Crop growth environment intelligent monitoring and irrigation control system based on big data

Abstract

The invention discloses a crop growth environment intelligent monitoring and irrigation control system based on big data, and relates to the technical field of intelligent irrigation, and the system comprises a cloud scheduling server which is used for issuing an irrigation scheduling instruction with an expected execution timestamp to an edge controller; the edge controller is used for entering a first control state and calculating a pressure change rate when a local clock reaches an expected execution timestamp; if the irrigation valve is opened in a second control state in a continuous preset number of sampling periods, adding a first state flag bit to the acquired sensing data and packaging the sensing data into a data packet to be reported in an isolation period from the opening of the irrigation valve; the cloud scheduling server is used for receiving a data packet, and if a first state flag bit is analyzed from the data packet, the data packet containing sensing data is intercepted and forbidden to be input into the built-in compensation algorithm module; according to the invention, the problem that the local clock of the edge controller is easily influenced by the environment to generate drifting, and pipe network pressure resonance is caused is solved.

Description

Technical Field

[0001] This invention relates to the field of intelligent irrigation technology, specifically to an intelligent monitoring and irrigation control system for crop growth environment based on big data. Background Technology

[0002] In smart agricultural irrigation systems for large-scale, contiguous planting, cloud servers typically generate rotational irrigation scheduling instructions based on crop water requirements and distribute these instructions, including expected execution timestamps, to field edge controllers. Each edge controller executes valve switching operations based on its local clock, while simultaneously uploading real-time sensor data such as pipeline pressure and flow rate to the cloud-based compensation algorithm module. This module dynamically adjusts control parameters based on the real-time sensor data to maintain stable pipeline operation. These systems are widely used in automated irrigation control in citrus orchards, field crops, and facility agriculture, achieving zoned rotational irrigation and integrated water and fertilizer management through centralized cloud scheduling and distributed edge execution.

[0003] In existing irrigation control systems that rely on clock-based scheduling commands, the local clock of the edge controller is susceptible to drift due to environmental influences, which can lead to overlapping valve switching timings and consequently cause pipeline pressure resonance. Simultaneously, transient fluctuations generated by valve switching directly enter the compensation algorithm module, causing algorithm miscompensation and further exacerbating pressure fluctuations, ultimately resulting in pipeline fatigue damage. Furthermore, traditional clock synchronization methods have poor reliability and high modification costs in weak network environments, failing to address the pipeline resonance and runaway problems caused by timing misalignment and data distortion. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an intelligent monitoring and irrigation control system for crop growth environment based on big data.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a big data-based intelligent monitoring and irrigation control system for crop growth environment, comprising: A cloud-based scheduling server is used to send irrigation scheduling instructions with expected execution timestamps to the edge controller; An edge controller is used to enter a first control state when the local clock reaches the expected execution timestamp. In the first control state, the instantaneous pressure value of the corresponding irrigation pipeline is collected at a set sampling frequency and the pressure change rate is calculated. If the pressure change rate is less than the first convergence threshold within a consecutive preset number of sampling periods, the system transitions to the second control state and opens the irrigation valve in the second control state. During the isolation period starting from when the irrigation valve is opened, the collected sensor data is appended with the first status flag and packaged into a data packet for reporting. The cloud scheduling server is also used to receive the data packet. If the first status flag is parsed from the data packet, the data packet containing the sensor data is intercepted and prohibited from being input into the built-in compensation algorithm module, so that the compensation algorithm module maintains the output parameters of the previous cycle during the isolation period. The isolation period terminates when the pressure change rate is again less than the first convergence threshold for a consecutive preset number of sampling periods.

[0006] As a preferred embodiment of the present invention, the edge controller includes a sampling and filtering module, a state machine control module, and a protocol encapsulation module, and the cloud scheduling server includes a message gateway and routing module and the compensation algorithm module; The sampling and filtering module is used to collect the instantaneous pressure value and calculate the pressure change rate; The state machine control module is used to control the transition of the edge controller between the first control state and the second control state; The protocol encapsulation module is used to attach the first status flag bit to the collected sensor data and package it into the data packet; The message gateway and routing module are used to parse the data packets and perform corresponding interception operations.

[0007] As a preferred embodiment of the present invention, the sampling and filtering module has a built-in Kalman filter unit or a moving average algorithm unit. The sampling and filtering module uses the Kalman filter unit or the moving average algorithm unit to filter out interference noise in the acquired signal in order to obtain the smoothed instantaneous pressure value.

[0008] As a preferred embodiment of the present invention, the first convergence threshold is calculated and determined based on the rated working pressure of the pipeline network, the pipeline aging attenuation factor, and the pipeline network's years of operation.

[0009] As a preferred embodiment of the present invention, the state machine control module is further configured to: Monitor the duration for which the edge controller remains in the first control state; If the duration reaches the maximum timeout threshold and the condition that the pressure change rate is less than the first convergence threshold is not met within a preset number of consecutive sampling periods, then the system transitions to the third control state. In the third control state, the opening of the irrigation valve is terminated, and an alarm message carrying an error code is sent to the cloud scheduling server.

[0010] As a preferred embodiment of the present invention, the state machine control module synchronously activates the fourth control state when transitioning to the second control state; The second control state is used to maintain the output of the opening signal to the irrigation valve, and the fourth control state is used to trigger the protocol encapsulation module to perform the operation of attaching the first state flag bit. The second control state and the fourth control state run in parallel without blocking each other.

[0011] As a preferred embodiment of the present invention, the protocol encapsulation module is further configured to: After the isolation period ends, the collected sensor data will be appended with a second status flag. The message gateway and routing module is further configured to: if the second status flag is parsed from the data packet, then normally input the data packet containing the sensing data into the compensation algorithm module, so that the compensation algorithm module can restore the iteration of control parameters based on the received sensing data.

[0012] As a preferred embodiment of the present invention, the cloud scheduling server further includes a time-series reconstruction engine; The time-series reconstruction engine is used to further optimize the irrigation scheduling instructions based on the hydraulic characteristics of the pipeline network, for the following purposes: Extract the actual valve opening timestamp generated by the edge controller performing valve opening action under the same irrigation scheduling command, and obtain the pressure response timestamp reported by the end sensor of the pipeline network; The actual response time is obtained by calculating the difference between the pressure response timestamp and the actual valve opening timestamp, and the time residual is obtained by subtracting the preset hydraulic delay baseline value from the actual response time. The time residual represents the amount of drift of the edge controller's local clock relative to the reference clock.

[0013] As a preferred embodiment of the present invention, the timing reconstruction engine is further used for: Obtain the historical time residuals of adjacent first and second irrigation zones, and calculate the absolute value of the difference between the historical time residuals of the first and second irrigation zones; add a safety margin value to the absolute value of the difference to generate a safety buffer duration. The cloud-based scheduling server is also used to insert the safety buffer duration between the valve closing instruction corresponding to the first irrigation partition and the valve opening instruction corresponding to the second irrigation partition when generating irrigation scheduling instructions for the next scheduling cycle.

[0014] As a preferred embodiment of the present invention, the protocol encapsulation module is used to write the first status flag bit into a specific bit in the header flag field of a message queue telemetry transmission protocol message or a restricted application protocol message. The compensation algorithm module includes a penetration rate inversion unit and a proportional-integral-derivative control unit. The compensation algorithm module is used to receive sensor data that is not intercepted by the message gateway and routing module and update the output control parameters.

[0015] The beneficial effects of this invention are: 1. This invention sets the edge controller to enter a first control state when the local clock reaches the expected execution timestamp. In the first control state, the pressure change rate is calculated. If the pressure change rate is less than a first convergence threshold within a consecutive preset number of sampling periods, the controller transitions to a second control state and opens the irrigation valve in the second control state. This effectively uses the actual steady-state characteristics of the physical pipe network as a rigid pre-gating condition for opening the irrigation valve. This solves the problems of the edge controller's local clock being easily affected by environmental factors and drifting, leading to overlapping valve opening and closing timings and causing pipe network pressure resonance. It also addresses the issues of poor reliability and high modification costs of traditional clock synchronization methods in weak network environments.

[0016] 2. This invention, by setting an isolation period after the irrigation valve is opened, reports the collected sensor data with a first status flag bit attached, and by having the cloud scheduling server intercept and prohibit data packets containing the sensor data from entering the built-in compensation algorithm module if the first status flag bit is parsed, thus ensuring that the compensation algorithm module maintains the output parameters of the previous cycle, effectively isolates transient feedback data during the valve operation transition period. This solves the problem that transient fluctuation data generated by valve switching directly enters the compensation algorithm module, causing algorithm miscompensation and further aggravating pressure fluctuations, ultimately leading to pipeline fatigue damage. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0018] Figure 1 This is a schematic diagram of the overall structure of the intelligent monitoring and irrigation control system for crop growth environment of the present invention; Figure 2 This is a schematic diagram illustrating the timing of the coordinated linkage between the pressure change rate and the first state flag. Detailed Implementation

[0019] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0020] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0021] like Figure 1 As shown, this embodiment takes the scenario of zoned rotational irrigation of a large area of ​​contiguous farmland as an example to explain in detail the collaborative working process of instruction issuance, status verification, data marking and algorithm isolation between the cloud scheduling server and the edge controller.

[0022] The cloud-based scheduling server generates irrigation scheduling instructions for the next scheduling cycle based on crop water requirement models, weather forecasts, and soil moisture monitoring data. These instructions are sent to the corresponding edge controllers in the form of structured messages. The message payload carries the device identifier of the target control node and the expected execution timestamp. After receiving the instructions, the edge controller stores them in its local non-volatile storage area and continuously compares the local clock with the expected execution timestamp. When the local clock reaches the expected execution timestamp, the edge controller does not immediately drive the irrigation valves but switches its internal program state to the first control state.

[0023] In the first control state, the edge controller drives the analog-to-digital converter interface at a set sampling frequency to collect the instantaneous pressure value at the front end of the corresponding irrigation branch pipe. The collected pressure signal is then processed by basic noise reduction to obtain the pressure change rate at the current moment. This pressure change rate is used to quantify the degree of attenuation of hydraulic fluctuations within the pipe network. The larger the value, the more likely the pressure standing wave caused by the upstream valve action has not yet dissipated; a value close to zero indicates that the hydraulic system of the pipeline has returned to stability.

[0024] The edge controller maintains a state counter in memory. In each sampling period, it compares the absolute value of the pressure change rate with a first convergence threshold. If the absolute value is less than the first convergence threshold, the counter is incremented; if it is greater than or equal to the threshold, the counter is cleared.

[0025] Only when the cumulative value of the counter reaches a preset number, that is, when the pressure change rate is less than the first convergence threshold for a consecutive preset number of sampling periods, the edge controller determines that the physical state of the pipeline network has met the conditions for safe switching. The state machine then transitions from the first control state to the second control state and outputs a level signal to drive the irrigation valve to open.

[0026] When the irrigation valve is opened, transient hydraulic disturbances will inevitably occur inside the pipeline due to fluid inertia. This will cause the sensor data collected at this time to contain non-steady-state fluctuation components. In order to prevent these fluctuation components from interfering with cloud decision-making, the edge controller enters the isolation period.

[0027] During this isolation period, when encapsulating and reporting data packets, the edge controller appends a first status flag to the protocol header or status field and sends the data packets with the appended flag to the cloud scheduling server at a preset period. This flag serves as a logical identifier, indicating to the receiving end that the current data packet is in a physical transition period.

[0028] After the network access layer of the cloud scheduling server receives the data packet, it executes the protocol parsing and routing distribution logic. When the parser identifies the first status flag bit through bitmask matching or field extraction, it triggers the data bypass interception strategy. The cloud scheduling server redirects the data packet to the log storage channel for archiving and cuts off its input path to the built-in compensation algorithm module.

[0029] During the isolation period, the compensation algorithm module automatically switches to cache read mode because the input channel is temporarily isolated. It directly calls the output parameters that have been persisted in the previous steady-state cycle to maintain the static issuance of control commands. This mechanism ensures that during the turbulent period of pipeline physical switching, the cloud-based decision-making logic is not contaminated by transient distorted data, and avoids the algorithm outputting incorrect compensation adjustment amounts due to misjudgment of pressure fluctuations.

[0030] The termination condition during the isolation period is determined by the pressure change rate status continuously monitored by the edge controller. During the isolation period, the edge controller continues to collect instantaneous pressure values ​​and calculate the pressure change rate. When the water hammer fluctuations inside the pipeline are naturally attenuated by the damping of the pipe wall, so that the pressure change rate again meets the condition that it is less than the first convergence threshold for a consecutive preset number of sampling periods, the system determines that the hydraulic system of the pipeline has returned to steady state, and the isolation period ends.

[0031] The edge controller then stops attaching the first state flag to subsequent data packets. After parsing a steady-state data packet that does not carry the flag, the cloud scheduling server automatically removes the routing interception and re-injects the data stream into the compensation algorithm module, allowing the system to seamlessly resume closed-loop dynamic adjustment based on real-time sensor data.

[0032] Through the aforementioned control links, this system forcibly binds valve opening authorization with the physical steady state of the pipeline network at the edge, eliminating the overlap of action timing caused by local clock drift; at the cloud side, the first state flag bit is used to classify and route transient data and isolate algorithm input, cutting off the positive feedback transmission path of hydraulic fluctuations to the control algorithm, effectively suppressing the risk of pipeline network resonance runaway caused by the coupling of clock asynchrony and hydraulic delay.

[0033] Furthermore, on the edge controller side, the sampling and filtering module establishes an electrical connection with the pressure transmitter at the front end of the branch pipe via an analog-to-digital converter interface. The continuous analog signal output by the pressure transmitter is first subjected to noise suppression by a digital filtering unit. This filtering unit can be configured with a moving average algorithm routine or a Kalman filter routine. Taking the moving average algorithm as an example, the module maintains a pre-set circular sampling array in memory. After each sampling, the latest pressure value is written to the end of the array and overwrites the oldest historical data. Then, the arithmetic mean of the effective elements in the array is calculated, and the smoothed instantaneous pressure value is output.

[0034] After obtaining the smoothed pressure value, the sampling and filtering module needs to calculate the pressure change rate to quantify the degree of transient fluctuation in the pipeline network. Considering the computational overhead of the edge microprocessor in continuous differentiation operations and the numerical stability of discrete signal processing, this embodiment uses a discrete central difference model for slope estimation. The formula for calculating the pressure change rate is as follows: ; Where D is the rate of change of pressure, P i P represents the instantaneous pressure value during the current sampling period. i−2 The instantaneous pressure values ​​are the values ​​from the first two sampling periods, Δt is the sampling interval period, and ε is the minimum zero-prevention bias constant (usually taken as 10). −6 (Order level), used to prevent zero overflow of the denominator in low-level floating-point division operations.

[0035] This differential model uses data spanning the current sampling point to approximate the derivative, and has low-pass filtering truncation characteristics, which can effectively suppress high-frequency noise caused by air bubbles carried by the water flow.

[0036] In this embodiment, the sampling frequency is preferably 50Hz, corresponding to Δt of 0.02 seconds. The propagation speed of water hammer pressure waves in drip irrigation networks is approximately 1200m / s. For typical irrigation branch pipes with a laying length of less than 500 meters, the main frequency component of the network pressure standing wave is usually lower than 25Hz. A sampling frequency of 50Hz is sufficient to completely capture the spectrum of the standing wave signal without aliasing, while avoiding excessive occupation of the edge controller's memory and computing resources by excessively high sampling rates.

[0037] The state machine control module receives the pressure change rate and drives the transition of the control state accordingly. The module internally maintains a state word register and a continuous stable counter.

[0038] When the local clock reaches the expected execution timestamp, the state machine control module sets the state word to the first control state. In this state, the module reads the pressure change rate at fixed intervals and compares its absolute value with the first convergence threshold.

[0039] If the absolute value is less than the first convergence threshold, the continuous stable counter performs an accumulation operation; if the absolute value is greater than or equal to the threshold, the counter performs a zeroing operation.

[0040] Only when the cumulative value of the counter reaches a preset number of consecutive times, the module determines that the hydraulic fluctuations in the pipeline have attenuated to a safe range, transitions the status word to the second control state, and outputs a level signal to drive the irrigation valve to open.

[0041] The protocol encapsulation module and the state machine control module are linked by a hardware interrupt signal. When the state machine transitions to the second control state, the protocol encapsulation module synchronously activates the data marking program. During the subsequent isolation period, the module calls the preset data frame template, writes the collected sensor data into the payload area, and performs a setting operation on a specific bit in the extended field of the protocol header to generate the first state flag bit. The packaged data frame is sent to the cloud scheduling server via the wireless communication link according to the preset reporting cycle.

[0042] On the cloud scheduling server side, the message gateway and routing module act as the data access hub, performing protocol header parsing on the uplink data frame. The module has built-in bitmask matching logic, and when it identifies that the data frame header carries the first status flag, it determines that the current data is in the isolation period.

[0043] At this point, the module triggers a bypass routing strategy, redirecting the data frame to a read-only log storage channel for offline auditing and historical data tracing, while simultaneously cutting off its input path to the compensation algorithm module. If the data frame does not carry this flag, it is determined to be steady-state data, and the module routes it normally to the consumption queue of the compensation algorithm module.

[0044] During the interruption of steady-state data input, the compensation algorithm module automatically activates the parameter freezing mechanism. The internal state cache unit of the module directly retrieves the control parameters that have been persisted in the previous steady-state cycle and uses them as the output reference for the current scheduling cycle. This mechanism ensures that the cloud decision logic remains static during the turbulent period of pipeline physical switching, avoiding the triggering of erroneous compensation adjustments by transient distortion data. When the message gateway and routing module resume input of steady-state data without the flag bit, the compensation algorithm module automatically unfreezes and re-enters closed-loop iterative calculation based on real-time data.

[0045] Furthermore, in the actual operating environment of farmland irrigation networks, the analog signal output by the pressure transmitter is often superimposed with high-frequency electrical interference and local fluid oscillations caused by the operation of the variable frequency pump motor, grid voltage fluctuations, and water flow through valve throttling devices. If the unprocessed raw sampled values ​​are directly used for discrete difference calculations, the high-frequency noise will be significantly amplified during the differentiation process, resulting in instantaneous spikes in the pressure change rate calculation results. Such spikes can easily cause the state machine control module to misjudge that the network is in a state of severe hydraulic fluctuations, thereby hindering the normal transition of the edge controller from the first control state to the second control state.

[0046] To address this, the sampling and filtering module utilizes the characteristic that the actual pressure change in the pipeline network is a low-frequency, slowly varying signal while the interference noise is a high-frequency, random signal. It uses a discrete-time filtering algorithm to effectively implement low-pass filtering frequency selection, thereby controlling the signal distortion error within the accuracy range allowed by the state machine threshold comparison.

[0047] When the sampling and filtering module calls the moving average algorithm unit, it maintains a fixed-length circular data buffer structure. This structure stores the continuously acquired raw pressure values ​​sequentially according to the sampling time. When a new sample value arrives, it is written to the end of the queue, overwriting the oldest stored data, thus ensuring that the buffer always contains valid sampled data from the most recent continuous period. At the end of each sampling period corresponding to the set sampling frequency, the algorithm unit performs an accumulation and summation operation on all currently stored valid sample values ​​in the buffer structure, and divides the sum by the number of sampling points involved in the operation. The quotient is the smoothed instantaneous pressure value for the current period. The number of sampling points involved in the averaging can be configured according to the on-site noise intensity and system response speed requirements, typically between five and ten sampling points, to achieve an engineering balance between high-frequency noise suppression and signal response hysteresis.

[0048] When the electromagnetic environment is complex or the noise statistics change with the equipment's operating status, the sampling and filtering module can switch to the Kalman filter unit. This unit recursively estimates the true pressure value based on a state-space model, internally maintaining the optimal estimate and error covariance variable for the current period. Within each sampling period, the unit first makes a priori prediction of the pressure state based on a preset process noise variance. Then, it reads the actual collected raw pressure value as the measurement input, calculates the dynamic Kalman gain coefficient by combining it with a preset measurement noise variance, and outputs the corrected and smoothed instantaneous pressure value by multiplying the deviation between the measured value and the priori prediction value by this gain coefficient and adding it to the priori estimate. This recursive process only performs scalar floating-point multiplication and addition operations in each period, eliminating the need to maintain a long historical data queue, and can adaptively adjust the filtering intensity according to the relative proportion of prediction error and measurement noise.

[0049] When the pipeline pressure is stable, the filter unit enhances noise suppression; when the pressure undergoes a real step change, the unit automatically reduces the filter damping to maintain rapid tracking of physical changes.

[0050] The smoothed instantaneous pressure value output by the sampling and filtering module is directly used as the input variable for the pressure change rate calculation formula. After filtering, the high-frequency interference components in the original signal are effectively attenuated, and the pressure change rate curve calculated by discrete difference tends to be stable, which can truly reflect the pressure change trend caused by hydraulic fluctuations inside the pipeline network. This processing mechanism eliminates the instantaneous over-limit phenomenon of pressure change rate caused by sensor noise spikes, so that the state machine control module's comparison and judgment of the first convergence threshold is based on reliable physical signals, ensuring the accuracy of control state transitions and the operational stability of the system in complex field electromagnetic environments.

[0051] Furthermore, in the actual operation of agricultural irrigation pipe networks, if a fixed constant is used as the criterion for determining the steady state of the pipe network, it will be difficult to adapt to the dynamic changes in the physical properties of the pipe material with the service cycle. The pipe walls of newly built pipe networks are smooth and have low damping, and the pressure standing wave decays slowly, requiring a stricter threshold to ensure sufficient waiting time. As the years of operation increase, scaling on the inner wall of the pipe, the propagation of microcracks, and the decay of viscoelasticity will change the damping characteristics of the pipe network, and the speed at which pressure fluctuations subside will change accordingly.

[0052] If the initial fixed threshold is maintained, the system is prone to misjudging transitions before hydraulic fluctuations have completely dissipated, or waiting excessively after the system has stabilized, both of which can lead to timing overlaps or decreased scheduling efficiency. Therefore, this embodiment adopts a dynamic calculation mechanism based on the rated working pressure of the pipeline network, the pipeline aging attenuation factor, and the pipeline network's operational years, so that the threshold is adaptively adjusted throughout the entire life cycle of the pipeline network.

[0053] The rated working pressure of the pipeline network is determined by the pipeline network design pressure rating and pump selection parameters, serving as the pressure benchmark for normal system operation. The pipeline aging attenuation factor is an empirical coefficient characterizing the degree to which the physical properties of the pipe material deteriorate over time affect the rate of pressure fluctuation attenuation. Its dimension is per year, and the specific value is determined based on the pipe material type through accelerated aging tests and on-site hydraulic tests. The pipeline network's operational lifespan is calculated by subtracting the current system timestamp from the pre-stored operating timestamp.

[0054] Before each irrigation scheduling instruction is generated, the cloud scheduling server performs an exact match query using the unique identifier of the target device as an index, extracts the above parameters and completes the year calculation. The calculation task can then be sent to the edge controller or calculated directly in the cloud and synchronized with the instruction.

[0055] The first convergence threshold, serving as the core criterion for state transition, is dynamically calculated by the system based on the physical characteristics of the pipeline network. Its calculation formula is as follows: ; Wherein, ΔP safe P is the first convergence threshold. nom The rated working pressure of the pipeline network is α, which is the pipeline aging attenuation coefficient. Preferably, the value of α is in the range of 0.015 to 0.06. In this embodiment, the optimal value for the PVC-U main pipeline is 0.03. T age The actual operational lifespan of the pipeline network is denoted by k, which is the hydraulic fluctuation tolerance coefficient. Preferably, the value of k ranges from 0.01 to 0.05. In this embodiment, the optimal value is 0.02. If k < 0.01, the safety convergence band is too narrow, and the state machine control module will misjudge normal fluid white noise as resonance. If k > 0.05, the safety convergence band is too wide, and the system may incorrectly determine that the pipeline network has converged before the actual destructive hydraulic fluctuations have dissipated, thereby prematurely releasing the state machine lock and causing the anti-resonance gating mechanism to fail.

[0056] Taking a drip irrigation branch pipe as an example: the rated working pressure P of the pipeline network nom =0.4MPa, pipe material is PVC-U, pipe aging degradation factor α=0.03 / year, pipeline operation life T age =5 years, hydraulic fluctuation tolerance coefficient k=0.02.

[0057] Substitute into the formula: ΔP safe =0.02×0.4×(1−0.03×5)=0.0068MPa / s, that is, when the pressure change rate decreases to below 0.0068MPa / s, it is determined that the hydraulic system of the pipeline has returned to steady state and the irrigation valve can be safely opened.

[0058] Newly built pipelines have smooth inner walls and low damping, requiring strict thresholds to prevent residual fluctuations from accumulating. As the years of operation increase, scaling on the pipe walls and material fatigue increase the damping of the pipeline, accelerating the attenuation of fluctuations. The threshold can be appropriately relaxed to balance irrigation scheduling efficiency. The parameter configuration unit of the cloud scheduling server periodically reads the pipeline attributes of each zone, calculates the corresponding threshold, and then synchronizes it to the non-volatile storage area of ​​the edge controller through control commands to achieve adaptive matching of the system's criteria.

[0059] Furthermore, in the actual operation of agricultural irrigation networks, after the edge controller enters the first control state, it needs to wait for the network pressure fluctuations to naturally decay. If abnormal conditions such as pressure transmitter probe blockage, intermittent interference introduced into the signal line, or physical rupture of the network cause continuous distortion of the collected pressure signal, the calculated pressure change rate will fail to meet the transition condition of being less than the first convergence threshold within a continuous preset period for a long time. This embodiment introduces continuous duration monitoring and maximum timeout threshold determination logic to map physical or sensor abnormalities into manageable timeout events, and executes safety interlocking and alarm reporting.

[0060] The state machine control module has an internal timing unit that accumulates the duration of the edge controller in the first control state. The timing unit accumulates based on the system clock or sampling period. Its count value is compared with the preset maximum timeout threshold in each control cycle. The maximum timeout threshold is not a fixed empirical value, but is calculated based on the hydraulic settlement characteristics of the pipeline network, the drought resistance limit of the crop growth period, and the on-site operation and maintenance response time.

[0061] During the state transition judgment process, the state machine control module performs composite condition detection. If the detected duration reaches the maximum timeout threshold, and the continuous convergence condition of the pressure change rate is still not met after confirmation by the logic state register, the module determines that the current operating condition has exceeded the normal attenuation range and triggers a forced state transition. The control logic switches the edge controller from the first control state to the third control state by overwriting the underlying state word.

[0062] After entering the third control state, the edge controller synchronously performs physical blocking and data reporting actions. At the hardware control level, the drive circuit immediately cuts off the control level to the irrigation valve, sets the general input / output pins to the safe default state, and terminates the subsequent action of opening the irrigation valve to prevent forced water injection from aggravating physical damage in the case of abnormal pressure or pipeline leakage. At the data communication level, the protocol encapsulation module generates alarm messages according to the preset format.

[0063] The payload of the alarm message includes the device's unique identifier, the system timestamp that triggered the timeout, the error code that indicates the type of anomaly, and a snapshot of the pressure change rate data in the current ring buffer.

[0064] Furthermore, during the state transition process, when the edge controller meets the conditions for transitioning from the first control state to the second control state, the state machine control module synchronously sets the activation flag of the fourth control state within the same processing cycle.

[0065] The second control state is responsible for maintaining the open signal output to the irrigation valve. Its execution unit is typically mapped to a high-priority hardware control task or a timer interrupt service routine. This task directly operates the microcontroller's general-purpose input / output registers to continuously output a stable level signal to the valve drive circuit, ensuring that water is injected into the pipeline as needed throughout the irrigation cycle.

[0066] Meanwhile, the fourth control state is used to trigger the protocol encapsulation module to perform the operation of attaching the first state flag. Its execution body runs in a separate software task or a background polling thread. After the state machine control module activates the fourth control state, it writes a state activation flag to the shared memory area.

[0067] During the periodic data framing process, the protocol encapsulation module reads the activation flag and immediately enters the flag working mode, writing the first status flag bit into the protocol header field of all subsequent sensor data frames.

[0068] Because the hardware level output of the second control state and the protocol data flag of the fourth control state belong to different execution contexts, they are completely isolated in terms of memory access and task scheduling. Even if the fourth control state experiences execution delays due to network retransmission, protocol stack processing time, or communication congestion, the blocking scope is limited to the data reporting queue and will not suspend or slow down the continuous output of the irrigation valve opening signal from the second control state.

[0069] Furthermore, after the isolation period ends, the protocol encapsulation module synchronously switches the data marking strategy. During the subsequent data framing process, the protocol encapsulation module stops writing the first status flag bit in the data packet protocol header and instead appends the second status flag bit.

[0070] In actual communication protocol implementations, the first status flag and the second status flag can reuse the same status field in the protocol header, and can be distinguished by different logic levels or enumeration values: The first state flag indicates that the data is in a transient transition period, while the second state flag indicates that the data has been restored to the steady-state reference and does not contain the non-steady-state fluctuation components introduced by valve switching.

[0071] The protocol encapsulation module encapsulates the sensor data with the second status flag bit attached into data packets according to a preset reporting period, and sends them to the cloud scheduling server via the wireless communication link.

[0072] When the message gateway and routing module of the cloud scheduling server receive and parse uplink data packets, they perform routing distribution based on the status field in the protocol header. When the parser identifies that the data packet carries a second status flag, it determines that the corresponding edge controller has emerged from the period of physical instability, and the message gateway and routing module immediately lifts the previous data interception policy targeting that node. The data stream that was originally redirected to the offline log storage channel is switched back to the standard message queue topic and pushed normally to the data consumption port of the compensation algorithm module.

[0073] Upon detecting continuous reception of steady-state sensor data without interception flags, the compensation algorithm module automatically exits the parameter freeze mode based on historical cache. The permeability inversion unit and proportional-integral-derivative control unit within the module restart their computational logic, inputting the latest reported pipeline pressure, flow rate, and soil moisture data into the mathematical control model. This enables dynamic iterative calculation of control parameters, with the iteratively generated latest control parameters replacing the previous cycle's baseline values ​​maintained during the isolation period, and then being sent to the irrigation scheduling command generation link.

[0074] Through the coordinated logic of flag switching and route release, the system rebuilds the complete data closed loop from edge data acquisition to cloud algorithm decision-making as soon as the physical state of the pipeline network returns to stability. This achieves a smooth transition from transient open-loop protection to steady-state closed-loop regulation, ensuring that irrigation control parameters can respond promptly to dynamic changes in soil permeability and crop water demand.

[0075] like Figure 2 As shown in the upper part, from the moment the irrigation valve opens (i.e., the start of the horizontal time axis at 0 seconds), an initial water hammer standing wave is generated inside the pipeline network. Within the initial 0 to 1.5 seconds of valve opening, the pressure change rate calculated by the sampling and filtering module exhibits significant alternating positive and negative oscillations. Its initial physical amplitude peak reaches approximately 0.025 MPa / s, far exceeding the system's set first convergence threshold of ±0.005 MPa / s. During this turbulent phase, the physical state is not up to standard, and the protocol encapsulation module continuously adds a first state flag bit (such as...) to the data packet. Figure 2 As shown in the lower half, the thick solid line representing the first state flag remains at a high level (1).

[0076] Combination Figure 2 The continuous judgment window with gray shading shows that when the water hammer fluctuation enters the final stage of decay (about 1.5 seconds on the horizontal axis in the figure), although the fluctuation curve of the pressure change rate has completely fallen into the first convergence threshold safety zone of ±0.005MPa / s, the system still maintains isolation. The state machine must wait for the pressure change rate to not exceed the limit within a consecutive preset number of sampling periods.

[0077] like Figure 2 As shown in the lower half, at the precise moment when the aforementioned continuous verification period of approximately 0.6 seconds completely ends, the underlying logic level produces a strict and synchronous mutually exclusive transition: the first state flag bit (thick solid line) representing "maintaining the output parameters of the previous cycle" is instantly removed and drops to low level 0; at the same time, the second state flag bit (dashed line) representing "resuming control parameter iteration" is synchronously attached and transitions to high level 1.

[0078] Furthermore, in the data extraction stage, the time-series reconstruction engine reads two key time nodes under the same irrigation scheduling command from the system operation log and event association library.

[0079] One is the actual valve opening timestamp recorded when the edge controller performs the valve opening action. This timestamp is generated by the edge controller based on its internal clock and uploaded to the cloud along with the execution status message. Secondly, there is the response timestamp of the pressure rise detected by the sensors at the end of the pipeline network. This timestamp is marked in real time by the cloud scheduling server based on the server's own reference clock when receiving and parsing the sensor data.

[0080] Since the two timestamps rely on different clock sources, the difference not only includes the physical time of water flow transmission, but also the error introduced by clock asynchrony.

[0081] When the time-series reconstruction engine performs time residual calculation, it adopts a step-by-step linear operation model, first calculating the actual response time T. a The expression is: ; Where T1 represents the actual valve opening timestamp, and T2 represents the pressure response timestamp. This step calculates the total surface time at the digital system level. Due to the coordinate system deviation between T1 and T2, T... a Clock drift error is inevitably mixed in.

[0082] Subsequently, the engine performs an error stripping operation to obtain the time residual E. t The expression is: ; Among them, T b Characterizing the preset hydraulic delay baseline value, the final output time residual E is obtained by subtracting the known physical conduction constant from the actual response time. t That is, it accurately characterizes the amount of drift of the edge controller's local clock relative to the reference clock, when E t A positive value indicates that the local clock record is too early; a negative value indicates that the local clock record is too late.

[0083] Preset hydraulic delay baseline value T b The data originates from on-site clean water calibration tests after initial system installation or pipeline maintenance. During the calibration phase, after ensuring that the edge controller and the cloud complete initial clock synchronization, the system continuously records the time intervals from valve opening to end pressure establishment multiple times. After removing outliers caused by instantaneous pressure fluctuations, the arithmetic mean is taken as the baseline value for that pipe section.

[0084] Considering that network pressure fluctuations, sensor response delays, or instantaneous communication jitter may introduce random noise during a single irrigation event, the timing reconstruction engine incorporates a sliding window smoothing mechanism.

[0085] The engine maintains a buffer of time residual sequences of a preset length. Each time a new residual is calculated, it is pushed into the window, and invalid extreme values ​​caused by communication packet loss or sensor failure are removed. The engine performs an arithmetic mean operation on the remaining valid residuals within the window, and the resulting average value is used as a stable estimate of the current cumulative clock drift of the edge controller.

[0086] Furthermore, in the actual deployment of agricultural IoT, the time residuals calculated from a single irrigation event are easily affected by instantaneous communication jitter or sensor noise, and directly using them for scheduling decisions may lead to distortion in the calculation of buffer duration. To address this, when the time series reconstruction engine obtains the historical time residuals of the first and second irrigation zones, it does not directly read the single measurement value, but instead retrieves the effective time residual sequence of each zone within a preset sliding time window from the cloud time series database.

[0087] The system first executes outlier removal logic, filtering outliers that deviate from the historical residual mean of the partition by more than three standard deviations. Then, it performs an arithmetic mean calculation on the remaining valid residuals to obtain the smoothed historical time residuals for the first and second irrigation partitions. This data preprocessing step effectively suppresses random noise, making the extracted drift more closely match the long-term trends of crystal oscillator temperature drift and aging accumulation, providing a stable benchmark input for subsequent time-series isolation.

[0088] After obtaining the smoothed historical time residuals, the time series reconstruction engine performs the core calculation of the anti-overlap duration. To avoid introducing highly complex real-time fluid dynamics simulations, this scheme uses linear operations to construct an equivalent calculation model, the specific expression of which is: ; Among them, T buf Characterizing the duration of the safety buffer, Δt A With Δt B T represents the historical time residuals of the first and second irrigation zones, respectively. margin This represents the safety margin value.

[0089] The edge controllers of adjacent zones operate in independent local clock domains. When the worst-case scenario occurs (i.e., one zone's clock is faster than the other's), the theoretical maximum risk time for overlapping actions between the two in the physical network is equal to the relative difference in their drift relative to the reference clock. By calculating the absolute value of the difference, the system accurately quantifies the extreme values ​​of the relative clock offset between the edge nodes of the two zones, providing a quantitative basis for reserving isolation space on the scheduling timeline.

[0090] The safety margin value T superimposed in the formula margin It is not a fixed empirical constant; its setting is based on the physical damping characteristics of the directly related pipe network and the mechanical response delay of the actuator. There is a mechanical travel time from receiving the shut-off command to the valve core fully closing, and the inertia of the water flow in the pipe network will generate a weak pressure aftershock after the valve closes. The safety margin value must cover the mechanical action delay, the natural decay period of the hydraulic aftershock, and the uncontrollable jitter tolerance of the wide-area communication network. In actual engineering configurations, T... margin The value is usually matched based on the pipe type and the length of the pipeline.

[0091] After calculating and generating the safety buffer duration, the cloud scheduling server performs a timeline reconstruction operation when arranging irrigation scheduling instructions for the next scheduling cycle. The server identifies the first and second irrigation zones that are physically adjacent in the pipeline topology, extracts the rotation interval duration between the two in the original scheduling plan, and forcibly inserts the calculated safety buffer duration between the valve closing instruction corresponding to the first irrigation zone and the valve opening instruction corresponding to the second irrigation zone.

[0092] Furthermore, when the message arrives at the cloud scheduling server, the message gateway and routing module parse the header flags. If the first status flag is not detected, it indicates that the sensor data is in a steady state, and the module routes the data packet to the input port of the compensation algorithm module.

[0093] The compensation algorithm module internally includes a permeability inversion unit and a proportional-integral-derivative (PID) control unit, which work together sequentially to perform iterative calculations of control parameters. The permeability inversion unit first receives uninterrupted steady-state sensor data, including pipeline pressure values ​​and flow feedback values.

[0094] Based on a pre-defined empirical model of soil moisture transport or historical hydraulic calibration data, the unit maps steady-state pressure and flow characteristics to an estimate of the current soil permeability. This inversion process can be implemented using multidimensional lookup table interpolation or a simplified hydrological balance algorithm to avoid the computational burden of solving complex fluid dynamics equations and ensure real-time response.

[0095] Subsequently, the proportional-integral-derivative (PID) control unit receives the estimated actual permeability value and performs closed-loop calculations based on the deviation signal between the target soil moisture threshold and the current actual moisture. The proportional, integral, and derivative components within the control unit perform weighted calculations on the instantaneous deviation value, historical cumulative value, and trend of change, synthesizing the final output control parameters. These parameters typically represent the expected opening duration or duty cycle of the irrigation valves in the next scheduling cycle, and are uniformly distributed from the cloud-based scheduling server to the edge for execution.

[0096] During the isolation period, if the message gateway and routing module identify that a message carries the first status flag, a data interception strategy is executed. At this time, the input channel of the compensation algorithm module is temporarily blocked, the penetration rate inversion unit stops receiving new data, and the proportional-integral-derivative control unit automatically switches to parameter holding mode, directly calling the control parameters calculated and cached in the previous steady-state cycle as the current output.

[0097] When the isolation period ends and the cloud resumes inputting steady-state data without the flag bit, the proportional-integral-derivative control unit releases the hold mode and restarts the deviation calculation and parameter iteration based on the latest sensor data.

[0098] By rapidly identifying protocol layer flags and dynamically freezing and restoring control loops at the algorithm layer, the system ensures real-time irrigation scheduling while effectively isolating the interference of valve switching transients on the core control algorithm, thus guaranteeing the continuity of control parameter output and the physical stability of pipeline operation.

[0099] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A big data-based intelligent monitoring and irrigation control system for crop growth environment, characterized in that: include: A cloud-based scheduling server is used to send irrigation scheduling instructions with expected execution timestamps to the edge controller; An edge controller is used to enter a first control state when the local clock reaches the expected execution timestamp. In the first control state, the instantaneous pressure value of the corresponding irrigation pipeline is collected at a set sampling frequency and the pressure change rate is calculated. If the pressure change rate is less than the first convergence threshold within a consecutive preset number of sampling periods, the system transitions to the second control state and opens the irrigation valve in the second control state. During the isolation period starting from when the irrigation valve is opened, the collected sensor data is appended with the first status flag and packaged into a data packet for reporting. The cloud scheduling server is also used to receive the data packet. If the first status flag is parsed from the data packet, the data packet containing the sensor data is intercepted and prohibited from being input into the built-in compensation algorithm module, so that the compensation algorithm module maintains the output parameters of the previous cycle during the isolation period. The isolation period terminates when the pressure change rate is again less than the first convergence threshold for a consecutive preset number of sampling periods.

2. The intelligent monitoring and irrigation control system for crop growth environment based on big data as described in claim 1, characterized in that, The edge controller includes a sampling and filtering module, a state machine control module, and a protocol encapsulation module; the cloud scheduling server includes a message gateway and routing module, and the compensation algorithm module. The sampling and filtering module is used to collect the instantaneous pressure value and calculate the pressure change rate; The state machine control module is used to control the transition of the edge controller between the first control state and the second control state; The protocol encapsulation module is used to attach the first status flag bit to the collected sensor data and package it into the data packet; The message gateway and routing module are used to parse the data packets and perform corresponding interception operations.

3. The intelligent monitoring and irrigation control system for crop growth environment based on big data as described in claim 2, characterized in that, The sampling and filtering module has a built-in Kalman filter unit or a moving average algorithm unit. The sampling and filtering module uses the Kalman filter unit or the moving average algorithm unit to filter out interference noise in the acquired signal in order to obtain the smoothed instantaneous pressure value.

4. The intelligent monitoring and irrigation control system for crop growth environment based on big data as described in claim 1, characterized in that, The first convergence threshold is calculated and determined based on the rated working pressure of the pipeline network, the pipeline aging attenuation factor, and the pipeline network's years of operation.

5. The intelligent monitoring and irrigation control system for crop growth environment based on big data as described in claim 2, characterized in that, The state machine control module is also used for: Monitor the duration for which the edge controller remains in the first control state; If the duration reaches the maximum timeout threshold and the condition that the pressure change rate is less than the first convergence threshold is not met within a preset number of consecutive sampling periods, then the system transitions to the third control state. In the third control state, the opening of the irrigation valve is terminated, and an alarm message carrying an error code is sent to the cloud scheduling server.

6. The intelligent monitoring and irrigation control system for crop growth environment based on big data as described in claim 2, characterized in that, When the state machine control module transitions to the second control state, it simultaneously activates the fourth control state; The second control state is used to maintain the output of the opening signal to the irrigation valve, and the fourth control state is used to trigger the protocol encapsulation module to perform the operation of attaching the first state flag bit. The second control state and the fourth control state run in parallel without blocking each other.

7. The intelligent monitoring and irrigation control system for crop growth environment based on big data as described in claim 2, characterized in that, The protocol encapsulation module is also used for: After the isolation period ends, the collected sensor data will be appended with a second status flag. The message gateway and routing module is further configured to: if the second status flag is parsed from the data packet, then normally input the data packet containing the sensing data into the compensation algorithm module, so that the compensation algorithm module can restore the iteration of control parameters based on the received sensing data.

8. The intelligent monitoring and irrigation control system for crop growth environment based on big data as described in claim 1, characterized in that, The cloud-based scheduling server also includes a time-series reconstruction engine; The time-series reconstruction engine is used to further optimize the irrigation scheduling instructions based on the hydraulic characteristics of the pipeline network, for the following purposes: Extract the actual valve opening timestamp generated by the edge controller performing valve opening action under the same irrigation scheduling command, and obtain the pressure response timestamp reported by the end sensor of the pipeline network; The actual response time is obtained by calculating the difference between the pressure response timestamp and the actual valve opening timestamp, and the time residual is obtained by subtracting the preset hydraulic delay baseline value from the actual response time. The time residual represents the amount of drift of the edge controller's local clock relative to the reference clock.

9. The intelligent monitoring and irrigation control system for crop growth environment based on big data as described in claim 8, characterized in that, The time-series reconstruction engine is also used for: Obtain the historical time residuals of adjacent first and second irrigation zones, and calculate the absolute value of the difference between the historical time residuals of the first and second irrigation zones; add a safety margin value to the absolute value of the difference to generate a safety buffer duration. The cloud-based scheduling server is also used to insert the safety buffer time between the valve closing instruction corresponding to the first irrigation partition and the valve opening instruction corresponding to the second irrigation partition when generating irrigation scheduling instructions for the next scheduling cycle.

10. The intelligent monitoring and irrigation control system for crop growth environment based on big data according to claim 2, characterized in that, The protocol encapsulation module is used to write the first status flag bit into a specific bit in the header flag field of a message queue telemetry transmission protocol message or a restricted application protocol message. The compensation algorithm module includes a penetration rate inversion unit and a proportional-integral-derivative control unit. The compensation algorithm module is used to receive sensor data that is not intercepted by the message gateway and routing module and update the output control parameters.

Images