An AI-based precision mixed variable irrigation and fertilization system and method

Abstract

The application discloses an AI-based precise mixed variable irrigation and fertilization system and method. The system comprises a water and fertilizer mixing part, a plot water and fertilizer monitoring and feedback device and a system self-adaptive coupling optimization control part, which are electrically connected; the plot terminal collects effective nitrogen, EC, pH, temperature, flow and pressure and uploads them. The method comprises the following steps: calling an AI-based water and fertilizer information analysis model library to analyze the concentrations of nitrogen fertilizer, phosphorus fertilizer and potassium fertilizer; taking the concentration and flow error as input, using Q-learning reinforcement learning to update the PID parameter set, and using fuzzy PID cooperative control to control the main pipeline pressure pump, the mixed pipeline electromagnetic flow regulating valve and the nitrogen fertilizer, phosphorus fertilizer and potassium fertilizer flow regulating pump; and determining the drip irrigation pipe network leakage or burst according to the pipe network pressure mutation. The system and method can realize online monitoring of the plot irrigation and fertilization process, multi-component fertilizer solution analysis and self-adaptive coupling optimization control.

Description

Technical Field

[0001] This invention relates to the field of agricultural irrigation equipment and technology, and in particular to an AI-based precision mixed variable irrigation and fertilization system and method. Background Technology

[0002] The integrated water and fertilizer technology plays an increasingly prominent role and has great application prospects in the development of modern, efficient, water-saving agriculture, large-scale facility agriculture, precision agriculture, and smart agriculture. Precision mixed variable fertilization irrigation, as the key to the application of integrated water and fertilizer technology, is an important technical means to couple the growth and development of crops throughout the entire growth period and to supply water and fertilizer in a timely, quantitative, high-quality and efficient manner at different growth stages of crops.

[0003] Currently, domestic and international research and development of integrated water and fertilizer variable fertilization irrigation systems mainly consist of pipes, sprinklers, filters, fertilizer applicators, and simple monitoring equipment. These systems prioritize hardware over software, focusing on initial system control while neglecting feedback during the irrigation and fertilization process. They also lack expertise in dynamic regulation and optimization of the integrated system hardware, failing to achieve real-time analysis of complex mixed fertilizer information at the end of the pipeline network and real-time online feedback of the irrigation and fertilization process and its effects. This results in lower irrigation and fertilization accuracy and lower irrigation and fertilization quality in residential areas.

[0004] Therefore, there is an urgent need to develop a simple, low-cost precision mixed variable irrigation and fertilization system and method. This system and method can realize real-time monitoring and feedback of the entire process of irrigation and fertilization operations and the final irrigation and fertilization effect, and can perform autonomous and continuous optimization and control based on current and historical operating data during operation. Summary of the Invention

[0005] The purpose of this invention is to provide an AI-based precision mixed variable irrigation and fertilization system and method to solve the problems of real-time monitoring and feedback of the entire process of irrigation and fertilization operations and the final irrigation and fertilization effect in production practice, as well as adaptive coupling optimization and control during system operation. The system and method have the advantages of adaptive coupling optimization and control, precision and efficiency, and ease of operation.

[0006] To achieve the above objectives, this invention provides an AI-based precision mixed variable irrigation and fertilization system, including a precision mixed variable irrigation and fertilization system, an AI-based water and fertilizer information analysis model library, and an adaptive coupling optimization and control algorithm for the precision mixed variable irrigation and fertilization system based on a combination of fuzzy PID and reinforcement learning.

[0007] Furthermore, the precision mixed variable irrigation and fertilization system of the present invention mainly consists of (1) a water and fertilizer mixing part, (2) a plot water and fertilizer monitoring and feedback device and (3) a system dynamic coupling optimization and control part.

[0008] The (1) water-fertilizer mixing section includes a water storage tank. The inner wall of the water storage tank is equipped with a water level sensor and a water outlet, which is connected to the main pipeline of the system. A main pipeline filter is connected to the main pipeline. A main pipeline solenoid valve and a main pipeline booster pump are sequentially installed after the main pipeline filter. The main pipeline booster pump is connected to a main pipeline check valve and a main pipeline pressure gauge. A smart water-fertilizer machine is connected after the main pipeline pressure gauge. The smart water-fertilizer machine and the mixing pipeline electromagnetic flow regulating valve are connected in parallel. The parallel smart water-fertilizer machine and the mixing pipeline electromagnetic flow regulating valve are connected to the mixing pipeline. An irrigation branch pipe is connected after the mixing pipeline.

[0009] The intelligent fertigation machine includes a fertilizer solution tank and a wireless transceiver. A fertilizer solution solenoid valve and a fertilizer solution flow regulating pump are installed on the outlet pipe of the fertilizer solution tank. The fertilizer solution solenoid valve, fertilizer solution flow regulating pump, and mixing pipeline flow regulating valve are all electrically connected to the intelligent fertigation machine's drive relay. The intelligent fertigation machine's drive relay is electrically connected to the intelligent fertigation machine's central controller. The intelligent fertigation machine's central controller is an embedded system or microcontroller with strong computing power, capable of deploying and running corresponding models and algorithms, such as Raspberry Pi, STM32 series, Intel Movidius series, ARM Cortex-A and Cortex-M series microcontrollers, Qualcomm Snapdragon Neural Processing Engine (NPE), Xilinx Zynq series, and NVIDIA Jetson series. The fertilizer solution flowing from the fertilizer solution tank passes through the fertilizer solution solenoid valve, fertilizer solution filter, fertilizer solution check valve, fertilizer solution flow regulating pump, and Venturi tube before mixing with the irrigation water in the main pipeline and flowing into the mixing pipeline.

[0010] The (2) community water and fertilizer monitoring feedback device includes a community water and fertilizer monitoring feedback device solenoid valve. The community water and fertilizer monitoring feedback device solenoid valve is electrically connected to the community water and fertilizer monitoring feedback device drive relay. The community water and fertilizer monitoring feedback device solenoid valve is provided with an irrigation liquid EC sensor, an irrigation liquid pH sensor, an irrigation liquid temperature sensor, an irrigation liquid effective nitrogen monitor, an electromagnetic flow meter, and an irrigation liquid pressure sensor in sequence on the outside of the community water and fertilizer monitoring feedback device solenoid valve. The community water and fertilizer monitoring feedback device drive relay, irrigation liquid EC sensor, irrigation liquid pH sensor, irrigation liquid temperature sensor, irrigation liquid effective nitrogen monitor, electromagnetic flow meter, irrigation liquid pressure sensor, and community water and fertilizer monitoring feedback device wireless signal transceiver are connected to the community water and fertilizer monitoring feedback device solar power supply system through a cable.

[0011] The community water and fertilizer monitoring feedback device includes a drive relay, an irrigation liquid EC sensor, an irrigation liquid pH sensor, an irrigation liquid temperature sensor, an irrigation liquid effective nitrogen monitor, an electromagnetic flow meter, and an irrigation liquid pressure sensor. These components are connected to the community water and fertilizer monitoring feedback device's wireless transceiver via cables. The wireless transceiver is fixed to the support frame of the community water and fertilizer monitoring feedback device's solar power supply system column.

[0012] The solar power supply system of the community water and fertilizer monitoring feedback device includes solar photovoltaic panels, which are mounted on photovoltaic panel brackets. The photovoltaic panel brackets are mounted on columns, and the columns are rigidly connected to the base. The solar photovoltaic panels are electrically connected to a power management module, which is also electrically connected to a rechargeable battery. The power management module and the rechargeable battery are encapsulated together within the protective casing of the community water and fertilizer monitoring feedback device.

[0013] The wireless transceiver of the community water and fertilizer monitoring feedback device and the wireless transceiver of the intelligent water and fertilizer machine adopt a data wireless transmission module based on NB-IoT technology to transmit and receive water and fertilizer information of the irrigation community, system status information of the irrigation community water and fertilizer monitoring feedback device, and control the opening and closing of the solenoid valve of the irrigation community in real time.

[0014] The (3) system dynamic coupling optimization and control part of the coupled water and fertilizer mixing hardware and the community water and fertilizer monitoring feedback device includes an intelligent water and fertilizer machine central controller. The intelligent water and fertilizer machine central controller is electrically connected to the water level sensor of the water storage tank, the main pipeline solenoid valve, the fertilizer liquid solenoid valve, the fertilizer liquid flow regulating pump, the mixing pipeline electromagnetic flow regulating valve, and the drive relays that control the main pipeline solenoid valve, the fertilizer liquid solenoid valve, the fertilizer liquid flow regulating pump, and the mixing pipeline electromagnetic flow regulating valve.

[0015] The intelligent water and fertilizer machine's central controller embeds an AI-based water and fertilizer information analysis model library and an adaptive coupling optimization control algorithm for a precise mixed variable irrigation and fertilization system based on a combination of fuzzy PID and reinforcement learning. Secondly, as the brain of the entire system's operation and scheduling, the intelligent water and fertilizer machine's central controller receives data from the water and fertilizer mixing section and the community water and fertilizer monitoring feedback device in real time. Based on the AI ​​water and fertilizer information analysis model in the AI ​​water and fertilizer information analysis model library, it analyzes and displays the water and fertilizer concentration in the operating community in real time, thereby calculating the irrigation and fertilization rate and progress. On the other hand, based on the system's preset water and fertilizer targets, current operating status data, and historical control data, and using the adaptive coupling optimization control algorithm for the precise mixed variable irrigation and fertilization system, the intelligent water and fertilizer machine's central controller makes system-level dynamic coupling optimization control decisions and sends commands to the main pipeline solenoid valve, main pipeline booster pump, fertilizer solution solenoid valve, fertilizer solution flow regulating pump, mixing pipeline electromagnetic flow regulating valve, and community water and fertilizer monitoring feedback device solenoid valve.

[0016] The central controller of the intelligent water and fertilizer machine is also equipped with a human-machine interface, a data storage device, and an embedded 4G communication module, which enables it to transmit data, train algorithms in the cloud, and receive instructions with the cloud system and mobile app.

[0017] This invention also provides an AI-based water and fertilizer information analysis model library. The model library is formed by combining commonly used soluble nitrogen fertilizers (urea (CO(NH2)2), ammonium nitrate (NH4NO3), ammonium sulfate ((NH4)2SO4) etc.), phosphate fertilizers (monoammonium phosphate (NH4H2PO4), diammonium phosphate ((NH4)2HPO4), superphosphate (Ca(H2PO4)2) etc.), and potassium fertilizers (potassium chloride (KCl), potassium sulfate (K2SO4), potassium nitrate (KNO3) etc.) to form various types of fertigation mixed fertilizers. Machine learning or deep learning models are constructed for each type of mixed fertilizer at different component concentrations, including the effective nitrogen content, EC, pH, temperature, and other physical parameters of the fertilizer solution, as well as the concentrations of different fertilizer components in the fertilizer solution. The collection of these models constitutes the AI-based water and fertilizer information analysis model library.

[0018] Specifically, taking the construction of a fertilizer and nutrient information analysis model for a mixed fertilizer of potassium dihydrogen phosphate and potassium chloride as an example, based on the needs of actual production, mixed fertilizer solutions of potassium dihydrogen phosphate and potassium chloride with a total concentration of 2-10 g / L at different ratios as shown in Table 1 were configured. Under strict temperature control conditions of 15-40 ℃, the readings of the effective nitrogen sensor, EC sensor, pH sensor, and temperature sensor of the fertilizer solution were simultaneously read, forming a dataset of effective nitrogen-EC-pH-temperature-potassium dihydrogen phosphate concentration-potassium chloride concentration. After cleaning the dataset, the data was analyzed using the machine learning library sklearn.StandardScaler The dataset was standardized and then randomly divided into two parts in a 7:3 ratio, one for training and one for testing. The model selection function from the machine learning library sklearn.model_selection was used. GridSearchCV and k-fold cross- validation The selected machine learning algorithm (e.g., Support Vector Machine, SVM) is trained and its hyperparameters optimized. After achieving satisfactory performance, the data standardization process and the trained model are saved to the model library. See the results below. Figure 6 , 7 Similarly, other mixed fertilizer and water information analysis models follow a similar process, using machine learning or deep learning to build corresponding models and save them to a model library.

[0019] Table 1. Formula for Mixed Fertilizer Solution of Potassium Dihydrogen Phosphate and Potassium Chloride

[0020] Note: 2, 4, 6, 8, 10 represent the concentration of the mixed fertilizer solution (g / L); C DP and C PC The values ​​represent the concentrations (g / L) of potassium dihydrogen phosphate and potassium chloride in the mixed fertilizer solution, respectively; the ratio value represents the ratio of different concentrations of potassium dihydrogen phosphate and potassium chloride at the same concentration in the mixed fertilizer.

[0021] The model library is embedded in the central controller of the intelligent water and fertilizer machine. After the central controller receives the effective nitrogen content, EC, pH and temperature feedback data of the fertilizer solution from the water and fertilizer monitoring feedback part of the irrigation and fertilization operation area, it performs effectiveness verification and inputs it into the corresponding AI water and fertilizer information analysis model. It analyzes the concentration of mixed fertilizer solution components in real time and displays information such as pipeline pressure, pipeline flow and operation progress of the irrigation and fertilization operation area in real time on the human-machine interface, and uploads it to the cloud system or mobile terminal.

[0022] This invention also provides an adaptive coupling optimization and control algorithm for a precise hybrid variable irrigation and fertilization system based on a combination of fuzzy PID and reinforcement learning. Specifically, a target nitrogen fertilizer concentration C is preset in the system before each operation. N Phosphate fertilizer concentration C P Potassium fertilizer concentration C K System traffic Q w The Q-learning algorithm provides an initialization strategy based on prior knowledge. Specifically, it provides a set of PID parameters {K} based on historical operating data and current preset data to regulate the speed of the main pipeline booster pump, the duty cycle of the mixing pipeline flow regulating valve, the speed of the nitrogen fertilizer flow regulating pump, the speed of the phosphate fertilizer flow regulating pump, and the speed of the potash fertilizer flow regulating pump. p w , Ki w , K d w ; K p V , K i V , K d V ; K p N , K i N , K d N ; K p P , K i P , K d P ; K p K , K i K , K d K}0. The real-time nitrogen fertilizer concentration C, analyzed by the AI ​​water and fertilizer information analysis model during system operation. t N Phosphate fertilizer concentration C t P Potassium fertilizer concentration C t K and system traffic Q t w The nitrogen fertilizer concentration C preset by the system for this operation N Phosphate fertilizer concentration C P Potassium fertilizer concentration C K and system traffic Q w Error variable E between t N =C t N -C N E t P =C t P -C P E t K =C t K -C K E t w =Q t w -Q w and the rate of change of error EC t N =dE t N / dt,EC t P =d E t P / dt, EC t K =d E t K / dt, EC t w =d E t w / dt is used as the input variable of the algorithm. The error variable E is divided into 7 fuzzy sets: negative large (NB), negative medium (NM), negative small (NS), zero (ZO), positive small (PS), positive medium (PM), and positive large (PB), with a value range of [-5, 5], and the data is fuzzified using a triangular membership function. Similarly, the error change rate EC is divided into 7 fuzzy sets: negative large (NB), negative medium (NM), negative small (NS), zero (ZO), positive small (PS), positive medium (PM), and positive large (PB), with a value range of [-30%, 30%], and the data is fuzzified using a triangular membership function. The magnitude of the output, i.e., the fuzzy inference value, is obtained by querying the inference tables Tables 2, 3, and 4 based on the membership degrees of E and EC. The fuzzy inference values ​​are defuzzified using the median method to obtain the PID parameters used to regulate the speed of the main pipeline booster pump, the duty cycle of the mixing pipeline flow regulating valve, the speed of the nitrogen fertilizer flow regulating pump, the speed of the phosphate fertilizer flow regulating pump, and the speed of the potash fertilizer flow regulating pump. The parameters are then input into the PID controller to obtain the control signal at time t.

[0023] Table 2. K p Rule table

[0024] Table 3. K i Rule table

[0025] Table 4. K d Rule table

[0026] Meanwhile, to address the challenge of parameter tuning in traditional PID control, especially in complex, nonlinear, or time-varying systems, an adaptive coupled optimization control algorithm for a precise hybrid variable irrigation and fertilization system is proposed. This algorithm utilizes an interactive learning process and reinforcement learning to adjust the strategy based on the interaction with the environment, further optimizing the system performance of the fuzzy PID multi-controller coupled control. Specifically, based on the system's dynamic performance indicators, considering overshoot, energy consumption, response time, and stability, the reward function is designed as follows: (1) in, α , β , γ and δ These are all weighting coefficients used to balance the importance of error, error rate of change, and control input. These coefficients can be adjusted according to specific application scenarios. E ( t ) 2 and EC ( t ) 2 These represent the error and the square of the rate of change of error, respectively, used to penalize large errors and rapidly changing errors. u ( t ) 2 This represents the square of the control input, used to penalize excessive control actions, which helps reduce system oscillations and energy consumption. RT It is the time required for the system to start responding and reach a certain preset steady-state error range (such as within ±2%).

[0027] Q-learning algorithm was chosen as the core algorithm for handling continuous action space and applying it to PID controller parameter adjustment; the PID parameters (K... p K i K d The action space is part of the action space, and the size and range of actions are determined based on the actual system state characteristics. During system operation, continuously collected system data, current running data, and interactive learning between the algorithm and the current environment are used. Based on the current state, actions (PID parameters) are selected, actions are executed, new states and rewards are observed, and the estimated value of the Q-function is updated using the Q-learning algorithm following the Bellman equation, in the following form: (2) in α The learning rate determines the extent to which new information influences old estimates; r It is an immediate reward, which is the reward that an agent receives from the environment after performing an action; γ It is a discount factor, used to balance the importance of immediate rewards and future rewards; s’ It is a new state, that is, the action is being performed. a The state of the post-environment; a’ In the new state s’ Any action that may be taken; max a’ Q ( s’ , a’ ) indicates the new state s’ The maximum of all possible actionsQ The goal of Q-learning is to find the optimal Q-function. Q * ( s , a This makes it possible for all state-action pairs Q * ( s , a ), which is equivalent to being in state s Take action below a And follow the optimal strategy to obtain the maximum expected return.

[0028] During the training process, gradient-based optimization methods (policy gradient methods) are used to adjust PID parameters and continuously perform adaptive optimization and control of the system. This process is adaptive, and the agent can optimize the control strategy according to changes in the system and external disturbances.

[0029] A method for using an AI-based precision hybrid variable irrigation and fertilization system includes the following steps: S1: Start the intelligent water and fertilizer machine; S2: Based on actual fertilization and irrigation needs, set detailed parameters such as fertilization type, fertilization amount, and irrigation amount for each irrigation area through the human-machine interface of the intelligent water and fertilizer machine, the cloud system web interface, or even the mobile app, complete the system self-check, and troubleshoot any possible faults. S3: After the self-inspection is passed, the fertilization strategy is preset for each fertilization plot, and the fertilization and irrigation operation of the plot is started. S4: Community fertilization and irrigation operation. The main pipeline solenoid valve opens, and according to the preset irrigation and fertilization plan for each community, the corresponding solenoid valves of the community water and fertilizer monitoring feedback device and the fertilizer tank solenoid valve are activated. The main pipeline booster pump starts. Irrigation water is thoroughly mixed with fertilizer solution in the mixing pipe after passing through the intelligent water and fertilizer machine to form irrigation solution, which flows into the branch pipe and then into the community water and fertilizer monitoring feedback device. The irrigation solution undergoes parameter detection in the community water and fertilizer monitoring feedback device. The device's wireless transceiver transmits the detection information back to the intelligent water and fertilizer machine's central controller in real time. The intelligent water and fertilizer machine's central controller, based on the fertilization type and detection information, calls the preset water and fertilizer information analysis model, analyzes the information in real time, and displays it on multiple devices. The system monitors the concentration of each fertilizer in the community's irrigation solution, the progress of irrigation and fertilization, and the system's operating status. Using analytical and preset information as input, and based on a system intelligent dynamic coupling control algorithm combining fuzzy PID and reinforcement learning, it dynamically and collaboratively controls the main pipeline booster pump, the electromagnetic flow regulating valve in the mixing pipeline, and the fertilizer solution flow regulating pump. This ensures the system's fertilization and irrigation modes remain within preset ranges. The monitored irrigation solution flows to designated locations through the community's drip irrigation system. Furthermore, based on real-time feedback from the community's water and fertilizer monitoring feedback device regarding changes in the community's pipeline pressure, it can monitor for sudden problems such as pipe leaks and capillary bursts in the community's drip irrigation system, promptly shutting down the electromagnetic valve of the community's water and fertilizer monitoring feedback device and issuing an alarm. S5: After the system completes the fertilization and irrigation target of the target community based on the information fed back by the community water and fertilizer monitoring feedback device, the system issues an instruction to close the solenoid valve of the target community water and fertilizer monitoring feedback device, and performs the fertilization and irrigation operation of the next community according to the preset task. After the fertilization and irrigation operation of all communities is completed, the system closes all solenoid valves, sends a hibernation instruction to the water and fertilizer monitoring feedback device of each community, and the system is powered off.

[0030] The present invention achieves the following technical effects compared to the prior art: 1. The present invention connects a community water and fertilizer monitoring and feedback device in series on the branch pipe of the irrigation community. It has the function of automatically detecting irrigation liquid parameters and feeding them back to the intelligent water and fertilizer machine in real time. The intelligent water and fertilizer machine can analyze the water and fertilizer information of the irrigation liquid in real time based on the real-time feedback information and the AI ​​water and fertilizer information analysis model. It can provide real-time feedback on the quality of fertilization and irrigation throughout the process and accurately and efficiently control the amount of irrigation water and fertilizer.

[0031] 2. This invention sets up a water and fertilizer monitoring feedback device in the fertilization and irrigation area. This device realizes real-time detection, feedback and AI intelligent analysis of irrigation liquid parameters during the fertilization and irrigation process. Based on the linkage of the main pipeline pressurization pump, the intelligent water and fertilizer machine fertilizer liquid flow regulating pump and the electromagnetic flow regulating valve of the mixing pipeline, it uses an intelligent dynamic coupling optimization and control algorithm based on the combination of fuzzy PID control and reinforcement learning to regulate water and fertilizer. It realizes system-level precise mixed variable fertilization irrigation and real-time multi-terminal feedback and display of the system operation process and the final effect, realizing intelligent precise mixed variable fertilization irrigation in a simple, intuitive and clear way.

[0032] 3. The system of this invention utilizes the strong adaptability and systematic self-optimization capabilities of AI when facing complex, nonlinear, and time-varying systems. Based on fuzzy PID and reinforcement learning, it continuously learns and adaptively optimizes, dynamically and collaboratively controlling the main pipeline pressurization pump, the intelligent water and fertilizer machine fertilizer liquid flow regulating pump, and the electromagnetic flow regulating valve in the mixing pipeline. Based on actual environmental feedback, it optimizes the control strategy, enabling adaptive optimization during system operation and continuously improving the system's control accuracy and response speed. It does not require manual adjustment or redevelopment of control algorithms based on changes in the system environment, thus enhancing the applicability of the entire system.

[0033] Therefore, the present invention adopts the above-mentioned AI-based precision mixed variable irrigation and fertilization system and method. Based on the existing design, it innovatively proposes to monitor the automatic precision mixed variable fertilization irrigation process by coupling AI algorithm capabilities with existing water and fertilizer monitoring hardware, and to combine fuzzy PID and reinforcement learning for adaptive coupling optimization and control. It solves the problem of real-time online monitoring and feedback of the effect of precision mixed variable irrigation and fertilization process in irrigation plots and the problem of adaptive optimization and control of the system in production practice. It has the advantages of outstanding performance, easy operation and strong applicability.

[0034] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of an embodiment of an AI-based precision mixed variable irrigation and fertilization system and method according to the present invention; Figure 2 This is a schematic diagram of the system hardware of an embodiment of the AI-based precision mixed variable irrigation and fertilization system and method of the present invention; Figure 3 This is a schematic diagram of a community water and fertilizer monitoring feedback device, which is an embodiment of an AI-based precision mixed variable irrigation and fertilization system and method according to the present invention. Figure 4This is a schematic diagram of the adaptive coupling optimization control algorithm of a precision mixed variable irrigation and fertilization system based on fuzzy PID and reinforcement learning, according to an embodiment of the present invention. Figure 5 This is a flowchart illustrating the operation of the human-machine interface of an intelligent water and fertilizer machine, representing an embodiment of an AI-based precision mixed variable irrigation and fertilization system and method according to the present invention. Figure 6 This is a diagram showing the effect of the AI ​​water and fertilizer information analysis model on the prediction of the concentration of the first fertilizer in the irrigation solution in an embodiment of the AI-based precision mixed variable irrigation and fertilization system and method of the present invention. Figure 7 This is a graph showing the effect of the AI ​​water and fertilizer information analysis model of the present invention, which is used to predict the concentration of the second type of fertilizer in the irrigation solution, in an embodiment of the AI-based precision mixed variable irrigation and fertilization system and method.

[0036] Attached Figure

[0037] 1. Water storage tank; 2. Water level sensor; 3. Main pipeline filter; 4. Main pipeline solenoid valve; 5. Main pipeline booster pump; 6. Main pipeline check valve; 7. Main pipeline pressure gauge; 8. Intelligent fertigation machine central controller; 9. Fertilizer tank; 10. Fertilizer solenoid valve; 11. Fertilizer filter; 12. Fertilizer check valve; 13. Fertilizer flow regulating pump; 14. Venturi tube; 15. Mixing pipeline solenoid flow regulating valve; 16. Mixing pipeline; 17. Irrigation branch pipe; 18. Community water and fertilizer monitoring feedback device; 19. Community drip irrigation system; 20. Community water and fertilizer monitoring feedback device solenoid valve; 21. Community water and fertilizer monitoring feedback device irrigation liquid EC sensor; 2 2. Irrigation solution pH sensor for community water and fertilizer monitoring feedback device; 23. Irrigation solution temperature sensor for community water and fertilizer monitoring feedback device; 24. Effective nitrogen monitor for irrigation solution for community water and fertilizer monitoring feedback device; 25. Electromagnetic flow meter for community water and fertilizer monitoring feedback device; 26. Irrigation solution pressure sensor for community water and fertilizer monitoring feedback device; 27. Photovoltaic panel bracket; 28. Solar photovoltaic panel; 29. ​​Cable; 30. Wireless transceiver for community water and fertilizer monitoring feedback device; 31. Solar power supply system column for community water and fertilizer monitoring feedback device; 32. Protective casing for community water and fertilizer monitoring feedback device; 33. Base of solar power supply system column for community water and fertilizer monitoring feedback device. Detailed Implementation

[0038] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0039] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "comprising" or "including," as used in this invention, mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. The terms "connected" or "linked," etc., are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. Example

[0040] This invention provides an AI-based precision mixed variable irrigation and fertilization system, which can be divided into three main parts: water and fertilizer mixing part, community water and fertilizer monitoring and feedback part, and system coupling dynamic optimization and control part.

[0041] The following sections will introduce these three parts in detail.

[0042] I. Water and Fertilizer Mixing Section like Figure 1 As shown, the water-fertilizer mixing section includes: a water storage tank 1, a water level sensor 2, a main pipeline filter 3, a main pipeline solenoid valve 4, a main pipeline booster pump 5, a main pipeline check valve 6, a main pipeline pressure gauge 7, a fertilizer solution tank 9, a fertilizer solution solenoid valve 10, a fertilizer solution filter 11, a fertilizer solution check valve 12, a fertilizer solution flow regulating pump 13, a venturi tube 14, a mixing pipeline electromagnetic flow regulating valve 15, and a mixing pipeline 16.

[0043] The water storage tank 1 is used to store water and contains a water level sensor 2. The water level sensor 2 is connected to the central controller 8 of the intelligent water and fertilizer machine via a cable. The outlet of the water storage tank 1 is connected to the main pipeline. Irrigation water flows through the main pipeline and passes sequentially through the main pipeline filter 3, the main pipeline solenoid valve 4, the main pipeline booster pump 5, the main pipeline check valve 6, and the main pipeline pressure gauge 7 before mixing with the fertilizer solution flowing out of the intelligent water and fertilizer machine and entering the mixing pipeline 16. The main pipeline solenoid valve 4 and the main pipeline booster pump 5 are both connected to the drive relay in the central controller 8 of the intelligent water and fertilizer machine via cables.

[0044] The fertilizer tank 9 contains the fertilizer solution required for fertilization. After the fertilizer solution in the fertilizer tank 9 passes through the outlet, it flows into the mixing pipe 16 in sequence through the fertilizer solenoid valve 10, fertilizer filter 11, fertilizer check valve 12, fertilizer flow regulating pump 13, and venturi tube 14 to mix with the irrigation water in the pipe to form irrigation solution.

[0045] II. Feedback from Community Water and Fertilizer Monitoring like Figure 2As shown, the community water and fertilizer monitoring feedback device includes: a solenoid valve 20, an irrigation liquid EC sensor 21, an irrigation liquid pH sensor 22, an irrigation liquid temperature sensor 23, an irrigation liquid effective nitrogen monitor 24, an electromagnetic flow meter 25, an irrigation liquid pressure sensor 26, a photovoltaic panel bracket 27, a solar photovoltaic panel 28, a cable 29, a wireless transceiver 30, a solar power supply system column 31, a protective casing 32, and encapsulated within it a battery, a power supply system control circuit board, a power supply system drive relay, and a solar power supply system column base 33.

[0046] The community water and fertilizer monitoring feedback device driver relay, the community water and fertilizer monitoring feedback device irrigation liquid EC sensor 21, the community water and fertilizer monitoring feedback device irrigation liquid pH sensor 22, the community water and fertilizer monitoring feedback device irrigation liquid temperature sensor 23, the community water and fertilizer monitoring feedback device irrigation liquid effective nitrogen monitor 24, the community water and fertilizer monitoring feedback device irrigation liquid electromagnetic flow meter 25, the community water and fertilizer monitoring feedback device irrigation liquid pressure sensor 26, and the community water and fertilizer monitoring feedback device wireless signal transceiver 30 are connected to the community water and fertilizer monitoring feedback device solar power supply system via cables to obtain power energy during operation. The community water and fertilizer monitoring feedback device driver receives signal commands from the community water and fertilizer monitoring feedback device signal transceiver system.

[0047] The community water and fertilizer monitoring feedback device includes a drive relay, an irrigation liquid EC sensor 21, an irrigation liquid pH sensor 22, an irrigation liquid temperature sensor 23, an irrigation liquid effective nitrogen monitor 24, an electromagnetic flow meter 25, an irrigation liquid pressure sensor 26, and a solar power supply system. These components are connected to the community water and fertilizer monitoring feedback device's wireless signal transceiver 30 via cables. The data reading module in the wireless signal transceiver transmits the data back to the intelligent water and fertilizer machine via a signal transmitter, enabling real-time feedback of the irrigation community water and fertilizer monitoring feedback device system's operating status.

[0048] The solar power supply system for the community water and fertilizer monitoring feedback device consists of a solar photovoltaic panel 28, a battery, a power supply system drive relay, an inverter, a power supply system control circuit board, a protective enclosure 32, and a column 31. The solar photovoltaic panel 28 is located on the photovoltaic panel support 27. The power supply system drive relay is connected to the battery via a cable. The solar photovoltaic panel is connected to the battery via an inverter. The battery, the power supply system control circuit board, the inverter, and the power supply system drive relay are encapsulated in the protective enclosure 32.

[0049] After the irrigation solution, which has been uniformly mixed in the mixing pipe, reaches the community water and fertilizer monitoring and feedback device through the irrigation branch pipe 17, it passes sequentially through the community water and fertilizer monitoring and feedback device solenoid valve 20, the community water and fertilizer monitoring and feedback device irrigation solution EC sensor 21, the community water and fertilizer monitoring and feedback device irrigation solution pH sensor 22, the community water and fertilizer monitoring and feedback device irrigation solution temperature sensor 23, the community water and fertilizer monitoring and feedback device irrigation solution effective nitrogen monitor 24, the community water and fertilizer monitoring and feedback device electromagnetic flowmeter 25, and the community water and fertilizer monitoring and feedback device irrigation solution pressure sensor 26 before flowing into the community drip irrigation system 19. The community drip irrigation system 19 mainly consists of the irrigation branch pipe 17 and the community capillary pipes, and is responsible for uniformly delivering water and fertilizer solution to the root zone of each crop, providing the water and nutrients required for crop growth in a timely and efficient manner.

[0050] Meanwhile, data such as irrigation solution EC, pH, temperature, available nitrogen, flow rate, and pressure collected by the community water and fertilizer monitoring feedback device are transmitted in real-time online to the smart water and fertilizer machine's wireless transceiver (encapsulated in the smart water and fertilizer machine's central controller 8) via the community water and fertilizer monitoring feedback device's wireless transceiver 31, realizing real-time feedback of community water and fertilizer information. Furthermore, the smart water and fertilizer machine's command signals to the community water and fertilizer monitoring feedback device are also transmitted via the smart water and fertilizer machine's wireless transceiver, received by the community water and fertilizer monitoring feedback device's wireless transceiver 31, and then transmitted to the community water and fertilizer monitoring feedback device to control the opening and closing of the community water and fertilizer monitoring feedback device's solenoid valve 20.

[0051] III. System Coupling Dynamic Optimization and Control Section like Figure 4 As shown, the system coupling dynamic optimization and control part includes: main pipeline pressurization pump 5, fertilizer solution flow regulating pump 13, mixing pipeline electromagnetic flow regulating valve 15, intelligent water and fertilizer machine central controller 8, intelligent water and fertilizer machine wireless signal transceiver (encapsulated in intelligent water and fertilizer machine central controller 8), intelligent water and fertilizer machine drive relay (encapsulated in intelligent water and fertilizer machine central controller 8), community water and fertilizer monitoring feedback device wireless signal transceiver 31, community water and fertilizer monitoring feedback device drive relay (encapsulated in community water and fertilizer monitoring feedback device encapsulation protective shell 33).

[0052] The irrigation solution EC, pH, temperature, available nitrogen, flow rate, and pressure data collected by the community water and fertilizer monitoring feedback device are transmitted in real-time online to the intelligent water and fertilizer machine's wireless transceiver (encapsulated in the intelligent water and fertilizer machine's central controller 8) via the device's wireless transceiver 31. The data is then processed by the AI ​​water and fertilizer information analysis model in the intelligent water and fertilizer machine's central controller 8 for real-time online analysis of water and fertilizer concentration, irrigation and fertilization rate, and irrigation and fertilization progress. This data is also displayed synchronously across multiple terminals, enabling human-computer interaction. Figure 4 As shown, the central controller 8 of the intelligent water and fertilizer machine compares the preset fertilization and irrigation strategy with the real-time information of the community. It uses an adaptive coupling optimization and control algorithm based on fuzzy PID and reinforcement learning to perform coupled dynamic optimization and control of the entire system operation process, continuously improving the system's accuracy, stability and reliability during operation.

[0053] The present invention also provides a method for using the above-mentioned AI-based precision mixed variable fertilization irrigation system, comprising the following steps: S1: Start the intelligent water and fertilizer machine; S2: Based on actual fertilization and irrigation needs, set the fertilization type, fertilization amount, and irrigation amount for each irrigation area on the human-machine interface of the intelligent water and fertilizer machine, complete the system self-check, and troubleshoot any possible faults. S3: After the self-inspection is passed, the fertilization strategy is preset for each fertilization plot, and the fertilization and irrigation operation of the plot is started. S4: Community fertilization and irrigation operation. The main pipeline solenoid valve opens, and the corresponding solenoid valves of the community water and fertilizer monitoring feedback device and fertilizer tank solenoid valves are activated according to the preset irrigation and fertilization plan for each community. The main pipeline booster pump starts. Irrigation water is fully mixed with fertilizer solution in the mixing pipeline through the intelligent water and fertilizer machine to form irrigation solution, which flows into the branch pipe and then into the community water and fertilizer monitoring feedback device. The irrigation solution undergoes parameter detection in the community water and fertilizer monitoring feedback device. The device's wireless transceiver transmits the detection information back to the intelligent water and fertilizer machine in real time. The intelligent water and fertilizer machine calls the pre-defined AI fertilizer solution information analysis model according to the fertilizer type and detection information. It analyzes and displays the fertilizer concentration and irrigation and fertilization progress of each fertilizer in the community in real time, and compares the analyzed information with the preset information. It uses an adaptive coupling optimization and control algorithm based on fuzzy PID and reinforcement learning to dynamically and collaboratively control the main pipeline booster pump, the mixing pipeline solenoid flow regulating valve, and the fertilizer solution flow regulating pump to make the system's fertilization and irrigation mode consistent with the preset mode. The tested irrigation solution flows to the designated location through the community drip irrigation system. In addition, based on the real-time feedback of the community-end pipeline pressure changes from the community water and fertilizer monitoring feedback device, sudden problems such as pipeline leaks and capillary bursts in the community drip irrigation system can be monitored in real time, and the solenoid valve of the community water and fertilizer monitoring feedback device can be shut off in a timely manner and an alarm can be issued.

[0054] S5: After the system completes the fertilization and irrigation target of the target community based on the information fed back by the community water and fertilizer monitoring feedback device, the system issues an instruction to close the solenoid valve of the target community water and fertilizer monitoring feedback device, and performs the fertilization and irrigation operation of the next community according to the preset task. After the fertilization and irrigation operation of all communities is completed, the system closes all solenoid valves, sends a hibernation instruction to the water and fertilizer monitoring feedback device of each community, and the system is powered off.

[0055] Therefore, the present invention adopts the above-mentioned AI-based precision mixed variable irrigation and fertilization system and method, which solves the problem of real-time online monitoring and feedback of the effect of mixed precision variable irrigation and fertilization process in irrigation plots and adaptive coupling optimization control in production practice. It has the advantages of outstanding performance, easy operation and strong applicability.

[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An AI-based precise mixed variable irrigation and fertilization method, characterized in that: It includes a water and fertilizer mixing section, a community water and fertilizer monitoring and feedback device, and a system adaptive coupling optimization and control section, including the following steps: S1, pre-sets the type of mixed fertilizer and the target nitrogen fertilizer concentration C for the irrigation and fertilization operation area. N Target phosphate fertilizer concentration C P Target potassium fertilizer concentration C K and target system traffic Q w ; S2, during the irrigation and fertilization operation, the community water and fertilizer monitoring feedback device set in the irrigation branch pipe collects the effective nitrogen content, EC, pH, temperature, flow rate and pressure of the irrigation liquid in real time, and sends the collected data to the central controller of the intelligent water and fertilizer machine. S3, the central controller of the intelligent water and fertilizer machine, based on the type of mixed fertilizer, calls a water and fertilizer information analysis model matching the type of mixed fertilizer from the AI-based water and fertilizer information analysis model library, and inputs the effective nitrogen content, EC, pH, and temperature into the water and fertilizer information analysis model to obtain the real-time nitrogen fertilizer concentration C of the irrigation mixture. t N Real-time phosphate fertilizer concentration C t P and real-time potassium fertilizer concentration C t K ; S4, based on the real-time nitrogen fertilizer concentration C t N Real-time phosphate fertilizer concentration C t P Real-time potassium fertilizer concentration C t K and real-time system traffic Q t w Calculate the concentration of nitrogen fertilizer C relative to the target concentration. N Target phosphate fertilizer concentration C P Target potassium fertilizer concentration C K and target system traffic Q w Error E t N E t P E t K E t w and the corresponding error change rate EC t N EC t P EC t K EC t w ; S5, the Q-learning algorithm provides an initialization strategy for the PID parameter set used to regulate the speed of the main pipeline booster pump, the duty cycle of the mixing pipeline flow regulating valve, the speed of the nitrogen fertilizer flow regulating pump, the speed of the phosphate fertilizer flow regulating pump, and the speed of the potash fertilizer flow regulating pump, based on historical operating data and current preset data; the error E t N E t P E t K E t w and the corresponding error change rate EC t N EC t P EC t K EC t w As input variables to the fuzzy PID controller, real-time PID control parameters are obtained through fuzzy inference and defuzzification; according to the reward function: Calculate the reward, where, α , β , γ and δ All are weighting coefficients. E ( t ) 2 and EC ( t ) 2 Let these represent the error and the square of the rate of change of the error, respectively. u ( t ) 2 This represents the square of the control input. RT It is the time required for the system to start responding and reach a certain preset steady-state error range, and the estimated value of the Q-function is updated according to the Bellman equation based on the reward to adjust the PID parameter set; the adjusted PID parameters are then input into the corresponding PID controllers to obtain control signals; S6, based on the control signal, coordinately adjust the speed of the main pipeline booster pump, the duty cycle of the mixing pipeline flow regulating valve, the speed of the nitrogen fertilizer flow regulating pump, the speed of the phosphate fertilizer flow regulating pump, and the speed of the potash fertilizer flow regulating pump to achieve the real-time nitrogen fertilizer concentration C. t N Real-time phosphate fertilizer concentration C t P Real-time potassium fertilizer concentration C t K and real-time irrigation fluid flow rate Q t w Keep within the corresponding preset range; The AI-based water and fertilizer information model analysis library is based on the combination of soluble nitrogen, phosphorus, and potassium fertilizers in fertigation to form various types of fertigation mixed fertilizer solutions. Machine learning models are constructed for each type of mixed fertilizer solution at different component concentrations to determine the effective nitrogen content, EC, pH, temperature, and physical parameters of the solution, as well as the concentrations of different fertilizer components in the solution. The collection of these models constitutes the AI-based water and fertilizer information analysis model library. The nitrogen fertilizers include urea, ammonium nitrate, and ammonium sulfate; the phosphorus fertilizers include monoammonium phosphate, diammonium phosphate, and superphosphate; and the potassium fertilizers include potassium chloride, potassium sulfate, and potassium nitrate.

Images