Method and system for optimizing operation of water network system for carbon reduction and efficiency improvement in agricultural and pastoral areas

Abstract

The present invention discloses a method and system for optimizing the operation of a water network system in agricultural and pastoral areas for carbon reduction and efficiency improvement. The method collects agricultural and pastoral data, constructs an agricultural and pastoral database, builds a dynamic coupling model, and optimizes model parameters. A crop planting structure optimization model is constructed, and the model is solved to obtain a quarterly planting plan. The plan is input into the dynamic coupling model to obtain a corresponding coordinated scheduling plan. A daily-scale scheduling model is constructed, and the model is solved to obtain a set of daily scheduling plans, and the optimal daily scheduling plan is screened. A water and power storage strategy is formulated to adjust the operation of water pumps, wind and solar storage equipment. A real-time monitoring platform for agricultural and pastoral water network systems based on digital twin technology is constructed to generate a control strategy and a carbon reduction and efficiency improvement evaluation report, and optimize and adjust the parameters in the dynamic coupling model. The present invention realizes the coordinated scheduling of water resources, power resources, and agricultural loads, improves the utilization rate of renewable energy such as wind and solar power, and enhances the overall operating efficiency of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of integrated management of energy and water resources in agricultural and pastoral areas, and specifically to a method and system for optimizing the operation of carbon reduction and efficiency improvement of agricultural and pastoral water network systems based on multi-energy synergy and dynamic feedback control. Background Art

[0002] In the production and life of agricultural and pastoral areas, the water network system is closely coupled with the power system. The traditional operation mode has problems such as low resource utilization efficiency, high carbon emissions, and poor system reliability. Existing technologies mostly adopt single-objective optimization or static scheduling strategies, which cannot adapt to dynamic scenarios such as fluctuations in wind and solar resources and changes in crop water demand. For example, some agricultural and pastoral areas rely too much on fossil energy to drive water pumps during peak irrigation season, resulting in increased carbon emissions; or due to the under-utilization of renewable energy, wind and solar resources are wasted. In addition, the traditional system lacks real-time feedback and dynamic adjustment mechanisms, making it difficult to cope with sudden meteorological changes or equipment failures, and unable to achieve efficient coordination of water, electricity, and load. There is an urgent need for a comprehensive optimization solution that can take into account economy, environmental benefits and reliability.

[0003] The present invention proposes a method and system for optimizing the operation of water network systems in agricultural and pastoral areas to reduce carbon emissions and increase efficiency, solve the above-mentioned problems that currently exist, achieve coordinated and efficient utilization of water, electricity, and load resources in agricultural and pastoral areas, reduce system operating costs and carbon emissions, and improve system operation reliability. Summary of the Invention

[0004] The invention aims to provide a method for optimizing the operation of water network systems in agricultural and pastoral areas to reduce carbon emissions and increase efficiency, in order to solve the above-mentioned problems existing in the prior art. On the other hand, a system for optimizing the operation of water network systems in agricultural and pastoral areas to reduce carbon emissions and increase efficiency is provided.

[0005] Technical solution: A method for optimizing the operation of water network systems in agricultural and pastoral areas to reduce carbon emissions and increase efficiency, including the following steps:

[0006] Step S1: Collect agricultural and pastoral area data, build an agricultural and pastoral area database, construct a dynamic coupling model, and optimize model parameters;

[0007] Step S2: construct a crop planting structure optimization model, solve the model to obtain a quarterly planting planning scheme, and input it into the dynamic coupling model to obtain the corresponding quarterly collaborative scheduling scheme;

[0008] Step S3: Build a daily-scale scheduling model based on the quarterly collaborative scheduling plan, solve the model to obtain a set of daily scheduling plans, and select the optimal daily scheduling plan using cost and carbon emissions as indicators;

[0009] Step S4: Formulate a water and electricity storage strategy and use it as input to the pre-built distributed model to adjust the operation of water pumps, wind, solar and storage equipment to obtain an optimized operation plan for carbon reduction and efficiency improvement of the pastoral water network system;

[0010] Step S5: Build a real-time monitoring platform for the rural water network system based on digital twin technology, map the actual operating status in real time, build an adaptive feedback control model, predict the deviation between the system operating status and the scheduling plan, and dynamically adjust the system operating parameters based on the prediction results, generate control strategies and carbon reduction and efficiency improvement evaluation reports, and optimize and adjust the parameters in the water-electricity-load dynamic coupling model.

[0011] According to one aspect of the present application, step S1 further comprises:

[0012] Step S11: Collect agricultural and pastoral area data and build an agricultural and pastoral area database;

[0013] Step S12: extracting data on water networks, wind and solar power station outputs, and crop water requirements, fitting the distributions of water networks, wind and solar power station outputs, and crop water requirements, respectively, and constructing a joint distribution of the three, and randomly sampling to obtain several agricultural and pastoral scenarios;

[0014] Step S13: constructing a dynamic coupling model;

[0015] Step S14: inputting several agricultural and pastoral area scenarios into the dynamic coupling model, training and optimizing the model parameters;

[0016] According to one aspect of the present application, step S13 is further as follows:

[0017] Step S13a: Retrieve water network data from the agricultural and pastoral database, use pipelines, water pumps, and reservoirs as nodes and connecting edges, calculate and draw the pump head-flow curve, calculate the head loss along the process, and construct a hydrodynamic model of the agricultural and pastoral water network;

[0018] Step S13b: retrieve wind and solar power station data from the agricultural and pastoral area database, and construct a relationship model between the output power of the photovoltaic power station and the light intensity and temperature, a relationship model between the wind power power and wind speed, and a charge and discharge efficiency curve of the energy storage device, to obtain a power network flow model for the agricultural and pastoral area;

[0019] Step S13c: determining the corresponding crop coefficient based on the crop growth cycle, calculating the crop water requirement using the Penman-Montes formula in combination with meteorological data, and constructing a dynamic calculation model for crop water requirement;

[0020] Step S13d, coupling the hydrodynamic model of the agricultural and pastoral water network, the power network flow model of the agricultural and pastoral area, and the dynamic calculation model of crop water demand, introducing the spatiotemporal attention mechanism, calculating the attention weights of the time series and spatial nodes and configuring the model, and obtaining the water-electricity-load dynamic coupling model.

[0021] According to one aspect of the present application, step S2 further comprises:

[0022] Step S21: extracting reservoir capacity, river flow, and groundwater level data in agricultural and pastoral areas, calculating the total available water resources and spatiotemporal distribution characteristics for each quarter, and calculating the solar radiation and wind power generation for each quarter based on meteorological station data;

[0023] Step S22: constructing a crop planting structure optimization model, solving the model to obtain the optimal crop planting combination and area allocation plan, that is, the quarterly planting plan;

[0024] Step S23: input the quarterly planting planning scheme into the dynamic coupling model to obtain the corresponding quarterly water-electricity-load coordinated scheduling scheme.

[0025] According to one aspect of the present application, step S3 is further:

[0026] Step S31: Decompose the quarterly collaborative scheduling plan by day, and construct a daily scheduling time series with a time step of 15 minutes to obtain a daily-scale scheduling model;

[0027] Step S32: extract several agricultural and pastoral scenarios and input them into the daily scheduling model. Use a distributed computing framework to solve them in parallel to generate a water pump start-stop plan, energy storage charging and discharging strategy, and grid interaction plan for the corresponding scenarios, i.e., a daily scheduling plan set.

[0028] Step S33: Construct a three-dimensional evaluation system including total cost, total carbon emissions, and reliability indicators, use the approximate ideal solution ranking method to calculate the relative closeness of each plan to the ideal solution, and select the comprehensive optimal daily scheduling plan.

[0029] According to one aspect of the present application, step S4 is further:

[0030] Step S41: Based on the dynamic coupling model and the daily scheduling plan, a water and electricity storage strategy is formulated using mixed integer nonlinear programming;

[0031] Step S42: Build a distributed model, convert the water and electricity storage strategy into control instructions that the distributed model can recognize, clarify the start time, stop time, and operating power of the water pump, the power adjustment parameters of the wind and solar power generation equipment, and the charge and discharge control signals of the energy storage equipment, and input them into the distributed model to adjust the operation of the water pump, wind and solar power generation equipment;

[0032] Step S43: Collect the water level flow, water level changes, grid power fluctuations, and energy storage equipment charging and discharging status data after equipment operation adjustment, and feed them back to the dynamic coupling model to evaluate the system operation status and optimize the quarterly water-electricity-load coordinated scheduling plan and daily scheduling plan to obtain the carbon reduction and efficiency improvement optimization operation plan for the pastoral water network system.

[0033] According to one aspect of the present application, step S41 is further as follows:

[0034] Step S41a: Taking the maximization of the system's comprehensive benefits as the core goal, a multi-objective function is constructed that includes economic benefits, environmental benefits, and operational reliability, and weight coefficients of carbon trading prices, time-of-use electricity prices, and equipment maintenance cost parameters are determined;

[0035] Step S41b: Combined with the dynamic coupling model, set the upper and lower limits of the water level of the water storage equipment, the state of charge range of the power storage equipment, the power limit for starting and stopping the water pump, and the power balance constraint conditions of the power grid;

[0036] Step S41c: Using a mixed integer nonlinear programming algorithm, the charging and discharging time and water volume of the water storage equipment, and the charging and discharging period and power of the power storage equipment are set as decision variables, and the objective function is solved to obtain the optimal water and power storage strategy.

[0037] According to one aspect of the present application, step S5 is further:

[0038] Step S51: Build a real-time monitoring platform for the agricultural and pastoral water network system based on digital twin technology, collect system operation data in real time through IoT devices, and map the actual operation status in the digital twin model in real time;

[0039] Step S52: construct an adaptive feedback control model, use a deviation prediction algorithm to predict the deviation between the system operating state and the scheduling plan, and use an adaptive robust control algorithm to dynamically adjust the system operating parameters based on the prediction results;

[0040] Step S53: Construct a multi-dimensional economic evaluation system including economic indicators, environmental indicators, energy indicators, and reliability indicators, use the entropy weight-grey correlation analysis method to conduct a comprehensive evaluation of the system, and generate a control strategy and carbon reduction and efficiency improvement evaluation report;

[0041] Step S54: Based on the control strategy and the carbon reduction and efficiency improvement assessment report, the Bayesian optimization algorithm is used to optimize and adjust the parameters in the dynamic coupling model.

[0042] According to one aspect of the present application, step S52 is further as follows:

[0043] Step S52a: Construct a data comparison module to extract the real-time collected water network flow, wind and solar power generation, and energy storage charge state data, compare them with the expected values ​​in the daily scheduling plan and water and power storage strategy on a time-by-time basis, use a sliding window algorithm to process the deviation data, and extract the deviation change trend and change rate;

[0044] Step S52b: using the recursive least squares method to estimate system parameters and adjust the pump power-flow model parameters in real time;

[0045] Step S52c: Based on the deviation characteristics and the calculation results of the control algorithm, generate water pump speed adjustment and wind, solar and storage device power adjustment instructions, and send the control instructions to the device controller. After the device executes, it will provide real-time feedback on the execution status.

[0046] According to another aspect of the present application, a carbon reduction and efficiency-enhancing optimization operation system for a water network system in agricultural and pastoral areas is provided, comprising:

[0047] at least one processor; and

[0048] a memory communicatively connected to at least one of the processors; wherein,

[0049] The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement the carbon reduction and efficiency improvement optimization operation method of the agricultural and pastoral water network system described in any of the above technical solutions.

[0050] Beneficial effects: By adopting the optimized operation method of reducing carbon and increasing efficiency of the water network system in rural and pastoral areas, through multi-dimensional data fusion, hierarchical optimization scheduling and dynamic feedback control, the coordinated and efficient utilization of water, electricity and load resources in rural and pastoral areas can be achieved, the system operating costs and carbon emissions can be reduced, and the system operation reliability can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a flow chart of the present invention.

[0052] Figure 2 It is a flow chart of step S1 of the present invention.

[0053] Figure 3 It is a flow chart of step S2 of the present invention.

[0054] Figure 4 It is a flow chart of step S3 of the present invention.

[0055] Figure 5 It is a flow chart of step S4 of the present invention.

[0056] Figure 6 It is a flow chart of step S5 of the present invention. DETAILED DESCRIPTION

[0057] like Figure 1 According to one aspect of the present application, a method for optimizing the operation of a water network system in agricultural and pastoral areas to reduce carbon emissions and increase efficiency is provided, which is characterized by comprising the following steps:

[0058] Step S1: Collect agricultural and pastoral data, build an agricultural and pastoral database, construct a water-power-load dynamic coupling model, and optimize model parameters;

[0059] Step S2: Construct a crop planting structure optimization model, solve the model to obtain a quarterly planting plan, and input the water-power-load dynamic coupling model to obtain a corresponding coordinated scheduling plan;

[0060] Step S3: Build a daily-scale scheduling model based on the quarterly collaborative scheduling plan, solve the model to obtain a set of daily scheduling plans, and select the optimal daily scheduling plan using cost and carbon emissions as indicators;

[0061] Step S4: Formulate a water and electricity storage strategy and use it as input to the pre-built distributed model to adjust the operation of water pumps, wind, solar and storage equipment to obtain an optimized operation plan for carbon reduction and efficiency improvement of the pastoral water network system;

[0062] Step S5: Build a real-time monitoring platform for the rural water network system based on digital twin technology, map the actual operating status in real time, build an adaptive feedback control model, predict the deviation between the system operating status and the scheduling plan, and dynamically adjust the system operating parameters based on the prediction results, generate control strategies and carbon reduction and efficiency improvement evaluation reports, and optimize and adjust the parameters in the dynamic coupling model.

[0063] like Figure 2 As shown, according to one aspect of the present application, the step S1 is further:

[0064] Step S11: Collect agricultural and pastoral area data and build an agricultural and pastoral area database;

[0065] Deploy a multi-dimensional sensing network consisting of IoT sensors, satellite remote sensing equipment, and meteorological monitoring stations to collect high-precision data on water networks, wind and solar power plants, and crop water demand in real time. Water network data includes real-time water network topology, pipeline pressure and flow dynamic parameters, pump energy efficiency curves, and the spatiotemporal distribution of reservoir water storage. Wind and solar power plant data includes ultra-short-term power forecasts and equipment health status monitoring. Crop water demand data includes soil moisture in the crop root zone and water demand intensity at different growth stages.

[0066] Data on water, electricity, agriculture, etc. in rural and pastoral areas are scattered across different departments, with different formats and a lack of systematicity, making them difficult to use directly for model building and analysis. Integrating multi-source data to form a unified and standardized database provides basic data support for subsequent data processing and model building. The use of a relational database can efficiently store structured data and facilitate data query, update, and management. By formulating unified data collection standards and format specifications, data consistency and compatibility can be guaranteed.

[0067] In a certain embodiment, specifically:

[0068] Deploy soil moisture sensors, weather stations, and livestock tracking collar IoT devices in agricultural and pastoral areas to collect real-time data on soil, climate, and livestock growth;

[0069] Organize staff to collect basic information of farmers and herders, breeding scale, and static data of plant varieties through mobile terminal apps, and access agricultural and pastoral policy and land planning data disclosed by government departments;

[0070] Develop agricultural and pastoral area management systems to provide government departments with macro-decision-making analysis, such as predicting agricultural and pastoral product output and planning resource allocation.

[0071] Step S12: extracting data on water networks, wind and solar power station outputs, and crop water requirements, fitting the distributions of water networks, wind and solar power station outputs, and crop water requirements, respectively, and constructing a joint distribution of the three, and randomly sampling to obtain several agricultural and pastoral scenarios;

[0072] Water, electricity, and agricultural load data in agricultural and pastoral areas have significant spatiotemporal variability and uncertainty. Traditional deterministic models are unable to accurately describe their changing patterns. By fitting data and constructing joint distributions, we can explore the probability distribution characteristics of the data and generate diverse agricultural and pastoral scenarios to enhance the model's adaptability to uncertainty. In one embodiment, specifically:

[0073] Historical data on water network flow, photovoltaic power station output, wind power output, and crop water demand were extracted from the constructed database. The kernel density estimation function in the Python Scikit-learn library was used to fit the distribution of each data, and the probability density function of water network flow, photovoltaic power station output, wind power output, and crop water demand was obtained.

[0074] The Gaussian Copula function is used to construct the joint distribution function between the water network, wind and solar power station output and crop water demand;

[0075] Based on the joint distribution, the random sampling function in Python's Numpy library was used to generate no less than 1,000 different agricultural and pastoral scenarios. Each scenario contained a set of random data on water network flow, photovoltaic power station output, wind power output, and crop water demand.

[0076] The generated agricultural and pastoral scenarios can fully reflect the uncertainty in the system operation process, so that the subsequently constructed models have better generalization capabilities in different scenarios and improve prediction accuracy.

[0077] Step S13: constructing a dynamic coupling model;

[0078] Step S14: Input several agricultural and pastoral area scenarios into the water-power-load dynamic coupling model to train and optimize the model parameters;

[0079] In some cases, the at least 1,000 agricultural and pastoral area scenarios generated in step S12 are randomly divided into a training set and a test set, and the meteorological data, water network parameters, power network parameters, and crop planting information data of the training set are input into the water-power-load dynamic coupling model, the output of the model is calculated, and compared with the actual observation data, and the loss function is calculated. The model parameters are updated, the gradient is calculated by the back propagation algorithm, and the weight and bias parameters in the model are adjusted to minimize the loss function;

[0080] After 100 rounds of iterative training, the test set data is input into the trained model, and the prediction error of the model on the test set is calculated. If the error meets the set threshold, the model training is considered qualified. Otherwise, the parameters are continued to be adjusted for training until the model performance meets the requirements.

[0081] According to one aspect of the present application, step S13 is further as follows:

[0082] Step S13a: Retrieve water network data from the agricultural and pastoral database, use pipelines, water pumps, and reservoirs as nodes and connecting edges, calculate and draw the pump head-flow curve, calculate the head loss along the process, and construct a hydrodynamic model of the agricultural and pastoral water network;

[0083] Parameters such as the diameter, length, and roughness of water network pipelines, as well as the model, rated power, and rated flow rate of water pumps, are obtained from the agricultural and pastoral database. The head loss along the pipeline is calculated according to the Darcy-Weisberg formula. The head-flow curve of the water pump is drawn using performance data provided by the pump manufacturer or field experiments. The pipelines, pumps, and reservoirs in the water network are abstracted into nodes and connecting edges using Python's NetworkX library to construct a water network topology. Combined with fluid mechanics equations, the finite element analysis method is used to solve the pressure, flow, and water level of each node in the water network, thereby constructing a fluid dynamics model of the agricultural and pastoral water network.

[0084] Step S13b: retrieve wind and solar power station data from the agricultural and pastoral area database, and construct a relationship model between the output power of the photovoltaic power station and the light intensity and temperature, a relationship model between the wind power power and wind speed, and a charge and discharge efficiency curve of the energy storage device, to obtain a power network flow model for the agricultural and pastoral area;

[0085] Wind and solar power generation is intermittent and volatile, and traditional power models are difficult to accurately predict their output, resulting in irrational power system scheduling and serious wind and solar power abandonment. Establishing an accurate power network flow model in rural areas to describe the dynamic relationship between wind and solar power generation, energy storage equipment and power load can solve this problem. Therefore, in this embodiment, a power model is established based on the working principles of photovoltaic cells and wind turbines, and the charging and discharging efficiency curve of the energy storage equipment is constructed through experiments or historical data analysis. The power system flow calculation method is used to solve the power network flow to ensure the accuracy and convergence of the calculation results.

[0086] Step S13c: determining the corresponding crop coefficient based on the crop growth cycle, calculating the crop water requirement using the Penman-Montes formula in combination with meteorological data, and constructing a dynamic calculation model for crop water requirement;

[0087] Growth cycle data of major crops in agricultural and pastoral areas were collected and divided into sowing period, seedling period, growing period, flowering period, and maturity period. The crop coefficient corresponding to each stage was determined according to the crop coefficient table recommended by FAO. Daily solar radiation, temperature, relative humidity, and wind speed data were obtained from the meteorological database. The Penman-Montes formula was used to calculate the daily reference crop evapotranspiration. The water requirement of crops at each growth stage was then calculated. The data were processed and analyzed using Python's Pandas and NumPy libraries to construct a dynamic calculation model for crop water demand.

[0088] In a certain embodiment, specifically:

[0089] Corn is a crop grown in a certain agricultural and pastoral area. First, the growth cycle and stage division of corn are clarified, specifically:

[0090] The initial stage is from sowing to 20 days after emergence, when vegetation cover is low, transpiration is weak, and the crop coefficient is 0.3-0.5;

[0091] The development stage is from 20 days after emergence to the tasseling stage, with rapid leaf growth, a coverage of 70%-80%, and a crop coefficient of 0.5-1.0;

[0092] The middle stage is from the tasseling stage to the grain filling stage, the peak water demand period, the coverage is close to 100%, and the crop coefficient is 1.0-1.2;

[0093] The late stage is from the grain filling stage to the harvest stage, when leaves senesce, water requirements decrease, and the crop coefficient is 1.2-0.5;

[0094] Data on daily maximum and minimum temperatures, sunshine hours, air humidity, wind speed, and atmospheric pressure in agricultural and pastoral areas were collected from meteorological stations, outliers were cleaned, and missing data were supplemented.

[0095] The Penman-Monteis formula was used to calculate daily reference crop evapotranspiration;

[0096] Among them, solar radiation is calculated by sunshine hours, for example, R s =(0.75+0.000189n)R a , where n is the number of sunshine hours, R a is the top-of-atmosphere radiation;

[0097] Long-wave radiation R nl =σT 4 (0.34-0.14e a ^0.5)(1.35R a ÷R s -0.35), where σ is the Stefan-Boltzmann constant, e a is the actual water vapor pressure, T is the surface temperature;

[0098] In the middle growth period, G≈0.05R n , where G is the soil heat flux, R n =R s (1-α)-R nl , α is the surface albedo, which is 0.23 for corn;

[0099] Calculate actual water requirement based on crop coefficient: ET c =K c ×ET0;

[0100] Among them, ET0 refers to crop evapotranspiration, K c As a coefficient, ET c is the actual water demand;

[0101] In this embodiment, ET0 is calculated to be 5.2 mm / day, K c =1.15; then ET c =1.15×5.2≈6.0mm / day, that is, each mu of land requires about 4000L of water per day;

[0102] Construct a dynamic calculation model for crop water demand, input crop sowing date and daily meteorological data into the dynamic calculation model for crop water demand, calculate the growing days based on the current date and match the crop coefficient;

[0103] Calculate reference evapotranspiration, output actual daily water demand and generate dynamic curve.

[0104] Step S13d, coupling the hydrodynamic model of the agricultural and pastoral water network, the power network flow model of the agricultural and pastoral area, and the dynamic calculation model of crop water demand, introducing the spatiotemporal attention mechanism, calculating the attention weights of the time series and spatial nodes and configuring the model, and obtaining the water-electricity-load dynamic coupling model.

[0105] Build a dynamic coupling model to achieve integrated modeling of water, electricity, and agricultural load systems, accurately describing the dynamic relationship and mutual influence among the three;

[0106] The spatiotemporal attention mechanism can automatically learn the dependencies between variables at different temporal and spatial scales, highlighting the impact of key information on the system state, enhancing the model's expressiveness and predictive accuracy. By coupling the three sub-models, it can achieve collaborative modeling of water, electricity, and agricultural load systems, reflecting the overall operating characteristics of the system.

[0107] In a certain embodiment, specifically:

[0108] The hydrodynamic model of the agricultural and pastoral water network, the flow model of the agricultural and pastoral power network, and the dynamic calculation model of crop water demand are integrated, the input variables and output variables are defined, and the spatiotemporal attention mechanism is introduced. For time series data, the multi-head self-attention mechanism is used to calculate the attention weights between data of different time steps, highlighting the influence of recent data and data at key time points; for spatial node data, the graph attention network is used to calculate the attention weights between water network nodes, power network nodes, and agricultural area nodes, reflecting the interaction between various spatial elements. The coupling model is optimized through training data, and the model parameters and attention weights are adjusted so that the model can accurately predict the operating status of the water-power-load system under different scenarios, and finally a dynamic coupling model is obtained.

[0109] In this embodiment, the spatiotemporal attention mechanism is used to solve the technical problem that the weights of mutual influence between subsystems in traditional coupling modeling are fixed and cannot be adaptively adjusted. Traditional methods use linear weighting or simple matrix multiplication for system coupling, which makes it difficult to capture the complex dependencies between various elements at different spatiotemporal scales.

[0110] Specifically, the time attention weight calculation adopts the multi-head self-attention mechanism. For the time series data X_t∈R T ×d , where T is the number of time steps, d is the feature dimension, and the temporal attention weight calculation formula is:

[0111] Attention_time(Q,K,V)=softmax(QK T / sqrt(d_k))V;

[0112] Among them, Q=X_tW_Q, K=X_tW_K, V=X_tW_V are query matrix, key matrix and value matrix respectively, W_Q, W_K, W_V∈R d×d_k is the learnable parameter matrix, d_k is the dimension of the attention head, and the multi-head attention mechanism uses h=8 attention heads, and the dimension of each head is d_k=64.

[0113] The spatial attention weight is calculated using the Graph Attention Network (GAT), and for the spatial node feature H∈R N×F , where N is the number of nodes, F is the node feature dimension, and the spatial attention coefficient calculation formula is:

[0114] e_ij=LeakyReLU(a T [Wh_i||Wh_j]);

[0115] Where W∈R F'×F is the weight matrix, a∈R 2F' is the attention mechanism parameter vector, || represents the vector concatenation operation, LeakyReLU is the activation function, and the negative slope is set to 0.2.

[0116] The normalized attention weight is calculated as:

[0117] α_ij=exp(e_ij) / Σ_(k∈N_i)exp(e_ik);

[0118] Among them, N_i represents the set of neighbor nodes of node i.

[0119] It should be noted that the coupling weight matrix W_coupling is dynamically updated through the attention weight:

[0120] W_coupling(t)=α_time(t) O α_spatial+βW_static; where O represents the outer product operation, β=0.3 is the static weight retention coefficient, and W_static is the preset static coupling matrix.

[0121] For example, in a certain agricultural and pastoral water network system, which includes 12 water network nodes, 8 power nodes, and 15 agricultural area nodes, during the peak irrigation period (14:00-16:00), the system automatically calculates that the attention weight of water network node 3 (main reservoir) to power node 5 (water pump station) is 0.85, and the attention weight to agricultural area node 8 (corn planting area) is 0.72, indicating that the reservoir status during this period has a significant impact on water pump operation and corn irrigation demand. Accordingly, the system dynamically adjusts the weight distribution in the coupling model to prioritize the calculation accuracy of the critical path.

[0122] In some optional implementations, an adaptive graph convolutional network (AGCN) can be used instead of a graph attention network to dynamically adjust the spatial attention weights by learning changes in the topological structure of the graph.

[0123] Preferably, the computational complexity of the attention mechanism can be reduced by using the sparse technology, which only retains the connection relationship with attention weight greater than the threshold 0.1, reducing the computational complexity from O(N 2) is reduced to O(NlogN).

[0124] like Figure 3 As shown, according to one aspect of the present application, step S2 is further:

[0125] Step S21: extracting reservoir capacity, river flow, and groundwater level data in agricultural and pastoral areas, calculating the total available water resources and spatiotemporal distribution characteristics for each quarter, and calculating the solar radiation and wind power generation for each quarter based on meteorological station data;

[0126] Water and solar resources in agricultural and pastoral areas vary significantly across seasons. Traditional planting and energy utilization methods fail to fully account for these differences, resulting in resource waste and inefficient utilization.

[0127] Accurately grasp the potential of water resources and wind and solar resources in agricultural and pastoral areas in each season, and provide basic data for crop planting structure optimization and energy scheduling. In this embodiment, a statistical analysis method is used to calculate the total amount of available water resources and the temporal and spatial distribution characteristics, which can intuitively reflect the changing patterns of water resources. Meteorological data and renewable energy power generation models are used to calculate solar radiation and wind power generation, which can accurately assess the potential of wind and solar resources. Specifically:

[0128] Extract water resource data from the agricultural and pastoral database over the past five years, calculate the average available water volume of reservoirs, the average flow of rivers, and the changing trend of groundwater levels in each quarter, and draw a heat map of the spatiotemporal distribution of water resources;

[0129] Collect daily solar radiation data from meteorological stations and use radiation transfer models to calculate the total radiation for each season;

[0130] Based on the wind speed data and the power curve of the wind turbine, the theoretical power generation for each quarter is calculated.

[0131] In a certain embodiment, specifically:

[0132] Collect quarterly monitoring data from water conservancy departments and calculate reservoir capacity by deducting dead storage capacity;

[0133] The river flow was obtained by using the measured flow data of the hydrological station and deducting the ecological flow;

[0134] Based on the groundwater depth and the allowable mining coefficient, the exploitable volume is calculated;

[0135] Calculate the total available water resources in a quarter = (reservoir storage capacity - dead storage capacity) + (river flow - ecological flow) × quarter length + groundwater exploitable volume;

[0136] Collect the quarterly sunshine hours and total horizontal radiation intensity recorded by the weather station, and calculate the quarterly radiation amount = sunshine hours × radiation intensity;

[0137] Collect the quarterly average wind speed measured at the meteorological station and combine it with the wind turbine power curve to calculate the theoretical power generation = wind turbine power × quarterly duration × utilization efficiency, where the utilization efficiency value range is 0.7-0.8;

[0138] Agricultural and pastoral area A covers an area of ​​2,000 square kilometers and includes two reservoirs, one major river, and five groundwater monitoring points;

[0139] In the first quarter of last year, water resources were mainly stored in reservoirs, accounting for 6.8%. River flow decreased by 99.1% due to the ice-covered period, and groundwater accounted for 1.5%. In the summer, river flow accounted for over 90%, and the proportion of reservoir water storage decreased.

[0140] The water resource density in the riverside area is 1.1765 million m³ / km², while in the pastoral areas far from the river, it is only 150,000 m³ / km².

[0141] Sunshine hours in the first quarter: 540h, irradiation intensity: 3.2kWh / m 2 h, calculate the quarterly radiation exposure: 540×3.2=1728kWh / m 2 ;

[0142] The seasonal average wind speed is 6 m / s. The power of the 1.5 MW wind turbine used is 400 kW at a wind speed of 6 m / s. The theoretical power generation is calculated as follows: 400 kW × 24 hours × 90 days × 0.75 (efficiency) = 6.48 million kWh.

[0143] The total amount of water resources in this agricultural and pastoral area was sufficient in the first quarter, but the spatial distribution was uneven and needed to be allocated to the pastoral area through irrigation pipelines. The solar radiation can support 1,000 acres of photovoltaic agricultural greenhouses, and the wind power generation can provide electricity for 500 farmers and herdsmen, with 100W installed capacity per square meter and an average monthly electricity consumption of 500kWh per household.

[0144] Step S22: constructing a crop planting structure optimization model, solving the model to obtain the optimal crop planting combination and area allocation plan, that is, the quarterly planting plan;

[0145] The traditional planting structure in agricultural and pastoral areas is single and does not fully integrate resource conditions and market demand, resulting in low economic benefits and high resource consumption. Formulate scientific planting plans to achieve efficient resource utilization and maximize economic benefits.

[0146] The multi-objective linear programming model can comprehensively consider multiple conflicting objectives and find the optimal solution under resource constraints. The introduction of dynamic parameters such as market prices and policy subsidies can make the model more in line with actual conditions.

[0147] Taking maximizing net benefits and minimizing water consumption and carbon emissions as the objective function, the constraints of land area, total water resources and labor force are set;

[0148] The planting area of ​​crops such as wheat, corn, and soybeans is set as the decision variable. A model is constructed by combining the market price, water demand per unit area, and carbon absorption data of each crop to obtain the quarterly planting planning plan.

[0149] In a certain embodiment, specifically:

[0150] A farming and pastoral area has 1,000 mu of arable land and plans to plant three crops: wheat, corn, and soybeans within a quarter. According to market conditions, wheat earns 1,200 yuan per mu, corn earns 1,000 yuan per mu, and soybeans earns 1,100 yuan per mu. The planting areas of wheat, corn, and soybeans are x1, x2, and x3, respectively. The goal is to maximize the total profit, and the objective function is: MaxZ = 1200x1 + 1000x2 + 1100x3;

[0151] Constraints include:

[0152] Land resource constraint: The total arable land area of ​​the farm is 1000 mu, so x1+x2+x3≦1000;

[0153] Water resource constraints: Planting wheat, corn, and soybeans requires 300 cubic meters, 250 cubic meters, and 200 cubic meters of water per acre, respectively. The total water resources available to the farm in a quarter are 250,000 cubic meters, i.e., 300x1+250x2+200x3≦250,000.

[0154] Labor Constraint: Planting wheat, corn, and soybeans requires 5, 4, and 3 labor days per acre, respectively. The total labor force that the farm can provide per quarter is 4,000 labor days, i.e., 5x1+4x2+3x3≦4,000.

[0155] Non-negative constraints: The crop planting area cannot be negative, x1≧0, x2≧0, x3≧0.

[0156] The optimal solution is: 200 mu of wheat planting area, 400 mu of corn planting area, and 400 mu of soybean planting area, with a maximum total income of 1,180,000 yuan;

[0157] The calculated quarterly planting plan maximizes the farm's total revenue while meeting resource constraints such as land, water resources, and labor. Compared with traditional empirical planting, this plan can allocate resources more scientifically and reasonably and improve the farm's economic benefits.

[0158] Step S23: input the quarterly planting planning scheme into the dynamic coupling model to obtain the corresponding quarterly collaborative scheduling scheme.

[0159] According to one aspect of the present application, in step S2, the calculation of the quarterly planting plan may also be:

[0160] Extract crop growth cycle data and meteorological data from the agricultural and pastoral database to establish a phase relationship model between crop transpiration intensity and photoperiod;

[0161] Calculating physiological phase modulation parameters based on the phase relationship model to generate a micro-irrigation pulse control strategy;

[0162] The physiological phase modulation parameters are integrated into the crop planting structure optimization model, the time-varying water requirement distribution of each crop is recalculated, and a revised quarterly planting plan is output.

[0163] The establishment of the phase relationship model includes:

[0164] The time-varying phase parameter φ_adj is introduced into the traditional Penman-Monteis formula to construct a dynamic evapotranspiration calculation model ET_dynamic(t);

[0165] Establish a mapping relationship between the biological clock regulation coefficient and the crop growth period, and calculate the modulation effect of micro-irrigation pulses on crop physiological rhythms;

[0166] The φ_adj parameter is optimized through a phase-locking algorithm to synchronize the modulated crop water requirement curve with the renewable energy output curve, and the synchronization optimization results are incorporated into the generation process of the quarterly planting plan.

[0167] Preferably, the dynamic evapotranspiration calculation model is:

[0168] ET_dynamic(t)=ET_PM×[1+α_diurnal×cos(2πt / 24+φ_adj)]×F_response(t);

[0169] Among them, ET_PM is the basic evapotranspiration calculated by the Penman-Montes formula, α_diurnal is the diurnal variation coefficient, φ_adj is the adjustable phase angle, and F_response(t) is the impulse response function. The temporal distribution characteristics of ET_dynamic(t) are incorporated into the optimization calculation of crop planting area and variety to generate a quarterly planting planning scheme that takes phase modulation into account.

[0170] Preferably, the parameters of the micro-irrigation pulse control strategy include pulse time and pulse water volume, wherein:

[0171] Pulse time: t_pulse = T_sunrise + (φ_base + Δφ_actual) × 24 / (2π) - Δt_advance;

[0172] Pulse water volume: V_pulse=V_daily×k_pulse×A_plot;

[0173] Where T_sunrise is the sunrise time, φ_base is the crop base phase, Δφ_actual is the actual modulation offset, Δt_advance is the advance, V_daily is the daily crop water requirement, k_pulse is the pulse coefficient, and A_plot is the plot area. The pulse control strategy is used as the basis for correcting the time-varying crop water demand distribution and output to the quarterly planting plan.

[0174] The specific form of the impulse response function F_response(t) is:

[0175] F_response(t)=1+Σ_(i=1)^nR_pulse(t-t_pulse,i)×I_pulse,i;

[0176] Among them, the single pulse response function R_pulse(τ) adopts a double exponential decay model:

[0177] R_pulse(τ)=β(GDD)×A_max×[exp(-τ / τ_decay)-exp(-τ / τ_rise)]×H(τ);

[0178] The meanings of the parameters are as follows: β(GDD) is the biological clock regulation coefficient, which is related to the accumulated temperature GDD; A_max is the maximum response amplitude, ranging from 0.15 to 0.35; τ_decay is the response decay time constant, in hours, ranging from 8 to 15 hours; τ_rise is the response rise time constant, in hours, ranging from 0.8 to 2.5 hours; H(τ) is the unit step function, H(τ) = 1 when τ ≥ 0, and H(τ) = 0 when τ < 0; τ is the time interval from the moment of pulse application, in hours;

[0179] Biological clock regulation coefficient: β(GDD)=k_bio×S_stage(GDD)×S_circadian(t)×(1-D_stress);

[0180] Where: k_bio is the biological clock sensitivity constant, which is 0.28 for corn, 0.19 for wheat, and 0.35 for soybean; S_stage(GDD) is the growth stage sensitivity function: S_stage(GDD)=1-exp(-1.8×GDD / GDD_total); S_circadian(t) is the circadian rhythm intensity factor: S_circadian(t)=0.5×[1+cos(2π(t-7.5) / 24)]; D_stress is the degree of water stress, calculated as: D_stress=max(0,(θ_fc-θ_current) / θ_fc); θ_fc is the field water holding capacity, and θ_current is the current soil moisture content;

[0181] The calculation of the pulse intensity I_pulse,i takes into account the soil moisture status and crop growth stage: I_pulse,i = V_pulse / (V_daily×k_norm)×[1-θ_current / θ_fc]×S_stage(GDD); where V_pulse is the water volume of a single pulse (liters / square meter), V_daily is the daily water requirement of the crop (liters / square meter), and k_norm is the normalization coefficient, which is set to 1.0.

[0182] It should be noted that the pulse time t_pulse,i is determined by the phase locking algorithm:

[0183] t_pulse, i=t_solar_peak-Δt_response-φ_adj×24 / (2π);

[0184] Where t_solar_peak is the peak time of solar power generation (usually 12:30-13:30), Δt_response is the physiological response delay time (1.5-2.8 hours), and φ_adj is the phase modulation angle (-π / 4 to π / 4).

[0185] For example, in a specific application in a corn-growing area, the accumulated temperature on June 15th, GDD, is 1250°C·d, and the total accumulated temperature during the entire growth period, GDD_total, is 2800°C·d. The calculated value is S_stage(1250)=1-exp(-1.8×1250 / 2800)=0.578. At 10:30 AM (t=10.5), the circadian rhythm intensity factor is:

[0186] S_circadian(10.5)=0.5×[1+cos(2π×(10.5-7.5) / 24)]=0.933;

[0187] Field water holding capacity θ_fc=0.35, current soil moisture content θ_current=0.28, and water stress degree D_stress=(0.35-0.28) / 0.35=0.2.

[0188] Calculate the biological clock regulation coefficient: β = 0.28 × 0.578 × 0.933 × (1-0.2) = 0.121;

[0189] Set the pulse parameters: A_max=0.25, τ_decay=10 hours, τ_rise=1.5 hours. After 2 hours of pulse application (τ=2), the single pulse response is:

[0190] R_pulse(2)=0.121×0.25×[exp(-2 / 10)-exp(-2 / 1.5)]=0.121×0.25×[0.819-0.264]=0.0168.

[0191] This response value indicates that pulse irrigation still has a 1.68% promoting effect on crop transpiration rate after 2 hours.

[0192] In some optional embodiments, a three-exponential model may be used instead of a two-exponential model, and a mid-term response term may be added to more accurately describe the physiological response process of the crop.

[0193] Preferably, the pulse intensity can be fine-tuned according to the real-time monitored leaf temperature. When the leaf temperature exceeds 30°C, the pulse intensity is increased by 15-25% to enhance the transpiration cooling effect.

[0194] According to another aspect of the present application, in step S2, the dynamic evapotranspiration calculation model may also be:

[0195] The traditional Penman-Monteis formula is dynamically modified, time-varying phase parameters are introduced, and a dynamic evapotranspiration calculation model is established.

[0196] In this embodiment, the traditional Penman-Monteis formula only considers the static influence of meteorological factors and cannot reflect the temporal periodicity and adjustability of crop transpiration behavior. The present invention introduces a dynamic phase modulation mechanism.

[0197] Specifically, the dynamic evapotranspiration model:

[0198] ET_dynamic(t)=ET_PM×[1+α_diurnal×cos(2πt / 24+φ_adj)]×F_response(t);

[0199] Among them, ET_PM=[Δ(R_n-G)+ρ_a×c_p×(e_s-e_a) / r_a] / [Δ+γ(1+r_s / r_a)]; where Δ is the slope of the saturated water vapor pressure curve, R_n is the net radiation, G is the soil heat flux, ρ_a is the air density, c_p is the constant-pressure specific heat, e_s and e_a are the saturated water vapor pressure and actual water vapor pressure, respectively, r_a is the aerodynamic impedance, r_s is the canopy impedance, γ is the hygrometer constant, α_diurnal is the diurnal variation coefficient with a value range of 0.2-0.5, and φ_adj is the adjustable phase angle. It should be noted that the introduction of φ_adj breaks through the traditional "demand rigidity" assumption and makes the temporal distribution of crop water demand a controllable parameter.

[0200] According to one aspect of the present application, it also includes establishing a mathematical model for crop circadian clock regulation, calculating the response characteristics of the circadian clock to external stimuli, and obtaining the circadian clock regulation coefficient.

[0201] In this embodiment, the biological clock regulation mechanism refers to the ability of the crop's internal circadian rhythm system to respond to external light and water signals. Based on the biological rhythm theory, the present invention establishes a quantitative biological clock response model.

[0202] Specifically, the biological clock regulation coefficient: β(t)=k_bio×S_stage(GDD)×S_circadian(t)×(1-D_stress);

[0203] Among them, the growth stage sensitivity function is: S_stage(GDD)=1-exp(-γ×GDD / GDD_total);

[0204] Where GDD is the current accumulated temperature, GDD_total is the total accumulated temperature during the entire growth period, and γ is the accumulated temperature influence coefficient, which ranges from 1.5 to 2.5.

[0205] Circadian strength factor: S_circadian(t)=0.5×[1+cos(2π(t-t_circadian) / 24)];

[0206] Where t_circadian is the peak time of the biological clock, usually 6:00-8:00.

[0207] Among them, the biological clock sensitivity constant k_bio is determined according to the crop variety: corn k_bio=0.25; wheat k_bio=0.18; soybean k_bio=0.32; rice k_bio=0.15;

[0208] It should be noted that the size of the β value directly determines the response degree of the crop to phase modulation. The larger the β value, the more obvious the phase modulation effect.

[0209] According to one aspect of the present application, it also includes: establishing a transfer function between a micro-irrigation pulse signal and a crop physiological response, and calculating the degree of influence of the pulse stimulation on the biological clock phase, so as to solve the problem that the existing technology ignores the adjustability of the physiological rhythm of crops and regards biological needs as rigid constraints.

[0210] In this embodiment, the micro-pulse signal refers to a short-term water stimulation that is 10-20% lower than the normal irrigation amount, which is used to regulate the physiological rhythm of crops without meeting their main water demand. The present invention establishes a quantitative relationship between the pulse signal and the physiological response.

[0211] Specifically, the impulse response function F_response(t)=1+ΣᵢR_pulse(t-t_pulse,i)×I_pulse,i;

[0212] Among them, the single pulse response function R_pulse(τ)=β×A_max×[exp(-τ / τ_decay)-exp(-τ / τ_rise)]×H(τ); where A_max is the maximum response amplitude (range: 0.1-0.3), τ_decay is the response decay time constant (6-12 hours), τ_rise is the response rise time constant (0.5-2 hours), and H(τ) is the step function.

[0213] Furthermore, the calculation of the pulse intensity I_pulse,i takes into account the soil moisture content and crop growth status:

[0214] I_pulse, i=V_pulse / (V_daily×k_norm)×[1-θ / θ_fc]×S_stage;

[0215] Where V_pulse is the pulse water volume, V_daily is the daily water requirement, k_norm is the normalization coefficient (taken as 1.0), θ is the current soil moisture content, and θ_fc is the field water holding capacity.

[0216] The optimization goal of the pulse time t_pulse,i is to synchronize the modulated transpiration peak with the solar power generation peak:

[0217] t_pulse, i=t_solar_peak-Δt_response-φ_adj×24 / (2π);

[0218] Where Δt_response is the physiological response delay time, usually 1-3 hours.

[0219] It should be noted that the pulse response has a cumulative effect, and the effects of multiple pulses will be superimposed, but the effect of a single pulse decays to a negligible level after 24-48 hours.

[0220] According to another aspect of the present application, a phase locking algorithm is also designed between the crop water demand curve and the renewable energy output curve, and the optimal modulation parameters are determined through iterative optimization.

[0221] In this embodiment, phase locking refers to the control process of synchronizing the crop transpiration peak time with the solar power generation peak time by adjusting the φ_adj parameter and pulse timing. The present invention uses adaptive control theory to achieve precise locking.

[0222] Specifically, the phase locking error function:

[0223] E_lock=∫0 24 φ_crop(t)-φ_solar(t)| 2 dt+λ×∫0 24 [dφ_crop / dt] 2 dt;

[0224] Among them, crop phase:

[0225] φ_crop(t)=arctan[∫ET_dynamic(τ)×sin(2πτ / 24)dτ / ∫ET_dynamic(τ)×cos(2πτ / 24)dτ];

[0226] Solar phase φ_solar(t)=arctan[∫P_solar(τ)×sin(2πτ / 24)dτ / ∫P_solar(τ)×cos(2πτ / 24)dτ];

[0227] Among them, the parameter optimization adopts the gradient descent algorithm:

[0228] φ_adj(k+1)=φ_adj(k)-η×dE_lock / dφ_adj; I_pulse(k+1)=I_pulse(k)-η×dE_lock / dI_pulse;

[0229] Where η is the learning rate, which ranges from 0.01 to 0.05.

[0230] Furthermore, to ensure control stability, the following constraints are introduced:

[0231] |φ_adj|≤π / 3 (phase modulation does not exceed 4 hours); 0.05≤I_pulse≤0.2 (pulse intensity limit);

[0232] E_lock≤ε_threshold (locking accuracy requirement, ε_threshold=0.1).

[0233] It should be noted that the convergence time of the phase locking algorithm is usually 3-5 days, during which continuous monitoring and adjustment are required.

[0234] In some optional implementations, a model predictive control (MPC) algorithm may be used instead of the gradient descent method to adjust the pulse timing in advance by predicting the solar power output changes in the next 24 hours.

[0235] like Figure 4 As shown, according to one aspect of the present application, step S3 is further:

[0236] Step S31: Decompose the quarterly collaborative scheduling plan by day, and construct a daily scheduling time series with a time step of 15 minutes to obtain a daily-scale scheduling model;

[0237] The quarterly scheduling plan is split into daily periods, and each day is divided into 96 time steps. For each time step, the expected flow of the water network, the expected power balance of the power system, and the expected demand parameters of the agricultural load are determined, and a daily-scale scheduling model is constructed. The model includes a water network fluid dynamics sub-model, a power network flow sub-model, and an agricultural load sub-model.

[0238] Step S32: extract several agricultural and pastoral scenarios and input them into the daily scheduling model. Use a distributed computing framework to solve them in parallel to generate a water pump start-stop plan, energy storage charging and discharging strategy, and grid interaction plan for the corresponding scenarios, i.e., a daily scheduling plan set.

[0239] Select 50 typical scenarios from the agricultural and pastoral scenario library of no less than 1,000 scenarios generated in step S12, including different weather and load change scenarios such as sunny days, cloudy days, strong winds, and sudden rainfall. Input these scenario data into the daily-scale scheduling model to generate a daily scheduling plan set including the start and stop status of the water pump at each time step, the charging and discharging power of the energy storage equipment, and the content of the power purchase and sales plan of the power grid.

[0240] Step S33: Construct a three-dimensional evaluation system including total cost, total carbon emissions, and reliability indicators, use the approximate ideal solution ranking method to calculate the relative closeness of each plan to the ideal solution, and select the comprehensive optimal daily scheduling plan.

[0241] A comprehensive multi-dimensional evaluation system comprehensively assesses the pros and cons of each solution and selects the daily scheduling solution that best meets the carbon reduction and efficiency goals. The three-dimensional evaluation system comprehensively covers economic, environmental, and reliability goals. The approach to ideal solution ranking method is data-driven and, in one embodiment, specifically:

[0242] A total cost evaluation index was constructed, including electricity purchase cost, equipment loss cost, maintenance cost, total carbon emission index, and reliability index. The various indicators of 50 daily scheduling plans were normalized, and then the relative closeness of each plan to the ideal solution was calculated based on the approximate ideal solution ranking method. Under the conditions of lowest total cost, low carbon emissions and high reliability, the plan with the highest relative closeness score was selected as the optimal daily scheduling plan.

[0243] like Figure 5 As shown, according to one aspect of the present application, step S4 is further:

[0244] Step S41: Based on the dynamic coupling model and the daily scheduling plan, a water and electricity storage strategy is formulated using mixed integer nonlinear programming;

[0245] Step S42: Build a distributed model, convert the water and electricity storage strategy into control instructions that the distributed model can recognize, clarify the start time, stop time, and operating power of the water pump, the power adjustment parameters of the wind and solar power generation equipment, and the charge and discharge control signals of the energy storage equipment, and input them into the distributed model to adjust the operation of the water pump, wind and solar power generation equipment;

[0246] Step S43: Collect the water level flow, water level changes, grid power fluctuations, and energy storage equipment charging and discharging status data after equipment operation adjustment, and feed them back to the dynamic coupling model to evaluate the system operation status and optimize the quarterly water-electricity-load coordinated scheduling plan and daily scheduling plan to obtain the carbon reduction and efficiency improvement optimization operation plan for the pastoral water network system.

[0247] The collected data is input into the dynamic coupling model and compared with the expected value of the original scheduling plan. If it is found that the actual reservoir water level drops faster than planned, the model will recalculate and adjust the subsequent pump operation plan and energy storage discharge strategy to optimize the daily scheduling plan; if the deviation is large and affects the overall quarterly goal, the quarterly water-electricity-load coordinated scheduling plan will be further optimized. After multiple iterative optimizations, an optimized operation plan for carbon reduction and efficiency improvement of the pastoral water network system that adapts to the current actual operating status is finally obtained.

[0248] According to one aspect of the present application, step S41 is further as follows:

[0249] Step S41a: Taking the maximization of the system's comprehensive benefits as the core goal, a multi-objective function is constructed that includes economic benefits, environmental benefits, and operational reliability, and weight coefficients of carbon trading prices, time-of-use electricity prices, and equipment maintenance cost parameters are determined;

[0250] Step S41b: Combined with the dynamic coupling model, set the upper and lower limits of the water level of the water storage equipment, the state of charge range of the power storage equipment, the power limit for starting and stopping the water pump, and the power balance constraint conditions of the power grid;

[0251] Step S41c: Using a mixed integer nonlinear programming algorithm, the charging and discharging time and water volume of the water storage equipment, and the charging and discharging period and power of the power storage equipment are set as decision variables, and the objective function is solved to obtain the optimal water and power storage strategy.

[0252] like Figure 6 As shown, according to one aspect of the present application, step S5 is further:

[0253] Step S51: Build a real-time monitoring platform for the agricultural and pastoral water network system based on digital twin technology, collect system operation data in real time through IoT devices, and map the actual operation status in the digital twin model in real time;

[0254] Digital twin technology can create virtual models that are highly consistent with physical systems, enabling real-time dynamic mapping; IoT technology can enable the automatic collection and transmission of device data;

[0255] Based on the actual layout and parameters of water networks, power systems and agricultural facilities in rural and pastoral areas, three-dimensional modeling is carried out in a virtual environment. Data is collected in real time through sensors deployed on site and transmitted to the monitoring platform. The platform associates the data with the digital twin model and updates the operating status, water flow path and power flow direction of the equipment in the model in real time. It is intuitively displayed in the form of three-dimensional animation and charts. When a water pump fails, the corresponding water pump icon in the digital twin model will turn red and flash, and the fault information will be displayed.

[0256] Step S52: construct an adaptive feedback control model, use a deviation prediction algorithm to predict the deviation between the system operating state and the scheduling plan, and use an adaptive robust control algorithm to dynamically adjust the system operating parameters based on the prediction results;

[0257] Step S53: Construct a multi-dimensional economic evaluation system including economic indicators, environmental indicators, energy indicators, and reliability indicators, use the entropy weight-grey correlation analysis method to conduct a comprehensive evaluation of the system, and generate a control strategy and carbon reduction and efficiency improvement evaluation report;

[0258] In a certain embodiment, specifically:

[0259] Based on the four dimensions of economy, environment, energy, and reliability, specific indicators are designed in combination with industry characteristics:

[0260] Economic indicators: investment cost, operating expenses, internal rate of return, and payback period;

[0261] Environmental indicators: carbon emission intensity, pollutant emission compliance rate, and renewable energy consumption rate;

[0262] Energy indicators: comprehensive energy efficiency, energy consumption per unit of output value, peak-valley load difference rate;

[0263] Reliability indicators: system average power outage time, power supply reliability, equipment failure rate;

[0264] The extreme value method is used to eliminate the dimension effect, calculate the index proportion, calculate the entropy value and determine the weight;

[0265] Determine the reference sequence X0={x01, x02, …, x0m} (ideal value) and calculate the correlation coefficient and correlation degree;

[0266] Optimize investment plans through equipment life cycle cost analysis, such as reducing energy costs through preheater modification;

[0267] Implement carbon capture and storage technologies, such as promoting low-nitrogen burners to reduce pollutant emissions;

[0268] Introducing smart microgrids to dynamically regulate energy distribution. In one implementation, the Whale Optimization algorithm improved the system COP by 18%, reducing carbon emissions by 1,200 tons annually.

[0269] By combining a multi-dimensional economic evaluation system with entropy weight-grey correlation analysis, system shortcomings can be systematically identified and precise control strategies can be generated. This method can significantly improve energy efficiency and reduce emissions in the industrial, power, and manufacturing sectors.

[0270] In this embodiment, the objective function of the mixed integer nonlinear programming algorithm is:

[0271] maxF=w1×R_economic+w2×R_environment+w3×R_reliability;

[0272] Among them, R_economic is the economic benefit, R_environment is the environmental benefit, R_reliability is the operational reliability, w1, w2, and w3 are weight coefficients, and the constraints include the upper and lower limits of the water level of the water storage equipment, the charge state range of the power storage equipment, and the power balance constraint of the power grid. The optimization solution results are used as the basis for generating control instructions.

[0273] Step S54: Based on the control strategy and the carbon reduction and efficiency improvement assessment report, the Bayesian optimization algorithm is used to optimize and adjust the parameters in the water-power-load dynamic coupling model.

[0274] According to one aspect of the present application, step S52 is further as follows:

[0275] Step S52a: Construct a data comparison module to extract the real-time collected water network flow, wind and solar power generation, and energy storage charge state data, compare them with the expected values ​​in the daily scheduling plan and water and power storage strategy on a time-by-time basis, use a sliding window algorithm to process the deviation data, and extract the deviation change trend and change rate;

[0276] Step S52b: using the recursive least squares method to estimate system parameters and adjust the pump power-flow model parameters in real time;

[0277] Step S52c: Based on the deviation characteristics and the calculation results of the control algorithm, generate water pump speed adjustment and wind, solar and storage device power adjustment instructions, and send the control instructions to the device controller. After the device executes, it will provide real-time feedback on the execution status.

[0278] The carbon reduction and efficiency improvement optimization operation method and system for the rural water network system proposed in this invention innovatively integrates the concept of multi-energy synergy with dynamic feedback control. Through a data-driven hierarchical optimization and scheduling system, it accurately solves the industry problems of low resource utilization efficiency and high carbon emissions in rural and pastoral areas. From multi-dimensional data fusion modeling to closed-loop dynamic optimization, every link demonstrates a deep understanding of and innovative breakthroughs in complex systems. It not only realizes the intelligent coordination of water resources, power resources and agricultural loads, improves the utilization rate of renewable energy and the reliability of system operation, but also provides efficient and low-carbon solutions for the sustainable development of rural and pastoral areas through scientific evaluation and iteration mechanisms.

[0279] According to another aspect of the present application, a carbon reduction and efficiency-enhancing optimization operation system for a water network system in agricultural and pastoral areas is provided, characterized by comprising:

[0280] at least one processor; and

[0281] a memory communicatively connected to at least one of the processors; wherein,

[0282] The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement any of the above-mentioned methods for optimizing operation of carbon reduction and efficiency improvement of agricultural and pastoral water network systems.

[0283] The preferred embodiments of the present invention are described in detail above. However, the present invention is not limited to the specific details in the above embodiments. Within the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all fall within the scope of protection of the present invention.

Claims

1. A method for optimizing the operation of water network systems in agricultural and pastoral areas to reduce carbon emissions and increase efficiency, characterized by: The steps include: Step S1: Collect agricultural and pastoral area data, build an agricultural and pastoral area database, construct a dynamic coupling model, and optimize model parameters; Step S2: construct a crop planting structure optimization model, solve the model to obtain a quarterly planting planning scheme, and input it into the dynamic coupling model to obtain the corresponding quarterly collaborative scheduling scheme; Step S3: Build a daily-scale scheduling model based on the quarterly collaborative scheduling plan, solve the model to obtain a set of daily scheduling plans, and select the optimal daily scheduling plan using cost and carbon emissions as indicators; Step S4: Formulate a water and electricity storage strategy and use it as input to the pre-built distributed model to adjust the operation of water pumps, wind, solar and storage equipment to obtain an optimized operation plan for carbon reduction and efficiency improvement of the pastoral water network system; Step S5: Build a real-time monitoring platform for the agricultural and pastoral water network system based on digital twin technology, perform real-time mapping of the actual operating status, build an adaptive feedback control model, predict the deviation between the system operating status and the scheduling plan, and dynamically adjust the system operating parameters based on the prediction results. Generate a control strategy and carbon reduction and efficiency improvement assessment report, and optimize and adjust the parameters in the water-power-load dynamic coupling model; The step S1 is further as follows: Step S11: Collect agricultural and pastoral area data and build an agricultural and pastoral area database; Step S12: extracting data on water networks, wind and solar power station outputs, and crop water requirements, fitting the distributions of water networks, wind and solar power station outputs, and crop water requirements, respectively, and constructing a joint distribution of the three, and randomly sampling to obtain several agricultural and pastoral scenarios; Step S13: Integrate the hydrodynamic model of the agricultural and pastoral water network, the power network flow model of the agricultural and pastoral area, and the dynamic calculation model of crop water demand, define input variables and output variables, introduce a spatiotemporal attention mechanism, and for time series data, use a multi-head self-attention mechanism to calculate the attention weights between data at different time steps, highlighting the influence of recent data and data at key time points; For spatial node data, a graph attention network is used to calculate the attention weights between water network nodes, power network nodes, and agricultural area nodes to reflect the interactions between various spatial elements. The coupling model is optimized through training data, and the model parameters and attention weights are adjusted to enable the model to accurately predict the operating status of the water-power-load system under different scenarios, thereby constructing a dynamic coupling model. Step S14: inputting several agricultural and pastoral area scenarios into the dynamic coupling model, training and optimizing the model parameters; The step S2 is further as follows: Step S21: extracting reservoir capacity, river flow, and groundwater level data in agricultural and pastoral areas, calculating the total available water resources and spatiotemporal distribution characteristics for each quarter, and calculating the solar radiation and wind power generation for each quarter based on meteorological station data; Step S22: constructing a crop planting structure optimization model, solving the model to obtain the optimal crop planting combination and area allocation plan, i.e., a quarterly planting plan, extracting crop growth cycle data and meteorological data from the agricultural and pastoral area database, establishing a phase relationship model between crop transpiration intensity and photoperiod, calculating physiological phase modulation parameters based on the phase relationship model, generating a micro-irrigation pulse control strategy, integrating the physiological phase modulation parameters into the crop planting structure optimization model, recalculating the time-varying water requirement distribution of each crop, and outputting a revised quarterly planting plan; Step S23: input the quarterly planting planning scheme into the dynamic coupling model to obtain the corresponding quarterly water-electricity-load coordinated scheduling scheme.

2. The carbon reduction and efficiency improvement optimization operation method of the agricultural and pastoral water network system according to claim 1 is characterized in that: The step S13 is further as follows: Step S13a: Retrieve water network data from the agricultural and pastoral database, use pipelines, water pumps, and reservoirs as nodes and connecting edges, calculate and draw the pump head-flow curve, calculate the head loss along the process, and construct a hydrodynamic model of the agricultural and pastoral water network; Step S13b: retrieve wind and solar power station data from the agricultural and pastoral area database, and construct a relationship model between the output power of the photovoltaic power station and the light intensity and temperature, a relationship model between the wind power power and wind speed, and a charge and discharge efficiency curve of the energy storage device, to obtain a power network flow model for the agricultural and pastoral area; Step S13c: determining the corresponding crop coefficient based on the crop growth cycle, calculating the crop water requirement using the Penman-Montes formula in combination with meteorological data, and constructing a dynamic calculation model for crop water requirement; Step S13d, coupling the hydrodynamic model of the agricultural and pastoral water network, the power network flow model of the agricultural and pastoral area, and the dynamic calculation model of crop water demand, introducing the spatiotemporal attention mechanism, calculating the attention weights of the time series and spatial nodes and configuring the model, and obtaining the water-electricity-load dynamic coupling model.

3. The carbon reduction and efficiency improvement optimization operation method of the agricultural and pastoral water network system according to claim 1 is characterized in that: The step S3 is further as follows: Step S31: Decompose the quarterly water-power-load coordinated scheduling plan by day, and construct a daily scheduling time series with a time step of 15 minutes to obtain a daily-scale scheduling model; Step S32: extract several agricultural and pastoral scenarios and input them into the daily scheduling model. Use a distributed computing framework to solve them in parallel to generate a water pump start-stop plan, energy storage charging and discharging strategy, and grid interaction plan for the corresponding scenarios, i.e., a daily scheduling plan set. Step S33: Construct a three-dimensional evaluation system including total cost, total carbon emissions, and reliability indicators, use the approximate ideal solution ranking method to calculate the relative closeness of each plan to the ideal solution, and select the comprehensive optimal daily scheduling plan.

4. The method for optimizing the operation of the water network system in agricultural and pastoral areas for carbon reduction and efficiency improvement according to claim 1, characterized in that: The step S4 is further as follows: Step S41: Based on the dynamic coupling model and the daily scheduling plan, a water and electricity storage strategy is formulated using mixed integer nonlinear programming; Step S42: Build a distributed model, convert the water and electricity storage strategy into control instructions that the distributed model can recognize, clarify the start time, stop time, and operating power of the water pump, the power adjustment parameters of the wind and solar power generation equipment, and the charge and discharge control signals of the energy storage equipment, and input them into the distributed model to adjust the operation of the water pump, wind and solar power generation equipment; Step S43: Collect the water level flow, water level changes, grid power fluctuations, and energy storage equipment charging and discharging status data after equipment operation adjustment, and feed them back to the dynamic coupling model to evaluate the system operation status and optimize the quarterly water-electricity-load coordinated scheduling plan and daily scheduling plan to obtain the carbon reduction and efficiency improvement optimization operation plan for the pastoral water network system.

5. The method for optimizing the operation of the water network system in agricultural and pastoral areas for carbon reduction and efficiency improvement according to claim 4 is characterized in that: The step S41 is further as follows: Step S41a: Taking the maximization of the system's comprehensive benefits as the core goal, a multi-objective function is constructed that includes economic benefits, environmental benefits, and operational reliability, and weight coefficients of carbon trading prices, time-of-use electricity prices, and equipment maintenance cost parameters are determined; Step S41b: Combined with the dynamic coupling model, set the upper and lower limits of the water level of the water storage equipment, the state of charge range of the power storage equipment, the power limit for starting and stopping the water pump, and the power balance constraint conditions of the power grid; Step S41c: Using a mixed integer nonlinear programming algorithm, the charging and discharging time and water volume of the water storage equipment, and the charging and discharging period and power of the power storage equipment are set as decision variables, and the objective function is solved to obtain the optimal water and power storage strategy.

6. The method for optimizing the operation of the water network system in agricultural and pastoral areas for carbon reduction and efficiency improvement according to claim 1, characterized in that: The step S5 is further as follows: Step S51: Build a real-time monitoring platform for the agricultural and pastoral water network system based on digital twin technology, collect system operation data in real time through IoT devices, and map the actual operation status in the digital twin model in real time; Step S52: construct an adaptive feedback control model, use a deviation prediction algorithm to predict the deviation between the system operating state and the scheduling plan, and use an adaptive robust control algorithm to dynamically adjust the system operating parameters based on the prediction results; Step S53: Construct a multi-dimensional economic evaluation system including economic indicators, environmental indicators, energy indicators, and reliability indicators, use the entropy weight-grey correlation analysis method to conduct a comprehensive evaluation of the system, and generate a control strategy and carbon reduction and efficiency improvement evaluation report; Step S54: Based on the control strategy and the carbon reduction and efficiency improvement assessment report, the Bayesian optimization algorithm is used to optimize and adjust the parameters in the water-power-load dynamic coupling model.

7. The method for optimizing the operation of the water network system in agricultural and pastoral areas for carbon reduction and efficiency improvement according to claim 6, characterized in that: The step S52 is further as follows: Step S52a: Construct a data comparison module to extract the real-time collected water network flow, wind and solar power generation, and energy storage charge state data, compare them with the expected values ​​in the daily scheduling plan and water and power storage strategy on a time-by-time basis, use a sliding window algorithm to process the deviation data, and extract the deviation change trend and change rate; Step S52b: using the recursive least squares method to estimate system parameters and adjust the pump power-flow model parameters in real time; Step S52c: Based on the deviation characteristics and the calculation results of the control algorithm, generate water pump speed adjustment and wind, solar and storage device power adjustment instructions, and send the control instructions to the device controller. After the device executes, it will provide real-time feedback on the execution status.

8. The carbon reduction and efficiency improvement optimization operation system of the agricultural and pastoral water network system is characterized by: include: at least one processor; as well as a memory communicatively connected to at least one of the processors; wherein, The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement the carbon reduction and efficiency improvement optimization operation method of the agricultural and pastoral water network system as described in any one of claims 1 to 7.

Images