Agricultural drought multi-scale monitoring and early warning and drought resistance resource scheduling system

Abstract

The invention discloses an agricultural drought multi-scale monitoring and early warning and drought resistance resource scheduling system, and the system comprises a data perception and execution layer which carries out the irrigation operation through an intelligent irrigation terminal, and collects and reports the irrigation behavior data; a data warehouse of the intelligent analysis and management layer stores multi-source information such as irrigation behavior data, weather forecast data and remote sensing monitoring data, a drought evaluation model module generates two types of drought distribution information based on the data, and model parameters are dynamically adjusted through deviation comparison. And the scheduling optimization engine module is combined with future precipitation forecast to calculate a predictive water saving amount, and a multi-objective optimization model is constructed to generate a preventive drought-resistant scheduling scheme containing a cross-regional water resource allocation scheme and an irrigation adjustment instruction. Closed-loop optimization is formed through irrigation behavior data feedback, drought evaluation accuracy and scheduling pertinence are improved, and meanwhile the water resource utilization efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of agricultural drought resistance technology, specifically to an agricultural drought multi-scale monitoring and early warning system and drought resistance resource allocation system. Background Technology

[0002] Agricultural drought is one of the major natural disasters threatening food security and sustainable agricultural development. Traditional drought monitoring and irrigation management are often disconnected in terms of data sources: on the one hand, they rely on external observation data such as meteorological stations and remote sensing satellites for macro-level drought assessment; on the other hand, they rely on preset thresholds or fixed soil moisture sensors for localized irrigation responses. This disconnect makes it difficult for early warning models to accurately reflect the complex water stress conditions in the fields, and irrigation decisions lack proactive prediction and coordination of regional drought evolution trends and future meteorological conditions, resulting in inefficient water resource allocation and delayed drought response.

[0003] To improve the efficiency of agricultural water resource utilization, existing technologies have attempted improvements from different perspectives. For example, invention patent CN110738196A provides a real-time irrigation forecasting system based on regional soil moisture monitoring and remote sensing data. This system generates irrigation forecasts by integrating remote sensing data, environmental data, and a soil moisture prediction module, and corrects the forecasts using an actual water usage module. However, its "correction" mechanism mainly relies on a simple comparison and adjustment between the forecast results and the actual water usage afterward. It lacks the use of dynamic operational data (such as opening and closing frequency, duration, and water volume) from widely distributed irrigation execution terminals (such as valves) as real-time, high-density ground verification information for continuous, closed-loop optimization and correction of the front-end drought early warning model itself. This results in a bottleneck in its early warning accuracy in farmland scenarios with complex spatiotemporal variations. Another invention patent, CN108764573B, relates to a multi-dimensional balanced allocation system for water resources in inland arid regions. This invention focuses on the multi-objective optimization and allocation of water resources at the macro-basin or regional level. However, its regulation and decision-making are mainly based on static or periodically updated hydrological and ecological data models, which fail to form a close loop with real-time and proactive irrigation behavior and demand feedback in the field. It cannot achieve dynamic synergy of "micro-level water use behavior driving macro-level model optimization and macro-level scheduling strategies guiding micro-level precise prevention".

[0004] In summary, the limitation of existing technologies lies in their failure to construct a closed-loop system that uses massive behavioral data from the irrigation system's execution end as feedback to drive the dynamic adaptive optimization of drought early warning models, and on this basis, achieves proactive and preventative water resource allocation. This results in a disconnect between the two key aspects of agricultural drought management: "perceiving the true drought situation" and "implementing efficient drought resistance," thus restricting the accuracy of early warnings and the timeliness and predictability of allocation measures.

[0005] Therefore, there is an urgent need to propose an innovative technical solution to address the problem of the lack of feedback loop between the existing agricultural drought monitoring and early warning and irrigation scheduling systems, which prevents the integration of drought assessment and drought relief scheduling into a dynamic and coordinated optimization. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide an agricultural drought multi-scale monitoring and early warning and drought relief resource scheduling system. Through dynamic optimization of dual drought assessment models, closed-loop collaboration and multi-objective scheduling, it can accurately monitor drought conditions and efficiently allocate drought relief resources.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: an agricultural drought multi-scale monitoring and early warning and drought relief resource scheduling system, the system comprising: a data perception and execution layer, and an intelligent analysis and management layer;

[0008] The data perception and execution layer includes multiple intelligent irrigation terminals deployed in the irrigation area. The intelligent irrigation terminals are configured to perform irrigation operations and collect and report the start and stop time, duration and water consumption data of each irrigation operation to form irrigation behavior data. The intelligent analysis and management layer includes a data warehouse, a drought assessment model module, a scheduling optimization engine module and a human-computer interaction interface.

[0009] The data warehouse is configured to receive and store irrigation behavior data from the data perception and execution layer, as well as weather forecast data and remote sensing monitoring data from external sources.

[0010] The drought assessment model module is configured to: run a first drought assessment model based on the meteorological forecast data and the remote sensing monitoring data to generate first drought distribution information; generate second drought distribution information reflecting actual water demand based on the irrigation behavior data; compare the first drought distribution information with the second drought distribution information and deviation data; and automatically adjust the parameters of the first drought assessment model when the deviation data continuously exceeds a preset threshold, so that the drought distribution information output by the adjusted first drought assessment model converges to the second drought distribution information.

[0011] The scheduling optimization engine module is configured to: receive drought distribution information and trend prediction information output by the drought assessment model module, and combine the future precipitation forecast in the meteorological forecast data to calculate the predictable water-saving amount for areas with rainfall forecasts. Taking the overall drought resistance benefit of the region as the objective function, the predictable water-saving amount as the constraint for new water sources, and the engineering water conveyance capacity as the physical constraint, a multi-objective optimization model is constructed and solved to generate a preventive drought resistance scheduling scheme that includes cross-regional water resource allocation schemes and specific plot irrigation adjustment instructions.

[0012] Furthermore, the scheduling optimization engine module is also configured to send the irrigation adjustment command to the corresponding smart irrigation terminal;

[0013] The intelligent irrigation terminal is also configured to: adjust the subsequent irrigation operations according to the received irrigation adjustment instruction, and generate new irrigation behavior data to be fed back to the data warehouse, thereby forming a closed-loop optimization between drought assessment and irrigation scheduling;

[0014] The drought assessment model module is configured to generate the second drought distribution information and adjust the model parameters according to the following mathematical relationship, for each preset geographic grid unit. Calculate the real-time water stress index of the geographic grid cell i. The real-time water stress index The calculation formula is:

[0015]

[0016] in, Represents the geographic grid cell The total irrigation water consumption of all smart irrigation terminals within the set time window. Represents the geographic grid cell The potential evapotranspiration within the same time window is calculated from the aforementioned meteorological forecast data. Represents the geographic grid cell Crop coefficients of major crops in the region, function Indicates to , and The standardized mapping relationship is used to eliminate the influence of different crops and growth stages, so that... It can characterize a uniform relative degree of water scarcity;

[0017] The first drought assessment model outputs data for the geographic grid unit. The drought index is recorded as The system is in continuous Within each evaluation period, the calculation of each of the geographic grid units is performed. Cumulative deviation: ,when At that time, it triggers the impact on the geographic grid units in the first drought assessment model. Key parameters of the drought index The adjustment is aimed at improving the calculation of subsequent cycles. Approaching .

[0018] This represents the geographic grid cell output by the first drought assessment model. The drought index Geographic grid cells representing the output of the second drought assessment model In the The drought index for each cycle, This represents the geographic grid cell output by the first drought assessment model. In the The drought index for each cycle, This indicates the preset number of consecutive evaluation periods. Represents geographic grid cells In continuous The cumulative deviation of the drought index within each cycle This indicates the preset deviation threshold. Indicating the geographical grid units affected in the first drought assessment model Key parameters for calculating the drought index.

[0019] Furthermore, in the scheduling optimization engine module, the predictive water saving... For areas where effective rainfall is forecast Calculations were performed to determine the amount of water saved through this predictive method. The mathematical expression is:

[0020]

[0021] in, For the forecast area The effective rainfall, For the area The area of ​​crop planting, For the area The average soil precipitation infiltration recharge coefficient, For the area The theoretical crop water requirement for the current period is obtained by multiplying the reference crop evapotranspiration by the crop coefficient. For the area The function represents the amount of irrigation water already supplied in the current time period. This means taking the smaller of the two values ​​to ensure that the predicted water saving does not exceed the smaller of the rainfall replenishment amount and the actual water shortage amount.

[0022] Furthermore, the intelligent irrigation terminal includes a controller, a solenoid valve electrically connected to the controller, a water flow sensor, and an Internet of Things (IoT) communication unit;

[0023] The water flow sensor is configured to measure the volume of water flowing through the solenoid valve.

[0024] The controller is configured to control the opening and closing of the solenoid valve according to a local preset program or instructions received from the intelligent analysis and management layer, and to upload irrigation behavior data including terminal identification code, timestamp, valve status, irrigation duration and water consumption through the Internet of Things communication unit.

[0025] Furthermore, the first drought assessment model running in the drought assessment model module is a comprehensive assessment model that integrates the meteorological drought index and the remote sensing drought index.

[0026] The remote sensing drought index includes at least a vegetation water supply index or a soil moisture index derived from satellite remote sensing data.

[0027] Furthermore, the multi-objective optimization model constructed by the scheduling optimization engine module has an objective function... Simultaneously minimizing expected water shortage losses and scheduling costs, expressed as:

[0028]

[0029] in, This represents the total number of areas affected by the drought. For the area The weight of crop output value per unit area In order to implement the aforementioned preventive drought relief scheduling plan in the region The expected production loss to be recovered is allocated to the region. A function of water volume and the initial drought level in the area. The total number of water sources participating in the scheduling, To obtain water source The unit cost coefficient for water intake, To obtain water source Total water volume transferred out;

[0030] The constraints of the optimization model include: water demand balance constraints for all areas, upper limit constraints on the water supply capacity of each water source, flow capacity constraints for water transmission channels, and predictive water conservation measures. As a region Water demand reduction constraints.

[0031] Furthermore, the closed-loop optimization is specifically manifested as follows: after each output of drought distribution information, the drought assessment model module will continuously monitor the new second drought distribution information generated by the subsequently generated irrigation behavior data;

[0032] If the new second drought distribution information shows that the drought relief trend is consistent with the previous scheduling plan, then the current model parameters will be maintained.

[0033] When a deviation occurs, it triggers a recalibration of the first drought assessment model or an adjustment of the optimization weights in the scheduling optimization engine.

[0034] Furthermore, the data warehouse is also configured to receive and store measured data from fixed soil moisture monitoring stations and automatic weather stations deployed in the irrigation area;

[0035] When generating the second drought distribution information, the drought assessment model module also integrates the measured data from the fixed soil moisture monitoring station for spatial interpolation calibration.

[0036] Furthermore, the irrigation adjustment instructions generated by the scheduling optimization engine module specifically include a suggested irrigation start time, a suggested irrigation duration, or a suggested maximum water consumption quota for the next irrigation cycle.

[0037] The intelligent irrigation terminal is configured to integrate the irrigation adjustment command with a locally preset irrigation program, and execute the adjusted irrigation operation in a higher priority manner.

[0038] Furthermore, the intelligent analytics and management layer is deployed on a cloud computing platform;

[0039] The data perception and execution layer and the intelligent analysis and management layer communicate with each other through at least one of narrowband Internet of Things, fourth-generation mobile communication technology, or fifth-generation mobile communication technology.

[0040] Compared with existing technologies, this multi-scale monitoring, early warning and drought relief resource allocation system for agricultural drought has the following advantages:

[0041] I. This invention collects field irrigation behavior data through the data perception and execution layer, and integrates this data with multi-source information such as weather forecasts and remote sensing monitoring through the intelligent analysis and management layer to construct a dual drought assessment model. By comparing the deviations of the two types of drought distribution information, the model parameters are dynamically adjusted, and a closed-loop optimization mechanism is formed by combining the irrigation behavior data after scheduling execution. This effectively solves the problems in the existing technology of drought early warning models being disconnected from the actual water stress conditions in the field, and the lack of dynamic coordination between drought assessment and drought relief scheduling. It makes drought distribution information more consistent with the actual field conditions, significantly improves the accuracy of drought assessment and the timeliness of drought relief scheduling, and ensures that the scheduling plan can accurately match the actual water demand in the field, avoiding delays in drought relief response.

[0042] Second, this invention uses a scheduling optimization engine module combined with future precipitation forecasts to calculate predictive water-saving amounts. With the overall regional drought resistance benefit as the goal, it constructs a multi-objective optimization model to achieve cross-regional water resource allocation and precise adjustment of irrigation for specific plots. This fully taps the potential of natural rainfall resources, rationally coordinates various water sources, effectively improves water resource allocation efficiency, and reduces unnecessary irrigation water consumption. At the same time, the system relies on a closed-loop optimization mechanism to continuously adapt to dynamic changes in crop growth and meteorological conditions, enhancing the system's environmental adaptability and operational stability, and reducing the labor costs of agricultural drought management.

[0043] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0045] Figure 1 This is a diagram illustrating the overall system architecture and data interaction of the present invention.

[0046] Figure 2 This is a flowchart of the drought assessment and dynamic adjustment process of the present invention;

[0047] Figure 3 This is a flowchart of the closed-loop optimization and scheduling execution process of the present invention. Detailed Implementation

[0048] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0049] Example 1

[0050] like Figures 1 to 3As shown, this embodiment aims to elaborate on the specific implementation of an agricultural drought multi-scale monitoring, early warning, and drought-resistant resource scheduling system. Through the collaborative work of the data perception and execution layer, and the intelligent analysis and management layer, it achieves accurate monitoring and early warning of drought conditions and efficient scheduling of drought-resistant resources. This embodiment is designed based on the actual application scenario of a large irrigation area in the North China Plain. The irrigation area has a total area of ​​100 km², mainly growing wheat, corn, and rice. It is divided into 100 1 km × 1 km geographical grid units, deploying 500 intelligent irrigation terminals, and is equipped with 20 fixed soil moisture monitoring stations and 5 automatic weather stations. The system operates stably and reliably, and can meet the actual needs of precise control of agricultural irrigation and optimal allocation of drought-resistant resources.

[0051] In this embodiment, the overall system architecture is implemented as follows:

[0052] The core of the data perception and execution layer is the intelligent irrigation terminal, which serves as the execution unit and data acquisition unit for field irrigation operations and directly affects the system's perception accuracy and execution efficiency.

[0053] Composition and Installation of Smart Irrigation Terminals: Each smart irrigation terminal includes a controller, solenoid valve, water flow sensor, and IoT communication unit. The selection and installation of each component are designed around the practicality and stability of the irrigation system control. The controller uses an industrial-grade embedded controller, featuring low power consumption and anti-interference characteristics, adaptable to high-temperature and high-humidity field environments. Its core chip uses a 32-bit ARM processor, with a processing speed sufficient for local irrigation logic control and data processing needs. The solenoid valve is a waterproof solenoid control valve, with a nominal diameter matched to the irrigation pipe specifications, typically DN25-DN50. It is installed at the inlet of each irrigation area and electrically connected to the controller via wires. The response time is no more than 0.5 seconds, ensuring timely irrigation start and stop. The water flow sensor uses a turbine-type water flow sensor, installed downstream of the solenoid valve and coaxially connected to the irrigation pipe. Its measurement range is 0.1-10 m³ / h, with a measurement accuracy of ±2%, accurately capturing the volume of water flowing through the pipe. The IoT communication unit integrates communication modules of narrowband IoT, fourth-generation mobile communication technology and fifth-generation mobile communication technology. It can automatically switch communication modes according to the network coverage of the irrigation area. In remote areas, narrowband IoT is used first to achieve low-power data transmission, while 4G or 5G technology is used in areas with good network coverage to ensure the real-time performance of data transmission.

[0054] Data Acquisition and Reporting: The core function of the intelligent irrigation terminal is to execute irrigation operations and collect irrigation behavior data. The controller pre-stores local irrigation programs, which are set based on the crop's normal water requirements, including basic parameters such as irrigation cycle and duration of each irrigation. During irrigation, the controller controls the opening and closing of solenoid valves according to the local program or instructions received from the intelligent analysis and management system. Water flow sensors measure the volume of water flowing through the solenoid valves in real time and transmit the measurement signals to the controller. The controller, based on the opening and closing status of the solenoid valves, records the start and stop time and duration of each irrigation operation and calculates the water consumption.

[0055] Irrigation behavior data is reported using a combination of scheduled and triggered reporting. The scheduled reporting cycle is set to one hour. The controller uploads the irrigation behavior data for this cycle to the intelligent analysis and management layer via the IoT communication unit. The data includes the terminal identification code, timestamp, valve status, irrigation duration, and water consumption. When an abnormality occurs during irrigation, an immediate reporting mechanism is triggered to ensure timely feedback of abnormal information. The terminal identification code is a unique 16-bit string used to distinguish different smart irrigation terminals, facilitating the system's identification of the equipment status and data source for specific irrigation areas.

[0056] The intelligent analysis and management layer is deployed on a cloud computing platform, using cloud servers with elastic computing capabilities. Resource configuration is dynamically adjusted based on system data processing volume and access demands, supporting simultaneous access and operation by multiple users. This layer includes a data warehouse, a drought assessment model module, a scheduling optimization engine module, and a human-computer interaction interface. These modules work together to achieve data storage, drought assessment, scheduling scheme generation, and human-computer interaction functions.

[0057] The core function of a data warehouse is to receive and store various types of data, providing data support for drought assessment and scheduling optimization. Its data storage adopts a distributed storage architecture to ensure data security and scalability.

[0058] The data warehouse receives three types of data: First, irrigation behavior data from the data perception and execution layer, received in real time via the Internet of Things (IoT) communication network, categorized and stored according to terminal identification codes and timestamps, with a data retention period of 3 years to meet the needs of historical data traceability and model optimization; second, external weather forecast data and remote sensing monitoring data. Weather forecast data is obtained from the National Meteorological Data Sharing Platform, acquired twice daily, and includes temperature, humidity, wind speed, sunshine duration, and precipitation forecasts. Remote sensing monitoring data comes from remote sensing images from MODIS and Sentinel series satellites, acquired every 3 days. The data includes vegetation coverage, vegetation water supply index, and soil moisture index; thirdly, measured data from fixed soil moisture monitoring stations and automatic weather stations. Fixed soil moisture monitoring stations are installed at the center of each geographic grid unit and use the time domain reflectometer method to measure soil moisture content at depths of 0-20cm, 20-40cm, and 40-60cm with a measurement accuracy of ±1%. The measured data is reported every 2 hours. Automatic weather stations are deployed at key nodes in the irrigation area to measure meteorological parameters such as air temperature, humidity, wind speed, and sunshine duration. The measurement accuracy meets national meteorological observation standards, and the measured data is reported every 1 hour.

[0059] All data is standardized before storage, using JSON format for easy reading and processing. The data warehouse has data validation capabilities, verifying the completeness and rationality of received data. If missing or abnormal data is found, it is automatically marked and relevant maintenance personnel are notified. Simultaneously, interpolation or historical data substitution methods are used for temporary supplementation to ensure the continuity of subsequent model calculations.

[0060] The drought assessment model module is the core of the system for achieving accurate drought monitoring. By running the first and second drought assessment models, it generates comprehensive and accurate drought distribution information and continuously optimizes the assessment accuracy by dynamically adjusting the model parameters.

[0061] The first drought assessment model is a comprehensive assessment model that integrates meteorological drought index and remote sensing drought index, aiming to reflect the regional drought situation from a macro perspective. The meteorological drought index uses the Standardized Precipitation Index (SPI), which is calculated based on precipitation forecasts from meteorological forecast data, with a time scale of 15 days, and can reflect short-term precipitation deficits. The remote sensing drought index uses the Vegetation Water Supply Index (VSWI) and the Soil Moisture Index (SMI). The Vegetation Water Supply Index is obtained by inverting the vegetation index and surface temperature from satellite remote sensing imagery, while the Soil Moisture Index is obtained by inverting microwave band data from remote sensing imagery. Both reflect the drought situation from the perspectives of vegetation growth status and soil moisture status, respectively.

[0062] The model fusion employs a weighted average method, with weights preset based on the climate characteristics and crop type of the irrigated area. Specifically, the standardized precipitation index has a weight of 0.4, the vegetation water supply index has a weight of 0.3, and the soil moisture index has a weight of 0.3. The fusion calculation process is as follows: First, each individual index is normalized to the 0-1 range, where an index value closer to 1 indicates a more severe drought. Then, a comprehensive drought index is calculated according to the preset weights. This index is the drought index output by the first drought assessment model for each geographic grid unit. Ultimately, the first drought distribution information is generated, which presents the regional drought differences in space using geographic grids as units.

[0063] The second drought assessment model, generated based on irrigation behavior data, aims to reflect actual water demand in the fields. Its core is to calculate the real-time water stress index for each geographic grid cell. The calculation formula is: ,in, Represents geographic grid cells The total irrigation water consumption of all smart irrigation terminals within the grid within a set time window, with the time window set to 1 day, is obtained by summarizing the daily irrigation water consumption reported by all smart irrigation terminals within the grid, which can directly reflect the actual water replenishment situation in the field. Represents geographic grid cells The potential evapotranspiration within the same time window is calculated using the Penman-Montes formula based on the temperature, humidity, wind speed, and sunshine duration in the meteorological forecast data, and can reflect the potential water consumption of crops. Represents geographic grid cells The crop coefficients of major crops are determined by consulting FAO recommended values ​​based on crop type and growth stage, such as wheat seedling stage. Jointing period Grouting period Corn seedling stage Jointing period Maturity Rice tillering stage , booting stage Grouting period .

[0064] function This is a standardized processing function, its purpose is to eliminate the influence of different crops and growth stages, so that... It can characterize a uniform relative water shortage level. The specific processing procedure is as follows: First, calculate... and The ratio, which reflects the degree of matching between actual water replenishment and potential water demand, is then normalized to the 0-1 range, where a smaller ratio indicates a higher water demand. The closer the ratio is to 1, the more severe the water shortage; the larger the ratio, the greater the severity of the water shortage. The closer the value is to 0, the more abundant the water supply. Furthermore, when generating the second drought distribution information, spatial interpolation calibration is performed using measured data from fixed soil moisture monitoring stations. Kriging interpolation is used to interpolate discrete measured soil moisture data across the entire geographic grid, combined with... The calculation results correct for spatial differences in drought distribution and improve the spatial accuracy of the second drought distribution information.

[0065] Model parameter adjustment: To ensure that the output of the first drought assessment model is consistent with the actual field conditions, the system calculates the deviation data and dynamically adjusts the model parameters by comparing the first and second drought distribution information. The specific process is as follows:

[0066] Set the number of consecutive evaluation periods Each evaluation period is one day. Within each period, each geographic grid unit is calculated separately. of and And calculate the deviation value within that period. ( Indicates the first (a cycle). In continuous After each cycle, calculate the cumulative deviation for each geographic grid cell i. .

[0067] Preset deviation threshold , The value is determined based on the drought resistance of the crop, for grain crops. Set to 0.2; for cash crops, adjustments can be made as needed based on actual conditions. When At this time, the parameter adjustment mechanism of the first drought assessment model is triggered. The key parameters to be adjusted are: That is, the geographical grid units that affect the first drought assessment model. The weighting coefficients for calculating the drought index. The adjustment method uses proportional-integral adjustment, determining the adjustment magnitude based on the magnitude and direction of the cumulative deviation, with each adjustment not exceeding 10% to ensure model stability. For example, when... This indicates that the first drought assessment model underestimated the drought situation and needs to be improved. This enables subsequent calculations Approaching ;when This indicates that the first drought assessment model overestimated the drought situation and needs to be reduced. Similarly, Towards convergence.

[0068] The core function of the scheduling optimization engine module is to generate preventive drought relief scheduling schemes. By combining drought information, weather forecasts and engineering constraints, it achieves optimal allocation of water resources. The key lies in the calculation of predictive water-saving volume and the construction and solution of multi-objective optimization models.

[0069] Predictive water-saving calculation: Predictive water-saving is for areas where effective rainfall is forecast. The calculation, aimed at making full use of natural rainfall and reducing unnecessary irrigation water, uses the following formula: ,in, For the forecast area The effective rainfall is obtained by multiplying the forecasted rainfall by the effective rainfall coefficient. The effective rainfall coefficient is determined according to the rainfall level: 0.8 for light rain (5-10 mm of daily rainfall), 0.9 for moderate rain (10-25 mm of daily rainfall), and 1.0 for heavy rain (≥25 mm of daily rainfall). For the area The crop planting area is obtained through remote sensing monitoring data inversion or agricultural department statistical data, with an accuracy controlled within ±5%. For the area The average soil precipitation infiltration recharge coefficient is determined according to soil type: 0.3-0.5 for sandy soil, 0.5-0.7 for loam, and 0.2-0.4 for clay soil. In this embodiment, the irrigated area is mainly composed of loam. Take 0.6; For the area The theoretical crop water requirement for the current period is obtained by multiplying the reference crop evapotranspiration by the crop coefficient. The reference crop evapotranspiration is calculated using the Penman-Montes formula. For the area The irrigation water supply already scheduled for the current period is retrieved from the data warehouse and allocated to the area for the current period. The irrigation water volume was obtained.

[0070] The function `min` ensures that the anticipated water saving amount does not exceed the smaller of the available rainfall and the actual water shortage, thus preventing excessive water conservation from exacerbating the drought. For example, if the available rainfall is 1000 m³ and the actual water shortage is 800 m³, then the anticipated water saving amount is 800 m³, meaning that 800 m³ of irrigation water is reduced in this area. If the available rainfall is 600 m³ and the actual water shortage is 800 m³, then the anticipated water saving amount is 600 m³, and the remaining 200 m³ of water shortage is supplemented by other water sources.

[0071] The multi-objective optimization model takes the overall drought resistance benefit of the region as the objective function and the predictable water saving and engineering water conveyance capacity as constraints. By solving the model, the cross-regional water resource allocation scheme and specific plot irrigation adjustment instructions can be obtained.

[0072] objective function The expression is: ,in, In this embodiment, the total number of areas affected by drought is represented. Each area; For the area The weight of crop output value per unit area is determined based on crop market prices and planting benefits, for example, rice. ,wheat ,corn ; Due to the implementation of the preventive drought relief dispatch plan, in the area The expected production loss to be recovered is allocated to the region. The water volume and the initial drought level of the area are calculated as a function based on the Jensen crop water production function. ,in For the area Potential output, For crops Actual evapotranspiration during the reproductive period For crops Critical evapotranspiration during the reproductive period For the first Moisture sensitivity index during the reproductive period; In this embodiment, the total number of water sources participating in the scheduling is [number missing]. Water sources (reservoirs, groundwater well groups, reclaimed water plants); To obtain water source The unit cost coefficient for water intake is determined based on the type of water source, such as surface water (reservoir). ,groundwater Reclaimed water ; To obtain water source The total amount of water transferred out shall not exceed the water source. The upper limit of water supply capacity.

[0073] The constraints of the optimization model include:

[0074] Water demand balance constraints: That is, the total water supply from all water sources is equal to the total water demand of all areas minus the planned irrigation water supply and the anticipated water saving;

[0075] Water supply capacity upper limit constraints: ,in For water source Maximum water supply capacity of the reservoir The capacity is determined based on the reservoir capacity and inflow rate; in this example, it is 50,000 m³ / d, using a group of groundwater wells. 30,000 m³ / d, reclaimed water plant It is 20,000 m³ / d;

[0076] Flow capacity constraints of water conveyance channels: ,in To obtain water source To the area Water flow rate, The design flow rate of the water conveyance channel is 5 m³ / s for the concrete water conveyance channel and 3 m³ / s for the water conveyance pipeline in this embodiment.

[0077] Water demand reduction constraints: That is, the area The adjusted water requirement should not be less than 0 to avoid excessive downward adjustment that could lead to crop water shortage.

[0078] The multi-objective optimization model is solved using the non-dominated sorting genetic algorithm (NSGA-II), with the following parameters: population size 100, number of iterations 200, crossover probability 0.8, and mutation probability 0.05. The Pareto optimal solution set is obtained by solving the model. The system selects the optimal solution based on regional drought relief priorities and generates cross-regional water resource allocation schemes and specific irrigation adjustment instructions for each plot.

[0079] Irrigation Adjustment Command Issuance and Execution: The irrigation adjustment command generated by the scheduling optimization engine module includes the suggested irrigation start time, suggested irrigation duration, and maximum water consumption quota for the next irrigation cycle. The command is sent to the corresponding smart irrigation terminal via the IoT communication network. Upon receiving the command, the smart irrigation terminal integrates it with its locally preset irrigation program, executing the irrigation operation based on the principle that the command has higher priority than the local program. For example, if the local preset program sets the wheat area to irrigate for 2 hours at 10:00 AM daily, with a maximum water consumption quota of 30 m³ / mu, and the command requires irrigation for 1.5 hours at 9:00 AM, with a maximum water consumption quota of 25 m³ / mu, then the smart irrigation terminal will adjust the irrigation operation according to the command, executing irrigation starting at 9:00 AM and lasting for 1.5 hours, with a water consumption not exceeding 25 m³ / mu.

[0080] The human-computer interface adopts a B / S architecture design, allowing users to log in to the system via computers, mobile phones, and other terminals without installing dedicated software. The interface mainly includes the following functional modules:

[0081] Drought monitoring module: Displays the regional drought distribution in map form, using different colors to indicate drought levels, such as mild drought, moderate drought, severe drought, and extreme drought. Clicking on a geographic grid cell allows you to view the details of that cell. , Detailed information such as soil moisture content and crop type;

[0082] The scheduling plan module displays cross-regional water resource allocation plans, irrigation adjustment instructions for each area, and water supply status. Users can view detailed parameters of the plan, such as water supply volume of each water source, water transmission path, and irrigation time.

[0083] Equipment Management Module: Displays the operating status, working parameters, and fault alarm information of all smart irrigation terminals, soil moisture monitoring stations, and automatic weather stations, and supports users to remotely control the start and stop of smart irrigation terminals;

[0084] Data statistics module: generates drought trend charts, irrigation water consumption statistics reports, water resource allocation benefit analysis reports, etc., and supports data querying and export by time, region, and crop type;

[0085] System settings module: Supports users to adjust model parameters, set user permissions, configure data acquisition frequency, etc.

[0086] The interface design follows the principle of simplicity and ease of use, using blue and green, which are commonly used in the agricultural field, as the main colors. The operation process is in line with the working habits of agricultural managers, and daily management work can be completed without professional computer operation skills.

[0087] In this embodiment, closed-loop optimization is the core mechanism to ensure the accuracy of drought assessment and the effectiveness of scheduling plans, and it runs through the entire process of data collection, drought assessment, scheduling execution, and data feedback.

[0088] The specific implementation process is as follows:

[0089] After the intelligent irrigation terminal executes the irrigation adjustment command, it collects new irrigation behavior data in real time, including the adjusted irrigation start and stop time, duration, and water consumption, and feeds it back to the data warehouse.

[0090] The drought assessment model module extracts new irrigation behavior data from the data warehouse and combines it with meteorological forecast data, remote sensing monitoring data and measured data to generate new secondary drought distribution information;

[0091] The drought assessment model module compares the new second drought distribution information with the expected drought mitigation trend of the previous dispatch plan. The expected drought mitigation trend is calculated based on the water replenishment amount of the dispatch plan, such as the expected drought index for a certain area. Decreased by 0.1;

[0092] If the new second drought distribution information shows that the drought relief trend is in line with expectations, then the current first drought assessment model parameters and the optimization weights of the scheduling optimization engine will be maintained.

[0093] If the new second drought distribution information deviates from the expectation, an adjustment mechanism is triggered: if the deviation is due to the accuracy of drought assessment, the parameters of the first drought assessment model are recalibrated; if the deviation is due to the rationality of the scheduling plan, the optimization weights in the scheduling optimization engine are adjusted, the scheduling plan is regenerated and sent to the smart irrigation terminal.

[0094] Repeating the above process forms a closed-loop optimization of "data collection - drought assessment - scheduling execution - data feedback - model / parameter adjustment", enabling the system to continuously adapt to changes in actual field conditions and continuously improve the accuracy of drought assessment and the effectiveness of scheduling plans.

[0095] In this embodiment, the data communication between the data perception and execution layer and the intelligent analysis and management layer adopts a combination of multiple communication technologies to ensure the stability and real-time performance of data transmission.

[0096] During data transmission, data with high real-time requirements, such as irrigation behavior data and measured data, will be transmitted using 4G or 5G technology, with transmission latency controlled to within 1 second. Non-real-time data, such as weather forecast data and remote sensing monitoring data, can be transmitted using narrowband IoT or 4G technology, with the transmission cycle determined based on the data update frequency. Data transmission employs encryption protocols (such as HTTPS) to encrypt the data and prevent theft or tampering. A retransmission mechanism is also employed; if data transmission fails, the system automatically retransmits, up to three times, ensuring data transmission reliability.

[0097] The intelligent analysis and management layer is deployed on a cloud computing platform, possessing powerful data processing and storage capabilities. It supports the simultaneous reception of data from multiple smart irrigation terminals and monitoring stations, and can dynamically expand computing resources based on data volume to prevent system lag or crashes due to data spikes. The cloud computing platform also features data backup capabilities, employing an off-site backup strategy to back up important data to servers in different regions, preventing data loss.

[0098] In this embodiment, the workflow can be divided into the following steps in chronological order:

[0099] Data collection phase (00:00-23:59 daily): Smart irrigation terminals collect irrigation behavior data in real time and report it every hour; fixed soil moisture monitoring stations report soil moisture content data every 2 hours; automatic weather stations report meteorological measurement data every 1 hour; meteorological forecast data is obtained from external platforms at 08:00 and 20:00 daily; remote sensing monitoring data is obtained every 3 days.

[0100] Data storage and preprocessing stage (24:00 daily): The data warehouse receives all collected data, performs format standardization and integrity verification, marks and completes abnormal data, and then stores it in categories.

[0101] Drought assessment phase (01:00-02:00 daily): The drought assessment model module runs the first drought assessment model, generating first drought distribution information based on meteorological forecast data and remote sensing monitoring data; at the same time, it runs the second drought assessment model, generating second drought distribution information based on irrigation behavior data and measured data; the deviation data of the two are compared, and if the deviation continues to exceed the threshold, the parameters of the first drought assessment model are adjusted.

[0102] Dispatch scheme generation phase (02:00-03:00 daily): The dispatch optimization engine module receives drought distribution information and trend prediction information, combines it with future precipitation forecasts in meteorological forecasts, calculates the predictable water-saving amount for each area, constructs a multi-objective optimization model, and solves the preventive drought relief dispatch scheme, including cross-regional water resource allocation schemes and irrigation adjustment instructions.

[0103] Command issuance and execution phase (03:00-04:00 daily): The scheduling optimization engine module issues irrigation adjustment commands to the corresponding smart irrigation terminals; after receiving the commands, the smart irrigation terminals integrate local preset programs and execute the adjusted irrigation operations.

[0104] Closed-loop optimization phase (daily cycle): The smart irrigation terminal feeds back new irrigation behavior data, the drought assessment model module monitors new secondary drought distribution information, determines whether it matches expectations, and adjusts model parameters or optimization weights if necessary to complete closed-loop optimization.

[0105] The agricultural drought multi-scale monitoring, early warning, and drought-resistant resource allocation system implemented in this embodiment achieves the following beneficial effects in irrigation system control through the collaborative work of the data perception and execution layer and the intelligent analysis and management layer, as well as the continuous operation of the closed-loop optimization mechanism:

[0106] It improves the accuracy of drought assessment. By integrating meteorological, remote sensing, irrigation behavior and measured data, a dual drought assessment model is constructed and parameters are dynamically adjusted. This effectively reduces the bias of assessment from a single data source, makes drought distribution information more consistent with the actual situation in the field, and provides a reliable basis for irrigation scheduling decisions.

[0107] It improves the efficiency of water resource allocation, makes full use of natural rainfall resources by calculating predictive water-saving amounts, and reduces unnecessary irrigation water use; at the same time, through a multi-objective optimization model, it realizes the rational allocation of water resources across regions, reduces water waste, and improves the utilization efficiency of limited water resources.

[0108] It enhances the predictability and targeting of irrigation scheduling, generates preventive scheduling plans based on future weather forecasts and drought trend predictions, and avoids the lag of traditional irrigation scheduling; it formulates personalized irrigation adjustment instructions according to the drought conditions and crop needs of different areas, thereby improving the targeting of irrigation scheduling.

[0109] The system achieves dynamic self-adaptation, and the closed-loop optimization mechanism enables the system to continuously adapt to changes in actual field conditions. Whether it is the change of crop growth period, the fluctuation of meteorological conditions, or the difference in soil properties, the system can maintain a good operating state by adjusting model parameters or scheduling strategies, thereby improving the applicability and stability of the system.

[0110] It reduces the labor costs of irrigation management. Through automated data collection, drought assessment and scheduling, it reduces the workload of manual inspection, data recording and manual control, improves the efficiency of irrigation management and reduces management costs.

[0111] This implementation, through reasonable module design, data fusion, and closed-loop optimization, achieves integrated coordination between drought monitoring and early warning and drought relief resource allocation. It has good practicality and operability and can meet the actual needs of precise control of agricultural irrigation.

[0112] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A multi-scale monitoring, early warning, and drought-resistant resource allocation system for agricultural drought, characterized in that, The system consists of: a data perception and execution layer, and an intelligent analysis and management layer; The data perception and execution layer includes multiple intelligent irrigation terminals deployed in the irrigation area. The intelligent irrigation terminals are configured to perform irrigation operations and collect and report the start and stop time, duration and water consumption data of each irrigation operation to form irrigation behavior data. The intelligent analysis and management layer includes a data warehouse, a drought assessment model module, a scheduling optimization engine module and a human-computer interaction interface. The data warehouse is configured to receive and store irrigation behavior data from the data perception and execution layer, as well as weather forecast data and remote sensing monitoring data from external sources. The drought assessment model module is configured to: run a first drought assessment model based on the meteorological forecast data and the remote sensing monitoring data to generate first drought distribution information; generate second drought distribution information reflecting actual water demand based on the irrigation behavior data; compare the first drought distribution information with the second drought distribution information and deviation data; and automatically adjust the parameters of the first drought assessment model when the deviation data continuously exceeds a preset threshold, so that the drought distribution information output by the adjusted first drought assessment model converges to the second drought distribution information. The scheduling optimization engine module is configured to: receive drought distribution information and trend prediction information output by the drought assessment model module, and combine the future precipitation forecast in the meteorological forecast data to calculate the predictable water-saving amount for areas with rainfall forecasts. Taking the overall drought resistance benefit of the region as the objective function, the predictable water-saving amount as the constraint for new water sources, and the engineering water conveyance capacity as the physical constraint, a multi-objective optimization model is constructed and solved to generate a preventive drought resistance scheduling scheme that includes cross-regional water resource allocation schemes and specific plot irrigation adjustment instructions.

2. The agricultural drought multi-scale monitoring, early warning, and drought relief resource allocation system according to claim 1, characterized in that, The scheduling optimization engine module is also configured to send the irrigation adjustment command to the corresponding smart irrigation terminal; The intelligent irrigation terminal is also configured to: adjust the subsequent irrigation operations according to the received irrigation adjustment instruction, and generate new irrigation behavior data to be fed back to the data warehouse, thereby forming a closed-loop optimization between drought assessment and irrigation scheduling; The drought assessment model module is configured to generate the second drought distribution information and adjust the model parameters through the following mathematical relationships, for each preset geographic grid unit. Calculate the real-time water stress index of the geographic grid cell i. The real-time water stress index The calculation formula is: ； in, Represents the geographic grid cell The total irrigation water consumption of all smart irrigation terminals within a set time window. Represents the geographic grid cell The potential evapotranspiration within the same time window is calculated from the aforementioned meteorological forecast data. Represents the geographic grid cell Crop coefficients of major crops in the region, function Indicates to , and The standardized mapping relationship is used to eliminate the influence of different crops and growth stages, so that... It can characterize a uniform relative degree of water scarcity; The first drought assessment model outputs data for the geographic grid unit. The drought index is recorded as The system is in continuous Within each evaluation period, the calculation of each of the geographic grid units is performed. Cumulative deviation: ,when At that time, it triggers the impact on the geographic grid units in the first drought assessment model. Key parameters of the drought index The adjustment is aimed at improving the calculation of subsequent cycles. Approaching .

3. The agricultural drought multi-scale monitoring, early warning, and drought relief resource allocation system according to claim 1, characterized in that, In the scheduling optimization engine module, the predictive water saving... For areas where effective rainfall is forecast Calculations were performed to determine the amount of water saved through this predictive method. The mathematical expression is: ； in, For the forecast area The effective rainfall, For the area The area of ​​crop planting, For the area The average soil precipitation infiltration recharge coefficient, For the area The theoretical crop water requirement for the current period is obtained by multiplying the reference crop evapotranspiration by the crop coefficient. For the area The function represents the amount of irrigation water already supplied in the current time period. This indicates taking the smaller of the two values.

4. The agricultural drought multi-scale monitoring, early warning, and drought relief resource allocation system according to claim 1, characterized in that, The intelligent irrigation terminal includes a controller, a solenoid valve electrically connected to the controller, a water flow sensor, and an Internet of Things (IoT) communication unit. The water flow sensor is configured to measure the volume of water flowing through the solenoid valve. The controller is configured to control the opening and closing of the solenoid valve according to a local preset program or instructions received from the intelligent analysis and management layer, and to upload irrigation behavior data including terminal identification code, timestamp, valve status, irrigation duration and water consumption through the Internet of Things communication unit.

5. The agricultural drought multi-scale monitoring, early warning, and drought-resistant resource allocation system according to claim 1, characterized in that, The first drought assessment model running in the drought assessment model module is a comprehensive assessment model that integrates the meteorological drought index and the remote sensing drought index. The remote sensing drought index includes at least a vegetation water supply index or a soil moisture index derived from satellite remote sensing data.

6. The agricultural drought multi-scale monitoring, early warning, and drought-resistant resource allocation system according to claim 1, characterized in that, The multi-objective optimization model constructed by the scheduling optimization engine module has an objective function. Simultaneously minimizing expected water shortage losses and scheduling costs, expressed as: ； in, This represents the total number of areas affected by the drought. For the area The weight of crop output value per unit area In order to implement the aforementioned preventive drought relief scheduling plan in the region The expected production loss to be recovered is allocated to the region. A function of water volume and the initial drought level in the area. The total number of water sources participating in the scheduling, To obtain water source The unit cost coefficient for water intake, To obtain water source Total water volume transferred out; The constraints of the optimization model include: water demand balance constraints for all areas, upper limit constraints on the water supply capacity of each water source, flow capacity constraints for water transmission channels, and predictive water conservation measures. As a region Water demand reduction constraints.

7. The agricultural drought multi-scale monitoring, early warning, and drought-resistant resource allocation system according to claim 2, characterized in that, The closed-loop optimization is specifically manifested as follows: after each output of drought distribution information, the drought assessment model module will continuously monitor the new second drought distribution information generated by the subsequently generated irrigation behavior data; If the new second drought distribution information shows that the drought relief trend is consistent with the previous scheduling plan, then the current model parameters will be maintained. When a deviation occurs, it triggers a recalibration of the first drought assessment model or an adjustment of the optimization weights in the scheduling optimization engine.

8. The agricultural drought multi-scale monitoring, early warning, and drought-resistant resource allocation system according to claim 1, characterized in that, The data warehouse is also configured to receive and store measured data from fixed soil moisture monitoring stations and automatic weather stations deployed in the irrigation area; When generating the second drought distribution information, the drought assessment model module also integrates the measured data from the fixed soil moisture monitoring station for spatial interpolation calibration.

9. The agricultural drought multi-scale monitoring, early warning, and drought-resistant resource allocation system according to claim 1, characterized in that, The irrigation adjustment instructions generated by the scheduling optimization engine module specifically include a suggested irrigation start time, a suggested irrigation duration, or a suggested maximum water consumption quota for the next irrigation cycle. The intelligent irrigation terminal is configured to integrate the irrigation adjustment command with a locally preset irrigation program, and execute the adjusted irrigation operation in a higher priority manner.

10. The agricultural drought multi-scale monitoring, early warning, and drought-resistant resource allocation system according to claim 1, characterized in that, The intelligent analytics and management layer is deployed on a cloud computing platform; The data perception and execution layer and the intelligent analysis and management layer communicate with each other through at least one of narrowband Internet of Things, fourth-generation mobile communication technology, or fifth-generation mobile communication technology.

Images