A water purification terminal flow rate self-adaptive adjusting system for non-steady pipe network environment

Abstract

This invention relates to the field of control or regulation systems for non-electrical variables, and discloses an adaptive flow rate regulation system for water purification terminals in non-steady-state pipe network environments. The system includes an input state sensing unit, a terminal response monitoring unit, a flow resistance modulation execution unit, and an adaptive control unit. The system utilizes the adaptive control unit to address the basic drive duty cycle in a preset pressure-flow characteristic spectrum and executes high-frequency chatter injection logic to generate a zero-mean chatter component with a frequency higher than the mechanical cutoff frequency of the execution unit. This component is then superimposed on the basic drive duty cycle to generate the final drive waveform. The chatter amplitude follows an adjustment strategy negatively correlated with the basic drive duty cycle. This invention, through chatter injection and dynamic amplitude adjustment, maintains the contact friction state of the actuator as dynamic friction, eliminating the static friction dead zone of the solenoid valve without increasing high-cost hardware, and achieving a hysteresis-free response to minute feedforward commands.

Description

Technical Field

[0001] This invention relates to an adaptive flow velocity control system for water purification terminals in non-steady-state pipe network environments, belonging to the technical field of control or regulation systems for non-electrical variables. Background Technology

[0002] In current urban water supply terminals and industrial water purification processes, maintaining a constant effluent flow rate to ensure filtration efficiency and membrane module lifespan often relies on direct-acting electromagnetic proportional valves based on throttling principles as actuators. These valves are often combined with PID closed-loop feedback or static feedforward control strategies to adjust valve opening based on flow or pressure sensor feedback signals, thus offsetting inlet pressure fluctuations. Such solutions achieve basic flow rate regulation under relatively stable inlet pressure conditions. However, in actual unsteady-state pipe network environments, high-frequency, large-amplitude transient pressure pulsations occur at the inlet due to factors such as pressurization from high-rise secondary water supply, interference from neighboring water usage, or water hammer effects. Due to cost constraints, the actuators in these systems are mostly low-cost direct-acting electromagnetic valves, rather than high-precision servo valves with position closed-loop control. These valves are limited by machining precision and sealing structure, and exhibit static friction and hysteresis nonlinearity at the physical level.

[0003] When the control system performs precise compensation for minute pressure fluctuations, the calculated incremental drive signal cannot generate sufficient electromagnetic force to overcome the static friction of the valve core. This causes the valve core to remain stationary, creating a control dead zone. Simply relying on the cumulative error of the integral term to exceed the friction threshold results in a sudden displacement of the valve core at the moment of exceedance, leading to flow rate overshoot or low-frequency limit loop oscillation. The nonlinear hysteresis determined by the physical characteristics of the hardware causes conventional linear control algorithms to fail when processing minute signals, making it impossible to achieve real-time smooth decoupling of high-frequency pressure disturbances without increasing hardware costs. In addition to the physical defects of the actuator itself, the adaptive strategy based on conventional logic judgment in the existing technology has the problem of too narrow a response bandwidth, making it difficult to adapt to transient pipeline environments. For example, Chinese invention patent application with publication number CN108946874A discloses a... The adaptive water purification method measures the water quality and actual pressure in the inlet pipe, classifying the operating conditions into different water quality types and optimal working pressure ranges. Based on this, it adjusts the booster pump power and drain flow rate accordingly. Although this method can adapt to the water quality and pressure benchmarks of different regions at the overall level, the control logic is essentially a step-by-step adjustment based on the steady-state range. The system only focuses on whether the pressure deviates from the preset static threshold range. Faced with millisecond-level high-frequency water hammer impacts or micro-amplitude rapid pulsations in the pipe network, this low-frequency control loop based on lookup tables and threshold judgments cannot provide real-time compensation by fine-tuning the valve opening in the early stage of disturbance transmission, nor can it provide a high-frequency excitation signal that can shake the valve core to eliminate the static friction dead zone. As a result, the system still has lag and nonlinear errors in fine flow control.

[0004] Therefore, the technical problem to be solved by this invention is how to overcome the dead zone of physical static friction of low-cost solenoid valves without using expensive servo actuators, and solve the problems of response lag and model mismatch in the control system under non-steady-state pipeline network environment. 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 water purification terminal flow velocity adaptive adjustment system for unsteady-state pipe network environments, comprising:

[0006] The inlet water pressure detection unit is located at the fluid input end of the water purification system and is used to collect inlet water pressure signals.

[0007] The flow regulation actuator is located in the flow path between the fluid input end and the fluid output end, and has a direct-acting electromagnetic proportional valve driven by a pulse width modulation signal.

[0008] The adaptive control unit is connected to both the inlet pressure detection unit and the flow regulation execution unit. The adaptive control unit executes flow control logic incorporating a negative amplitude correlation adjustment strategy: it determines the base drive duty cycle by querying a preset pressure-flow characteristic spectrum based on the inlet pressure signal; this base drive duty cycle is used to set the target opening of the electromagnetic proportional valve; it executes high-frequency chatter injection logic to generate a zero-mean chatter component with a frequency higher than the mechanical cutoff frequency of the electromagnetic proportional valve, and calculates the dynamic amplitude of the chatter component based on the current base drive duty cycle; the calculation of the dynamic amplitude follows a negative amplitude correlation adjustment strategy: when the base drive duty cycle is in the high flow resistance range, the dynamic amplitude is set to the first amplitude; when the base drive duty cycle is in the low flow resistance range, the dynamic amplitude is set to the second amplitude, with the first amplitude being greater than the second amplitude; the base drive duty cycle is superimposed with the chatter component with the dynamic amplitude to generate the final drive waveform, which is output to the flow regulation execution unit. This drives the electromagnetic proportional valve to maintain a small reciprocating motion while maintaining the target opening, thus maintaining the contact friction state of the electromagnetic proportional valve core as dynamic friction.

[0009] Preferably, when generating the zero-mean flutter component, the adaptive control unit sets the frequency of the flutter component to a fixed frequency of 200 Hz to 300 Hz. This frequency is set to maintain a non-multiplication synchronization relationship with the carrier frequency of the pulse width modulation signal, and this frequency is higher than the mechanical resonance frequency of the water purification system, so as to suppress fluid pulsation noise while eliminating the static friction dead zone of the electromagnetic proportional valve.

[0010] Preferably, the adaptive control unit is also used to execute asymmetric gain scheduling logic based on the direction of pressure change rate: calculating the time change rate of the inlet pressure signal; when the time change rate is positive and the absolute value is greater than a preset sudden change judgment threshold, the system is determined to be in a water hammer impact state, the filtering link for the inlet pressure signal is bypassed and the dynamic damping correction is calculated using a first gain coefficient, and the dynamic damping correction is superimposed on the basic drive duty cycle to form a fast response control loop; when the time change rate is negative or the absolute value is less than or equal to the sudden change judgment threshold, the system is determined to be in a normal fluctuation state, the filtering link for the inlet pressure signal is activated and the dynamic damping correction is calculated using a second gain coefficient to form a smooth response control loop; wherein, the value of the first gain coefficient is greater than that of the second gain coefficient.

[0011] Preferably, the adaptive control unit is also used to execute the online correction logic of the pressure-flow characteristic spectrum: under steady-state operation, it acquires a three-dimensional data set including the inlet pressure signal, the final drive waveform, and the actual outlet flow velocity signal; calculates the flow resistance residual between the measured flow resistance based on the three-dimensional data set and the theoretical flow resistance obtained from the pressure-flow characteristic spectrum query; and extracts the physical characteristic drift of the flow regulation actuator according to the following aging trend factor calculation formula: ,in, As an aging trend factor, The sampling window length, For the first The measured flow resistance during the second sampling. For the first The theoretical flow resistance at the time of the second sampling; when the absolute value of the aging trend factor exceeds the preset update threshold, the mapping node data in the pressure-flow characteristic spectrum is weighted and corrected using the aging trend factor.

[0012] Preferably, the pressure-flow characteristic spectrum preset in the adaptive control unit is constructed as a three-dimensional discrete data table. The three axes of this data table are the inlet pressure axis, the target flow velocity axis, and the basic drive duty cycle axis. When querying, the adaptive control unit calculates the corresponding basic drive duty cycle in the three-dimensional discrete data table based on the real-time collected inlet pressure signal and the preset target flow velocity value using a bilinear interpolation algorithm.

[0013] Preferably, the system further includes an outlet flow velocity detection unit located at the fluid output end for acquiring the actual outlet flow velocity signal; the adaptive control unit is also used to calculate the steady-state deviation between the actual outlet flow velocity signal and the target flow velocity, calculate the steady-state correction using a proportional-integral control algorithm, and add the steady-state correction to the basic drive duty cycle; when calculating the steady-state correction, the adaptive control unit only starts the integral term calculation of the proportional-integral control algorithm when the rate of change of the inlet pressure signal is less than the steady-state determination threshold, so as to prevent integral saturation during transient pressure disturbances.

[0014] Preferably, when the adaptive control unit executes the amplitude negative correlation adjustment strategy, it defines the high flow resistance range as the range where the basic drive duty cycle is less than 30%, and the low flow resistance range as the range where the basic drive duty cycle is greater than 70%; the flutter signal intensity corresponding to the first amplitude is set to the duty cycle amplitude corresponding to the minimum current increment that can overcome the maximum static friction of the electromagnetic proportional valve; the flutter signal intensity corresponding to the second amplitude is set to 50% to 80% of the first amplitude.

[0015] Preferably, when the adaptive control unit determines that the system is in a water hammer impact state, it is also used to execute forced cutoff logic: if the sum of the calculated basic drive duty cycle and the dynamic damping correction is less than the minimum maintenance duty cycle, the duty cycle signal output to the flow regulation execution unit is forcibly set to zero until the time change rate of the inlet pressure signal recovers to within the sudden change judgment threshold.

[0016] Preferably, the inlet pressure detection unit includes a ceramic piezoresistive pressure sensor, which is directly installed in the upstream flow path of the flow regulation actuator; the electromagnetic proportional valve of the flow regulation actuator is a normally closed structure, and its valve core cuts off the flow path by the force of the reset spring when there is no drive signal; during the system power-on initialization phase, the adaptive control unit outputs a fully open pulse signal with a preset duration to eliminate the viscous resistance generated by the valve core after a long period of static placement.

[0017] Preferably, the adaptive control unit integrates a microprocessor and a power drive circuit; the microprocessor is used to perform query, calculation and signal generation logic; the power drive circuit is used to receive the final drive waveform and convert it into a drive current output to the coil of the electromagnetic proportional valve; the system also includes a regulated power supply module for supplying power to the power drive circuit, the regulated power supply module having an overcurrent protection circuit to prevent instantaneous current peaks generated during the superposition of flutter components from damaging circuit components.

[0018] Compared with the prior art, the beneficial effects of the present invention are:

[0019] 1. A high-frequency zero-mean chatter component negatively correlated with the basic duty cycle is superimposed on the main drive signal of the flow resistance modulation actuator to maintain the valve core's micro-amplitude reciprocating motion. The vibration is used to change the contact properties between the valve core and the valve body from high-resistance static friction to low-resistance dynamic friction, eliminating the mechanical start-up dead zone and hysteresis effect of the direct-acting electromagnetic proportional valve. Based on the system's microsecond-level micro-amplitude feedforward compensation command instantaneous execution capability, general-purpose hardware is used to improve the small signal input displacement resolution of the actuator, ensuring real-time decoupling accuracy for high-frequency micro-pressure pulsations in the pipeline network.

[0020] 2. A pressure gradient polarity asymmetric gain scheduling strategy is adopted. The control loop response characteristics are dynamically reconstructed according to the sign of the influent pressure change rate. When the forward pressure surges, the bypass filter is applied and a high feedforward gain is added to construct a high-damping fast-response channel to cut off the mechanical impact of water hammer pressure waves on the reverse osmosis membrane. When the pressure drops or fluctuates steadily, the channel is switched to a low-gain channel with filtering to keep the flow rate regulation gentle. The risk asymmetric control logic is used to solve the conflict between rapidly blocking sudden high pressure and maintaining a constant bandwidth for steady-state flow rate, ensuring the mechanical safety of the core filter components and avoiding flow path oscillation caused by over-regulation.

[0021] 3. An online correction module for the hydraulic admittance spectrum is set up. A closed-loop observer is constructed using the steady-state operating pressure, flow rate, and duty cycle data of the system. The residual between the physically measured hydraulic resistance and the theoretical hydraulic resistance of the spectrum is calculated and the time series trend is extracted. The slow-changing physical drift factors that characterize valve core wear, spring fatigue, or scale deposition are identified. The nodes of the preset hydraulic admittance spectrum are updated nonlinearly with weights. The control model is adaptively tracked and compensated for the physical aging of the actuator by secondary mining of historical data, so as to ensure that the feedforward control accuracy does not decrease with the decay of hardware characteristics during the long-term service life of the system. Attached Figure Description

[0022] Figure 1 This is a block diagram of the flow rate adaptive control logic for dynamic flutter injection in this invention;

[0023] Figure 2 This is a comparison chart of the flow velocity response and fluctuation suppression effect of the present invention under unsteady pipeline pressure.

[0024] Figure 3 This is a flowchart illustrating the execution of the asymmetric gain scheduling strategy for the pressure gradient polarity of this invention. Detailed Implementation

[0025] The following examples are intended to illustrate the present invention and are not intended to limit the scope of protection of the present invention in any way.

[0026] This invention proposes an adaptive flow velocity adjustment system for end-point water purification in unsteady-state pipe network environments. It comprises an input state sensing unit at the fluid input end, a flow resistance modulation execution unit in the middle of the flow path, an end-point response monitoring unit at the fluid output end, and an adaptive control unit. Addressing the physical defect of static friction dead zone in ordinary direct-acting electromagnetic proportional valves during small-aperture adjustment, the adaptive control unit incorporates dynamic chattering injection logic based on flow resistance status. Given the inherent static friction between the valve core and body of the electromagnetic valve, which is typically greater than dynamic friction, when the incremental electromagnetic force corresponding to the small adjustment command output by the control unit is insufficient to overcome the maximum static friction, the valve core remains stationary. To solve this problem, the adaptive control unit integrates a high-frequency signal generator, whose generation frequency is set to [missing information]. to The zero-mean periodic chatter component, the frequency of which is selected and set to be higher than the cutoff frequency of the electromagnetic proportional valve core mechanical components, that is, usually lower than... The frequency range is set to ensure that the signal energy only causes high-frequency jitter of the valve core at the microscopic level, without generating an overall displacement that can change the fluid flux. At the same time, the frequency is set to maintain an asynchronous relationship with the carrier frequency of the pulse width modulation signal without being a multiple of the carrier frequency. In order to realize this asynchronous feature and suppress beat frequency noise in digital control, the adaptive control unit adopts a dual-rate interrupt scheduling mechanism: the microprocessor's main PWM timer triggers the control algorithm update at 100Hz, while the sub-timer responsible for generating the jitter component looks up the preset 233Hz sine wave discrete point table at a high sampling rate of 1kHz. When the main PWM cycle is updated, the system directly reads the current jitter value of the sub-timer and accumulates it with the base duty cycle. Since the jitter update frequency (1kHz) is much higher than the carrier frequency (100Hz) and the two are not integer multiples of each other, the periodicity of phase superposition is broken in the time domain. That is, when the pulse width modulation carrier is 100Hz, the jitter frequency is set to a non-harmonic frequency such as 233Hz.

[0027] Regarding the amplitude modulation of the flutter signal, the system implements an amplitude negative correlation adjustment strategy, and the adaptive control unit monitors the current base drive duty cycle in real time. ,when Less than When the valve is in a high flow resistance range, the unbalanced hydrodynamic force on the valve core and the clamping force of the sealing ring cause the static friction barrier to rise. The system then sets the dynamic amplitude of the chatter component to the first amplitude, and the corresponding peak value of the electromagnetic force fluctuation generated by the current is set to be greater than the static friction force of the valve core under maximum working pressure. This forces the valve core to always be in a state of micro-amplitude reciprocating motion, maintaining the frictional properties between the contact surfaces as low-resistance dynamic friction. Greater than When the valve is determined to be in the low flow resistance range, the system switches the dynamic amplitude to the second amplitude and sets it back to the first amplitude. to ,for In the transition range between 30% and 70%, the system executes a linear interpolation mapping strategy: as the basic drive duty cycle increases linearly, the amplitude of the chatter component decreases uniformly from the first amplitude to the second amplitude. Specifically, in each control cycle, the system calculates the position ratio of the current point in the 30%-70% range, and takes a weighted average between two preset amplitude constants based on this ratio to ensure a smooth transition of the friction state of the solenoid valve core during the opening switching process, eliminating the risk of flow jump caused by abrupt amplitude changes, so as to reduce coil heat generation and electromagnetic noise while maintaining the dynamic friction state. The chatter component of this dynamic amplitude is linearly superimposed on the basic drive duty cycle to generate a composite drive waveform and output to the flow resistance modulation execution unit.

[0028] To address the asymmetry of pressure fluctuation risks in unsteady pipeline environments—that is, sudden pressure surges caused by water hammer may lead to physical damage to the filter membrane, while pressure drops only affect water production efficiency—the system establishes an asymmetric gain scheduling logic based on the polarity of the pressure gradient. The input state sensing unit employs a response time superior to... Ceramic piezoresistive pressure sensors, with The sampling rate is used to collect the inlet water pressure signal. The differentiator within the adaptive control unit calculates the time rate of change of the signal. ,Right now Before performing the differential operation, the system performs a 5-point moving average smoothing process on the pressure signal acquired at a 1kHz sampling rate, that is, takes the average pressure value of the current moment and the previous four moments as the effective input to eliminate high-frequency sensor noise; when calculating the rate of change of time, the system takes the current effective input value minus the effective input value 20ms ago and divides it by the time interval; when the calculated value exceeds 0.05MPa / s for two consecutive calculation cycles and the direction is positive, the system determines that water hammer impact has occurred and triggers a high-gain fast response loop. For example, a positive value whose absolute value exceeds the preset mutation determination threshold. When the system determines that it is currently in a water hammer impact state, it activates the fast response control loop, in which the system bypass drops the target The low-pass filter stage directly uses the instantaneous pressure value and calls the first gain coefficient, which has a relatively large value. Calculate the dynamic damping correction, where the first gain coefficient is... The preset value is a constant of 1.2, used to directly convert the pressure change rate into a negative correction offset of the duty cycle; under water hammer conditions, the system... Multiplying by the pressure change rate yields the required reduction in duty cycle, which is then immediately added to the basic drive command; in contrast, in the smooth response loop, the second gain coefficient... The default value is 0.3; this fixed-coefficient scaling ensures that the system can generate a driving torque sufficient to cover the hysteresis of the electromagnetic coil when facing a sudden high-pressure pulse, allowing the actuator to shut down before the pressure peak arrives; execute the instruction-level reset procedure of the FIFO buffer queue: set the read / write pointer of the moving average filter to zero, using the current instantaneous pressure value. The entire historical data in the buffer is overwritten to eliminate group delay; the shadow register preloading function of the PWM timer is synchronously disabled, forcing the calculated dynamic damping correction to take effect immediately in the comparison match event of the current clock cycle, avoiding control lag in the next cycle. This correction is directly added to the drive instruction, driving the valve to close to cut off the pressure wave transmission. When the value is negative or its absolute value does not exceed the above threshold, the system determines that it is in a normal fluctuation or steady state. At this time, the smooth response control loop is activated. Perform a moving average filter and use a smaller second gain coefficient. ,in Less than 0.5 times .

[0029] To address the open-loop model mismatch issue caused by valve core wear, spring fatigue, or scale buildup during long-term operation, the system incorporates online correction logic for the hydraulic admittance map. The system is pre-loaded with a three-dimensional discrete data table as the initial pressure-flow characteristic map, i.e., an H-Map. The three axes of this map represent the inlet pressure, target flow velocity, and basic drive duty cycle, respectively. When the system is in steady-state operation—that is, when both the inlet pressure change rate and the outlet flow velocity fluctuation rate are below their respective steady-state thresholds—the adaptive control unit periodically collects a set of three-dimensional data: current inlet pressure... The actual total drive duty cycle of the output and actual outflow velocity The system calculates the current physical measured flow resistance. At the same time, based on the current The theoretical flow resistance can be obtained by reverse lookup and interpolation in the graph. The system calculates the aging tendency factor, which characterizes the slow drift of physical properties, based on the following formula. ,in, The set sampling window length, For sampling sequence number, For the first The measured flow resistance during the second sampling. For the first The theoretical flow resistance at the next sampling point is used to execute a micro-update logic for the graph nodes based on local bilinear weights: locking the envelope at the current operating point. Four three-dimensional discrete nodes Calculate spatial mapping weights based on Euclidean distance Perform nonlinear iteration of node values Set the learning rate And force a single iteration to update the amplitude limit. To prevent the spectral model from diverging due to transient sampling noise, when the calculated... When the absolute value exceeds the preset update threshold, it indicates that the physical characteristics of the actuator have shifted. The system uses this factor to weight and correct the mapping node data in the H-Map, adjusting the duty cycle baseline values ​​of the corresponding pressure and velocity nodes in the map according to... The ratio is shifted in the opposite direction to offset the control errors caused by hardware aging at the software level.

[0030] To address the potential issue of prolonged static stagnation during system startup, the adaptive control unit executes a valve core activation procedure during each power-on initialization process, with an output duration of [duration missing]. to The fully open pulse signal, i.e. The duty cycle signal, using maximum electromagnetic force, breaks through the adhesion layer formed by scale or impurities between the valve core and valve seat, and the system enters soft-start mode. The duty cycle begins by linearly increasing the drive signal at a preset slope until the base drive duty cycle obtained from the current inlet pressure is reached. In the hardware deployment, the flow resistance modulation execution unit uses a normally closed direct-acting electromagnetic proportional valve. Its coil drive circuit is equipped with a current closed-loop control module. The adaptive control unit uses a microprocessor, such as a microcontroller based on the ARM Cortex-M0 core, to generate a pulse width modulation drive waveform using a timer. After being amplified by a power metal-oxide-semiconductor field-effect transistor, the waveform drives the solenoid valve coil. The power input terminal of the drive circuit is connected in series with an overcurrent protection circuit with a response time in the microsecond range. The end response monitoring unit uses a Hall effect water flow sensor to provide a low-frequency closed-loop feedback signal to correct the steady-state error caused by sensor zero-point drift. Its integral regulation function is only activated when the system determines that it is in steady-state mode to prevent integral saturation during transient disturbances.

[0031] Example 1: In a 400G reverse osmosis water purification system applied to the secondary water supply environment of high-rise residential buildings, the inlet water pressure signal is collected in real time by the input status sensing unit. It exhibits characteristics of high-frequency, large-amplitude pulsation, often accompanied by water hammer and pressure drops caused by the water usage behavior of nearby users. Faced with such unsteady conditions, the low-cost direct-acting electromagnetic proportional valve used in the flow resistance modulation actuator is limited by its physical structure. Static friction and hysteresis exist between its valve core and body, causing it to often fail to produce effective displacement when receiving small opening adjustment commands, creating a response dead zone. To overcome this physical hysteresis, the adaptive control unit executes dynamic chatter injection logic based on the flow resistance state, adjusting the basic drive duty cycle... The superposition frequency is to The zero-mean chatter component, when the system is in the high flow resistance range, i.e. the valve opening is small, the adaptive control unit automatically outputs a higher amplitude chatter component according to the amplitude negative correlation adjustment strategy. The electromagnetic force fluctuation generated by this signal drives the valve core to maintain a small reciprocating motion near the equilibrium position, changing the contact friction between the valve core and the valve seat from high-resistance static friction to low-resistance dynamic friction. This physical reconstruction of the friction state enables the actuator to make an instantaneous response to the millisecond-level tiny feedforward compensation command.

[0032] When a violent water hammer impact occurs at the inlet, it causes the inlet pressure signal to be lost. rate of change over time Exceed When the mutation threshold is reached, the system faces a bandwidth conflict between mechanical damage to the filter membrane and constant flow rate control. At this time, the adaptive control unit activates the asymmetric gain scheduling logic based on the pressure gradient polarity, automatically bypasses the low-pass filter for the inlet water pressure, and calls the first gain coefficient with a larger value. The dynamic damping correction is calculated, and the high-gain feedforward control loop drives the flow resistance modulation actuator to rapidly reduce the opening, completing the flow path cutoff before the pressure peak reaches the reverse osmosis membrane surface. This rapid actuator action mitigates the risk of transient overpressure. When the pressure fluctuation is gentle or in a negative gradient drop, the system automatically switches to using the second gain coefficient. The smooth response control loop ensures a gentle and overshoot-free adjustment of the outlet flow rate, maintaining the stability of the flow field at the water supply end. After months of continuous operation, scale gradually accumulates at the valve port of the flow resistance modulation actuator due to water hardness, causing irreversible slow drift in its physical flow resistance characteristics. This leads to model mismatch in the preset pressure-flow characteristic spectrum. The adaptive control unit uses online correction logic based on the hydraulic admittance spectrum to periodically collect the inlet water pressure during steady-state operation. Total drive duty cycle and actual outflow velocity And calculate the physical measured flow resistance. Theoretical flow resistance obtained from the table The system generates an aging trend factor based on the deviation between them. This factor is then used to perform nonlinear weighted correction on the mapping node data in the H-Map.

[0033] Example 2: In the test platform used to verify the effectiveness of the technical solution of the present invention, a test loop simulating the unsteady water supply network environment of a high-rise residential building was constructed. The fluid input end of the loop was connected to a programmable pressure pulsation generator, which could generate frequency coverage. to Amplitude range is to The complex pressure waveform, superimposed with random water hammer impact signals, is used in the flow resistance modulation actuator with a nominal diameter of [missing information]. The direct-acting electromagnetic proportional valve has a coil resistance of... The adaptive control unit operates based on a 32-bit microcontroller, and the PWM carrier frequency is set to... The core objective of the experiment is to quantitatively evaluate the system's response characteristics to minute pressure fluctuations and transient shocks under the action of dynamic flutter injection and asymmetric gain scheduling strategies, and to verify the compensation effect of the online correction logic of the hydraulic admittance spectrum on long-term performance degradation; to verify the effectiveness of the dynamic flutter injection mechanism in eliminating the static friction dead zone, a frequency was set at... Amplitude Using a sinusoidal pressure perturbation signal as input, in the control group without flutter injection, the valve opening response exhibits a step-like hysteresis. Only when the incremental driving force generated by the accumulated integral error exceeds the static friction threshold does the valve core undergo a sudden displacement, resulting in an amplitude of approximately [missing value]. The low-frequency oscillation activates the dynamic flutter injection function of this invention, and the flutter frequency is set to... Amplitude based on Dynamic adjustment, test data shows that the valve opening change curve becomes smoother and more continuous, and the response delay to small pressure fluctuations is reduced from that of the control group. Reduce to The fluctuation range of the outflow velocity was suppressed within Within.

[0034] In addition to its defense performance against transient water hammer impacts, the test circuit generates a rise time slope of... The sudden pressure surge signal, in the comparison group using conventional PID control, results in a valve closing delay of approximately [missing information] due to the phase lag of the low-pass filter and integral element. This causes the peak pressure to be transmitted to the end through the valve, and the instantaneous outflow velocity overshoot reaches the target value. In the test group where the asymmetric gain scheduling strategy based on pressure gradient polarity of the present invention was enabled, the system detected... Exceed After reaching the threshold, the bypass filter is switched to a high-gain fast-response mode. Data shows that the valve responds well to pressure surges. The shutdown action is initiated immediately, and the overshoot of the terminal flow rate is limited to [a certain value]. Within a certain range, and after the pressure stabilizes, it quickly and without oscillations returns to a steady state, verifying the dual effectiveness of this strategy in ensuring system safety and maintaining a constant flow rate; to simulate the scale deposition effect caused by long-term operation, a layer with a thickness of [thickness missing] was artificially coated on the valve core surface in the experiment. The epoxy resin coating increases flow resistance. Without enabling online correction logic, the open-loop feedforward accuracy of the system decreases, and the steady-state flow velocity deviates from the target value by approximately [value missing]. Furthermore, the closed-loop adjustment time is extended, activating the online correction logic for the hydraulic admittance spectrum and setting the sampling window length. The aging trend factor was calculated by the system after approximately 2 hours of steady-state operation, with a value of 1000. The system gradually converges and automatically corrects the mapping nodes in the H-Map. After correction, the system undergoes another cold start test, and the deviation between its initial effluent flow rate and the target value is restored to normal. Within this range, it is demonstrated that the mechanism can effectively identify and compensate for the slow drift of the physical characteristics of the actuator. See Table 1 for a summary of the key performance index data of each group in the above tests.

[0035] Table 1: Comparison of Key Performance Indicators

[0036] ;

[0037] The above experimental results show that, by introducing dynamic dithering and asymmetric gain scheduling at the signal level, this invention improves the system's decoupling capability from unsteady pipeline network pressure and its long-term operational stability without changing the low-cost hardware architecture.

[0038] Example 3: This example combines Figures 1 to 3 This document describes an adaptive flow velocity control system for water purification terminals in unsteady-state pipe network environments. Figure 1 As shown, this adaptive flow rate regulation system for water purification terminals in unsteady pipe network environments collects inlet pressure signals through an inlet pressure detection unit. The adaptive control unit then performs a pressure-flow characteristic spectrum query to determine the basic drive duty cycle. The control loop does not directly output this duty cycle but instead introduces a negative amplitude correlation regulation strategy in parallel. Based on the flow resistance state, the high resistance range is set as the first amplitude and the low resistance range is set as the second amplitude. This drives the high-frequency flutter injection logic to generate a zero-mean flutter component with a frequency higher than the cutoff frequency of the actuator. In the drive waveform generation stage, the system superimposes the basic drive duty cycle and the flutter component. The resulting final drive waveform, which contains micro-amplitude oscillation characteristics, is sent to the direct-acting electromagnetic proportional valve, which serves as the flow regulation actuator.

[0039] like Figure 2As shown in the waveform comparison graph reflecting the dynamic performance of the system, the horizontal axis represents time, the left vertical axis represents the amplitude of inlet pressure fluctuation in MPa, and the right vertical axis represents the percentage of flow velocity fluctuation in %. The dashed line in the graph represents the inlet pressure fluctuation signal, simulating the unsteady characteristics of the pipeline network environment. It exhibits periodic large-amplitude changes and instantaneous abrupt changes. The solid line represents the flow velocity fluctuation curve of the control group, which shows a large-amplitude oscillation following characteristic under the influence of pressure disturbance. Its peak deviation is close to 8%, indicating that conventional control cannot effectively decouple pressure disturbance. The dotted line represents the flow velocity fluctuation curve of the present invention, which always maintains a small range near the zero axis under the same pressure disturbance, and its fluctuation amplitude is lower than that of the control group. Figure 3 As shown, the logic flow begins with a real-time pressure signal input from a pressure sensor. The pressure change rate calculation module calculates the value and feeds it back to the adaptive control unit for branch determination. When the system determines that it is in a water hammer impact state (i.e., the pressure change rate is greater than a preset threshold and the direction is positive), the logic path automatically bypasses the low-pass filter. The adaptive control unit then uses a high-gain coefficient with a larger value. The dynamic damping correction is calculated and the flow regulation actuator is driven to quickly close the valve, thereby physically cutting off the transmission of the pressure wave. Conversely, when the system determines that it is in a normal fluctuation state, i.e., the pressure change rate is less than the threshold or the direction is negative, the logic path uses a low-pass filter to process the signal and return a smoothed pressure value. The adaptive control unit then switches to a lower gain coefficient with a smaller value. .

[0040] Example 4: Addressing the issue of nonlinear, slow increases in filter cartridge impedance with the accumulation of trapped contaminants in unsteady-state pipe network environments, this example constructs an online correction mechanism for the hydraulic admittance spectrum based on load impedance characteristic drift. This mechanism is integrated into the adaptive control unit. Its aim is to achieve adaptive calibration of the pressure-flow characteristic spectrum through in-depth analysis of existing pressure and flow rate data without adding additional physical sensors such as valve position feedback sensors. This is particularly relevant when the water purification system is in steady-state operation, i.e., when the inlet water pressure signal... The time change rate is lower than the steady-state determination threshold, and the effluent flow velocity signal When the fluctuation range is within the allowable error band, the adaptive control unit initiates online correction logic. The system operates at a preset sampling period, for example, every [time range missing]. Simultaneously collect a set of data including the current inlet water pressure Total drive duty cycle and actual outflow velocity Based on the three-dimensional operating data and fluid mechanics principles, the system calculates the equivalent fluid resistance of the physical actuator under actual operating conditions. The calculation is based on At the same time, the system adjusts the total drive duty cycle based on the current output. Addressing and interpolation are performed in a preset three-dimensional hydraulic admittance map (H-Map) to obtain the theoretical equivalent hydraulic resistance corresponding to this duty cycle under standard factory conditions. .

[0041] The system calculates the flow resistance residual between the measured equivalent fluid resistance and the theoretical equivalent fluid resistance. ,Right now To eliminate random errors caused by sensor noise and transient fluid disturbances, the system performs moving average filtering or exponential weighted moving average processing on the flow resistance residuals from multiple consecutive sampling periods to extract aging trend factors that characterize the slow-changing drift trend of the actuator's physical properties. When the absolute value of the aging trend factor exceeds a preset update threshold, it indicates that the physical characteristics of the actuator, such as the degree of fouling at the valve port or the spring stiffness, have changed, causing the original control model to no longer match the current physical state. Under this triggering condition, the adaptive control unit utilizes the aging trend factor... The system performs nonlinear weighted correction on the mapping node data in the three-dimensional hydraulic admittance map. Specifically, the system relies on... The size and polarity of the baseline driving duty cycle values ​​for the corresponding pressure and velocity nodes in the graph are scaled or shifted proportionally to generate the updated graph. The mapping relationship, this online self-healing process, enables the control model to dynamically follow and compensate for changes in system impedance caused by filter blockage, valve core wear, or scale deposition, ensuring that accurate initial opening commands can be output based on the current real physical state during the feedforward control stage.

[0042] Example 5: To address the accuracy issues of open-loop control models under production batch variations, long-term aging drift, and sudden environmental changes, a standardized offline calibration and data filling procedure is executed before actual system deployment or shipment to construct a high-confidence initial pressure-flow characteristic map (H-Map). This procedure is performed on a controlled test bench equipped with a high-precision pressure source and flow standard. The flow resistance modulation actuator is placed in a constant-temperature environment to eliminate the interference of temperature on coil resistance and fluid viscosity. The system traverses all possible inlet pressure points according to a preset step size. and basic drive duty cycle For each pressure-duty cycle node, the system records its corresponding steady-state effluent velocity. The equivalent liquid resistance under this working condition is calculated. By performing polynomial fitting or spline interpolation on the data points in the entire working domain, a continuous and smooth three-dimensional surface model is generated and discretized and stored as an initial H-Map lookup table.

[0043] After the system is installed on-site and powered on for the first time, and after each replacement of the core filter element, the adaptive control unit automatically executes the pre-deployment calibration procedure to eliminate systematic errors caused by installation location, pipeline impedance differences, and individual tolerances of the filter element. The system control valves are fully open, and the actual flow resistance is measured under maximum flow conditions and compared with the factory calibration value to calculate the global flow resistance correction coefficient. The system then enters the self-learning phase, dynamically capturing the steady-state operating point during normal water use and fine-tuning the proportional coefficient and integral time constant in the PID control parameters to adapt to the specific water hammer frequency characteristics and pressure fluctuation amplitude of the current pipeline network.

[0044] Example 6: Addressing the issue of inherent differences in physical properties such as coil resistance, spring stiffness, and valve core mass among different batches of direct-acting electromagnetic proportional valves due to machining tolerances, thus affecting open-loop control accuracy, this example constructs a standardized offline generation and verification procedure for adaptive parameter matrices. Through a controlled automated test sequence, it generates a unique initial control parameter set for each individual actuator. After system assembly and initial power-on, the adaptive control unit automatically enters parameter generation mode, performs dead-zone calibration, and controls the drive circuit to output a linearly increasing PWM duty cycle signal, with an initial value of... Step size is Meanwhile, the end-response monitoring unit uses Sampling rate monitoring of water flow velocity When detected The first time the minimum measurable flow threshold is exceeded, such as At that moment, the system records the duty cycle value and defines it as the static opening threshold of the valve. The system performs a full-process scan. to Within the scope, To output a series of stepped drive signals at intervals, for each duty cycle node, the system waits for the flow velocity to stabilize and then records the corresponding steady-state flow velocity value, thereby constructing the discretized flow response curve of the valve under the current inlet pressure. The system uses a piecewise linear interpolation algorithm to generate the baseline data of the initial pressure-flow characteristic map (H-Map) specific to this individual.

[0045] After the baseline layer is constructed, the system performs a dynamic response verification step, and the adaptive control unit outputs an amplitude value corresponding to the duty cycle of the baseline layer. , frequency is The square wave disturbance signal is monitored, and the rise time and overshoot of the flow velocity response are monitored in real time. If the rise time exceeds the preset standard, the signal is processed. The system automatically increases the proportional coefficient in the PID control algorithm. If the overshoot exceeds Then decrease And increase the differential coefficient The parameter optimization process follows the debugging logic of proportional, then integral, and then derivative, until the dynamic response index of the system meets the preset engineering acceptance criteria. The final determined PID parameter set and dedicated H-Map are then stored in non-volatile memory as the baseline parameters for the subsequent operation of the device.

[0046] 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.

[0047] 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 purification terminal flow velocity adaptive adjustment system for unsteady-state pipe network environments, characterized in that, include: The inlet water pressure detection unit is located at the fluid input end of the water purification system and is used to collect inlet water pressure signals. The flow regulation actuator is located in the flow path between the fluid input end and the fluid output end, and has a direct-acting electromagnetic proportional valve driven by a pulse width modulation signal. The adaptive control unit is connected to the inlet pressure detection unit and the flow regulation execution unit respectively. The adaptive control unit is used to execute flow control logic that includes a negative correlation adjustment strategy: it determines the basic drive duty cycle by querying a preset pressure-flow characteristic spectrum based on the inlet pressure signal. This basic drive duty cycle is used to set the target opening of the electromagnetic proportional valve. The high-frequency chatter injection logic is executed to generate a zero-mean chatter component with a frequency higher than the mechanical cutoff frequency of the electromagnetic proportional valve. The dynamic amplitude of the chatter component is calculated based on the current basic drive duty cycle. The calculation of the dynamic amplitude follows a negative amplitude correlation adjustment strategy: when the basic drive duty cycle is in the high flow resistance range, the dynamic amplitude is set to the first amplitude; when the basic drive duty cycle is in the low flow resistance range, the dynamic amplitude is set to the second amplitude, and the first amplitude is greater than the second amplitude. The basic drive duty cycle and the chatter component with the dynamic amplitude are superimposed to generate the final drive waveform and output to the flow regulation execution unit. This drives the electromagnetic proportional valve to maintain a small reciprocating motion while maintaining the target opening, so as to maintain the contact friction state of the electromagnetic proportional valve core as dynamic friction. When the adaptive control unit executes the amplitude negative correlation adjustment strategy, it defines the high flow resistance range as the range where the basic drive duty cycle is less than 30%, and the low flow resistance range as the range where the basic drive duty cycle is greater than 70%. The flutter signal intensity corresponding to the first amplitude is set to the duty cycle amplitude corresponding to the minimum current increment that can overcome the maximum static friction of the electromagnetic proportional valve. The flutter signal intensity corresponding to the second amplitude is set to 50% to 80% of the first amplitude.

2. The adaptive flow velocity adjustment system for water purification terminals in non-steady-state pipe network environments according to claim 1, characterized in that, When generating the zero-mean flutter component, the adaptive control unit sets the frequency of the flutter component to a fixed frequency of 200 Hz to 300 Hz. This frequency is set to maintain a non-multiplication synchronization relationship with the carrier frequency of the pulse width modulation signal, and this frequency is higher than the mechanical resonance frequency of the water purification system, so as to suppress fluid pulsation noise while eliminating the static friction dead zone of the electromagnetic proportional valve.

3. The adaptive flow velocity adjustment system for water purification terminals in non-steady-state pipe network environments according to claim 1, characterized in that, The adaptive control unit is also used to execute asymmetric gain scheduling logic based on the direction of pressure change rate: calculating the time change rate of the inlet pressure signal; when the time change rate is positive and its absolute value is greater than a preset abrupt change threshold, the system is determined to be in a water hammer impact state, and the system bypasses the target... The low-pass filter stage directly uses the instantaneous pressure value and calls the first gain coefficient, which has a relatively large value. Calculate the dynamic damping correction and execute the instruction-level reset procedure of the FIFO buffer queue: reset the read / write pointers of the moving average filter to zero, using the current instantaneous pressure value. The system covers all historical data in the buffer zone, bypasses the filtering stage for the inlet pressure signal, and calculates the dynamic damping correction using the first gain coefficient. The dynamic damping correction is then superimposed on the basic drive duty cycle to form a fast-response control loop. When the rate of change of time is negative or its absolute value is less than or equal to the abrupt change threshold, the system is determined to be in a normal fluctuation state. The filtering stage for the inlet pressure signal is then activated, and the dynamic damping correction is calculated using the second gain coefficient to form a smooth-response control loop. The value of the first gain coefficient is greater than that of the second gain coefficient.

4. The adaptive flow velocity adjustment system for water purification terminals in non-steady-state pipe network environments according to claim 1, characterized in that, The adaptive control unit is also used to execute the online correction logic of the pressure-flow characteristic spectrum: under steady-state operating conditions, it acquires a three-dimensional data set including the inlet pressure signal, the final drive waveform and the actual outlet flow rate signal; Calculate the flow resistance residual between the measured flow resistance based on the three-dimensional data set and the theoretical flow resistance obtained from the pressure-flow characteristic spectrum query; extract the physical characteristic drift of the flow regulation actuator according to the following aging trend factor calculation formula: ,in, As an aging trend factor, The sampling window length, For the first The measured flow resistance during the second sampling. For the first The theoretical flow resistance at the time of the second sampling; When the absolute value of the aging trend factor exceeds the preset update threshold, the current operating point of the envelope is locked. Four three-dimensional discrete nodes Calculate spatial mapping weights based on Euclidean distance Perform nonlinear iteration of node values Set the learning rate And force a single iteration to update the amplitude limit. When the calculated When the absolute value of the factor exceeds the preset update threshold, it indicates that the physical characteristics of the actuator have shifted. The system uses this factor to perform weighted correction on the mapping node data in the H-Map.

5. The adaptive flow velocity adjustment system for water purification terminals in non-steady-state pipe network environments according to claim 1, characterized in that, The pressure-flow characteristic spectrum preset in the adaptive control unit is constructed as a three-dimensional discrete data table. The three axes of this data table are the inlet pressure axis, the target flow velocity axis, and the basic drive duty cycle axis. When querying, the adaptive control unit calculates the corresponding basic drive duty cycle in the three-dimensional discrete data table based on the real-time collected inlet pressure signal and the preset target flow velocity value using a bilinear interpolation algorithm.

6. The adaptive flow velocity adjustment system for water purification terminals in non-steady-state pipe network environments according to claim 1, characterized in that, The system also includes an outlet flow velocity detection unit located at the fluid output end, used to collect the actual outlet flow velocity signal; the adaptive control unit is also used to calculate the steady-state deviation between the actual outlet flow velocity signal and the target flow velocity, calculate the steady-state correction using a proportional-integral control algorithm, and add the steady-state correction to the basic drive duty cycle; When calculating the steady-state correction, the adaptive control unit only initiates the integral term calculation of the proportional-integral control algorithm when the rate of change of the inlet pressure signal is less than the steady-state determination threshold.

7. The adaptive flow velocity adjustment system for water purification terminals in non-steady-state pipe network environments according to claim 3, characterized in that, When the adaptive control unit determines that the system is in a water hammer impact state, it is also used to execute the forced cutoff logic: if the sum of the calculated basic drive duty cycle and the dynamic damping correction is less than the minimum maintenance duty cycle, the duty cycle signal output to the flow regulation execution unit is forcibly set to zero until the time change rate of the inlet pressure signal recovers to within the sudden change judgment threshold.

8. The adaptive flow velocity adjustment system for water purification terminals in non-steady-state pipe network environments according to claim 1, characterized in that, The inlet pressure detection unit includes a ceramic piezoresistive pressure sensor, which is directly installed in the upstream flow path of the flow regulation actuator; the electromagnetic proportional valve of the flow regulation actuator is a normally closed structure, and its valve core cuts off the flow path by the force of the reset spring when there is no drive signal. During the system power-on initialization phase, the adaptive control unit outputs a fully open pulse signal with a preset duration.

9. The adaptive flow velocity adjustment system for water purification terminals in non-steady-state pipe network environments according to claim 1, characterized in that, The adaptive control unit integrates a microprocessor and a power drive circuit; the microprocessor is used to perform query, calculation and signal generation logic; the power drive circuit is used to receive the final drive waveform and convert it into a drive current output to the coil of the electromagnetic proportional valve; the system also includes a regulated power supply module for supplying power to the power drive circuit, which has an overcurrent protection circuit.

Images