Multi-source tracer data-driven water quantity optimization control method

Abstract

The present application relates to the technical field of ecological hydrological regulation and water quantity optimization control, and discloses a multi-source tracer data driven water quantity optimization control method, comprising: mapping a water dynamics parameter sequence into a state space vector, combining multi-source tracer analysis data to construct a low-frequency hydrological mechanism constraint, and calculating a soil negative pressure variation rate by derivation on soil negative pressure data; constructing an asynchronous decoupling control architecture containing a slow optimization loop and a fast blocking loop; collecting time-consuming parameters of the slow optimization loop, and if not overtime, outputting an ecological water supplement collaborative configuration scheme containing a target water supplement quantity, a water supplement time period and a water supplement intensity as a water quantity optimization control instruction, and if overtime, cutting off the loop and generating a transient blocking control instruction by searching a discrete mapping table according to the negative pressure variation rate; and outputting corresponding instructions to convert into adjustment signals to control the valve opening degree of the flow mechanism, wherein the present application relies on a double-time-scale decoupling and dimension reduction switching mechanism to eliminate deadlocks and oscillations caused by clock mismatch.

Description

Technical Field

[0001] This invention relates to a multi-source tracer data-driven water quantity optimization control method, belonging to the field of eco-hydrological regulation and water quantity optimization control technology. Background Technology

[0002] Currently, the industry generally adopts a closed-loop control method based on sensor state sampling. The industrial control computer treats the collected hydrological parameters as homogeneous inputs and constructs a feedback loop based on the uniform transport assumption to output regulation commands. However, when the controlled area encounters extreme precipitation pulses or other conditions, the objectively existing preferential flow channels in the soil pores will cause rapid water loss. At the same time, when introducing stable isotopes for water source analysis, existing technologies often directly mix isotope data with ordinary hydrodynamic sensor data, performing indiscriminate normalization or synchronous high-frequency iteration, which destroys the true isotope end-member ratio relationship between different water bodies such as river water, permafrost water, soil water, and plant water. The existing single-clock reference control architecture forcibly combines low-frequency hysteresis isotope parameters with high-frequency negative pressure disturbance parameters with frequency band gaps into the same nonlinear objective function for synchronous iteration. This not only fails to accurately reflect the ecological hydrological mechanism of replenishment-storage-utilization, but also causes the control loop to diverge in gradient optimization or get trapped in local optima when dealing with sudden preferential flows, causing the flow regulation hardware to miss the blocking window and induce ecological water loss.

[0003] To address the aforementioned phenomena, simply shortening the overall sampling period or increasing the control gain can easily transmit transient physical disturbances directly to the objective function, causing the regulating valve to fall into high-frequency mechanical chatter. At the control method level, for example, Chinese invention patent application CN122238615A discloses a method, system, equipment, and medium for estimating atmospheric chlorine deposition flux. This method extrapolates water transport by measuring the vertical distribution profile of tritium and chlorine in the vadose zone. This technique relies on the long-term historical accumulation of static equilibrium in the soil profile and belongs to a steady-state parameter inversion model. When faced with highly time-varying conditions such as soil priority flow activated by extreme precipitation pulses in ecological water quantity optimization scheduling, the model based on the historical steady-state preset premise has a fundamental mismatch with the actual dynamic high-frequency abrupt boundary conditions. It cannot capture the hidden danger of transient water loss, and lacks multi-band adaptive decoupling and real-time circuit breaking logic. This causes the control loop to get stuck in the quagmire of high-dimensional optimization when encountering strong external disturbances. The aforementioned defects reveal that traditional data-driven systems have fundamental architectural limitations when dealing with multi-source heterogeneous parameters, and cannot provide deterministic anti-disturbance guarantee for dynamic scheduling in complex time-varying environments.

[0004] Therefore, how to reconstruct the underlying clock scheduling and weight allocation logic of the optimization control system, and on the basis of stripping the frequency bands of hysteresis parameters and abrupt parameter response, establish a dimension reduction circuit breaker to prevent optimization divergence and a feedforward pre-correction network to resist mechanical dead zones, thereby eliminating the hidden dangers of system stall and regulation oscillation under extreme conditions, has become the technical problem to be solved by this invention. Summary of the Invention

[0005] To address the problems in the background art, the technical solution of the present invention is as follows: A water quantity optimization control method driven by multi-source tracer data, comprising the following steps: Step S1: Obtain multi-source tracer analysis data, hydrodynamic parameter sequence and soil negative pressure data, map the hydrodynamic parameter sequence into a state space vector, and construct low-frequency hydrological mechanism constraints in combination with multi-source tracer analysis data, and calculate the soil negative pressure variation rate by taking the derivative of the soil negative pressure data. Step S2: Construct an asynchronous decoupled control architecture that includes a slow benchmark optimization loop and a fast transient blocking loop, and deploy the optimization control model within the slow benchmark optimization loop; Step S3: Collect the optimization time parameter when the optimization control model calculates the state space vector through time monitoring logic, and compare the optimization time parameter with the preset closed-loop response time threshold. Step S4: When the optimization time parameter does not exceed the closed-loop response time threshold, the optimization control model is driven by the slow benchmark optimization loop to solve the state space vector, and the output ecological water replenishment collaborative configuration scheme containing the target water replenishment volume, water replenishment period and water replenishment intensity is used as the water quantity optimization control command. Step S5: When the optimization time parameter exceeds the closed-loop response time threshold, the slow benchmark optimization loop is cut off, and the preset discrete mapping table is retrieved by the soil negative pressure change rate through the fast transient blocking loop, and the matched discrete water volume control limit is used as the transient blocking control command. Step S6: Output water volume optimization control command or transient interruption control command, and convert it into a regulating signal to control the valve opening of the ecological flow control mechanism.

[0006] Preferably, the following steps are included between steps S1 and S2: Step S201, decompose the state space vector into a base current component and an impulse component; Step S202, calculate the information entropy of the impulse component; Step S203, when the information entropy exceeds a preset noise threshold, lower the response coefficient of the risk term in the optimization control model and raise the inertia coefficient of the water balance term.

[0007] Preferably, step S4, which uses a slow benchmark optimization loop to drive the optimization control model to solve the state space vector and outputs an ecological water replenishment coordination configuration scheme containing the target water replenishment amount, water replenishment period, and water replenishment intensity as a water quantity optimization control instruction, includes the following sub-steps: Step S41, performing end-member hybrid segmentation model calculations on multi-source tracer analysis data containing river water, permafrost water, soil water, and plant water to determine the contribution ratio and replenishment relationship of different water bodies within the controlled area; Step S42, determining the water body replenishment relationship within the controlled area based on a dual isotope mass balance model combined with hydrological observations; Step S43, within the preset ecological boundary constraints, using the contribution ratio and water body replenishment relationship as boundary conditions, solving a multivariate nonlinear objective function to obtain an ecological water replenishment coordination configuration scheme containing the target water replenishment amount, water replenishment period, and water replenishment intensity as a water quantity optimization control instruction.

[0008] Preferably, before step S1, the following steps are also included: step S401, recording the historical evolution data of the state space vector within the continuous control cycle; step S402, calculating the state change rate based on the historical evolution data; step S403, when the state change rate shows a continuous decrease and reaches a preset lower limit threshold, outputting a correction factor to the slow benchmark optimization loop and increasing the benchmark water replenishment in the water quantity optimization control command.

[0009] Preferably, the multi-source tracer analysis data includes hydrogen isotope deviation data and oxygen isotope deviation data of river water, permafrost water, soil water and plant water, and the hydrodynamic parameter sequence includes soil saturated hydraulic conductivity and surface runoff rate.

[0010] Preferably, ecological boundary constraints include lower limits for groundwater level and lower limits for soil moisture content.

[0011] Preferably, the step S1 of mapping the hydrodynamic parameter sequence to a state space vector includes the following sub-steps: step S11, importing the hydrodynamic parameter sequence into a filter for filtering to generate a filtered hydrodynamic parameter sequence; step S12, mapping the filtered hydrodynamic parameter sequence to a state space vector.

[0012] Preferably, step S6, controlling the valve opening of the ecological flow control mechanism, includes the following sub-steps: step S61, converting the water volume optimization control command or transient interruption control command into the corresponding electrical signal adjustment increment; step S62, sending the electrical signal adjustment increment to the electric regulating valve of the ecological flow control mechanism to drive the electric regulating valve to adjust the valve opening.

[0013] Preferably, the discrete mapping table stores the correspondence between multiple soil negative pressure variation rate intervals and water volume control limits; step S5 involves retrieving the preset discrete mapping table based on the soil negative pressure variation rate and using the matched discrete water volume control limit as a transient blocking control command, including the following sub-steps: step S51, determining the interval to which the soil negative pressure variation rate belongs; step S52, reading the water volume control limit associated with the interval in the discrete mapping table; step S53, extracting the water volume control limit as a transient blocking control command.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention uses multi-source tracer data as a constraint on low-frequency hydrological mechanisms, enabling the segmentation of end-member mixtures of multiple water bodies, including river water, permafrost water, soil water, and plant water, and the identification of water replenishment relationships. This avoids the blind standardization of raw isotope data, allowing the optimization of ecological water volume to be based on the real replenishment-storage-utilization ecological hydrological mechanisms. It can accurately solve ecological water replenishment synergistic configuration schemes that include target water replenishment volume, water replenishment time period, and water replenishment intensity, significantly improving the scientificity and accuracy of ecological water resource regulation.

[0015] 2. This invention constructs an asynchronous decoupled control architecture for low-frequency hydrological mechanism optimization and high-frequency transient blocking. When external sudden environmental changes cause timeouts in high-dimensional objective solutions or drastic fluctuations in soil negative pressure, the controller can adaptively cut off the slow optimization loop and switch to discrete mapping blocking control based on the soil negative pressure variation rate. This not only ensures the rationality of the high-level ecological water replenishment coordination configuration scheme, but also eliminates the risk of the underlying hardware getting stuck in computational deadlock or control oscillation when dealing with sudden priority flows, thus realizing the deep integration of ecological hydrological mechanisms and automated control. Attached Figure Description

[0016] Figure 1 This is a flowchart of the water quantity optimization control method of the present invention; Figure 2 This is a diagram of the water quantity optimization control loop structure of the present invention; Figure 3 This is a graph showing the solution of the objective function parameters under the multi-source mechanism constraint of this invention. Figure 4 This is a schematic diagram of the multi-dimensional architecture of the core technical elements for optimization and control in this invention.

[0017] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0019] A water quantity optimization control method driven by multi-source tracer data includes the following steps: Step S1: Obtain multi-source tracer analysis data, hydrodynamic parameter sequence and soil negative pressure data, map the hydrodynamic parameter sequence into a state space vector, and construct low-frequency hydrological mechanism constraints in combination with multi-source tracer analysis data, and calculate the soil negative pressure variation rate by taking the derivative of the soil negative pressure data. Step S2: Construct an asynchronous decoupled control architecture that includes a slow benchmark optimization loop and a fast transient blocking loop, and deploy the optimization control model within the slow benchmark optimization loop; Step S3: Collect the optimization time parameter when the optimization control model calculates the state space vector through time monitoring logic, and compare the optimization time parameter with the preset closed-loop response time threshold. Step S4: When the optimization time parameter does not exceed the closed-loop response time threshold, the optimization control model is driven by the slow benchmark optimization loop to solve the state space vector, and the output ecological water replenishment collaborative configuration scheme containing the target water replenishment volume, water replenishment period and water replenishment intensity is used as the water quantity optimization control command. Step S5: When the optimization time parameter exceeds the closed-loop response time threshold, the slow benchmark optimization loop is cut off, and the preset discrete mapping table is retrieved by the soil negative pressure change rate through the fast transient blocking loop, and the matched discrete water volume control limit is used as the transient blocking control command. Step S6: Output water volume optimization control command or transient interruption control command, and convert it into a regulating signal to control the valve opening of the ecological flow control mechanism.

[0020] Preferably, the following steps are included between steps S1 and S2: Step S201, decompose the state space vector into a base current component and an impulse component; Step S202, calculate the information entropy of the impulse component; Step S203, when the information entropy exceeds a preset noise threshold, lower the response coefficient of the risk term in the optimization control model and raise the inertia coefficient of the water balance term.

[0021] Preferably, step S4, which uses a slow benchmark optimization loop to drive the optimization control model to solve the state space vector and outputs an ecological water replenishment coordination configuration scheme containing the target water replenishment amount, water replenishment period, and water replenishment intensity as a water quantity optimization control instruction, includes the following sub-steps: Step S41, performing end-member hybrid segmentation model calculations on multi-source tracer analysis data containing river water, permafrost water, soil water, and plant water to determine the contribution ratio and replenishment relationship of different water bodies within the controlled area; Step S42, determining the water body replenishment relationship within the controlled area based on a dual isotope mass balance model combined with hydrological observations; Step S43, within the preset ecological boundary constraints, using the contribution ratio and water body replenishment relationship as boundary conditions, solving a multivariate nonlinear objective function to obtain an ecological water replenishment coordination configuration scheme containing the target water replenishment amount, water replenishment period, and water replenishment intensity as a water quantity optimization control instruction.

[0022] Preferably, before step S1, the following steps are also included: step S401, recording the historical evolution data of the state space vector within the continuous control cycle; step S402, calculating the state change rate based on the historical evolution data; step S403, when the state change rate shows a continuous decrease and reaches a preset lower limit threshold, outputting a correction factor to the slow benchmark optimization loop and increasing the benchmark water replenishment in the water quantity optimization control command.

[0023] Preferably, the multi-source tracer analysis data includes hydrogen isotope deviation data and oxygen isotope deviation data of river water, permafrost water, soil water and plant water, and the hydrodynamic parameter sequence includes soil saturated hydraulic conductivity and surface runoff rate.

[0024] Preferably, ecological boundary constraints include lower limits for groundwater level and lower limits for soil moisture content.

[0025] Preferably, the step S1 of mapping the hydrodynamic parameter sequence to a state space vector includes the following sub-steps: step S11, importing the hydrodynamic parameter sequence into a filter for filtering to generate a filtered hydrodynamic parameter sequence; step S12, mapping the filtered hydrodynamic parameter sequence to a state space vector.

[0026] Preferably, step S6, controlling the valve opening of the ecological flow control mechanism, includes the following sub-steps: step S61, converting the water volume optimization control command or transient interruption control command into the corresponding electrical signal adjustment increment; step S62, sending the electrical signal adjustment increment to the electric regulating valve of the ecological flow control mechanism to drive the electric regulating valve to adjust the valve opening.

[0027] Preferably, the discrete mapping table stores the correspondence between multiple soil negative pressure variation rate intervals and water volume control limits; step S5 involves retrieving the preset discrete mapping table based on the soil negative pressure variation rate and using the matched discrete water volume control limit as a transient blocking control command, including the following sub-steps: step S51, determining the interval to which the soil negative pressure variation rate belongs; step S52, reading the water volume control limit associated with the interval in the discrete mapping table; step S53, extracting the water volume control limit as a transient blocking control command.

[0028] Example 1: The method of the present invention runs on an industrial control computer and is applied to the scenario of precise water scheduling in water conservancy and ecological regulation areas. In the daily automated water control practice for this area, the overall water source contribution rate generated by multi-source tracer data analysis has a long transmission physical lag. Sudden soil water priority flow paths often lead to rapid water loss under millisecond-level transient conditions. The synchronous optimization architecture with a single clock reference is prone to control delay overshoot and system oscillation. The industrial control computer establishes an asynchronous decoupled control architecture that includes a slow reference optimization loop and a fast transient blocking loop. Each loop is allocated an independent operating clock, and steady-state regulation and transient risk blocking are achieved through data frequency band stripping.

[0029] The industrial control computer reads in real time from various monitoring terminals via a data acquisition bus the hydrogen and oxygen isotope deviation data (ppmD), soil moisture content data, and soil negative pressure data at typical sites for river water, permafrost water, soil water, and plant water at different stages of the watershed's water cycle. Filtering is then performed to eliminate transmission noise. To preserve the true end-member ratios of stable isotopes, the ppmD data is kept in its original numerical form as a constraint for low-frequency hydrological mechanisms and is not subjected to indiscriminate standardization. Only hydrodynamic parameters such as soil saturated hydraulic conductivity and surface runoff rate are processed. Zero-to-one standardization mapping was performed to eliminate the dimensional and scale gap. The mapping rule was as follows: the real-time collected soil saturated hydraulic conductivity and surface runoff rate were subtracted from their corresponding 7-day moving minimum values ​​in the historical database, and then divided by the difference between the corresponding 7-day moving maximum and minimum values, thus transforming them into dimensionless scalar values ​​between 0 and 1, which were then arranged to construct the system state space vector. The slow benchmark optimization loop ran with a low-frequency sampling period, calling the endmember hybrid segmentation model to analyze the contribution ratio and recharge relationship of different water bodies such as river water, permafrost water, soil water, and plant water, and combined with hydrological data. The observational inference of water body recharge relationships is used as a constraint input to the nonlinear objective programming function, constrained by low-frequency hydrological mechanisms. Based on the preset ecological boundaries of the lower limit of groundwater level and lower limit of soil moisture content, an ecological water replenishment coordination configuration scheme including target water replenishment volume, water replenishment period, water replenishment area, and water replenishment intensity under the current working conditions is solved, and the scheme is written into a shared register while monitoring the time consumption parameter of the model calculation. Specifically, the nonlinear objective programming function is composed of a weighted superposition of a water demand balance term and an ecological risk penalty term, wherein the water demand balance term is defined as the ratio of the target ecological water storage in the controlled area to the actual measured water storage. The absolute value of the difference is obtained by multiplying the current soil moisture content, groundwater level, and the geometric area of ​​the region by the actual measured water storage. The ecological risk penalty is a unilateral constraint penalty: when the measured groundwater level is higher than or equal to the preset lower limit of the groundwater level ecological boundary constraint, no penalty is triggered, and the ecological risk penalty is 0. When the measured groundwater level is lower than the preset lower limit of the groundwater level ecological boundary constraint, a penalty is triggered, defined as the square of the difference between the current groundwater level and the preset lower limit of the groundwater level, multiplied by a preset nonlinear penalty coefficient. The calculation formula for the nonlinear objective programming function is as follows: When hour: ;when hour: In the formula, This represents the overall cost of the target control model, with the dimension being volume. ; This represents the pre-set target ecological water storage capacity of the controlled area, in units of... ; This represents the average soil moisture content measured at multiple points in the controlled area. It is a volumetric moisture content and is dimensionless. This indicates the measured groundwater level, in units of... ; Represents the geometric distribution area of ​​the controlled region, in units of ,in Characterized by actual measured water storage, in units of ; This represents the preset lower limit of the groundwater level ecological boundary constraint, in units of... ; Characterized by the deviation of the measured groundwater level from the lower limit, in meters; This represents the preset nonlinear penalty coefficient, in units of... ,Right now It is used to convert the square of the groundwater level deviation into a volume dimension consistent with water storage, and to characterize the equivalent ecological water storage risk value corresponding to a unit square of water level deviation. The inertia coefficient representing the water balance term is initially set to 0.4; The response coefficient of the ecological risk penalty item is initially set to 0.6. When the environmental conditions are stable and the calculation time is kept within the preset closed-loop response time threshold (such as the 80ms algorithm timeout threshold), the industrial control computer reads the ecological water replenishment coordination configuration scheme in the shared register, converts it into the corresponding electrical signal adjustment command and sends it to the electric regulating valve of the flow regulation actuator to drive the electric regulating valve to adjust the valve opening and realize the overall coordinated control of water demand and water replenishment.

[0030] The fast transient blocking loop operates at a high frequency with a sampling period of 10ms, continuously reading the soil negative pressure change rate. The soil negative pressure change rate is calculated based on continuous sampling difference, and the calculation formula is as follows: In the formula, This indicates the rate of change of soil negative pressure; This represents the measured value of soil negative pressure collected at the current sampling time; This represents the measured value of soil negative pressure collected at the previous sampling time; The sampling period of the fast transient blocking loop is set to [value]. It compares the value with the built-in priority flow activation threshold. When the absolute value of the soil negative pressure change rate exceeds the threshold, it indicates that the priority flow channel is active. The fast transient blocking loop searches the discrete mapping table to obtain the corresponding transient feedforward compensation amount, and sends a high-priority hardware interrupt through the asynchronous instruction arbitrator to forcibly overwrite the output instruction of the slow reference optimization loop, generating the final flow regulation control instruction. This drives the servo motor of the flow regulation actuator to complete the fast response action. When the industrial control computer monitors that the optimization time parameter of the slow reference optimization loop exceeds the 80ms threshold, it automatically triggers the output fuse control signal, switching the control to the fast transient blocking loop. The quasi-safety compensation threshold output by the scatter mapping table is used as a control command to ensure that the system does not fall into computational deadlock under extreme conditions. The flow regulation actuator executes the command to maintain the water level in the target area. This process calculates the position execution residual vector and performs feedforward incremental compensation for the mechanical hysteresis dead zone of the regulating valve stem to eliminate steady-state tracking error. After the above regulation and control, the water dispatch response speed of this water conservancy ecological regulation area during the period of severe ablation is improved to the millisecond level, which effectively alleviates the resource leakage caused by priority flow. There are no abnormal high-frequency oscillations in the closed-loop control process of the system. The tracking error between the target value of vegetation water demand and the actual water storage is kept within the preset index, realizing the dynamic balance of ecological water resources under complex disturbance environment.

[0031] Example 2: The method of the present invention was operated in the same ecological regulation area as the previous example. The experimental group was configured with an asynchronous decoupled control architecture including a slow benchmark optimization loop and a fast transient blocking loop, while the control group was configured with a synchronous optimization control architecture with a single sampling period. The experiment evaluated the tracking response accuracy and system operation stability of the two architectures to the target water demand under the condition of transient soil preferential flow interference. The experimental platform consisted of a multi-source tracer data acquisition terminal, an industrial control computer, and an electric regulating valve actuator. The sampling period for hydrogen isotope abundance data was set to 1 hour, the sampling period for soil negative pressure data was set to 10 ms, and the threshold for negative pressure change rate was set to 1.5 kPa / s. Here, the threshold settings of the above parameters have clear system-level engineering boundary limitations: the closed-loop response time threshold is set to 80 ms as a calculation protection threshold to prevent the industrial control computer from timeout divergence. The fast loop sampling period is set to 10ms as the clock sampling period of the high-frequency sensor acquisition card to achieve high-frequency continuous monitoring of soil negative pressure. The priority flow judgment threshold is set to 1.5kPa / s, which is the extreme critical point for activation of macropore flow in the soil medium in this specific control area. If it is lower than 1.0kPa / s, the control valve will be frequently falsely triggered due to small environmental disturbances. If it is higher than 3.0kPa / s, the best window for blocking will be missed. The absolute value of the position execution residual is set to 0.3% to match the inherent mechanical backlash and fit tolerance of the large-diameter regulating valve stem. If it is lower than this value, the feedforward compensation will cause the actuator to fall into ineffective high-frequency micro-oscillation loss. When the absolute value of the soil negative pressure change rate is greater than 1.5kPa / s, the priority flow blocking action is triggered. The test signal source is superimposed with random noise with a peak amplitude of 0.5kPa to simulate the high-frequency fluctuation of the on-site hydrological monitoring environment.

[0032] During the operation of the test group, the slow benchmark optimization loop calculated the steady-state water volume with a 1-hour cycle. When the precipitation event caused the soil moisture to infiltrate rapidly, and the soil negative pressure change rate increased from 2.2 kPa / s to 4.8 kPa / s within 12.5 minutes, the fast transient interruption loop detected this rate of change and output the corresponding transient feedforward compensation amount of 15.6 m³ / h according to the discrete mapping table. By interrupting the logic to instantaneously overwrite the steady-state control command, the electric regulating valve of the flow regulating actuator acted with a response time of 200 ms, adjusting the valve stem opening from 35.2% to 42.8%. During this period, the system offset the mechanical hysteresis effect of the regulating valve stem through the feedforward incremental compensation mechanism, and the deviation between the measured valve position feedback and the control command decreased from 4.2% to 0.5%.

[0033] In the control group, under the same operating conditions, because the single-loop optimization calculation cycle was fixed at 1 hour, the controller still output based on historical steady-state data during the first 30 minutes of the soil negative pressure fluctuation rate increase, failing to capture the priority flow burst signal. It wasn't until the 45th minute, when a new round of optimization calculation was completed, that the flow regulation actuator activated, resulting in a water loss of 45.8 m³ in the target area during that period. The experimental data shows that when the negative pressure fluctuation rate threshold is between 1.0 kPa / s and 2.0 kPa / s, the average tracking error of the experimental group remains within 0.6%. Within the system, there was no abnormal oscillation. When the threshold was below 0.5 kPa / s, the frequent triggering of blocking actions caused the flow regulation actuator to operate 12 times per hour, resulting in mechanical wear of the regulating valve stem. When the threshold was above 3.0 kPa / s, the system's response to small priority flows was delayed, and the water volume scheduling deviation increased to 8.5%. The experimental group achieved high-precision water demand regulation under complex disturbances through asynchronous decoupling mechanism and mechanical dead zone feedforward compensation. All data confirmed the stability and response effectiveness of the method of this invention in engineering applications.

[0034] Example 3: This example combines Figures 1 to 4 This paper describes a water quantity optimization control method driven by multi-source tracer data, such as... Figure 1 As shown, in step S1, multi-source tracer analysis data, hydrodynamic parameter sequences, and soil negative pressure data are acquired. The hydrodynamic parameter sequences are mapped into state space vectors, and low-frequency hydrological mechanism constraints are constructed in conjunction with the multi-source tracer analysis data. The soil negative pressure variation rate is calculated by differentiating the soil negative pressure data. In step S2, an asynchronous decoupled control architecture containing a slow benchmark optimization loop and a fast transient blocking loop is constructed. An optimization control model is deployed within the slow benchmark optimization loop. In step S3, the optimization time consumption parameter when the optimization control model calculates the state space vector is collected through time monitoring logic. The optimization time consumption parameter is compared with the preset closed-loop response time threshold. In step S4, the optimization time consumption parameter is further analyzed. When the time parameter does not exceed the closed-loop response time threshold, the state space vector is solved by driving the optimization control model through the slow benchmark optimization loop. The output ecological water replenishment coordinated configuration scheme, which includes the target water replenishment volume, water replenishment period, and water replenishment intensity, is used as the water quantity optimization control command. In step S5, when the optimization time parameter exceeds the closed-loop response time threshold, the slow benchmark optimization loop is cut off. The preset discrete mapping table is retrieved through the fast transient blocking loop using the soil negative pressure change rate. The matched discrete water quantity control limit is used as the transient blocking control command. In step S6, the water quantity optimization control command or the transient blocking control command is output and converted into an adjustment signal to control the valve opening of the ecological flow control mechanism.

[0035] like Figure 2As shown, the state space vector is transmitted to the slow reference optimization loop when the time consumed does not exceed the response time threshold, and to the fast transient blocking loop when the time consumed exceeds the response time threshold. The slow reference optimization loop transmits the data to the fast transient blocking loop when the soil negative pressure change rate exceeds the judgment threshold. The slow reference optimization loop outputs a water quantity optimization control command to the electric regulating valve to actuate, and the fast transient blocking loop outputs a transient blocking control command to the electric regulating valve to actuate. After the electric regulating valve actuates, the system acquires the current evolution data and records it in the historical database. The historical evolution data is transmitted to the state space vector within the continuous control cycle.

[0036] like Figure 3 As shown, the state space vector records historical evolution data within a continuous control period to calculate the state change rate. When the state change rate is determined to be continuously decreasing and reaches a preset lower limit threshold, a correction factor is triggered to execute an upward adjustment of the baseline water replenishment. Simultaneously, hydrogen isotope deviation data and oxygen isotope deviation data are extracted from the multi-source tracer analysis data. Based on the dual isotope mass balance model and combined with hydrological observations, the water body replenishment relationship is determined. At the same time, the tracer data is input into the end-member hybrid segmentation model to determine the contribution ratio of different water bodies and their replenishment relationship. In the solution execution stage, the dynamic control factor formed by the risk term response coefficient downward adjustment command and the water demand balance term inertia coefficient upward adjustment command, as well as the ecological boundary constraints including the groundwater level lower limit constraint and the soil moisture content lower limit constraint, are jointly input into the multivariate nonlinear objective function. Finally, the solution is performed using the contribution ratio and water body replenishment relationship as boundary conditions, and the output is an ecological water replenishment collaborative configuration scheme including the target water replenishment volume, water replenishment period, and water replenishment intensity.

[0037] like Figure 4 As shown, the multi-source tracer data-driven water optimization control method is composed of multi-dimensional technical elements. The multi-source tracer analysis data dimension covers river water, permafrost water, soil water, and plant water, and is characterized by hydrogen isotope deviation data and oxygen isotope deviation data. The slow benchmark optimization loop mechanism integrates the end-member hybrid segmentation model and the water replenishment connection solution logic to generate an ecological water replenishment collaborative configuration scheme. The adaptive damping tuning mechanism dynamically adjusts the inertia coefficient of the water demand balance term and the response coefficient of the risk term based on the information entropy of the pulse component. The hydrodynamic parameter sequence dimension maps the soil saturated hydraulic conductivity and surface runoff rate into a state space vector. The fast transient blocking loop mechanism uses the soil negative pressure change rate and discrete mapping table to generate transient blocking control commands. The ecological flow control mechanism dimension calculates the residual and adjusts the electrical signal increment according to the location to physically adjust the valve opening.

[0038] Example 4: In the water level and quantity monitoring system of the ecological regulation area, the existing control architecture often deviates from the optimal water demand threshold due to the lag in updating the historical baseline when facing sudden changes in surface runoff and dynamic fluctuations in groundwater level. The system deploys a time-based reconstruction mechanism based on the negative pressure baseline within the industrial control computer. It calibrates the reference input of the flow regulation model in real time through a sliding time window, dynamically correcting the steady-state deviation under long-term environmental changes. The industrial control computer is configured with a rolling buffer in the non-volatile storage area, updating the historical database on a 24-hour cycle, and setting the sliding time window to 7 days. The system reads the soil negative pressure data every 10 milliseconds, stores it in the rolling buffer, and calculates the statistical mean and standard deviation of the soil negative pressure change rate within the window period. When the standard deviation exceeds the preset value for three consecutive sampling periods, the system will take action. When the dispersion tolerance is reached, the logic controller determines that the historical model baseline is invalid and triggers the parameter update process. Here, the preset dispersion tolerance is specifically set to 0.5 kPa / s. When the calculated standard deviation is greater than 0.5 kPa / s for three consecutive sampling periods, it is determined that the historical model baseline cannot truly reflect the current drastic changes in operating conditions. In the parameter update process, the so-called regression fitting reconstruction logic is as follows: the controller extracts the scalar of the soil negative pressure change rate of each group in the rolling buffer in the last 7 days as the input independent variable, and uses the corresponding isotope analysis water volume data as the dependent variable. The univariate linear regression method is used to calculate the weight of the first derivative of the dependent variable deviating from the independent variable, thereby solving for the new slope coefficient and intercept coefficient, and directly overwriting them into the control register as the updated water volume characteristic regression coefficient to complete the time-sensitive reconstruction of the baseline.

[0039] The controller stops iteratively calling the current flow regulation model and instead executes the reconstruction logic. The system uses the soil negative pressure change rate and isotope-analyzed water volume data from the past 7 days within the rolling buffer as a sample set. It updates the model's water volume characteristic regression coefficients using a regression fitting method, writing the obtained regression coefficients into the control register as the baseline input for subsequent control logic. Addressing the response hysteresis and execution deviation of the flow regulation actuator, the controller activates the actuator state feedback loop before generating the final command. The control logic collects the actual state feedback signal of the flow regulation actuator in real time, compares it with the theoretical opening of the command, calculates the position execution residual vector, and applies a directional dead zone compensation signal based on the sign of the residual vector. When the residual... When the absolute value of the vector is greater than 0.3%, the controller sends an incremental compensation signal to drive the ecological flow control mechanism to overcome dead zone friction and ensure that the control accuracy of the flow regulation actuator is maintained within 0.5%. The above-mentioned reconstruction mechanism is executed in a 7-day cycle to ensure that the flow regulation model always matches the current geological and institutional conditions. When updating the model parameters, the controller performs data verification in parallel, hashes and anonymizes all negative pressure values ​​to ensure that the control operation is based only on the feature values ​​and does not involve geographical coordinates or operator identification. Through the dynamic translation of the model baseline and the real-time closed-loop compensation of the control mechanism, the flow regulation accuracy remains steady throughout the entire life cycle, and the control logic and physical execution layer achieve synchronous convergence.

[0040] Example 5: In this water conservancy ecological regulation system, the industrial control computer establishes a closed-loop control architecture consisting of a host computer, a multi-source tracer data acquisition terminal, and electric actuators. Before system deployment, the industrial control computer performs dead zone calibration procedures on each electric actuator. The controller sends incrementally small voltage signals to the electric actuators until displacement of the valve stem is detected. The minimum increment value of the voltage signal is recorded and used as the mechanical dead zone parameter of the electric actuator and stored in the memory. At the same time, the industrial control computer fits a regional water transport benchmark feature library based on historical isotope abundance and soil moisture transport data. This feature library contains hydrogen and oxygen isotope end-member ratio data for different seasonal operating conditions.

[0041] During system operation, the slow benchmark optimization loop acquires multi-source tracer analysis data within the current control area, including hydrogen isotope deviation data and oxygen isotope deviation data for river water, permafrost water, soil water, and plant water, as well as hydrodynamic parameters such as soil saturated hydraulic conductivity and surface runoff rate. The industrial control computer calls the end-member hybrid segmentation model to perform end-member segmentation on river water, permafrost water, soil water, and plant water, identifying the source composition, contribution ratio, and water recharge relationships of each type of water body. Furthermore, the system performs time-series moving average decomposition on the constructed state space vector, where the moving average of the past 10 periods is defined as the slowly varying baseflow component, and the current real-time state... The residual part of the state vector minus the base current component is defined as the abrupt pulse component. To quantify the intensity of the pulse component's fluctuation, this invention calculates its information entropy. The specific steps are as follows: Within the current control cycle, 60 pulse component sample values ​​are collected at a sampling interval of 1 minute. The frequency of each sample value falling within 10 preset equidistant numerical intervals is counted, and the frequency is divided by the total number of samples (60) to obtain the probability of occurrence of each interval. All non-zero probability values ​​among these 10 probability values ​​are multiplied by their base-2 logarithm, and the product results are summed and then multiplied by -1. The final scalar value is the information entropy of the pulse component. The mathematical formula for calculating information entropy is as follows: In the formula, Information entropy representing the pulse component; This represents the total number of preset equally spaced numerical intervals, with a value of 10. This indicates that the pulse component sample value falls within the first... The probability of occurrence of each interval, i.e. ,in For falling in the first The sample frequency of each interval The total number of samples within the sampling period, with a value of 60; when When, define Based on the contribution ratio of various water bodies, the water replenishment relationship and the ecological water level boundary constraints, the controller performs multivariate nonlinear objective function optimization to solve for an ecological water replenishment coordination configuration scheme that includes target water replenishment volume, water replenishment period, water replenishment area and water replenishment intensity.

[0042] In the specific calculation, the initial response coefficient of the risk term in the optimization control model is set to 0.6, and the initial inertia coefficient of the water balance term is set to 0.4. When the information entropy of the calculated pulse component exceeds the preset noise threshold of 1.2, the controller activates the adaptive damping tuning mechanism: the response coefficient of the risk term is directly reduced from the initial 0.6 to a fixed value of 0.2 to reduce the system's active response sensitivity to high-frequency pulse disturbances; simultaneously, the inertia coefficient of the water balance term is increased from the initial 0.4 to a fixed value of 0.8, thereby increasing the virtual damping to forcibly maintain the smoothness of the system output and avoid frequent jumps and reversals of commands under strong noise. If the optimization calculation cycle exceeds 80ms, the controller activates the timeout lockout logic, cuts off the output of this loop, and switches to fast transient response. The system employs a transient blocking loop control mechanism. A rapid transient blocking loop operates with a 10ms cycle, real-time reading of soil negative pressure data at monitoring sites within the control area. This data is then differentially processed to obtain the soil negative pressure fluctuation rate. The system compares this rate with a preset 1.5kPa / s priority flow threshold. When the absolute value of the soil negative pressure fluctuation rate exceeds the threshold, the system determines that a transient change in priority flow has occurred. The controller suspends the slow benchmark optimization loop and directly retrieves the built-in discrete mapping table. This mapping table pre-defines the logical correspondence between the soil negative pressure fluctuation rate and water volume control limits. Based on the current negative pressure fluctuation rate, the controller outputs a transient blocking control command to ensure rapid water volume restriction during sudden priority flow events, preventing the leakage of ecological water resources through non-target channels.

[0043] During the water flow regulation execution phase, the system introduces a mechanical hysteresis compensation mechanism based on real-time position residuals. After the electric actuator receives the control command, the industrial control computer reads the actual position feedback of the regulating valve stem in real time. The controller calculates the difference between the actual position and the theoretical target position to obtain the position execution residual. The system calls the stored dead zone parameters to calculate the compensation voltage increment corresponding to the position execution residual. Specifically, the cross-dimensional conversion of the residual vector to the regulation signal increment relies on the built-in feedforward dead zone compensation mapping gain term. The dead zone parameters record the minimum trigger signal threshold required to overcome the dead zone friction of the ecological flow control mechanism. This threshold is set to 0.2V. When the absolute value of the calculated position execution residual is greater than 0.3%, the controller multiplies the residual value by a preset proportional gain coefficient to calculate the basic signal regulation increment, and adds it to the minimum trigger signal threshold to obtain the final compensation signal increment. At the same time, the polarity of the signal increment is determined by the sign of the residual vector, thereby completing the bridging mapping from geometric displacement deviation to control signal quantity. The calculation formula for the feedforward dead zone compensation signal increment is as follows: In the formula, This represents the calculated increment of the compensation signal; The position execution residual is obtained by subtracting the actual feedback position from the theoretical target position of the control valve stem; Represents the absolute value of the position execution residual, when Calculations are performed when >0.3%; when ≤0.3% ; The sign function is used to determine the polarity of the compensation signal. This represents the preset dead-time ratio gain coefficient; This represents the minimum trigger signal threshold, with a value of [value missing]. The industrial control computer superimposes the compensation signal increment into the final electric actuator drive control signal, outputting a corrected control signal. The ecological flow control mechanism acts according to the corrected control signal, smoothly adjusting the valve stem opening to the target value. When the system detects that the calculation time continuously exceeds the preset time limit of 80ms, the controller triggers the fuse protection logic, locking the control logic within the safe water volume limit range output by the discrete mapping table, ensuring that the water level in the water conservancy ecological regulation area remains under control under emergency conditions.

[0044] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A water quantity optimization control method driven by multi-source tracer data, characterized in that, Includes the following steps: Step S1: Obtain multi-source tracer analysis data, hydrodynamic parameter sequence and soil negative pressure data, map the hydrodynamic parameter sequence into a state space vector, and construct low-frequency hydrological mechanism constraints in combination with multi-source tracer analysis data, and calculate the soil negative pressure variation rate by taking the derivative of the soil negative pressure data. Step S2: Construct an asynchronous decoupled control architecture that includes a slow benchmark optimization loop and a fast transient blocking loop, and deploy the optimization control model within the slow benchmark optimization loop; Step S3: Collect the optimization time parameter when the optimization control model calculates the state space vector through time monitoring logic, and compare the optimization time parameter with the preset closed-loop response time threshold. Step S4: When the optimization time parameter does not exceed the closed-loop response time threshold, the optimization control model is driven by the slow benchmark optimization loop to solve the state space vector, and the output ecological water replenishment collaborative configuration scheme containing the target water replenishment volume, water replenishment period and water replenishment intensity is used as the water quantity optimization control command. Step S5: When the optimization time parameter exceeds the closed-loop response time threshold, the slow benchmark optimization loop is cut off, and the preset discrete mapping table is retrieved by the soil negative pressure change rate through the fast transient blocking loop, and the matched discrete water volume control limit is used as the transient blocking control command. Step S6: Output water volume optimization control command or transient interruption control command, and convert it into a regulating signal to control the valve opening of the ecological flow control mechanism.

2. The water quantity optimization control method driven by multi-source tracer data according to claim 1, characterized in that, The steps between steps S1 and S2 are as follows: Step S201, decompose the state space vector into a base current component and an impulse component; Step S202, calculate the information entropy of the impulse component; Step S203, when the information entropy exceeds a preset noise threshold, lower the response coefficient of the risk term in the optimization control model and raise the inertia coefficient of the water balance term.

3. The water quantity optimization control method driven by multi-source tracer data according to claim 1, characterized in that, In step S4, the state space vector is solved by driving the optimization control model through a slow benchmark optimization loop. The output ecological water replenishment coordination configuration scheme, which includes the target water replenishment volume, water replenishment period, and water replenishment intensity, is used as the water quantity optimization control instruction. The sub-steps are as follows: Step S41, the end-member hybrid segmentation model is calculated on the multi-source tracer analysis data including river water, permafrost water, soil water, and plant water to determine the contribution ratio of different water bodies in the controlled area and their replenishment relationship; Step S42, the water replenishment relationship in the controlled area is determined based on the dual isotope mass balance model combined with hydrological observation. Step S43: Within the preset ecological boundary constraints, using the contribution ratio and water replenishment relationship as boundary conditions, solve the multivariate nonlinear objective function to obtain an ecological water replenishment coordination configuration scheme that includes the target water replenishment volume, water replenishment period, and water replenishment intensity as a water quantity optimization control instruction.

4. The water quantity optimization control method driven by multi-source tracer data according to claim 1, characterized in that, Before step S1, the following steps are also included: step S401, recording the historical evolution data of the state space vector in the continuous control cycle; step S402, calculating the state change rate based on the historical evolution data; step S403, when the state change rate shows a continuous decrease and reaches the preset lower limit threshold, outputting a correction factor to the slow benchmark optimization loop and increasing the benchmark water replenishment in the water quantity optimization control command.

5. The water quantity optimization control method driven by multi-source tracer data according to claim 1, characterized in that, The multi-source tracer analysis data includes hydrogen isotope deviation data and oxygen isotope deviation data from river water, permafrost water, soil water, and plant water. The hydrodynamic parameter series includes soil saturated hydraulic conductivity and surface runoff rate.

6. The water quantity optimization control method driven by multi-source tracer data according to claim 3, characterized in that, Ecological boundary constraints include lower limits for groundwater levels and lower limits for soil moisture content.

7. The water quantity optimization control method driven by multi-source tracer data according to claim 1, characterized in that, The step S1, which maps the hydrodynamic parameter sequence to a state space vector, includes the following sub-steps: Step S11, importing the hydrodynamic parameter sequence into a filter for filtering to generate a filtered hydrodynamic parameter sequence; Step S12, mapping the filtered hydrodynamic parameter sequence to a state space vector.

8. The water quantity optimization control method driven by multi-source tracer data according to claim 1, characterized in that, Step S6, controlling the valve opening of the ecological flow control mechanism, includes the following sub-steps: Step S61, converting the water quantity optimization control command or transient interruption control command into the corresponding electrical signal adjustment increment; Step S62: The electrical signal adjustment increment is sent to the electric regulating valve of the ecological flow control mechanism to drive the electric regulating valve to adjust the valve opening.

9. The water quantity optimization control method driven by multi-source tracer data according to claim 1, characterized in that, The discrete mapping table stores the correspondence between multiple soil negative pressure variation rate intervals and water volume control limits. Step S5 involves retrieving the preset discrete mapping table based on the soil negative pressure variation rate and using the matched discrete water volume control limits as transient blocking control commands, including the following sub-steps: Step S51, determine the interval to which the soil negative pressure variation rate belongs; Step S52, read the water volume control limits associated with the interval in the discrete mapping table; Step S53, extract the water volume control limits as transient blocking control commands.

Images