishes a mapping relationship between the numbers, and generates a flow-evapotranspiration correlation map.

[0084] Based on the center coordinates of plots included in the main evapotranspiration path numbering group and the coordinates of control nodes included in the path-related node numbering set, a spatial distance matching operation is performed. The module iterates through each "plot-node" pair and calculates the Euclidean distance between them. The calculation logic is as follows: first, obtain the difference between the plot's x-coordinate and the node's x-coordinate and square it; then obtain the difference between the plot's y-coordinate and the node's y-coordinate and square it; add the two squared values ​​and take the arithmetic square root to obtain the distance value. The module sets a valid physical connectivity distance threshold (e.g., 500 meters), which is set based on the average control radius of the canal water conveyance. If the calculated distance is less than the threshold and there is a connectivity relationship in the canal system topology, the pairing is retained. The module establishes a strong mapping relationship between the selected plot numbers and control node numbers that meet the conditions, constructs a graph structure containing spatial location and logical correspondence, and generates a flow-evapotranspiration correlation map. A specific example is as follows: the center coordinates of plot ID_205 are (1000, 2000), and the coordinates of control node Gate_E03 are (1200, 2100). Euclidean distance calculation steps: Step 1: Calculate the difference in x-coordinates. Subtract the node's x-coordinate from the plot's x-coordinate of 1000, resulting in -200. Step 2: Calculate the square of the difference in x-coordinates. Multiply -200 by -200, resulting in 40000. Step 3: Calculate the difference in y-coordinates. Subtract the node's y-coordinate from the plot's y-coordinate of 2000, resulting in -100. Step 4: Calculate the square of the difference in y-coordinates. Multiply -100 by -100, resulting in 10000. Step 5: Calculate the sum of squares. Add the result from Step 2 (40000) to the result from Step 4 (10000), resulting in 50000. Step 6: Calculate the distance. Take the square root of the result from Step 5 (50000), resulting in approximately 223.6 meters. Step 7: Perform threshold determination. The calculated distance of 223.6 meters was compared with the preset threshold of 500 meters. Since 223.6 is less than 500, the match was deemed successful. This calculation verified the physical proximity of the control node and the controlled plot, ensuring the operability of the generated map in actual hydraulic control.

[0085] Specifically, such as Figure 6 As shown, the regulatory map generation module includes:

[0086] The path number parsing submodule reads the number combination in the flow-evapotranspiration correlation map, extracts the set of plot numbers corresponding to the intersecting paths, retrieves the evapotranspiration density sequence of the path segment in the evapotranspiration difference thermal layer, sorts them in descending order of density value, and generates a path evapotranspiration density ranking table.

[0087] The module reads the evapotranspiration correlation map and extracts the set of plot numbers involving all intersecting paths. Intersecting paths refer to areas where multiple high-intensity paths converge or overlap in the evapotranspiration difference heatmap; these areas typically represent key nodes of disturbance superposition. For the plots corresponding to these intersection points, the module retrieves the original density data from the evapotranspiration difference heatmap layer. The module obtains the average evapotranspiration density value of all rasters on each path segment and sorts all path segments in descending order based on this value. Path segments with higher average density values ​​indicate a higher degree of evapotranspiration anomaly and require priority processing. The module generates a path evapotranspiration density ranking table containing path IDs, corresponding plot numbers, and average evapotranspiration densities. A specific example is as follows: Assume there are two intersecting paths in the system, path P1 passes through plots A and B, and path P2 passes through plots B and C. Plot B is the intersection point. Average density sorting calculation steps: Step 1: Calculate the density values ​​of all rasters covered by path P1 and calculate their arithmetic mean, which is 85. Step 2: Calculate the arithmetic mean of the density values ​​of all grid cells covered by path P2, which is 95. Step 3: Compare the average density values ​​of the two paths. Since 95 is greater than 85, P2 is determined to have a higher priority than P1. Step 4: Generate the ranking results. Rank P2 first and P1 second, forming the ranking table: First place: P2 (density 95), Second place: P1 (density 85). This process, by quantitatively comparing the disturbance intensity of different paths, provides a basis for subsequent resource allocation and control priority decisions.

[0088] The control element screening submodule locates the plot number corresponding to the top-ranked path segment based on the path evapotranspiration density ranking table, calls the control element coordinate set recorded in the canal system record, compares the position of each path segment with the position of the control element according to the spatial overlap principle, filters the element numbers with spatial overlap relationship, and generates the valve number set of the control zone.

[0089] Based on the path evapotranspiration density ranking table, the plot numbers corresponding to the top-ranked (e.g., top 30%) path segments are selected sequentially. The module calls the coordinate set of control elements in the canal system record, which contains the precise locations of all electric valves and gates. The module performs spatial overlap analysis to determine whether the plot range of each high-density path segment overlaps with the effective control range of the control element (usually a polygonal buffer zone around the element coordinates). Specifically, the module uses the ray casting method to determine whether the valve coordinate point is located inside the polygonal boundary of the plot. If the determination point is in-plane, or the distance between the two boundaries is less than the preset construction error range (e.g., 5 meters), then a spatial overlap relationship is confirmed. The module filters out all element numbers that meet the conditions and generates a set of valve numbers for the control zone. A specific example is as follows: As shown in Table 3, the top-ranked plot ID_206 is selected, and its boundary is defined as a series of vertex coordinates. The coordinates of the control element Gate_E03 are (1200, 2100). Spatial overlap filtering steps: Step 1: Construct a polygonal region model of plot ID_206. Step 2: Input the coordinates (1200, 2100) of valve Gate_E03 into a point-to-surface relationship judgment algorithm (e.g., ray casting). Step 3: The algorithm performs the calculation. If the result is "True", it means the valve coordinates are inside the polygon. Step 4: If the result of Step 3 is "False", calculate the vertical distance from the valve point to the nearest boundary of the polygon. Step 5: If the distance is 3 meters, compare it with the preset error threshold of 5 meters. Since 3 is less than 5, it is still considered an effective overlap. Step 6: Add Gate_E03 to the control set. If the coordinates of another valve, Gate_X09, are outside the plot and the distance exceeds 5 meters, it will not be included. This step precisely identifies the specific hardware device requiring physical operation, transforming the digital analysis results into actual engineering control commands.

[0090] Table 3. Results of Control Component Screening and Matching

[0091]

[0092] The map output submodule calls the flow-evaporation correlation map and the valve number set of the control zone. Based on the number mapping relationship, it constructs a graphical data structure including path structure, control unit location and control number to generate a smart management and control map of the entire irrigation area.

[0093] The module comprehensively utilizes the flow-evapotranspiration correlation map and the valve number set for the controlled zones. Centered on the number mapping relationship, it constructs a hierarchical graphical data structure. The first layer is the geographic background layer, loading satellite imagery or vector maps of the irrigation area; the second layer is the path structure layer, drawing the extracted high evapotranspiration disturbance paths and thermal distribution; the third layer is the control unit layer, highlighting the selected valve icons at corresponding coordinate locations and labeling their control parameters. The module merges these three layers of data, adds metadata (such as generation time and disturbance level), and renders an interactive intelligent management map of the entire irrigation area. This map allows users to click on valve icons to view details of their associated downstream disturbed plots and recommended valve opening adjustment strategies. A specific example is described below: Data structure construction steps: Step 1: Create a JSON root object and define an array named "MapLayers". Step 2: Construct the "ControlLayer" node object and set the attribute key-value pairs: {"ValveID": "Valve_007", "Status": "Critical", "LinkedPatch": "ID_B05", "Coordinates": [1200, 2100]}. Step 3: The rendering engine reads the data from Step 2 and draws a flashing red valve symbol at map coordinates (1200, 2100). Step 4: Draw guide lines. Draw a dashed line connecting the valve coordinates (1200, 2100) to the center point of plot ID_B05. The final generated map file is 5 megabytes in size and has a resolution of 4000 x 3000 pixels, clearly displaying a panoramic view of the "disturbance source - transmission path - control point," directly assisting managers in making accurate decisions.

[0094] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of protection of the technical solution.

Claims

1. A smart management and control platform for the entire irrigation district based on digital twins, characterized in that, The system includes: The disturbance structure construction module collects the water level time series from the headwater level sensor, the inflow and outflow times series from the boundary flow meter, and the pressure difference change images from the upstream and downstream pressure sensors. It calculates the vertical fluctuation of water level, the boundary frequency change value, and the number of water potential reversals, and combines them to generate a three-dimensional disturbance feature matrix. The rhythm segmentation module reads the disturbance time series in the three-dimensional disturbance feature matrix, performs amplitude comparison on the number and spacing of fluctuation peaks, divides the disturbance into dense and sparse areas, numbers and identifies high-disturbance areas and low-disturbance areas, and generates a time-series disturbance zoning map of the irrigation area. The evapotranspiration layer generation module obtains the irrigation area time-series disturbance zoning map, extracts the evapotranspiration record sequence, calculates the evapotranspiration difference curve for each pair of plots, filters areas with frequent difference reversals, and constructs an evapotranspiration difference thermal layer. The fluctuation mapping module extracts the path number with the maximum evaporation intensity from the evaporation difference thermal layer, obtains the corresponding water flow node number in the canal system map, compares the connectivity of the numbers, filters the connectivity combinations, and generates a flow-evaporation correlation map. The regulation map generation module reads the numbered paths in the flow-evapotranspiration correlation map, extracts the evapotranspiration density ranking, compares the positions of regulation elements in the canal system records, filters the control valve numbers of the fields to be controlled, and constructs a smart management and control map of the entire irrigation area.

2. The intelligent management and control platform system for the entire irrigation district based on digital twins as described in claim 1, characterized in that: The three-dimensional disturbance feature matrix includes water level vertical fluctuation index, boundary frequency change index, and water potential reversal frequency index. The irrigation district time-series disturbance zoning map includes high disturbance area number, low disturbance area number, and disturbance interval type label. The evapotranspiration difference thermal layer includes evapotranspiration difference amplitude distribution, evapotranspiration difference reversal frequency, and evapotranspiration dynamic distribution characteristics. The flow-evapotranspiration correlation map includes evapotranspiration path plot number, water flow change node number, and spatial connectivity matching relationship. The irrigation district-wide intelligent management and control map includes evapotranspiration density ranking sequence, control element location information, and field control valve number set.

3. The intelligent management and control platform system for the entire irrigation district based on digital twins as described in claim 1, characterized in that: The dense disturbance region refers to the area in the disturbance time series where there are many fluctuation peaks with small intervals, while the sparse disturbance region refers to the area where there are few fluctuation peaks with large intervals.

4. The intelligent management and control platform system for the entire irrigation district based on digital twins as described in claim 1, characterized in that: The frequently inverted difference region refers to the spatiotemporal region in the evaporation difference curve where positive and negative changes are frequent and the direction switching rate is high.

5. The intelligent management and control platform system for the entire irrigation district based on digital twins as described in claim 1, characterized in that, The disturbance structure construction module includes: The data receiving submodule collects the water level time series from the headwater level sensor, the inflow and outflow sequence from the boundary flow meter, and the pressure difference trend image from the upstream and downstream pressure sensors. It then performs time alignment and missing data imputation to generate the boundary disturbance input dataset. The disturbance extraction submodule calculates the adjacent differences in water level time series based on the boundary disturbance input dataset to obtain the vertical fluctuation of water level, counts the inflow and outflow frequency per unit time to obtain the boundary frequency change value, extracts the slope change symbol sequence count in the pressure difference image to obtain the number of water potential reversals, and generates a disturbance response index set. The feature matrix generation submodule calls the vertical fluctuation of water level, the change of boundary frequency and the number of water potential reversals in the disturbance response index set, and combines them in dimensional order to fill the axes and generate a three-dimensional disturbance feature matrix.

6. The intelligent management and control platform system for the entire irrigation district based on digital twins as described in claim 1, characterized in that, The rhythm segmentation module includes: The time series extraction submodule reads the numerical sequence of the perturbation index along the time axis in the three-dimensional perturbation feature matrix, establishes a unified index structure according to the time step, integrates them into a time series data group, and generates a time series set of perturbation index. The disturbance interval identification submodule extracts the location of local extreme points of continuous fluctuations in each sequence based on the disturbance index time series set, calculates the amplitude difference and time interval between adjacent extreme points, counts the fluctuation density according to a preset time window, assigns high-density windows as high-disturbance areas and low-density windows as low-disturbance areas, and generates a disturbance area classification sequence. The time-series partitioning output submodule calls the perturbation area classification sequence, maps it to a two-dimensional region structure according to the time index, constructs a corresponding numbered matrix layer, and generates a time-series perturbation partitioning map of the irrigation area.

7. The intelligent management and control platform system for the entire irrigation district based on digital twins as described in claim 1, characterized in that, The evaporation layer generation module includes: The area extraction submodule obtains the plot numbers marked as high-disturbance areas in the irrigation district time-series disturbance zoning map, retrieves the continuous evapotranspiration record sequence of the corresponding plot, aligns the sequence according to the time index and establishes a mapping relationship to generate an evapotranspiration sequence mapping set. The difference curve generation submodule performs time synchronization difference operation on any two evaporation record sequences based on the evaporation sequence mapping set, calculates the evaporation difference between adjacent time steps, detects the number of reversals of the difference sign in continuous segments and filters segments according to the reversal density, and generates evaporation difference reversal interval groups. The heat map output submodule calls the evapotranspiration difference inversion interval group, establishes a two-dimensional raster structure according to the segment position and inversion density, assigns numerical weights to the raster and forms a thermal distribution layer, and generates an evapotranspiration difference thermal layer.

8. The intelligent management and control platform system for the entire irrigation district based on digital twins as described in claim 1, characterized in that, The fluctuation mapping module includes: The path extraction submodule obtains the evapotranspiration intensity value of the grid area in the evapotranspiration difference thermal layer, traverses the continuous grid path along the direction of the maximum intensity value, extracts the plot number sequence on the path and stores it in order, and generates the evapotranspiration main path number group. The node association extraction submodule calls the main evapotranspiration path number group, finds the water flow change record position corresponding to the path according to the preset canal system map, extracts the canal system control node identifier corresponding to the number, records the coordinate information, and generates a path association node number set. The connectivity matching output submodule, based on the main evapotranspiration path number group and the path-associated node number set, calls the two-dimensional coordinates of the plots and nodes respectively, performs Euclidean distance determination and filters number pairs with a positional distance less than a threshold, establishes a mapping relationship between the numbers, and generates a flow-evapotranspiration correlation map.

9. The intelligent management and control platform system for the entire irrigation district based on digital twins as described in claim 1, characterized in that, The regulatory map generation module includes: The path number parsing submodule reads the number combination in the flow-evapotranspiration correlation map, extracts the set of plot numbers corresponding to the intersecting paths, retrieves the evapotranspiration density sequence of the path segment in the evapotranspiration difference thermal layer, sorts them in descending order of density value, and generates a path evapotranspiration density ranking table. The control element screening submodule locates the plot number corresponding to the top-ranked path segment according to the path evapotranspiration density ranking table, calls the control element coordinate set recorded in the canal system record, compares the position of each path segment with the position of the control element according to the spatial overlap principle, filters the element numbers with spatial overlap relationship, and generates the valve number set of the control zone. The map output submodule calls the flow-evaporation correlation map and the set of valve numbers for the control zone. Based on the number mapping relationship, it constructs a graphical data structure including path structure, control unit location and control number to generate a smart management and control map of the entire irrigation area.

10. The intelligent management and control platform system for the entire irrigation district based on digital twins as described in claim 9, characterized in that: The principle of spatial overlap means that when the geographical location of a path segment coincides with or intersects with the coordinate area of ​​the control element, it is considered that there is a spatial relationship.

Images