Agricultural and pastoral light complementary three-dimensional system in high and cold steppe region and construction method thereof

Abstract

The invention discloses a high and cold grassland area agriculture and pasture light complementary three-dimensional system and a construction method thereof, and relates to the technical field of new energy agriculture comprehensive utilization. A sensor network covering the whole area and a grassland dynamic growth model based on LSTM are constructed, and through multi-dimensional sensors of illumination, soil and weather, the grassland dynamic growth model is constructed; the spatial heterogeneity of a non-uniform photo-thermal environment under a photovoltaic panel is perceived and digitally reproduced in real time, the LSTM model learns and memorizes a complex nonlinear relationship between an environment sequence and biomass growth through a unique gating mechanism of the LSTM model, and a growth suitability index of a specific partition in the future can be predicted in advance; the predicted value drives a decision library to execute a prospective decision, water and fertilizer intervention is performed on a stress partition in advance, and the transformation from passive reaction to active prediction realizes extreme utilization and conservation of grassland resources in time and space dimensions.

Description

Technical Field

[0001] This invention relates to the field of new energy agricultural comprehensive utilization technology, specifically to a three-dimensional system of agricultural and pastoral complementary photovoltaic system and its construction method in alpine grassland areas. Background Technology

[0002] The ecosystems of alpine grasslands are fragile, and the carrying capacity of pastures is limited. Traditional extensive animal husbandry easily leads to grassland degradation and soil desertification, highlighting the contradiction between economic development and ecological protection. At the same time, the region possesses abundant solar resources, providing excellent natural conditions for the development of the photovoltaic industry. As the "photovoltaic + agriculture" model is introduced to the region as a new form of integrated land use, it aims to achieve the dual benefits of green electricity and agricultural output by generating electricity from photovoltaic panels above and carrying out agricultural and livestock production below. However, this simple physical superposition model faces severe challenges in practical application. The core issue lies in the complex interaction between the unique ecological environment of alpine grasslands and the operation of photovoltaic facilities. If not handled properly, it may not only fail to achieve complementarity but also exacerbate ecological risks.

[0003] Existing photovoltaic agricultural and pastoral systems have significant shortcomings, mainly in two aspects: First, they lack refined and intelligent management of the microclimate environment under the photovoltaic panels. The shading effect of the photovoltaic panels leads to severe uneven distribution of light and moisture under the panels, forming unique patchy microenvironments. Current technologies often only plant a single shade-tolerant crop or forage grass under the panels, failing to plan and precisely control the zones based on the gradient differences of environmental factors such as light, water, air, and heat, resulting in low efficiency in the utilization of land and solar energy resources. Second, the three links of photovoltaic power generation, agricultural planting, and livestock breeding usually operate independently, lacking data interaction and coordination mechanisms, which seriously restricts the promotion efficiency and ecological benefits of this model in high-altitude grassland areas.

[0004] Therefore, there is an urgent need in this field for a three-dimensional system that can deeply integrate intelligent management technology and truly realize the efficient and coordinated operation of photovoltaic power generation, ecological planting and smart breeding, so as to ensure the benefits of photovoltaic power generation while promoting the healthy restoration of grasslands and the sustainable development of animal husbandry. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a three-dimensional system for complementary agriculture and animal husbandry with photovoltaics in alpine grassland areas and its construction method. It can construct a sensor network covering the entire area and a grassland dynamic growth model based on LSTM. Through multi-dimensional sensors of light, soil, and weather, it can perceive and digitally reproduce the spatial heterogeneity of the non-uniform light and heat environment under photovoltaic panels in real time. The LSTM model, through its unique gating mechanism, learns and memorizes the complex nonlinear relationship between environmental sequence and biomass growth, and can predict the growth suitability index of specific areas in advance. This prediction value drives the decision base to make forward-looking decisions and pre-intervene water and fertilizer in stressed areas. This transformation from passive reaction to active prediction realizes the ultimate utilization and conservation of grassland resources in both time and space.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: On the one hand, a three-dimensional system for complementary agriculture and animal husbandry in alpine grassland areas. The system is constructed in alpine grassland areas and adopts a spatial three-dimensional structure, which includes, in sequence, a photovoltaic layer for converting solar energy into electrical energy, a planting layer for agricultural production, and a breeding layer for animal husbandry.

[0007] The system also includes an integrated intelligent management platform for coordinating the operation of the photovoltaic layer, planting layer, and aquaculture layer.

[0008] The intelligent management platform includes an environmental monitoring module, a data analysis and decision-making module, and an execution module, wherein:

[0009] The environmental monitoring module consists of various sensors deployed at different locations within the system space, used to collect physical environmental parameters and biological activity parameters in real time and continuously.

[0010] The data analysis and decision-making module is connected to the environmental monitoring module and integrates a grassland dynamic growth model, livestock behavior units, and a decision database. The grassland dynamic growth model is built based on machine learning algorithms, using processed environmental data as input to predict and output the water suitability index for different zones within the planting layer in real time. The livestock behavior unit determines the amount of exercise and health status of livestock in the breeding layer based on livestock location and activity data. The decision base receives the output results from the pasture dynamic growth model and the livestock behavior unit, performs fusion analysis, and generates a comprehensive control strategy.

[0011] The execution module is connected to the data analysis and decision-making module and is used to physically execute the decision base regulation in the data analysis and decision-making module;

[0012] The photovoltaic layer, planting layer, and breeding layer form a three-dimensional agricultural-livestock-solar complementary system through the intelligent management platform.

[0013] Furthermore, the sensor network in the environmental monitoring module adopts a heterogeneous networking approach, including:

[0014] Light monitoring nodes are distributed under the photovoltaic layer, and each node integrates a photosynthetically active radiation sensor and a light intensity sensor;

[0015] Soil moisture monitoring nodes are embedded in the planting layer soil. Each node integrates a soil temperature and humidity sensor, a soil pH sensor, and a soil electrical conductivity sensor.

[0016] The meteorological monitoring sensor array installed in the system integrates air temperature and humidity sensors, wind speed and direction instruments, rain gauges and atmospheric pressure sensors.

[0017] The livestock vital signs monitoring tags, which are attached to the livestock in the breeding layer, integrate GPS / BeiDou positioning, a three-axis accelerometer, and a body temperature sensor.

[0018] Furthermore, the grassland dynamic growth model in the data analysis and decision-making module is a machine learning model based on multivariate time series prediction, which includes a Long Short-Term Memory (LSTM) network structure with at least one hidden layer. The structure consists of an input layer, at least one LSTM hidden layer, and a fully connected output layer.

[0019] The input layer receives a multi-dimensional time-series environmental feature vector uploaded by the environmental monitoring module;

[0020] The hidden layer contains three gating mechanisms: a forget gate, an input gate, and an output gate, which are used to learn and remember the long-term dependency between environmental factors and grassland growth.

[0021] The fully connected output layer maps the hidden layer output state at the last time step to a single scalar value, namely the moisture suitability index for the next τ days. in, h represents the water suitability index predicted τ days from now at time step t. t W is the hidden state vector output at time step t after computation by the LSTM hidden layer. y It is a weight vector used to weight the high-dimensional hidden state h. t A linear mapping to a scalar value, b y It is a bias term that inputs the water suitability index calculated by the grassland dynamic growth model into the decision library for irrigation regulation.

[0022] Furthermore, the livestock behavior unit in the data analysis and decision-making module takes data from the environmental monitoring module as input, including the livestock's GPS / BeiDou positioning coordinates P. t Triaxial acceleration data At and body temperature data Calculate and output livestock behavioral characteristics:

[0023] Motion volume calculation: based on acceleration data A t Calculate the sum of the magnitudes of the dynamic acceleration vectors within each time window, as an indicator of motion intensity. Among them, a x,i a y,i a z,i The original values ​​of the triaxial acceleration within one sampling period are given, and n is the number of sampling points within a time window.

[0024] Foraging activity identification: based on location data P t Calculate the displacement distance D of the livestock per unit time;

[0025] Health status assessment: based on body temperature data Determine if livestock have abnormal body temperatures.

[0026] Furthermore, the decision base is linked to the predicted water suitability index output by the grassland dynamic growth model. Based on the behavioral characteristics output by the livestock behavior model, a comprehensive control strategy is implemented:

[0027] Water and fertilizer regulation in grazing areas based on moisture suitability index: The decision base maintains a list of predicted moisture suitability indices for all zones, and assigns each zone... Compared with the preset moisture suitability index threshold Y th When comparing, When the system determines that the area is under water and nutrient stress and has insufficient growth suitability, it generates irrigation and fertilization control instructions for the planting area and initiates water and fertilizer operations.

[0028] Grazing strategy regulation based on livestock behavior: The decision base receives the daily displacement distance and body temperature status of livestock from the livestock behavior unit. When the displacement distance of livestock is detected to be lower than the average of livestock of the same type, the decision base, in conjunction with the execution module, drives the livestock guidance unit to guide the livestock to the backup zone. When abnormal body temperature is detected in the livestock group, a special instruction is immediately generated to drive the sound and light induction device of the execution module to isolate the individual with abnormal body temperature to the designated observation area, and at the same time send an alarm message.

[0029] The comprehensive control measures generated by the decision base are sent to the execution module for execution.

[0030] Furthermore, the execution module includes:

[0031] The livestock guiding unit consists of a movable fence, an electronic access control system, and an audio-visual guidance device. After receiving the control strategy from the decision database, it reconstructs the physical boundary of the breeding layer by moving the fence and opening and closing the electronic access control system, and at the same time uses the audio-visual guidance device to guide and drive the livestock herd.

[0032] The agricultural operation unit consists of a zone-controlled irrigation solenoid valve group. After receiving the irrigation and fertilization control commands, it controls the solenoid valve group to perform precise and quantitative irrigation on the planting zone.

[0033] Furthermore, when the agricultural operation unit performs quantitative irrigation, it follows this logic:

[0034] The daily reference crop evapotranspiration is calculated by calling the moisture suitability index output by the grassland dynamic growth model. Among them, R n Let G be the net surface radiation, T be the soil heat flux density, T be the daily average temperature, u2 be the wind speed, and e be the average temperature. s e is the saturated vapor pressure. a Δ is the actual water vapor pressure, Δ is the slope of the saturated water vapor pressure curve, and γ is the wet / dry gauge constant.

[0035] Determine the crop coefficient K based on the crop type. c And calculate the actual water requirement ET of the crop. c =K c ×ET0 and daily water deficit D w =max(0,(ET) c -P eff )×A×10), where P eff The effective precipitation for the day is represented by A, and the area of ​​the irrigated zone is represented by A.

[0036] The decision base will include the water deficit D w With the preset irrigation trigger threshold D th Compare; when D w >D th If so, an irrigation instruction is generated. The irrigation instruction is structured data and includes: target irrigation zone number, irrigation start time, and irrigation duration.

[0037] On the other hand, the method for constructing a three-dimensional agricultural-pastoral complementary solar system in alpine grassland regions includes the following specific steps:

[0038] S100, Systematic Survey and Design Phase: Conduct a detailed survey of the target alpine grassland site; based on the survey data, optimize the layout design of the photovoltaic array, and simultaneously complete the zoning planning of the planting layer, the architecture design of the intelligent management platform, and the wiring planning of the infrastructure.

[0039] S200, Infrastructure Layered Construction Phase: Install photovoltaic brackets and modules according to the design, construct the photovoltaic layer, lay irrigation pipe network, and deploy sensor network;

[0040] S300, Planting Layer Implementation and Deployment Phase: According to the zoning plan, plant plants with different characteristics, and complete the installation, power supply and network connection of all sensors and actuators;

[0041] S400, Platform Integration and Optimization Phase: Establish a unified communication protocol, input local historical climate data, soil data and livestock behavior data, and train and calibrate the grassland dynamic growth model and rule parameters in the decision base.

[0042] S500, System Integration and Operation Phase: Livestock are introduced, the entire system is started, the intelligent management platform begins operation, and the system achieves intelligent and adaptive operation.

[0043] Compared with existing technologies, this integrated agricultural-pastoral solar power system and its construction method in alpine grassland regions have the following advantages:

[0044] I. This invention constructs a sensor network covering the entire area and a grassland dynamic growth model based on LSTM. Through multi-dimensional sensors of light, soil, and weather, it perceives and digitally reproduces the spatial heterogeneity of the non-uniform light and heat environment under photovoltaic panels in real time. The LSTM model, through its unique gating mechanism, learns and memorizes the complex nonlinear relationship between environmental sequences and biomass growth, and can predict the growth suitability index of specific zones in advance. This prediction drives the decision base to make forward-looking decisions and intervene in water and fertilizer in advance in stressed zones. This shift from passive reaction to active prediction realizes the ultimate utilization and conservation of grassland resources in both time and space.

[0045] Second, this invention achieves deep coupling and closed-loop flow of matter and energy among four subsystems—photovoltaics, planting, aquaculture, and water and fertilizer—through an intelligent management platform. It redistributes and coordinates the three-dimensional space and resources. The photovoltaic layer not only generates electricity, but its surface also serves as a rain collection surface, and the space beneath the panels forms a microenvironment suitable for the growth of shade-tolerant crops. The planting layer is vertically partitioned according to the light environment gradient to maximize the efficiency of light energy utilization. The livestock activities in the aquaculture layer are guided by the intelligent system and accurately fed back to the planting layer under the instructions of the decision-making module. The entire process is data-driven, forming a complete ecological cycle chain that is interdependent and mutually reinforcing, achieving the internal balance and self-sustaining of the entire system's material cycle and energy flow.

[0046] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0048] Figure 1 A flowchart of a solar-agricultural complementary system in alpine grassland regions;

[0049] Figure 2 A flowchart illustrating the construction steps of an integrated agricultural-pastoral-solar hybrid system in alpine grassland regions;

[0050] Figure 3 This is a schematic diagram of the overall structure of a three-dimensional system for complementary agriculture and animal husbandry in alpine grassland regions. Detailed Implementation

[0051] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0052] Example 1

[0053] This embodiment addresses the characteristics of high-altitude grassland regions, such as fragile ecosystems, limited pasture carrying capacity, and abundant sunlight resources. It constructs a three-dimensional agricultural-pastoral solar-powered complementary system. This system adopts a spatial three-dimensional structure, including a photovoltaic layer, a planting layer, a livestock layer, and an integrated intelligent management platform. The intelligent management platform collects multi-dimensional parameters in real time through an environmental monitoring module, predicts the water suitability index using the LSTM pasture dynamic growth model in the data analysis and decision-making module, analyzes livestock status through livestock behavior units, generates comprehensive control strategies from a decision database, and finally executes the control operations through an execution module. This achieves deep synergy between photovoltaic power generation, agricultural planting, and livestock breeding, improving resource utilization efficiency and promoting a win-win situation for ecological protection and economic development.

[0054] (I) Overall System Setup

[0055] The photovoltaic layer, planting layer, and breeding layer form a closed-loop feedback system through the intelligent management platform. The arrangement structure of the photovoltaic layer determines the heterogeneity of the light microenvironment of the planting layer. The environmental monitoring module senses this heterogeneous environment and the biological responses within it. The data analysis and decision-making module analyzes the data and makes decisions. The execution module acts on the breeding and planting layers, changing livestock behavior and plant growth conditions. This change is then sensed again by the environmental monitoring module, thereby achieving adaptive optimization and steady-state operation of the entire system.

[0056] Photovoltaic layer: The photovoltaic support adopts a special reinforced structural design, which has excellent wind pressure and snow load resistance. The height of the support is customized according to the balance between the needs of the lower agricultural and pastoral activity space and the photovoltaic power generation efficiency. The tilt angle of the support is determined through professional analysis of the local solar radiation characteristics to ensure that the photovoltaic panels receive the best sunlight. The spacing of the photovoltaic array is calculated by scientific light simulation to ensure that the lower planting area receives sufficient sunlight duration.

[0057] Planting layer: Select cold-resistant, drought-resistant, and shade-tolerant specialty economic crops for planting, including high-quality forage grasses and medicinal plants such as Notopterygium incisum and Astragalus membranaceus. Equipped with advanced irrigation solenoid valve sets, combined with precise irrigation commands, dynamic irrigation is implemented according to the crop growth needs and soil moisture.

[0058] Aquaculture layer: Utilizing the shaded space created by photovoltaic panels, ecological free-range grazing of poultry or small livestock can be carried out to achieve an ecological cycle of planting and breeding.

[0059] (II) Operational Logic of the Intelligent Management Platform (e.g.) Figure 1 (As shown)

[0060] The environmental monitoring module consists of multiple sensors deployed at different locations within the system space. Data acquisition is performed using a heterogeneous networking approach. The light monitoring nodes distributed under the photovoltaic layer have photosynthetically active radiation sensors that collect the photosynthetically active radiation intensity at different locations under the photovoltaic layer in real time. This parameter reflects the light energy that plants can utilize for photosynthesis and is an important factor affecting plant growth. The illuminance sensor collects light intensity data to analyze the light distribution in different areas.

[0061] Soil moisture monitoring nodes are embedded in the planting layer soil. Soil temperature and humidity sensors collect temperature and humidity data at different soil depths. Soil temperature affects the growth and absorption function of plant roots, while soil humidity is directly related to the water supply of plants. Soil pH sensors collect soil acidity and alkalinity data.

[0062] The meteorological monitoring sensor array installed in the system includes air temperature and humidity sensors that collect air temperature and humidity data. Air temperature affects plant transpiration and respiration, while air humidity is related to plant water evaporation and disease occurrence. Anemometers collect wind speed and direction data. Wind speed affects plant transpiration and pollen dispersal, while wind direction has a certain impact on pollutant diffusion and the distribution of meteorological conditions. Rain gauges collect precipitation data to calculate effective precipitation, providing a basis for irrigation decisions. Atmospheric pressure sensors collect atmospheric pressure data. Changes in atmospheric pressure are related to weather changes and can be used as a reference factor for meteorological forecasting.

[0063] Animal health monitoring tags attached to the livestock in the breeding layer use GPS / BeiDou positioning to collect the animals' location coordinates in real time, which is used to analyze the animals' activity range and displacement distance; a three-axis accelerometer collects the animals' acceleration data, and the amount of movement of the animals is measured by calculating the sum of the magnitudes of the acceleration vectors; and a body temperature sensor collects the animals' body temperature data, which is used to determine the animals' health status.

[0064] The environmental monitoring module preprocesses the collected raw data, including data cleaning, noise reduction, and format conversion, to remove abnormal data and interference signals and convert the data into a unified format for subsequent processing by the data analysis and decision-making module. After preprocessing, the processed environmental data and biological activity parameters are transmitted to the data analysis and decision-making module in real time.

[0065] The data analysis and decision-making module receives processed data transmitted from the environmental monitoring module. This module integrates a grassland dynamic growth model, livestock behavior units, and a decision-making database.

[0066] The grassland dynamic growth model operates as follows: Based on a multivariate time-series prediction machine learning model, it employs a Long Short-Term Memory (LSTM) network structure with at least one hidden layer. This structure consists of an input layer, at least one LSTM hidden layer, and a fully connected output layer. The input layer receives multi-dimensional time-series environmental feature vectors uploaded by the environmental monitoring module. These vectors include time-series data such as light monitoring data (photosynthetically active radiation intensity, illuminance), soil moisture data (soil temperature and humidity, soil pH, soil electrical conductivity), and meteorological data (air temperature and humidity, wind speed, wind direction, precipitation, atmospheric pressure). The input layer passes these multi-dimensional time-series environmental feature vectors to the LSTM hidden layer. The LSTM hidden layer contains three gating mechanisms: a forget gate, an input gate, and an output gate. The forget gate determines which information from the previous hidden state to discard based on the input environmental feature vector and the previous hidden state. For example, for... For short-term environmental fluctuations that have little long-term dependence on grassland growth, the forget gate discards them. The input gate determines which new information to store in the cell state based on the input environmental feature vector and the hidden state of the previous time step. For example, when key environmental factors such as light and water change significantly, the input gate will include this important information in the cell state. The output gate determines which information to output as the hidden state of the current time step based on the input environmental feature vector and the current cell state. This hidden state contains information about the long-term dependence between environmental factors and grassland growth. Through these three gating mechanisms, the LSTM hidden layer can learn and remember the long-term dependence between environmental factors and grassland growth, effectively handling the long-term dependence problem in multivariate time series data. The fully connected output layer maps the hidden layer output state of the last time step to a single scalar value, namely the water suitability index for the next τ days, which is calculated as: Y t =W y ·h t +b y , where Y t h represents the water suitability index predicted τ days from now at time step t. t W is the hidden state vector output at time step t after computation by the LSTM hidden layer. y It is a weight vector used to weight the high-dimensional hidden state h. t A linear mapping to a scalar value, b y This is a bias term. Moisture suitability index Y t It reflects the water suitability of different grassland zones in the planting layer within the next τ days. The higher the index value, the more suitable the water conditions for grassland growth in that zone. The water suitability index calculated by the grassland dynamic growth model is input into the decision database to provide a basis for subsequent grazing regulation and irrigation and fertilization regulation.

[0067] Livestock behavior unit operation: It takes data transmitted from the environmental monitoring module as input, including the livestock's GPS / BeiDou positioning coordinates. t Triaxial acceleration data A t and body temperature data By analyzing and calculating these data, livestock behavioral characteristics are output, specifically including exercise volume calculation, feeding activity identification, and health status assessment. Specifically, in terms of exercise volume calculation, acceleration data A is used as the basis. t Calculate the sum of the magnitudes of the dynamic acceleration vectors within each time window, and use this as the motion quantity index M. t = In the formula, a x,i a y,i a z,i Let M be the raw value of the triaxial acceleration within one sampling period, and n be the number of sampling points within a time window. The motion index M is calculated using this formula. t This can intuitively reflect the activity intensity of livestock within a given time window; the greater the amount of movement, the more active the livestock. Regarding grazing activity identification, based on location data P... t To calculate the displacement distance D of livestock per unit time, the distance between two adjacent sampling times is calculated by continuously collecting the livestock's position coordinates. The distances between all adjacent sampling times within a unit time are then summed to obtain the displacement distance D. For health status assessment, body temperature data is used as a basis. To determine if livestock are experiencing abnormal body temperatures, the first step is to establish a normal body temperature range for the livestock. This range is determined based on normal body temperature data for different breeds, ages, and physiological stages of livestock. The collected livestock body temperature data is then... Compared with the normal body temperature range, if If the animal's body temperature exceeds the upper limit of the normal range or falls below the lower limit of the normal range, it is determined that the animal has an abnormal body temperature and may have a health problem. If the animal's body temperature is within the normal range, it is considered to have a normal body temperature and be in good health. The animal behavior unit will then calculate the activity level index M. t Livestock behavioral characteristics data, such as displacement distance D and abnormal body temperature, are transmitted to the decision-making database to provide data support for the database to formulate grazing strategies and control measures.

[0068] Decision database operation: The decision database receives the moisture suitability index Y from the grassland dynamic growth model. t The livestock behavior characteristic data output by the livestock behavior unit are fused and analyzed to generate comprehensive control strategies, specifically including water and fertilizer control in grazing areas based on the water suitability index and grazing strategy control based on livestock behavior:

[0069] In terms of water and fertilizer regulation in grazing areas based on the water suitability index, the decision base first maintains a list of predicted water suitability indices for all planting zones. This list is updated in real time with the water suitability index Y of each planting zone output by the grassland dynamic growth model. t Then, the moisture suitability index Y for each planting zone was calculated. t Compared with the preset moisture suitability index threshold Y th Comparison, when Y t <Y th When the decision base determines that the zone is under water and nutrient stress and has insufficient growth suitability, it generates irrigation and fertilization control instructions for the planting zone. The instructions clearly indicate the target planting zone number, the expected irrigation amount and fertilization amount, etc., to alleviate the water and nutrient stress of the zone and promote grassland growth. The generated irrigation and fertilization control instructions are then sent to the agricultural operation unit of the execution module.

[0070] In terms of grazing strategy regulation based on livestock behavior, the decision base receives daily displacement distance and body temperature data of livestock from the livestock behavior unit. First, it analyzes the daily displacement distance of livestock, calculates the average displacement distance of similar livestock in the same time period, and compares the daily displacement distance of each livestock with the average displacement distance. When the displacement distance of a certain livestock is found to be lower than the average of similar livestock, it indicates that the livestock may have decreased activity, poor health, or other abnormal conditions. The decision base, in conjunction with the execution module, drives the livestock guidance unit to guide the livestock to a backup zone. The backup zone usually has a more suitable growth environment and sufficient pasture resources, which facilitates the observation and care of the livestock, while avoiding its impact on the activities and grazing of other normal livestock.

[0071] When an individual with abnormal body temperature is detected in the livestock herd, the decision-making database immediately generates a special instruction. This instruction activates the audio-visual induction device in the execution module. The device uses specific frequency sounds and specific color lights to lure the individual with abnormal body temperature and isolate them in a designated observation area. Simultaneously, the decision-making database sends an alarm message to the system administrator's terminal device, such as a mobile phone or computer, reminding the administrator to promptly examine, diagnose, and treat the isolated livestock with abnormal body temperature to prevent the spread of disease and protect the health of the entire livestock herd.

[0072] The decision-making database will generate comprehensive control strategies, including irrigation and fertilization control instructions, livestock guidance instructions, and isolation alarm instructions, and distribute them to the execution module in the form of structured data to ensure that the execution module can accurately understand and execute the control instructions.

[0073] The execution module connects to the data analysis and decision-making module, receives comprehensive control strategies from the decision database, and performs physical execution. The execution module includes a livestock guidance unit and an agricultural operation unit, wherein:

[0074] The livestock guidance unit consists of a movable fence, electronic access control, and an audio-visual guidance device. When it receives a livestock guidance instruction from the decision-making database, it first determines the movement path of the fence and the opening and closing sequence of the electronic access control based on the target livestock and target backup zone specified in the instruction. The movable fence moves along a preset path, gradually guiding the target livestock towards the backup zone. During the fence's movement, the electronic access control system opens and closes in a timely manner based on the fence's progress and the target livestock's position, ensuring the target livestock can smoothly enter the backup zone while preventing other normal livestock from following. When a special instruction from the decision-making database to isolate livestock with abnormal body temperatures is received, the sound and light guidance device is immediately activated. The sound device emits a specific frequency sound, which is somewhat inducing to attract livestock with abnormal body temperatures to move in the designated direction. The light device emits light of a specific color as a visual guidance signal, assisting the sound guidance to guide the livestock with abnormal body temperatures to the designated observation area. During the sound and light guidance process, the movable fence and electronic access control system work together to adjust the physical boundaries, preventing livestock with abnormal body temperatures from escaping the guidance route and ensuring that they are successfully isolated to the observation area.

[0075] The agricultural operation unit consists of a zonal controlled irrigation solenoid valve assembly. When it receives irrigation and fertilization control instructions from the decision base, it first calls the water suitability index Y of the target irrigation zone output by the grassland dynamic growth model. t Based on the relevant meteorological data collected by the environmental monitoring module, the daily reference crop evapotranspiration ET0 is calculated using the following formula: Among them, R n The net surface radiation is collected and calculated by relevant sensors in the meteorological monitoring sensor array. It reflects the net radiation received by the surface after subtracting reflected and emitted radiation. G is the soil heat flux density, collected and calculated by relevant sensors in the soil moisture monitoring node, reflecting the heat transfer in the soil. T is the daily average air temperature, obtained by averaging the air temperature and humidity data collected at multiple times throughout the day by the air temperature and humidity sensor. u2 is the wind speed, collected by the anemometer and wind direction instrument. s The saturated vapor pressure (e) reflects the maximum amount of water vapor that the air can hold at that temperature. a The actual water vapor pressure, calculated from air humidity data, reflects the actual water vapor content in the air. Δ is the slope of the saturated water vapor pressure curve, and γ is the ecliptic constant. Then, the crop coefficient K is determined according to the type of crop within the planting zone. c Crop coefficient Kc This reflects the differences in water requirements of different crops at different growth stages, using the calculated reference crop evapotranspiration ET0 and crop coefficient K. c Calculate the actual water requirement ET of the crop. c =K c ×ET0, and simultaneously calculate the daily water deficit D. w =max(0,(ET) c -P eff )×A×10), where P eff The effective precipitation for the day is calculated by combining precipitation data collected by rain gauges with factors such as soil infiltration rate. It represents the actual precipitation absorbed and utilized by the soil after deducting surface runoff and deep seepage. A represents the area of ​​the irrigation zone, and ET represents the actual water requirement of the crop. c Greater than the effective precipitation P of the day eff At that time, there is a water deficit, and the water deficit is (ET). c -P eff )×A×10, when the actual water requirement of the crop is ET c Less than or equal to the effective precipitation P of the day e At time ff, there is no water deficit, and the water deficit is D. w If the value is 0, the decision base will calculate the water deficit D. w With the preset irrigation trigger threshold D th When comparing, when D w >D th When the water deficit in the planting zone reaches a level requiring irrigation, the decision base generates an irrigation instruction. This irrigation instruction is structured data and includes information such as the target irrigation zone number, irrigation start time, and irrigation duration. After receiving the irrigation instruction, the agricultural operation unit controls the corresponding irrigation solenoid valve group to open according to the target irrigation zone number specified in the instruction, and performs precise and quantitative irrigation according to the irrigation start time and irrigation duration to ensure that the crops receive sufficient water supply while avoiding water waste.

[0076] During irrigation, agricultural operation units also collect data such as irrigation flow and irrigation pressure in real time and feed this data back to the decision database. The decision database monitors and adjusts the irrigation process based on the feedback data. If abnormal irrigation flow or unstable irrigation pressure is detected, the opening degree of the irrigation solenoid valve group is adjusted in time or irrigation is suspended to ensure the stability of the irrigation process and the accuracy of the irrigation amount.

[0077] In summary, this embodiment details the construction and operation of a solar-agricultural complementary system in alpine grassland regions. From multi-dimensional data acquisition and preprocessing in the environmental monitoring module, to the prediction of the LSTM grassland dynamic growth model in the data analysis and decision-making module, the analysis of livestock behavior units, and the generation of comprehensive control strategies in the decision base, to the precise execution operation in the execution module, each link is closely connected and works collaboratively. It makes full use of the abundant local solar resources, achieves green electricity production, promotes the healthy recovery of grasslands and the sustainable development of animal husbandry, and provides an effective solution for ecological protection and economic development in alpine grassland regions.

[0078] Example 2

[0079] Based on Example 1, this example elaborates on the construction method of an integrated agricultural-pastoral-solar system in alpine grassland regions, such as... Figure 2 As shown, the specific steps of this method are as follows:

[0080] S100, Systematic Survey and Design Phase

[0081] Site survey and planning: Conduct a detailed survey of the target alpine grassland site, including topography, light resources, soil composition and original vegetation conditions.

[0082] System Design: Based on survey data, optimize the layout of the photovoltaic array (determine the support height, tilt angle and spacing), and simultaneously complete the zoning planning of the planting layer (delineate strong light area, moderate shade area, etc.).

[0083] S200, Infrastructure Layered Construction Phase

[0084] Infrastructure installation:

[0085] Install wind- and earthquake-resistant photovoltaic brackets and components to construct a photovoltaic layer.

[0086] Lay irrigation pipelines and deploy sensor networks (light, soil, and weather sensors).

[0087] Install smart fence base tracks or anchor points and construct sewage collection channels.

[0088] S300, Planting Layer Implementation and Deployment Phase

[0089] Planting implementation: According to the zoning plan, plant the appropriate plants in different light areas (e.g., plant pasture in strong light areas and shade-tolerant cash crops in shade areas).

[0090] Intelligent deployment: Complete the installation, power supply and network connection of all sensors and actuators (solenoid valves, fence robots, water pumps, etc.).

[0091] S400, Platform Integration and Optimization Phase

[0092] Platform integration: Integrate modules such as environmental monitoring, data analysis and decision-making, and execution into a unified communication protocol.

[0093] S500, System Integration and Operation Phase

[0094] Model optimization: Input local historical climate, soil and livestock behavior data to train and calibrate the grassland dynamic growth model and the rule parameters in the expert decision database.

[0095] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A three-dimensional system for complementary agriculture and animal husbandry with solar power in alpine grassland regions, characterized in that: The system is built in a high-altitude grassland region and adopts a three-dimensional spatial structure, which includes a photovoltaic layer for converting solar energy into electrical energy, a planting layer for agricultural production, and a breeding layer for livestock breeding. The system also includes an integrated intelligent management platform for coordinating the operation of the photovoltaic layer, planting layer, and aquaculture layer. The intelligent management platform includes an environmental monitoring module, a data analysis and decision-making module, and an execution module, wherein: The environmental monitoring module consists of various sensors deployed at different locations within the system space, used to collect physical environmental parameters and biological activity parameters in real time and continuously. The data analysis and decision-making module is connected to the environmental monitoring module and integrates a grassland dynamic growth model, livestock behavior units, and a decision database. The grassland dynamic growth model is built based on machine learning algorithms, using processed environmental data as input to predict and output the water suitability index for different zones within the planting layer in real time. The livestock behavior unit determines the amount of exercise and health status of livestock in the breeding layer based on livestock location and activity data. The decision base receives the output results from the pasture dynamic growth model and the livestock behavior unit, performs fusion analysis, and generates a comprehensive control strategy. The execution module is connected to the data analysis and decision-making module and is used to physically execute the decision base regulation in the data analysis and decision-making module; The photovoltaic layer, planting layer, and breeding layer form a three-dimensional agricultural-livestock-solar complementary system through the intelligent management platform.

2. The agro-pastoral-solar complementary three-dimensional system for alpine grassland regions according to claim 1, characterized in that, The sensor network in the environmental monitoring module adopts a heterogeneous networking approach, including: Light monitoring nodes are distributed under the photovoltaic layer, and each node integrates a photosynthetically active radiation sensor and a light intensity sensor; Soil moisture monitoring nodes are embedded in the planting layer soil. Each node integrates a soil temperature and humidity sensor, a soil pH sensor, and a soil electrical conductivity sensor. The meteorological monitoring sensor array installed in the system integrates air temperature and humidity sensors, wind speed and direction instruments, rain gauges and atmospheric pressure sensors. The livestock vital signs monitoring tags, which are attached to the livestock in the breeding layer, integrate GPS / BeiDou positioning, a three-axis accelerometer, and a body temperature sensor.

3. The agro-pastoral-solar complementary three-dimensional system for alpine grassland regions according to claim 1, characterized in that, The grassland dynamic growth model in the data analysis and decision-making module is a machine learning model based on multivariate time series prediction, which includes a Long Short-Term Memory (LSTM) network structure with at least one hidden layer. The structure consists of an input layer, at least one LSTM hidden layer, and a fully connected output layer. The input layer receives a multi-dimensional time-series environmental feature vector uploaded by the environmental monitoring module; The hidden layer contains three gating mechanisms: a forget gate, an input gate, and an output gate, which are used to learn and remember the long-term dependency between environmental factors and grassland growth. The fully connected output layer maps the hidden layer output state at the last time step to a single scalar value, namely the moisture suitability index for the next τ days. in, h represents the water suitability index predicted τ days from now at time step t. t W is the hidden state vector output at time step t after computation by the LSTM hidden layer. y It is a weight vector used to weight the high-dimensional hidden state h. t A linear mapping to a scalar value, b y It is a bias term that inputs the water suitability index calculated by the grassland dynamic growth model into the decision library for irrigation regulation.

4. The agro-pastoral-solar complementary three-dimensional system for alpine grassland regions according to claim 1, characterized in that, The livestock behavior unit in the data analysis and decision-making module takes data from the environmental monitoring module as input, including the livestock's GPS / BeiDou positioning coordinates P. t Triaxial acceleration data A t and body temperature data Calculate and output livestock behavioral characteristics: Motion volume calculation: based on acceleration data A t Calculate the sum of the magnitudes of the dynamic acceleration vectors within each time window, as an indicator of motion intensity. Among them, a x,i a y,i a z,i The original values ​​of the triaxial acceleration within one sampling period are given, and n is the number of sampling points within a time window. Foraging activity identification: based on location data P t Calculate the displacement distance D of the livestock per unit time; Health status assessment: based on body temperature data Determine if livestock have abnormal body temperatures.

5. The agro-pastoral-solar complementary three-dimensional system for alpine grassland regions according to claim 4, characterized in that, The decision base is linked to the predicted water suitability index output by the grassland dynamic growth model. Based on the behavioral characteristics output by the livestock behavior model, a comprehensive control strategy is implemented: Water and fertilizer regulation in grazing areas based on moisture suitability index: The decision base maintains a list of predicted moisture suitability indices for all zones, and assigns each zone... Compared with the preset moisture suitability index threshold Y th When comparing, When the zone is determined to be under water and nutrient stress, irrigation and fertilization control instructions are generated for that planting zone, and water and fertilizer operations are initiated. Grazing strategy regulation based on livestock behavior: The decision base receives the daily displacement distance and body temperature status of livestock from the livestock behavior unit. When the displacement distance of livestock is detected to be lower than the average of livestock of the same type, the decision base, in conjunction with the execution module, drives the livestock guidance unit to guide the livestock to the backup zone. When abnormal body temperature is detected in the livestock group, a special instruction is immediately generated to drive the sound and light induction device of the execution module to isolate the individual with abnormal body temperature to the designated observation area, and at the same time send an alarm message. The comprehensive control measures generated by the decision base are sent to the execution module for execution.

6. The agro-pastoral-solar complementary three-dimensional system for alpine grassland regions according to claim 5, characterized in that, The execution module includes: The livestock guiding unit consists of a movable fence, an electronic access control system, and an audio-visual guidance device. After receiving the control strategy from the decision database, it reconstructs the physical boundary of the breeding layer by moving the fence and opening and closing the electronic access control system, and at the same time uses the audio-visual guidance device to guide and drive the livestock herd. The agricultural operation unit consists of a zone-controlled irrigation solenoid valve group. After receiving the irrigation and fertilization control commands, it controls the solenoid valve group to perform precise and quantitative irrigation on the planting zone.

7. The agro-pastoral-solar complementary three-dimensional system for alpine grassland regions according to claim 6, characterized in that, When the agricultural operation unit performs quantitative irrigation, it follows the following logic: The daily reference crop evapotranspiration is calculated by calling the moisture suitability index output by the grassland dynamic growth model. Among them, R n Let G be the net surface radiation, T be the soil heat flux density, T be the daily average temperature, u2 be the wind speed, and e be the average temperature. s e is the saturated vapor pressure. a Δ is the actual water vapor pressure, Δ is the slope of the saturated water vapor pressure curve, and γ is the wet / dry gauge constant. Determine the crop coefficient K based on the crop type. c And calculate the actual water requirement ET of the crop. c =K c ×ET0 and daily water deficit D w =max(0,(ET) c -P eff )×A×10), where P eff The effective precipitation for the day is represented by A, and the area of ​​the irrigated zone is represented by A. The decision base will include the water deficit D w With the preset irrigation trigger threshold D th Compare; when D w >D th If so, an irrigation instruction is generated. The irrigation instruction is structured data and includes: target irrigation zone number, irrigation start time, and irrigation duration.

8. A method for constructing a three-dimensional agro-pastoral-photovoltaic complementary system in alpine grassland regions, applicable to the three-dimensional agro-pastoral-photovoltaic complementary system in alpine grassland regions as described in any one of claims 1-7, characterized in that, The method includes the following steps performed in sequence: S100, Systematic Survey and Design Phase: Conduct a detailed survey of the target alpine grassland site, optimize the layout of the photovoltaic array based on the survey data, and simultaneously complete the zoning planning of the planting layer, the architecture design of the intelligent management platform, and the wiring planning of the infrastructure. S200, Infrastructure Layered Construction Phase: Install photovoltaic brackets and modules according to the design, construct the photovoltaic layer, lay irrigation pipe network, and deploy sensor network; S300, Planting Layer Implementation and Deployment Phase: According to the zoning plan, plant plants with different characteristics, and complete the installation, power supply and network connection of all sensors and actuators; S400, Platform Integration and Optimization Phase: Establish a unified communication protocol, input local historical climate data, soil data and livestock behavior data, and train and calibrate the grassland dynamic growth model and rule parameters in the decision base. S500, System Integration and Operation Phase: Livestock are introduced, the entire system is started, the intelligent management platform begins operation, and the system achieves intelligent and adaptive operation.

Images