Plateau digital twin irrigation district dynamic construction system and method
By collecting, cleaning, and standardizing multi-source data from plateau irrigation areas, a plateau irrigation area model was established and optimized. Combined with AI-driven and real-time data correction, the problem of insufficient applicability of plateau irrigation area models was solved, realizing high-precision simulation and accurate pre-showing and visualization of irrigation schemes, supporting efficient management and ecological protection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- POWERCHINA BEIJING ENG CORP
- Filing Date
- 2025-11-17
- Publication Date
- 2026-07-31
AI Technical Summary
The existing digital twin technology is not applicable enough to plateau irrigation areas. The model simulation accuracy is insufficient and its actual applicability is inadequate. The ability to integrate multi-source data needs to be improved, making it difficult to support the construction of high-fidelity models and the needs of refined management in plateau irrigation areas.
By collecting multi-source datasets from plateau irrigation areas, performing data cleaning and standardization, extracting features, establishing plateau irrigation area models, introducing AI data-driven model parameter optimization, and monitoring data integration into the model in real time for dynamic correction, combined with irrigation schemes for simulation and prediction, and achieving three-dimensional visualization output.
It improves the simulation accuracy and multi-source data fusion capability of the plateau irrigation area model, adapts to the special plateau environment, supports efficient management and ecological protection, provides high-fidelity consistency between the virtual model and the physical entity, and realizes accurate pre-simulation and visualization of irrigation schemes.
Smart Images
Figure CN121615064B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of dynamic construction technology of plateau irrigation areas, specifically relating to a dynamic construction system and method for plateau digital twin irrigation areas. Background Technology
[0002] As a vital ecological barrier and distinctive agricultural production area in my country, the stable operation and efficient management of irrigation districts in plateau regions are crucial for ensuring regional food security and maintaining ecosystem balance. They serve as core infrastructure supporting local agricultural development and ecological protection. Currently, digital twin technology has been applied in conventional scenarios such as plain irrigation districts. By integrating multi-dimensional data including geospatial, hydrological, and meteorological data to construct simulation models, preliminary simulations and management optimization of the irrigation process have been achieved, providing a mature technical framework and practical experience for the digital transformation of irrigation districts. However, existing digital twin technology still has significant limitations due to the unique characteristics of plateau irrigation districts: (1) Existing models are mostly based on climate and soil characteristics parameters in plains areas, which fail to fully adapt to the unique environmental characteristics of plateaus such as low temperature, low air pressure and complex soil texture, resulting in insufficient simulation accuracy and practical applicability of the models in plateau scenarios. (2) The ability to integrate multi-source data needs to be improved. The existing system's integration of geospatial data, hydrological and meteorological data, and crop data is mostly limited to simple overlay. It lacks a parameter optimization and completion mechanism for areas with scarce data in the plateau region, making it difficult to support the construction of high-fidelity models and meet the needs of refined management in plateau irrigation areas. In view of this, the present invention is hereby proposed. Summary of the Invention
[0003] To address the aforementioned technical problems in existing technologies, this invention provides a dynamic construction system and method for digital twin irrigation districts in plateau regions, thus solving the problem of insufficient applicability of existing technologies in plateau irrigation districts.
[0004] To achieve the above objectives, the technical solution of the present invention is as follows: Firstly, a dynamic construction system for a plateau digital twin irrigation district includes: Data processing end: Used to collect multi-source datasets from plateau irrigation areas, clean the data of each sub-data item in the collected multi-source datasets, standardize the data of each sub-data item in the cleaned multi-source datasets, and extract features from the standardized multi-source data; Model building end: used to build a plateau irrigation area model based on the collected multi-source datasets and feature-extracted data of the plateau irrigation area, introduce data-driven model to optimize the parameters of the plateau irrigation area model, and integrate real-time monitoring data of the plateau irrigation area into the plateau irrigation area model to dynamically correct the simulation results, so that the digital twin is consistent with the physical entity of the plateau irrigation area. Model pre-simulation terminal: Used to import the set irrigation scheme into the established digital twin model of the plateau irrigation area for simulation and prediction of the impact of irrigation on the target area environment of the plateau irrigation area. Combined with the environmental correlation data of the plateau irrigation area, it simulates the impact of irrigation on specific land types in the region. The simulation process and results are compared with the measured data. The simulation results are output through visualization technology, and the relevant data are stored.
[0005] Furthermore, the multi-source dataset for the plateau irrigation area includes: a geospatial dataset, a hydro-meteorological dataset, and a plateau crop dataset; the data processing terminal includes a data acquisition module and a data preprocessing module. The data acquisition module includes: a geospatial acquisition unit, a hydro-meteorological acquisition unit, and a plateau crop acquisition unit; The geospatial data acquisition unit is used to collect geospatial datasets, which provide a basis for locating the boundary of the plateau irrigation area model. The geospatial datasets include soil data, topographic data, and vegetation coverage. The hydro-meteorological data acquisition unit is used to collect hydro-meteorological datasets. These datasets provide a data source for the dynamic correction of water flow simulation and real-time fusion in the plateau irrigation area model. The hydro-meteorological datasets include rainfall data, temperature data, and irrigation area runoff data. The plateau crop data collection unit is used to collect plateau crop datasets, which provide time-series data for the ecological effect assessment of crop growth simulation and environmental prediction in plateau irrigation district models. The plateau crop datasets include crop leaf area index and crop water requirement.
[0006] Furthermore, the data preprocessing module includes a data cleaning unit and a feature extraction unit; The data cleaning unit cleans the basic dataset by removing outliers and normalizes all sub-data items in the basic dataset to the range of [0, 1] according to the normalization formula. The feature extraction unit is used to extract features from geospatial datasets and hydro-meteorological datasets.
[0007] Furthermore, the steps for feature extraction from geospatial datasets and hydro-meteorological datasets include: Soil moisture data were obtained by using the moving average method based on a geospatial dataset. The specific calculation formula is as follows:
[0008] in, To adjust the sliding window size, For time-series indexes, For the time index of the target smoothing moment, Here is the measured soil moisture data at time i; Calculations are performed based on topographic data to obtain surface slope data. The specific calculation formula is as follows:
[0009] in, This is the arctangent function, used to convert the output value within the immediately following parentheses to radian format. This represents the rate of change of terrain height in the horizontal direction. The horizontal and vertical rates of change of terrain height. This is the conversion factor for radian angles; Rainfall in the blank monitoring area was calculated using the Thiessen polygon formula based on hydrological and meteorological data. The specific calculation formula is as follows:
[0010] in, The number of blank monitoring areas included in the calculation. Let the area of the Thiessen polygon be the rainfall amount corresponding to the i-th blank monitoring area. The total Thiessen polygon area represents the total rainfall in the blank monitoring area. Let be the measured rainfall at the i-th monitoring station. This is an index for monitoring sites.
[0011] Furthermore, the model building module includes: a plateau irrigation area model building module, an AI model coupling module, and a real-time data fusion module; The plateau irrigation district model establishment module includes a plateau basic model establishment unit, a hydrological model establishment unit, and a plateau crop model establishment unit. The plateau basic model building unit is used to construct the model topology based on the geospatial dataset, which clarifies the spatial range of the model and the interaction logic of its internal components, and provides a basic framework for subsequent sub-model coupling. The hydrological model building unit is used to calculate the channel water conveyance based on surface slope data and hydrological and meteorological datasets. The plateau crop model establishment unit calculates crop biomass based on the plateau crop dataset.
[0012] Furthermore, the water delivery volume of the channel is obtained. The calculation formula is:
[0013] in, The cross-sectional area of the channel through which water flows. For hydraulic radius, The roughness coefficient is Manning's coefficient. Obtaining crop biomass The calculation formula is:
[0014] in, The maximum biomass of the crop. It is a natural constant. All are weighting coefficients. Accumulated temperature during the crop's growing season.
[0015] Furthermore, the AI model coupling module includes an AI digital driving unit and a multi-model coupling unit; The AI digital driving unit is used to determine sensitive parameters based on the plateau irrigation area model through sensitivity analysis, and to construct an AI parameter optimization model based on the preprocessed feature data and the simulation results of the plateau irrigation area model. The multi-model coupling unit is used to substitute the optimized parameters output by AI into the calculation formula of the plateau irrigation area model, and at the same time, it uses SHAP value to analyze the influence contribution of AI parameters, thereby improving the interpretability of the model.
[0016] Furthermore, the specific steps for building an AI parameter optimization model include: Obtain a set of feature data stored in the database and the simulation results of the plateau irrigation area model, and compare them with the current time according to the timestamp. Group the comparison results from small to large and use the labeled results as the sample set. The sample set was divided into a 70% training set, a 15% test set, and a 15% validation set. Based on the sample set, an original AI parameter optimization model was built. The original AI parameter optimization model is iterated continuously until the required confidence level is reached and then output, thus obtaining the AI parameter optimization model. Transfer learning is used to transfer AI model parameters from areas with abundant data to areas with scarce data, calculate the feature similarity between the two areas, and then correct the AI output parameters in areas with scarce data based on the similarity weighting, thus solving the parameter calibration problem caused by insufficient data in some plateau areas.
[0017] Furthermore, the real-time data fusion module includes a dynamic data import unit and an algorithm import unit; The dynamic data import unit is used to import soil moisture data, rainfall data and crop leaf area index, and sets the assimilation frequency and triggering conditions. The algorithm introduction unit is used to calculate the Karman gain based on the error range of the coupled model's predicted state and the error range of the observed data, and to fuse and correct the coupled model's predicted state with the observed data to generate the optimal model state at the current moment.
[0018] Furthermore, the model pre-simulation terminal includes an environment prediction module and a rendering output module; The environmental prediction module includes a scheme setting unit and a scheme simulation unit; The scheme setting unit is used to set different irrigation schemes based on plateau crop datasets and hydrological and meteorological datasets to cover different management needs, and at the same time provide direction for scheme derivation; The scheme deduction unit is used to deduce the scheme generated by the scheme setting unit at the current time based on the optimal model state of the scheme, and obtain the scheme with the highest score. The rendering output module includes a rendering processing unit and a data comparison unit; The rendering processing unit is used to convert the input data into a three-dimensional rendering adaptation format based on the dataset collected by the data acquisition module, the current optimal model state output by the model building end, and the irrigation scheme deduction results output by the environmental prediction module. The input data is then input into the current optimal model state and rendered on a large scale using a cloud rendering engine. At the same time, a dual-layer visualization design of surface and underground is used to display the three-dimensional visualization model data in a three-dimensional way. The data comparison unit is used to overlay and compare real-time data and predicted data, realize time-sliding interaction and difference calculation and display, and establish a database based on system data for storage.
[0019] Furthermore, based on the schemes generated by the scheme setting unit, the steps to obtain the highest-scoring scheme include: The downstream runoff forecast is obtained by calculating based on the current optimal model state. The specific calculation formula is as follows:
[0020] in, This represents the total irrigation water consumption of the irrigation district. This refers to the amount of loss due to evaporation in the channel. This represents the total deep seepage in the irrigation area. Effective rainfall; by quantifying the balance of water input, consumption and output in the irrigation district, the remaining ecological flow downstream under different irrigation schemes is predicted; Soil salinity predictions are obtained based on the current optimal model state and crop water requirement. The specific calculation formula is as follows:
[0021] in, Initial soil salinity, This represents the initial soil moisture content. To rinse the irrigation water volume, The rinsing efficiency coefficient is... This refers to the amount of water used for a single irrigation. For crop water requirements, To predict the time step; The predicted groundwater level depth is calculated based on the current optimal model state and hydrological and meteorological dataset. Based on the current optimal model state, the synergistic impact of irrigation schemes on crop productivity and ecology is calculated, and the predicted net primary productivity reflects the amount of carbon fixed by crop photosynthesis. The predicted downstream runoff, soil salinity, groundwater level depth, and net primary productivity are packaged to obtain the irrigation scheme simulation results, which are then sent to the rendering output module.
[0022] Secondly, a method for dynamically constructing a digital twin irrigation district in a plateau region includes: S1. Collect geospatial datasets, hydro-meteorological datasets, and plateau crop datasets to construct a multi-source dataset; S2. Perform data cleaning and feature extraction on the multi-source dataset; S3. Based on multi-source datasets and feature-extracted data, establish a plateau irrigation area model; S4. Based on the plateau irrigation area model, a coupled model is obtained by integrating AI data-driven approaches. S5. Based on the coupled model, real-time data is incorporated for correction to obtain the optimal model state at the current moment; S6. Based on the current optimal model state and irrigation scheme, perform a deduction to obtain the irrigation scheme deduction results; S7. Perform 3D rendering based on the irrigation scheme simulation results and store the system data.
[0023] Compared with existing technologies, the present invention provides a dynamic construction system and method for a plateau digital twin irrigation area. The system includes: a data processing end, a model building end, and a model pre-simulation end. The data processing end collects geospatial datasets, hydrological and meteorological datasets, and plateau crop datasets of the plateau irrigation area. It cleans and normalizes the sub-data items in each dataset by filling in missing values and extracts features from the datasets. The model building end constructs a plateau irrigation area model based on the collected datasets and feature-extracted data. It introduces AI data-driven model optimization mechanism and model parameters, and integrates real-time monitoring data into the model to dynamically correct the simulation results, enabling the digital twin to be consistent with the model. The physical entities of the plateau irrigation area remain consistent. The model pre-simulation end imports the set irrigation scheme into the established digital twin model for simulation, simulating the impact of irrigation on the downstream ecological environment and saline-alkali land. The simulation process and results are compared with measured data, and three-dimensional visualization output is achieved through cloud rendering technology, and relevant data is stored. The method includes the steps of data acquisition and processing, plateau irrigation area model construction and optimization, irrigation scheme simulation and result output. This invention effectively adapts to the special environment of plateau low temperature and low air pressure, improves the multi-source data fusion capability and model simulation accuracy, and provides technical support for the stable operation, efficient management and ecological protection of plateau irrigation areas. Attached Figure Description
[0024] Figure 1 This is an architecture diagram of the dynamic construction system for a plateau digital twin irrigation district provided in an embodiment of the present invention; Figure 2 An architecture diagram of the data processing end of the dynamic construction system for a digital twin irrigation district in plateau provided in an embodiment of the present invention; Figure 3 This is an architecture diagram of the model establishment end of the dynamic construction system for a digital twin irrigation district in plateau provided in an embodiment of the present invention. Figure 4 This is an architecture diagram of the model pre-simulation terminal of the dynamic construction system for digital twin irrigation districts in plateau regions provided in an embodiment of the present invention; Figure 5 A flowchart of a method for dynamically constructing a digital twin irrigation district in a plateau region, provided in an embodiment of the present invention. Detailed Implementation
[0025] The technical solution of the present invention will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are not all embodiments of the present invention. All other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.
[0026] It should be noted that, unless otherwise specifically stated, the relative arrangement and numerical expressions of the components and steps described in these embodiments should not be construed as limiting the scope of the invention.
[0027] The following description of exemplary embodiments is merely illustrative and is not intended to limit the invention or its application or use in any way. Techniques, methods, and apparatus known to those skilled in the art may not be discussed in detail herein, but where applicable, such techniques, methods, and apparatus should be considered part of this specification.
[0028] Example 1 See Figure 1 , Figure 1 This is an architecture diagram of a dynamic construction system for a digital twin irrigation district in a plateau region proposed in this invention. The system is set up in a cloud server. In terms of implementation, it can be used as one or more service devices, or as an application installed in the cloud (such as a mobile service operator's server, server cluster, etc.), or it can be developed into a website.
[0029] Based on the implemented functions, the dynamic construction system for a plateau digital twin irrigation district based on multi-source data fusion may include a geospatial data acquisition module, a hydrological and meteorological data acquisition module, a plateau crop data acquisition module, a data preprocessing module, a plateau model building module, an AI model coupling module, a real-time data fusion module, an environmental prediction module, and a rendering output module. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by an electronic device's processor and perform a fixed function, stored in the electronic device's memory. Specifically, they may include: M1, Data Processing Terminal: Used to collect geospatial datasets, hydro-meteorological datasets, and plateau crop datasets. It performs data cleaning on all sub-data items in these datasets by filling in missing values, and normalizes all sub-data items to the range [0, 1] using a normalization formula. It also performs feature extraction on the data from these datasets. (See also...) Figure 2 Specifically, it includes: M11, Data Acquisition Module: Includes geospatial acquisition unit, hydro-meteorological acquisition unit and plateau crop acquisition unit; M111, Geospatial Data Acquisition Unit, collects geospatial datasets to provide a basis for boundary positioning of the plateau irrigation area model. The geospatial dataset includes soil data, topographic data, and vegetation coverage. M112, Hydrological and Meteorological Data Acquisition Unit, collects hydrological and meteorological datasets to provide a data source for dynamic correction of water flow simulation and real-time fusion in plateau irrigation area models. Hydrological and meteorological data include rainfall data, temperature data, and irrigation area runoff data. M113, Plateau Crop Data Collection Unit, collects plateau crop datasets to provide time-series data for ecological effect assessment of crop growth simulation and environmental prediction in plateau irrigation area models. The plateau crop dataset includes crop leaf area index and crop water requirement.
[0030] M12, Data Preprocessing Module: Includes a data cleaning unit and a feature extraction unit; M121, Data Cleaning Unit, is used to clean the basic dataset by removing outliers and normalize all sub-data items in the basic dataset to the range of [0, 1] according to the normalization formula; Outliers are removed, for example, when crop water requirement is negative; the basic datasets include geospatial datasets, hydro-meteorological datasets, and plateau crop datasets; the specific expression of the normalization formula is as follows:
[0031] in, Normalized value Any sub-data item of the basic data, This represents the historical maximum value of any given sub-data item. This is the historical minimum value of this arbitrary sub-data item.
[0032] M122, Feature Extraction Unit, is used to extract features from the geospatial dataset and the hydro-meteorological dataset; further, the steps for feature extraction from the geospatial dataset and the hydro-meteorological dataset include: M1221. Soil moisture data were obtained by using the moving average method based on a geospatial dataset. The specific calculation formula is as follows:
[0033] in, To adjust the sliding window size, For time-series indexes, For the time index of the target smoothing moment, The soil moisture data is the measured data at time i. The calculated soil moisture data provides a stable input for the plateau irrigation area model, avoiding frequent changes in the simulated irrigation water demand due to data fluctuations, and ensuring the continuity of the irrigation plan. M1222. Calculations are performed based on terrain data to obtain surface slope data. The specific calculation formula is as follows:
[0034] in, This is the arctangent function, used to convert the output value within the immediately following parentheses to radian format. This represents the rate of change of terrain height in the horizontal direction. The horizontal and vertical rates of change of terrain height. This is the conversion factor for radian angle; surface slope calculation is used to convert the spatial changes in terrain height into intuitive slope values. It is a key parameter for calculating surface runoff, ensuring that the simulation of irrigation water distribution and loss conforms to actual terrain characteristics. M1223. Based on hydrological and meteorological data, the Thiessen polygon formula was used to calculate the rainfall in the blank monitoring area. The specific calculation formula is as follows:
[0035] in, The number of blank monitoring areas included in the calculation. Let the area of the Thiessen polygon be the rainfall amount corresponding to the i-th blank monitoring area. The total Thiessen polygon area represents the total rainfall in the blank monitoring area. Let be the measured rainfall at the i-th monitoring station. This serves as an index for monitoring stations; the calculation of rainfall in blank monitoring areas avoids data gaps caused by uneven rainfall distribution in plateau irrigation areas, which would lead to deviations in the field water accumulation and infiltration volume simulated by the plateau irrigation area model.
[0036] M2, Model Building End: Used to build a plateau irrigation area model based on collected geospatial datasets, hydrological and meteorological datasets, plateau crop datasets, and feature-extracted data. It introduces AI data-driven modeling to optimize the parameters of the mechanistic model, improving the model's interpretability and prediction accuracy. Simultaneously, real-time monitoring data is continuously integrated into the model to dynamically correct simulation results, enabling the digital twin to evolve synchronously and maintain consistency with the physical entity. (See also...) Figure 3 Specifically, it includes: M21, Plateau Irrigation District Model Establishment Module: This module includes a plateau basic model establishment unit, a hydrological model establishment unit, and a plateau crop model establishment unit. The sub-models of these units are coupled through a data interface, and historical irrigation district data is used to verify the model's confidence level. If the error between the simulated and measured soil moisture values is less than 3%, and the deviation between the simulated and measured yield values is less than 10%, the model passes and outputs the plateau irrigation district model. If the model fails, the parameters in the sub-models are adjusted and recalculated until the required confidence level is achieved. Specifically, this includes: M211, the plateau basic model building unit, is used to construct the model topology based on geospatial datasets. It clarifies the spatial range of the model and the interaction logic of its internal components, providing a basic framework for subsequent sub-model coupling. M212, the hydrological model building unit, is used to calculate the channel water conveyance based on surface slope data and hydrological and meteorological datasets. The specific calculation formula is as follows:
[0037] in, The cross-sectional area of the channel through which water flows. For hydraulic radius, The roughness coefficient is Manning's coefficient; the channel water conveyance is combined with the physical characteristics of the channel and the terrain conditions to simulate the water conveyance capacity of the water conservancy facility and provide the water inflow boundary conditions for field water flow. M213, the plateau crop model establishment unit, is calculated based on the plateau crop dataset to obtain crop biomass. The specific calculation formula is as follows:
[0038] in, The maximum biomass of the crop. It is a natural constant. All are weighting coefficients. Accumulated temperature during the crop growth period; crop biomass is directly related to crop yield and is the core basis for evaluating the economic benefits of irrigation schemes in plateau irrigation areas; M22, AI Model Coupling Module: Includes AI digital driving unit and multi-model coupling unit; The M221 AI digital drive unit is used to determine sensitive parameters based on a plateau irrigation area model using sensitivity analysis, and to construct an AI parameter optimization model based on preprocessed feature data and simulation results from the plateau irrigation area model. Further, the specific steps for determining sensitive parameters based on a plateau irrigation area model using sensitivity analysis and constructing an AI parameter optimization model based on preprocessed feature data and simulation results from the plateau irrigation area model include: M2211. Obtain a set of feature data stored in the database and the simulation results of the plateau irrigation area model, and compare them with the current time according to the timestamp. Group the comparison results from smallest to largest and label them accordingly. The labeling results are as follows: The labeled results are used as the sample set; M2212. Divide the sample set into 70% training set, 15% test set and 15% validation set, and build an original AI parameter optimization model based on the sample set. M2213. Iterate the original AI parameter optimization model until the required confidence level is reached and then output to obtain the AI parameter optimization model. M2214. Transfer learning is used to transfer the AI model parameters from areas with sufficient data to areas with scarce data, calculate the feature similarity between the two areas, and then correct the AI output parameters in areas with scarce data based on the similarity weighting, thus solving the parameter calibration problem caused by insufficient data in some areas of the plateau. M222, a multi-model coupling unit, is used to substitute the optimized parameters output by AI into the calculation formula of the plateau irrigation area model, so as to realize the deep coupling between AI and plateau irrigation area model. At the same time, SHAP value is used to analyze the influence contribution of AI parameters and improve the interpretability of the model. During model validation, real-time monitoring data is used to validate the coupled model. If the error between the simulated and measured soil moisture values decreases from less than 3% in the plateau irrigation area model to less than 2.5% in the coupled model, the model meets the standard and is output as a coupled model. Otherwise, new monitoring data is added to train the AI parameter optimization model, and the AI parameter optimization model is iteratively trained until the model parameters meet the standard, ensuring the long-term accuracy of the coupled model in the plateau environment.
[0039] M23, Real-time Data Fusion Module: Includes dynamic data import unit and algorithm import unit; M231, the dynamic data import unit, is used to import soil moisture data, rainfall data, and crop leaf area index. These three data directly reflect the current state of the irrigation area and provide a basis for model correction. At the same time, the assimilation frequency and triggering conditions are set to adapt to the characteristics of sudden rainfall and rapid changes in soil moisture in the plateau, thus avoiding the accumulation of model bias. M232, the algorithm introduction unit, is used to calculate the Karman gain based on the error range of the predicted state and the error range of the observed data from the coupled model. When the error in the observed data is small, the weights are tilted towards the observed data; when the error in the coupled model prediction is small, the weights are tilted towards the coupled model prediction, avoiding bias caused by relying solely on the model or observation, and ensuring that the correction is more in line with reality; at the same time, the coupled model prediction state and the observed data are fused and corrected. If the Karman gain is tilted towards the observed data, the prediction state is adjusted significantly using the observed data; if the Karman gain is tilted towards the coupled model prediction, the prediction state is finely adjusted. Through fusion correction, the optimal model state at the current moment is generated. This state includes both the physical mechanism logic of the model and the real-time actual situation of the physical irrigation area, solving the model simulation deviation problem caused by parameter drift and sudden environmental changes. After correction, based on the deviation between the optimal model state and the predicted state of the coupled model and the observation data during this fusion process, the initial error range of the next round of algorithm is updated. The algorithm enters the next loop based on the optimal model state after the previous round of correction, repeating the process of "model prediction - observation input - Karman gain calculation - state correction - error update" and running continuously at a set frequency. This loop mechanism can respond to the dynamic changes of the plateau irrigation area in real time, ensuring that the optimal model state at the current moment is always consistent with the physical irrigation area, providing highly reliable input data for the subsequent environmental prediction module.
[0040] M3, Model Pre-Drilling Terminal: Used to import the set plan into the established digital twin model for simulation and prediction of the impact of irrigation on the downstream ecological environment, ensuring the health of the fragile plateau ecosystem. Simultaneously, it combines groundwater level, water quality, and soil moisture data to simulate the impact of irrigation on saline-alkali land, optimize improvement measures, and compare the simulation process and results with measured data. It achieves efficient output of 3D visualization effects through cloud rendering technology, and completes data storage for convenient and timely data retrieval. (See also...) Figure 4 Specifically, it includes: M31, Environmental Prediction Module: Includes a scheme setting unit and a scheme simulation unit; M311, the scheme setting unit, is used to set different irrigation schemes based on the plateau crop dataset and hydro-meteorological dataset to cover different management needs and provide direction for scheme simulation. The steps involved in deriving the highest-scoring scheme from the scheme definition unit's generated schemes include: M3111. Based on the current optimal model state, the predicted downstream runoff value is obtained. The specific calculation formula is as follows:
[0041] in, This represents the total irrigation water consumption of the irrigation district. This refers to the amount of loss due to evaporation in the channel. This represents the total deep seepage in the irrigation area. Effective rainfall; by quantifying the balance of water input, consumption and output in the irrigation district, the remaining ecological flow downstream under different irrigation schemes is predicted; M3112. Based on the current optimal model state and crop water requirement, the predicted soil salinity value is obtained. The specific calculation formula is as follows:
[0042] in, Initial soil salinity, This represents the initial soil moisture content. To rinse the irrigation water volume, The rinsing efficiency coefficient is... This refers to the amount of water used for a single irrigation. For crop water requirements, To predict the time step, the predicted soil salinity value dynamically simulates the dilution or accumulation trend of soil salinity after irrigation. When the predicted soil salinity value is less than the initial soil salinity, it indicates that the irrigation scheme can effectively reduce soil salinity and improve salinization. When the predicted soil salinity value is greater than or equal to the initial soil salinity, it indicates that the irrigation scheme may aggravate salinization, and it is necessary to adjust the irrigation amount or irrigation frequency to protect the fragile soil ecology of the plateau. M3113. Based on the current optimal model state and hydrological and meteorological dataset, the predicted groundwater level depth is calculated. The predicted groundwater level depth reflects the rise and fall of the groundwater level after irrigation. If the depth is shallow, it may lead to secondary salinization. If the depth is deep, it may affect the water supply to vegetation. M3114. Based on the current optimal model state, the synergistic impact of irrigation schemes on crop productivity and ecology is calculated, and the predicted net primary productivity reflects the amount of carbon fixed by crop photosynthesis, which is the core indicator for assessing the productivity and carbon sink capacity of the irrigation area ecosystem. M3115, package the predicted downstream runoff, predicted soil salinity, predicted groundwater level depth and predicted net primary productivity to obtain the irrigation scheme simulation results, and send the irrigation scheme simulation results to the rendering output module. M312, the scheme deduction unit, is used to deduce the scheme generated by the scheme setting unit at the current time in the optimal model state to obtain the scheme with the highest score; M32, Rendering Output Module: Includes a rendering processing unit and a data comparison unit; M321, the rendering processing unit, is used to convert the input data into a 3D rendering adaptation format based on the dataset collected by the data acquisition module, the current optimal model state output by the model building end, and the irrigation scheme deduction results output by the environmental prediction module. The input data is then input into the current optimal model state and rendered on a large scale using a cloud rendering engine. At the same time, a dual-layer visualization design of surface and underground is used to display the 3D visualization model data in a three-dimensional way. M322, the data comparison unit, is used to overlay and compare real-time data with predicted data, enabling time-sliding interaction and difference calculation and display. It also stores data in a database built from system data. System data includes datasets and feature extraction data collected by the data processing end, model data built by the model building end, and data derived from model pre-testing. By specifically collecting geospatial, hydrological, meteorological, and crop-specific data of the plateau region, the problem of data acquisition difficulties and poor adaptability in plateau areas was solved, laying a high-quality data foundation for subsequent model construction. In model construction and optimization, physical mechanisms and AI technology were integrated to not only create a high-fidelity virtual model that is highly consistent with the physical irrigation area, but also to effectively improve the interpretability and prediction accuracy of the model in the data-scarce plateau scenario through AI parameter optimization and real-time data fusion and dynamic correction, achieving synchronous evolution of the model and the physical entity. From an ecological and management perspective, relying on the model's extrapolation function, the impact of irrigation schemes on the downstream ecological environment and saline-alkali land can be accurately predicted, ensuring the health and stability of the fragile plateau ecosystem and providing a scientific basis for soil improvement and efficient water resource utilization. Ultimately, the synergistic improvement of ecological protection and production benefits in plateau irrigation areas is achieved, providing innovative and practical technical support for the digital and refined management of plateau irrigation areas.
[0043] In this embodiment, in the dynamic construction system of a plateau digital twin irrigation area based on multi-source data fusion, each of the above modules can be implemented independently and can call other modules. Calling can be understood as a module connecting to multiple modules of another type and providing corresponding services to those connected modules. For example, the sharing and evaluation module can call the same information collection module to obtain the information collected by that module. Based on the above characteristics, in the dynamic construction system of a plateau digital twin irrigation area based on multi-source data fusion provided by this embodiment, the applicable scope of the system architecture can be adjusted by adding modules and directly calling them without modifying the program code, achieving cluster-based horizontal expansion to quickly and flexibly expand the system. In practical applications, the above modules can be set in the same device or different devices, or in virtual devices, such as service instances in a cloud server.
[0044] Example 2 See Figure 5 , Figure 5 This is a flowchart of a method for dynamically constructing a digital twin irrigation district in a plateau region, as proposed in this invention. Specific steps may include: S1. Collect geospatial datasets, hydro-meteorological datasets, and plateau crop datasets to construct a multi-source dataset; S2. Perform data cleaning and feature extraction on the multi-source dataset; S3. Based on multi-source datasets and feature-extracted data, establish a plateau irrigation area model; S4. Based on the plateau irrigation area model, a coupled model is obtained by integrating AI data-driven approaches. S5. Based on the coupled model, real-time data is incorporated for correction to obtain the optimal model state at the current moment; S6. Based on the current optimal model state and irrigation scheme, perform a deduction to obtain the irrigation scheme deduction results; S7. Perform 3D rendering based on the irrigation scheme simulation results and store the system data.
[0045] In summary, the present invention has the following advantages: 1. By setting up the data processing terminal in the data acquisition module, and through the use of equipment and scientific deployment adapted to the special environment of the plateau, comprehensive data covering geospatial, hydrological and meteorological, and crop fields are acquired, ensuring the accuracy and timeliness of the data in the complex plateau environment and providing the system with a high-quality raw data source. The data preprocessing module extracts and standardizes features, removes noise, fills in gaps, and generates feature data that is adapted to the subsequent model, reducing the data processing burden of the subsequent modules and improving the quality of the model input data. 2. By setting up a model building module in the plateau irrigation area, multi-dimensional plateau-specific data are systematically collected to provide comprehensive and accurate basic support for subsequent modules, ensuring the accuracy and completeness of the digital twin model construction. The AI data-driven module combines physical mechanisms with AI technology to optimize model parameters, improve model interpretability and prediction accuracy, and solve the problem of scarce plateau data. The real-time data fusion module dynamically integrates real-time monitoring data to correct simulation results, ensuring that the digital twin and the physical entity evolve synchronously. The three work together to help build a high-fidelity model, providing reliable technical support for ecological protection and land improvement in the plateau irrigation area. 3. By setting up a model pre-simulation terminal and relying on real-time fused data in the environmental prediction module, multiple irrigation schemes can be simulated to accurately simulate the impact of irrigation on downstream ecology, groundwater level, and saline-alkali land. Ecological risks can be identified in advance, providing a scientific basis for the protection of fragile plateau ecosystems. At the same time, saline-alkali land improvement measures can be optimized to help improve land quality. The rendering output module uses cloud rendering to achieve 3D visualization, clearly presenting the difference between the prediction and the current environmental data, intuitively displaying the model simulation results, and can also systematically store various types of data to provide data support for subsequent analysis and decision-making, improving the system's practicality and decision-making efficiency.
[0046] The above specific embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to examples, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.
Claims
1. A dynamic construction system for a digital twin irrigation district in a plateau region, characterized in that, include: Data processing end: Used to collect multi-source datasets from plateau irrigation areas, clean the data of each sub-data item in the collected multi-source datasets, standardize the data of each sub-data item in the cleaned multi-source datasets, and extract features from the standardized multi-source data; Model building module: This module is used to build a model of the plateau irrigation area based on the collected multi-source datasets and feature-extracted data. It introduces data-driven model optimization to optimize the parameters of the plateau irrigation area model, and integrates real-time monitoring data of the plateau irrigation area into the model to dynamically correct the simulation results, ensuring consistency between the digital twin and the physical entity of the plateau irrigation area. The model building module includes: a plateau irrigation area model building module, an AI model coupling module, and a real-time data fusion module. The plateau irrigation district model establishment module includes a plateau basic model establishment unit, a hydrological model establishment unit, and a plateau crop model establishment unit. The plateau basic model building unit is used to construct the model topology based on the geospatial dataset, which clarifies the spatial range of the model and the interaction logic of its internal components, and provides a basic framework for subsequent sub-model coupling. The hydrological model building unit is used to calculate the channel water conveyance based on surface slope data and hydrological and meteorological datasets. The plateau crop model establishment unit calculates crop biomass based on the plateau crop dataset; The AI model coupling module includes an AI digital driving unit and a multi-model coupling unit; The AI digital driving unit is used to determine sensitive parameters based on a plateau irrigation area model using sensitivity analysis, and to construct an AI parameter optimization model based on preprocessed feature data and the simulation results of the plateau irrigation area model. The specific steps for constructing the AI parameter optimization model include: Obtain a set of feature data stored in the database and the simulation results of the plateau irrigation area model, and compare them with the current time according to the timestamp. Group the comparison results from small to large and use the labeled results as the sample set. The sample set was divided into a 70% training set, a 15% test set, and a 15% validation set. Based on the sample set, an original AI parameter optimization model was built. The original AI parameter optimization model is iterated continuously until the required confidence level is reached and then output, thus obtaining the AI parameter optimization model. Transfer learning is used to transfer AI model parameters from areas with abundant data to areas with scarce data, calculate the feature similarity between the two areas, and then correct the AI output parameters in areas with scarce data based on the similarity weighting, thus solving the parameter calibration problem caused by insufficient data in some areas of the plateau. The multi-model coupling unit is used to substitute the optimized parameters output by AI into the calculation formula of the plateau irrigation area model, and at the same time, the SHAP value is used to analyze the influence contribution of AI parameters, thereby improving the interpretability of the model. The real-time data fusion module includes a dynamic data import unit and an algorithm import unit; The dynamic data import unit is used to import soil moisture data, rainfall data and crop leaf area index, and sets the assimilation frequency and triggering conditions. The algorithm introduction unit is used to calculate the Kalman gain based on the error range of the coupled model's predicted state and the error range of the observed data, and to fuse and correct the coupled model's predicted state with the observed data to generate the optimal model state at the current moment. Model pre-simulation terminal: Used to import the set irrigation scheme into the established digital twin model of the plateau irrigation area for simulation and prediction of the impact of irrigation on the target area environment of the plateau irrigation area. Combined with the environmental correlation data of the plateau irrigation area, it simulates the impact of irrigation on specific land types in the region. The simulation process and results are compared with the measured data. The simulation results are output through visualization technology, and the relevant data are stored.
2. The dynamic construction system for a plateau digital twin irrigation district according to claim 1, characterized in that, The multi-source dataset for the plateau irrigation area includes: a geospatial dataset, a hydro-meteorological dataset, and a plateau crop dataset; the data processing terminal includes a data acquisition module and a data preprocessing module. The data acquisition module includes: a geospatial acquisition unit, a hydro-meteorological acquisition unit, and a plateau crop acquisition unit; The geospatial data acquisition unit is used to collect geospatial datasets, which provide a basis for locating the boundary of the plateau irrigation area model. The geospatial datasets include soil data, topographic data, and vegetation coverage. The hydro-meteorological data acquisition unit is used to collect hydro-meteorological datasets. These datasets provide a data source for the dynamic correction of water flow simulation and real-time fusion in the plateau irrigation area model. The hydro-meteorological datasets include rainfall data, temperature data, and irrigation area runoff data. The plateau crop data collection unit is used to collect plateau crop datasets. The plateau crop datasets provide time-series data for the ecological effect assessment of crop growth simulation and environmental prediction in plateau irrigation area models. The plateau crop datasets include crop leaf area index and crop water requirement.
3. The dynamic construction system for a plateau digital twin irrigation district according to claim 2, characterized in that, The data preprocessing module includes a data cleaning unit and a feature extraction unit; The data cleaning unit cleans the basic dataset by removing outliers and normalizes all sub-data items in the basic dataset to the range of [0, 1] according to the normalization formula. The feature extraction unit is used to extract features from geospatial datasets and hydro-meteorological datasets.
4. The dynamic construction system for plateau digital twin irrigation districts according to claim 3, characterized in that, The steps for feature extraction from geospatial datasets and hydro-meteorological datasets include: Soil moisture data were obtained by using the moving average method based on a geospatial dataset. The specific calculation formula is as follows: in, To adjust the sliding window size, For time-series indexes, For the time index of the target smoothing moment, The measured soil moisture data is given at time i. Calculations are performed based on topographic data to obtain surface slope data. The specific calculation formula is as follows: in, This is the arctangent function, used to convert the output value within the immediately following parentheses to radian format. This represents the rate of change of terrain height in the horizontal direction. The horizontal and vertical rates of change of terrain height. This is the conversion factor for radian angles; Rainfall in the blank monitoring area was calculated using the Thiessen polygon formula based on hydrological and meteorological data. The specific calculation formula is as follows: in, The number of blank monitoring areas included in the calculation. Let the area of the Thiessen polygon be the rainfall amount corresponding to the i-th blank monitoring area. The total Thiessen polygon area represents the total rainfall in the blank monitoring area. Let be the measured rainfall at the i-th monitoring station. This is an index for monitoring sites.
5. The dynamic construction system for a plateau digital twin irrigation district according to claim 1, characterized in that, Obtain the water conveyance of the channel The calculation formula is: in, The cross-sectional area of the channel through which water flows. For hydraulic radius, This is the Manning roughness coefficient; Obtaining crop biomass The calculation formula is: in, The maximum biomass of the crop. It is a natural constant. All are weighting coefficients. Accumulated temperature during the crop's growing season.
6. The dynamic construction system for a plateau digital twin irrigation district according to claim 1, characterized in that, The model pre-simulation terminal includes an environment prediction module and a rendering output module; The environmental prediction module includes a scheme setting unit and a scheme simulation unit; The scheme setting unit is used to set different irrigation schemes based on plateau crop datasets and hydrological and meteorological datasets to cover different management needs, and at the same time provide direction for scheme derivation; The scheme deduction unit is used to deduce the scheme generated by the scheme setting unit at the current time based on the optimal model state of the scheme, and obtain the scheme with the highest score. The rendering output module includes a rendering processing unit and a data comparison unit; The rendering processing unit is used to convert the input data into a three-dimensional rendering adaptation format based on the dataset collected by the data acquisition module, the current optimal model state output by the model building end, and the irrigation scheme deduction results output by the environmental prediction module. The input data is then input into the current optimal model state and rendered on a large scale using a cloud rendering engine. At the same time, a dual-layer visualization design of surface and underground is used to display the three-dimensional visualization model data in a three-dimensional way. The data comparison unit is used to overlay and compare real-time data and predicted data, realize time-sliding interaction and difference calculation and display, and establish a database based on system data for storage.
7. The dynamic construction system for a plateau digital twin irrigation district according to claim 6, characterized in that, The steps involved in deriving the highest-scoring scheme from the scheme definition unit's generated schemes include: The downstream runoff forecast is obtained by calculating based on the current optimal model state. The specific calculation formula is as follows: in, This represents the total irrigation water consumption of the irrigation district. This refers to the amount of loss due to evaporation in the channel. This represents the total deep seepage in the irrigation area. Effective rainfall; by quantifying the balance of water input, consumption and output in the irrigation district, the remaining ecological flow downstream under different irrigation schemes is predicted; Soil salinity predictions are obtained based on the current optimal model state and crop water requirement. The specific calculation formula is as follows: in, Initial soil salinity, This represents the initial soil moisture content. To rinse the irrigation water volume, The rinsing efficiency coefficient is... This refers to the amount of water used for a single irrigation. For crop water requirements, To predict the time step; The predicted groundwater level depth is calculated based on the current optimal model state and hydrological and meteorological dataset. Based on the current optimal model state, the synergistic impact of irrigation schemes on crop productivity and ecology is calculated, and the predicted net primary productivity reflects the amount of carbon fixed by crop photosynthesis. The predicted downstream runoff, soil salinity, groundwater level depth, and net primary productivity are packaged to obtain the irrigation scheme simulation results, which are then sent to the rendering output module.
8. A method for dynamically constructing a digital twin irrigation district in a plateau region, characterized in that, The system applied to the dynamic construction system of the plateau digital twin irrigation district according to any one of claims 1-7 includes: S1. Collect geospatial datasets, hydro-meteorological datasets, and plateau crop datasets to construct a multi-source dataset; S2. Perform data cleaning and feature extraction on the multi-source dataset; S3. Based on multi-source datasets and feature-extracted data, establish a plateau irrigation area model; S4. Based on the plateau irrigation area model, a coupled model is obtained by integrating AI data-driven approaches. S5. Based on the coupled model, real-time data is incorporated for correction to obtain the optimal model state at the current moment; S6. Based on the current optimal model state and irrigation scheme, perform a deduction to obtain the irrigation scheme deduction results; S7. Perform 3D rendering based on the irrigation scheme simulation results and store the system data.