Water and fertilizer integrated machine irrigation control method and system based on internet of things

Abstract

The application belongs to the technical field of irrigation control, and particularly relates to a water and fertilizer integrated machine irrigation control method and system based on the Internet of Things, which comprises the following steps: calculating equivalent hydraulic load by collecting real-time operation data of an irrigation system and combining a pre-calibrated partition load coefficient; predicting a dynamic pressure benchmark based on the equivalent hydraulic load and water pump frequency by using a dynamic prediction model, combining real-time flow to predict an expected pressure standard deviation, normalizing the deviation of the measured pressure from the dynamic benchmark by standard deviation to generate a fault feature index; and statistically monitoring the index by an exponentially weighted moving average control chart, determining a fault once the index exceeds a control limit, identifying a fault type according to the index deviation direction, and locating a fault partition in combination with a current valve state vector. The application can effectively distinguish pressure fluctuations caused by normal operation switching from abnormal deviations caused by end blockage, leakage and other faults, and realize intelligent fault diagnosis and positioning of the irrigation system.

Description

Technical Field

[0001] This invention relates to the field of irrigation control technology. More specifically, this invention relates to an irrigation control method and system for an integrated water and fertilizer machine based on the Internet of Things (IoT). Background Technology

[0002] The integrated water and fertilizer irrigation system is the core equipment of modern precision agriculture. It delivers precisely proportioned water and nutrients directly to the roots of crops through a pipeline network. The stable operation of this system is the key to ensuring healthy crop growth and conserving water and fertilizer resources.

[0003] However, irrigation networks are complex and operate in harsh environments. The drippers or branch pipes at the ends often become clogged due to impurities, fertilizer crystals, or biological accumulation, leading to irrigation failure in some areas and seriously affecting crop yield and quality.

[0004] To achieve real-time monitoring and fault diagnosis of system operation status, the Statistical Process Control (SPC) method has been introduced. SPC identifies faults by constructing control charts and using statistical principles to distinguish between normal fluctuations in the process and fluctuations caused by abnormal causes. Typical SPC methods, such as the Exponential Weighted Moving Average (EWMA) control chart, determine abnormalities by monitoring key process parameters, such as the pressure of the main pipeline of the pipeline network, to see if they exceed preset statistical control limits.

[0005] However, irrigation systems are not a steady-state production process; their normal operating conditions are dynamic. When the system switches irrigated plots or adjusts valve combinations in different irrigation zones, the fluid dynamics of the entire network changes, causing planned steps or drifts in the normal operating pressure and flow of the main pipeline. Traditional SPC methods use fixed centerlines and control limits, which cannot distinguish between such planned normal state switching and abnormal state deviations caused by faults such as blockages. Therefore, a large number of false alarms are generated, making them unreliable and unusable in practical applications. Summary of the Invention

[0006] To address the technical problem that traditional SPC methods, which use fixed centerlines and control limits, cannot distinguish between planned normal state transitions and abnormal state deviations caused by faults such as blockages, resulting in a large number of false alarms, this invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides an irrigation control method for an integrated water and fertilizer system based on the Internet of Things, comprising: acquiring real-time operating data of the irrigation system, including real-time pressure value, real-time flow value, current operating frequency of the water pump, and valve state vectors describing the opening and closing states of valves in each irrigation zone; calculating the equivalent hydraulic load characterizing the current system operating state based on the valve state vectors and pre-calibrated load coefficients of each irrigation zone; predicting the dynamic pressure benchmark under the current operating condition based on a dynamic prediction model using the equivalent hydraulic load and the current operating frequency of the water pump, and predicting the expected pressure standard deviation under the current operating condition using the real-time flow value of the main pipeline; normalizing the difference between the real-time pressure value and the dynamic pressure benchmark using the expected pressure standard deviation to obtain fault characteristic indicators; statistically monitoring the fault characteristic indicators using an exponentially weighted moving average control chart, and determining that a system fault has occurred when the statistical quantity exceeds a preset control limit, and distinguishing the fault type according to the deviation direction and determining the fault zone according to the current valve state vector.

[0008] Preferably, the method for obtaining the pre-calibrated load factor of each irrigation zone includes: closing all irrigation zone valves while the water pump is running and recording the main pipeline pressure at this time as a reference pressure; opening each irrigation zone individually in sequence, and after the water flow stabilizes, measuring and recording the stable pressure and stable flow rate of the main pipeline under that irrigation zone; dividing the difference between the reference pressure and the stable pressure when individually opened by the square of the corresponding stable flow rate to obtain the load factor of that irrigation zone.

[0009] Preferably, the formula for calculating the equivalent hydraulic load is: In the formula, Equivalent hydraulic load; For the first Valve status of each irrigation zone; For the first Load factor of each irrigation zone; This represents the total number of irrigation zones.

[0010] This invention transforms the discrete multidimensional valve switching state vector into a single and continuous equivalent hydraulic load scalar. By weighted averaging of the load coefficients of each open zone, it not only considers the number of open zones but also their quality, i.e., the difference in hydraulic impedance between different zones. This allows for a more accurate characterization of the overall hydraulic condition of the system and provides input variables for the accuracy of subsequent dynamic prediction models.

[0011] Preferably, the method for obtaining the dynamic prediction model includes: using a training dataset, fitting a dynamic pressure benchmark for prediction through a multivariate nonlinear regression algorithm. The model, in its specific form, is as follows: In the formula, The predicted dynamic pressure baseline serves as the dynamic centerline of the SPC control chart. This is the current operating frequency of the water pump; Equivalent hydraulic load; These are the model coefficients obtained through regression fitting.

[0012] This invention utilizes pump frequency and equivalent hydraulic load as inputs to more accurately capture the nonlinear physical laws of pressure changes in the system under different operating conditions. This allows the pressure benchmark to be dynamically adjusted in real time according to the operating conditions, thus providing a dynamic normal reference standard for fault diagnosis. This is the core of achieving high-precision diagnosis.

[0013] Preferably, the step of predicting the expected pressure standard deviation under the current operating condition using the real-time flow value of the main pipeline of the pipeline network includes: using the training dataset, analyzing the pressure data within each stable operating condition window. Calculate its standard deviation and pair it with the average flow rate within that window; then, fit the predicted standard deviation of the expected pressure using linear regression analysis. The model, in its specific form, is as follows: In the formula, The expected standard deviation of the predicted stress; This is a real-time traffic value; , These are the model coefficients obtained through linear regression.

[0014] This invention establishes a linear model of the expected standard deviation of the predicted flow rate and the real-time flow rate based on the natural fluctuation amplitude of pressure. This model allows the denominator used for normalization to be adaptively adjusted according to the operating conditions. The tolerance is relaxed when the flow rate is large and the fluctuation is large, and the tolerance is tightened when the flow rate is small and the system is stable. This improves the signal-to-noise ratio of the fault characteristic indicators and the reliability of the diagnosis.

[0015] Preferably, the step of using an exponentially weighted moving average control chart to statistically monitor the fault characteristic indicators includes: calculating the exponentially weighted moving average statistic for the real-time generated fault characteristic indicators. In the formula, This is the EWMA statistic for the current moment; The fault characteristic indicators at the current moment; This is the EWMA statistic from the previous time step; This is the smoothing coefficient.

[0016] This invention integrates information from current and historical values ​​through EWMA statistics, which has a memory effect. This makes the method extremely sensitive to small but persistent system deviations, such as initial small leaks or slow blockages. It can amplify fault signals through cumulative effects, thereby achieving early fault detection. At the same time, it can effectively smooth out accidental or instantaneous sensor noise or disturbances, enhancing the stability of monitoring.

[0017] Preferably, the preset control limits of the exponentially weighted moving average control chart include an upper control limit and a lower control limit, whose values ​​are set to 3 and -3, respectively.

[0018] Preferably, the step of distinguishing the fault type based on the deviation direction includes: if the EWMA statistic at the current moment exceeds the upper control limit, the fault type is determined to be a blockage fault; if the EWMA statistic at the current moment is lower than the lower control limit, the fault type is determined to be a leakage fault.

[0019] Preferably, determining the fault zone based on the current valve state vector includes: reading and parsing the current valve state vector. Find the irrigation partitions corresponding to all elements with a state of "1" and use them as fault partitions.

[0020] Secondly, the present invention provides an IoT-based water and fertilizer integrated irrigation control system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned IoT-based water and fertilizer integrated irrigation control method is implemented.

[0021] By adopting the above technical solution, a computer program is generated from the above-mentioned IoT-based water and fertilizer integrated irrigation control method and stored in the memory so that it can be loaded and executed by the processor. In this way, a terminal device can be made based on the memory and the processor for convenient use.

[0022] The beneficial effects of this invention are as follows:

[0023] This invention, by constructing an equivalent hydraulic load index and a dynamic pressure benchmark model, can accurately predict the pressure performance of the system under any normal operating conditions, thereby effectively distinguishing pressure fluctuations caused by normal operation switching from abnormal deviations caused by terminal blockages, leaks, and other faults, and solving the problem of high false alarm rate of traditional SPC methods under dynamic operating conditions.

[0024] Furthermore, by standardizing the pressure deviation, the fault characteristic index constructed in this invention amplifies weak fault signals. Combined with the sensitivity of the EWMA control chart to continuous small deviations, this method can detect early or minor faults such as single or a small number of dripper blockages, providing a time window for preventive maintenance. Attached Figure Description

[0025] The above and other objects, features, and advantages of the present invention will become readily apparent from the following detailed description of exemplary embodiments, accompanied by the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein:

[0026] Figure 1 This is a schematic flowchart illustrating an IoT-based water and fertilizer integrated irrigation control method according to the present invention.

[0027] Figure 2 This is a flowchart illustrating step S4;

[0028] Figure 3 A schematic diagram illustrating dynamic reference pressure monitoring under different operating conditions is shown.

[0029] Figure 4 This diagram illustrates the performance of the traditional SPC method in irrigation system fault diagnosis tasks.

[0030] Figure 5 The schematic diagram illustrates the performance of the solution of the present invention in the task of diagnosing irrigation system faults. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0033] This invention discloses an irrigation control method for an integrated water and fertilizer machine based on the Internet of Things, referring to... Figure 1 This includes steps S1 to S4:

[0034] S1: Acquire real-time operating data of the irrigation system and preprocess the data.

[0035] It should be noted that in order to fully and accurately grasp the operating status of the irrigation system, it is necessary to simultaneously acquire various types of data that can reflect the system's input, output, and control status. Furthermore, in order to eliminate the interference of sensor noise and transient disturbances on subsequent analysis, the raw data needs to undergo preliminary purification processing to ensure data quality.

[0036] Specifically, after the system starts up, it synchronously acquires multi-source heterogeneous data from the integrated water and fertilizer machine control system and pipeline network sensors through the IoT data acquisition interface. This dataset includes at least: the real-time pressure value of the main pipeline network. Real-time flow rate of the main pipeline The current operating frequency of the water pump And a description of the opening and closing status of valves in each irrigation zone. 3D binary valve state vector The valve state vector S is in the form of ,in, This represents the total number of irrigation zones. Indicates the first The valves in each irrigation zone are open. This indicates that the device is in the off state.

[0037] Furthermore, the collected real-time pressure values and real-time traffic value A moving average filter is applied to smooth the data and eliminate high-frequency noise, with the sliding window size set to 5.

[0038] S2: Calculate the equivalent hydraulic load characterizing the current system operating state based on the valve status and pre-calibrated load coefficient of each irrigation zone.

[0039] It should be noted that traditional methods directly monitor raw data such as pressure, which cannot adapt to changes in operating conditions. Therefore, this invention constructs a comprehensive index that can accurately quantify the current hydraulic requirements of the system, namely the equivalent hydraulic load. This index transforms the discrete valve opening and closing state vector into a continuously changing value to characterize the current normal operating condition of the system.

[0040] Specifically, during the initial system deployment or maintenance, a one-time calibration procedure is performed: while the water pump is running, all irrigation zone valves are first closed, and the main pipeline pressure at this time is measured and recorded as a reference pressure. Subsequently, each irrigation zone was activated individually in turn. After the water flow in the pipeline network stabilizes, measure and record the stable pressure of the main pipeline in that irrigation zone. and stable traffic .

[0041] Furthermore, based on the measured data, the first... Load factor of each irrigation zone The specific calculation formula is as follows:

[0042]

[0043] In the formula, For the first Load factor of each irrigation zone; This is the reference pressure when all valves are closed. To open the first one separately Stable pressure in each irrigation zone; To open the first one separately Stable flow rate for each irrigation zone.

[0044] Among them, load factor Approximately characterizes the first The equivalent hydraulic impedance of the irrigation zone network is the value of the load factor of the irrigation zone with the greater impedance.

[0045] Specifically, during normal system operation, based on the valve state vector acquired in real time... Calculate the equivalent hydraulic load at the current moment using the pre-calibrated load factors of each irrigation zone. The specific calculation formula is as follows:

[0046]

[0047] In the formula, Equivalent hydraulic load; For the first The valve status of each irrigation zone is set to either 0 or 1. For the first Load factor of each irrigation zone; This represents the total number of irrigation zones.

[0048] The numerator reflects the current total load by linearly superimposing the load coefficients of all activated irrigation zones; the denominator is a square root normalization of the number of activated irrigation zones. When the number of activated irrigation zones increases, the denominator grows more slowly than the numerator, resulting in... The value primarily reflects the increase in total load; however, when the number of irrigation zones remains unchanged, but the load factor is activated... When the irrigation zone is higher, The value will increase significantly, making the index more sensitive to the quality-to-quantity ratio of the irrigation zone opening, thus enabling it to more stably characterize the system's operating conditions.

[0049] It should be noted that this step proposes an equivalent hydraulic load index, which maps the complex and discrete valve combination states into a single and continuous physical quantity, providing a key standardized input variable for the subsequent construction of a dynamic benchmark model that can adapt to any operating condition.

[0050] S3: Based on the equivalent hydraulic load and system operating parameters, construct a dynamic prediction model to obtain the dynamic pressure benchmark and the expected pressure standard deviation under the current operating conditions.

[0051] It should be noted that since the normal pressure value and its fluctuation range of the irrigation system change dynamically with the operating conditions, fixed monitoring thresholds are no longer applicable. Therefore, it is necessary to establish a model that can predict the normal pressure performance in real time based on the current operating conditions, including its expected center value and reasonable fluctuation range, as a dynamic adaptive evaluation benchmark.

[0052] Specifically, once the system is confirmed to be fault-free, the control system drives the integrated water and fertilizer machine through a series of preset and representative normal operating conditions, including different water pump operating frequencies. Combination with different irrigation zones Under each operating condition, allow the system to run stably for a period of time and collect the corresponding data. The data points eventually form a training dataset that covers the normal operating range of the system.

[0053] Furthermore, using the training dataset, a benchmark for predicting dynamic pressure is fitted through a multivariate nonlinear regression algorithm, such as the Levenberg-Marquardt algorithm. The model, in its specific form, is as follows:

[0054]

[0055] In the formula, The predicted dynamic pressure baseline serves as the dynamic centerline of the SPC control chart. This is the current operating frequency of the water pump; Equivalent hydraulic load; These are the model coefficients obtained through regression fitting.

[0056] It should be noted that the pressure reference is approximately positively correlated with the square of the pump frequency, which is consistent with the pump characteristic curve. At the same time, the pressure will decrease as the equivalent hydraulic load increases, because a larger load means that energy is distributed to more end outlets, resulting in a drop in the pressure of the main pipeline.

[0057] For example, Figure 3 The diagram illustrates the dynamic reference pressure monitoring under different operating conditions (normal operating condition switching, blockage failure period, leakage failure period, etc.). The curve corresponding to the filtered pressure data changes stepwise with the opening / closing of different irrigation zones. The curve corresponding to the calculated dynamic pressure reference automatically adjusts with the changes in normal operating conditions: when the system normally switches zones, causing a pressure drop, the dynamic pressure reference also drops synchronously. During the blockage failure period, the actual pressure begins to be consistently higher than its normal level, i.e., the dynamic pressure reference, and the gap between the two becomes larger and larger. During the leakage failure period, the actual pressure is significantly lower than the dynamic pressure reference.

[0058] Simultaneously, using the training dataset, pressure data within each stable operating condition window were analyzed. Calculate its standard deviation and pair it with the average flow rate within that window; then, fit the predicted expected pressure standard deviation using linear regression analysis. The model, in its specific form, is as follows:

[0059]

[0060] In the formula, The expected standard deviation of the predicted stress; This is a real-time traffic value; , These are the model coefficients obtained through linear regression. It primarily reflects the pressure fluctuations caused by flow-related fluid turbulence. This represents the system's inherent, flow-independent, baseline noise level.

[0061] When the system flow rate increases, the turbulence of the fluid inside the pipe usually increases, which leads to a linear increase in the natural fluctuation amplitude of the pressure signal.

[0062] It should be noted that this step, by establishing a data-driven dynamic prediction model, replaces the static centerline and fixed control limit width in traditional SPC, enabling the monitoring benchmark to follow the changes in the normal operating conditions of the system in real time and accurately. This is the core prerequisite for achieving high-precision, low-false-alarm-rate fault diagnosis.

[0063] S4: Statistical monitoring of fault characteristic indicators is carried out through exponentially weighted moving average control charts, and fault diagnosis and alarms are performed based on the abnormality judgment rules of the control charts.

[0064] The flowchart for step S4 is shown below. Figure 2 The process includes steps S401 to S403, specifically as follows:

[0065] S401: The difference between the real-time pressure value of the main pipeline and the dynamic pressure benchmark is normalized using the expected pressure standard deviation to obtain the fault characteristic index.

[0066] It should be noted that directly comparing the difference between real-time pressure and dynamic benchmark will still result in the magnitude of the difference changing with the operating conditions, which is not conducive to setting a uniform alarm threshold. Therefore, this deviation needs to be standardized to generate a fault characteristic index that follows a standard normal distribution, thereby simplifying and standardizing the subsequent statistical monitoring process.

[0067] Specifically, during the real-time monitoring of the system, the dynamic pressure benchmark predicted in step S3 is utilized. and expected stress standard deviation The real-time pressure values ​​obtained from S1 are standardized, and fault characteristic indicators are calculated. The specific calculation formula is as follows:

[0068]

[0069] In the formula, These are fault characteristic indicators; This is the real-time pressure value; This serves as a benchmark for predicted dynamic pressure. This represents the predicted standard deviation of expected pressure.

[0070] It should be noted that the deviation between the actual pressure and its theoretical normal value under this operating condition was normalized using the normal fluctuation range that should exist under this condition. Therefore, when the system is fault-free, regardless of changes in operating conditions, The values ​​of the indicators should fluctuate slightly around 0; when a malfunction occurs, the actual pressure will deviate significantly from its dynamic benchmark, leading to... The value deviates significantly from 0.

[0071] S402: Use EWMA control charts to monitor fault characteristic indicators.

[0072] Specifically, for the time series of fault characteristic indicators generated in real time, their exponentially weighted moving average statistic is calculated. The specific calculation formula is as follows:

[0073]

[0074] In the formula, This is the EWMA statistic for the current moment; The fault characteristic indicators at the current moment; This is the EWMA statistic from the previous time step; Smoothing coefficient The range of values ​​is , The smaller the value, the more sensitive EWMA is to historical data and the stronger its ability to detect small, persistent drifts, but the slower its response to step changes. Therefore, this invention sets... It equals 0.2.

[0075] Furthermore, a fixed upper control limit (UCL) and lower control limit (LCL) are set for the EWMA control chart, because... Since it has been standardized to a dimensionless characteristic, it can be directly set according to statistical principles. For example, the upper control limit UCL=3 and the lower control limit LCL=-3 are set, which correspond to the range of 3 times the standard deviation of the standard normal distribution.

[0076] S403: Perform fault diagnosis and alarm based on the abnormality judgment rules of the control chart.

[0077] It should be noted that the EWMA statistic has a strong memory effect; a single or transient fault characteristic indicator, such as a random sensor malfunction, is insufficient to represent the EWMA statistic. Pulling out the control limits [LCL, UCL], only when the fault characteristic index value remains on one side of 0 for multiple consecutive data acquisition cycles, i.e., continuously positive or continuously negative, will this unidirectional and continuous pressure, through the cumulative effect of EWMA, cause... The value gradually drifts and eventually exceeds the control limit. Therefore, once the system issues an alarm, it is equivalent to confirming that the deviation is continuous rather than instantaneous.

[0078] It should be further noted that when a fault is detected, the system can not only issue an alarm, but also analyze fault characteristic indicators. By observing the positive and negative values, a preliminary judgment can be made as to whether the fault type is blockage (positive over-limit) or leakage (negative over-limit). Combined with the current status of the active irrigation zone, this provides crucial contextual information for maintenance personnel to quickly locate the fault point.

[0079] Specifically, real-time determination of EWMA statistic Whether it exceeds the control limit, when or When this happens, the system determines that a malfunction has occurred: if This means that the actual pressure remains significantly higher than the dynamic baseline, and the system initially classifies the fault type as a blockage fault; if This means that the actual pressure is consistently significantly lower than the dynamic baseline, and the system initially identifies the fault type as a leakage fault.

[0080] Furthermore, at the same moment the fault type is determined, the system immediately reads and parses the current valve state vector. Find the irrigation partitions corresponding to all elements with a state of "1" and use them as fault partitions.

[0081] Finally, the final alarm information is generated and output by combining the fault type and fault zone; for example, when the detected fault type is a blockage fault, the alarm information is generated in the valve state vector. The values ​​at indices 3 and 5 are 1, indicating that irrigation partitions 3 and 5 are currently running. Therefore, the final output alarm message is: "Alarm: Blockage fault detected during the operation of irrigation partitions 3 and 5".

[0082] It should be noted that this step, by constructing a standardized index that is insensitive to changes in normal operating conditions but sensitive to fault signals, and combining it with EWMA control charts, enables early detection of minor and progressive faults. At the same time, by analyzing the direction of deviation of the index, it provides an important basis for fault type judgment and rapid location.

[0083] For example, Figure 4 The diagram illustrates the performance of the traditional SPC method in irrigation system fault diagnosis. When the system normally switches irrigation zones, causing a pressure drop, the EWMA statistic immediately falls below the lower control limit because the control limit is fixed, resulting in a large number of false alarms. Although the method also triggers alarms during actual blockage and leakage faults, these real alarms are drowned out by a large amount of false information because alarms are already frequent during normal operation.

[0084] For example, Figure 5 The diagram illustrates the performance of the present invention in irrigation system fault diagnosis. During normal operation and normal condition switching, the EWMA statistic remains stable within the control limits without any false alarms. Upon entering the blockage fault period, as the actual pressure remains consistently lower than the dynamic pressure benchmark, the EWMA statistic begins to steadily decline and eventually exceeds the lower control limit at the initial stage of blockage, thus correctly triggering a fault alarm. Upon entering the leakage fault period, the EWMA statistic rises sharply and quickly exceeds the upper control limit, again correctly issuing a fault alarm. In summary, the present invention successfully identifies two actual fault periods and remains stable during normal condition switching, successfully distinguishing between normal changes and abnormal faults, achieving accurate and error-free diagnosis.

[0085] In summary, by constructing an equivalent hydraulic load and a dynamic prediction model, this invention can calculate the pressure benchmark under the current operating conditions in real time, thereby successfully separating normal pressure fluctuations within the plan from abnormal pressure deviations caused by faults. This enables the invention to achieve zero false alarms and high-sensitivity fault diagnosis under complex dynamic operating conditions, solving the technical defects of traditional SPC methods in such scenarios.

[0086] This invention also discloses an IoT-based water and fertilizer integrated irrigation control system, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement an IoT-based water and fertilizer integrated irrigation control method according to the present invention.

[0087] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

Claims

1. A water and fertilizer integrated irrigation control method based on the Internet of Things, characterized in that, include: Acquire real-time operating data of the irrigation system, including real-time pressure and flow rates of the main pipeline, current operating frequency of the pumps, and valve state vectors describing the opening and closing status of valves in each irrigation zone. Based on the valve state vector and the pre-calibrated load coefficients of each irrigation zone, the equivalent hydraulic load characterizing the current system operating state is calculated. The formula for calculating the equivalent hydraulic load is: In the formula, Equivalent hydraulic load; For the first Valve status of each irrigation zone; For the first Load factor of each irrigation zone; This represents the total number of irrigation zones; Based on the dynamic prediction model, the dynamic pressure benchmark under the current operating condition is predicted by using the equivalent hydraulic load and the current operating frequency of the pump, and the expected pressure standard deviation under the current operating condition is predicted by using the real-time flow value of the main pipeline of the pipeline network. The method for obtaining the dynamic prediction model includes: using a training dataset, fitting a dynamic pressure benchmark using a multivariate nonlinear regression algorithm. The model, in its specific form, is as follows: In the formula, The predicted dynamic pressure baseline serves as the dynamic centerline of the SPC control chart. This is the current operating frequency of the water pump; These are the model coefficients obtained through regression fitting; The difference between the real-time pressure value and the dynamic pressure benchmark of the main pipeline of the pipeline network is normalized using the expected pressure standard deviation to obtain fault characteristic indicators. The fault characteristic indicators are statistically monitored using an exponentially weighted moving average control chart. When the statistical value exceeds the preset control limit, the system is determined to have a fault. The fault type is distinguished according to the direction of deviation, and the fault zone is determined according to the current valve state vector.

2. The irrigation control method for an integrated water and fertilizer machine based on the Internet of Things according to claim 1, characterized in that, The method for obtaining the pre-calibrated load coefficients of each irrigation zone includes: With the water pump running, close all irrigation zone valves and record the main pipeline pressure as a reference pressure; open each irrigation zone individually in sequence, and after the water flow stabilizes, measure and record the stable pressure and stable flow rate of the main pipeline under that irrigation zone; divide the difference between the reference pressure and the stable pressure when it is opened individually by the square of the corresponding stable flow rate to obtain the load factor of that irrigation zone.

3. The irrigation control method for an integrated water and fertilizer machine based on the Internet of Things according to claim 1, characterized in that, The method of predicting the expected pressure standard deviation under current operating conditions using the real-time flow value of the main pipeline of the pipeline network includes: Using the training dataset, pressure data within each steady-state operating window was analyzed. Calculate its standard deviation and pair it with the average flow rate within that window; then, fit the predicted standard deviation of the expected pressure using linear regression analysis. The model, in its specific form, is as follows: ； In the formula, The expected standard deviation of the predicted stress; This is a real-time traffic value; , These are the model coefficients obtained through linear regression.

4. The irrigation control method for an integrated water and fertilizer machine based on the Internet of Things according to claim 1, characterized in that, The statistical monitoring of the fault characteristic indicators using an exponentially weighted moving average control chart includes: For the fault characteristic indicators generated in real time, calculate their exponentially weighted moving average statistic: ； In the formula, This is the EWMA statistic for the current moment; The fault characteristic indicators at the current moment; This is the EWMA statistic from the previous time step; This is the smoothing coefficient.

5. The irrigation control method for an integrated water and fertilizer machine based on the Internet of Things according to claim 1, characterized in that, The preset control limits of the exponentially weighted moving average control chart include an upper control limit and a lower control limit, with values ​​set to 3 and -3, respectively.

6. The irrigation control method for an integrated water and fertilizer machine based on the Internet of Things according to claim 5, characterized in that, The method of distinguishing fault types based on the direction of deviation includes: If the current EWMA statistic exceeds the upper control limit, the fault type is determined to be a blockage fault; if the current EWMA statistic is below the lower control limit, the fault type is determined to be a leakage fault.

7. The irrigation control method for an integrated water and fertilizer machine based on the Internet of Things according to claim 1, characterized in that, The step of determining the fault zone based on the current valve state vector includes: Read and parse the current valve state vector Find the irrigation partitions corresponding to all elements with a state of "1" and use them as fault partitions.

8. An integrated water and fertilizer irrigation control system based on the Internet of Things, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement an IoT-based water and fertilizer integrated irrigation control method according to any one of claims 1-7.

Images