Tail end purification node self-adaptive operation decision-making system driven by multi-dimensional water quality characteristics

Abstract

The invention relates to the technical field of industrial process control, and discloses a multi-dimensional water quality characteristic-driven terminal purification node adaptive operation decision system, which comprises the following steps of: acquiring fluid state parameters, an actuating mechanism driving instruction and a response measurement value, calculating a ratio of the two to construct a transient hydraulic impedance index, and calculating a transient hydraulic impedance value; according to the method, the physical characteristic drift of the assembly is identified by monitoring the ratio of the driving energy to the fluid response in real time, the problem that the static look-up table cannot adapt to the structural variation is solved, and the reliability of the assembly is improved. And realizing stable self-adaptive control on the nonlinear fluid system by using the existing computing power.

Description

Technical Field

[0001] This invention relates to an adaptive operation decision system for end-of-pipe purification nodes driven by multi-dimensional water quality characteristics, belonging to the field of industrial process control technology. Background Technology

[0002] In current applications of semiconductor wet process, medical dialysis, and precision laboratory water supply, the end-of-line purification node is a key execution unit for ensuring water quality indicators. The stability of pressure and flow control determines the yield of downstream processes and equipment safety. Existing technologies typically employ single-loop PID feedback or a static gain scheduling strategy based on explicit state variables to maintain constant system output indicators.

[0003] Conventional control methods stop at monitoring thresholds of single physical parameters, failing to delve into adaptive adjustment of dynamic parameters at the execution level. For example, Chinese invention patent CN101786698A discloses a one-way pipeline membrane purification device for drinking water. Although it introduces a differential pressure sensor to monitor the pressure gradient at the inlet and outlet of the membrane module in real time, and determines blockage and triggers backup pipeline switching when the differential pressure exceeds a set value, the control core is still limited to threshold-triggered actions executing discrete logic. The system cannot continuously drift and correct the controller's proportional gain and integral time in real time based on flow resistance characteristics during the long decay cycle from membrane module installation to the need for cleaning. Furthermore, it cannot adapt to the changing raw water temperature... Fluctuations cause changes in fluid viscosity, and there is a lack of dynamic compensation mechanisms based on physical models. This control method is based on the assumption that the controlled object is a linear time-invariant system or that the characteristic drift can be characterized by sensors. In actual long-term operation, the physical transfer function of the purification component is dynamically modulated by multidimensional unobservable variables. In addition to measurable fluid parameters, there are microscopic pore caking, biofilm deposition, local air resistance in the flow channel, and structural physical variations due to material thermal inertia hysteresis inside the component. These variations cause the system's sensitivity to control energy to drift. Under the same temperature and flow sensor readings, the optimal control gain required by the component at different aging stages or clogging states varies by orders of magnitude.

[0004] Therefore, the technical problem to be solved by this invention is how to identify the unobserved structural state drift of the controlled object in real time using the existing control loop without adding additional physical sensor hardware, and dynamically reconstruct the control law accordingly to achieve adaptive control throughout the entire life cycle. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-of-pipe purification nodes, the system comprising:

[0006] The data acquisition module is used to synchronously acquire the input-side fluid state parameters of the end-of-line purification node in each control cycle, and obtain the current drive command value of the actuator and the output-side fluid response measurement value; the input-side fluid state parameters include raw water temperature, raw water flow rate and raw water conductivity;

[0007] The transient hydraulic impedance calculation module is used to calculate the ratio of the drive command value to the output-side fluid response measurement value, and obtain the transient hydraulic impedance index characterizing the current flow resistance characteristics of the fluid loop;

[0008] The dynamic gain scheduling module stores a mapping table between the transient hydraulic impedance index and the reference proportional gain coefficient. This module is used to find the reference proportional gain coefficient corresponding to the current operating condition from the mapping table based on the real-time calculated transient hydraulic impedance index, and determine the adaptive PID control parameters at the current moment in combination with the input fluid state parameters.

[0009] The composite drive control module is used to calculate the deviation between the output-side fluid response measurement value and the set value based on the adaptive PID control parameters to generate a feedback control component. The feedback control component is then superimposed with the feedforward control component calculated based on the input-side fluid state parameters to generate a drive command value for adjusting the actuator.

[0010] Preferably, the transient hydraulic impedance calculation module is also used to calculate the time rate of change of the transient hydraulic impedance index; the dynamic gain scheduling module includes an integral anti-saturation unit, which is used to set the integral term gain in the adaptive PID control parameters to zero when the time rate of change of the transient hydraulic impedance index exceeds a preset threshold, until the transient hydraulic impedance index stabilizes within a preset range.

[0011] Preferably, the transient hydraulic impedance calculation module calculates the transient hydraulic impedance index using the following formula: ,in, For the first Transient hydraulic impedance index for each control cycle, For the first The drive command value applied to the actuator in each control cycle, For the first The output-side fluid response measurement values ​​are collected in each control cycle. A preset constant is used to prevent the denominator from being zero.

[0012] Preferably, the system further includes a thermal hysteresis compensation module; this module stores a thermal response time constant characterizing the thermal capacity of the terminal purification node components, which is used to perform a first-order low-pass filtering on the raw water temperature in the input fluid state parameters based on the thermal response time constant to obtain the component equivalent temperature; the dynamic gain scheduling module is used to use the component equivalent temperature to replace the raw water temperature as the basis for correcting the adaptive PID control parameters.

[0013] Preferably, the mapping table adopts a piecewise nonlinear setting: when the transient hydraulic impedance index is higher than the preset high resistance threshold, a higher reference proportional gain coefficient is applied; when the transient hydraulic impedance index is lower than the preset low resistance threshold, a lower reference proportional gain coefficient is applied; the high resistance threshold and the low resistance threshold are determined based on the calibrated impedance value of the end purification node.

[0014] Preferably, the system also includes a maintenance effect evaluation module; this module is used to obtain performance recovery indicators after cleaning and maintenance is performed at the end purification node, compare them with preset benchmark values, and decompose the long-term drift of the drive command value into recoverable fouling components and unrecoverable aging components; the dynamic gain scheduling module is used to correct the base value of the mapping table only based on the aging component, and reset the mapping table when the fouling component is detected to exceed the preset reset threshold.

[0015] Preferably, the data acquisition module includes a data synchronization unit; this unit is used to time-align the input-side fluid state parameters and the output-side fluid response measurements according to the sampling frequency and transmission delay of different sensors.

[0016] Preferably, the composite drive control module includes a feedforward compensation unit; this unit stores a flow-pressure correspondence table based on a hydraulic model, which is used to look up the pressure compensation value as a feedforward control component according to the rate of change of the raw water flow in the input fluid state parameters.

[0017] Preferably, the actuator is a variable frequency pump or a regulating valve; the drive command value is the PWM duty cycle for controlling the variable frequency pump or the voltage value for controlling the regulating valve; the output side fluid response measurement value is the pipeline pressure value or the product water flow rate value.

[0018] Preferably, the dynamic gain scheduling module is also used to execute an impedance anomaly protection strategy; when the transient hydraulic impedance index is lower than the minimum safe value for a certain period of time, it is determined that there is a pipe disconnection or sensor failure in the fluid circuit, and a shutdown command is output to cover the drive command value.

[0019] Compared with the prior art, the beneficial effects of the present invention are:

[0020] 1. Construct a real-time observation logic based on the ratio of drive command to system response in the control loop, establish an adaptive adjustment mechanism that senses the unobserved physical stiffness changes of the controlled object, and utilize the transient ratio between the energy applied by the actuator and the actual response of the fluid loop to integrate multi-dimensional physical disturbances such as fluid viscosity changes, component micro-fouling degree, and flow channel geometric deformation into a single system impedance characteristic. This enables the controller to identify operating conditions where sensor readings are the same but physical drive sensitivities are vastly different, even in the absence of full-dimensional sensor coverage. The control gain is scaled in real time according to the actual throughput efficiency of the controlled object, eliminating the control blind spot caused by the inability of the static lookup table model to adapt to the structural time-varying characteristics of the controlled object. This ensures that the control command maintains dynamic matching with the actual load characteristics of the system when facing unmodeled physical variations, and effectively controls the nonlinear fluid system using the existing computing power of the system.

[0021] 2. By utilizing signal processing logic based on thermal response time constant, the phase relationship between control decisions and component physical response is aligned in the time domain. By applying inertial filtering processing that matches the component material and mass to the fluid temperature observation value, equivalent control variables characterizing the actual thermal state of the component are generated. This allows the controller gain adjustment trajectory to deviate from the transient step of fluid temperature and remain synchronized with the gradual curve of component physical permeability. This suppresses false judgments of operating conditions caused by switching between hot and cold fluids. By digitally adapting to physical thermal inertia, the system avoids overshooting of pipeline pressure or oscillation of actuators due to excessive control gain adjustment during thermal shock transients, ensuring a smooth transition of precision water supply under variable temperature conditions.

[0022] 3. Introduce performance recovery arbitration logic triggered by maintenance events to decouple reversible fouling and irreversible aging states of components in the full life cycle control strategy. By comparing the system's baseline response characteristics before and after cleaning or regeneration operations, distinguish between temporary impedance increases caused by surface deposits and permanent performance drift caused by material wear. Only the irreversible aging component is solidified as a long-term correction factor for control parameters. This prevents the adaptive algorithm from misjudging recoverable fouling as equipment aging, thus blocking the parameter locking path after maintenance. It also eliminates the risk of overcompensation due to excessively high historical gains after the system performs cleaning operations, ensuring that the evolution direction of the control system converges to the true physical life characteristics of the components, and realizing a closed loop between the adaptive strategy and the equipment maintenance cycle logic. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the system control logic flow for transient hydraulic impedance feedback of the present invention;

[0024] Figure 2 This is a comparison diagram of the adaptive strategy under step disturbance and the response of traditional PID control in this invention;

[0025] Figure 3 This is a diagram illustrating the closed-loop interaction architecture between the physical entity domain and the digital computing domain of this invention. Detailed Implementation

[0026] This specific embodiment is only used to illustrate the present invention and is not intended to limit the scope of protection of the present invention. 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 present invention, and such modifications and substitutions should all be included within the scope of protection of the present invention.

[0027] This invention provides a multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-of-pipe purification nodes, comprising a data acquisition module, a transient hydraulic impedance calculation module, a dynamic gain scheduling module, a feedforward compensation unit, a composite drive control module, and a maintenance effect evaluation module. In physical implementation, these modules run on an embedded microprocessor, interacting with memory and I / O interfaces via an internal bus, and connecting to the sensor group and actuators of the end-of-pipe purification node. The actuators include variable frequency pumps or regulating valves, and the sensor group covers both the raw water side and the product water side. The data acquisition module is responsible for constructing a synchronized time-series state vector. This module has a built-in circular buffer queue and performs zero-order hold interpolation on low-frequency sampled data based on the sampling cycle of the sensor with the highest sampling frequency in the system. The data acquisition module performs time-series alignment of the acquired input-side fluid state parameters and output-side fluid response measurements according to the pre-calibrated signal transmission delay time of each sensor, generating the... Synchronous status data for each control cycle, including input-side fluid state parameters such as raw water temperature, raw water flow rate, and raw water conductivity; and output-side fluid response measurements such as pipeline pressure or product water flow rate.

[0028] The transient hydraulic impedance calculation module is used to quantify the physical load characteristics of the fluid loop in each control cycle. This module obtains the current drive command value applied to the actuator. and the output-side fluid response measurement after synchronous processing ,in, In variable frequency pump control, it is the PWM duty cycle value; in regulating valve control, it is the DAC voltage value. For the corresponding pressure or flow sensor readings, this module uses the formula... Calculate the transient hydraulic impedance index In the formula To prevent the denominator from being zero, a preset constant is used, which is set to 0.1% of the sensor's range. This transient hydraulic impedance index... This characterizes the degree of obstruction of the current fluid loop to the driving energy; the dynamic gain scheduling module adjusts the control parameters in real time based on the transient hydraulic impedance index. The module's memory contains a preset impedance-gain mapping table, which stores the transient hydraulic impedance index. The numerical domain is divided into multiple discrete intervals, and a corresponding reference proportional gain coefficient is assigned to each interval. When the transient hydraulic resistance index When it is in the high-order range, the mapping table corresponds to a larger reference proportional gain coefficient; when When in the low-order range, the mapping table corresponds to a smaller reference proportional gain coefficient, which the processor utilizes in real-time calculations. Linear interpolation is performed in the mapping table to determine the reference proportional gain coefficient at the current time. This module calculates the time rate of change of the transient hydraulic impedance index. ,when When the preset mutation threshold is exceeded, the module triggers the integral anti-saturation logic, reducing the integral term gain in the PID controller. Set the value to zero until the rate of change falls back to the preset stable range.

[0029] The system corrects the thermal response lag caused by the component's heat capacity through a thermal hysteresis compensation module, which stores the thermal response time constant of the core purification component. This constant is obtained based on the specific heat capacity and total mass of the component material. The module reads the raw water temperature observation value from the input-side fluid state parameters at each sampling time. and using the formula Calculate the equivalent temperature of the components ,in, These are the filter coefficients. The system sampling period is The dynamic gain scheduling module uses the component's equivalent temperature from the previous moment as its reference. The observed raw water temperature is used as the basis for correcting the temperature correlation coefficient in the PID parameters; the composite drive control module is used to synthesize the final actuator control command; the feedforward compensation unit queries a preset flow-pressure correspondence table based on the rate of change of the input raw water flow and outputs the feedforward control component. Simultaneously, the module calculates the deviation between the measured value and the set value of the output fluid response, and, in conjunction with the adaptive PID parameters determined by the dynamic gain scheduling module, calculates the feedback control component. The module will provide feedback control components. With feedforward control components The values ​​are superimposed to generate the final drive instruction value. The output is then sent to the actuator; the maintenance effect evaluation module is used to decouple reversible fouling and irreversible aging of components. After the system performs cleaning and maintenance operations and resumes operation, this module obtains the performance recovery indicators under standard operating conditions. ,like If the impedance exceeds the preset reset threshold, the system determines that the increase is due to reversible blockage and performs a reset operation on the correction coefficient in the dynamic gain scheduling module; if If the impedance is below the preset reset threshold, the system determines that the increase in impedance is due to irreversible aging, and updates the base value of the mapping table based on the current impedance residual to solidify the impact of aging on the control model.

[0030] Example 1: In a 12-inch semiconductor wafer wet cleaning process, the water supply circuit needs to cope with high-frequency and drastic fluctuations in operating conditions during process step switching. This condition is characterized by the raw water flow rate needing to jump from 5 L / min to 20 L / min within 200 ms, and the raw water temperature needing to simultaneously increase from 25°C. Quickly switch to 60 The pressure fluctuation in the pipeline is controlled within ±5 kPa of the set value. Otherwise, pressure overshoot will cause physical wear of the precision megasonic nozzle, while pressure drop will cause uneven cleaning and a decrease in wafer yield. Under such temperature and current coupling disturbances, traditional control strategies often fail to analyze the nonlinear changes in fluid viscosity and the physical impedance drift caused by thermal expansion and contraction of the membrane module in real time, resulting in pressure oscillations or response lags that exceed the allowable range.

[0031] When the system faces the above-mentioned mixed operating conditions, the feedforward compensation unit, based on the rate of change of the raw water flow rate on the input side, queries a preset flow-pressure correspondence table and outputs an open-loop feedforward control component within 100ms of the step change in the raw water flow rate. To the actuator, this action pre-increases the output torque of the variable frequency pump before the pipeline pressure physically drops due to increased flow demand. It utilizes feedforward energy to counteract the pressure fluctuation potential energy caused by the flow step, and adjusts the raw water temperature readings before the injection of hot fluid. Simultaneously with the abrupt change, the thermal hysteresis compensation module reads the observed value and combines it with the pre-stored thermal response time constant, calibrated to 120 seconds. Using the formula Calculate the equivalent temperature of the components When the fluid temperature has reached 60 The membrane module itself is still in the physical transition period of low temperature. It exhibits a slow upward trend along an exponential curve, and the dynamic gain scheduling module adjusts accordingly. Determine the reference proportional gain coefficient to prevent the controller from misjudging that the component is in a high temperature and low viscosity state and prematurely reducing the control gain due to directly referencing the sudden change in fluid temperature. This ensures that the control stiffness of the system always matches the actual high impedance physical state of the component during the thermal conduction hysteresis period.

[0032] During the continuous operation of the above process, the transient hydraulic impedance calculation module performs calculations in each control cycle. Calculate the driving instruction value Compared with the output-side fluid response measurement value The ratio, i.e., the transient hydraulic resistance index When the filter element surface traps tiny particles, causing an increase in flow resistance, As the gain gradually increases and enters the high-level range, the dynamic gain scheduling module automatically matches a larger reference proportional gain coefficient through the impedance-gain mapping table. It outputs enhanced correction energy to overcome the additional physical damping caused by clogging, in case of airbag discharge in the pipeline. Rate of change over time When a step drop exceeds a preset threshold, the dynamic gain scheduling module triggers integral anti-saturation logic, adjusting the integral term gain in the PID controller. Setting it to zero cuts off the cumulative path of the error at the moment of impedance change, thus preventing reverse overshoot of pipeline pressure caused by integral overcharging after the airbag is discharged.

[0033] Example 2: This example aims to quantitatively verify the control performance and stability of a multi-dimensional water quality characteristic-driven adaptive operation decision-making system under real complex operating conditions through comparative experiments. The experimental platform is built on a standard test bench simulating a semiconductor ultrapure water terminal supply loop. The test loop includes a 2.2kW stainless steel vertical multistage centrifugal pump as the actuator, whose speed is driven by an industrial-grade frequency converter; a simulated load unit containing a reverse osmosis membrane module is connected in series in the pipeline; the sensor group includes a temperature sensor (accuracy ±0.1) installed on the raw water side. The system includes a response time T90 < 5s, a turbine flow meter (accuracy ±0.5%), and a conductivity meter; a high-frequency pressure sensor (sampling frequency 1kHz, accuracy ±0.25%FS) is installed on the output side. The data acquisition and control algorithm runs in an embedded controller based on an ARM Cortex-M7 core, with a control cycle set to 50ms. To comprehensively evaluate the system's response characteristics under different disturbance types and intensities, this experiment includes a composite test scenario encompassing three dimensions: flow step, temperature shock, and simulated clogging. To quantitatively verify the advantages of this invention's technical solution compared to traditional control strategies, a control group and an experimental group are set up. The control group uses a standard incremental PID control algorithm with calibrated parameters (parameters set to...). The experimental group adopted the adaptive control strategy of the present invention, which includes transient hydraulic impedance calculation, dynamic gain scheduling and thermal hysteresis compensation. During the experiment, the electric regulating valve and the constant temperature water tank were controlled by the programmable logic controller (PLC) to inject a disturbance signal with specific waveform characteristics into the system.

[0034] A flow step disturbance test was conducted. With the system in steady-state operation (set pressure 300 kPa), the raw water flow rate was rapidly increased from 10 L / min to 15 L / min within 2 seconds by quickly adjusting the inlet valve. Data recording showed that in the control group, the pipeline pressure dropped by -18 kPa at the instant of the flow surge. After approximately 15 seconds of oscillation adjustment under PID control, it recovered to steady state, accompanied by an overshoot of +8 kPa. In contrast, in the experimental group, the feedforward compensation unit output a preset voltage increment within the first control cycle after the flow sensor detected the rate of change, causing the pump speed to increase rapidly. The maximum pressure drop in the experimental group was only -4 kPa, with no significant overshoot, and the steady-state recovery time was shortened to less than 3 seconds. Subsequently, a temperature shock test was conducted, simulating the switching between hot and cold fluids. The raw water temperature increased from 20 kPa to 15 kPa within 60 seconds. linearly up to 45 As the water temperature rises, the fluid viscosity decreases, the permeability of the reverse osmosis membrane increases, and the physical impedance of the loop decreases. The control group, unable to perceive this change in physical characteristics, maintains its original control gain, causing the pressure to gradually deviate from the set value. After the temperature stabilizes, a pressure fluctuation of ±12 kPa occurs for about 20 seconds. The experimental group calculates the equivalent temperature of the components through the thermal hysteresis compensation module and corrects the PID gain in real time accordingly. The data shows that during the entire heating process, the pressure fluctuation of the experimental group remains within the range of ±3 kPa and converges rapidly after the temperature stabilizes. Table 1 shows a comparison of the key performance indicators of the two control strategies under different temperature conditions.

[0035] Table 1: Comparison of Control Performance under Variable Temperature Conditions ; As shown in Table 1, the peak pressure fluctuation of the strategy of this invention under variable temperature conditions is only about 25% of that of traditional PID, and the steady-state settling time is shortened by more than 80%, verifying the effectiveness of the thermal hysteresis compensation mechanism in handling systems with large thermal inertia. Finally, a simulated fouling test was conducted, and the system impedance was artificially constructed by adjusting the opening of the damping valve in the simulated load unit. The gradually increasing operating conditions simulate the clogging process of a filter element during long-term operation, when the impedance index... When the pressure was increased to 150% of the initial value, the control group experienced a sustained negative deviation in pressure due to the slow accumulation of the integral term, and its response to the setpoint became sluggish. The dynamic gain scheduling module of the experimental group detected this. When the value enters the high impedance range, the reference proportional gain is automatically adjusted. When the pressure was increased to 1.8 times the initial value, test results showed that under simulated severe fouling conditions, the response time of the test group to a step change in the pressure setpoint increased by only 10%, while that of the control group increased by 45% and exhibited a significant tailing phenomenon. Furthermore, under conditions where an airbag was artificially vented (resistance momentarily dropped), the test group monitored… The integral anti-saturation logic is triggered, limiting the pressure reverse overshoot to within 2 kPa, while the control group shows a dangerous overshoot of more than 25 kPa.

[0036] Example 3: This example combines Figures 1 to 3 This paper describes an adaptive operation decision-making system for end-point purification nodes driven by multi-dimensional water quality characteristics, as follows: Figure 1 As shown, the operating logic of this system begins with the fluid state parameters and response measurements provided by the sensor group at the end purification node, as well as the current drive command value fed back by the actuator. These signals are fed into the data acquisition module, and after synchronous acquisition and timing alignment processing, the generated synchronous state data is transmitted to the transient hydraulic impedance calculation module, which calculates the ratio of drive to response to construct the transient hydraulic impedance index. This index, along with the fluid state parameters, is sent to the dynamic gain scheduling module to map the reference gain and synthesize adaptive PID parameters. Simultaneously, the fluid state parameters are transmitted in parallel to the feedforward compensation unit, which generates feedforward control components based on the fluid state parameters. Finally, the composite drive control module receives the adaptive PID parameters and the feedforward control components, generates drive commands through superposition calculations, and sends them to the actuator to complete the execution of the adjustment commands.

[0037] like Figure 2 As shown, in a coordinate system constructed with the time axis (seconds) and the pressure deviation axis (kPa), the response curves of the conventional PID control and the adaptive control of this invention are compared under the same disturbance conditions. The dashed line represents the conventional PID control, which exhibits a significant negative pressure drop in the early stages of the disturbance, along with obvious positive overshoot and a prolonged oscillating convergence process. The solid line represents the adaptive control, which suppresses the peak pressure fluctuations, shows no obvious overshoot throughout the adjustment process, and returns to the zero-deviation steady-state baseline more quickly. Figure 3 As shown, the system architecture is divided into two interactive parts: a physical entity domain and a digital computing domain. The physical entity domain on the left contains real fluids and hardware components, specifically consisting of actuators composed of variable frequency pumps or valves, end-point purification components composed of reverse osmosis membranes or pipelines, and input sensor groups for temperature / flow / conductivity and response sensor groups for pressure / product water. The digital computing domain on the right is based on an embedded microprocessor ARM / Core, which internally runs algorithm models and decision logic, covering data synchronization and preprocessing, transient hydraulic impedance calculation models, dynamic gain mapping table look-up, thermal hysteresis compensation algorithms, and composite PID control decision modules. The two domains transmit raw sensor data through a multi-dimensional state mapping stream and transmit voltage or PWM commands through an adaptive energy drive stream, thereby constructing a real-time closed-loop control loop.

[0038] Example 4: This example illustrates the specific engineering procedures for the aforementioned multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-of-pipe purification nodes. After initial deployment or replacement of core components such as the reverse osmosis membrane column, the system performs parameterized characterization of system characteristics and initialization of the control model. This eliminates control model mismatch caused by individual component differences and obtains accurate thermal response time constants through a standardized offline self-learning calibration mode. An impedance-gain mapping table is constructed. In the initial state where the system is cold-shutdown and the fluid loop is filled with raw water, thermal inertia characteristic calibration is performed. The control system drives the constant temperature source on the inlet side to generate an amplitude of not less than 10. A temperature step input, i.e., the raw water temperature drops from its initial steady-state value. Instantly switch to Simultaneously with the fluid temperature jump, the system continuously records the input-side fluid temperature observations at a sampling period of 100ms. And the actual temperature response curve of the component body (monitored by a standard contact temperature probe temporarily attached to the surface of the membrane housing, for calibration purposes only). The processor calculates the normalized temperature difference ratio in real time. ,when The first time it reached 0.632 (i.e.) When this happens, the system records the duration of that moment and defines it as the thermal response time constant for that specific component. The processor is based on the formula (in The filter coefficients are automatically calculated and fixed during the system's control cycle. This completes the parameterized configuration of the thermal hysteresis compensation module, ensuring the equivalent temperature of the components during subsequent operation. The calculations have a real physical basis.

[0039] Next, the system enters the impedance-gain mapping table construction stage. In this stage, the electric regulating valve on the end outlet side is used as an active load simulator to artificially construct a flow resistance state covering the entire operating range. The processor controls the variable frequency pump to run at a constant speed and fully opens the regulating valve, measuring the drive command at this time. With fluid response The reference hydraulic impedance of the system was calculated. The system follows a preset gradient (e.g., with...). (Using 10% as the step size), gradually reduce the opening of the regulating valve until the resistance value reaches... (Set as) (simulating severe fouling conditions) at each discrete impedance test point. The system suspends valve operation and triggers a relay feedback self-tuning process under this condition; during this self-tuning process, the controller temporarily disconnects the PID loop, replacing it with an amplitude of The relay nonlinear element drives the variable frequency pump, inducing constant-amplitude critical oscillations in the fluid pressure. The processor analyzes the oscillation waveform to extract the critical gain of the system. and critical period Based on the Ziegler-Nichols frequency domain response rule, the system calculates the impedance point. The reference proportional gain coefficient that can ensure system stability and has the fastest response is the following. The system iterates through all impedance test points and obtains a series of paired data. The data is stored in non-volatile memory and a continuous impedance-gain mapping table is generated through piecewise linear fitting.

[0040] The system is constructed in a non-volatile storage space with transient hydraulic resistance index For index key, reference proportional gain coefficient The array is a two-dimensional discrete array in the numerical domain. The index step size is set according to the minimum resolution of the actuator and the noise tolerance of the sensor, with a minimum interval of 3% to 5% of the rated total impedance of the system. During the real-time control cycle, the processor performs a first-order linear interpolation operation based on two adjacent points on the measured values ​​of the transient hydraulic impedance index between two discrete index keys to calculate the reference proportional gain coefficient for the current operating condition. This eliminates the step of control parameters caused by the switching of the lookup interval. When the extreme impedance exceeds the coverage range of the array, the corresponding gain value of the boundary index is called to trigger the state flag. The controller freezes the accumulation of the integral term accordingly and switches the control strategy to a conservative mode containing only proportional and derivative terms to prevent divergent oscillations caused by model mismatch in the uncalibrated physical interval. The value of the reference proportional gain coefficient is determined according to the digital implementation path of the Ziegler-Nichols frequency domain response criterion. In the calibration mode, the actuator is driven to generate amplitude setpoints at each discrete impedance point. The processor extracts the oscillation period for critical constant amplitude oscillation using a zero-point detection algorithm. With critical gain According to the formula The gain at this point is calculated to accommodate the physical characteristic drift caused by component aging. During the first hour of operation after each cleaning and maintenance cycle, the deviation ratio between the current average impedance and the original average impedance is automatically collected and calculated. When the deviation ratio exceeds the preset dead zone threshold of 5%, the processor performs an overall translation correction on all gain coefficients of the two-dimensional discrete array. The correction magnitude is proportional to the impedance deviation ratio.

[0041] Example 5: This example details the defensive control strategies and emergency response procedures of the aforementioned adaptive operation decision system under non-ideal operating conditions such as extreme boundary conditions, hardware failures, and network anomalies. Through risk contingency plan deployment, it ensures that the system can maintain the operation of the minimum safe function set under unforeseen physical or information shocks, preventing catastrophic shutdowns or water quality safety accidents. The system has a built-in sensor data confidence assessment mechanism to defend against control logic misjudgments caused by the failure of a single sensor. In each control cycle, the processor reads the raw water temperature, flow rate, and conductivity data from the input fluid state parameters and calculates the first-order difference of these parameters. If the instantaneous rate of change of a parameter exceeds a reasonable threshold set based on physical inertia, for example, the rate of change of water temperature exceeds 5... If the flow rate changes by / s or exceeds the pump's maximum response speed, the system will determine that the sensor data is unreliable and activate the data freeze and extrapolation mode. In this mode, the system temporarily blocks the real-time reading of the faulty sensor and uses the effective value of the previous moment or the linear extrapolation value based on historical trends as the control input. At the same time, the system activates the backup sensor (if it exists) or triggers an audible and visual alarm to prompt maintenance personnel to intervene.

[0042] In addition, to address the potential risks of jamming or loss of control of the actuator, the system is configured with driver-response consistency verification logic, and the transient hydraulic impedance calculation module monitors the driver command values ​​in real time. Compared with the output-side fluid response measurement value The correlation, if in continuous Within each control cycle, such as The transient hydraulic impedance index is a phenomenon where the driving command changes drastically while the fluid response remains constant. If the value approaches infinity or zero, the system determines that the actuator has suffered a mechanical failure or the pipeline is completely blocked. At this time, the system executes the emergency shutdown procedure, cuts off the power supply to the pump and closes the inlet valve to prevent the pump body from overheating or the pipeline from bursting. In addition, to cope with the risk of network communication interruption or loss of upper computer instructions, the embedded controller is configured to have independent autonomous operation capability. When the system detects that the heartbeat signal with the upper computer has been lost for more than a preset time limit, such as 30 seconds, it automatically switches to the safety maintenance mode. In this mode, the system abandons the pursuit of the ultimate dynamic response speed and instead adopts a set of preset conservative control parameters to ensure that the water quality index of the output water is not lower than the minimum qualified line until communication is restored or manual intervention is required.

[0043] Example 6: This example details the adaptive operation decision-making system for end-of-pipe purification nodes driven by multi-dimensional water quality characteristics. After initial deployment or major component replacement, it executes standardized system initialization and full-condition benchmark calibration procedures. Through active physical excitation and data traversal, it acquires the fingerprint characteristics of the system under the current physical configuration, completes the adaptive filling of core control parameters, and eliminates control model mismatch caused by differences in pipeline geometry and the discreteness of pump and valve characteristics. After the system completes physical connection and electrical self-test, it enters the flow-pressure hydraulic model construction stage. The controller puts the variable frequency pump in open-loop control mode and sets the initial PWM duty cycle to 10%. While keeping the outlet regulating valve fully open, the controller gradually increases the PWM duty cycle in 5% increments until it reaches 95%. At each duty cycle level, the system waits for 3 seconds to ensure fluid stability, and then synchronously collects the turbine flow meter readings. With the reading of the circuit pressure sensor The system will record discrete data points The data is stored in non-volatile memory, and the flow-pressure characteristic curve equation of the current pipeline characteristics is fitted using the least squares method. This curve serves as the feedforward compensation unit to query the feedforward control components during subsequent operation. The physical reference ensures that the feedforward quantity is precisely matched with the actual tube resistance characteristics.

[0044] During the impedance-gain mapping table optimization phase of the system execution, the controller places the system in closed-loop pressure control mode, sets the target pressure to a process standard value such as 300 kPa, and uses the outlet regulating valve as an active load simulator. Starting from the fully open position, the opening degree is gradually reduced in steps of 2%, artificially constructing a global transient hydraulic impedance index that ranges from low resistance and rapid response to high resistance and sluggish response. The sequence, at each valve opening corresponding to the impedance test point Upon activation, the controller automatically triggers a critical oscillation test based on relay feedback. The system temporarily disconnects the PID loop and replaces it with an amplitude of [missing value]. The relay nonlinear element drives the variable frequency pump, inducing constant-amplitude critical oscillations in the pipeline pressure. The processor analyzes the oscillation waveform in real time and extracts the critical gain of the system. and critical period Based on the Ziegler-Nichols frequency domain response rule, the system calculates the impedance point. The reference proportional gain coefficient that can ensure system stability and has the fastest response is the following. The system iterates through all impedance test points, constructs and solidifies a continuous impedance-gain mapping table, thereby improving the dynamic gain scheduling module's adaptability to component physical characteristic drift throughout its lifecycle. Finally, the system enters the abnormal boundary fingerprinting stage, where the controller drives the system at extreme flow rates, such as 110% of the pump's rated flow, and extreme temperatures, such as 60°C. During a short-term run, the transient hydraulic impedance index was recorded. and its first derivative The system sets 1.2 times the peak value as the threshold for mechanical fault judgment and airbag collapse judgment during operation, thus establishing the quantitative boundary of the system's fault diagnosis logic and preventing missed reports due to excessively wide threshold settings or false reports due to excessively narrow threshold settings.

[0045] 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.

[0046] 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 multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-of-pipe purification nodes, characterized in that the system... include: The data acquisition module is used to synchronously acquire the input-side fluid state parameters of the end-of-line purification node in each control cycle, and obtain the current drive command value of the actuator and the output-side fluid response measurement value; the input-side fluid state parameters include raw water temperature, raw water flow rate and raw water conductivity; The transient hydraulic impedance calculation module is used to calculate the ratio of the drive command value to the output-side fluid response measurement value, and obtain the transient hydraulic impedance index characterizing the current flow resistance characteristics of the fluid loop; The dynamic gain scheduling module stores a mapping table between the transient hydraulic impedance index and the reference proportional gain coefficient. This module is used to find the reference proportional gain coefficient corresponding to the current operating condition from the mapping table based on the transient hydraulic impedance index calculated in real time, and to determine the adaptive PID control parameters at the current moment in combination with the fluid state parameters on the input side. The composite drive control module is used to calculate the deviation between the output-side fluid response measurement value and the set value based on the adaptive PID control parameters to generate a feedback control component. The feedback control component is then superimposed with the feedforward control component calculated based on the input-side fluid state parameters to generate a drive command value for adjusting the actuator.

2. The multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-point purification nodes according to claim 1, characterized in that, The transient hydraulic impedance calculation module is also used to calculate the time rate of change of the transient hydraulic impedance index; the dynamic gain scheduling module includes an integral anti-saturation unit, which is used to set the integral term gain in the adaptive PID control parameters to zero when the time rate of change of the transient hydraulic impedance index exceeds a preset threshold, until the transient hydraulic impedance index stabilizes within the preset range.

3. The multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-of-pipe purification nodes according to claim 1, characterized in that, The transient hydraulic impedance calculation module calculates the transient hydraulic impedance index using the following formula: ,in, For the first Transient hydraulic impedance index for each control cycle, For the first The drive command value applied to the actuator in each control cycle, For the first The output-side fluid response measurement values ​​are collected in each control cycle. A preset constant is used to prevent the denominator from being zero.

4. The multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-of-pipe purification nodes according to claim 1, characterized in that, The system also includes a thermal hysteresis compensation module; this module stores the thermal response time constant characterizing the thermal capacity of the terminal purification node components, and is used to perform first-order low-pass filtering on the raw water temperature in the input fluid state parameters based on the thermal response time constant to obtain the component equivalent temperature; the dynamic gain scheduling module is used to replace the raw water temperature with the component equivalent temperature as the basis for correcting the adaptive PID control parameters.

5. The multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-point purification nodes according to claim 1, characterized in that, The mapping table adopts a piecewise nonlinear setting: when the transient hydraulic impedance index is higher than the preset high resistance threshold, a higher reference proportional gain coefficient is applied; when the transient hydraulic impedance index is lower than the preset low resistance threshold, a lower reference proportional gain coefficient is applied; the high resistance threshold and low resistance threshold are determined based on the calibrated impedance value of the end-of-pipe purification node.

6. The multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-of-pipe purification nodes according to claim 1, characterized in that, The system also includes a maintenance effect evaluation module; this module is used to obtain performance recovery indicators after cleaning and maintenance are performed at the end purification node, compare them with preset benchmark values, and decompose the long-term drift of the drive command value into recoverable fouling components and unrecoverable aging components. The dynamic gain scheduling module is used to correct the base value of the mapping table based only on the aging component, and to reset the mapping table when the detected contamination component exceeds the preset reset threshold.

7. The multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-point purification nodes according to claim 1, characterized in that, The data acquisition module includes a data synchronization unit; this unit is used to time-align the input-side fluid state parameters and the output-side fluid response measurements according to the sampling frequency and transmission delay of different sensors.

8. The multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-of-pipe purification nodes according to claim 1, characterized in that, The composite drive control module includes a feedforward compensation unit; this unit stores a flow-pressure correspondence table based on a hydraulic model, which is used to look up the pressure compensation value as a feedforward control component based on the rate of change of the raw water flow in the input fluid state parameters.

9. The multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-of-pipe purification nodes according to claim 1, characterized in that, The actuator is a variable frequency pump or a regulating valve; the drive command value is the PWM duty cycle that controls the variable frequency pump or the voltage value that controls the regulating valve; the output side fluid response measurement value is the pipeline pressure value or the product water flow rate value.

10. The multi-dimensional water quality characteristic-driven adaptive operation decision-making system for end-of-pipe purification nodes according to claim 1, characterized in that, The dynamic gain scheduling module is also used to execute impedance anomaly protection strategies; when the transient hydraulic impedance index is lower than the minimum safe value for a certain period of time, it is determined that there is a pipe disconnection or sensor failure in the fluid circuit, and a shutdown command is output to cover the drive command value.

Images