Method for scheduling water and fertilizer integration resources in orange and pomelo orchard based on multi-objective optimization

Abstract

The application discloses a method for water and fertilizer integration resource scheduling of orange and pomelo orchard based on multi-objective optimization, and particularly relates to the technical field of irrigation and fertilization control. The conductivity, flow, pressure, running quantity of the fertilizer injection pump and opening and closing feedback of the partition electromagnetic valve are collected, and the normalized concentration trajectory is obtained by aligning the slices under the unified time reference. The equivalent retention volume and mixing time constant of the main pipe are identified online by triggering the partition switching and pump start-stop events, the dynamic identifiable and reliable comprehensive quantity of the fertilizer injection is constructed and is gated. When the gating is satisfied, the partition fertilizer injection process is planned as the switching buffer section, the steady-state fertilizer injection section and the end compensation section, the multi-objective optimization model considering the fertilizer liquid string area risk and the partition realization deviation risk is established, and the partition fertilizer injection timing scheduling scheme is generated. After the execution, the effective fertilization dose and the fertilizer liquid impact are calculated, the risk comprehensive quantity is formed, the online compensation is triggered or the fertilization arrears record is generated, and the traceable closed-loop scheduling is realized.

Description

Technical Field

[0001] This invention relates to the field of irrigation and fertilization control technology, and more specifically, to a multi-objective optimization-based method for integrated water and fertilizer resource scheduling in citrus orchards. Background Technology

[0002] Smart agriculture's integrated water and fertilizer control technology is mainly used for the automated scheduling and closed-loop control of orchard irrigation and fertilization processes, enabling water and fertilizer supply to be allocated on demand, stably delivered, and traceably accounted for in different irrigation zones. This type of technology is widely used in large-scale orchard water-saving irrigation, precision fertilization in facility agriculture, and field water and fertilizer management. Among these, the scheme of using a single fertilizer injection pump with a mixing tank and rotating irrigation according to irrigation zones is widely adopted due to its low equipment investment and ease of implementation. Typically, a main EC sensor, flow meter, and pressure sensor are configured on the pump station side, and zone switching control is achieved through PLC and zone solenoid valves.

[0003] The existing technology has the following shortcomings: In the past, orchard fertigation systems often relied on experience-based control during zone switching. A common practice was to immediately close the solenoid valve for zone A and open the solenoid valve for zone B after fertilization in zone A, or simply add a fixed-duration pre-flushing / post-flushing section. Because the fertilization pump and mixing process have dynamic responses, and there is a significant amount of residual volume in the main and branch pipes, valve switching can carry the high-EC fertilizer solution remaining in the pipeline into the next zone. This results in a short-term fertilizer surge at the beginning of zone B, while the effective fertilizer solution may be cut off before reaching the root zone at the end of zone A, leading to fertilizer cross-zone and zone-specific delivery discrepancies. Furthermore, differences in pipeline length and volume between different zones, as well as real-time flow fluctuations, further amplify this discrepancy, making it difficult to maintain consistent effectiveness under multiple operating conditions using a fixed-duration strategy. On the other hand, EC is mostly measured only at the pump station side on-site. There is a lack of methods to back-calculate the EC process curve into the effective fertilizer dosage for each zone and to carry out closed-loop compensation accordingly. The scheduling layer also often simplifies the decision to the total fertilizer amount for each zone, without incorporating concentration, time trajectory and switching transients into the constraint expression, which makes it difficult for the execution layer to achieve calculable and correctable process control.

[0004] To address the above problems, this invention proposes a solution. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a multi-objective optimization-based water and fertilizer integrated resource scheduling method for citrus orchards to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A multi-objective optimization-based method for integrated water and fertilizer resource allocation in citrus orchards includes the following steps: Triggered by the switching event of adjacent irrigation zones, the equivalent retention volume of the main pipe is determined based on the process data set and the normalized concentration trajectory; triggered by the start-stop event of the fertilizer injection pump, the mixing time constant is determined based on the rising segment of the normalized concentration trajectory after the start of the fertilizer injection pump; and the comprehensive quantity of identifiable reliability of fertilizer injection dynamics is determined based on the process data set and the normalized concentration trajectory. When the overall reliability of the dynamic identification of fertilization is not lower than the preset threshold, the fertilization execution process is planned into a switching buffer section, a steady-state fertilization section and an end compensation section based on the equivalent retention volume of the main pipe, the mixing time constant and the target fertilization dose of each irrigation zone, and a candidate zone fertilization timing scheduling scheme is generated. The risk of fertilizer cross-contamination is determined based on the volume of water flowing through the switching buffer zone and the equivalent retention volume of the main pipe. The risk of deviation in the effective fertilization rate of each zone is determined based on the deviation between the planned effective fertilization dose and the target fertilization dose. Based on the risks of fertilizer cross-contamination and the risk of deviation in the effective fertilization rate of each zone, the fertilization timing scheduling scheme for each zone is determined, and the fertilization pump and the solenoid valves of each irrigation zone are controlled.

[0007] In a preferred embodiment, determining the equivalent retention volume of the main pipe based on the process data set and the normalized concentration trajectory includes: The process variable sampling sequence, fertilizer injection pump operation volume, and opening / closing receipts of solenoid valves in each irrigation zone are acquired to form a process data set. The process variable sampling sequence includes at least one of the main pipe conductivity sampling sequence, main pipe flow rate sampling sequence, and main pipe pressure sampling sequence. The sampling sequences are aligned with a unified time standard, and the irrigation zones are divided into rotational irrigation segments based on the opening / closing receipts. The baseline conductivity of clear water and steady-state conductivity of fertilizer injection are determined based on the main pipe conductivity sampling sequence to obtain the normalized concentration trajectory.

[0008] In a preferred embodiment, determining the equivalent retention volume of the main pipe includes: Within the corresponding time window following the switching event of adjacent irrigation zones, candidate intervals for residual impact segments are determined based on the sudden rise and fall of the normalized concentration trajectory; the sampling sequence of main flow in the candidate intervals for residual impact segments is numerically accumulated to obtain the propulsion flow volume, which is then used as the calculated value of the equivalent retention volume of the main flow corresponding to the current switching event, and the equivalent retention volume of the main flow is recursively updated based on the calculated values ​​of the equivalent retention volume of the main flow corresponding to multiple switching events.

[0009] In a preferred embodiment, determining the candidate interval of the residual impact segment includes: The average value of the normalized concentration trajectory within multiple consecutive sampling periods before the switching event is used as the reference value before the switching. When the normalized concentration trajectory after the switching event continues to rise relative to the reference value before the switching, and the rise is not less than 0.1 and is maintained for at least two consecutive sampling periods, the moment when the normalized concentration trajectory first simultaneously satisfies the conditions of continuous rise, rise not less than 0.1, and maintenance for at least two consecutive sampling periods is determined as the starting moment of the candidate interval for the residual impact segment. After the normalized concentration trajectory reaches its peak value, when the difference between it and the reference value before the switching is not greater than 0.05 and is maintained for at least two consecutive sampling periods, the moment when the difference is first simultaneously satisfied for at least two consecutive sampling periods is determined as the ending moment of the candidate interval for the residual impact segment.

[0010] In a preferred embodiment, the recursive update of the supervisor's equivalent retention volume includes: The equivalent retention volume of the supervisor before the update is retained according to the first weight, and the calculated value of the equivalent retention volume of the supervisor corresponding to the current switching event is added according to the second weight to determine the updated equivalent retention volume of the supervisor. The sum of the first weight and the second weight is 1, and the second weight is used as the update coefficient. When the data availability corresponding to the current switching event decreases or the fluctuation of the supervisor flow and supervisor pressure increases, the update coefficient is reduced. When the calculated value of the equivalent retention volume of the supervisor corresponding to multiple consecutive switching events remains stable, the update coefficient is increased. When the current switching event has not formed a complete candidate interval of the residual impact segment or the comprehensive amount of the dynamic identifiable confidence of fertilizer injection is lower than the preset threshold, the calculated value of the equivalent retention volume of the supervisor corresponding to the current switching event is not used to update the equivalent retention volume of the supervisor.

[0011] In a preferred embodiment, determining the mixing time constant includes: The normalized concentration trajectory rise segment after the fertilizer injection pump starts is dynamically fitted to obtain the mixing time constant; the steady-state arrival time is determined based on the mixing time constant, the effective metering start point of the steady-state fertilizer injection segment is defined by the steady-state arrival time, and the duration of the switching buffer segment and the end compensation segment is determined.

[0012] 7. The method for integrated water and fertilizer resource scheduling in citrus orchards based on multi-objective optimization according to claim 6, characterized in that determining the comprehensive quantity of identifiable reliability of fertilizer injection dynamics includes: The comprehensive quantity of identifiable reliability of fertilizer injection dynamics is determined based on at least three of the following: main flow stability, main pressure stability, response segmentability of normalized concentration trajectory, switching event consistency, and data availability. Among these, the switching event consistency is determined based on the arrival delay and opening / closing jitter of the solenoid valve opening / closing receipts for each irrigation zone, and the data availability is determined based on the missing measurement marker, time alignment residual, and sensor saturation status.

[0013] In a preferred embodiment, fertilizer injection scheduling is performed based on the comprehensive identifiable reliability of the fertilizer injection dynamics, including: When the overall reliability of the dynamic identification of fertilizer injection is lower than the preset threshold, the identification result of this scheduling cycle is marked as low reliability. The preset conservative switching buffer time and conservative end compensation time are adopted, and the switching frequency of the zone is reduced or the number of consecutive switching is limited. When the overall reliability of the dynamic identification of fertilizer injection is not lower than the preset threshold, the fertilizer injection pump is kept in the off state or low output state in the switching buffer section, and the cumulative water volume in the switching buffer section is not less than the equivalent retention volume of the main pipe. After the steady state reaches the time, it enters the steady state fertilizer injection section. After the fertilizer injection pump stops, it enters the end compensation section and promotes the fertilizer solution at the end of this irrigation zone through clean water.

[0014] In a preferred embodiment, determining the partitioned fertilizer injection timing scheduling scheme includes: Under the constraints of the main flow rate limit, the main pressure limit, the maximum fertilization capacity of the fertilization pump, and the irrigation rotation sequence, multiple candidate fertilization timing schemes for each irrigation zone are generated by adjusting the irrigation rotation sequence, the opening and closing times of the solenoid valves in each irrigation zone, the start and stop times of the fertilization pumps, and the duration of the switching buffer section, the steady-state fertilization section, and the end compensation section. Weights are assigned to the fertilizer solution cross-regional risk and the zone effective fertilization fulfillment deviation risk corresponding to each candidate fertilization timing scheme, and a weighted average is performed to obtain a comprehensive evaluation result. The candidate fertilization timing scheme with the smallest comprehensive evaluation result is selected as the fertilization timing scheme for each zone. When the comprehensive evaluation results of two or more candidate fertilization timing schemes are the same, the candidate fertilization timing scheme with the shorter total execution time is selected.

[0015] In a preferred embodiment, after the partitioned fertilizer injection timing scheduling scheme is executed, the method further includes: When the overall reliability of dynamic fertilization injection is not lower than a preset threshold, a steady-state fertilization calculation window is determined based on the equivalent retention volume of the main pipe and the mixing time constant. Within the steady-state fertilization calculation window, the effective fertilization dose is calculated based on the main pipe flow rate and normalized concentration trajectory. Within the residual impact segment window, the fertilizer liquid impact area is determined based on the portion of the main pipe conductivity exceeding the impact threshold conductivity. The overall risk of effective fertilization fulfillment deviation in a given area is determined based on the target fertilization dose, effective fertilization dose, fertilizer liquid impact area, and overall reliability of dynamic fertilization injection. When the overall risk of effective fertilization fulfillment deviation in a given area exceeds a preset risk threshold or the effective fertilization dose is lower than the allowable lower limit, online compensation is performed. When online compensation cannot be completed due to time window, pump station capacity, or the order of rotational irrigation in a given area, the fertilization deficit is determined, and in the next scheduling cycle, the fertilization deficit is used as a correction item for the target fertilization dose of the corresponding irrigation area.

[0016] The technical effects and advantages of this invention are as follows: This invention elevates the fertilization process under zonal rotation irrigation from an experience-based switching method to executable timing and concentration trajectory control. Under a unified time reference, it utilizes operational variables such as mains EC (excessive EC), flow rate, and pressure to drive online identification of the mains equivalent retention volume and mixing time constant. A comprehensive measure of the identifiable reliability of fertilization dynamics is used as a gating condition, enabling the stable generation and distribution of PLC-executable control sequences even in real orchard environments with fluctuating data quality and operating conditions. Consequently, residual high EC fertilizer solution during zonal switching is preferentially confined to the switching buffer zone for consumption, suppressing the initial impact on the next zonal, and avoiding the fulfillment deviation caused by premature truncation at the end of the previous zonal application. This significantly reduces the risk of fertilizer cross-zonal application and improves the consistency and interpretability of the fertilization process across zonals.

[0017] Furthermore, this invention integrates the scheduling results with the segmented calculation and recharge closed loop: after execution, the effective fertilizer dosage and fertilizer liquid impact characteristics of each irrigation zone are calculated based on the process curve, and a comprehensive risk quantity of effective fertilizer delivery deviation is formed. When the threshold is exceeded or insufficient delivery occurs, online compensation is triggered. Priority is given to extending the steady-state fertilizer injection section or the end compensation section to promote the delivery of fertilizer liquid at the tail end, or to weaken the impact by extending the switching buffer section and reducing the switching frequency. If the time window or pump station capacity constraints prevent complete compensation in this cycle, the fertilizer backlog is calculated and written into the recharge record. The target fertilizer dosage is corrected in the next scheduling cycle. At the same time, the data of this cycle is used to continuously update the retention volume and mixed dynamic parameters, so that the system has the ability to continuously optimize in a calculable, correctable, and traceable manner. Attached Figure Description

[0018] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings; Figure 1 This is a flowchart illustrating the integrated water and fertilizer resource scheduling method for citrus orchards based on multi-objective optimization according to the present invention. Figure 2 This is a schematic diagram of the zonal fertilization control timing when the rotation irrigation sequence constraint and gating conditions are met. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1: This invention provides a multi-objective optimization-based method for integrated water and fertilizer resource scheduling in citrus orchards, such as... Figure 1As shown, it includes the following steps: Step 1: Collect time-synchronized slices; This step focuses on the integrated water and fertilizer management process in a citrus orchard with zoned rotational irrigation. Addressing the typical engineering problem of fertilizer solution cross-contamination during zone switching, the system first collects and organizes key operational quantities reflecting the fertilization process and pipeline status in a traceable manner into a process data set based on the same time reference. Specifically, the system deploys conductivity EC sensors at the main pump station to continuously measure the change in solution conductivity over time; flow meters are deployed at the main pump station to measure instantaneous flow rate; pressure sensors are deployed at the main pump or key pressure points to characterize water supply pressure levels and fluctuations; simultaneously, the PLC controller reads the fertilization pump operation data to characterize the pump's start / stop status and pump speed or instantaneous fertilization flow rate. For the timing boundaries on the execution side, the PLC controller records the opening and closing receipts of the solenoid valves in each irrigation zone, using the time corresponding to each valve opening and closing as the start and end boundaries of the zoned rotational irrigation segment. This allows for accurate segmentation, alignment, and backtracking of the subsequent fertilization process for each irrigation zone.

[0021] Since the aforementioned sensors and controllers may originate from different devices and communication links, their sampling periods, reporting delays, and clock errors are common in engineering sites. This step further performs unified time-scale alignment processing on each data source. Preferably, the system uses the local clock of the PLC controller as a unified time reference to resample the sampling sequences corresponding to the EC sensor, flow meter, pressure sensor, and fertilizer pump operating quantities, so that they form aligned time sequences under a preset sampling period Δt. Missing measurement points are filled in using nearest-neighbor hold or linear interpolation, while retaining missing measurement markers for subsequent reliability assessment. For example, the sampling period Δt is 5 seconds.

[0022] After time alignment is completed, the system divides the continuously sampled data into several operational segments according to the solenoid valve feedback of the irrigation zone, that is, it establishes a segment interval for each irrigation zone i. ,in This is the receipt time for the solenoid valve in this irrigation zone to switch from closed to open. This refers to the feedback time of the solenoid valve in the irrigation zone from opening to closing. Through this segmentation process, the EC measurement value, main flow rate, main pressure, and fertilizer pump operation at any given time can be mapped to the corresponding operation segment of the irrigation zone, thus providing accurate boundaries for subsequent switching event trigger identification and segmented calculation.

[0023] To reduce comparability issues caused by different clean water baselines, conductivity sensor zero-point bias, and different fertilizer formulations, this step further converts the conductivity EC time series into a normalized concentration trajectory. This ensures that the system can be stably used for process analysis under different operating days, fertilizer ratios, and water source conditions. The system first calculates the baseline conductivity of the main water source during periods of clean water operation or when the fertilizer injection pump is confirmed to be not running. Furthermore, during the period when the fertilizer injection process enters a stable phase and the fertilizer injection pump operates stably, the steady-state conductivity of the fertilizer injection system is statistically analyzed. The normalized concentration trajectory is then calculated using the following formula. : ; in, The main conductivity is represented by the EC sensor at time t. This represents the baseline conductivity of the clean water, used to reflect the difference between the water source conductivity and the sensor baseline under conditions without fertilizer injection. This indicates the steady-state electrical conductivity of fertilizer injection, which reflects the electrical conductivity level in the main pipe after stable mixing under the current formula and fertilizer injection conditions. This is a normalized concentration characterization quantity, whose value can be understood in engineering as the normalized concentration ratio relative to the clear water baseline, and is used to subsequently distinguish between the residual shock phase and the steady-state fertilization phase. Specifically, the conductivity of the clear water baseline... The determination method is as follows: before each irrigation begins or during the clean water flushing phase when the fertilizer pump is not running, the system continuously collects the conductivity of the main pipe and takes continuous values. The average conductivity over a period of time (e.g., 5 minutes) is taken as the period. Steady-state conductivity of fertilizer injection The method for determining this is as follows: During a period when the fertilizer injection pump is running continuously and the main flow rate is stable, select a continuous period of time. The average conductivity over a time period (e.g., 3 minutes) is used as . and It can be adjusted according to the actual project, and usually the period in which the conductivity fluctuation is less than 5% is selected.

[0024] Through the above processing, this step ultimately outputs a unified time-base aligned set of process data. , , , and Irrigation Zone Event Receipt Collection And simultaneously output the normalized concentration trajectory The mapping relationship between the operation segments of the irrigation zones and their zones is established. By associating sensor readings, hydraulic processes, fertilizer pump operation, and zone valve events on the same time base, this step lays the data foundation for subsequent online identification of pipeline stagnation volume and mixed dynamic parameters, as well as the interpretable calculation of effective fertilizer dosage for each irrigation zone. It also provides an actionable prerequisite for suppressing fertilizer-liquid cross-regional switching caused by zone switching.

[0025] Step 2: Identification of Residual Mixtures; This step is carried out based on the unified time-base process data formed in step one. The object of processing is the main conductivity response caused by zoning switching during zoning rotation irrigation. The input includes the normalized concentration trajectory. Main conductivity Main flow rate Q(t), main pressure P(t), fertilizer injection pump operating rate And the opening and closing receipt times of the solenoid valves in each irrigation zone. and Since the essence of fertilizer solution cross-contamination is the presence of a certain volume of residual fertilizer solution within the main and branch pipes, when the previous irrigation zone ends and the next irrigation zone begins, this residual fertilizer solution is carried into the next irrigation zone along with the clean water, resulting in a short-term high EC shock at the beginning of the next irrigation zone. Simultaneously, there is a dynamic lag between the start-up of the fertilizer injection pump and the stabilization of the main pipe concentration, meaning that effective fertilizer injection and the mixing transition do not coincide in time. To ensure that subsequent scheduling can implement the timing constraints of the switching buffer segment, steady-state fertilizer injection segment, and end-of-pipe compensation segment into an executable control sequence, this step first identifies the above two types of mechanism parameters online. Then, through multi-source consistency and data availability, a comprehensive quantity of identifiable reliability for fertilizer injection dynamics is constructed for gating subsequent segmented calculations and compensation.

[0026] In identifying the retained volume, the system uses the switching event between adjacent irrigation zones as the trigger point for calculation. Let the previous irrigation zone at time... End and switch to the next irrigation zone, where The time can be taken as the closing receipt time of the previous irrigation zone. Or the start receipt time of the next irrigation zone The earlier value is selected to ensure the switching boundary is traceable in engineering. After a switching event occurs in an adjacent irrigation zone, the system compares the normalized concentration trajectory c(t) before and after the switching event. Specifically, the average value of the normalized concentration trajectory c(t) over multiple consecutive sampling periods before the switching event is used as the reference value before the switching event. When the normalized concentration trajectory c(t) continuously increases relative to the reference value before the switching event after the switching event, and the increase is not less than 0.1 and is maintained for at least two consecutive sampling periods, the moment when the above condition is first met is determined as the starting moment of the candidate interval for the residual impact segment. After the normalized concentration trajectory c(t) reaches its peak value, when the difference between it and the reference value before the switching event is not greater than 0.05 and is maintained for at least two consecutive sampling periods, the moment when this condition is first met is determined as the ending moment of the candidate interval for the residual impact segment. The values ​​of 0.1 and 0.05 are exemplary values ​​and can be adjusted according to the normal fluctuation range of the main conductivity. The same determination standard is used within the same scheduling cycle.

[0027] The main flow rate Q(t) at each sampling time within the candidate interval of the residual impact section is accumulated according to the sampling period Δt. The main flow rate at each sampling time is added to the water volume corresponding to the sampling period in sequence to obtain the propulsion water volume within the candidate interval of the residual impact section. The propulsion water volume is then used as the calculated value of the equivalent retention volume of the main flow rate corresponding to this switching event.

[0028] For multiple consecutive switching events, the corresponding equivalent retention volume of the main control is calculated, and the equivalent retention volume of the main control is then... A recursive update is performed. Specifically, during each recursive update, the pre-update equivalent retention volume of the supervisor is retained according to the first weight, and the calculated value of the equivalent retention volume of the supervisor obtained from the current switching event is added according to the second weight. The two are then added together to obtain the updated equivalent retention volume of the supervisor. The sum of the first weight and the second weight is 1. The second weight is used as an update coefficient, preferably between 0.2 and 0.5. The update coefficient is decreased when the data availability corresponding to the current switching event decreases or when the main flow or main pressure fluctuates more, and the update coefficient is increased when the calculated value of the main equivalent retention volume obtained from multiple consecutive switching events remains stable.

[0029] If the current switching event does not form a complete candidate interval of residual impact segment that meets the above conditions, or if the corresponding dynamic identifiable confidence level of fertilizer injection is lower than a preset threshold. When this happens, the equivalent retention volume of the main pipe is not updated using the calculated value of the main pipe equivalent retention volume for that time. The reliability of the most recent fertilization dynamic identification should not be lower than the preset threshold. The equivalent retention volume of the supervisor obtained at that time This avoids abnormal changes in the equivalent retention volume of the main pipe caused by a single abnormal switch, instantaneous water hammer, or fluctuations in the main pipe flow rate.

[0030] In obtaining the equivalent retention volume of the supervisor Then, starting from the reference time when the adjacent irrigation zone switch occurs, the propulsion flow volume corresponding to the main flow rate Q(t) is accumulated. When the accumulated propulsion flow volume first reaches the equivalent retention volume of the main flow, the propulsion flow volume is calculated. The duration from the reference time to that time is then determined as the residual propulsion duration. Through the equivalent retention volume of the main body. and residual propulsion time Determine the required propulsion volume and corresponding time scale for the subsequent switching buffer section, so that the residual fertilizer solution in the main pipe is preferentially propelled by clean water in the switching buffer section.

[0031] In identifying the mixing dynamics, the system focuses on the temporal coupling relationship between the start-up and shutdown of the fertilizer pump and the concentration response in the main pipe. Since the fertilizer solution does not reach stable mixing instantaneously after the pump starts, the concentration in the main pipe typically exhibits a gradual increase from low to high. To quantify this process, this step utilizes the normalized concentration trajectory formed in step one. The start time of the fertilizer injection pump For reference, during the upward phase after fertilizer injection begins... By performing dynamic fitting, the mixing time constant, which characterizes the speed of the mixing response, is obtained. Preferably, a first-order dynamic approximation can be used to describe the rising segment: ; in, This indicates the moment when the fertilizer injection pump switches from stopped to running. Indicates the mixing time constant. This represents the normalized concentration trajectory. When... A larger value indicates that it will take longer to reach a stable concentration range after fertilizer injection begins; when A smaller value indicates faster mixing and earlier attainment of steady state. During the fitting process, the system simultaneously calculates the fitting residual and the steady-state entry time to determine when to begin including the effective fertilizer dose in the steady-state fertilization phase. This avoids miscounting the mixing transition phase as effective fertilization, preventing deviations in the calculation of zonal fertilization fulfillment.

[0032] After completion and Following the identification, this step further generates a comprehensive credibility measure of identifiable dynamic fertilizer injection. This is used to characterize whether the identification driven by this switching event is usable and interpretable, and serves as a gating quantity for subsequent segmented calculation and execution compensation. The comprehensive quantity of fertilizer injection dynamic identifiability reliability... It is not a single sensor indicator, but rather based on the main flow stability quantile. , Supervisory pressure stability quanta The response of the normalized concentration trajectory can be piecewise quantum. Switching event consistency subquantity and data availability subquantity At least three of the above are determined comprehensively. Preferably, It can be determined in the following form: ; Among them, the main flow stability subquantity The calculation method is as follows: A time window is set before and after the switching event between adjacent irrigation zones, and the coefficient of variation of the main flow sequence Q(t) within this window is calculated, i.e.: ;in, The standard deviation of the flow series within this window. This represents the average value of the flow sequence within this window.

[0033] Supervisory pressure stability quantile The calculation method is as follows: within the switching window, calculate the coefficient of variation of the main pressure sequence P(t), that is: ;in, The standard deviation of the pressure series within this window. This represents the average value of the pressure sequence within this window.

[0034] The response of the normalized concentration trajectory can be piecewise quantum. The determination includes: peak detection of conductivity EC(t) or normalized concentration trajectory c(t) within the switching window; when the normalized concentration trajectory c(t) forms a peak, the increase of the peak relative to the reference value before switching is not less than 0.1, and there is a continuous rising process before the peak and a continuous falling process after the peak, it is determined that there exists a peak that can be used for segmentation, and the peak detection sub-quantity is set as follows. Set the value to 1; when there is no peak value as described above, there are two or more peak values ​​that satisfy the above conditions, or the fallback process after the peak value is not completed, set the peak detection subquantity to 1. It is 0.

[0035] A first-order dynamic fit is performed on the rising segment of the conductivity or normalized concentration curve after the switch to obtain the sum of squared fitting residuals. And calculate the goodness-of-fit quantum. ,in Let be the variance of the conductivity sequence within this window; take the average of the two as the piecewise quantum quantity of the EC response, i.e.: ; Switching event consistency subquant The calculation method is as follows: based on the arrival delay and switch jitter of the solenoid valve opening and closing receipts for the irrigation zone, if there is no delay or jitter in the receipt arrival, then... If the delay is within the allowable range and the number of jitters is less than the set threshold, then ;otherwise The allowable range can be within 2 seconds, and the threshold for the number of jitters can be 2.

[0036] Data availability sub-quantity The calculation method is as follows: Based on a comprehensive judgment of the missing measurement rate of conductivity, flow rate, pressure sequence, time alignment residual, and sensor saturation status, if there are no missing measurements, the time alignment residual is less than the sampling period, and there is no sensor saturation, then... If the missing rate is less than 5%, the time alignment residual is less than twice the sampling period, and the saturation time ratio is less than 5%, then ;otherwise .

[0037] The weight to The method for determining this is as follows: it is obtained through training with historical data, or it is set based on expert experience, and it meets the following conditions. For example, it is advisable to , , , , .

[0038] The The function is defined as: if If, then return 0; if If the condition is met, return 1; otherwise, return x.

[0039] By integrating multi-source information The system can proactively reduce its reliance on the current identification results when there are jitters, drastic fluctuations in traffic pressure, or when the EC curve is not segmentable, thus avoiding the reliance on unreliable results. and This introduces a secondary error for subsequent segmented calculations and compensation.

[0040] The output of this step is the equivalent retention volume of the main tube. Mixing time constant and the comprehensive reliability of fertilizer injection dynamic identification and will With preset threshold Comparisons are made to form gating decisions, for example .when Not less than At that time, the system allows and This serves as the basis for the constraints of the generation process in step three and the segmented calculation in step four; when Below At this point, the system marks the identification result for this cycle as low confidence and enters a conservative strategy, allowing subsequent scheduling to prioritize longer switching buffers and end-point compensation to suppress the risk of fertilizer-liquid cross-regional application. Simultaneously, the data from this cycle is retained for parameter updates when conditions improve. In this way, this step achieves a closed-loop pre-processing of online identification and reliable gating in engineering, providing a stable, interpretable, and continuously iterative foundation for subsequent zoned fertilization timing scheduling and zoned effective fertilization fulfillment accounting.

[0041] Step 3: Gating and timing scheduling; The equivalent retention volume of the main tube output in step two is used in this step. Mixing time constant and the comprehensive reliability of fertilizer injection dynamic identification Based on this, the processing targets the fertilization execution process of each irrigation zone within the scheduling cycle, and the inputs include the target fertilization dose for each irrigation zone. Planned irrigation duration Or planned irrigation water volume The pump station's water supply capacity parameters and the controllable rotational irrigation sequence constraints of the irrigation zone's solenoid valves are considered. The pump station's water supply capacity parameters include the upper limit of the main flow rate, the lower limit of the main pressure, and the maximum fertilizer injection capacity of the fertilizer injection pump. The core processing mechanism of this step is to extend the total fertilizer dosage for each zone from static allocation to executable control of the zone's fertilizer injection timing and concentration trajectory. Furthermore, the switching buffer segment, steady-state fertilizer injection segment, and end-compensation segment required to suppress fertilizer liquid cross-region switching are solidified as part of the scheduling result, enabling the scheduling output to be directly issued and executed by the PLC controller and consistent with subsequent closed-loop calculations.

[0042] Because in the zoned irrigation mode with a single fertilizer injection pump and mixing tank, residual fertilizer solution in the main pipe will be carried into the next irrigation zone when switching zones. This step first depends on... The basic size of the switching buffer section is determined so that after the switch, a short flow volume is initially propelled with clean water, thereby absorbing as much residual shock as possible within the buffer section and reducing the risk of high EC shock at the beginning of the next irrigation zone. Specifically, the system sets a switching buffer duration for each zone switch. This ensures that the cumulative water volume propelled within the buffer section after the switch is not less than [amount missing]. Preferred to satisfy: ; in, The opening time of the solenoid valve for irrigation zone i. Let Q(t) be the switching buffer time for irrigation zone i, and Q(t) be the main flow rate. The equivalent retention volume of the main pipe identified in step two. During engineering implementation, the system uses a representative flow rate under the current operating conditions of the pumping station. Approximate Q(t), thus making Take as In addition, a safety margin is added to ensure that the buffer section is feasible and takes into account on-site flow fluctuations.

[0043] After switching the buffer section, this step determines the starting conditions of the steady-state fertilization section based on the mixing time constant τ_m. This avoids mistakenly including the mixing transition section after fertilization initiation in the effective fertilization, thus preventing the combined problem of incomplete fertilization in the previous section and fertilizer shock in the next section. Specifically, the system sets a steady-state entry threshold η, which is used to determine whether to enter the steady-state fertilization section. For example, η = 0.95, and is determined by... Calculate the steady-state arrival time This makes the normalized concentration trajectory after fertilizer injection start-up... To achieve a stable level, the following conditions should be met: ; in, For the steady-state arrival time, Here, η is the mixing time constant, and η is the steady-state entry threshold. When η takes a larger value, it indicates a stricter requirement that the main concentration must approach the steady state before the steady-state fertilization phase begins to be calculated. Through this setting, the system clearly defines the effective metering starting point of the steady-state fertilization phase at the scheduling level, ensuring a consistent standard for subsequent effective fertilization dosage calculations.

[0044] Under the above constraints, the system generates a fertilization timing structure for each irrigation zone i, consisting of a switching buffer section, a steady-state fertilization section, and an end-compensation section. Within the switching buffer section, the system keeps the fertilization pump off or at a low output, using only clean water to propel the residual fertilizer in the pipe. Within the steady-state fertilization section, the system starts the fertilization pump and controls the main pipe conductivity within the target range, allowing this section to handle the primary fertilization. The end-compensation section continues to propagate fertilizer with clean water after the fertilization pump stops, ensuring that the fertilizer solution at the end of the zone reaches the target root zone while reducing the probability of residual fertilizer entering the next irrigation zone. The minimum length of the end-compensation section... Similarly, it can be used with In this regard, it is preferable to make the volume of the compensation section cover the main part of the equivalent retention volume of the main pipe, thereby transferring the risk of the series area into a controllable flushing volume, so that the scheduling has an adjustable trade-off space between fertilizer saving and steady-state risk.

[0045] Equivalent retention volume of the supervisor Used to determine the basic propulsion volume corresponding to the switching buffer section and the end compensation section, and to determine the corresponding basic duration in combination with the main flow rate; mixing time constant Used to characterize the mixed dynamics of the current fertilization process, the switching buffer duration is adjusted based on the basic duration. and end-of-pipe compensation duration Make corrections. When the mixing time constant... An increase in this value indicates that the time required for the fertilizer solution in the main pipe to reach stable mixing has increased, and the switching buffer time should be increased accordingly. or end-of-pipe compensation duration Safety margin; when the mixing time constant When decreasing, according to the equivalent retention volume of the main tube. The corresponding basic propulsion volume determines the duration of the switching buffer section and the end compensation section. This, in turn, determines the equivalent retention volume of the main pipe. Characterization of residual propulsion process and mixing time constant The hybrid dynamics of the representation are used together to define the zonal fertilization execution process.

[0046] like Figure 2As shown, the switching buffer section is used to consume residual shock, the steady-state fertilization section is used to measure effective fertilization, and the end compensation section is used to advance the tail end to the correct position and reduce the probability of cross-regional application.

[0047] Under the premise of meeting process constraints, this step will determine the target fertilizer dosage for each irrigation zone. This is mapped to the target application intensity and duration of the steady-state fertilization phase. Preferably, the system utilizes a normalized concentration trajectory. Construct the planned effective fertilizer dosage for the steady-state fertilization section using the main flow rate Q(t), and then convert the dosage using a conversion factor. The integral of flow rate and concentration is converted into the effective dosage of fertilizer solution, thus extending the duration of the steady-state fertilization phase. satisfy: ; in, The target fertilizer dosage for irrigation zone i, This is a conversion factor from electrical conductivity to the effective dosage of fertilizer solution, for example... Q(t) represents the main flow rate. For the normalized concentration trajectory, This refers to the steady-state fertilizer injection period duration. For the project scheduling generation phase, the system can correlate Q(t) with... Approximating it with its steady-state representative value in order to quickly obtain an executable value. At the same time, space is reserved for the realization calculation and deviation compensation based on the actual EC curve and flow receipt in step four.

[0048] To ensure that the aforementioned switching buffer section, steady-state fertilization section, and end-compensation section form a unified zonal fertilization timing scheduling scheme under the constraints of pump station water supply capacity and irrigation rotation sequence, this step further constructs a multi-objective optimization model. The adjustment objects of the multi-objective optimization model include the irrigation rotation sequence of irrigation zones, the opening and closing times of the solenoid valves in each irrigation zone, the start and stop times of the fertilization pumps, and the corresponding switching buffer duration for each irrigation zone. Steady-state fertilizer injection period duration and end-of-pipe compensation duration The irrigation rotation sequence for each irrigation zone can only be adjusted within the range that satisfies the rotation sequence constraints.

[0049] For any zonal fertilization timing scheme that satisfies the rotation irrigation sequence constraint, the first step is to determine the equivalent retention volume of the main pipe. The risk of fertilizer solution cross-contamination is determined by the planned main flow rate within the switching buffer zone. Specifically, starting from the opening time of the irrigation zone solenoid valve, the risk is determined based on the switching buffer duration. The propulsion water passage volume within the cumulative switching buffer section is then compared with the equivalent retention volume of the main pipe. Comparison. When the propulsion flow volume within the switching buffer section is not less than the equivalent retention volume of the main pipe. At that time, it is determined that the risk of fertilizer-liquid cross-contamination in the irrigation zone under the current fertilization timing scheme is at its lowest level; when the volume of water flowing through the switching buffer section is less than the equivalent retention volume of the main pipe. At that time, based on the difference between the two relative to the equivalent retention volume of the main body. The proportion of fertilizer-liquid cross-contamination is used to determine the risk of fertilizer-liquid cross-contamination; the higher the proportion, the higher the risk. The fertilizer-liquid cross-contamination risks for each irrigation zone are then combined to obtain the fertilizer-liquid cross-contamination risk corresponding to the current fertilization timing schedule for that zone.

[0050] Meanwhile, based on the duration of the steady-state fertilization period corresponding to each irrigation zone... The planned effective fertilizer dosage is determined by the target application intensity, planned main flow rate, and steady-state representative value of the normalized concentration trajectory during the steady-state fertilization phase. Specifically, within the time range corresponding to the steady-state fertilization phase, the fertilizer application rate represented by the planned main flow rate and the steady-state representative value of the normalized concentration trajectory is converted using a conversion factor kd and accumulated over the duration of the steady-state fertilization phase to obtain the planned effective fertilizer dosage for the corresponding irrigation zone. This planned effective fertilizer dosage is then compared with the target fertilizer dosage for that irrigation zone. The comparison is made based on the absolute value of the difference between the two values ​​relative to the target fertilizer dosage. The proportion determines the corresponding risk of deviation in the effective fertilization implementation of each irrigated zone; the higher the proportion, the higher the risk of deviation. The risk of deviation in the effective fertilization implementation of each irrigated zone is then aggregated to obtain the risk of deviation in the effective fertilization implementation corresponding to the current irrigated zone's fertilization timing schedule.

[0051] The multi-objective optimization model aims to simultaneously reduce the risk of fertilizer cross-contamination and the risk of deviation in the effective fertilization implementation of each zone. It ensures that the fertilization timing scheme for each candidate zone meets at least the following conditions: main pipeline flow rate does not exceed the upper limit, main pipeline pressure is not lower than the lower limit, fertilization pump operation does not exceed the maximum fertilization capacity, and the switching buffer time for each irrigation zone is [not specified]. Steady-state fertilizer injection period duration and end-of-pipe compensation duration All values ​​are non-negative, and the solenoid valves of each irrigation zone open and close according to the allowed rotational irrigation sequence.

[0052] In engineering implementation, the sampling period Δt is used as the switching buffer duration. Steady-state fertilizer injection period duration and end-of-pipe compensation duration The minimum adjustment step size is used to generate multiple candidate fertilization timing schemes for different zones within the range that meets the constraints of pump station water supply capacity and irrigation sequence. The fertilizer liquid cross-zone risk and the zone effective fertilization fulfillment deviation risk are determined for each candidate fertilization timing scheme.

[0053] Weights are assigned to the risks of fertilizer-solvent cross-regional application and the risk of deviation in effective fertilization delivery within a given region. The two risks are then weighted according to their respective weights and summed to obtain a comprehensive evaluation result of the candidate region's fertilization timing scheduling scheme. The sum of the two weights is 1. Preferably, the two risks have the same weight. For irrigation regions more sensitive to transient high conductivity, the weight corresponding to the risk of fertilizer-solvent cross-regional application is increased; for irrigation regions with high requirements for the fulfillment of target fertilization doses, the weight corresponding to the risk of deviation in effective fertilization delivery within the region is increased.

[0054] The system selects the candidate fertilization timing scheme with the smallest comprehensive evaluation result from the candidate fertilization timing schemes that meet the constraints of pump station water supply capacity and irrigation sequence as the fertilization timing scheme for this scheduling cycle; when the comprehensive evaluation results of two or more candidate fertilization timing schemes are the same, the candidate fertilization timing scheme with the shorter total execution time is selected.

[0055] Using the above method, the multi-objective optimization model, before the actual implementation of zoned fertilization, determines the equivalent retention volume of the main pipeline. Mixing time constant The first step involves pre-evaluating the risks of fertilizer-liquid cross-regional application and deviations in effective fertilization delivery for different zonal fertilization timing schemes based on target fertilizer dosage and pump station water supply capacity parameters. This determines the actual zonal fertilization timing scheme to be implemented. The fourth step calculates the effective fertilizer dosage and fertilizer-liquid impact area based on the mains conductivity EC(t), mains flow rate Q(t), and normalized concentration trajectory c(t) obtained during the actual execution of the zonal fertilization timing scheme. This is used to determine whether online compensation is needed for the actual execution results, thus distinguishing the risk assessment during the scheduling generation stage from the actual calculation after execution.

[0056] To ensure the stable implementation of this scheduling process under different data quality conditions, this step uses the comprehensive metric of identifiable reliability of fertilizer injection dynamics. As a gating condition. When Not lower than the preset threshold At that time, the system considers this period to be... and It is available, therefore based on the above. and Derived switching buffer duration Steady-state arrival time With end-of-line compensation duration Generate a refined zonal fertilization timing scheduling scheme and output a control sequence that can be directly sent to the PLC controller, including the irrigation zone rotation sequence, the opening and closing times of the solenoid valves in each zone, the start and stop times of the fertilization pumps, and the duration of each segment; when Below When the system considers the identification result of this cycle as low reliability, in order to avoid erroneous compensation or incorrect segmentation under unreliable identification conditions, the system enters a conservative scheduling mode. This mode uses a preset conservative switching buffer duration and conservative end-compensation duration, and reduces the frequency of partition switching or limits the number of consecutive switching. This prioritizes the suppression of fertilizer-liquid cross-region risks at the execution layer, while retaining the data from this operation for updates in subsequent cycles. and .

[0057] Through the above method, the zonal fertilization timing scheme output in this step is no longer simply about allocating a certain amount of fertilizer to each irrigation zone, but rather an executable control sequence with a basic structure of switching buffer sections, steady-state fertilization sections, and end-of-pipe compensation sections. Furthermore, this sequence is subject to… The gating system can operate stably in a real orchard environment with fluctuating data quality, providing a consistent and traceable execution basis for the calculation of effective fertilization dosage and closed-loop correction in step four.

[0058] Step 4: Segmented calculation of recharge; This step focuses on the actual execution process of zoned rotational irrigation. The object of processing is the zoned fertilizer injection timing schedule plan issued in step three and executed by the PLC controller. Inputs include the main conductor conductivity EC(t) and normalized concentration trajectory collected during execution. Main flow rate Q(t), main pressure P(t), fertilizer injection pump operating rate Irrigation zone solenoid valve opening / closing receipt ( ), and the equivalent retention volume of the main tube output in step two. Mixing time constant Combined with the identifiable reliability of fertilizer injection dynamics The core of this step lies in: When the gating conditions are met, use and The segmented approach provides interpretable calculations of the fertilization process for each irrigation zone, yielding the actual effective fertilization dose and switching impact index for that irrigation zone. Furthermore, it constructs a comprehensive risk measure for the effective fertilization fulfillment deviation of the zone, thereby driving online compensation or backfilling, ultimately forming a traceable closed-loop scheduling execution result.

[0059] In actual orchard operations, during the instant of zone switching, residual fertilizer solution in the main pipe is propelled into the next irrigation zone by clean water. If the fertilization status is judged solely based on the fertilizer pump setting or the zone setting duration, the residual impact can easily be miscalculated as effective fertilization, leading to inaccurate accounting. To avoid this type of error, this step first uses a comprehensive measure of the identifiable reliability of fertilization dynamics. As an accounting threshold condition, when Below the preset threshold When the system determines that the current handover event and EC response lack stable and segmentable characteristics, it will not perform fine-grained segmentation calculation and online compensation for this cycle, but will instead adopt a conservative approach: only recording anomalies and deficit trends, and prioritizing the increase of handover buffer and end-of-cycle compensation safety margins in the next scheduling cycle to reduce cross-region risk; Not less than At that time, the system believes and It is available and allows for the development of based on and The segmented calculation ensures a consistent standard for evaluating the effectiveness of fertilization. Therefore, subsequent calculations of the effective dose calculation window, impact window, and comprehensive risk all rely on the output identified in step two; without this, the assessment is incomplete. , and Therefore, it is impossible to reliably define the boundary between the residual impact segment and the steady-state fertilization segment, and thus it is impossible to obtain an effective fertilization dose that can be used for closed-loop compensation.

[0060] When satisfied After the gating condition is met, the system first issues a confirmation message indicating that the irrigation zone solenoid valve has opened. As a reference starting point for residual propulsion, and based on the equivalent retention volume of the main tube. The residual propulsion time is determined by the main flow rate Q(t) after the switch. Specifically, starting from the moment the solenoid valve of irrigation zone i opens, the propulsion water volume corresponding to the main flow rate Q(t) is accumulated. When the accumulated propulsion water volume first reaches the equivalent retention volume of the main flow rate... The residual propulsion time will be determined from the moment the solenoid valve of irrigation zone i is opened to that moment. .

[0061] Subsequently, the start-up and stop times of the fertilizer injection pumps corresponding to irrigation zone i are determined based on the fertilizer injection pump operation rate uf(t), and the steady-state arrival time is determined based on the mixing time constant τm. Specifically, starting from the moment the fertilizer injection pump starts, the time required to reach the steady-state threshold η is determined based on the first-order dynamic approximation relationship of the normalized concentration trajectory c(t) in step two, and this time is defined as the steady-state arrival time. .

[0062] The system determines the end time of residual propulsion and the steady-state arrival time after the fertilizer injection pump starts. The end time of residual propulsion is the opening time of the irrigation zone solenoid valve and the duration of residual propulsion. The sum of the steady-state arrival time after the fertilizer injection pump starts is the sum of the fertilizer injection pump start-up time and the steady-state arrival time. The sum. Compare the two times mentioned above, and use the later time as the steady-state fertilizer injection calculation window. The start time is used, and the stop time of the fertilizer injection pump is used as the steady-state fertilizer injection calculation window. The end moment.

[0063] Through the above methods, the residual impact phase will not be included in the effective fertilizer dosage before the residual fertilizer solution in the main pipe completes its main propulsion. The mixing transition phase after the fertilizer pump starts will also not be included in the effective fertilizer dosage before the concentration in the main pipe reaches stability. At the same time, the end compensation phase after the fertilizer pump stops will be distinguished from the steady-state fertilizer injection phase, so that the switching buffer phase, steady-state fertilizer injection phase and end compensation phase formed in step three will adopt the same time boundary as the effective fertilizer dosage calculation in step four.

[0064] In steady-state fertilizer injection accounting window Within the system, the effective fertilizer dosage is calculated based on the main flow rate and normalized concentration trajectory to obtain the effective fertilizer dosage for irrigation zone i. The formula for its calculation is: ; in, Effective fertilizer dosage for irrigation zone i Here, Q(t) is the conversion factor from electrical conductivity to effective fertilizer dosage, and Q(t) is the main flow rate. For the normalized concentration trajectory, This is the steady-state fertilizer injection accounting window. The formula can be pre-calibrated using fertilizer formulation, mixing ratio, and EC calibration relationship. This formula is used to convert the integral of flow rate and normalized concentration into an effective fertilizer dosage that can be used for agricultural management. Through this calculation, the system can obtain the actual effective fertilizer dosage obtained for each irrigation zone based on the execution receipt and process curve, thereby distinguishing between the set fertilizer injection amount and the actual effective fertilizer application.

[0065] To quantify the impact of fertilizer and liquid treatment on irrigation zone switching and identify the high EC impact risk at the beginning of the next irrigation zone, the system further calculates the fertilizer and liquid treatment impact area within the residual impact window. The preferred method is determined by the following formula: ; in, Let EC(t) be the fertilizer solution impact area of ​​irrigation zone i, and EC(t) be the measured value of the main pipe conductivity. The impact threshold conductivity is used to characterize the upper limit of acceptable transient conductivity for this irrigation zone. It is determined by taking the baseline conductivity of clean water. With fertilizer injection steady-state conductivity The weighted average, i.e. This ensures that the impact threshold conductivity ECth is located at the clear water baseline conductivity. With fertilizer injection steady-state conductivity The 0.5 is an intermediate level between these values. The 0.5 is an exemplary coefficient and can be adjusted based on the age of the citrus tree, soil conditions, and the corresponding irrigation zone's tolerance to transient high conductivity. Used to accumulate only the impact contribution exceeding the threshold. (By...) The system transforms the impact of fertilizer solution from a subjective description into comparable quantitative indicators, enabling subsequent closed-loop control to take clear measures to address the impact risk.

[0066] In obtaining and This step then generates a comprehensive risk assessment of deviations in the effective fertilization implementation for each zone. This is used to uniformly characterize two types of negative impacts: whether fertilization is complete and whether the impact is too strong, and to correlate them with zoning sensitivity and identification reliability, thereby achieving a controllable closed-loop compensation strategy in engineering. Preferably, Determine by the following formula: ; in, The comprehensive risk of effective fertilization fulfillment deviation in irrigation zone i; The target fertilizer dosage for irrigation zone i; This is the zoning sensitivity coefficient, used to reflect the differences in the sensitivity of sapling areas, salt-sensitive areas, or low-permeability soil areas to shocks and deviations. It is set according to the crop type, tree age, and soil characteristics within the zoning. For example, it is set to 1.5 for sapling areas, salt-sensitive areas, or low-permeability soil areas, and to 1.0 for mature tree areas, salt-tolerant areas, or high-permeability soil areas. and These are weighting coefficients used to balance the contributions of the realization deviation term and the impact term. They are determined through historical data or set based on expert experience, for example... , ; This is a normalization constant for impact area, used for dimensional uniformity under different working conditions. It is taken as the average of the historical maximum impact area, or set as a fixed value based on engineering experience, for example... ; For example, a credibility penalty coefficient. ; This is the comprehensive identifiable reliability measure of fertilizer injection dynamics output from step two. This is achieved by introducing [a certain factor] into the comprehensive risk measure. When the identification conditions weaken but are still usable, the system automatically increases the risk weight, making the compensation strategy more conservative, thereby reducing erroneous compensation caused by segment uncertainty.

[0067] To determine whether the effective fertilizer dosage for an irrigation zone has reached the allowable range of the target fertilizer dosage, the system uses the target fertilizer dosage for that irrigation zone. Determine the lower limit of the target fertilizer dosage. Preferably, the target fertilizer dosage... 90% is determined as the lower limit of the target fertilizer dosage, when the effective fertilizer dosage Below the target fertilizer dosage When 90% of the target fertilizer application rate is reached, it is determined that there is insufficient effective fertilization in that irrigation zone. This 90% is an exemplary percentage and can be adjusted based on the age of the citrus trees, soil conditions, and allowable deviations in the target fertilizer dosage, while remaining constant within the same scheduling cycle.

[0068] when Exceeding the preset risk threshold Or when When the fertilizer dosage falls below the lower limit of the target allowable dosage, the system will prioritize triggering online compensation before the execution window closes, based on the risk threshold. Based on engineering experience, for example For cases of insufficient delivery, the system, while ensuring the lower limit of pump station pressure and the capacity constraints of the fertilizer injection pump, extends the steady-state fertilizer injection period corresponding to the steady-state fertilizer injection calculation window, or extends the end compensation period to advance the delivery of fertilizer solution at the tail end, so that... Returning to the acceptable range; for cases of excessive impact, the system prioritizes extending the switching buffer or reducing the partition switching frequency, ensuring that residual impacts mainly occur within the buffer and are reduced. .

[0069] When online compensation cannot be completed within the current scheduling cycle due to time window limitations, pump station capacity constraints, or zonal irrigation rotation sequence limitations, the system determines the fertilizer backlog for the corresponding irrigation zone. This information is then written into the recharge record for correction in the next scheduling cycle. Specifically, when the effective fertilizer dose... Below the target fertilizer dosage At that time, apply the target fertilizer dose. With effective fertilizer dosage The difference between them is used as the amount of fertilizer owed. When the effective fertilizer dosage No less than the target fertilizer dosage At that time, the amount of fertilizer owed will be... The value is determined to be 0.

[0070] When generating the target fertilizer dose in the next scheduling cycle, the corresponding irrigated zone's fertilizer deficit will be considered. When the value is greater than 0, the amount of fertilizer owed will be... As a correction to the target fertilizer dosage for this irrigation zone, any unfulfilled fertilizer dosage in this scheduling cycle will be compensated for in subsequent scheduling cycles; when the fertilizer shortfall... When the target fertilizer dose is 0, it will not be increased based on the amount of fertilizer deficit for that irrigation zone. The revised target fertilizer dose will continue to be used to generate the zonal fertilization timing schedule for the next scheduling cycle, and will continue to meet the pump station water supply capacity constraints and rotation irrigation sequence constraints. Simultaneously, the execution data for this scheduling cycle will be used to update the equivalent retention volume of the main pipe. Mixing time constant And the comprehensive reliability of fertilizer injection dynamic identification This allows the identification results used in subsequent scheduling cycles to be continuously corrected as the pipeline status, flow conditions, and fertilizer injection response change.

[0071] Based on the above calculations and closed-loop compensation, this step outputs the effective fertilizer dosage for each irrigation zone. Fertilizer liquid impact area Comprehensive risk assessment of deviations in effective fertilization implementation across zones Online compensation execution receipt, fertilizer arrears It also records the backfilling and outputs the updated data synchronously. , and Therefore, the system can not only suppress fertilizer and liquid cross-regional flow caused by zonal switching under zonal rotation irrigation conditions, but also calculate, correct, and trace the actual fertilizer received by each irrigation zone in a closed-loop process data manner, so that the scheduling results can be realized and continuously optimized in the project execution.

[0072] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0073] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0074] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0075] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0076] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A multi-objective optimization-based method for integrated water and fertilizer resource scheduling in citrus orchards, characterized in that: Includes the following steps: Triggered by the switching event of adjacent irrigation zones, the equivalent retention volume of the main pipe is determined based on the process data set and the normalized concentration trajectory; triggered by the start-stop event of the fertilizer injection pump, the mixing time constant is determined based on the rising segment of the normalized concentration trajectory after the start of the fertilizer injection pump; and the comprehensive quantity of identifiable reliability of fertilizer injection dynamics is determined based on the process data set and the normalized concentration trajectory. When the overall reliability of the dynamic identification of fertilization is not lower than the preset threshold, the fertilization execution process is planned into a switching buffer section, a steady-state fertilization section and an end compensation section based on the equivalent retention volume of the main pipe, the mixing time constant and the target fertilization dose of each irrigation zone, and a candidate zone fertilization timing scheduling scheme is generated. The risk of fertilizer cross-contamination is determined based on the volume of water flowing through the switching buffer zone and the equivalent retention volume of the main pipe. The risk of deviation in the effective fertilization rate of each zone is determined based on the deviation between the planned effective fertilization dose and the target fertilization dose. Based on the risks of fertilizer cross-contamination and the risk of deviation in the effective fertilization rate of each zone, the fertilization timing scheduling scheme for each zone is determined, and the fertilization pump and the solenoid valves of each irrigation zone are controlled.

2. The method for integrated water and fertilizer resource scheduling in citrus orchards based on multi-objective optimization according to claim 1, characterized in that, The determination of the equivalent retention volume of the main pipe based on the process data set and the normalized concentration trajectory includes: The process variable sampling sequence, fertilizer injection pump operation volume, and opening / closing receipts of solenoid valves in each irrigation zone are acquired to form a process data set. The process variable sampling sequence includes at least one of the main pipe conductivity sampling sequence, main pipe flow rate sampling sequence, and main pipe pressure sampling sequence. The sampling sequences are aligned with a unified time standard, and the irrigation zones are divided into rotational irrigation segments based on the opening / closing receipts. The baseline conductivity of clear water and steady-state conductivity of fertilizer injection are determined based on the main pipe conductivity sampling sequence to obtain the normalized concentration trajectory.

3. The method for integrated water and fertilizer resource scheduling in citrus orchards based on multi-objective optimization according to claim 2, characterized in that, Determining the equivalent retention volume of the main tube includes: Within the corresponding time window following the switching event of adjacent irrigation zones, candidate intervals for residual impact segments are determined based on the sudden rise and fall of the normalized concentration trajectory; the sampling sequence of main flow in the candidate intervals for residual impact segments is numerically accumulated to obtain the propulsion flow volume, which is then used as the calculated value of the equivalent retention volume of the main flow corresponding to the current switching event, and the equivalent retention volume of the main flow is recursively updated based on the calculated values ​​of the equivalent retention volume of the main flow corresponding to multiple switching events.

4. The method for integrated water and fertilizer resource scheduling in citrus orchards based on multi-objective optimization according to claim 3, characterized in that, The determination of the candidate interval for the residual impact segment includes: The average value of the normalized concentration trajectory within multiple consecutive sampling periods before the switching event is used as the reference value before the switching. When the normalized concentration trajectory after the switching event continues to rise relative to the reference value before the switching, and the rise is not less than 0.1 and is maintained for at least two consecutive sampling periods, the moment when the normalized concentration trajectory first simultaneously satisfies the conditions of continuous rise, rise not less than 0.1, and maintenance for at least two consecutive sampling periods is determined as the starting moment of the candidate interval for the residual impact segment. After the normalized concentration trajectory reaches its peak value, when the difference between it and the reference value before the switching is not greater than 0.05 and is maintained for at least two consecutive sampling periods, the moment when the difference is first simultaneously satisfied for at least two consecutive sampling periods is determined as the ending moment of the candidate interval for the residual impact segment.

5. The method for integrated water and fertilizer resource scheduling in citrus orchards based on multi-objective optimization according to claim 4, characterized in that, The recursive update of the supervisor's equivalent retention volume includes: The equivalent retention volume of the supervisor before the update is retained according to the first weight, and the calculated value of the equivalent retention volume of the supervisor corresponding to the current switching event is added according to the second weight to determine the updated equivalent retention volume of the supervisor. The sum of the first weight and the second weight is 1, and the second weight is used as the update coefficient. When the data availability corresponding to the current switching event decreases or the fluctuation of the supervisor flow and supervisor pressure increases, the update coefficient is reduced. When the calculated value of the equivalent retention volume of the supervisor corresponding to multiple consecutive switching events remains stable, the update coefficient is increased. When the current switching event has not formed a complete candidate interval of the residual impact segment or the comprehensive amount of the dynamic identifiable confidence of fertilizer injection is lower than the preset threshold, the calculated value of the equivalent retention volume of the supervisor corresponding to the current switching event is not used to update the equivalent retention volume of the supervisor.

6. The method for integrated water and fertilizer resource scheduling in citrus orchards based on multi-objective optimization according to claim 1, characterized in that, The determination of the mixing time constant includes: The normalized concentration trajectory rise segment after the fertilizer injection pump starts is dynamically fitted to obtain the mixing time constant; the steady-state arrival time is determined based on the mixing time constant, the effective metering start point of the steady-state fertilizer injection segment is defined by the steady-state arrival time, and the duration of the switching buffer segment and the end compensation segment is determined.

7. The method for integrated water and fertilizer resource scheduling in citrus orchards based on multi-objective optimization according to claim 6, characterized in that, The comprehensive measure for determining the identifiable reliability of fertilizer injection dynamics includes: The comprehensive quantity of identifiable reliability of fertilizer injection dynamics is determined based on at least three of the following: main flow stability, main pressure stability, response segmentability of normalized concentration trajectory, switching event consistency, and data availability. Among these, the switching event consistency is determined based on the arrival delay and opening / closing jitter of the solenoid valve opening / closing receipts for each irrigation zone, and the data availability is determined based on the missing measurement marker, time alignment residual, and sensor saturation status.

8. The method for integrated water and fertilizer resource scheduling in citrus orchards based on multi-objective optimization according to claim 7, characterized in that, Fertilizer injection scheduling is performed based on the comprehensive reliability of the dynamic identifiable fertilization injection, including: When the overall reliability of the dynamic identification of fertilizer injection is lower than the preset threshold, the identification result of this scheduling cycle is marked as low reliability. The preset conservative switching buffer time and conservative end compensation time are adopted, and the switching frequency of the zone is reduced or the number of consecutive switching is limited. When the overall reliability of the dynamic identification of fertilizer injection is not lower than the preset threshold, the fertilizer injection pump is kept in the off state or low output state in the switching buffer section, and the cumulative water volume in the switching buffer section is not less than the equivalent retention volume of the main pipe. After the steady state reaches the time, it enters the steady state fertilizer injection section. After the fertilizer injection pump stops, it enters the end compensation section and promotes the fertilizer solution at the end of this irrigation zone through clean water.

9. The method for integrated water and fertilizer resource scheduling in citrus orchards based on multi-objective optimization according to claim 8, characterized in that, The method for determining the zonal fertilizer injection timing schedule includes: Under the constraints of the main flow rate limit, the main pressure limit, the maximum fertilization capacity of the fertilization pump, and the irrigation rotation sequence, multiple candidate fertilization timing schemes for each irrigation zone are generated by adjusting the irrigation rotation sequence, the opening and closing times of the solenoid valves in each irrigation zone, the start and stop times of the fertilization pumps, and the duration of the switching buffer section, the steady-state fertilization section, and the end compensation section. Weights are assigned to the fertilizer solution cross-regional risk and the zone effective fertilization fulfillment deviation risk corresponding to each candidate fertilization timing scheme, and a weighted average is performed to obtain a comprehensive evaluation result. The candidate fertilization timing scheme with the smallest comprehensive evaluation result is selected as the fertilization timing scheme for each zone. When the comprehensive evaluation results of two or more candidate fertilization timing schemes are the same, the candidate fertilization timing scheme with the shorter total execution time is selected.

10. The method for integrated water and fertilizer resource scheduling in citrus orchards based on multi-objective optimization according to claim 9, characterized in that, After the partitioned fertilizer injection timing scheduling scheme is executed, the method further includes: When the overall reliability of dynamic fertilization injection is not lower than a preset threshold, a steady-state fertilization calculation window is determined based on the equivalent retention volume of the main pipe and the mixing time constant. Within the steady-state fertilization calculation window, the effective fertilization dose is calculated based on the main pipe flow rate and normalized concentration trajectory. Within the residual impact segment window, the fertilizer liquid impact area is determined based on the portion of the main pipe conductivity exceeding the impact threshold conductivity. The overall risk of effective fertilization fulfillment deviation in a given area is determined based on the target fertilization dose, effective fertilization dose, fertilizer liquid impact area, and overall reliability of dynamic fertilization injection. When the overall risk of effective fertilization fulfillment deviation in a given area exceeds a preset risk threshold or the effective fertilization dose is lower than the allowable lower limit, online compensation is performed. When online compensation cannot be completed due to time window, pump station capacity, or the order of rotational irrigation in a given area, the fertilization deficit is determined, and in the next scheduling cycle, the fertilization deficit is used as a correction item for the target fertilization dose of the corresponding irrigation area.

Images