Intelligent water and fertilizer integrated precise irrigation control method

Abstract

The invention discloses an intelligent water and fertilizer integrated precise irrigation control method. The method comprises the following steps: step 1, establishing a multi-modal data acquisition network; step 2, inputting the data obtained in the step 1 into a trained LSTM neural network, and predicting future crop transpiration water demand; 3, according to nutrient concentration distribution fed back by a soil conductivity sensor in real time, combining with a water and fertilizer coupling migration model, dynamically calculating a nitrogen-phosphorus-potassium ratio of each irrigation zone, realizing precise mixing of a fertilizer mother solution and irrigation water through a three-channel proportioning pump, and implementing a variable irrigation decision; 4, dynamically adjusting the pulse irrigation duration of each branch through an MPC algorithm; and step 5, recovering the residual liquid at the tail end of the pipe network to a buffer tank, detecting the EC value of the recovered liquid through an electrochemical sensor, and taking the EC value as a reference parameter automatic correction proportioning algorithm of the next irrigation period. Accurate control over the irrigation process can be achieved, so that the water and fertilizer utilization efficiency is improved, and resource waste and environmental pollution are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent water and fertilizer precise irrigation control, and particularly to an intelligent integrated water and fertilizer precise irrigation control method. Background Art

[0002] At present, with the increasing global water shortage and the challenges faced by agricultural sustainable development, traditional irrigation and fertilization methods are difficult to meet the requirements of modern agriculture for high efficiency, environmental protection and sustainability. Traditional irrigation mostly adopts flood irrigation, which not only wastes a large amount of water resources, but also easily causes problems such as soil salinization and nutrient loss; fertilization depends on empirical judgment, resulting in over-application and unbalanced nutrients, which not only increases the agricultural production cost, but also pollutes the environment.

[0003] Therefore, the integrated water and fertilizer technology has emerged. This technology organically combines the irrigation and fertilization processes, and through an intelligent control system, realizes the synchronous supply and precise management of water and nutrients according to the growth needs of crops and the soil environment conditions. However, there is still room for improvement in the intelligence and precision of existing integrated water and fertilizer systems. For example, although some systems can carry out irrigation and fertilization according to a preset plan, it is difficult to accurately monitor soil environment parameters and crop growth conditions in real time and adjust the plan in a timely manner. Summary of the Invention

[0004] In view of this, the present invention proposes an intelligent integrated water and fertilizer precise irrigation control method to solve the problem of insufficient intelligence and precision in integrated water and fertilizer irrigation in the prior art.

[0005] The specific technical solution of the present invention is as follows:

[0006] An intelligent integrated water and fertilizer precise irrigation control method, comprising:

[0007] Step 1, constructing a multi-modal data acquisition network covering physiology, soil, meteorology and canopy temperature through a plant stem micro-deformation sensor, a root layer soil profile water sensor array, a weather station and a crop canopy infrared camera;

[0008] Step 2, inputting the data obtained in Step 1 into a pre-trained LSTM neural network model, which integrates the crop variety genotype feature library, the historical growth stage water response curve, and the historical plant physiological parameters and environmental data, and predicts the crop transpiration water demand in the next 6 hours by analyzing the correlation between the genetic characteristics of the crop variety and the transpiration rate; the LSTM model optimizes the prediction accuracy through genotype feature matching and dynamically adjusts the prediction value based on the real-time input data;

[0009] Step 3: Based on the water demand prediction result of step 2 and the nitrogen, phosphorus and potassium nutrient concentration distribution fed back by the soil conductivity sensor in real time, the nitrogen, phosphorus and potassium ratio of each irrigation zone is calculated using the water-fertilizer coupling migration model; the water-fertilizer coupling migration model trains historical soil nutrient migration data through a machine learning algorithm and outputs a nutrient ratio function that meets the needs of crops; the flow of fertilizer mother solution and irrigation water is dynamically adjusted through a three-channel proportional pump to achieve precise mixing;

[0010] Step 4: Using a distributed pressure compensation dripper array, based on the real-time data of the pipeline flow meter and pressure sensor, the MPC algorithm is used to dynamically adjust the pulse irrigation duration of each branch, so that the deviation between the actual soil moisture content and the predicted value is controlled within ±3%; the dripper array realizes pressure adaptation through the elastic diaphragm and vortex flow channel design, and combines the fractal topology network architecture to compensate for local pressure anomalies;

[0011] Step 5: After the irrigation cycle ends, the negative pressure siphon device is started to recover the residual liquid at the end of the pipe network to the buffer tank, and the EC value of the recovered liquid is detected by the electrochemical sensor. Based on the mapping relationship between the EC value and the nitrogen, phosphorus and potassium concentrations, the ratio algorithm for the next irrigation cycle is corrected.

[0012] Specifically, in step 1, the tiny deformation of crop stems is monitored in real time through plant stem micro-deformation sensors, and the crop transpiration intensity is obtained based on the nonlinear relationship between stem deformation and transpiration rate; the three-dimensional moisture distribution data of the soil layer is obtained in combination with the root layer soil profile moisture sensor array, and the light intensity, evaporation and crop canopy infrared thermal imaging data of the meteorological station are collected simultaneously.

[0013] Specifically, in step 1, the plant stem micro-deformation sensor is directly installed on the crop stem and is made of strain gauges or piezoelectric materials. The stem micro-deformation is converted into a digital signal and transmitted to a data centralized management device; the root layer soil profile moisture sensor array is arranged in a layered and zoned manner in the 0-60cm soil layer, and the soil moisture is measured using a capacitive or time domain reflection principle, and the three-dimensional moisture distribution data of the soil layer is obtained through data fusion and three-dimensional modeling technology.

[0014] Specifically, in step 1, the stem shape and transpiration rate Satisfies the nonlinear relationship:

[0015]

[0016] in, is the stem shape variable, indicating the degree of slight deformation of the crop stem due to transpiration; k is the crop variety specific coefficient; r is the stem radius; represents the transpiration rate; represents the time variable; It is a non-linear function of transpiration rate T and time t, which is determined by fitting according to actual experimental data.

[0017] Specifically, in step 2, the transpiration water requirement of the crop in the next 6 hours The prediction formula is expressed as:

[0018]

[0019] Where, Represents the input vector, including the multi-modal data collected in step 1, the stem deformation , soil water content W, light intensity I, evaporation E, crop canopy temperature ; Represents the parameter set of the LSTM neural network, including the weight matrix and bias vector, which are obtained by training historical data, including the crop variety genotype feature library, historical growth stage water response curve, historical plant physiological parameters and environmental data; Represents the output generated by the LSTM neural network after a series of calculations given the input And model parameters , and this output is used to predict the transpiration water requirement of the crop in the next 6 hours.

[0020] Specifically, in step 3, the three-channel proportional pump is installed on the irrigation pipeline, connected to the fertilizer mother liquor storage container and the irrigation water pipeline, and connected to the control system through the control line, and controls the flow rates of the fertilizer mother liquor and irrigation water according to the calculated nitrogen, phosphorus and potassium ratios.

[0021] Specifically, in step 3, let , , Be the nitrogen, phosphorus and potassium nutrient contents required for the i-th irrigation zone respectively, with the unit of gram, then

[0022] ,

[0023] ,

[0024] ,

[0025] Where, Is the transpiration water requirement of the crop in the next 6 hours predicted in step 2, with the unit of liter; , , Respectively represent the nitrogen, phosphorus and potassium nutrient concentrations fed back by the soil conductivity sensors in the i-th irrigation zone, with the unit of gram per liter; , , They are functions of crop water requirement and soil nutrient concentration respectively.

[0026] Specifically, in step 3, let 、 、 be the flow rates of nitrogen, phosphorus, and potassium in the fertilizer mother liquor respectively, with the unit of liter per hour; be the flow rate of irrigation water, with the unit of liter per hour, then:

[0027] ,

[0028] ,

[0029] ,

[0030] Among them, 、 、 are the concentrations of nitrogen, phosphorus, and potassium in the fertilizer mother liquor respectively, with the unit of gram per liter.

[0031] Specifically, in step 4, the distributed pressure-compensated dripper array is installed on the branch of the irrigation pipeline, adopting a fractal topology network architecture, integrating pressure adaptive technology and vortex pressure stabilization mechanism to ensure uniform and stable water output of the drippers under different terrain and pressure conditions. The MPC algorithm automatically optimizes the irrigation strategy according to the deviation between the actual soil water content and the predicted value, and controls the deviation within ±3%.

[0032] Specifically, in step 5, let the required nitrogen, phosphorus, and potassium nutrient contents in the corrected ith irrigation zone be 、 、 respectively, and the EC value of the recycled liquid be ,then:

[0033] ,

[0034] ,

[0035] ,

[0036] Among them, 、 、 respectively represent the required nitrogen, phosphorus, and potassium nutrient contents in the ith irrigation zone calculated in step 3; 、 、 are the correction coefficients related to the EC value of the recycled liquid, obtained by measuring the nitrogen, phosphorus, and potassium contents in the recycled liquid at different EC values through experiments; represents the conductivity of the recycled liquid.

[0037] The beneficial effects of the present invention are as follows:

[0038] (1) Construct a multi-modal data acquisition network, integrating devices such as plant stem micro-deformation sensors, root layer soil profile moisture sensor arrays, weather stations, and infrared cameras, to comprehensively and accurately collect crop physiological and environmental data, providing a reliable basis for subsequent decision-making;

[0039] Based on the real-time collected multi-modal data, use the trained long short-term memory network (LSTM) neural network, combined with crop historical data, to dynamically predict the crop transpiration water requirement in the next 6 hours and can be adjusted in a timely manner with new data;

[0040] According to the water requirement prediction and soil nutrient concentration distribution, use the water-fertilizer coupling migration model, and through a three-channel proportional pump, achieve precise irrigation and fertilization, improve resource utilization efficiency, promote crop growth and reduce pollution;

[0041] (4) Adopt a distributed pressure-compensated drip head array, combined with the data of pipeline flow meters and pressure sensors, and dynamically adjust the pulse irrigation duration of each branch through the MPC algorithm, controlling the deviation between the actual soil moisture content and the predicted value within ±3%, to adapt to complex agricultural environments;

[0042] After the irrigation cycle ends, use a negative pressure siphon device to recover the residual liquid at the end of the pipe network, automatically correct the mixing ratio algorithm for the next cycle by detecting the EC value, realize the recycling of resources, reduce waste and environmental pressure. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0044] Figure 1 is a schematic flow chart of the intelligent water and fertilizer integrated precise irrigation control method of the present invention Figure 1 ;

[0045] Figure 2 is a schematic flow chart of the intelligent water and fertilizer integrated precise irrigation control method of the present invention Figure 2 . DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the following further describes the present invention in detail with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0047] The present invention proposes an intelligent water-fertilizer integrated precision irrigation control method. Figure 1-2 shown.

[0048] The method of the present invention comprises: step 1, constructing a multimodal data acquisition network covering physiology, soil, meteorology and canopy temperature through plant stem micro-deformation sensors, root layer soil profile moisture sensor arrays, meteorological stations and crop canopy infrared cameras; step 2, inputting the data obtained in step 1 into a pre-trained LSTM neural network model, the model integrates the crop variety genotype feature library, the historical growth stage moisture response curve, and the historical plant physiological parameters and environmental data, and predicts the crop transpiration water requirement in the next 6 hours by analyzing the correlation between the crop variety genetic characteristics and the transpiration rate; the LSTM model optimizes the prediction accuracy through genotype feature matching, and dynamically adjusts the prediction value based on real-time input data; step 3, based on the water demand prediction result of step 2, combined with the nitrogen, phosphorus and potassium nutrient concentration distribution fed back in real time by the soil conductivity sensor, the water-fertilizer coupling transport model is used to calculate each irrigation partition The water-fertilizer coupling migration model uses machine learning algorithms to train historical soil nutrient migration data and outputs a nutrient ratio function that meets crop needs. The flow of fertilizer mother solution and irrigation water is dynamically adjusted by a three-channel proportional pump to achieve precise mixing. Step 4: A distributed pressure compensation dripper array is used. Based on the real-time data of the pipeline flow meter and pressure sensor, the pulse irrigation duration of each branch is dynamically adjusted through the MPC algorithm to control the deviation between the actual soil moisture content and the predicted value within ±3%. The dripper array realizes pressure adaptation through the design of elastic diaphragm and vortex flow channel, and combines the fractal topology network architecture to compensate for local pressure anomalies. Step 5: After the irrigation cycle ends, the negative pressure siphon device is started to recover the residual liquid at the end of the pipe network to the buffer tank, and the EC value of the recovered liquid is detected by the electrochemical sensor. Based on the mapping relationship between the EC value and the nitrogen, phosphorus and potassium concentration, the ratio algorithm for the next irrigation cycle is corrected.

[0049] For step 1, the micro-deformation of crop stems is monitored in real time through plant stem micro-deformation sensors, and the crop transpiration intensity is obtained based on the nonlinear relationship between stem deformation and transpiration rate. The three-dimensional humidity distribution of the 0-60 cm soil layer is obtained in combination with the root layer soil profile moisture sensor array. The light intensity, evaporation and crop canopy infrared thermal imaging data of the meteorological station are simultaneously collected to establish a multimodal data acquisition network.

[0050] In order to comprehensively and accurately obtain relevant information on the crop growth environment and physiological status, the present invention constructs a multi-modal data acquisition network by collecting multi-modal data covering crop physiological characteristics, soil conditions, meteorological information, and crop canopy characteristics. To collect the above-mentioned multi-modal data, the present invention uses a plant stem micro-deformation sensor, a root layer soil profile moisture sensor array, and a weather station, and integrates the above-mentioned various sensors and devices to form a network system that works in cooperation. This system can realize the synchronous acquisition of different types of data, ensuring the real-time and accuracy of the data.

[0051] The transpiration of plants is an important process of their water metabolism and is closely related to the water demand of crops. The plant stem micro-deformation sensor can monitor the minute deformation of the crop stem in real time, and this deformation is directly related to the intensity of crop transpiration. By monitoring the micro-deformation of the stem, the transpiration of the crop can be indirectly reflected, providing an important basis for subsequent water demand prediction. The data of the plant stem micro-deformation sensor comes from the physiological activities of the crop itself. When the crop undergoes transpiration, the stem will undergo minute deformations, which are captured by the sensor and converted into digital signals.

[0052] The plant stem micro-deformation sensor is directly installed on the crop stem and is in close contact with the stem by means of fitting or winding, etc., to accurately monitor the minute deformation of the stem. It is connected to the data transmission line and transmits the collected digital signals to the data centralized management device.

[0053] The plant stem micro-deformation sensor can be a strain gauge, a sensitive element that can convert mechanical strain into an electrical signal. The strain gauge is pasted on the surface of the plant stem. When the stem undergoes minute deformation due to transpiration, the strain gauge will also deform accordingly, and its resistance value will change accordingly. By measuring the change in the resistance value and using circuits such as a Wheatstone bridge to convert the resistance change into a voltage or current change, after signal processing such as amplification and filtering, it is finally converted into a digital signal.

[0054] The plant stem micro-deformation sensor can also be made of piezoelectric materials. Piezoelectric materials generate electric charges when subjected to mechanical stress. A sensor is made using piezoelectric materials and is in contact with the plant stem. When the stem undergoes minute deformation, the piezoelectric material is stressed to generate electric charges. By measuring the change in the electric charges to reflect the deformation of the stem, and then through circuits such as a charge amplifier to convert the charge signal into a voltage signal, it is finally converted into a digital signal.

[0055] Transpiration causes changes in the water potential gradient in the xylem vessels of plants, leading to the contraction or expansion of the vascular tissue, and the amount of stem deformation and the transpiration rate satisfy a non-linear relationship:

[0056]

[0057] Among them, is the stem deformation amount, representing the degree of slight deformation of the crop stem due to transpiration; k represents the crop variety specific coefficient; r represents the stem radius; represents the transpiration rate; represents the time variable; is a non-linear function of the transpiration rate T and time t, which can be determined by fitting according to actual experimental data.

[0058] The crop roots are mainly distributed in the soil layer of 0 - 60 cm. The moisture condition of this soil layer is crucial for crop growth. The root layer soil profile moisture sensor array adopted in the present invention measures soil moisture by using the capacitance or time domain reflectometry (TDR) principle. The sensors are arranged in the 0 - 60 cm soil layer in a layered and zoned manner, and the sensors in each layer are evenly distributed in the horizontal direction to obtain soil moisture information at different depths and positions. Through data fusion and three-dimensional modeling technology, the data collected by each sensor is converted into a three-dimensional humidity distribution of the 0 - 60 cm soil layer. The data collected by the sensors is transmitted to the centralized data management device by wired or wireless means. After preprocessing and fusion analysis, it provides detailed data support for irrigation decision-making.

[0059] The capacitive moisture sensor measures soil moisture by using the relationship between the soil dielectric constant and soil water content. The dielectric constant of the soil changes with the change of water content. When the electrodes of the sensor are placed in the soil, the soil acts as the dielectric of the capacitor, and the change of its dielectric constant will cause the change of the capacitance value. The sensor can obtain the soil water content by measuring the capacitance value and through calibration and conversion. For example, when the soil water content increases, the dielectric constant of the soil increases and the capacitance value also increases accordingly.

[0060] The TDR sensor emits high-frequency electromagnetic pulses into the soil. The pulses propagate in the metal rod. When encountering the soil, due to the different dielectric properties of the soil, part of the pulses will be reflected back. The sensor measures the time difference between the emission and reflection of the pulses, and calculates the dielectric constant of the soil according to the relationship between the propagation speed of electromagnetic waves in the medium and the dielectric constant, and then obtains the soil water content.

[0061] Taking the capacitive moisture sensor as an example, the relationship between the soil water content W and the capacitance value C is expressed as:

[0062]

[0063] Among them, W is the soil moisture content, representing the percentage of the mass of water in the soil to the mass of dry soil; C is the capacitance value, which is the capacitance value measured by the capacitive moisture sensor; a, b, and c are calibration coefficients obtained through calibration experiments on known soil moisture content and corresponding capacitance values.

[0064] Meteorological conditions such as light intensity and evaporation have a significant impact on the transpiration and water demand of crops. The meteorological station can collect meteorological data such as light intensity and evaporation in real time, and combine the infrared thermal imaging data of the crop canopy to comprehensively analyze the environmental conditions of the crops. The infrared thermal imaging data of the crop canopy can reflect the temperature distribution of the crop canopy, and then infer the water stress status and growth state of the crops. Parameters such as light intensity and evaporation are collected through the sensors of the meteorological station, while the infrared thermal imaging data of the crop canopy is obtained through an infrared camera.

[0065] The meteorological station is set in an open and unobstructed position in the crop planting area to ensure accurate collection of meteorological data such as light intensity and evaporation. The meteorological station is equipped with various sensors, which are connected to the data processing unit of the meteorological station through internal lines, and the data processing unit then transmits the data to the data centralized management device through the network. The infrared camera is installed at a position where it can overlook the crop canopy, such as on a high bracket, and transmits the collected infrared thermal imaging data of the crop canopy to the data centralized management device.

[0066] Step 2: Based on the plant physiological parameters and environmental data collected in real time in Step 1, construct a dynamic water demand prediction model. First, input the multi-modal data obtained in Step 1 into a trained long short-term memory (LSTM) neural network. This network has been trained with a large amount of data, and its model training data includes the crop variety genotype feature library, the historical growth stage water response curve, and historical plant physiological parameters and environmental data. The LSTM neural network predicts the transpiration water demand of the crops in the next 6 hours based on the current plant physiological parameters and environmental data, combined with the relationship between crop water demand and various factors learned during the training process.

[0067] In practical applications, the growth state and environmental conditions of crops are constantly changing. Therefore, it is necessary to input the collected plant physiological parameters into the trained LSTM neural network in real time. As new data is continuously input, the model can timely adjust the predicted value of the transpiration water demand of the crops in the next 6 hours to adapt to the dynamic changes during the crop growth process. For example, if sudden rainfall occurs within 2 hours after prediction and the soil humidity increases, input the new soil humidity data into the model, and the model will correspondingly adjust the water demand prediction value for the subsequent 4 hours.

[0068] In this step, the long short-term memory (LSTM) neural network is selected as the prediction model. LSTM has a powerful ability to process sequential data and can effectively handle the long-term dependencies in time series data. In crop water requirement prediction, plant physiological parameters and environmental data are sequential data that change over time. LSTM can capture the long-term trends and periodic changes in the data through its internal memory units and gating mechanisms, and can learn the complex relationships between physiological parameters and water requirements during the crop growth process, making it very suitable for predicting crop water requirements.

[0069] The genotype feature library of crop varieties records the genetic characteristics of different crop varieties, and these characteristics will affect the growth and development of crops and their water requirement patterns. For example, different wheat varieties have differences in root development, leaf morphology, etc., and these differences will lead to different abilities of them to absorb and utilize water. By collecting genotype data of a large number of different crop varieties and establishing a genotype feature library, the present invention can provide variety-specific information for the model.

[0070] The present invention can collect seed samples of different crop varieties from sources such as agricultural research institutions and germplasm resource banks, obtain genotype data by combining technologies such as whole-genome sequencing and reduced-representation genome sequencing, and integrate information such as genetic characteristics and water requirement patterns recorded in historical literature. At the same time, relying on the public platforms of the National Long-Term Genebank for Crop Germplasm and the DNA Fingerprint Database, a genotype feature library of crop varieties is constructed to realize the standardized management of the "molecular identity cards" of varieties and ensure the traceability and sharing of data.

[0071] Bioinformatics methods are used to clean and extract features from the original gene data, such as screening gene markers related to stress resistance and water requirement characteristics (such as SNP loci, gene expression levels, etc.). The interaction relationship between genotype and environmental factors is analyzed by LSTM to predict the water requirement response of different varieties under specific environments. In addition, combined with the precise phenotypic identification technology, a genotype-phenotype association database is established to clarify the influence of genetic background on the water requirements at different growth stages of crops. Link the genotype feature library with multi-source databases such as meteorology and soil to provide variety-specific parameters for the dynamic water requirement prediction model. For example, optimize the prediction accuracy of the LSTM model for different varieties by matching the genotype features with the historical growth stage water response curves.

[0072] In irrigation decision-making, the genotype feature library can assist in identifying the genetic drought resistance of crops. For example, the gene characteristics of root development of a certain corn variety show that its deep water absorption ability is strong, and combined with soil humidity data, the shallow irrigation frequency can be reduced; while for a certain wheat variety, due to the differences in leaf stomatal regulation genes, the predicted value of transpiration water requirement needs to be dynamically adjusted according to the canopy temperature data.

[0073] The historical growth stage water response curve reflects the changing water requirements of crops at different growth stages. In the initial stage of crop growth, the root system is not fully developed and the water requirement is relatively small. As the growth process progresses and enters the flowering and fruiting stage, the water requirement increases significantly. In the later stage of growth, the water requirement gradually decreases. The present invention utilizes multi-modal data of soil moisture, meteorology, and crop growth collected by an agricultural Internet of Things system for a long time, identifies the growth stage division nodes through a time series clustering algorithm, and combines with the machine learning model LSTM to mine the dynamic relationship between water requirements and environmental factors. For example, by analyzing the correlation between canopy infrared thermal imaging data and transpiration intensity, the water stress threshold can be inverted and the response curve can be generated.

[0074] The above crop variety genotype feature library and the historical growth stage water response curve, combined with the collected historical plant physiological parameters and environmental data, are used as training data to be input into the LSTM neural network for training. By continuously adjusting the model parameters, the model learns the complex relationship between the crop water requirement and various factors.

[0075] The transpiration water requirement of the crop in the next 6 hours The prediction formula can be expressed as:

[0076]

[0077] Where, represents the input vector, , which includes the multi-modal data collected in step 1, such as the stem deformation , soil moisture content W, light intensity I, evaporation E, crop canopy temperature , etc.; represents the parameter set of the LSTM neural network, including the weight matrix and bias vector, etc., which are obtained by training a large amount of historical data, including the crop variety genotype feature library, the historical growth stage water response curve, the historical plant physiological parameters and environmental data; represents the output generated by the LSTM neural network after a series of calculations under the given input and model parameters , and this output is used to predict the transpiration water requirement of the crop in the next 6 hours.

[0078] Step 3: Based on the predicted results of crop transpiration water requirement obtained in Step 2 and combined with the nutrient concentration distribution real-time feedback by the soil conductivity sensor, implement variable irrigation decision-making. Using the water-fertilizer coupling transport model, dynamically calculate the nitrogen, phosphorus, and potassium ratios in each irrigation zone according to the crop water requirement and soil nutrient distribution. Through a three-channel proportional pump, accurately mix the fertilizer mother liquor and irrigation water according to the calculated ratio to ensure that the water and fertilizer for irrigation can meet the actual needs of crops in different regions. Through accurate mixing, achieve fertilization according to demand and improve the utilization efficiency of water and fertilizer. The specific steps are as follows:

[0079] Step 31: Install soil conductivity sensors in different regions of the soil to make them in full contact with the soil, and real-time monitor the concentration distribution of nutrients in the soil. The sensors are connected to a data collector through a data cable, and the data collector transmits the data to the control system. The control system analyzes the data feedback by the sensors to understand the contents of nutrients such as nitrogen, phosphorus, and potassium in the soil in different regions. The soil conductivity is closely related to the nutrient concentration in the soil. Through the soil conductivity sensor, the concentration distribution of nutrients in the soil can be real-time monitored, and the contents of nutrients such as nitrogen, phosphorus, and potassium in the soil in different regions can be understood.

[0080] The soil conductivity sensor mainly works based on the conductive characteristics of the soil. The soil is a complex system composed of solid particles, water, air, and various ions dissolved in water. When an electric current passes through the soil, the ions in the soil will move directionally under the action of the electric field, thus generating a conductive phenomenon. The soil conductivity (EC) reflects the concentration and types of ions in the soil, and the nutrients in the soil (such as nitrogen, phosphorus, potassium, etc.) usually exist in the soil solution in the form of ions, so the soil conductivity is closely related to the nutrient concentration in the soil. The measurement method of the soil conductivity sensor of the present invention can adopt electrode measurement or electromagnetic induction measurement.

[0081] Step 32: Calculate the nitrogen, phosphorus, and potassium ratios. Using the water-fertilizer coupling transport model, comprehensively consider the migration laws and interactions of water and nutrients in the soil, and calculate the ratios of nutrients such as nitrogen, phosphorus, and potassium required in each irrigation zone according to the nutrient concentration distribution feedback by the soil conductivity sensor. The water-fertilizer coupling transport model can simulate the movement process of water and fertilizer in the soil, predict the nutrient distribution in the soil under different irrigation and fertilization methods, and thus provide a scientific basis for precise fertilization. The input data of the water-fertilizer coupling transport model comes from the nutrient concentration distribution feedback by the soil conductivity sensor and the predicted crop water requirement in Step 2.

[0082] The water-fertilizer coupling transport model simulates the dynamic migration process of water and nutrients in the soil-crop system through a mathematical model. By integrating multi-source data such as historical soil moisture, nutrient concentration, meteorological data, and crop growth data, a large-scale dataset is formed. Machine learning algorithms, such as neural networks, decision trees, and support vector machines, are used to train the dataset. With the crop water requirement and soil nutrient distribution as the input and the nitrogen, phosphorus, and potassium ratios in each irrigation zone as the output, the water-fertilizer coupling transport model is trained.

[0083] Step 33: Precisely mix the fertilizer mother liquor and irrigation water. According to the calculated nitrogen, phosphorus, and potassium ratios, set the parameters of the three-channel proportioning pump to precisely mix the fertilizer mother liquor and irrigation water. The three-channel proportioning pump can accurately control the flow rates of the fertilizer mother liquor and irrigation water, ensuring that the nutrient content in the mixed water-fertilizer solution meets the requirements of each irrigation zone, thereby achieving precise fertilization for crops in different regions, meeting the nutrient requirements of crops at different growth stages, promoting the healthy growth of crops, and improving crop yield and quality.

[0084] The three-channel proportioning pump is installed on the irrigation pipeline and connected to the fertilizer mother liquor storage container and the irrigation water pipeline. It is connected to the control system through a control line and accurately controls the flow rates of the fertilizer mother liquor and irrigation water according to the nitrogen, phosphorus, and potassium ratios calculated by the control system to achieve precise mixing.

[0085] Through variable irrigation decision-making, the irrigation and fertilization strategies can be precisely adjusted according to the actual nutrient status of the soil and the water requirement of the crops, avoiding the waste of fertilizers and the pollution of the environment caused by over-application. At the same time, it also improves the absorption efficiency of water and fertilizers by crops and promotes crop growth.

[0086] The formulas involved in the process of calculating the nitrogen, phosphorus, and potassium ratios using the water-fertilizer coupling operation model are as follows: Let 、 、 be the nitrogen, phosphorus, and potassium nutrient contents required for the i-th irrigation zone, respectively, in grams. Then

[0087] ,

[0088] ,

[0089] ,

[0090] where is the transpiration water requirement of the crop predicted in the next 6 hours in step 2, in liters; 、 、 represent the nitrogen, phosphorus, and potassium nutrient concentrations feedback by the soil conductivity sensors in the i-th irrigation zone, respectively, in grams per liter; 、 、 They are functions of crop water requirement and soil nutrient concentration respectively. As an example:

[0091] ,

[0092]

[0093] ,

[0094] Among them, , , and , , are coefficients obtained by training historical data through machine learning algorithms such as neural networks, decision trees, support vector machines, etc.

[0095] The formulas involved in the flow control process of the three-channel proportional pump are:

[0096] Let , , be the flow rates of nitrogen, phosphorus, and potassium in the fertilizer mother liquor respectively, with the unit of liters per hour; is the flow rate of irrigation water, with the unit of liters per hour, then:

[0097] ,

[0098] ,

[0099] ,

[0100] Among them, , , are the concentrations of nitrogen, phosphorus, and potassium in the fertilizer mother liquor respectively, with the unit of grams per liter.

[0101] Step 4, perform closed-loop feedback control, adopt a distributed pressure-compensated dripper array, and based on the real-time data of the pipeline flowmeter and pressure sensor, dynamically adjust the pulse irrigation duration of each branch through the MPC (Model Predictive Control) algorithm, which can monitor and adjust the irrigation process in real time, and control the deviation between the actual soil water content and the predicted value in Step 2 within ±3%. Through closed-loop feedback control, the irrigation process can be monitored and adjusted in real time to ensure the accuracy of irrigation and avoid the adverse effects on crop growth caused by insufficient or excessive irrigation.

[0102] The distributed pressure-compensating dripper array can ensure uniform and stable water output from the drippers and ensure irrigation uniformity under different terrain and pressure conditions. Its structural design combines pressure adaptive technology with vortex pressure stabilization mechanism. Specifically, the array achieves pressure adaptation through dynamic pressure regulation of elastic diaphragms and vortex pressure stabilization technology of three-dimensional flow channels. Each dripper has a built-in elastic silicone diaphragm, and its deformation characteristics enable the water outlet aperture to be dynamically adjusted with the pipeline pressure: under high pressure, the diaphragm expands and contracts the cross-sectional area of ​​the flow channel to suppress the surge in flow, and under low pressure, the diaphragm rebounds and expands the channel to compensate for the flow attenuation. Combined with the vortex flow channel designed with a spiral guide groove, the water flow forms a forced vortex motion inside the dripper, using centrifugal force to offset the kinetic energy changes caused by pressure fluctuations, and working with the elastic diaphragm to form a dual pressure stabilization mechanism, ensuring that each dripper can stably discharge water at a preset flow rate under complex terrain and different pressure environments, thereby avoiding the problem of uneven irrigation caused by terrain undulations or pressure changes.

[0103] The distributed pressure compensation dripper array is installed on the branch of the irrigation pipeline and evenly distributed in the crop planting area. The dripper is tightly connected to the irrigation pipeline through a joint to ensure that the water can flow stably from the pipeline to the dripper. The dripper array adopts a fractal topology network architecture, dividing the irrigation area into four sub-networks (main pipe-branch pipe-capillary tube-dripper unit), and deploys pressure sensor nodes and electromagnetic pulse valve groups with a response time of less than 50ms at intervals of 10 meters. This architecture supports local pressure anomalies through linkage compensation of adjacent nodes to avoid system-level oscillation. The MPC (model predictive control) algorithm is dynamically coupled with the dripper array, and the rolling optimizer is driven by real-time soil moisture deviation data to generate a pressure target value and trigger a multi-mode compensation decision: the branch booster pump is started when the terrain elevation difference is greater than 2%, the pulse valve opening time is extended when the instantaneous flow is insufficient, and the redundant flow channel is automatically switched when local blockage is detected to ensure the stable operation of the irrigation system.

[0104] Pipeline flow meters and pressure sensors are installed on irrigation pipes to collect flow and pressure data in the irrigation pipes in real time, reflecting the operating status of the irrigation system. Pipeline flow meters can accurately measure the amount of water passing through the pipe per unit time, while pressure sensors monitor the pressure changes in the pipe in real time. They are connected to the control system through data cables, and transmit flow and pressure data to the control system in real time for MPC algorithm analysis.

[0105] The MPC algorithm can optimize the control strategy based on the current state and future predictions of the irrigation system. This algorithm not only considers the current system state but also makes optimal control decisions in advance by predicting the system state over a period of time in the future, thus achieving more precise control. The data collected by the pipeline flowmeter and pressure sensor are input into the MPC algorithm. By continuously analyzing the input data, the MPC algorithm can accurately judge the deviation between the current soil moisture content and the predicted value, and dynamically adjust the pulse irrigation duration of each branch according to the preset control target, control the irrigation water volume, enabling the irrigation system to have an adaptive ability to adapt to different soil conditions and crop growth requirements. By continuously adjusting the irrigation duration, the actual soil moisture content gradually approaches the predicted value, realizing precise irrigation.

[0106] Whether it is the change in soil texture or the difference in crop growth stages, the MPC algorithm can quickly respond according to real-time data, adjust the irrigation strategy, and ensure that the crops are always in the best growth environment. For example, for soils with poor water retention, the MPC algorithm will appropriately increase the irrigation frequency and duration to meet the water requirements of the crops. This adaptive ability makes the irrigation system more flexible and efficient, and can adapt to various complex agricultural environments.

[0107] Step 5, implement water and fertilizer recycling compensation. After the irrigation cycle ends, start the negative pressure siphon device to recycle the residual liquid at the end of the pipe network to the buffer tank. Detect the EC value (electrical conductivity) of the recycled liquid through an electrochemical sensor, and automatically correct the mixing ratio algorithm as the reference parameter for the next irrigation cycle. The specific steps are as follows:

[0108] Step 51, start the negative pressure siphon device to recycle the residual liquid at the end of the pipe network. In the irrigation system, the residual liquid at the end of the pipe network contains a certain amount of fertilizer and water. If directly discharged, it will not only cause waste of fertilizer and water resources but also may cause environmental pollution. For example, nitrogen, phosphorus and other elements in the fertilizer entering the water body may cause eutrophication of the water body. The negative pressure siphon device is based on the siphon principle, that is, using the pressure difference generated by the liquid at different heights to make the liquid flow from the end with high pressure to the end with low pressure. After the irrigation cycle ends, start the negative pressure siphon device to suck the residual liquid at the end of the pipe network into the buffer tank, realizing the recycling and reuse of resources, reducing production costs and environmental pressure.

[0109] The negative pressure siphon device is installed at the end of the pipe network and is connected to the pipe network through a pipeline. When the irrigation cycle ends, start the negative pressure siphon device to suck the residual liquid at the end of the pipe network into the buffer tank through the connecting pipeline.

[0110] Step 52, the EC value can reflect the nutrient concentration in the recycled liquid. By detecting the EC value of the recycled liquid, the nutrient content in the residual liquid can be understood, providing a basis for correcting the mixing ratio algorithm for the next cycle.

[0111] The EC value (electrical conductivity) is closely related to the ion concentration in the solution. In the water-fertilizer solution, various ions are generated after the fertilizer dissolves, such as nitrogen ions, phosphorus ions, potassium ions, etc. The presence of these ions makes the solution have a certain electrical conductivity. The higher the ion concentration, the greater the electrical conductivity of the solution. Therefore, by detecting the EC value of the recycled liquid with an electrochemical sensor, the nutrient concentration in the recycled liquid can be indirectly reflected. Electrochemical sensors usually work based on the principle of ion-selective electrodes, have a selective response to specific ions, and can convert the ion concentration into an electrical signal for measurement, so as to accurately obtain the nutrient content information in the recycled liquid.

[0112] The electrochemical sensor is installed in the buffer tank and is in full contact with the recycled liquid in the buffer tank. Based on the principle of ion-selective electrodes, it detects the EC value of the recycled liquid. The electrochemical sensor is connected to the control system through a data cable, and transmits the detected EC value of the recycled liquid to the control system for correcting the mixing ratio algorithm of the next cycle.

[0113] Step 53: According to the EC value of the recycled liquid, correct the nitrogen, phosphorus, and potassium mixing ratio algorithm of the next cycle. By adjusting the fertilizer input amount, the irrigation and fertilization in the next cycle can be made more accurate, and the utilization efficiency of water and fertilizer can be improved.

[0114] In step 3, the water-fertilizer coupling migration model is used to calculate the nitrogen, phosphorus, and potassium mixing ratios of each irrigation zone according to the crop water demand and soil nutrient distribution. However, since the residual liquid at the end of the pipe network in the previous irrigation cycle contains certain nutrients, if these residual nutrients are not considered, the fertilization in the next cycle may lead to excessive or insufficient nutrients. According to the EC value of the recycled liquid, the nutrient content in the residual liquid can be determined and used as a reference parameter for the next irrigation cycle. When calculating the nitrogen, phosphorus, and potassium mixing ratio algorithm, the nutrient content in the residual liquid is taken into account, and the fertilizer input amount is adjusted accordingly. For example, if the nitrogen element content in the recycled liquid is high, the input amount of nitrogen fertilizer can be appropriately reduced in the mixing ratio algorithm of the next cycle, so as to make the irrigation and fertilization more accurate, improve the utilization efficiency of water and fertilizer, and avoid the waste of fertilizer and adverse effects on the environment.

[0115] The formulas involved in the water-fertilizer recycling compensation process are as follows: Let the required nitrogen, phosphorus, and potassium nutrient contents in the corrected i-th irrigation zone be 、 、 , the EC value of the recycled liquid is , then:

[0116] ,

[0117] ,

[0118] ,

[0119] Among them, , , respectively represent the nitrogen, phosphorus, and potassium nutrient contents required for the i-th irrigation zone calculated in step 3; , , are the correction coefficients related to the EC value of the recycled liquid, obtained by experimentally measuring the nitrogen, phosphorus, and potassium contents in the recycled liquid at different EC values; represents the electrical conductivity of the recycled liquid.

[0120] In summary, in the data collection stage, the multi-modal data collection network monitors the transpiration intensity of crops in real time through the plant stem micro-deformation sensor, and its data directly reflects the physiological activity state of crops; at the same time, the root layer soil profile moisture sensor array obtains the three-dimensional humidity distribution of the 0-60 cm soil layer, accurately depicting the water supply situation of crop roots. These two types of data provide the core basis for water demand prediction from the physiological and soil levels respectively. Environmental parameters such as light intensity and evaporation collected by the weather station, together with the crop canopy infrared thermal imaging data, constitute an environmental dynamic monitoring system.

[0121] In the water demand prediction stage, the real-time collected multi-modal data is input into the pre-trained LSTM neural network. This model captures long-term dependence relationships by integrating the historical water response curves in the growth stage, the genotype feature library, and the historical plant physiological parameters and environmental data. For example, when it detects a decrease in soil humidity and a sudden increase in light intensity, the model can combine the transpiration volume change trend in similar situations in the historical data to dynamically predict the water demand in the next 6 hours.

[0122] In the variable irrigation decision-making stage, the predicted crop water demand and the nutrient concentration distribution feedback by the soil conductivity sensor jointly drive the water-fertilizer coupling migration model. This model simulates the migration law of water and fertilizer in the soil and calculates the customized nitrogen, phosphorus, and potassium ratios for each irrigation zone. For example, in areas where nitrogen is scarce but potassium content is high, the model will dynamically adjust the ratio based on the current growth stage requirements of the crops, and then accurately mix the fertilizer mother liquor and irrigation water through a three-channel proportional pump to achieve "one zone, one policy" precise supply. This process not only avoids the homogenized fertilization defect of traditional irrigation but also improves the resource utilization rate by responding to soil nutrient changes in real time.

[0123] In the closed-loop control stage, the distributed pressure-compensated drip head array combines the pipeline flow rate and pressure sensor data to dynamically optimize the pulse irrigation duration through the MPC algorithm. For example, when the actual soil water content is lower than the predicted value, the algorithm will adjust the branch irrigation duration based on the terrain water retention and real-time pressure data, controlling the deviation within ±3%.

[0124] During the recovery and compensation stage, the negative pressure siphon device recovers the residual liquid in the pipeline network. The electrochemical sensor detects its EC value and corrects the mixing ratio algorithm for the next cycle, forming a complete closed-loop of "monitoring - execution - recovery - correction", effectively solving the lag and resource waste problems of traditional irrigation systems.

[0125] The beneficial effects of the present invention are as follows:

[0126] (1) By constructing a multi-modal data acquisition network, the present invention can monitor the intensity of crop transpiration in real time, obtain the three-dimensional humidity distribution of the soil layer from 0 to 60 cm, and simultaneously collect the light intensity, evaporation amount of the weather station and the infrared thermal imaging data of the crop canopy. These data directly reflect the physiological activity state of the crops, the root water supply situation and the environmental conditions, providing the core basis for subsequent precise irrigation.

[0127] (2) The present invention uses a pre-trained LSTM neural network for dynamic water demand prediction. This model integrates the historical water response curve of the growth stage, the crop variety genotype feature library and the historical plant physiological parameters and environmental data, and can capture the long-term dependence relationship in the data to accurately predict the transpiration water demand of the crops in the next 6 hours. This prediction result provides a scientific basis for variable irrigation decision-making.

[0128] (3) In the variable irrigation decision-making stage, the present invention uses a water and fertilizer coupling migration model, combines the predicted crop water demand and the nutrient concentration distribution feedback by the soil conductivity sensor, and calculates the customized nitrogen, phosphorus and potassium ratios for each irrigation zone. The fertilizer mother liquor and irrigation water are precisely mixed by a three-channel proportional pump, realizing precise supply, avoiding the homogenized fertilization defect of traditional irrigation, and improving the water and fertilizer utilization efficiency.

[0129] (4) The present invention also adopts a closed-loop feedback control mechanism. The distributed pressure compensation drip head array combines the pipeline flow rate and pressure sensor data, and dynamically optimizes the pulse irrigation duration through the MPC algorithm to ensure that the deviation between the actual soil water content and the predicted value is controlled within ±3%. This control mechanism improves the precision of irrigation and avoids the adverse effects on crop growth caused by insufficient or excessive irrigation.

[0130] (5) The present invention also implements a water and fertilizer recovery and compensation strategy. The negative pressure siphon device recovers the residual liquid in the pipeline network. The electrochemical sensor detects its EC value and corrects the mixing ratio algorithm for the next cycle, forming a complete closed-loop of "monitoring - execution - recovery - correction". This strategy effectively solves the lag and resource waste problems of traditional irrigation systems and further improves the overall efficiency of the irrigation system.

[0131] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An intelligent water-fertilizer integrated precision irrigation control method, characterized in that: include: Step 1: Build a multimodal data collection network covering physiology, soil, meteorology and canopy temperature through plant stem micro-deformation sensors, root layer soil profile moisture sensor arrays, weather stations and crop canopy infrared cameras; Step 2: Input the data obtained in step 1 into a pre-trained LSTM neural network model. The model integrates the crop variety genotype feature library, the historical growth stage water response curve, and the historical plant physiological parameters and environmental data. By analyzing the correlation between the genetic characteristics of the crop variety and the transpiration rate, the crop transpiration water requirement in the next 6 hours is predicted. The LSTM model optimizes the prediction accuracy through genotype feature matching and dynamically adjusts the prediction value based on real-time input data. Step 3: Based on the water demand prediction result of step 2 and the nitrogen, phosphorus and potassium nutrient concentration distribution fed back by the soil conductivity sensor in real time, the nitrogen, phosphorus and potassium ratio of each irrigation zone is calculated using the water-fertilizer coupling migration model; the water-fertilizer coupling migration model trains historical soil nutrient migration data through a machine learning algorithm and outputs a nutrient ratio function that meets the needs of crops; the flow of fertilizer mother solution and irrigation water is dynamically adjusted through a three-channel proportional pump to achieve precise mixing; Step 4: Using a distributed pressure compensation dripper array, based on the real-time data of the pipeline flow meter and pressure sensor, the MPC algorithm is used to dynamically adjust the pulse irrigation duration of each branch, so that the deviation between the actual soil moisture content and the predicted value is controlled within ±3%; the dripper array realizes pressure adaptation through the elastic diaphragm and vortex flow channel design, and combines the fractal topology network architecture to compensate for local pressure anomalies; Step 5: After the irrigation cycle ends, the negative pressure siphon device is started to recover the residual liquid at the end of the pipe network to the buffer tank, and the EC value of the recovered liquid is detected by the electrochemical sensor. Based on the mapping relationship between the EC value and the nitrogen, phosphorus and potassium concentrations, the ratio algorithm for the next irrigation cycle is corrected.

2. The intelligent water-fertilizer integrated precision irrigation control method according to claim 1, characterized in that: In step 1, the micro deformation of crop stems is monitored in real time by plant stem micro deformation sensors, and the intensity of crop transpiration is obtained based on the nonlinear relationship between stem deformation and transpiration rate; the three-dimensional moisture distribution data of the soil layer is obtained in combination with the root layer soil profile moisture sensor array, and the light intensity, evaporation and crop canopy infrared thermal imaging data of the meteorological station are simultaneously collected.

3. The intelligent water-fertilizer integrated precision irrigation control method according to claim 2, characterized in that: In step 1, the plant stem micro-deformation sensor is directly installed on the crop stem and is made of strain gauges or piezoelectric materials. The stem micro-deformation is converted into a digital signal and transmitted to a data centralized management device; the root layer soil profile moisture sensor array is arranged in a layered and partitioned manner in the 0-60cm soil layer, and the soil moisture is measured using a capacitive or time domain reflection principle, and the three-dimensional moisture distribution data of the soil layer is obtained through data fusion and three-dimensional modeling technology.

4. The intelligent water-fertilizer integrated precision irrigation control method according to claim 2, characterized in that: In step 1, the stem shape and transpiration rate Satisfies the nonlinear relationship: in, is the stem shape variable, indicating the degree of slight deformation of the crop stem due to transpiration; k is the crop variety specific coefficient; r is the stem radius; represents the transpiration rate; represents the time variable; It is a nonlinear function of transpiration rate T and time t, which is determined by fitting based on actual experimental data.

5. The intelligent water-fertilizer integrated precision irrigation control method according to claim 2, characterized in that: In step 2, the transpiration water requirement of the crop in the next 6 hours The prediction formula is expressed as: in, Represents the input vector, including the multimodal data collected in step 1, the stem shape variable , soil moisture content W, light intensity I, evaporation E, crop canopy temperature ; It represents the parameter set of the LSTM neural network, including the weight matrix and the bias vector, which is obtained by training the historical data, including the crop variety genotype feature library, the historical growth stage water response curve, the historical plant physiological parameters and environmental data; Indicates that given input and model parameters In the case of , the output generated by the LSTM neural network after a series of calculations is used to predict the transpiration water requirement of crops in the next 6 hours.

6. The intelligent water-fertilizer integrated precision irrigation control method according to claim 1, characterized in that: In step 3, the three-channel proportional pump is installed on the irrigation pipeline, connected to the fertilizer mother solution storage container and the irrigation water pipeline, connected to the control system through the control line, and controls the flow of the fertilizer mother solution and the irrigation water according to the calculated nitrogen, phosphorus and potassium ratio.

7. The intelligent water-fertilizer integrated precision irrigation control method according to claim 1, characterized in that: In step 3, it is assumed that , , are the nitrogen, phosphorus and potassium nutrient contents required for the i-th irrigation zone, in grams. ， ， ， in, The transpiration water demand of the crop in the next 6 hours predicted in step 2, in liters; , , They represent the nitrogen, phosphorus and potassium nutrient concentrations fed back by the soil conductivity sensor in the i-th irrigation zone, in g / L; , , are functions of crop water requirement and soil nutrient concentration, respectively.

8. The intelligent water-fertilizer integrated precision irrigation control method according to claim 7, characterized in that: In step 3, it is assumed that , , are the flow rates of nitrogen, phosphorus and potassium in the fertilizer mother solution, respectively, in liters per hour; is the flow rate of irrigation water in liters per hour, then: ， ， ， in, , , They are the concentrations of nitrogen, phosphorus and potassium in the fertilizer mother solution, respectively, in grams per liter.

9. The intelligent water-fertilizer integrated precision irrigation control method according to claim 1, characterized in that: In step 4, the distributed pressure-compensating dripper array is installed on the branch of the irrigation pipe, adopts a fractal topology network architecture, and integrates pressure adaptation technology and vortex pressure stabilization mechanism to ensure that the water output of the drippers is uniform and stable under different terrain and pressure conditions. The MPC algorithm automatically optimizes the irrigation strategy according to the deviation between the actual soil moisture content and the predicted value, and controls the deviation within ±3%.

10. The intelligent water-fertilizer integrated precision irrigation control method according to claim 7, characterized in that: In step 5, the nitrogen, phosphorus and potassium nutrient contents required for the i-th irrigation zone after correction are respectively , , The EC value of the recovered solution is ,but: ， ， ， in, , , They represent the nitrogen, phosphorus and potassium nutrient contents required for the i-th irrigation zone calculated in step 3 respectively; , , are the correction coefficients related to the EC value of the recovery solution, respectively, obtained by experimentally measuring the contents of nitrogen, phosphorus and potassium in the recovery solution at different EC values; Indicates the conductivity of the recovery liquid.

Images