Planting structure driven irrigation district water resource dynamic allocation method and system

Abstract

The application discloses a kind of planting structure driven irrigation water resource dynamic allocation method and system.Irrigation planting structure and crop growth stage are identified based on multi-source remote sensing image and crop phenology information, agricultural irrigation water requirement is predicted by combining weather forecast, crop leaf area index and real-time soil moisture content;Multi-objective optimization allocation model is established with the minimum comprehensive water shortage penalty, the minimum energy consumption of water supply system and the maximum comprehensive economic benefit as the goal, and the water deficit in the scheduling period is mapped to the yield economic loss of the whole growth period according to the crop stage water sensitive index;Solve by stage according to industry type level, water use unit level and time sequence daily level, trigger agricultural centralized water supply according to root zone soil moisture content daily, and when channel capacity is limited, priority is given to crops in critical period of water.The method can improve the real-time and engineering executability of water distribution decision, realize water allocation according to planting, key period priority and low energy consumption allocation.

Description

Technical Field

[0001] This application relates to the field of irrigation district water resource optimization and scheduling technology, specifically to a planting structure-driven method and system for dynamic allocation of irrigation district water resources. Background Technology

[0002] Irrigation districts typically supply water for domestic, industrial, and agricultural purposes simultaneously, and the available water supply, as well as the water conveyance capacity of canals and pipelines, are all constrained. When the available water supply or water conveyance capacity cannot meet the various water demands, it is necessary to ensure basic water supply while also considering operational energy consumption and water use efficiency, and to rationally determine the water allocation and supply sequence for each water-using unit.

[0003] The current irrigation districts rely heavily on statistical planting area, experience-based irrigation quotas, or fixed crop coefficients for water allocation. This makes it difficult to obtain timely information on the dynamic changes in plot-level planting structure, crop growth stage, growth status, and soil moisture. Consequently, there are discrepancies between agricultural water demand forecasts and actual water demand, and water allocation plans are also unable to provide differentiated protection for crops in critical water periods.

[0004] Furthermore, the irrigation scheduling period is typically shorter than the entire crop growth period, making it difficult for existing methods to link water deficits during the scheduling period with economic losses in yield over the entire growth period. The daily coordinated allocation of water resources across multiple water-using units (domestic, industrial, and agricultural) also presents a high-dimensional nonlinear optimization problem, making it difficult to simultaneously address real-time solutions and engineering constraints such as canal rotation irrigation. Therefore, a dynamic water resource allocation method that integrates planting structure identification, dynamic water demand prediction, stage sensitivity assessment, and engineered water supply and drainage scheduling is needed. Summary of the Invention

[0005] To address the aforementioned issues, this application provides a planting structure-driven method and system for dynamic allocation of irrigation water resources. This method combines planting structure and crop growth stage identification, dynamic water demand prediction, stage-sensitive water shortage loss calculation, three-level dimensionality reduction and phased solution, and trigger-based centralized water supply. This connects short-term water allocation decisions with the yield benefits of the entire crop growth period, while also taking into account water supply security, operational energy consumption, economic benefits, and engineering feasibility.

[0006] The first aspect of this application provides a method for dynamic allocation of irrigation water resources driven by planting structure, including: Based on multi-source remote sensing images, the planting structure of the irrigation area and the current growth stage of crops in each field are identified. Based on the planting structure and crop growth stage, the irrigation water demand of agricultural irrigation areas during the future scheduling period is predicted, and the planned water demand reported by water plants and factories is obtained. A multi-objective optimization allocation model is constructed, with the objectives of minimizing the comprehensive water shortage penalty, minimizing the energy consumption of the water supply system, and maximizing the comprehensive economic benefits. Specifically, based on the crop stage water sensitivity index determined by the planting structure and crop growth stage, the water deficit of agricultural irrigation areas during the scheduling period is mapped to the yield and economic loss of the entire crop growth period. The multi-objective optimization allocation model is solved in stages according to industry type, water use unit, and daily time series. The decision variables obtained from the previous level are used as water quantity constraints for the next level. The total water allocation and daily water allocation plan of each water use unit during the scheduling period are determined. The water use unit includes the water plant, the factory, and the agricultural irrigation area. Among them, agricultural water use is triggered by the daily projection of soil moisture content in the root zone; when the total water demand of multiple sub-irrigation areas in the agricultural irrigation area that trigger water supply on the same day exceeds the maximum water conveyance capacity of the main canal, the water supply priority is determined according to the crop stage water sensitivity index.

[0007] Optionally, the step of identifying the planting structure of the irrigation area and the current growth stage of crops in each field based on multi-source remote sensing imagery includes: Multi-temporal optical remote sensing images and synthetic aperture radar images of the irrigation area are acquired. Spectral and texture features are extracted from the optical remote sensing images, and backscattering features are extracted from the synthetic aperture radar images to construct a multi-source heterogeneous feature dataset. The vegetation index is calculated by combining the crop phenological calendar of the irrigation area and a vegetation index time series curve is constructed. The multi-source heterogeneous feature dataset is fused with the vegetation index time series curve and then input into a deep learning classification model for classification to obtain the initial planting structure and the prediction confidence of each field. A confidence threshold is set, and fields with a prediction confidence level lower than the threshold are selected and a manual verification sample set is generated. The deep learning classification model is iteratively optimized based on the labeled manual verification sample set to output the planting structure of the irrigation area. At the same time, based on the inflection point and peak characteristics of the vegetation index time series curve and combined with the local crop phenological calendar, the current growth stage of crops in each field is identified as the basis for subsequent water demand calculation and water sensitivity index determination.

[0008] Optionally, the prediction of irrigation water demand for agricultural irrigation areas during the future scheduling period includes: Based on the planting structure, crop growth stage, and forecast meteorological data during the scheduling period, the theoretical baseline water requirement of crops in each field at different growth stages is calculated using the Penman-Montes formula and crop coefficients determined by looking up tables based on crop type and growth stage. Obtain the crop leaf area index and real-time soil moisture content, input the theoretical baseline water requirement, the crop leaf area index and the real-time soil moisture content into the neural network model, and output the actual net irrigation water requirement of crops in each field. The water demand for each field is obtained by subtracting the predicted effective rainfall and the capillary recharge of groundwater from the actual net irrigation water demand. The water demand for each agricultural irrigation area is then obtained by summarizing the data level by level according to the water supply network topology.

[0009] Optionally, the multi-objective optimization allocation model includes the following objective function: The objective function that minimizes the overall water shortage penalty is: In the formula, As a comprehensive water shortage penalty; , , These represent the number of water plants, factories, and agricultural irrigation areas, respectively. , The first The planned water demand and actual water distribution of each water plant; , The first Planned water demand and actual water distribution for each factory; , These are the economic loss coefficients per unit of water shortage for each water plant and each factory, respectively. , , These are the industry-specific water shortage penalty coefficients for domestic, industrial, and agricultural sectors, respectively. Water shortage penalties for agricultural irrigation areas; The objective function for minimizing the energy consumption of the water supply system is: In the formula, Energy consumption of the water supply system; The total number of pumping stations and booster stations supplying water to various water plants, factories, and agricultural irrigation areas; For the first The water pumping flow rate of each pumping station; To lift water; To improve the overall efficiency of the pump station units; The density of water; It is the acceleration due to gravity; For the first p The operating time of each pumping station during the scheduling period; The objective function that maximizes overall economic benefits is: In the formula, For comprehensive economic benefits; For the first Industrial output value coefficient per unit water volume of a factory; The conversion factor for converting soil water storage into economic benefits in the next period; For the end of the scheduling period The increase in soil water storage in the root zone of an agricultural irrigation area, measured in terms of water volume.

[0010] Optionally, in the objective function that minimizes the overall water shortage penalty, the water shortage penalty of agricultural irrigation plots is included. The calculation formula is: In the formula, For the first Unit price of agricultural irrigation crops; For the first The maximum yield per unit area of ​​an agricultural irrigation crop; For the first The planting area of ​​each agricultural irrigation area; and The first The first agricultural irrigation crop Daily actual evaporation and maximum potential evaporation; For the first The daily water sensitivity index of crops within a specific agricultural irrigation area during the specified scheduling period is taken as the [number]th [item]. The stage water sensitivity index of a crop's growth stage is calculated by dividing the total number of days in that growth stage by the stage water sensitivity index of that growth stage. This stage water sensitivity index reflects the crop's sensitivity to water deficiency at a specific growth stage. The moisture sensitivity index for the aforementioned stage, and the critical period for moisture. Large value, non-sensitive period The value is small; T is the number of days included in the scheduling period.

[0011] Optionally, the constraints of the multi-objective optimization allocation model include water source constraints, water user balance constraints, minimum guarantee rate constraints for water users, water transmission capacity constraints of agricultural water supply pipelines, water transmission capacity constraints of domestic and industrial water supply networks, and water intake capacity constraints of agricultural water intakes. When the water source is a surface reservoir, the water volume constraint on the water source side is: In the formula, , These represent the initial and final reservoir capacities, respectively. This represents the total inflow into the reservoir. This refers to the total water supply from the reservoir used for power generation. , , These represent the total water supply for domestic, industrial, and agricultural use, respectively. This refers to the total flood discharge from the reservoir. The amount of water lost through evaporation and seepage from the reservoir; and the reservoir capacity at the end of the period shall not be lower than the dead storage capacity and not higher than the usable storage capacity; , For a given value; The , The parameters are provided by the reservoir dispatching department before the dispatching period according to the power generation plan and flood control requirements, and are used as input parameters for this model, but are not involved in the optimization of this model.

[0012] When the water source is a surface river or groundwater, the water volume constraint on the water source side is: In the formula, The total available water volume is the amount of water allocated during the scheduling period, and is not greater than the sum of the planned water demand of each water-using unit. The water supply capacity constraint of the agricultural water supply pipeline is: In the formula, Number the water outlets on each main and branch canal from bottom to top; For the first The water flow rate at each branch point; Each main and branch canal is numbered from bottom to top, corresponding to the water distribution point, with each main and branch canal numbered starting from 1; For the first The maximum flow rate corresponding to the channel segment; The water intake capacity constraint of the agricultural water intake is: In the formula, For the first The maximum flow rate of each branch point; The water balance constraint for each water user is: the actual water allocation of each water user during the scheduling period shall not exceed its planned water demand. The minimum guarantee rate constraint for water users is: the actual water allocation of each water user unit shall not be less than the product of its planned water demand and the corresponding guarantee rate; The water supply capacity constraint of the domestic and industrial water supply network is: the sum of the water supply flow of the downstream nodes of each pipe section in the domestic and industrial water supply network shall not exceed the maximum design flow of that pipe section.

[0013] Optionally, the phased solution includes: With the total available water supply as a constraint, and the total water allocation for domestic, industrial, and agricultural sectors as decision variables at the industry type level, the multi-objective optimization allocation model is solved using a multi-objective optimization algorithm to obtain a Pareto optimal solution set. The total water allocation for domestic, industrial, and agricultural sectors during the scheduling period is then determined from the Pareto optimal solution set according to preset decision rules. The total water allocation for each sector serves as the water quantity constraint at the next level of water-using unit. Under the constraint of the total water allocation for each industry, the total water allocation for each water plant, each factory, and each agricultural irrigation area is used as the decision variable to allocate water to each water-using unit. The total water allocation for each agricultural irrigation area is further allocated to each sub-irrigation area under its jurisdiction to obtain the total water allocation for each sub-irrigation area. The total water allocation for each water-using unit and the total water allocation for each sub-irrigation area serve as the next level of time-series daily water allocation constraints. Among them, agricultural water use is prioritized to ensure the sub-irrigation areas in the critical water period based on the crop stage water sensitivity index, while also meeting the minimum guarantee rate constraint for each water-using unit. Under the constraints of the total water allocation of each water-using unit and the total water allocation of each sub-irrigation area, the daily water allocation of each water-using unit is used as the decision variable. Domestic water and industrial water are continuously allocated daily according to the daily water demand ratio. Agricultural water is allocated using a trigger-based centralized water supply schedule. The water transmission capacity constraint is checked daily to generate the daily water allocation plan for each water-using unit and the centralized agricultural water supply schedule. The total available water supply constraint remains consistent across the three levels, the minimum guarantee rate constraint remains consistent at the water-using unit level and the time-series daily level, and the water transmission capacity constraint is checked daily at the time-series daily level.

[0014] Optionally, the trigger-based centralized water supply scheduling for agricultural water use includes: Using the measured soil moisture content in the root zone as the initial state, and combining the daily crop water requirement and daily effective rainfall, the change process of soil moisture content in the root zone of each agricultural sub-irrigation area is deduced day by day; When the soil moisture content in the root zone of a sub-irrigation patch drops to the lower limit threshold of suitable soil moisture content, a water supply command for that sub-irrigation patch is triggered, and that day is designated as the water supply time for that sub-irrigation patch. At the water supply time, the net irrigation amount required to restore the soil moisture content in the root zone to the upper limit of the suitable soil moisture content and the net irrigation amount required to meet the remaining water demand of this growth stage are calculated. The smaller of the two values ​​is taken, and the one-time concentrated gross water supply of the sub-irrigation area is calculated in combination with the channel water conveyance loss. The sum of the one-time concentrated gross water supply of the sub-irrigation area during the scheduling period shall not exceed the total water distribution of the sub-irrigation area. When multiple sub-irrigation areas trigger water supply on the same day and the total water demand exceeds the maximum water conveyance capacity of the main canal, priority is given to the sub-irrigation areas with higher water sensitivity indices based on the crop stage water sensitivity index, and the remaining sub-irrigation areas are supplied water in turn.

[0015] Optionally, the method further includes converting the planting structure and crop growth stage into model parameters, specifically including: Establish a table of correspondence between crop type, growth stage and crop stage water sensitivity index. Determine the crop stage water sensitivity index of each sub-irrigation area by referring to the table based on the identified crop type and current growth stage. When the scheduling period spans multiple growth stages, the crop stage water sensitivity index of each stage is weighted and summed according to the proportion of days occupied by each growth stage in the scheduling period to obtain the equivalent water sensitivity index of the sub-irrigation area in the scheduling period.

[0016] The crop stage water sensitivity index or the equivalent water sensitivity index is used as a parameter of the allocation model and input into the multi-objective optimization allocation model so that the water allocation priority is dynamically adjusted according to the changes in the identified crop type and growth stage.

[0017] The second aspect of this application provides a planting structure-driven dynamic water resource allocation system for irrigation districts, comprising: The planting structure identification module is used to identify the planting structure of irrigation areas and the current growth stage of crops in each field based on multi-source remote sensing images. The water demand prediction module is used to predict the irrigation water demand of agricultural irrigation areas during the future scheduling period based on the planting structure and crop growth stage, and to obtain the planned water demand reported by water plants and factories. The allocation model module is used to establish a multi-objective optimization allocation model with the goals of minimizing the comprehensive water shortage penalty, minimizing the energy consumption of the water supply system, and maximizing the comprehensive economic benefits. Based on the crop stage water sensitivity index determined by the planting structure and crop growth stage, the water deficit of agricultural irrigation plots during the scheduling period is mapped to the yield and economic loss of the entire crop growth period. The allocation and solution module is used to solve the multi-objective optimization allocation model in stages according to industry type level, water use unit level, and time series daily level. The decision variables obtained from the solution of the previous layer are used as water quantity constraints for the next layer to generate daily water allocation plans for each water use unit. Among them, agricultural water use triggers centralized water supply based on the daily extrapolation results of root zone soil moisture content. When the total water demand of multiple sub-irrigation areas in the agricultural irrigation area that triggers water supply on the same day exceeds the maximum water conveyance capacity of the main canal, the water supply priority is determined according to the crop stage water sensitivity index.

[0018] The first aspect of this application uses the planting structure and crop growth stage of the irrigation area as the information source for allocation. Based on the crop stage water sensitivity index, the water deficit during the scheduling period is converted into economic loss of yield throughout the entire growth period. It also performs synergistic optimization among water supply security, water supply energy consumption and economic benefits, so that the limited water volume is prioritized for water-using units with high water shortage costs and agricultural irrigation areas in the critical water period.

[0019] Optionally, the fusion of optical remote sensing, synthetic aperture radar imagery, phenological time series, and active learning can improve the stability of planting structure identification in complex plots; crop leaf area index and real-time soil moisture content are used to correct theoretical water demand, so that water demand prediction is adapted to crop growth and soil moisture; phased solutions at the industry type level, water use unit level, and time series daily level can reduce the solution complexity of high-dimensional models; centralized water supply and channel capacity verification based on soil moisture content thresholds can connect the allocation results with irrigation district rotation irrigation and gate control methods.

[0020] Optionally, this application directly converts the crop type, planting area, and current growth stage obtained from the planting structure identification into allocation model parameters: the crop type and growth stage are used to determine the water sensitivity index of the crop stage by looking up the table. When the scheduling period spans multiple growth stages, the equivalent sensitivity index is obtained by weighting the proportion of days in each stage, so that the water allocation priority is dynamically adjusted according to the identification results, thereby realizing the closed-loop drive of the identification results for water allocation decision, rather than just using the identification results as a static statistical output.

[0021] Optionally, the three-level dimensionality reduction phased solution of this application—industry-level, unit-level, and daily-level—has clear inter-level variable transfer and constraint consistency: the industry type level uses the total water allocation of the industry as the decision variable; the water-using unit level uses the total water allocation of each water-using unit as the decision variable and the total amount of the higher-level industry as the constraint; and the time-series daily level uses the daily water allocation of each water-using unit as the decision variable and the total amount of the higher-level unit as the constraint. The water source constraint remains consistent across the three levels, the minimum guarantee rate constraint remains consistent in the latter two levels, and the water transmission capacity constraint is checked daily in the final level. After the three-level dimensionality reduction, the number of decision variables is reduced, significantly reducing the solution complexity, and it differs from ordinary step-by-step calculations because the result of each level serves as a hard constraint for the next level.

[0022] Optionally, this application establishes a continuous control chain for soil moisture content estimation, water supply timing triggering, centralized water supply calculation, channel capacity verification, water sensitivity index ranking, and gate rotation irrigation and drainage schedule. This control chain is then embedded into a three-level time-series daily micro-allocation sub-problem that is solved in stages with reduced dimensionality: the total water volume limit for the trigger-based drainage schedule is constrained by the meso-level allocation results at the water-using unit level; the water sensitivity index used for water supply priority ranking originates from the same parameter as the sensitivity index used for water shortage loss assessment and water-using unit level allocation; and the engineering parameters used for channel capacity verification are the same parameter as the water conveyance capacity constraint of the allocation model. This creates a coupled combination feature between the three-level results of trigger-based centralized water supply and multi-objective optimization.

[0023] It is understandable that the beneficial effects of the second aspect mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A schematic flowchart of the planting structure-driven dynamic allocation method for irrigation water resources in an embodiment of this application; Figure 2 This is a logical diagram illustrating the three-level dimensionality reduction phased solution strategy adopted in the embodiments of this application. Detailed Implementation

[0026] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0027] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0028] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0029] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0030] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0031] The planting structure-driven dynamic allocation method for irrigation district water resources provided in this application can be applied to irrigation district water resource scheduling and management platforms, servers, industrial control computers, cloud computing nodes, and other computing devices with data processing capabilities in water conservancy information scheduling centers. It is used to dynamically optimize and allocate water resources in irrigation districts facing conflicts between domestic, industrial, and agricultural water use or limited water delivery capacity. Specifically, it can be executed by the processor of the computing device when running a computer program with corresponding functions. The irrigation district typically includes water sources such as reservoirs, rivers, or groundwater; water supply channels or pipelines such as main canals, branch canals, and pipelines; and water-using units such as water plants, factories, and agricultural irrigation areas. This application does not impose any limitations on the scale of the irrigation district, the type of water source, or the number of water-using units. The technical solution of this application will be further described in detail below with reference to the accompanying drawings and embodiments.

[0032] This application provides a planting structure-driven method for dynamic allocation of irrigation water resources, including the following steps S101 to S103: Step S101: Identify the planting structure of the irrigation area and the current growth stage of crops in each field based on multi-source remote sensing images, and predict the irrigation water demand of agricultural irrigation areas during the future scheduling period based on the planting structure and crop growth stage, and obtain the planned water demand reported by water plants and factories. Step S102: Construct a multi-objective optimization allocation model with the objectives of minimizing comprehensive water shortage penalty, minimizing water supply system energy consumption, and maximizing comprehensive economic benefits; and based on the crop stage water sensitivity index determined by the planting structure and crop growth stage, map the water deficit of agricultural irrigation areas during the scheduling period into the economic loss of crop yield throughout the entire growth period. Step S103: Solve the multi-objective optimization allocation model in stages according to industry type level, water use unit level, and time series daily level. The decision variables obtained from the solution of the previous level are used as water quantity constraints for the next level. Determine the total water allocation and daily water allocation plan for each water use unit during the scheduling period. Among them, agricultural water use is triggered by centralized water supply based on the daily extrapolation results of root zone soil moisture content. When the total water demand of multiple sub-irrigation areas in the agricultural irrigation area that triggers water supply on the same day exceeds the maximum water conveyance capacity of the main canal, the water supply priority is determined according to the crop stage water sensitivity index.

[0033] The water-using units include water plants, factories, and agricultural irrigation areas.

[0034] In application, the agricultural irrigation patches are divided according to crop type and irrigation system. Within the same agricultural irrigation patch, each sub-irrigation patch and each field contains the same crop type and grows at the same stage. Each agricultural irrigation patch includes one or more sub-irrigation patches, and each sub-irrigation patch includes one or more fields, thus forming a three-level spatial hierarchy of agricultural irrigation patch - sub-irrigation patch - field. The crop stage water sensitivity index is determined by the crop type and the current growth stage of the crop. The crop stage water sensitivity index of each field, each sub-irrigation patch, and each agricultural irrigation patch is the water sensitivity index of the crop at its current growth stage, and all three have the same value. At the sub-irrigation patch level, it is used for priority ranking between the sub-irrigation patches at the water use unit level and the daily time-series level. At the agricultural irrigation patch level, after being divided by the total number of days in the growth stage to convert to a daily water sensitivity index, it becomes the water shortage penalty calculation formula for the agricultural irrigation patch. .

[0035] In application, firstly, multi-source remote sensing images of irrigation areas and crop phenological information are used to identify the field-level planting structure and crop growth stage; then, forecasted weather, crop leaf area index and real-time soil moisture content are integrated to predict the water demand of the field, and the water demand of each water-using unit is summarized step by step according to the water supply network topology; then, the water demand prediction results, water supply network topology and engineering capacity parameters are input into the multi-objective optimization and allocation model to generate the total water allocation, daily water allocation plan and agricultural centralized water supply schedule.

[0036] In this application, this embodiment further elucidates the physical topology and hydraulic correspondence between irrigation area water sources, water supply channels or pipelines, and water-using units. It determines the physical connections between water sources, main canal systems, branch canals, water outlets, and water-using units, constructs a directed graph topology network, and clarifies the design flow capacity of each pipe or canal segment, the maximum diversion capacity of each water outlet, and the head and flow-efficiency curves of each pumping station or booster station. This provides fundamental parameters for subsequent hydraulic constraints. The water-using units include water plants, factories, and agricultural irrigation areas within the irrigation area. Based on this, for areas with water supply conflicts, it coordinates domestic, industrial, and agricultural water demands, constructing a multi-objective optimization allocation model aimed at minimizing comprehensive water shortage penalties, minimizing water supply system energy consumption, and maximizing comprehensive economic benefits. It also determines constraints such as water quantity constraints on the water source side, water conveyance capacity constraints on agricultural water supply pipelines, and water intake capacity constraints on agricultural water inlets. Specifically, for agricultural water use, water shortages during the allocation period are mapped to yield economic losses throughout the entire growth period based on crop stage water sensitivity indices, ensuring that short-term allocation decisions are aligned with the economic benefits of the entire crop growth period. When solving the multi-objective optimization allocation model, the high-dimensional allocation problem is decomposed into three cascaded sub-problems within a unified scheduling period: macro-allocation by industry type, meso-allocation by water-using unit, and micro-allocation by time series. These sub-problems are solved sequentially to generate water allocation plans for each water-using unit within the scheduling period and agricultural centralized water supply schedules. The scheduling period can be set according to the actual management needs of the irrigation district. In a specific application, the scheduling period is taken as a unified scheduling period of the next fifteen days.

[0037] In application, Figure 1 This embodiment illustrates the overall process of the planting structure-driven dynamic allocation method for irrigation water resources in an irrigation district, including planting structure and growth stage identification, dynamic correction of irrigation water demand based on neural networks, cross-stage equivalent transformation of water sensitivity index, water shortage loss mapping and multi-objective allocation model construction, industry-unit-day three-level dimensionality reduction and phased solution, and triggered centralized water supply and rotational irrigation and drainage. Specifically, planting structure and growth stage identification and dynamic correction of irrigation water demand based on neural networks correspond to step S101, which is specifically implemented by steps S201-S203 and S301-S303, respectively. Cross-stage equivalent transformation of water sensitivity index corresponds to steps S601-S603, water shortage loss mapping and multi-objective allocation model construction corresponds to step S102, and industry-unit-day three-level dimensionality reduction and phased solution and triggered centralized water supply and rotational irrigation and drainage correspond to step S103, which is specifically implemented by steps S401-S403 and S501-S504, respectively. The results of planting structure and growth stage identification are used to determine agricultural water demand and stage water sensitivity index; water demand prediction results and water supply network constraints are jointly fed into the allocation model; the final output of the daily water allocation plan and agricultural centralized water supply schedule for each water-using unit is solved in stages.

[0038] This embodiment integrates planting information throughout the entire process of water demand forecasting, water shortage loss assessment, and water allocation decision-making. It differentiates the water shortage costs for different crops and different growth stages using a stage-based water sensitivity index, coordinates water supply pathways and water use benefits through water supply energy consumption and economic benefit targets, and transforms the optimization results into executable rotational irrigation and drainage schedules through staged solutions and triggered centralized water supply.

[0039] In one embodiment, the step of identifying the planting structure of the irrigation area and the current growth stage of crops in each field based on multi-source remote sensing images includes the following steps S201 to S203: Step S201: Acquire multi-temporal optical remote sensing images and synthetic aperture radar images of the irrigation area, extract spectral and texture features from the optical remote sensing images, extract backscattering features from the synthetic aperture radar images, and construct a multi-source heterogeneous feature dataset. Step S202: Calculate the vegetation index by combining the crop phenological calendar of the irrigation area and construct the vegetation index time series curve. After fusing the multi-source heterogeneous feature dataset with the vegetation index time series curve, input it into the deep learning classification model for classification to obtain the initial planting structure and the prediction confidence of each field. Step S203: Set a confidence threshold, filter out fields with a prediction confidence level lower than the confidence threshold and generate a manual verification sample set, iteratively optimize the deep learning classification model based on the labeled manual verification sample set, and output the planting structure of the irrigation area; at the same time, based on the inflection point and peak characteristics of the vegetation index time series curve and combined with the local crop phenological calendar, identify the current growth stage of crops in each field.

[0040] In application, the process of identifying planting structures in irrigation areas in this embodiment can be divided into three stages: multi-source remote sensing data acquisition and feature extraction, phenological time-series feature fusion and preliminary identification, and model iterative optimization based on active learning. In the multi-source remote sensing data acquisition and feature extraction stage, multi-temporal optical remote sensing images and synthetic aperture radar (SAR) images of the irrigation area are acquired. Spectral and textural features are extracted from the optical remote sensing images, and backscattering features are extracted from the SAR images. These features from different sources and with varying dimensions are then jointly organized into a multi-source heterogeneous feature dataset. Since optical remote sensing images are easily affected by cloud cover, while SAR images are not limited by cloud or rainy weather, the two complement each other at the feature level, jointly characterizing the land cover attributes of the fields, thus providing sufficient feature support for the accurate identification of complex plots and subdivided crops.

[0041] In the application, during the phenological time-series feature fusion and preliminary identification stage, vegetation indices are calculated by combining the local crop phenological calendar of the irrigation area, and a vegetation index time-series curve reflecting the crop growth rhythm is constructed. The multi-source heterogeneous feature dataset and the vegetation index time-series curve are fused at the feature level and then input into a deep learning classification model for classification to obtain the initial planting structure and the prediction confidence of each field. Specifically, the deep learning classification model can adopt the spatiotemporal Transformer model, which takes into account both the spatial features of remote sensing images and the temporal features of phenological curves, and can accurately depict the spatiotemporal change patterns of different crops during the growth process, thereby improving the ability to distinguish complex plots and subdivided crop categories. In the iterative optimization phase of the model based on active learning, a confidence threshold is set, and fields with a prediction confidence level lower than the threshold are selected as complex fields with higher recognition difficulty, and a manual verification sample set is generated. After manual on-site verification or high-resolution image annotation, the deep learning classification model is fine-tuned and iterated using the annotated manual verification sample set until the model converges, outputting the planting structure and crop growth stage of the irrigation area. The active learning strategy concentrates limited manual verification resources on fields with lower confidence levels, reducing the cost of manual verification while ensuring recognition accuracy.

[0042] This embodiment overcomes the limitation of stable imaging due to cloud cover in single optical remote sensing by fusing multi-temporal optical remote sensing images and synthetic aperture radar images, enabling planting structure identification to continue even in cloudy and rainy weather conditions. Spectral and texture features are extracted from optical remote sensing images, and backscattering features are extracted from synthetic aperture radar images to construct a multi-source heterogeneous feature dataset, enriching the feature representation of field features and improving the separability of complex plots and subdivided crops. Vegetation indices are calculated by combining crop phenological calendars and constructing vegetation index time-series curves. The spatial and phenological time-series features are then fused and input into a deep learning classification model, enabling the model to classify crops based on their growth rhythms, further improving the identification accuracy of subdivided crops. An active learning iterative optimization mechanism based on confidence threshold screening and manual verification sample labeling is introduced, concentrating manual verification resources on plots with higher identification difficulty. This improves identification accuracy while reducing manual verification costs, providing highly reliable basic data for water resource allocation.

[0043] In one embodiment, predicting the water demand of each water-using unit in the irrigation district during the future scheduling period includes predicting the irrigation water demand of each agricultural irrigation area, specifically including the following steps S301 to S303: Step S301: Based on the planting structure, crop growth stage and forecast meteorological data during the scheduling period, the theoretical baseline water requirement of crops in each field at different growth stages is calculated using the Penman-Montes formula and combined with the crop coefficient determined by looking up the table based on crop type and growth stage. Step S302: Obtain the crop leaf area index and real-time soil moisture content, and input the theoretical baseline water requirement, the crop leaf area index and the real-time soil moisture content into the neural network model to obtain the actual net irrigation water requirement of crops in each field. Step S303: Subtract the predicted effective rainfall and groundwater capillary recharge from the actual net irrigation water demand to obtain the field water demand, and summarize it level by level according to the water supply network topology to obtain the irrigation water demand of each agricultural irrigation area.

[0044] In application, the process of predicting irrigation water demand for each agricultural irrigation area in this embodiment can be divided into four stages: theoretical baseline water demand calculation, multi-source data acquisition from the air, ground, and air, dynamic correction of water demand based on neural networks, and effective replenishment deduction and hierarchical aggregation. In the theoretical baseline water demand calculation stage, based on the planting structure, crop growth stage, and forecast meteorological data during the scheduling period, the Penman-Montes formula is used to calculate the theoretical baseline water demand for different crops at different growth stages in each field, obtaining a crop water demand baseline reflecting the influence of meteorological conditions and crop coefficients. In the multi-source data acquisition stage, crop leaf area index retrieved from UAV multispectral imagery and real-time soil moisture content data collected by ground-based IoT soil moisture sensors are acquired. The UAV multispectral imagery characterizes the actual crop growth from an aerial perspective, while the IoT soil moisture sensors characterize the actual soil moisture from a ground perspective. These two, together with planting structure information identified by remote sensing, constitute an integrated air-ground observation data source.

[0045] In the application, during the dynamic correction of water demand based on neural networks, the theoretical baseline water demand, the crop leaf area index, and the real-time soil moisture content are input into the neural network model as input features, and the corrected actual net irrigation water demand for each field is output. Specifically, the neural network model can be a dynamic correction model of water demand constructed using a long short-term memory network. This model uses the theoretical baseline water demand as the baseline term and the leaf area index, which reflects the actual growth of the crop, and the real-time soil moisture content, which reflects the actual soil moisture, as correction terms. It learns the nonlinear mapping relationship between the three terms and the actual net irrigation water demand, thereby correcting the theoretical water demand obtained solely from meteorological data and empirical formulas to the actual net irrigation water demand that takes into account both crop growth and soil moisture, overcoming the problem of the disconnect between theoretical and actual water demand. In the effective supply deduction and hierarchical aggregation stage, the predicted effective rainfall and groundwater capillary recharge are deducted from the actual net irrigation water demand to obtain the final field water demand. Then, according to the topological relationship of the water supply channel or pipeline, the data is summed up level by level to first obtain the total irrigation water demand of each sub-irrigation area, and then further summed to obtain the irrigation water demand of each agricultural irrigation area, thus forming hierarchical water demand data consistent with the water supply network structure.

[0046] This embodiment uses the theoretical baseline water demand calculated by the Penman-Montes formula as the physical baseline, and dynamically corrects it using crop leaf area index and real-time soil moisture content. It also deducts effective rainfall and groundwater capillary recharge to avoid the repeated inclusion of natural recharge in irrigation water demand. The water demand is summarized step by step according to the water supply network topology to ensure that the spatial hierarchy of water demand is consistent with the structure of the water supply project.

[0047] In one embodiment, the multi-objective optimization allocation model includes the following objective function: The objective function that minimizes the overall water shortage penalty is: In the formula, As a comprehensive water shortage penalty; , , These represent the number of water plants, factories, and agricultural irrigation areas, respectively. , The first The planned water demand and actual water distribution of each water plant; , The first Planned water demand and actual water distribution for each factory; , These are the economic loss coefficients per unit of water shortage for each water plant and each factory, respectively. , , These are the industry-specific water shortage penalty coefficients for domestic, industrial, and agricultural sectors, respectively. Water shortage penalties for agricultural irrigation areas; The objective function for minimizing the energy consumption of the water supply system is: In the formula, Energy consumption of the water supply system; The total number of pumping stations and booster stations supplying water to various water plants, factories, and agricultural irrigation areas; For the first The water pumping flow rate of each pumping station; To lift water; To improve the overall efficiency of the pump station units; The density of water; It is the acceleration due to gravity; For the first p The operating time of each pumping station during the scheduling period; The objective function that maximizes overall economic benefits is: In the formula, For comprehensive economic benefits; For the first Industrial output value coefficient per unit water volume of a factory; The conversion factor for converting soil water storage into economic benefits in the next period; For the end of the scheduling period The increase in soil water storage in the root zone of an agricultural irrigation area, measured in terms of water volume.

[0048] The objective function described above is calculated by aggregating the decision variables at each solution level: at the industry type level, the water allocation of each water-using unit is calculated from the total water allocation of the industry according to a preset pre-allocation rule, and the water pumping flow of each pumping station is obtained by summarizing the pre-allocated water allocation along the water supply network topology. The pre-allocation rule is a component of the multi-objective optimization allocation model. At the water-using unit level and the time-series daily level, the objective function is calculated by aggregating the decision variables at the corresponding level.

[0049] In this application, this embodiment addresses the conflict of water competition among domestic, industrial, and agricultural sectors in the irrigation area by constructing a multi-objective collaborative water resource optimization and allocation model. Its objective function consists of three mutually synergistic objectives. With the goal of minimizing the overall water shortage penalty, we aim to ensure water supply security and food security. For domestic and industrial water use, the degree of water shortage is measured by multiplying the economic loss coefficient corresponding to a unit of water shortage by the amount of water shortage, and then multiplying by the industry-specific water shortage penalty coefficient for domestic and industrial use respectively. , The greater the water shortage, the greater the water shortage penalty, thereby reducing the water shortage in each water-using unit; for agricultural water use, the water shortage penalty is based on the agricultural irrigation area. Multiply by the agricultural sector's water shortage penalty coefficient This is measured by setting industry-wide water shortage penalty coefficients and unit water shortage economic loss coefficients for each water plant and factory, to achieve the water rights allocation principle of prioritizing domestic water use, considering the efficiency of industrial water use, and flexibly allocating agricultural water use. Objective function With the goal of minimizing the total energy consumption of the water supply system and reducing carbon emissions and operating costs, the energy consumed by each pumping station and booster station supplying water to various water plants, factories, and agricultural irrigation areas during the water lifting process is calculated based on the overall unit efficiency and accumulated over the operating time. This includes both water lifting energy consumption and water lifting flow rate. Water lifting and pumping Proportional to, and related to the overall efficiency of the pumping station unit This is inversely proportional, thus guiding the allocation scheme to prioritize the use of lower energy-consuming water supply paths while meeting water supply demand. Objective function With the goal of maximizing overall economic benefits, this approach integrates industrial output value and implicit agricultural income. The first factor is the industrial output value coefficient per unit of water used by each factory. Multiply by the actual water distribution volume The first term represents the explicit output value created by industrial water use, while the second term uses the soil water storage conversion factor. Multiply by the increase in soil water storage in the root zone of each agricultural irrigation area at the end of the scheduling period. This characterizes the implicit economic benefits of current water allocation in improving soil moisture.

[0050] In application, specifically for agricultural water use, this embodiment focuses on the objective function. The paper introduces a stage-based dimensionality reduction mapping mechanism based on crop stage water sensitivity index to reduce water shortage penalties in agricultural irrigation areas. The model expresses water sensitivity through a crop stage index, mapping short-term water shortages during the scheduling period to economic losses in yield across the entire growth cycle. Since the model's allocation period is a relatively short future scheduling period, and agricultural economic benefits depend on the entire crop growth cycle, the model uses a dynamic water sensitivity index to capture the crop's current growth stage: when the scheduling period covers the crop's critical water period, the model automatically amplifies the water shortage penalty, prioritizing water use to maintain the final yield; when the crop is in a non-sensitive stage, the model allows for moderate deficit irrigation, allocating water resources to water-using units with higher marginal economic benefits. Simultaneously, the objective function... Soil water storage carryover project introduced in China The improvement effect of water allocation during the current scheduling period on soil moisture is quantified into implicit economic benefits, so as to achieve seamless connection between short-term rolling optimization and long-term economic benefits throughout the entire growth period, and ensure the scientific nature and consistency of the water allocation strategy based on planting in the time dimension.

[0051] In application, this embodiment integrates the three objectives of minimizing comprehensive water shortage penalty, minimizing water supply system energy consumption, and maximizing comprehensive economic benefits into a unified optimization framework. It constructs a multi-objective collaborative optimization model covering domestic, industrial, and agricultural aspects, comprehensively considering pumping station energy consumption and comprehensive economic benefits while ensuring water supply security, thus achieving synergistic optimization of the comprehensive benefits of water resources, energy, and food. For domestic and industrial water shortages, a linear measurement method is adopted, multiplying the unit water shortage economic loss coefficient by the water shortage amount and weighting them separately. This ensures that the water shortage penalty increases linearly with the increase in water shortage amount, thereby prioritizing domestic water use. The energy consumption of the water supply system is also considered... Expressed as a value directly proportional to the pumping flow rate and pumping head, and inversely proportional to the overall efficiency of the generating unit, this guides the allocation scheme to prioritize low-energy water supply routes to reduce operating costs and carbon emissions. The water shortage penalty item incorporates a crop stage water sensitivity index, and the economic benefit item incorporates a soil water storage carryover item. This maps short-term water shortages to economic losses throughout the entire growth period and converts the improvement of soil moisture through water allocation into implicit economic benefits. This connects short-term allocation with long-term benefits on a time scale, ensuring priority for production preservation during critical water periods and moderate deficits during non-sensitive periods, achieving optimal allocation of water resources in both time and space.

[0052] In one embodiment, the objective function for minimizing the overall water shortage penalty includes the water shortage penalty of agricultural irrigation plots. The calculation formula is: In the formula, For the first Unit price of agricultural irrigation crops; For the first The maximum yield per unit area of ​​an agricultural irrigation crop; For the first The planting area of ​​each agricultural irrigation area; and The first The first agricultural irrigation crop Daily actual evaporation and maximum potential evaporation; For the first The daily water sensitivity index of crops within a specific agricultural irrigation area during the specified scheduling period is taken as the [number]th [item]. The stage water sensitivity index of a crop's growth stage is calculated by dividing the total number of days in that growth stage by the stage water sensitivity index of that growth stage. This stage water sensitivity index reflects the crop's sensitivity to water deficiency at a specific growth stage. The moisture sensitivity index for the aforementioned stage, and the critical period for moisture. Large value, non-sensitive period The value is small; T is the number of days included in the scheduling period.

[0053] In application, this embodiment demonstrates the water shortage penalty for agricultural irrigation patches. Based on the idea of ​​a water-based production function, this method multiplies and accumulates the impact of daily water deficits on total yield during the entire growing season, then converts this into an economic loss. In the formula, , and The first The product unit price, maximum yield per unit area, and planting area of ​​a crop in an agricultural irrigation area represent the maximum output value that the area can obtain under fully irrigated conditions; the product term in parentheses represents the actual daily transpiration evaporation. With maximum potential evaporation The ratio is used as the relative evapotranspiration ratio for that day, and the moisture sensitivity index is used as the metric. The index is used for weighted multiplication to accumulate the relative impact of water deficit on final yield for each day during the scheduling period. Subtracting this multiplication term from the index gives the relative yield reduction ratio for the entire growing season corresponding to the water deficit during the scheduling period. Multiplying this ratio by the maximum yield gives the economic loss of yield caused by water shortage in the agricultural irrigation area. For days outside the scheduling period, the relative evapotranspiration ratio is taken as 1 based on full water supply. The ratio is also taken as 1 when the maximum potential evapotranspiration is zero on a certain day.

[0054] In application, the moisture sensitivity index The water sensitivity index is dynamically set according to the crop's growth stage, with larger values ​​during the critical water period and smaller values ​​during the non-sensitive period. This results in crops in the critical water period experiencing a greater yield reduction and economic loss under the same relative evapotranspiration deficit, while crops in the non-sensitive period experience relatively smaller yield reductions and economic losses even with some deficit. Therefore, this embodiment uses the calculation formula for agricultural irrigation area water deficit penalty to quantitatively map the water deficit during the agricultural scheduling period to the yield and economic loss throughout the entire growth period. By introducing a water sensitivity index that dynamically changes with the growth stage, the model can identify and amplify the water deficit cost during the critical water period, thus prioritizing water supply for agricultural irrigation areas in the critical water period in multi-objective optimization. The product of the actual evapotranspiration and the maximum potential evapotranspiration is used as a yield reduction measure, quantifying the cumulative impact of daily water deficit on the final yield during the scheduling period. This provides a unified and comparable economic measurement scale for the flexible allocation of agricultural water at different growth stages.

[0055] In one embodiment, the constraints of the multi-objective optimization allocation model include water source constraints, water user balance constraints, minimum guarantee rate constraints for water users, water transmission capacity constraints for agricultural water supply pipelines, water transmission capacity constraints for domestic and industrial water supply networks, and water intake capacity constraints for agricultural water intakes. When the water source is a surface reservoir, the water volume constraint on the water source side is: In the formula, , These represent the initial and final reservoir capacities, respectively. This represents the total inflow into the reservoir. This refers to the total water supply from the reservoir used for power generation. , , These represent the total water supply for domestic, industrial, and agricultural use, respectively. This refers to the total flood discharge from the reservoir. The amount of water lost through evaporation and seepage from the reservoir; and the reservoir capacity at the end of the period shall not be lower than the dead storage capacity and not higher than the usable storage capacity; , For a given value; The , The parameters are provided by the reservoir dispatching department before the dispatching period according to the power generation plan and flood control requirements, and are used as input parameters for this model, but are not involved in the optimization of this model.

[0056] When the water source is a surface river or groundwater, the water volume constraint on the water source side is: In the formula, The total available water volume is the amount of water allocated during the scheduling period, and is not greater than the sum of the planned water demand of each water-using unit. The water supply capacity constraint of the agricultural water supply pipeline is: In the formula, Number the water outlets on each main and branch canal from bottom to top; For the first The water flow rate at each branch point; Each main and branch canal is numbered from bottom to top, corresponding to the water distribution point, with each main and branch canal numbered starting from 1; For the first The maximum flow rate corresponding to the channel segment; The water intake capacity constraint of the agricultural water intake is: In the formula, For the first The maximum flow rate of each branch point; The water balance constraint for each water user is: the actual water allocation of each water user during the scheduling period shall not exceed its planned water demand. The minimum guarantee rate constraint for water users is: the actual water allocation of each water user unit shall not be less than the product of its planned water demand and the corresponding guarantee rate; The water supply capacity constraint of the domestic and industrial water supply network is: the sum of the water supply flow of the downstream nodes of each pipe section in the domestic and industrial water supply network shall not exceed the maximum design flow of that pipe section.

[0057] In application, this embodiment defines constraints for the multi-objective optimization allocation model, including water source-side water quantity constraints, agricultural water supply pipeline conveyance capacity constraints, and agricultural water intake capacity constraints, to ensure the physical feasibility of the allocation scheme. Regarding water source-side water quantity constraints, constraints are established for different water source types: when the water source is a surface reservoir, the final reservoir capacity is constrained by the water balance relationship, i.e., the final reservoir capacity. equal to the initial storage capacity Including the total inflow of water from the reservoir Subtract the total water supply used for power generation, domestic use, industry, and agriculture. , , , and total flood discharge With evaporation and leakage This ensures the balance of revenue and expenditure of the reservoir during the scheduling period; when the water source is a surface river or groundwater, the total available water volume of the source will be used. To allocate the total amount, the total water supply for domestic, industrial, and agricultural use is constrained. , , The sum of these amounts equals the available water volume, thus completely distributing the available water supply among the three industries during the scheduling period and maintaining a balance between income and expenditure. The total available water volume is the water volume allocated during the scheduling period, and its value is not greater than the sum of the planned water demands of each water-using unit. Furthermore, the constraints also include water balance constraints for water users, used to constrain the income and expenditure relationship between the actual water allocation and the water demand of each water-using unit during the scheduling period, ensuring that the allocation plan also satisfies water balance at the level of each water-using unit. Regarding the minimum guarantee rate constraint for water users, the actual water allocation of each water-using unit is not less than the product of its planned water demand and the corresponding guarantee rate. The guarantee rate for domestic water use is to ensure basic livelihood, the guarantee rate for industrial water use is to maintain production continuity, and the guarantee rate for agricultural water use is to ensure the basic survival water needs of crops, thereby ensuring the minimum water supply floor for each industry in a multi-industry water competition scenario. Regarding the constraints on the water transmission capacity of domestic and industrial water supply networks, the sum of the water supply flow of each downstream node in the domestic and industrial water supply network shall not exceed the maximum design flow of that pipeline segment, thereby ensuring that the domestic and industrial water supply networks do not exceed their water transmission capacity.

[0058] In application, regarding the water conveyance capacity constraints of agricultural water supply pipelines, each main and branch canal is numbered sequentially from bottom to top and corresponds to the water distribution point. Each main and branch canal is numbered starting from 1. The downstream water distribution flow of the channel section is borne by each branch point. The sum must not exceed the maximum flow rate corresponding to that channel section. This ensures that the water conveyance of the channel does not exceed its designed flow capacity; regarding the water intake capacity constraints at agricultural water intakes, the project sequentially numbers each branch water intake from bottom to top, the number of which is... The water distribution flow rate of each branch point The flow rate must not exceed the maximum flow rate of the branch outlet. This ensures that the water intake at each branch point does not exceed its designed diversion capacity. In this embodiment, the maximum water conveyance capacity of the main canal refers to the maximum flow rate corresponding to the section at the head of the main canal, i.e., the maximum flow rate corresponding to the section with the largest number in the agricultural water supply pipeline water conveyance capacity constraint. Whether the total water demand of each sub-irrigation area triggering water supply on the same day exceeds the maximum water conveyance capacity of the main canal is checked daily based on this constraint. The above constraints, along with the designed flow rate of each pipe section or canal section and the maximum diversion capacity of each branch point in the aforementioned directed graph topology network, together constitute the hydraulically feasible region of the allocation scheme. Therefore, this embodiment establishes water source-side water quantity constraints for surface reservoirs and surface rivers or groundwater respectively, ensuring that the allocation scheme meets the water quantity balance and the upper limit of available water quantity under different water source conditions, avoiding over-extraction and imbalance. By constraining the water transmission capacity of agricultural water supply pipelines and the water intake capacity of agricultural water inlets, the design flow capacity of canal sections and the design diversion capacity of water outlets are introduced into the optimization model, so that the generated water allocation scheme can be carried by the actual water transmission and distribution capacity of the canals and water outlets, avoiding infeasible schemes that exceed the engineering water transmission and distribution capacity, and ensuring the engineering feasibility of the allocation results.

[0059] In one embodiment, the phased solution includes the following steps S401 to S403: Step S401: Using the total available water supply as a constraint, perform macro-level water allocation for domestic, industrial, and agricultural sectors at the industry type level. Use a multi-objective optimization algorithm to obtain the Pareto optimal solution set, and determine the total industry water allocation for domestic, industrial, and agricultural sectors during the scheduling period from the Pareto optimal solution set according to preset decision rules. Step S402: Under the constraint of the total water allocation of each industry, the total water allocation of each water plant, each factory, and each agricultural irrigation area is used as the decision variable to allocate water to each water-using unit, and the total water allocation of each agricultural irrigation area is further allocated to each sub-irrigation area under the jurisdiction of the agricultural irrigation area to obtain the total water allocation of each sub-irrigation area; among which, agricultural water use is prioritized to ensure the sub-irrigation areas in the critical water period according to the crop stage water sensitivity index, while meeting the minimum guarantee rate constraint of each water-using unit; Step S403: Based on the total water allocation of each water-using unit and the total water allocation of each sub-irrigation area, the domestic water and industrial water are continuously allocated on a daily basis according to the daily water demand ratio. The agricultural water is allocated using a trigger-based centralized water supply schedule, and the water transmission capacity constraints are checked daily to generate the water allocation plan for each water-using unit and the centralized agricultural water supply schedule.

[0060] In applications, such as Figure 2As shown, this embodiment addresses the problem of high-dimensional daily allocation variables in large irrigation districts with multiple water sources and users, which is difficult for traditional multi-objective algorithms to converge. It employs a phased dimensionality reduction strategy to decompose the high-dimensional allocation problem into three cascaded sub-problems: industry-type macro-allocation, water-use unit-level meso-allocation, and time-series daily-level micro-allocation, which are solved sequentially. In the industry-type macro-allocation sub-problem, the water demand of each water plant, factory, and agricultural irrigation area obtained from the aforementioned water demand prediction stage during the scheduling period is extracted and summarized to obtain the total water demand for domestic, industrial, and agricultural sectors. This is combined with the total available water supply from the water source during the scheduling period. Using the total available water supply as a constraint, and with the multi-objective functions of minimizing the comprehensive water shortage penalty, minimizing water supply system energy consumption, and maximizing comprehensive economic benefits, a non-dominated sorting genetic algorithm with elite retention is used for optimization to generate a Pareto optimal solution set for industry water allocation. When calculating the agricultural water shortage penalty, the total agricultural water allocation is divided according to the water demand of each agricultural irrigation area. Water is pre-allocated to each irrigation area. The water shortage penalty for each irrigation area is calculated and summed based on the stage water sensitivity index of the crop in each irrigation area (converted to a daily water sensitivity index by dividing by the total number of days in the stage). This allows the crop stage water sensitivity index to influence the macro-water allocation among industries through the magnitude of the water shortage penalty. Subsequently, the weights of the three objective functions—comprehensive water shortage penalty, total energy consumption, and comprehensive economic benefits—are determined by combining the analytic hierarchy process (AHP) and the entropy weight method. The optimal industry total water allocation scheme is then selected from the Pareto optimal solution set using the approximation ideal solution ranking method according to preset decision rules. This determines the total industry water allocation for domestic, industrial, and agricultural sectors during the scheduling period.

[0061] In applications, at the industry-type level, the objective functions use the water distribution volume of each water-using unit and the operating parameters of each pumping station as direct variables, while at the industry-type level, only the total water distribution volume of the three industries of domestic, industrial, and agricultural is considered. , , Since the total water allocation for the industry is a decision variable, before evaluating the objective function, the pre-allocation water volume for each water-using unit is first converted according to a preset pre-allocation rule. This pre-allocated water volume is only used for objective function evaluation and constraint verification at the industry level and is not considered the final water allocation result at the water-using unit level. For domestic water use, the total water allocation for the domestic industry is... Water demand is pre-allocated to each water plant according to the proportion of its planned water demand to the total water demand of the domestic sector. i The pre-allocated water volume of each water plant is ,and No more than For industrial water use, the total water allocation for industrial sectors will be... Water is pre-allocated to each factory according to the proportion of each factory's planned water demand to the total water demand of the industrial sector. k The pre-allocated water volume for each factory is ,and No more than For agricultural water use, as mentioned above, water is pre-allocated to each irrigation area according to its water demand ratio. Through this pre-allocation, the water allocation rate is the same for all water-using units within the same industry, and each unit shares the industry's water shortage in the same proportion, preventing situations where individual water-using units experience severe water shortages while others receive sufficient water. Therefore, the objective function... The penalties for domestic water shortage and industrial water shortage depend only on and For the corresponding monotonically decreasing function, when the total water allocation for the industry is not less than the total water demand for the industry, each water-using unit within the industry will fully allocate water according to the planned water demand, and the corresponding water shortage penalty term will be zero. In another implementation, domestic water use and industrial water use can also be calculated separately according to the water shortage economic loss coefficient. , The product of the planned water demand and the total water demand in the same industry is used for pre-allocation, so that the pre-allocation result is adapted to the water shortage penalty weight.

[0062] In application, the objective function for minimizing the energy consumption of the water supply system and the decision variables for the three-stage dimensionality reduction solution are quantitatively mapped as follows: Based on the aforementioned directed graph topology network, the water supply path of each water user unit and the pumping stations and booster stations along the path are determined. The total water allocation for each industry is decomposed to each water user unit according to the aforementioned pre-allocation rules. Then, the water is accumulated step-by-step from downstream to upstream along the water supply path of each water user unit to obtain the [then...] p The amount of water pumped by each pumping station during the scheduling period is calculated by dividing the water pumped by the operating time of the pumping station during the scheduling period to obtain the first pumping station. p The pumping flow rate of each pumping station is calculated as follows: at the industry type and water user unit levels, the pumping station operating time is taken as the duration of the scheduling period, and the resulting pumping flow rate is the average flow rate during the scheduling period; at the daily time series level, the pumping station operating time and pumping flow rate are determined by the daily water distribution plan and the triggered water supply schedule; the pumping head is determined by the water level difference between the inlet and outlet pools of the pumping station and the friction head loss and local head loss of the pipeline section; the overall efficiency of the pumping station unit is determined by interpolation of the aforementioned flow-efficiency curve according to the pumping flow rate. Therefore, the total industry water distribution at the industry type level, the total water distribution of water user units at the water user unit level, and the daily water distribution at the daily time series level can all be converted into the pumping flow rate and operating time of each pumping station at each level, making the energy consumption of the water supply system quantitatively calculable at all three levels depending on the values ​​of the decision variables; among them, the multi-objective optimization of the water supply system energy consumption is completed at the industry type level.

[0063] In application, the objective function that minimizes the comprehensive water shortage penalty and the agricultural-related terms in the objective function that maximizes the comprehensive economic benefit are quantitatively mapped to the decision variables obtained through the three-level dimensionality reduction and phased solution as follows: For the first... jFor each agricultural irrigation plot, the daily water allocation for that plot during the specified scheduling period is converted into net irrigation water entering the field using the canal system water utilization coefficient. This net water allocation, along with the crop water requirement and effective rainfall for that day, is then substituted into the daily calculation of root zone soil moisture content to obtain the root zone soil moisture content for that plot on day t. Based on this root zone soil moisture content, the soil moisture stress coefficient for that day is determined, and this coefficient is multiplied by the maximum potential transpiration evaporation for that day, obtained using the Penman-Montes formula. That is, to obtain the actual evaporation on that day. By substituting the formula for calculating the water shortage penalty of an agricultural irrigation area into the given formula, the water shortage penalty of that irrigation area can be quantitatively calculated from the water distribution amount. The scheduling period ends at the end of the period. j Increase in soil water storage in the root zone of each agricultural irrigation area The difference between the root zone soil moisture content at the end of the scheduling period and the measured root zone soil moisture content at the beginning of the scheduling period, obtained through daily extrapolation, is multiplied by the planned wetting layer depth and planting area of ​​the agricultural irrigation area to convert it into water volume. Thus, the total water allocation for the industry at the industry type level, the total water allocation for the water-using unit at the water-using unit level, the total water allocation for the sub-irrigation area, and the daily water allocation at the time-series level can all be converted level by level into the actual evaporation and evaporation of each agricultural irrigation area and the increase in soil water storage at the end of the period. This allows the two objective functions of comprehensive water shortage penalty and comprehensive economic benefit to be quantitatively calculable at all three levels, depending on the values ​​of the decision variables. Together with the aforementioned water supply system energy consumption objective function, this constitutes a multi-objective optimization allocation model that can be fully evaluated by the decision variables.

[0064] In the application, in the water-use unit-level meso-level allocation sub-problem, under the constraint of the total water allocation for each industry, the total water allocation for each industry is further allocated to specific water-use units: for domestic and industrial water use, linear programming is used to solve the problem according to the water shortage penalty weight and minimum guarantee rate constraint of each water plant and factory to obtain the total water allocation for each water plant and factory during the scheduling period; for agricultural water use, the pre-allocation method based on the water demand ratio is broken, and a crop stage water sensitivity index is introduced when allocating the total agricultural water allocation to each sub-irrigation area, giving priority to sub-irrigation areas in the critical water period, and allowing moderate deficits in sub-irrigation areas in the non-sensitive period, thereby realizing a refined secondary allocation of water rights within agriculture. In the time-series daily-level micro-allocation subproblem, based on the determined total water allocation for each water-using unit, and combined with the daily water demand process obtained from water demand forecasting, rainfall forecasts, and the aforementioned soil water storage carry-over mechanism, differentiated time-series allocation strategies are adopted for different water use types: For domestic and industrial water use in each water plant and factory, under the premise of meeting the daily water delivery capacity constraints of the water supply network, the total water allocation is continuously allocated daily according to the daily water demand ratio, generating daily continuous water supply instructions for domestic and industrial use; For agricultural water use, a trigger-based centralized water supply schedule is adopted to determine the optimal water supply time for each sub-irrigation area and meet its stage water demand at once, ultimately generating the water allocation plan for each water-using unit and the centralized agricultural water supply schedule.

[0065] In application, Figure 2 This embodiment illustrates the logical relationship of the phased dimensionality reduction solution, which presents three levels from top to bottom: macro-allocation at the industry type level, meso-allocation at the water-using unit level, and micro-allocation at the time-series daily level. The total water allocation for the industry and the total water allocation for the water-using unit obtained at the previous level are passed down sequentially to form the water constraints for the next level; the crop stage water sensitivity index ( Figure 2 Chinese abbreviation is When substituting into the water shortage penalty calculation formula, the daily water sensitivity index is converted by dividing by the total number of days in the stage. This process runs through three levels: at the industry type level, water shortage penalties affect water allocation between industries; at the water use unit level, it prioritizes water supply to sub-irrigation areas in the critical water period; and at the time-series daily level, it determines the water supply order of multiple sub-irrigation areas on the same day. At the time-series daily level, water is distributed according to water use type: domestic and industrial water use is continuously allocated according to the daily water demand ratio, while agricultural water use enters a triggered centralized water supply period. The criterion is whether the root zone soil moisture content obtained from daily calculations has dropped to the appropriate lower limit threshold. If it has not dropped to the lower limit, daily calculations continue. When it drops to the lower limit, a one-time centralized water supply volume is calculated and irrigation is grouped according to the crop stage water sensitivity index. The irrigation grouping is determined by whether the total water demand of each sub-irrigation area triggered for water supply on the same day exceeds the maximum water conveyance capacity of the main canal. If it exceeds, water is supplied in descending order of crop stage water sensitivity index, with the sub-irrigation areas ranked lower being postponed to the next day. At the industry-type macro-allocation level, the water demand during the scheduling period for each water plant, factory, and agricultural irrigation area, along with the total available water supply from the water source, are used as inputs. A Pareto optimal solution set is obtained through a non-dominated sorting genetic algorithm with elite retention. Then, the comprehensive weights are determined using the analytic hierarchy process and the entropy weight method, and the total water allocation for domestic, industrial, and agricultural sectors is obtained by screening using a sorting method that approximates the ideal solution. At the water-use unit-level meso-allocation level, the total water allocation for domestic and industrial water use is obtained by solving a linear programming problem under constraints of water shortage penalty weights and minimum guarantee rates. Agricultural water use... Based on the crop stage water sensitivity index, priority is given to ensuring the water supply of sub-irrigation areas in the critical water period. In the time-series daily micro-allocation level, domestic and industrial water use is continuously allocated daily according to the daily water demand ratio. Agricultural water use is generated after soil moisture dynamic simulation, water supply timing trigger judgment, one-time centralized water supply calculation, and rotation irrigation grouping and water supply scheduling. Finally, the daily water use allocation plan and agricultural centralized water supply schedule of each water use unit are output. The three levels form a solution link from top to bottom, which refines the macro decision-making to the micro execution and transmits the constraints level by level.

[0066] This embodiment employs a phased solution strategy at the industry type level, water-using unit level, and daily time-series level. It sequentially processes inter-industry water allocation, specific water-using unit allocation, and daily water supply scheduling, reducing the complexity of solving a high-dimensional multi-objective problem all at once. Simultaneously, the solution results from the previous level serve as water constraints for the next level, ensuring consistency between macro-level allocation, meso-level distribution, and micro-level execution. Agricultural water allocation is further constrained by crop stage water sensitivity indices, soil moisture content projections, and canal water conveyance capacity constraints, enabling the solution results to be converted into specific water supply dates and centralized water supply volumes.

[0067] In application, the changes in decision variables, constraint propagation, and complexity of the three-stage dimensionality reduction solution in this embodiment are as follows: Industry-type level: The decision variables are the total water allocation for three industries: residential, industrial, and agricultural, with a total number of variables of 3; the constraints are the total available water supply and the minimum guarantee rate for each industry; the solution method is a non-dominated sorting genetic algorithm with elite retention combined with an approximation ideal solution sorting method to obtain the Pareto optimal solution set and to select and determine the total water allocation for each industry according to the preset decision rules.

[0068] Water use unit level: The decision variable is the total water allocation for each water plant, each factory, and each agricultural irrigation area; the number of variables is [number missing]. Nw + Nf + Na ,in Nw , Nf , Na The number of water plants, factories, and agricultural irrigation areas are respectively; the constraints are the total water allocation for each industry, the minimum guarantee rate for each water-using unit, and the water balance constraint for each water-using unit, which are determined at the industry type level; domestic and industrial water use are solved by linear programming according to the water shortage penalty weight and the minimum guarantee rate constraint; agricultural water use is based on the crop stage water sensitivity index to prioritize water-critical period sub-irrigation areas and allow moderate deficits in non-sensitive period sub-irrigation areas.

[0069] Daily time series: The decision variable is the water allocation volume of each water-using unit on each day during the scheduling period, and the number of variables is ( Nw + Nf )× T + Na × Ttrig ,in T The number of days in the scheduling period. Ttrig The actual number of days for triggering water supply in agriculture; the constraints are the total water allocation of each water-using unit, the daily channel water conveyance capacity, the daily water intake capacity of the branch water outlet, and the water conveyance capacity of the domestic and industrial water supply network determined at the water-using unit level; domestic and industrial water use is continuously allocated according to the daily water demand ratio, and agricultural water use adopts a trigger-based centralized water supply schedule.

[0070] Regarding the transmission of constraints between levels, the total water allocation for each industry determined at the industry type level serves as the total water allocation constraint at the water-using unit level, and the total water allocation for each water-using unit determined at the water-using unit level serves as the total water allocation constraint at the time-series daily level. Higher-level decisions impose hard constraints on lower levels rather than soft suggestions. Regarding the consistency of constraints across levels, the water source-side water allocation constraint remains consistent across all three levels; the minimum guarantee rate constraint for water-using units remains consistent at both the water-using unit level and the time-series daily level; and the water transmission capacity constraint is checked daily at the time-series daily level and uses the same parameters as the maximum flow rate of the channel section and the maximum flow rate of the branch outlet in the allocation model constraints.

[0071] Regarding the change in complexity, if the original problem is solved directly, the number of decision variables is ( Nw + Nf + Na )× T ,when Nw =5、 Nf =10、 Na =20、 T When the dimensionality is reduced to 15, the number of variables reaches 525, making it difficult for the multi-objective genetic algorithm to converge within a reasonable time. After three-stage dimensionality reduction, the first stage has only 3 variables, the second stage has 35 variables, and the third stage has approximately (5+10)×15+20×15 variables. Ttrig 1 variable ( Ttrig (Typically 3 to 5); in the second level, domestic and industrial water use is solved using linear programming; in the third level, domestic and industrial water use is directly allocated according to the daily water demand ratio, and agricultural water use is determined according to the rules of soil moisture content threshold triggering and sensitivity index ranking. Neither requires multi-objective population optimization. Therefore, the actual decision variables that need to be optimized by the multi-objective genetic algorithm are only the three from the first level. The number of variables at each level is significantly reduced, and suitable solution algorithms (genetic algorithm, linear programming, trigger scheduling) can be used, significantly reducing the solution scale and convergence difficulty. Because the solution results of each level serve as the water quantity constraint for the next level, this three-level structure ensures consistency in macro-level configuration, meso-level allocation, and micro-level execution, unlike ordinary step-by-step calculations that simply splice together solutions from each level independently.

[0072] In one embodiment, the trigger-based centralized water supply scheduling for agricultural water use includes steps S501 to S504: Step S501: Using the measured soil moisture content in the root zone as the initial state, and combining the daily crop water requirement and daily effective rainfall, the change process of soil moisture content in the root zone of each agricultural sub-irrigation area is deduced day by day. Step S502: When the soil moisture content in the root zone of the sub-irrigation patch drops to the lower limit threshold of suitable soil moisture content, the water supply command for the sub-irrigation patch is triggered, and the day is determined as the water supply time for the sub-irrigation patch. Step S503: Calculate the net irrigation amount required to restore the soil moisture content in the root zone to the upper limit of the suitable soil moisture content and the net irrigation amount required to meet the remaining water demand of this growth stage. Take the smaller value of the two and calculate the one-time concentrated gross water supply of the sub-irrigation area in combination with the channel water conveyance loss. The sum of the one-time concentrated gross water supply of the sub-irrigation area during the scheduling period shall not exceed the total water distribution of the sub-irrigation area. Step S504: When multiple sub-irrigation areas trigger water supply on the same day and the total water demand exceeds the maximum water conveyance capacity of the main canal, priority is assigned according to the crop stage water sensitivity index, and water is supplied to the sub-irrigation areas with higher water sensitivity indices first, while the remaining sub-irrigation areas are supplied water in turn.

[0073] In this application, this embodiment addresses the unsuitability of daily micro-water supply for agricultural irrigation areas. It abandons the assumption of daily micro-water supply that is detached from the actual conditions of irrigation projects and adopts a trigger-based centralized water supply scheduling mechanism based on soil moisture balance. The execution process can be divided into four stages: dynamic soil moisture projection, water supply timing trigger judgment, one-time centralized water supply calculation, and rotational irrigation grouping and water supply scheduling. In the dynamic soil moisture projection stage, using the current measured soil moisture content as the initial state, and combining the daily crop water requirements and daily effective rainfall, the change process of soil moisture content in the root zone of each agricultural sub-irrigation area is projected daily, thus characterizing the evolution of soil moisture balance in each sub-irrigation area over time. In the water supply timing triggering stage, a lower and upper threshold for suitable soil moisture content for crops are preset. During the daily simulation, when the soil moisture content of a certain sub-irrigation area drops to or falls below the lower threshold, a water supply command for that sub-irrigation area is triggered, and that day is determined as the water supply timing for that sub-irrigation area. Thus, the soil moisture balance is used as the criterion for whether or not to supply water.

[0074] In the application, during the calculation of the one-time centralized water supply, after the water supply opportunity is triggered, the net irrigation amount required to restore the soil moisture content of the root zone of the sub-irrigation area to the upper limit of the suitable soil moisture content and the net irrigation amount required to meet the remaining water demand of the growth stage are calculated respectively. The smaller value of the two is taken, and the net irrigation amount is converted into the one-time centralized gross water supply amount according to the canal water utilization coefficient and the canal water conveyance loss. At the same time, it is constrained that the sum of the one-time centralized gross water supply amounts of the sub-irrigation area during the scheduling period does not exceed the total water allocation amount of the sub-irrigation area, so as to replenish the soil moisture in the one-time centralized water supply without exceeding the total water volume constraint determined by the meso-level allocation. In the rotation irrigation grouping and water supply scheduling process, when multiple sub-irrigation areas trigger water supply on the same day and their total water demand exceeds the maximum water conveyance capacity of the main canal, the triggering sub-irrigation areas are prioritized according to the crop stage water sensitivity index. Sub-irrigation areas with higher sensitivity indices are given priority for water supply, while sub-irrigation areas with lower sensitivity indices are postponed to the next day. This ensures that sub-irrigation areas in the critical water period are prioritized when the water conveyance capacity of the main canal is limited, and generates the final agricultural gate opening and closing and centralized water supply scheduling instructions.

[0075] This embodiment uses the daily change in soil moisture content in the root zone as the trigger for water supply. When the moisture content falls below the appropriate lower limit, water is replenished in a concentrated manner. The amount of water supplied at one time is determined by combining the remaining water demand during the growth stage, water loss in the canal, and the upper limit of the total water distribution during the scheduling period. When multiple sub-irrigation areas are triggered simultaneously and exceed the water conveyance capacity of the main canal, the crops are sorted according to the water sensitivity index of the crop stage, so as to take into account the protection of the critical period of the crop, the stability of canal operation, and the requirements of rotational irrigation management.

[0076] In one embodiment, the method further includes converting the planting structure and crop growth stage into adjustment model parameters, specifically including the following steps S601 to S603: Step S601: Based on the crop type of each field in the planting structure, determine the crop coefficient of the corresponding crop and the stage water sensitivity index of each growth stage by looking up a table of the correspondence between crop type, growth stage and crop stage water sensitivity index; wherein, the correspondence table stores the correspondence between different crop types and their crop coefficients, the start and end dates of each growth stage and the stage water sensitivity index. Step S602: Based on the current growth stage of crops in each field and the start and end dates of the scheduling period, determine the range of growth stages spanned by the scheduling period, and calculate the proportion of days occupied by each growth stage within the scheduling period. Step S603: Using the proportion of days in each growth stage as weight, the stage water sensitivity index of each stage is weighted and summed to obtain the equivalent water sensitivity index of the sub-irrigation area in the scheduling period, which serves as the basis for ranking the water supply priority among sub-irrigation areas in the multi-objective optimization allocation model; when substituting into the agricultural water shortage penalty formula, it is converted into a daily water sensitivity index by dividing by the total number of days in the corresponding growth stage.

[0077] In this application, after the planting structure identification is completed and before the allocation model is established, an explicit parameter conversion step is added. The input of this step is the crop type and current growth stage of each field output by the planting structure identification module, and the output is the crop coefficient and water sensitivity index corresponding to each field. According to the aforementioned spatial hierarchy of agricultural irrigation area - agricultural sub-irrigation area - field, the water sensitivity index of each field is used as the crop stage water sensitivity index of the agricultural sub-irrigation area to which the field belongs. The crop coefficient is used to correct the theoretical baseline water requirement in the Penman-Montes formula; the water sensitivity index is used to prioritize water supply among sub-irrigation areas, and when substituted into the calculation formula for agricultural water shortage penalty, it is converted into a daily water sensitivity index by dividing by the total number of days of the corresponding growth stage.

[0078] In application, when the scheduling period is short and falls entirely within a single growth stage, the stage water sensitivity index of that stage is directly used as the equivalent water sensitivity index for that scheduling period. When the scheduling period spans two or more growth stages, such as when the scheduling period includes the late jointing stage and the early heading stage, the equivalent water sensitivity index for that sub-plot within the scheduling period is calculated using the proportion of days in the jointing and heading stages within the scheduling period as weights. This method maintains the sensitivity differentiation ability of the stage water sensitivity index at critical water periods and solves the problem of incomplete alignment between the scheduling period and the growth stage boundary, ensuring that short-term allocation can still reflect the overall stage-specific water requirements of the crop.

[0079] This embodiment transforms unstructured planting information obtained from remote sensing identification into structured parameters that can be directly called by the allocation model through two steps: table lookup conversion and cross-stage weighting. This opens up the data link from remote sensing identification to optimized allocation, enabling the model to automatically adjust the water shortage penalty weight for different crops and different growth stages, thereby achieving dynamic allocation driven by planting structure.

[0080] In one embodiment, this application provides a planting structure-driven dynamic allocation system for irrigation district water resources, comprising: The planting structure identification module is used to identify the planting structure of irrigation areas and the current growth stage of crops in each field based on multi-source remote sensing images. The water demand prediction module is used to predict the irrigation water demand of agricultural irrigation areas during the future scheduling period based on the planting structure and crop growth stage, and to obtain the planned water demand reported by water plants and factories. The allocation model module is used to establish a multi-objective optimization allocation model with the goals of minimizing the comprehensive water shortage penalty, minimizing the energy consumption of the water supply system, and maximizing the comprehensive economic benefits; and based on the crop stage water sensitivity index determined by the planting structure and crop growth stage, the water deficit of agricultural irrigation areas during the scheduling period is mapped to the yield and economic loss of the entire crop growth period. The allocation and solution module is used to solve the multi-objective optimization allocation model in stages according to industry type level, water use unit level, and time series daily level. The decision variables obtained from the solution of the previous layer are used as water quantity constraints for the next layer to generate daily water allocation plans for each water use unit. Among them, agricultural water use triggers centralized water supply based on the daily extrapolation results of root zone soil moisture content. When the total water demand of multiple sub-irrigation areas in the agricultural irrigation area that triggers water supply on the same day exceeds the maximum water conveyance capacity of the main canal, the water supply priority is determined according to the crop stage water sensitivity index.

[0081] In application, the planting structure-driven irrigation area water resource dynamic allocation system is used to execute the steps in the above-described planting structure-driven irrigation area water resource dynamic allocation method embodiment. The system can be a virtual device within the computing device, run by the processor of the computing device, or it can be the computing device itself. The planting structure identification module, water demand prediction module, allocation model module, and allocation solution module are sequentially connected and transmit data according to the order of data processing: The planting structure identification module takes multi-source remote sensing images of the irrigation area as input, extracts spectral and texture features from optical remote sensing images, extracts backscattering features from synthetic aperture radar images, and constructs a vegetation index time series curve by combining phenological calendars. This is input into a deep learning classification model, and after active learning iterative optimization, the planting structure and crop growth stage of the irrigation area are output, and this result is output to the water demand prediction module. The water demand prediction module, based on the planting structure and crop growth stage, integrates crop leaf area index and real-time soil moisture content, and calculates the result based on theoretical benchmarks. After calculating water demand and dynamically correcting it using a neural network, and deducting effective rainfall and capillary recharge from groundwater, the water demand of each water-using unit is aggregated level by level according to the water supply network topology. This water demand is then output to the allocation model module. The allocation model module takes the water demand of each water-using unit and the correspondence between water sources, channels, or pipelines and the water-using units as input. It constructs a multi-objective optimization allocation model with corresponding constraints, aiming to minimize comprehensive water shortage penalties, minimize water supply system energy consumption, and maximize comprehensive economic benefits. This model is then output to the allocation solution module. The allocation solution module performs staged dimensionality reduction on the multi-objective optimization allocation model to generate a water allocation plan for each water-using unit during the scheduling period. The modules collaborate through the aforementioned data transmission relationships to jointly realize the dynamic allocation function of irrigation district water resources driven by planting structure.

[0082] This application embodiment provides information on planting structure and crop growth stage through a planting structure identification module, hierarchical water demand corresponding to the water supply network topology through a water demand prediction module, water supply security, water supply energy consumption and economic benefits through an allocation model module, and daily water allocation plan and agricultural centralized water supply schedule through an allocation solution module. The data from each module are transmitted sequentially and form a closed loop from crop identification to water supply execution.

[0083] In application, the modules in the planting structure-driven dynamic allocation system for irrigation water resources provided in this application embodiment can be software program modules, or they can be implemented by different logic circuits integrated in a processor, or they can be implemented by multiple distributed processors.

[0084] In one embodiment, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the planting structure-driven dynamic allocation method for irrigation water resources as described in any of the above method embodiments.

[0085] In one embodiment, this application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the planting structure-driven dynamic allocation method for irrigation water resources as described in any of the above method embodiments.

[0086] In one embodiment, this application also provides a computer program product that, when run on an electronic device, causes the electronic device to execute the steps in the various method embodiments described above.

[0087] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for dynamic allocation of irrigation water resources in an irrigation district driven by planting structure, characterized in that, include: Based on multi-source remote sensing images, the planting structure of the irrigation area and the current growth stage of crops in each field are identified. Based on the planting structure and crop growth stage, the irrigation water demand of agricultural irrigation areas during the future scheduling period is predicted, and the planned water demand reported by water plants and factories is obtained. A multi-objective optimization allocation model is constructed, with the objectives of minimizing the comprehensive water shortage penalty, minimizing the energy consumption of the water supply system, and maximizing the comprehensive economic benefits. Specifically, based on the crop stage water sensitivity index determined by the planting structure and crop growth stage, the water deficit of agricultural irrigation areas during the scheduling period is mapped to the yield and economic loss of the entire crop growth period. The multi-objective optimization allocation model is solved in stages according to industry type, water use unit, and daily time sequence. The decision variables obtained from the previous layer are used as water quantity constraints for the next layer. The total water allocation and daily water allocation plan of each water use unit during the scheduling period are determined. The water use unit includes the water plant, the factory, and the agricultural irrigation area. The agricultural irrigation area includes at least one sub-irrigation area. Among them, agricultural water use is triggered by the daily projection of soil moisture content in the root zone of the sub-irrigation area; when the total water demand of multiple sub-irrigation areas that trigger water supply on the same day exceeds the maximum water conveyance capacity of the main canal supplying water to the sub-irrigation area, the water supply priority is determined according to the crop stage water sensitivity index.

2. The method for dynamic allocation of irrigation water resources driven by planting structure as described in claim 1, characterized in that, The multi-source remote sensing images include multi-temporal optical remote sensing images and synthetic aperture radar images; the identification of the planting structure of the irrigation area and the current growth stage of crops in each field based on the multi-source remote sensing images includes: Multi-temporal optical remote sensing images and synthetic aperture radar images of the irrigation area are acquired. Spectral and texture features are extracted from the optical remote sensing images, and backscattering features are extracted from the synthetic aperture radar images to construct a multi-source heterogeneous feature dataset. The vegetation index is calculated by combining the crop phenological calendar and a vegetation index time series curve is constructed. The multi-source heterogeneous feature dataset is fused with the vegetation index time series curve and then input into a deep learning classification model to obtain the initial planting structure and the prediction confidence of each field. A confidence threshold is set, and fields with prediction confidence levels below the threshold are selected and a manual verification sample set is generated. The deep learning classification model is iteratively optimized based on the labeled manual verification sample set to obtain the planting structure of the irrigation area. At the same time, based on the inflection point and peak characteristics of the vegetation index time series curve and combined with the local crop phenological calendar, the current growth stage of crops in each field is identified as the basis for subsequent water demand calculation and water sensitivity index determination.

3. The method for dynamic allocation of irrigation water resources driven by planting structure as described in claim 1, characterized in that, The predicted irrigation water demand for agricultural irrigation areas during the future scheduling period includes: Based on the planting structure, crop growth stage, and forecast meteorological data during the scheduling period, the theoretical baseline water requirement of crops in each field at different growth stages is calculated using the Penman-Montes formula and crop coefficients determined by looking up tables based on crop type and growth stage. Obtain the crop leaf area index and real-time soil moisture content, input the theoretical baseline water requirement, the crop leaf area index and the real-time soil moisture content into the neural network model to obtain the actual net irrigation water requirement of crops in each field. The predicted effective rainfall and groundwater capillary recharge are deducted from the actual net irrigation water demand to obtain the field water demand. The water demand of each agricultural irrigation area is then obtained by summarizing the data level by level according to the water supply network topology.

4. The method for dynamic allocation of irrigation water resources driven by planting structure as described in claim 1, characterized in that, The multi-objective optimization allocation model includes the following objective function: The objective function that minimizes the overall water shortage penalty is: In the formula, As a comprehensive water shortage penalty; , , These represent the number of water plants, factories, and agricultural irrigation areas, respectively. , The first The planned water demand and actual water distribution of each water plant; , The first Planned water demand and actual water distribution for each factory; , These are the economic loss coefficients per unit of water shortage for each water plant and each factory, respectively. , , These are the industry-specific water shortage penalty coefficients for domestic, industrial, and agricultural sectors, respectively. Water shortage penalties for agricultural irrigation areas; The objective function for minimizing the energy consumption of the water supply system is: In the formula, Energy consumption of the water supply system; The total number of pumping stations and booster stations supplying water to various water plants, factories, and agricultural irrigation areas; For the first The water pumping flow rate of each pumping station; To lift water; To improve the overall efficiency of the pump station units; The density of water; It is the acceleration due to gravity; For the first p The operating time of each pumping station during the scheduling period; The objective function that maximizes overall economic benefits is: In the formula, For comprehensive economic benefits; For the first Industrial output value coefficient per unit water volume of a factory; The conversion factor for converting soil water storage into economic benefits in the next period; For the end of the scheduling period The increase in soil water storage in the root zone of an agricultural irrigation area, measured in terms of water volume.

5. The method for dynamic allocation of irrigation water resources driven by planting structure as described in claim 4, characterized in that, In the objective function that minimizes the overall water shortage penalty, the water shortage penalty for agricultural irrigation plots is... The calculation formula is: In the formula, For the first Unit price of agricultural irrigation crops; For the first The maximum yield per unit area of ​​an agricultural irrigation crop; For the first The planting area of ​​each agricultural irrigation area; and The first The first agricultural irrigation crop Daily actual evaporation and maximum potential evaporation; For the first The daily water sensitivity index of crops within a specific agricultural irrigation area during the specified scheduling period is taken as the [number]th [item]. The moisture sensitivity index of the reproductive stage to which the day belongs is divided by the total number of days in that reproductive stage. The moisture sensitivity index for the aforementioned stage, and the critical period for moisture. Large value, non-sensitive period The value is small; T is the number of days included in the scheduling period.

6. The method for dynamic allocation of irrigation water resources driven by planting structure as described in claim 1, characterized in that, The constraints of the multi-objective optimization allocation model include water source constraints, water user balance constraints, minimum guarantee rate constraints for water users, water transmission capacity constraints of agricultural water supply pipelines, water transmission capacity constraints of domestic and industrial water supply networks, and water intake capacity constraints of agricultural water intakes. When the water source is a surface reservoir, the water volume constraint on the water source side is: In the formula, , These represent the initial and final reservoir capacities, respectively. This represents the total inflow into the reservoir. This refers to the total water supply from the reservoir used for power generation. , , These represent the total water supply for domestic, industrial, and agricultural use, respectively. This refers to the total flood discharge from the reservoir. The amount of water lost through evaporation and seepage from the reservoir; and the reservoir capacity at the end of the period shall not be lower than the dead storage capacity and not higher than the usable storage capacity; , For a given value; When the water source is a surface river or groundwater, the water volume constraint on the water source side is: In the formula, The total available water volume is the amount of water allocated during the scheduling period, and is not greater than the sum of the planned water demand of each water-using unit. The water supply capacity constraint of the agricultural water supply pipeline is: In the formula, Number the water outlets of each main and branch canal from bottom to top; For the first The water flow rate at each branch point; Each main and branch canal is numbered from bottom to top, corresponding to the water distribution point, with each main and branch canal numbered starting from 1; For the first The maximum flow rate corresponding to the channel segment; The water intake capacity constraint of the agricultural water intake is: In the formula, For the first The maximum flow rate of each branch point; The water balance constraint for each water user is: the actual water allocation of each water user during the scheduling period shall not exceed its planned water demand. The minimum guarantee rate constraint for water users is: the actual water allocation of each water user unit shall not be less than the product of its planned water demand and the corresponding guarantee rate; The water supply capacity constraint of the domestic and industrial water supply network is: the sum of the water supply flow of the downstream nodes of each pipe section in the domestic and industrial water supply network shall not exceed the maximum design flow of that pipe section.

7. The method for dynamic allocation of irrigation water resources driven by planting structure as described in claim 1, characterized in that, The phased solution includes: With the total available water supply as a constraint, and the total water allocation for domestic, industrial, and agricultural sectors as decision variables at the industry type level, the multi-objective optimization allocation model is solved using a multi-objective optimization algorithm to obtain a Pareto optimal solution set. The total water allocation for domestic, industrial, and agricultural sectors during the scheduling period is then determined from the Pareto optimal solution set according to preset decision rules. The total water allocation for each sector serves as the water quantity constraint at the next level of water-using unit. Under the constraint of the total water allocation for each industry, the total water allocation for each water plant, each factory, and each agricultural irrigation area is used as the decision variable to allocate water to each water-using unit. The total water allocation for each agricultural irrigation area is further allocated to each sub-irrigation area under its jurisdiction to obtain the total water allocation for each sub-irrigation area. The total water allocation for each water-using unit and the total water allocation for each sub-irrigation area serve as the next level of time-series daily water allocation constraints. Among them, agricultural water use is prioritized to ensure the sub-irrigation areas in the critical water period based on the crop stage water sensitivity index, while also meeting the minimum guarantee rate constraint for each water-using unit. Based on the total water allocation of each water-using unit and the total water allocation of each sub-irrigation area, and using the daily water allocation of each water-using unit as the decision variable, domestic water and industrial water are continuously allocated daily according to the daily water demand ratio. For agricultural water, a trigger-based centralized water supply schedule is adopted, and the water transmission capacity constraint is checked daily to generate the daily water allocation plan for each water-using unit and the centralized agricultural water supply schedule. Among them, the total available water supply constraint remains consistent across the three levels, the minimum guarantee rate constraint remains consistent at the water-using unit level and the time-series daily level, and the water transmission capacity constraint is checked daily at the time-series daily level.

8. The method for dynamic allocation of irrigation water resources driven by planting structure as described in claim 7, characterized in that, The aforementioned trigger-based centralized water supply scheduling for agricultural water use includes: Using the measured soil moisture content in the root zone as the initial state, and combining the daily crop water requirement and daily effective rainfall, the change process of soil moisture content in the root zone of each agricultural sub-irrigation area is deduced day by day; When the soil moisture content in the root zone of a sub-irrigation patch drops to the lower limit threshold of suitable soil moisture content, a water supply command for that sub-irrigation patch is triggered, and that day is designated as the water supply time for that sub-irrigation patch. At the water supply time, the net irrigation volume required to restore the soil moisture content in the root zone to the upper limit of the suitable soil moisture content and the net irrigation volume required to meet the remaining water demand of this growth stage are calculated respectively. The smaller value of the two is taken, and the one-time concentrated gross water supply volume of the sub-irrigation area is calculated in combination with the channel water conveyance loss. The sum of the one-time concentrated gross water supply volumes of the sub-irrigation area during the scheduling period shall not exceed the total water distribution volume of the sub-irrigation area. When multiple sub-irrigation areas trigger water supply on the same day and the total water demand exceeds the maximum water conveyance capacity of the main canal, priority is given to the sub-irrigation areas with higher water sensitivity indices based on the crop stage water sensitivity index, and the remaining sub-irrigation areas are supplied water in turn.

9. The method for dynamic allocation of irrigation water resources driven by planting structure as described in claim 1, characterized in that, It also includes converting the planting structure and crop growth stages into parameters for adjusting the model, specifically including: Establish a table of correspondence between crop type, growth stage and crop stage water sensitivity index, and determine the crop stage water sensitivity index of each sub-plot by referring to the table based on the identified crop type and current growth stage. When the scheduling period spans multiple growth stages, the crop stage water sensitivity index of each stage is weighted and summed based on the proportion of days occupied by each growth stage within the scheduling period to obtain the equivalent water sensitivity index of the sub-irrigation patch within the scheduling period. The crop stage water sensitivity index or the equivalent water sensitivity index is used as a parameter of the allocation model and input into the multi-objective optimization allocation model so that the water allocation priority is dynamically adjusted according to the changes in the identified crop type and growth stage.

10. A planting structure-driven dynamic water resource allocation system for irrigation districts, characterized in that, include: The planting structure identification module is used to identify the planting structure of irrigation areas and the current growth stage of crops in each field based on multi-source remote sensing images. The water demand prediction module is used to predict the irrigation water demand of agricultural irrigation areas during the future scheduling period based on the planting structure and crop growth stage, and to obtain the planned water demand reported by water plants and factories. The allocation model module is used to establish a multi-objective optimization allocation model with the goals of minimizing the comprehensive water shortage penalty, minimizing the energy consumption of the water supply system, and maximizing the comprehensive economic benefits. Based on the crop stage water sensitivity index determined by the planting structure and crop growth stage, the water deficit of agricultural irrigation plots during the scheduling period is mapped to the yield and economic loss of the entire crop growth period. The allocation and solution module is used to solve the multi-objective optimization allocation model in stages according to industry type level, water use unit level, and time series daily level. The decision variables obtained from the solution of the previous layer are used as water quantity constraints for the next layer to generate daily water allocation plans for each water use unit. Among them, agricultural water use triggers centralized water supply based on the daily extrapolation results of root zone soil moisture content. When the total water demand of multiple sub-irrigation areas in the agricultural irrigation area that triggers water supply on the same day exceeds the maximum water conveyance capacity of the main canal, the water supply priority is determined according to the crop stage water sensitivity index.

Images