An internet of things irrigation multi-parameter intelligent regulation and control system and method

Abstract

The application discloses a kind of Internet of Things irrigation multi-parameter intelligent control system and method, it is related to intelligent control technical field, including dynamic performance module, parameter identification module and intelligent control module;Dynamic performance module obtains the actual response time of electromagnetic valve and variable frequency pump flow-pressure characteristic coefficient in real time;Parameter identification module is based on pipe network pressure variation, on-line identification system equivalent time constant and pure lag time;Intelligent control module then according to the above parameters, adaptive timing compensation, feedforward calculation and feedback parameter adjustment are carried out to control instruction, generate final control signal, realize the adaptive matching of control system to actuator performance and the dynamic characteristics of controlled object, significantly improve the control precision, stability and irrigation efficiency of system in long-term operation.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, and in particular to an Internet of Things (IoT) irrigation multi-parameter intelligent control system and method. Background Technology

[0002] With the development of agricultural automation and informatization, automatic irrigation technology based on environmental sensing has become an important means to achieve water-saving agriculture. Existing technical solutions generally adopt the deployment of multi-source sensor networks such as soil and meteorology, and use data analysis models to generate irrigation decisions. These systems have reduced the burden of manual management to a certain extent. However, in long-term actual operation, existing systems have the following two main problems at the level of control command generation and execution:

[0003] First, there is a mismatch between the control timing and the dynamic response characteristics of the actuators: the control sequences for valve opening and closing, pump start and stop, etc., generated by the system's host computer based on static models (such as fixed delay and rated flow) do not fully consider the dynamic performance degradation of the actuators (such as solenoid valves and variable frequency pumps) due to factors such as mechanical wear, electromagnetic characteristic drift, and power supply voltage fluctuations during actual operation. For example, the action time of a solenoid valve from coil energization to valve core movement to rated opening will increase with the duration of use; after impeller wear, the speed-flow-pressure curve of a water pump will shift. If the control commands are still executed according to the initial calibration parameters, it will directly lead to hydraulic timing disorder in multi-valve coordinated irrigation: if the preceding valve fails to reach the predetermined opening due to response lag, and the subsequent valves have already started according to the original sequence, it will cause a sudden drop in local pressure in the pipeline network and insufficient flow at the end; or it will be forced to extend the overall irrigation cycle to wait for the preceding action to complete, missing the optimal soil moisture compensation window. This disconnect between "plan" and "reality" makes it impossible to accurately implement refined irrigation strategies, resulting in decreased irrigation uniformity and hidden waste of water resources.

[0004] Secondly, there is the problem of controller parameters failing to adapt to the time-varying characteristics of the controlled object: To achieve stable regulation of flow and pressure, the system often uses PID or its improved algorithms (such as fuzzy PID) as the core of closed-loop control. Its control parameters (proportional coefficient Kp, integral time Ti, derivative time Td) are usually tuned and fixed during the system commissioning phase based on the initial characteristics of the controlled object (such as a hydraulic system composed of pipe network, valves, and pumps). However, during long-term operation, the characteristics of the controlled object will change slowly or abruptly due to pipe scaling, filter blockage, changes in valve flow resistance, and load increases or decreases. A controller with fixed parameters cannot adapt to these time-varying characteristics, easily leading to a deterioration in system regulation quality, manifested as: increased response overshoot, prolonged transient time, inability to eliminate steady-state error, and even continuous pressure oscillations. This not only affects irrigation accuracy but may also exacerbate equipment fatigue due to frequent adjustment actions. While existing systems can monitor the health status of equipment through sensors, they lack a mechanism to feed back real-time operating parameters such as equipment performance degradation and pipeline impedance changes to the controller parameter self-tuning stage in an online, closed-loop manner. This results in the system being unable to maintain optimal control performance throughout its entire lifecycle. Summary of the Invention

[0005] To address the above issues, this invention solves the control mismatch problem by real-time monitoring of the dynamic performance parameters of the actuator and online identification of the pipeline model parameters, and by adaptive timing compensation and composite correction of the control commands.

[0006] The technical solution is an IoT-based intelligent control system for irrigation with multiple parameters, including a dynamic performance module, a parameter identification module, and an intelligent control module.

[0007] The dynamic performance module acquires the dynamic performance parameters of the irrigation actuator and the current pipeline pressure. The dynamic performance parameters include the actual response time T_act of the solenoid valve and the current flow-pressure characteristic coefficient K_v of the variable frequency pump.

[0008] The parameter identification module uses the recursive prediction error method to perform online parameter identification based on the change of the current pipeline pressure P_now and the system control input, and obtains the current parameters of the controlled object model. The controlled object model is a first-order inertial plus pure time delay model. The current parameters include the system equivalent time constant T_sys and the pure time delay τ.

[0009] The intelligent control module adaptively adjusts the basic control commands determined by the actual response time T_act and the current flow-pressure characteristic coefficient K_v based on the system's equivalent time constant T_sys and the pure time delay τ. This adjustment includes the following steps:

[0010] Receive raw control commands, which include the planned action time t_plan, target flow rate Q_set, target pipeline pressure P_target, and the rated response time T_nom of the solenoid valve;

[0011] Based on the actual response time T_act, the rated response time T_nom, and the pure delay time τ, the command transmission time t_cmd is calculated, where... ;

[0012] Based on the current flow-pressure characteristic coefficient K_v, the target flow rate Q_set, and the current pipeline pressure P_now, calculate the feedforward control quantity U_ff;

[0013] Based on the system's equivalent time constant T_sys, adjust the proportional coefficient K_p and integral time T_i of the feedback controller;

[0014] At the command transmission time t_cmd, the final control signal U_final is output. The final control signal U_final is obtained by adding the feedforward control quantity U_ff to the feedback correction quantity U_fb(t) calculated by the feedback controller based on the real-time pressure deviation e(t), i.e., U_final = U_ff + U_fb(t). The integral variable η is time, t_0 is the start time of the current irrigation cycle, t is the current time, and the real-time pressure deviation e(t) = P_target - P_now.

[0015] Furthermore, the composite correction satisfies the following: when the increase in the system equivalent time constant T_sys compared to the baseline parameter T_sys_base exceeds 20%, the recalculation of the feedforward control quantity U_ff is triggered; and when the actual response time T_act exceeds the rated response time T_nom by 30%, the re-identification of the pure time delay τ is triggered; wherein, the threshold setting of 20% and 30% is based on ensuring that the system maintains control stability within the range of variation of T_sys and T_act.

[0016] Furthermore, the dynamic performance module inputs the current flow-pressure characteristic coefficient K_v and the current pipeline pressure P_now to the parameter identification module to drive the online update of model parameters; the parameter identification module inputs the system equivalent time constant T_sys and the pure time delay τ to the intelligent control module as the basis for its adaptive adjustment for composite correction. The composite correction control uses the pure time delay τ to correct the instruction transmission timing determined based on the actual response time T_act, uses the current flow-pressure characteristic coefficient K_v to calculate the feedforward control quantity, and uses the system equivalent time constant T_sys to adjust the parameters of the feedback controller, ultimately generating a control signal in a coordinated manner; the control signal output by the intelligent control module drives the actuator to act and affects the pipeline pressure. The changed pressure value is collected by the dynamic performance module and participates in the calculation of the parameter identification module again.

[0017] Furthermore, the step of the dynamic performance module to obtain the actual response time T_act includes: detecting the rising edge start time t_rise of the solenoid valve drive current signal, detecting the moment t_open_90 when the valve opening feedback signal reaches 90% of the rated opening, and calculating T_act = t_open_90 - t_rise;

[0018] The step of the dynamic performance module to obtain the current flow-pressure characteristic coefficient K_v includes: collecting the corresponding data of the outlet pressure P and flow rate Q of the variable frequency pump at different speeds, and determining K_v by linearly fitting the relationship P = P_0 -K_v * Q, where P_0 is the zero flow pressure;

[0019] The transfer function of the first-order inertial plus pure time delay model is G(s) = K * e^{-τ s} / (T_sys * s +1), where K is the static gain and s is the Laplace operator; the feedback controller is a PID controller, and the integral time T_i is the time constant of the integral element of the PID controller.

[0020] Further, based on the current flow-pressure characteristic coefficient K_v, the target flow rate Q_set, and the current pipeline pressure P_now, the feedforward control quantity U_ff is calculated, where U_ff = K_u * (Q_set / (C_v * sqrt((P_now + K_v * Q_set) / ΔP_std))), K_u is the coefficient obtained by fitting the valve's factory flow characteristic curve, C_v is the valve's rated flow coefficient, and ΔP_std is the valve's standard pressure difference for flow testing.

[0021] Furthermore, based on the system's equivalent time constant T_sys, the proportional coefficient K_p and integral time T_i of the feedback controller are adjusted, where K_p = K_p_base * (T_sys_base / T_sys), T_i = T_i_base * (T_sys / T_sys_base), and K_p_base, T_i_base, and T_sys_base are baseline parameters determined through step response testing during the initial commissioning of the system.

[0022] Further, the steps for determining the reference parameters K_p_base, T_i_base, and T_sys_base are as follows: In the initial state where the system pipeline is unloaded and all valves are closed, a unit step speed signal is input to the variable frequency pump, and the step response curve of the current pipeline pressure P_now is recorded. The time required for the pressure to rise to 63.2% of the steady-state value is obtained from the curve as T_sys_base, and K_p_base and T_i_base are obtained by tuning based on T_sys_base according to the critical proportional method or the attenuation curve method.

[0023] For a typical irrigation system with a rated head of 50 meters and a pipe diameter of DN100, the reference parameters are: K_p_base = 1.0, T_i_base = 30 seconds, and T_sys_base = 10 seconds.

[0024] A control method for an IoT-based multi-parameter intelligent irrigation control system includes the following steps:

[0025] S1, obtain the dynamic performance parameters of the irrigation actuator and the current pipeline pressure. The dynamic performance parameters include the actual response time T_act of the solenoid valve and the current flow-pressure characteristic coefficient K_v of the variable frequency pump.

[0026] S2, based on the change of the current pipeline pressure P_now and the system control input, the recursive prediction error method is used to identify online parameters to obtain the current parameters of the controlled object model. The controlled object model is a first-order inertial plus pure time delay model. The current parameters include the system equivalent time constant T_sys and the pure time delay τ.

[0027] S3, by using the system equivalent time constant T_sys and the pure time delay τ as the basis for adaptive adjustment, the basic control command determined by the actual response time T_act and the current flow-pressure characteristic coefficient K_v is compounded and modified to generate and output the final control signal.

[0028] Furthermore, step S3 includes feedforward and timing control sub-steps:

[0029] Receive raw control commands, which include the planned action time t_plan, target flow rate Q_set, target pipeline pressure P_target, and the rated response time T_nom of the solenoid valve;

[0030] Based on the actual response time T_act, the rated response time T_nom, and the pure delay time τ, calculate the instruction sending time t_cmd, where t_cmd = t_plan + max(0, (T_act - T_nom)) - τ;

[0031] Based on the current flow-pressure characteristic coefficient K_v, the target flow rate Q_set, and the current pipeline pressure P_now, the feedforward control quantity U_ff is calculated, where U_ff = K_u * (Q_set / (C_v * sqrt((P_now +K_v * Q_set) / ΔP_std))), K_u is the coefficient obtained by fitting the valve's factory flow characteristic curve, C_v is the valve's rated flow coefficient, and ΔP_std is the standard differential pressure for valve flow testing.

[0032] Furthermore, step S3 also includes a feedback adjustment and synthesis output sub-step:

[0033] Based on the system's equivalent time constant T_sys, the proportional coefficient K_p and integral time T_i of the feedback controller are adjusted, where K_p = K_p_base * (T_sys_base / T_sys), T_i = T_i_base * (T_sys / T_sys_base), and K_p_base, T_i_base, and T_sys_base are baseline parameters determined by step response testing during the initial commissioning of the system.

[0034] At the command sending time t_cmd, the final control signal U_final is output. The final control signal U_final is obtained by adding the feedforward control quantity U_ff and the feedback correction quantity U_fb(t) calculated by the feedback controller based on the real-time pressure deviation e(t), that is, U_final = U_ff + U_fb(t), where U_fb(t) = K_p* e(t) + (K_p / T_i) * ∫_{t_0}^{t} e(η) dη, the integral variable η is time, t_0 is the start time of the current irrigation cycle, t is the current time, and the real-time pressure deviation e(t) = P_target - P_now.

[0035] Due to the adoption of the above technical solutions, the present invention has the following advantages compared with the prior art;

[0036] The dynamic performance module and parameter identification module work together to acquire and identify in real time the actuator performance parameters (T_act, K_v) and controlled pipeline model parameters (T_sys, τ) that characterize the dynamic characteristics of the system. The intelligent control module compensates for the timing of T_act and τ to offset physical lag and ensure that the control action is strictly synchronized with the irrigation plan. Based on the model feedforward control of the K_v parameter, the dominant control quantity (U_ff) is calculated to achieve high-speed and efficient response to the setpoint and major disturbances. Based on the adaptive adjustment of the feedback parameter of T_sys, the controller gain (K_p, T_i) is optimized in real time to ensure that the system maintains the best damping and stability margin under various operating conditions. This realizes the adaptive matching of the control system to the performance of the actuator and the dynamic characteristics of the controlled object, which significantly improves the control accuracy, stability and irrigation efficiency of the system in long-term operation. Attached Figure Description

[0037] Figure 1 This is a system flowchart of an IoT-based intelligent control system and method for irrigation with multiple parameters.

[0038] Figure 2 This is a flowchart of a multi-parameter intelligent control system and method for irrigation based on the Internet of Things (IoT) of the present invention.

[0039] Figure 3 This is a practical application diagram of the IoT-based intelligent control system and method for irrigation with multiple parameters according to the present invention. Detailed Implementation

[0040] The foregoing and other technical contents, features and effects of the present invention are described in conjunction with the appendix below. Figures 1 to 3 The detailed description of the embodiments will make this clear. All structural details mentioned in the following embodiments are based on the accompanying drawings.

[0041] Based on existing technology, the dynamic performance module obtains the dynamic performance parameters of the irrigation actuator and the current pipeline pressure. The dynamic performance parameters include the actual response time T_act of the solenoid valve and the current flow-pressure characteristic coefficient K_v of the variable frequency pump.

[0042] The parameter identification module uses the recursive prediction error method to identify parameters online based on the change of the current pipeline pressure P_now and the system control input, and obtains the current parameters of the controlled object model. The controlled object model is a first-order inertial plus pure time delay model. The current parameters include the system equivalent time constant T_sys and the pure time delay τ.

[0043] The intelligent control module adaptively adjusts the basic control commands determined by the actual response time T_act and the current flow-pressure characteristic coefficient K_v based on the system's equivalent time constant T_sys and pure time delay τ. This adjustment includes the following steps:

[0044] Receive raw control commands, which include the planned action time t_plan, target flow rate Q_set, target pipeline pressure P_target, and the rated response time T_nom of the solenoid valve;

[0045] Based on the actual response time T_act, the rated response time T_nom, and the pure delay time τ, the command transmission time t_cmd is calculated, where... ;

[0046] Calculate the feedforward control quantity U_ff based on the current flow-pressure characteristic coefficient K_v, the target flow rate Q_set, and the current pipeline pressure P_now;

[0047] Based on the system's equivalent time constant T_sys, adjust the proportional coefficient K_p and integral time T_i of the feedback controller;

[0048] At the command transmission time t_cmd, the final control signal U_final is output. The final control signal U_final is obtained by adding the feedforward control quantity U_ff and the feedback correction quantity U_fb(t) calculated by the feedback controller based on the real-time pressure deviation e(t), that is, U_final = U_ff + U_fb(t), where The integral variable η is time, t_0 is the start time of the current irrigation cycle, t is the current time, and the real-time pressure deviation e(t) = P_target - P_now.

[0049] The composite correction satisfies the following conditions: when the increase in the system's equivalent time constant T_sys compared to the baseline parameter T_sys_base exceeds 20%, the feedforward control quantity U_ff is recalculated; and when the actual response time T_act exceeds the rated response time T_nom by 30%, the pure time delay τ is reidentified. The 20% and 30% thresholds are set to ensure that the system maintains control stability within the range of variation of T_sys and T_act.

[0050] The dynamic performance module inputs the current flow-pressure characteristic coefficient K_v and the current pipeline pressure P_now to the parameter identification module to drive the online update of model parameters. The parameter identification module inputs the system equivalent time constant T_sys and the pure time delay τ to the intelligent control module as the basis for its adaptive adjustment for composite correction. The composite correction control uses the pure time delay τ to correct the command transmission timing determined based on the actual response time T_act, uses the current flow-pressure characteristic coefficient K_v to calculate the feedforward control quantity, and uses the system equivalent time constant T_sys to adjust the parameters of the feedback controller, ultimately generating a control signal in a coordinated manner. The control signal output by the intelligent control module drives the actuator to act and affects the pipeline pressure. The changed pressure value is collected by the dynamic performance module and participates in the calculation of the parameter identification module again.

[0051] The steps for the dynamic performance module to obtain the actual response time T_act include: detecting the rising edge start time t_rise of the solenoid valve drive current signal, detecting the moment when the valve opening feedback signal reaches 90% of the rated opening t_open_90, and calculating T_act = t_open_90 - t_rise;

[0052] The steps for the dynamic performance module to obtain the current flow-pressure characteristic coefficient K_v include: collecting the corresponding data of outlet pressure P and flow rate Q of the variable frequency pump at different speeds, and determining K_v by linearly fitting the relationship P = P_0 - K_v * Q, where P_0 is the zero flow pressure;

[0053] The transfer function of the first-order inertial plus pure time delay model is G(s) = K * e^{-τ s} / (T_sys * s + 1), where K is the static gain and s is the Laplace operator; the feedback controller is a PID controller, and the integral time T_i is the time constant of the integral element of the PID controller.

[0054] Based on the current flow-pressure characteristic coefficient K_v, the target flow rate Q_set, and the current pipeline pressure P_now, the feedforward control quantity U_ff is calculated, where U_ff = K_u * (Q_set / (C_v * sqrt((P_now + K_v * Q_set) / ΔP_std))), K_u is the coefficient obtained by fitting the valve's factory flow characteristic curve, C_v is the valve's rated flow coefficient, and ΔP_std is the standard differential pressure of the valve flow test.

[0055] Based on the system's equivalent time constant T_sys, the proportional coefficient K_p and integral time T_i of the feedback controller are adjusted, where K_p = K_p_base * (T_sys_base / T_sys), T_i = T_i_base * (T_sys / T_sys_base), and K_p_base, T_i_base, and T_sys_base are baseline parameters determined through step response testing during the initial commissioning of the system.

[0056] The steps for determining the baseline parameters K_p_base, T_i_base, and T_sys_base are as follows: In the initial state where the system pipeline is unloaded and all valves are closed, a unit step speed signal is input to the variable frequency pump, and the step response curve of the current pipeline pressure P_now is recorded. The time required for the pressure to rise to 63.2% of the steady-state value is obtained from the curve as T_sys_base, and K_p_base and T_i_base are obtained by tuning based on T_sys_base according to the critical proportional method or the attenuation curve method.

[0057] For a typical irrigation system with a rated head of 50 meters and a pipe diameter of DN100, the baseline parameters are: K_p_base = 1.0, T_i_base = 30 seconds, and T_sys_base = 10 seconds.

[0058] A control method for an IoT-based multi-parameter intelligent irrigation control system includes the following steps:

[0059] S1, obtain the dynamic performance parameters of the irrigation actuator and the current pipeline pressure. The dynamic performance parameters include the actual response time T_act of the solenoid valve and the current flow-pressure characteristic coefficient K_v of the variable frequency pump.

[0060] S2, based on the change of the current pipeline pressure P_now and the system control input, the recursive prediction error method is used to identify the online parameters and obtain the current parameters of the controlled object model. The controlled object model is a first-order inertial plus pure time delay model. The current parameters include the system equivalent time constant T_sys and the pure time delay τ.

[0061] S3 uses the system equivalent time constant T_sys and pure time delay τ as adaptive adjustment criteria to perform composite correction on the basic control commands determined by the actual response time T_act and the current flow-pressure characteristic coefficient K_v, respectively, in order to generate and output the final control signal.

[0062] Step S3 includes feedforward and timing control sub-steps:

[0063] Receive raw control commands, which include the planned action time t_plan, target flow rate Q_set, target pipeline pressure P_target, and the rated response time T_nom of the solenoid valve;

[0064] Based on the actual response time T_act, the rated response time T_nom, and the pure delay time τ, calculate the command transmission time t_cmd, where t_cmd = t_plan + max(0, (T_act - T_nom)) - τ;

[0065] Based on the current flow-pressure characteristic coefficient K_v, the target flow rate Q_set, and the current pipeline pressure P_now, the feedforward control quantity U_ff is calculated, where U_ff = K_u * (Q_set / (C_v * sqrt((P_now + K_v * Q_set) / ΔP_std))), K_u is the coefficient obtained by fitting the valve's factory flow characteristic curve, C_v is the valve's rated flow coefficient, and ΔP_std is the standard differential pressure of the valve flow test.

[0066] Step S3 also includes a feedback adjustment and output synthesis sub-step:

[0067] Based on the system's equivalent time constant T_sys, the proportional coefficient K_p and integral time T_i of the feedback controller are adjusted, where K_p = K_p_base * (T_sys_base / T_sys), T_i = T_i_base * (T_sys / T_sys_base), and K_p_base, T_i_base, and T_sys_base are baseline parameters determined by step response testing during the initial commissioning of the system.

[0068] At the command sending time t_cmd, the final control signal U_final is output. The final control signal U_final is obtained by adding the feedforward control quantity U_ff and the feedback correction quantity U_fb(t) calculated by the feedback controller based on the real-time pressure deviation e(t), that is, U_final = U_ff + U_fb(t), where U_fb(t) = K_p * e(t) + (K_p / T_i) * ∫_{t_0}^{t} e(η) dη, the integral variable η is time, t_0 is the start time of the current irrigation cycle, t is the current time, and the real-time pressure deviation e(t) = P_target - P_now.

[0069] To facilitate understanding of this system, further detailed explanation is provided below:

[0070] 1.1 System initialization and baseline parameter acquisition;

[0071] After the system is first put into operation or overhauled, an initial step response test must be performed to determine the baseline parameters K_p_base and T_i_base of the feedback controller and the baseline time constant T_sys_base of the system.

[0072] The specific steps are as follows:

[0073] Ensure the system piping is unloaded and all terminal solenoid valves are closed.

[0074] Send a unit step signal from 0% to 50% of the rated speed to the variable frequency pump control system.

[0075] High-speed acquisition of pressure sensor data on the main pipeline network, recording the step response curve of pressure P_now as a function of time t.

[0076] From this curve, find the time point corresponding to the pressure rising from the initial value to 63.2% of the final steady-state value. This time interval is the baseline equivalent time constant T_sys_base of the system under the current configuration.

[0077] Based on the obtained T_sys_base, the baseline proportional coefficient K_p_base and baseline integral time T_i_base of the PID controller are determined using engineering tuning methods (such as the Ziegler-Nichols critical proportional gain method or the decay curve method). For a typical system with a rated head of 50 meters and a main pipeline diameter of DN100, a set of empirical baseline values ​​are: T_sys_base = 10s, K_p_base = 1.0, T_i_base = 30s. These baseline values ​​will serve as a reference for subsequent online adaptive adjustments.

[0078] 1.2 Implementation of online parameter identification using the recursive prediction error method;

[0079] This system employs a recursive prediction error method to perform online real-time identification of the parameters of the first-order inertial plus pure time-delay model of the controlled object (i.e., the "pump-pipeline" system). This algorithm can continuously update the estimates of T_sys and τ using the latest input and output data.

[0080] Model Discretization: First, the continuous-time model G(s) = K * e^{-τ s} / (T_sys * s + 1) is discretized with pure time delay over a sampling period T, resulting in the difference equation form:

[0081] y(k) = a * y(k-1) + b * u(k - d - 1);

[0082] Where y(k) is the pipeline pressure P_now at time k, u(k) is the controller's total output U_final or pump speed command at time k, d = floor(τ / ​​T) is the discrete lag step, and the parameters a and b are related to the continuous time parameter T_sys,K.

[0083] Parameter vector and data vector: Define the parameter vector to be identified as θ(k) = [a, b]^T, and the data vector as φ(k) = [y(k-1), u(kd-1)]^T. Then the prediction model is: y_hat(k|θ) = φ^T(k) θ(k-1).

[0084] Recursive update law: The algorithm performs the following steps in each control cycle:

[0085] Calculate the prediction error: e(k) = y(k) - y_hat(k|θ);

[0086] Calculate the gain matrix: L(k) = [ P(k-1) φ(k) ] / [ λ + φ^T(k) P(k-1) φ(k) ];

[0087] Update parameter estimate: θ(k) = θ(k-1) + L(k)e(k);

[0088] Update the covariance matrix: P(k) = (1 / λ) * [I - L(k) φ^T(k)] P(k-1);

[0089] Where P(k) is the covariance matrix, and λ (0.95 < λ < 1) is the forgetting factor, used to make the algorithm focus more on recent data, thereby tracking time-varying parameters. Finally, continuous T_sys and K can be solved from the converged θ, while the pure time lag τ can be determined by monitoring the cross-correlation peak between the pressure response and the control command or by searching for the model error as a function of d.

[0090] 1.3 Detailed calculation of the feedforward control quantity U_ff;

[0091] Feedforward control aims to directly calculate a basic actuator drive quantity based on the target flow rate Q_set and the current operating conditions, in order to quickly approximate the target. The formula is:

[0092] U_ff = K_u * ( Q_set / ( C_v * sqrt( ( P_now + K_v * Q_set ) / ΔP_std ) ) );

[0093] 1.4 Timing correction logic for instruction sending time t_cmd;

[0094] Timing correction is used to compensate for the slow response of the solenoid valve and the pure time delay of the system, ensuring that the control action takes effect at the desired time t_plan. The formula is:

[0095] t_cmd = t_plan + max(0, (T_act - T_nom))–τ;

[0096] Where: T_act - T_nom: indicates how much slower the actual response of the solenoid valve is compared to the rated value due to aging, wear, etc. If it is actually faster, this item is negative, but taking max(0, ...) means that we only compensate for the delay and do not send it in advance (because sending it in advance may cause malfunctions).

[0097] + max(0, (T_act - T_nom)): This means that if the valve slows down, the command needs to be sent in advance to "wait" for the valve to move slowly.

[0098] - τ: τ is the pressure response lag time of the signal transmission in the pipeline network. To compensate for this lag, control commands need to be sent even earlier by τ.

[0099] 1.5 Adaptive adjustment of feedback controller parameters;

[0100] The feedback controller (PID) adjusts its parameters online based on the identified system equivalent time constant T_sys to maintain optimal closed-loop performance. The adjustment formula is:

[0101] K_p = K_p_base * (T_sys_base / T_sys);

[0102] T_i = T_i_base * (T_sys / T_sys_base);

[0103] For first-order inertial plus pure time-delay models, many tuning rules (such as Cohen-Coon and IMC) indicate that the proportional gain K_p should be inversely proportional to the system time constant T_sys, and the integral time T_i should be directly proportional to T_sys. When T_sys increases (the system inertia increases and the response slows down), decreasing K_p can prevent overshoot, while increasing T_i can mitigate the integral action and avoid oscillations; conversely, the opposite is also true. In essence, this is a model-based adaptive control method.

[0104] The key terms, symbols, and formulas are explained below:

[0105] I. Key Terms and Symbols;

[0106] T_act (actual response time of the solenoid valve): The unit is seconds (s), which refers to the time it takes for the valve to reach 90% of its rated opening after the drive current starts to rise. This parameter directly reflects the valve's mechanical wear, lubrication condition, and other health conditions.

[0107] K_v (current flow-pressure characteristic coefficient of the variable frequency pump): The unit is usually bar / (m³ / h), which characterizes the slope of the decrease in pump outlet pressure as the flow rate increases under the current operating conditions. The relationship is P_pump = P_0 - K_v * Q, where P_0 is the zero flow head. This coefficient varies with pump speed and wear condition.

[0108] P_now (current pipeline pressure): The unit is bar or MPa, referring to the real-time pressure value at key monitoring points in the system pipeline (such as pump outlets and main pipes). It is the most critical feedback variable in the control system.

[0109] T_sys (System Equivalent Time Constant): Measured in seconds (s), it is a parameter characterizing the system's response speed when the complex hydraulic system consisting of pumps, pipe networks, and valves is equivalent to a first-order inertial element. A larger value indicates greater system inertia and slower pressure changes.

[0110] τ (pure time delay): The unit is seconds (s), which refers to the time required for the pressure fluctuation caused by a control signal (such as a change in valve opening) to propagate to the monitored point after it is issued. It is related to the pipeline length and the speed of sound in water.

[0111] t_plan (planned action time) and t_cmd (command transmission time): The unit is seconds (s). t_plan is the expected time point in the irrigation strategy to change; t_cmd is the actual time point when the command is sent after timing correction.

[0112] Q_set (target flow rate) and P_target (target network pressure): These are the flow rate unit (e.g., m³ / h) and pressure unit (e.g., bar), respectively, and are the expected values ​​set according to irrigation needs.

[0113] U_ff (feedforward control input) and U_fb(t) (feedback correction input): Both are dimensionless quantities (usually normalized). U_ff is a coarse adjustment command based on the static model; U_fb(t) is a dynamic correction value calculated by the PID controller based on the real-time deviation.

[0114] K_p (proportional gain) and T_i (integration time): K_p is the dimensionless proportional gain; T_i is in seconds (s) and is the time constant of the integral element.

[0115] K (Static Gain): Dimensionless, it is the steady-state gain of the system model, representing the steady-state pressure change caused by a unit control input. In this system, its impact has been mainly compensated by the feedforward element.

[0116] e^{-τ s} represents delaying a time-domain signal by τ time units. In the transfer function of a control system, it means that the input signal will appear unchanged at the output after a time of τ.

[0117] II. Formula;

[0118] Timing correction formula: t_cmd = t_plan + max(0, (T_act - T_nom)) – τ;

[0119] Function: To enable the "early transmission" of control commands to compensate for valve action delays and pressure propagation lags.

[0120] In the formula, the part "max(0, (T_act - T_nom))" is used to compensate for the valve's own delay relative to its rated value; the part "-τ" is used to compensate for the pure lag time of the pressure wave propagation in the pipeline network. The two work together to ensure that the control effect occurs near the planned time t_plan.

[0121] Feedforward control formula: U_ff = K_u * ( Q_set / ( C_v * sqrt( ( P_now + K_v * Q_set ) / ΔP_std ) ) );

[0122] Function: Based on the valve flow equation and the real-time characteristics of the pump, directly calculate the theoretical control quantity (valve opening or pump speed) required to meet the target flow rate Q_set.

[0123] The formula is essentially the inverse operation of the valve flow equation Q = C_v * sqrt(ΔP / ΔP_std), where ΔP is approximated by adding the pump pressure drop K_v * Q_set caused by the flow rate to the current pump outlet pressure P_now. K_u is a correction coefficient fitted based on the actual factory characteristic curve of the valve, used to improve the calculation accuracy.

[0124] Feedback parameter adaptive formula:

[0125] K_p = K_p_base * (T_sys_base / T_sys);

[0126] T_i = T_i_base * (T_sys / T_sys_base);

[0127] Function: To automatically adjust the parameters of the PID controller according to changes in the dynamic characteristics (T_sys) of the controlled object, thereby maintaining optimal performance of the control system. Specifically, when the system response slows down (T_sys increases), the proportional gain K_p is automatically reduced and the integral time T_i is increased to make the control action gentler and avoid overshoot and oscillation; when the system response speeds up, K_p is increased and T_i is decreased to strengthen the control action and accelerate the response speed. This is a model-based adaptive tuning strategy.

[0128] The formula for calculating the feedback control quantity is: U_fb(t) = K_p * e(t) + (K_p / T_i) * ∫_{t_0}^{t} e(η) dη;

[0129] Function: To calculate the output of the PID controller, U_fb(t).

[0130] This is the standard form of the positional PI control algorithm. The first term, K_p * e(t), is the proportional term, and its output is proportional to the current pressure deviation e(t) = P_target - P_now, providing rapid correction. The second term is the integral term, which eliminates static error by accumulating historical deviations; its strength is determined by K_p / T_i. This embodiment does not include a derivative term, mainly because the pipeline pressure signal may contain noise, and the derivative term is prone to causing control command oscillations.

[0131] In specific use, this invention is based on existing technology, such as... Figure 3As shown, it is solar-powered and has a built-in dynamic performance module, parameter identification module and intelligent control module. The dynamic performance module obtains the dynamic performance parameters of the irrigation actuator and the current pipeline pressure. The dynamic performance parameters include the actual response time T_act of the solenoid valve and the current flow-pressure characteristic coefficient K_v of the variable frequency pump.

[0132] The parameter identification module uses the recursive prediction error method to perform online parameter identification based on the change of the current pipeline pressure P_now and the system control input, and obtains the current parameters of the controlled object model. The controlled object model is a first-order inertial plus pure time delay model. The current parameters include the system equivalent time constant T_sys and the pure time delay τ.

[0133] The intelligent control module adaptively adjusts the basic control commands determined by the actual response time T_act and the current flow-pressure characteristic coefficient K_v based on the system's equivalent time constant T_sys and the pure time delay τ. This adjustment includes the following steps:

[0134] Receive raw control commands, which include the planned action time t_plan, target flow rate Q_set, target pipeline pressure P_target, and the rated response time T_nom of the solenoid valve;

[0135] Based on the actual response time T_act, the rated response time T_nom, and the pure delay time τ, the command transmission time t_cmd is calculated, where... ;

[0136] Based on the current flow-pressure characteristic coefficient K_v, the target flow rate Q_set, and the current pipeline pressure P_now, calculate the feedforward control quantity U_ff;

[0137] Based on the system's equivalent time constant T_sys, adjust the proportional coefficient K_p and integral time T_i of the feedback controller;

[0138] At the command transmission time t_cmd, the final control signal U_final is output. The final control signal U_final is obtained by adding the feedforward control quantity U_ff to the feedback correction quantity U_fb(t) calculated by the feedback controller based on the real-time pressure deviation e(t), i.e., U_final = U_ff + U_fb(t). The integral variable η is time, t_0 is the start time of the current irrigation cycle, t is the current time, and the real-time pressure deviation e(t) = P_target - P_now.

[0139] The above description is a further detailed explanation in conjunction with specific embodiments, and it should not be considered that the specific embodiments of the present invention are limited to this. For those skilled in the art to which this invention pertains and related fields, any extensions, operation methods, and data substitutions made based on the technical solution concept of this invention should fall within the protection scope of this invention.

Claims

1. An IoT-based intelligent multi-parameter control system for irrigation, characterized in that, It includes a dynamic performance module, a parameter identification module, and an intelligent control module; The dynamic performance module acquires the dynamic performance parameters of the irrigation actuator and the current pipeline pressure. The dynamic performance parameters include the actual response time T_act of the solenoid valve and the current flow-pressure characteristic coefficient K_v of the variable frequency pump. The parameter identification module uses the recursive prediction error method to perform online parameter identification based on the change of the current pipeline pressure P_now and the system control input, and obtains the current parameters of the controlled object model. The controlled object model is a first-order inertial plus pure time delay model. The current parameters include the system equivalent time constant T_sys and the pure time delay τ. The intelligent control module adaptively adjusts the basic control commands determined by the actual response time T_act and the current flow-pressure characteristic coefficient K_v based on the system's equivalent time constant T_sys and the pure time delay τ. Specifically, this includes the following steps: Receive raw control commands, which include the planned action time t_plan, target flow rate Q_set, target pipeline pressure P_target, and the rated response time T_nom of the solenoid valve; Based on the actual response time T_act, the rated response time T_nom, and the pure delay time τ, the command transmission time t_cmd is calculated, where... ; Based on the current flow-pressure characteristic coefficient K_v, the target flow rate Q_set, and the current pipeline pressure P_now, calculate the feedforward control quantity U_ff; Based on the system's equivalent time constant T_sys, adjust the proportional coefficient K_p and integral time T_i of the feedback controller; At the command transmission time t_cmd, the final control signal U_final is output. The final control signal U_final is obtained by adding the feedforward control quantity U_ff to the feedback correction quantity U_fb(t) calculated by the feedback controller based on the real-time pressure deviation e(t), i.e., U_final = U_ff + U_fb(t). The integral variable η is time, t_0 is the start time of the current irrigation cycle, t is the current time, and the real-time pressure deviation e(t) = P_target - P_now.

2. The system according to claim 1, characterized in that, Based on the system's equivalent time constant T_sys, the proportional coefficient K_p and integral time T_i of the feedback controller are adjusted, where K_p = K_p_base * (T_sys_base / T_sys), T_i = T_i_base * (T_sys / T_sys_base), and K_p_base, T_i_base, and T_sys_base are baseline parameters determined through step response testing during the initial commissioning of the system.

3. The system according to claim 2, characterized in that, The steps for determining the baseline parameters K_p_base, T_i_base, and T_sys_base are as follows: In the initial state where the system pipeline is unloaded and all valves are closed, a unit step speed signal is input to the variable frequency pump, and the step response curve of the current pipeline pressure P_now is recorded. The time required for the pressure to rise to 63.2% of the steady-state value is obtained from the curve as T_sys_base, and K_p_base and T_i_base are obtained based on T_sys_base according to the critical proportionality method or the attenuation curve method. For a typical irrigation system with a rated head of 50 meters and a pipe diameter of DN100, the reference parameters are: K_p_base = 1.0, T_i_base = 30 seconds, and T_sys_base = 10 seconds.

4. The system according to claim 3, characterized in that, The composite correction satisfies the following conditions: when the increase in the system equivalent time constant T_sys compared to the reference parameter T_sys_base exceeds 20%, the recalculation of the feedforward control quantity U_ff is triggered; and when the actual response time T_act exceeds the rated response time T_nom by 30%, the re-identification of the pure time delay τ is triggered.

5. The system according to claim 1, characterized in that, The dynamic performance module inputs the current flow-pressure characteristic coefficient K_v and the current pipeline pressure P_now to the parameter identification module to drive the online update of model parameters. The parameter identification module inputs the system equivalent time constant T_sys and the pure time delay τ to the intelligent control module as the basis for its adaptive adjustment for composite correction. The composite correction uses the pure time delay τ to correct the instruction transmission timing determined based on the actual response time T_act, uses the current flow-pressure characteristic coefficient K_v to calculate the feedforward control quantity, and uses the system equivalent time constant T_sys to adjust the parameters of the feedback controller, ultimately generating a control signal in a coordinated manner. The control signal output by the intelligent control module drives the actuator to act and affects the pipeline pressure. The changed pressure value is collected by the dynamic performance module and participates in the calculation of the parameter identification module again.

6. The system according to claim 1, characterized in that, The steps for the dynamic performance module to obtain the actual response time T_act include: detecting the rising edge start time t_rise of the solenoid valve drive current signal, detecting the moment t_open_90 when the valve opening feedback signal reaches 90% of the rated opening, and calculating T_act = t_open_90 - t_rise; The step of the dynamic performance module to obtain the current flow-pressure characteristic coefficient K_v includes: collecting the corresponding data of outlet pressure P and flow rate Q of the variable frequency pump at different speeds, and determining K_v by linearly fitting the relationship P = P_0 - K_v * Q, where P_0 is zero flow pressure; The transfer function of the first-order inertial plus pure time-delay model is G(s) = K * e^{-τ s} / (T_sys * s + 1), where K is the static gain and s is the Laplace operator; the feedback controller is a PID controller, and the integral time T_i is the time constant of the integral element of the PID controller.

7. The system according to claim 1, characterized in that, Based on the current flow-pressure characteristic coefficient K_v, the target flow rate Q_set, and the current pipeline pressure P_now, the feedforward control quantity U_ff is calculated, where U_ff = K_u * (Q_set / (C_v * sqrt((P_now + K_v * Q_set) / ΔP_std))), K_u is the coefficient obtained by fitting the valve's factory flow characteristic curve, C_v is the valve's rated flow coefficient, and ΔP_std is the standard differential pressure for valve flow testing.

8. A control method executed by the IoT irrigation multi-parameter intelligent control system according to any one of claims 1 to 7, characterized in that, Includes the following steps: S1, obtain the dynamic performance parameters of the irrigation actuator and the current pipeline pressure. The dynamic performance parameters include the actual response time T_act of the solenoid valve and the current flow-pressure characteristic coefficient K_v of the variable frequency pump. S2, based on the change of the current pipeline pressure P_now and the system control input, the recursive prediction error method is used to identify online parameters to obtain the current parameters of the controlled object model. The controlled object model is a first-order inertial plus pure time delay model. The current parameters include the system equivalent time constant T_sys and the pure time delay τ. S3, by using the system equivalent time constant T_sys and the pure time delay τ as the basis for adaptive adjustment, the basic control command determined by the actual response time T_act and the current flow-pressure characteristic coefficient K_v is compounded and modified to generate and output the final control signal.

9. The method according to claim 8, characterized in that, Step S3 includes feedforward and timing control sub-steps: Receive raw control commands, which include the planned action time t_plan, target flow rate Q_set, target pipeline pressure P_target, and the rated response time T_nom of the solenoid valve; Based on the actual response time T_act, the rated response time T_nom, and the pure delay time τ, calculate the instruction sending time t_cmd, where t_cmd = t_plan + max(0, (T_act - T_nom)) - τ; Based on the current flow-pressure characteristic coefficient K_v, the target flow rate Q_set, and the current pipeline pressure P_now, the feedforward control quantity U_ff is calculated, where U_ff = K_u * (Q_set / (C_v * sqrt((P_now + K_v * Q_set) / ΔP_std))), K_u is the coefficient obtained by fitting the valve's factory flow characteristic curve, C_v is the valve's rated flow coefficient, and ΔP_std is the standard differential pressure for valve flow testing.

10. The method according to claim 9, characterized in that, Step S3 further includes a feedback adjustment and output synthesis sub-step: Based on the system's equivalent time constant T_sys, the proportional coefficient K_p and integral time T_i of the feedback controller are adjusted, where K_p = K_p_base * (T_sys_base / T_sys), T_i = T_i_base * (T_sys / T_sys_base), and K_p_base, T_i_base, and T_sys_base are baseline parameters determined by step response testing during the initial commissioning of the system. At the command sending time t_cmd, the final control signal U_final is output. The final control signal U_final is obtained by adding the feedforward control quantity U_ff and the feedback correction quantity U_fb(t) calculated by the feedback controller based on the real-time pressure deviation e(t), that is, U_final = U_ff + U_fb(t), where U_fb(t) = K_p* e(t) + (K_p / T_i) * ∫_{t_0}^{t} e(η) dη, the integral variable η is time, t_0 is the start time of the current irrigation cycle, t is the current time, and the real-time pressure deviation e(t) = P_target - P_now.

Images