Water purification system capable of adjusting mineral concentration

Abstract

The invention relates to the technical field of water purification, and discloses a water purification system capable of adjusting mineral concentration, which comprises a pure water conveying subsystem, a mineralization injection subsystem, a signal sensing subsystem and a control subsystem. Wherein the control subsystem is configured to construct a mineralized liquid concentration attenuation model according to the shutdown duration and the environment temperature when water production is started; the injection execution module accumulates mineral mass requirements through an accumulation register, and drives the mineralization injection subsystem to act when the calculated opening duration exceeds a valve dead zone; and the feedback calibration module aligns the flow and water quality data to calculate an execution error after the circulation is finished, and updates the health factor of the filter element according to the calculation execution error. Through closed-loop control of model prediction and feedback calibration, the problem that the injection precision is reduced due to concentration fluctuation of a solid mineralizing medium and aging of a filter element is solved, and accurate mineralizing adjustment in the whole life cycle is achieved.

Description

Technical Field

[0001] This invention relates to the field of water purification technology, specifically to a water purification system with adjustable mineral concentration. Background Technology

[0002] Reverse osmosis water purification technology is widely used due to its excellent filtration precision; however, while removing harmful impurities, it also filters out most of the beneficial minerals in the water. To improve the taste of drinking water and supplement trace elements, adding a mineralization treatment unit after the reverse osmosis membrane has become the mainstream solution.

[0003] Existing mineralization technologies mainly rely on the passive dissolution of solid mineralizing media as water flows through. The dissolution concentration in this method is highly dependent on the contact time between water and the media. This results in a significant increase in the concentration of liquid minerals accumulated inside the filter cartridge after the system has been shut down for a long time. During continuous water production, the concentration of the effluent decreases rapidly due to the limited solid-liquid diffusion rate. This non-linear dissolution characteristic leads to huge fluctuations in the mineral content of the final effluent, making it difficult to maintain a constant level.

[0004] Although some technical solutions attempt to introduce electrically controlled valves for active regulation, they still face limitations in both hardware physical characteristics and control logic in practical applications. On the one hand, due to cost and size constraints, the electromagnetic regulating valves commonly used in household appliances have inherent response dead zones, making it difficult to accurately respond to the micro-injection requirements under low flow rate conditions, resulting in insufficient regulation precision. On the other hand, during long-term use, the leaching efficiency of solid mineralizing media undergoes irreversible changes as its volume decreases or its surface becomes passivated. Existing control systems mostly use factory-preset fixed parameters for open-loop control, which cannot detect and compensate for this characteristic drift caused by media aging, leading to a significant deviation between the actual mineralization effect and the set target in the later stages of equipment operation.

[0005] Therefore, this invention proposes a water purification system with adjustable mineral concentration to address the shortcomings of existing technologies. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a water purification system with adjustable mineral concentration, which solves the problems of inaccurate and unstable mineral concentration addition caused by nonlinear changes in the dissolution characteristics of the medium, hardware response dead zones, and filter aging in existing mineralized water purification systems.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a water purification system with adjustable mineral concentration, characterized in that it comprises: a pure water delivery subsystem configured to output filtered base water flow; a mineralization injection subsystem connected to the pure water delivery subsystem, configured to store mineral concentrate and inject the mineral concentrate into the base water flow via valve action; a signal sensing subsystem configured to detect fluid flow rate, temperature, and conductivity parameters; and a control subsystem connected to the pure water delivery subsystem, the mineralization injection subsystem, and the signal sensing subsystem, wherein the control subsystem includes a model prediction module, an injection execution module, and a feedback calibration module; the model prediction module is configured to activate the water purification cycle upon startup. The system reads the downtime of the water purification system and the ambient temperature, and combines this with the filter cartridge health factor to construct a decay function model describing the change in the concentration of the concentrate inside the mineralization injection subsystem as the outflow volume changes. The injection execution module is configured to calculate the mineral mass requirement based on the flow signal and the set target concentration and store it in an accumulation register. Based on the decay function model, it calculates the valve opening time required to consume the value in the accumulation register and drives the mineralization injection subsystem to operate. The feedback calibration module is configured to align the flow data and the mixed water quality data after the water purification cycle is completed, calculate the execution error between the actual total injected mineral mass and the theoretical target total mass, and update the filter cartridge health factor based on the execution error for the model prediction module to call next time.

[0008] Preferably, the mineralization injection subsystem adopts a parallel topology architecture, including: a multi-channel fluid switching unit, the inlet of which is connected to a pure water delivery subsystem and configured to open a designated flow channel according to the instructions of the control subsystem; a first mineralization branch, a second mineralization branch, a third mineralization branch, and a fourth mineralization branch, which are connected in parallel to the outlet of the multi-channel fluid switching unit, and the first, second, third, and fourth mineralization branches are filled with differentiated mineralization media configured for different population patterns; and a high-frequency electromagnetic regulating valve, which is located at the common outlet after the first, second, third, and fourth mineralization branches converge, and configured to pulse-type cut-off and release of the mineral concentrate passing through.

[0009] Preferably, the mineralization injection subsystem includes a mineral dissolution chamber, which is configured as a closed container filled with a solid mineralization medium for generating a mineral concentrate through diffusion at the solid-liquid interface under static conditions. The model prediction module is pre-set with dissolution characteristic parameters and is configured to determine the initial peak concentration and steady-state dissolution concentration based on the downtime of the water purification system and the ambient temperature by querying a pre-set data mapping table or executing a dissolution kinetics calculation model. The model prediction module then constructs a decay function model using the initial peak concentration and steady-state dissolution concentration as boundary conditions to characterize the physical process of nonlinearly decreasing concentration as the cumulative outflow volume increases.

[0010] Preferably, the injection execution module establishes a step-triggered mechanism based on flow pulses to discretize the continuous water flow into single-step injection volumes of equal volume. The injection execution module is configured to: calculate the mineral mass increment required for the single-step injection volume and accumulate it to the accumulation register when each single-step injection volume is triggered; output a drive signal only when the calculated valve opening duration is greater than the hardware dead zone threshold, and deduct the corresponding mass value from the accumulation register after the action is completed; otherwise, maintain the value of the accumulation register until the next calculation cycle.

[0011] Preferably, the injection execution module is further configured with pressure compensation logic; the injection execution module calculates the real-time flow velocity of the main channel by monitoring the time interval of adjacent flow pulses, and corrects the theoretical volumetric flow velocity parameters of the solenoid valve according to a preset pressure coupling model; the injection execution module uses the corrected theoretical volumetric flow velocity parameters of the solenoid valve to calculate the theoretical valve opening time required to consume the current value of the accumulation register, so as to compensate for the inhibitory effect of the change in the main channel flow velocity on the injection capacity of the mineralization injection subsystem.

[0012] Preferably, the feedback calibration module is configured with a delay alignment algorithm based on a first-in-first-out queue; the feedback calibration module determines the transmission delay parameter based on the physical pipeline volume between the junction point of the pure water delivery subsystem and the mineralization injection subsystem and the detection point of the signal sensing subsystem; the feedback calibration module uses the first-in-first-out queue to perform phase shifting on the pure water baseline concentration data, so that the pure water baseline concentration data and the final effluent concentration data after mixing are strictly corresponding in the time domain, and calculates the total mass of the actual injected minerals accordingly.

[0013] Preferably, the feedback calibration module is configured with an exponentially weighted moving average algorithm to update the filter element health factor; the feedback calibration module calculates the ratio of the actual total mass of injected minerals to the theoretical target total mass, and uses a preset iterative convergence coefficient to weight and correct the current filter element health factor, so that the updated filter element health factor represents the true dissolution efficiency of the current mineralized medium relative to the standard state; the control subsystem is configured to prohibit the update operation and trigger an alarm when the relative error exceeds a preset safety range.

[0014] Preferably, the control subsystem includes a non-volatile memory, which is configured with independent data storage areas corresponding to the first mineralization branch, the second mineralization branch, the third mineralization branch, and the fourth mineralization branch, respectively. The feedback calibration module is configured to update only the filter health factors corresponding to the first, second, third, or fourth mineralization branch that is activated in the current water production cycle, and write the updated data into the corresponding independent data storage area, thereby realizing independent tracking and storage of the aging status of multiple parallel filter elements.

[0015] Preferably, the system further includes a static mixing unit; the static mixing unit is located downstream of the physical confluence point of the pure water delivery subsystem and the mineralization injection subsystem, and is configured to generate turbulence using its internal baffle structure to uniformly disperse the pulsed mineral concentrate injected by the mineralization injection subsystem into the base water flow output by the pure water delivery subsystem.

[0016] Preferably, the system further includes a human-computer interaction unit configured to receive a user's crowd mode selection command; the control subsystem is configured to respond to the child mode command by locking the third mineralization branch and activating the safety current limiting logic, the safety current limiting logic forcibly limiting the maximum single injection duration calculated by the injection execution module to prevent the effluent concentration from exceeding the physiological safety threshold for children.

[0017] This invention provides a water purification system with adjustable mineral concentration. It has the following beneficial effects:

[0018] 1. This invention constructs a decay function model based on downtime and ambient temperature through a control subsystem, and, in conjunction with a feedback calibration module, updates the filter cartridge health factor in real time. This solves the control problem caused by the nonlinear concentration decay of solid mineralized media after settling. The system can predict the initial concentration peak and subsequent decline trend of each water production cycle, and corrects the model parameters in reverse through error analysis after each cycle. This eliminates the impact of filter cartridge aging on dissolution efficiency and ensures the stability and accuracy of mineral addition concentration throughout the entire life cycle of the filter cartridge.

[0019] 2. This invention introduces an accumulation register and a hardware dead-time threshold determination mechanism into the injection execution module, breaking through the physical response bottleneck of conventional solenoid valves under low flow rates. By accumulating single-step micro-demands, the action is triggered only when the valve opening time corresponding to the demand exceeds the hardware dead time. This control strategy avoids ineffective action or half-open state of the solenoid valve under extremely short pulses, achieving high-precision quantitative injection under low flow rate conditions without increasing hardware costs.

[0020] 3. This invention utilizes a delay alignment algorithm based on a first-in-first-out (FIFO) queue to eliminate data time-domain deviations caused by physical pipelines between the pure water delivery subsystem and the signal sensing subsystem. By precisely shifting the flow data according to the pipeline volume, the system can strictly correspond the flow fluctuations at the front end with the concentration changes at the back end, eliminating the interference of fluid transmission delays on the control loop and improving the accuracy of feedback calibration calculations and the robustness of closed-loop control. Attached Figure Description

[0021] Figure 1 This is a structural block diagram of the water purification system with adjustable mineral concentration according to the present invention;

[0022] Figure 2 This is a schematic flowchart of the mineral concentration adjustment method of the present invention;

[0023] Figure 3 This is a schematic diagram of the injection execution module logic of the present invention;

[0024] Figure 4 This is a schematic diagram of the feedback calibration module of the present invention;

[0025] Figure 5 This is a comparative diagram of concentration control during a single water sampling process according to the present invention.

[0026] Figure 6 This is a schematic diagram of the long-term error convergence curve of the present invention.

[0027] Among them, 110 is the model prediction module; 120 is the injection execution module; and 130 is the feedback calibration module. Detailed Implementation

[0028] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] See attached document Figure 1 The present invention provides a water purification system with adjustable mineral concentration, including a pure water delivery subsystem, a mineralization injection subsystem, a signal sensing subsystem, and a control subsystem.

[0030] The pure water delivery subsystem provides the base water flow after reverse osmosis filtration. The mineralization injection subsystem, connected to the pure water delivery subsystem, stores the mineral concentrate and injects it into the base water flow via a solenoid valve. The signal sensing subsystem detects fluid flow rate, temperature, and conductivity parameters. The control subsystem connects to the pure water delivery subsystem, mineralization injection subsystem, and signal sensing subsystem, performing data processing and device driving.

[0031] The control subsystem includes a model prediction module 110, an injection execution module 120, and a feedback calibration module 130.

[0032] The model prediction module 110 is used to read the system shutdown time and ambient temperature data when the water production cycle starts, query the preset dissolution characteristic data table, and combine the filter health factor to construct a decay function model describing the change of the concentration of the concentrate inside the mineralization injection subsystem with the outflow volume.

[0033] The injection execution module 120 receives the flow pulse signal from the signal sensing subsystem, calculates the mineral mass requirement corresponding to the current flow volume based on the set target concentration, and stores the mass requirement in the accumulation register. The injection execution module 120 calculates the valve opening time required to consume the value in the accumulation register according to the decay function model, and drives the mineralization injection subsystem to perform the injection action when the opening time meets the hardware dead zone threshold.

[0034] The feedback calibration module 130 is used to acquire flow rate data and mixed water quality data for the entire process after the water production cycle is completed. The feedback calibration module 130 calculates the actual total mass of injected minerals using a delay alignment algorithm and compares it with the theoretical target total mass to determine the execution error. Based on the execution error, the feedback calibration module 130 updates the filter cartridge health factor and writes the updated factor into non-volatile memory for future use by the model prediction module 110.

[0035] See attached document Figure 2 This invention provides a water purification method with adjustable mineral concentration, comprising the following steps:

[0036] In response to the user's crowd mode selection command, the target mineralization branch and target effluent concentration are determined. Before or at the moment of start-up of the water purification cycle, the control subsystem first reads the setting status of the human-machine interface unit and performs initialization configuration according to the preset "crowd mode-control strategy mapping table":

[0037] If a "male mode" command is received, the control subsystem locks the first mineralization branch as the current working flow channel, drives the multi-channel fluid switching unit to only open the valve port corresponding to that branch, and retrieves the first preset concentration value (e.g., 3.0 mg / L to 5.0 mg / L) from the memory as the target effluent concentration for this cycle;

[0038] If a "female mode" command is received, the control subsystem locks the second mineralization branch and sets the target effluent concentration to the second preset concentration value (e.g., 2.0 mg / L to 4.0 mg / L).

[0039] If a "Child Mode" command is received, the control subsystem locks the third mineralization branch and sets the target effluent concentration to a lower third preset concentration value (e.g., 0.5 mg / L to 1.0 mg / L). Specifically, in Child Mode, the system automatically activates safety flow limiting logic, forcibly limiting the maximum single injection duration calculated in subsequent steps to prevent excessive injection of high-concentration mineralization solution due to hardware failure.

[0040] If the "elderly mode" command is received, the control subsystem locks the fourth mineralization branch and sets the target effluent concentration to a suitable fourth preset concentration value.

[0041] After completing the above-mentioned flow channel locking and target value setting, the control subsystem retrieves the static leaching characteristic table, scouring attenuation coefficient and filter health factor that uniquely correspond to the currently locked mineralized branch from the non-volatile memory as model input parameters for subsequent steps S100 and S200.

[0042] S100, in response to the water intake trigger signal, obtains the system shutdown duration and current fluid temperature, queries the preset static dissolution characteristic table to determine the initial peak concentration and steady-state concentration, and establishes an instantaneous concentration decay model of the mineral concentrate in combination with the current filter element health factor.

[0043] S200 monitors the main flow pulse signal. When the flow pulse count reaches the preset step threshold, it calculates the single-step mineral quality requirement based on the difference between the target effluent concentration and the pure water baseline concentration, and adds the quality requirement to the quality accumulation register.

[0044] S300 calculates the theoretical valve opening time required to consume the current value of the mass accumulation register based on the instantaneous concentration decay model and the valve flow rate corrected by pressure compensation. It then determines whether the theoretical valve opening time is greater than the minimum physical response time of the solenoid valve. If it is greater, it drives the solenoid valve to open for the corresponding time and deducts the corresponding value of the mass accumulation register. If it is less, it keeps the solenoid valve closed and retains the value of the mass accumulation register until the next calculation cycle.

[0045] After the water intake process is completed, the S400 uses the first-in-first-out queue to align the concentration data of the main flow channel and the mixing flow channel, integrates to calculate the total mass of the actual injected minerals, and iteratively updates the filter health factor based on the relative error between the actual total mass of injected minerals and the theoretical target total mass.

[0046] To further clarify the implementation of each technical aspect of the present invention, the following will provide a detailed description of the implementation of each functional module involved above and its internal processing flow.

[0047] See attached document Figure 1 The adjustable mineral concentration water purification system provided in this embodiment is based on a dual-channel parallel topology design and adopts a control strategy that combines passive physical mixing with active pulse injection to achieve precise dynamic control of the mineral concentration of the terminal effluent.

[0048] In this embodiment, the water purification system with adjustable mineral concentration mainly consists of a pure water delivery subsystem, a mineralization injection subsystem, a signal sensing subsystem, and a control subsystem. The pure water delivery subsystem forms the main liquid pathway of the system, with an inlet booster pump, a reverse osmosis membrane filter assembly, and a pure water delivery pipeline arranged sequentially along the fluid flow direction. The reverse osmosis membrane filter assembly is configured to remove dissolved solids and impurities from the raw water, outputting basic pure water with a conductivity within a preset low threshold range. In this embodiment, this preset low threshold range is typically set to 0 to 50 μS / cm, with the specific value depending on the desalination characteristics of the reverse osmosis membrane. The pure water delivery pipeline is made of food-grade polymer materials such as polypropylene (PP) or polyethylene (PE), and its pipe diameter design must meet the hydrodynamic requirements of the system's rated flux. To ensure unidirectional fluid flow and maintain pipeline pressure, a one-way check valve is installed on the pure water delivery pipeline. The valve's opening pressure is set to 0.01 MPa to 0.03 MPa to prevent downstream fluid from flowing back to the reverse osmosis membrane filter assembly during shutdown and pressure loss.

[0049] The mineralization injection subsystem is connected to the main system as a controlled secondary channel. Its inlet end connects to the pure water output end of the pure water delivery subsystem, while its outlet end merges into the main channel to form a bypass parallel structure. The mineralization injection subsystem includes a mineral dissolution chamber and a high-frequency electromagnetic regulating valve. The mineral dissolution chamber is a sealed, static cavity container filled with sintered strontium-rich spheres, zinc spheres, or composite trace element particles, etc., as solid mineralizing media. In a static state, the water remaining in the chamber comes into contact with the solid mineralizing media, and solid-liquid interface diffusion occurs. This process follows Fick's diffusion law, causing the mineral concentration in the water within the chamber to increase non-linearly over time. It should be noted that the mineral dissolution chamber in this embodiment is a passive device, meaning it does not contain a motor, agitator, or any form of active mechanical stirring device; it relies entirely on natural diffusion and fluid scouring to achieve concentration changes. The high-frequency electromagnetic regulating valve is located at the downstream outlet of the mineral leaching chamber. This valve is a direct-acting or pilot-operated solenoid valve with a physical response time (i.e., the time from energization to full opening) of less than 20ms. It supports pulse modulation action with a frequency of 1Hz to 10Hz, thereby enabling it to respond to control commands and accurately release high-concentration mineralized liquid in the form of pulse flow.

[0050] To meet the diverse mineral requirements of different user groups, the mineralization injection subsystem in this embodiment adopts a multi-channel parallel topology. Specifically, the subsystem includes a multi-channel fluid switching unit and four parallel mineralization branches: a first mineralization branch, a second mineralization branch, a third mineralization branch, and a fourth mineralization branch. The inlet of the multi-channel fluid switching unit (e.g., a multi-position multi-way solenoid valve group or an integrated manifold valve island) is connected to the output of the pure water delivery subsystem, and its multiple outlets independently control the on / off state of each mineralization branch.

[0051] Each mineralization pathway is filled with a differentiated mineralization medium configured to meet specific physiological needs:

[0052] The first mineralization branch corresponds to the "male mode," and its interior is filled with a zinc-rich mineralization medium, configured to provide a high concentration of zinc element solution to meet the needs of male physiological function and immune enhancement.

[0053] The second mineralization branch corresponds to the "female mode," which is filled with a complex medium rich in calcium and iron ions, designed to replenish the trace element loss that women experience during daily life and special physiological cycles.

[0054] The third mineralization branch corresponds to the "children's mode". It is filled with a low-solubility balanced medium and is configured to output low-concentration, low-load balanced mineral water to adapt to the immature kidney metabolic function of young children.

[0055] The fourth mineralization branch corresponds to the "elderly mode," and its interior is filled with a high-calcium, easily absorbed medium, configured to provide a moderate concentration of calcium supplementation to maintain bone health.

[0056] In addition, this system includes a human-machine interface unit, which establishes a communication connection with the control subsystem. The human-machine interface unit can be a touch panel or physical button module located on the water purifier body, or a mobile application (APP) based on Wi-Fi / Bluetooth communication. This unit receives the user's human mode selection command and transmits the command to the control subsystem in real time.

[0057] The signal sensing subsystem is distributed at various key nodes in the flow path to collect the physical quantities required for system control. Specifically, the flow metering unit is located at the system's main inlet or on the pure water pipeline after the reverse osmosis membrane, preferably using a Hall pulse flow meter. Its internal impeller rotates under the pressure of the fluid, cutting magnetic field lines and generating a square wave signal with a frequency proportional to the flow velocity. The flow metering unit has a fixed flow coefficient K, which is the number of pulses generated per liter of fluid flowing through it, used to output the cumulative pulse signal in real time. The temperature sensing unit uses an NTC thermistor probe that directly contacts the fluid to collect real-time water temperature T. This temperature data is used in subsequent algorithms to query the solubility curve and compensate for conductivity temperature drift. The water quality monitoring unit contains two sets of conductivity probes. The first probe is located between the reverse osmosis membrane outlet and the mineralization branch intake point to detect the baseline concentration of pure water. The second probe is located at the end of the total effluent outlet after the main and secondary flow channels merge, and is used to detect the final effluent concentration after mixing. The above concentration values ​​were obtained by measuring the resistivity of the solution and converting it according to the linear conversion factor, and the units are ppm or mg / L.

[0058] At the physical confluence of the main and secondary flow channels, this embodiment includes a static mixing unit. This static mixing unit is a pipe section with spiral blades or baffles on its inner wall, or a component employing a labyrinthine flow channel structure. This structure utilizes the kinetic energy of the fluid itself. When the continuous pure water flow in the main flow channel and the pulsed high-concentration mineralized liquid flow in the secondary flow channel pass through the internal baffles, a Karman vortex street or micro-turbulence is generated, forcing the two fluids to rapidly exchange substances in the radial direction. The function of this physical structure is to uniformly disperse the temporally discretely injected pulsed mineralized liquid into the pure water base flow in space, eliminating instantaneous fluctuations in the concentration gradient and ensuring the uniformity of the final effluent.

[0059] The control subsystem is the core of the system's computation and execution, and it is electrically connected to the inlet booster pump, high-frequency electromagnetic regulating valve, and signal sensing subsystem via wiring harnesses. The control subsystem incorporates a microprocessor (MCU) and non-volatile memory (such as Flash or EEPROM). In this embodiment, the microprocessor is an ARM Cortex-M series chip or a chip with equivalent computing power, and it is equipped with a floating-point unit (FPU) to handle exponential decay model calculations. The microprocessor is configured to execute computational logic based on the fluid dynamics model, driving the high-frequency electromagnetic regulating valve to operate according to real-time data collected by sensors; the non-volatile memory is used to store preset leaching characteristic data tables and dynamically updated filter health factors. Through the connection and cooperation of the above physical modules, the system constructs a closed-loop fluid control network with adaptive adjustment capabilities.

[0060] See attached document Figure 2 In this embodiment, when the system initiates the water production cycle in response to the water intake trigger signal, the model prediction module 110 is configured to execute initialization logic, aiming to quantify the current physical state of the mineralization tank and construct a concentration benchmark for subsequent pulse control. This initialization process specifically includes:

[0061] S110, Get system downtime duration With fluid ambient temperature In this embodiment, the control subsystem uses a built-in real-time clock (RTC) unit to record time data. Each time the system finishes water production and the solenoid valve closes, the microprocessor writes the current timestamp to non-volatile memory. When the current water intake trigger signal is generated, the microprocessor reads the current timestamp and the stored previous end timestamp, calculating the difference to obtain the system downtime. To prevent data overflow or anomalies, the system has a maximum effective statistical duration. (In this embodiment, it is set to 48 hours), when the calculated Exceed At that time, Forced assignment This logic is based on the experimental observation of mineral diffusion saturation, that is, in a closed system, when the settling time exceeds a certain threshold, the concentration gradient in the chamber tends to reach equilibrium and no longer increases significantly with time.

[0062] Simultaneously, the model prediction module 110 reads the analog signal from the temperature sensing unit via an analog-to-digital converter (ADC). The microprocessor uses the temperature resistance characteristic curve of the thermistor to convert the electrical signal into a Celsius value, thus obtaining the fluid environment temperature. To eliminate the white noise interference from the sensor itself, the microprocessor performs a moving average filter on the sampled data within the last 50ms to ensure the acquired data is accurate and consistent. The values ​​are smooth and stable, and the measurement accuracy is controlled within ±0.5℃.

[0063] S120, determine the boundary concentration parameters based on a preset data mapping table or a dissolution kinetics calculation model. In this embodiment, because the liquid concentration state within the mineralization injection subsystem is affected by multiple coupling effects of time, thermodynamic environment, and spatial geometric parameters, it is difficult to measure directly in real time using conventional sensors. Therefore, the control subsystem is configured to obtain the current concentration boundary through indirect numerical mapping.

[0064] The system's non-volatile memory pre-stores dissolution characteristic parameters that uniquely correspond to the currently locked mineralization branch. These parameters are constructed based on the principles of solid-liquid interface dissolution kinetics (e.g., the Noyes-Whitney equation modified model): that is, under static and closed conditions, the solute on the surface of the mineralization medium diffuses into the surrounding water, and its concentration increases exponentially over time and tends to saturate; while under flowing conditions, the dissolution rate and the rate at which the fluid carries away the solute will reach a dynamic equilibrium.

[0065] Based on the real-time parameters obtained in step S110, the model prediction module 110 determines two key state boundary values:

[0066] The first boundary value is the static peak concentration. It characterizes the current temperature and downtime The combined effect results in the highest theoretical concentration of minerals leaching from the water in the leaching chamber.

[0067] As a basic implementation method, the model prediction module 110 can directly query a preset data mapping table. This table consists of multiple sets of discrete "temperature-time-concentration" mapping data. When the measured values ​​are between sampling points, the model prediction module 110 performs bilinear interpolation: specifically, it first performs linear interpolation between adjacent temperature points, and then performs secondary interpolation between adjacent time points based on the result.

[0068] As a preferred embodiment of this invention, in order to accommodate the geometric differences of filter cartridges of different specifications and reduce the storage space occupied, the model prediction module 110 is configured to calculate the static peak concentration using an analytical formula based on dissolution kinetics. This calculation logic specifically introduces a geometric loading factor for the mineralization chamber to correct the influence of the effective water volume on the concentration rise rate, and the calculation follows the following functional relationship:

[0069] ;

[0070] In the formula, This represents the theoretical saturation concentration at the current temperature, expressed in mg / L. Within the operating temperature range of this embodiment (5°C to 40°C), this parameter is set as a linear function of temperature. .in The solubility temperature coefficient is preferably in the range of 0.1 to 0.3. The basic solubility intercept is preferably taken in the range of 5.0 to 8.0. This represents the system downtime, expressed in hours (h). The reference time constant is expressed in hours (h). This parameter characterizes the rate at which solute diffusion reaches equilibrium. Based on the Arrhenius dissolution kinetics principle, within the water purification operating temperature range of this embodiment, this parameter is inversely proportional to temperature, and the preferred calculation formula is... .in The solution thermodynamic constant is preferably taken in the range of 500 to 800. For the temperature correction compensation term, a value range of 20 to 40 is preferred. This formula indicates that the lower the ambient temperature T, the larger the time constant. The larger the value, the longer it takes to reach the saturation concentration. The geometric loading correction factor is a dimensionless parameter. This factor is used to quantify the effect of the solid-liquid ratio on the concentration gradient within the mineralization chamber, and its calculation formula is as follows: ;in The filling rate of the mineralizing medium is the ratio of the medium volume to the total volume of the silo. In this embodiment... The typical design range is 0.2 to 0.6. This physical relationship indicates that the fill rate... The higher the concentration, the less effective water is retained in the chamber and the larger the solid-liquid contact area, which leads to a faster rate of increase in static concentration.

[0071] Using the above formula, the system can determine the specific filter cartridge filling parameters. It adaptively calculates concentration curves without requiring full-condition recalibration for each filter cartridge specification.

[0072] To facilitate a more intuitive understanding of the above logic, Table 1 presents examples of discrete data generated based on the algorithm of this embodiment. The data trends indicate that the static peak concentration... Depending on downtime The increase of is non-linear and increases with increasing temperature T.

[0073] Table 1: Static Dissolution Characteristics Data (Example)

[0074]

[0075] It should be noted that the values ​​in Table 1 are only examples illustrating the data structure and trends. The values ​​used in actual applications need to be calculated based on the aforementioned standard calibration experiments and specific formula parameters.

[0076] Furthermore, the second boundary value is the steady-state dissolution concentration. This parameter represents the dynamic equilibrium concentration maintained at the outlet when fluid continuously flows through the mineral leaching chamber at its rated flow rate at the current temperature T. Under the premise of constant flow rate, this parameter can be simplified to a single-variable function of temperature. The query logic follows the functional relationship below:

[0077] ;

[0078] In the formula, This is the steady-state solubility concentration, expressed in mg / L. This represents a one-dimensional lookup table and linear interpolation operation for the temperature dimension, or the use of a linear regression equation. Perform calculations, where and This is a preset steady-state characteristic constant.

[0079] Through the above steps, the system determines the "concentration limit" of the mineralization bin at the current moment. ) and "lower limit of concentration" This effectively defines the dynamic range of the subsequent exponential decay model, ensuring that the model's predicted values ​​conform to objective physical and chemical laws.

[0080] S130, Construct an instantaneous concentration decay model for the mineral concentrate. In this embodiment, when the mineralization injection subsystem activates the solenoid valve for pulse injection, fresh pure water generated by the pure water delivery subsystem enters the mineral dissolution chamber in a pulsed manner. Since the mineral dissolution chamber contains a high-concentration mineralization soaking solution formed over a long period of settling, and the chamber structure is typically designed as a columnar shape with a certain depth-to-diameter ratio or containing an internal baffle structure, the newly introduced pure water cannot instantly and completely replace the existing liquid. Instead, it undergoes a continuous mixing and dilution process with the stagnant liquid within the chamber. From a fluid dynamics perspective, this physical process approximates a non-ideal fully mixed flow reactor (CSTR) model, where the liquid replacement within the chamber follows first-order reaction kinetics. Therefore, the mineral concentration at the outlet of the mineral dissolution chamber does not undergo a step-like abrupt change with fluid flow, but rather exhibits a smooth exponential decay characteristic with increasing water flow, until the high-concentration liquid within the chamber is completely replaced, and the outlet concentration gradually approaches the steady-state dissolved concentration under dynamic equilibrium.

[0081] Based on the above physical mechanism, the model prediction module 110 uses the static peak concentration determined in the aforementioned steps. and steady-state dissolution concentration The following instantaneous concentration decay model describing the change in concentration with volume is constructed:

[0082] ;

[0083] In the formula, This indicates that the cumulative volume of liquid flowing through the mineral leaching chamber during the current water intake cycle is [amount missing]. At that time, the theoretical instantaneous concentration at the outlet of the mineral leaching chamber, in mg / L; The cumulative net volume of water flowing through the mineral dissolution chamber from the start of this water production cycle is expressed in liters (L). This value is reset to zero at the start of each water production cycle and is accumulated by the flow pulse count during the water production process. It is a natural constant; This is the scour attenuation coefficient, in L. -1 .

[0084] It should be noted that the scour attenuation coefficient It is a key physical parameter characterizing the fluid displacement rate of the mineral leaching chamber, and its value is related to the effective volume of the mineral leaching chamber. The net water volume (i.e., the volume of the solid medium after deducting the volume of the water) is inversely related to the given volume, thus satisfying the condition. In this embodiment, the effective volume of the mineral leaching chamber... The design range is 0.2L to 0.5L, therefore the corresponding scour attenuation coefficient is... The value range is set to 2.0 to 5.0. This parameter is fixed in the memory according to the hardware specifications when the system is shipped from the factory; The larger the value, the smaller the volume of the container, the faster the high-concentration liquid is emptied, and the steeper the concentration decay curve.

[0085] S140 introduces a filter element health factor correction model to establish the final estimation benchmark. The aforementioned basic instantaneous concentration decay model is a theoretical curve derived from mineralized media under standard conditions (i.e., the filter element is brand new and its surface is free of contamination). However, in practical applications, as the usage time increases, the solid mineralized media undergoes irreversible physicochemical changes: on the one hand, the total mass of the media decreases due to continuous leaching, leading to a decrease in the total solid-liquid contact surface area; on the other hand, impurities in the water may form a coating layer on the media surface, or the media surface may become passivated, resulting in a decrease in its leaching efficiency.

[0086] To compensate for the concentration prediction deviation caused by material aging, the system introduces a filter health factor. Filter element health factors It is a dimensionless normalization coefficient with a value range of (0,1). It is used when the system is initially installed or when a new filter element is replaced. It is initialized to 1.0 by default. As the system runs, this factor is dynamically updated based on the actual execution error calculated by the feedback calibration module and stored in non-volatile memory. The model prediction module 110 reads the current filter health factor. The overall gain of the above instantaneous concentration decay model is corrected to generate the final instantaneous concentration estimation model for real-time control:

[0087] ;

[0088] In the formula, The concentration of the mineral concentrate after aging correction is estimated in real time, in mg / L.

[0089] Through this step, the control subsystem abstracts the complex nonlinear dissolution process into a computable mathematical formula with the outflow volume v as the independent variable. In subsequent water production, regardless of when or how much water is drawn by the user, the system only needs to input the current cumulative outflow volume v to accurately determine the mineral concentration of the liquid injected into the main channel from the mineralization branch at that moment. This provides a mathematical benchmark for accurately calculating the injection pulse width and eliminates proportioning errors caused by concentration fluctuations.

[0090] See attached document Figure 3In this embodiment, the injection execution module 120 abandons the traditional fixed-time-period polling control method and instead adopts an event-driven architecture based on flow volume. The core logic of this architecture lies in discretizing the continuously flowing fluid in the spatial domain into a series of tiny, equal-volume control units, and calculating the mass deficit of minerals for each volume unit, thereby achieving high-precision feedforward control completely decoupled from pipeline flow velocity fluctuations. This process specifically includes:

[0091] S210 establishes an interrupt triggering mechanism based on flow pulses and defines single-step volume. The microprocessor of the control subsystem is equipped with an external interrupt input pin, which is electrically connected to the flow metering unit in the signal sensing subsystem. When the water purification cycle starts, the fluid in the pure water delivery subsystem drives the impeller inside the flow metering unit to rotate, generating continuous Hall pulse signals. The microprocessor's hardware counter is configured to count the effective edges (rising or falling edges) of this signal in real time.

[0092] To balance control precision and processor load, the system sets trigger threshold parameters. . It is a positive integer representing the cumulative number of pulses corresponding to a single control cycle. When the hardware counter reaches... When the microprocessor's interrupt service routine is triggered or the flag is set, a calculation process for the mineral injection volume is initiated. The counter is then automatically reset and restarts counting. At this point, the volume of pure water flowing through the microprocessor corresponding to this calculation process is defined as the single-step injection volume. .

[0093] Single-step injection volume The physical values ​​are determined by the following formula:

[0094] ;

[0095] In the formula, This refers to the injection volume in a single step, expressed in liters (L). The preset threshold for the number of interrupt trigger pulses is set to a range of 10 to 50 in this embodiment. The selection principle of this range is to ensure that the calculation frequency is not less than 1Hz at the minimum rated flow rate of the system (e.g., 0.5L / min) to ensure the real-time performance of the control. This represents the fluid volume resolution per pulse, measured in L / Pulse. It should be noted that... Depending on the hardware characteristics of the selected flow meter (i.e., the reciprocal of the flow coefficient K), the typical value of this parameter in the Hall flow meter selected in this embodiment is between 0.001L and 0.003L.

[0096] Through the above mechanism, the system logically divides the continuously flowing pure water into a series of segments with constant volume. The system uses discrete volume units. Regardless of how drastic the flow rate fluctuations are caused by the influent pressure, the system always performs proportioning calculations for this fixed volume unit, thereby eliminating the interference of nonlinear flow rate changes on concentration control accuracy at the physical level.

[0097] S220 calculates the single-step mineral quality requirement and performs an accumulation operation. Within each interrupt-triggered cycle, the injection execution module 120 performs a feedforward calculation to determine the required single-step injection volume. Theoretically, the physical mass of solid minerals needs to be added to achieve the user-defined target taste.

[0098] The microprocessor reads the target concentration value set by the user. Simultaneously, real-time signals from the first conductivity probe located behind the reverse osmosis membrane were acquired, and the baseline concentration of pure water was obtained after temperature compensation and analog-to-digital conversion. To prevent probe noise from introducing calculation errors, the system... Median filtering was employed. Next, the system calculated the required mineral mass increment for the current discrete volume unit and stored it in a software-defined mass accumulation register. middle.

[0099] Mass accumulation register This is a floating-point variable space allocated in the microprocessor's random access memory (RAM). Its initial state is set to zero at the start of each water production cycle. It stores the total mass of minerals for which the demand has been calculated but not yet released through the valve. This accumulation process follows the following discrete integral logic:

[0100] ;

[0101] In the formula, Indicates the first At the end of each calculation cycle, the total mass of the minerals to be injected stored in the cumulative register, in mg; This represents the register residual value at the end of the previous cycle; The target effluent mineral concentration set for the system, in mg / L; This represents the current measured baseline concentration of pure water, in mg / L.

[0102] Introduced in the formula The physical meaning of the function lies in the boundary constraints: when the original water baseline concentration is detected... The concentration is already higher than or equal to the target concentration. At times (such as during the initial replacement of the filter cartridge or when the water source fluctuates), the system determines that no minerals need to be injected at this time, and the mass increment is zero to prevent logical errors caused by negative values.

[0103] It should be noted that setting the mass accumulation register... This is the key technical means by which this embodiment solves the "actuator minimum response limit". Due to the single-step injection volume Typically, the volume is very small (e.g., only 20ml), requiring an extremely small amount of minerals. The corresponding valve opening time may be much shorter than the physical opening dead time of a high-frequency solenoid valve. Direct actuation would render the valve ineffective. By accumulating mass, the system can temporarily store multiple small mass demands until the accumulated value reaches the minimum threshold required for the linear operating range of the solenoid valve, thereby achieving precise pulse release.

[0104] S230, perform pressure compensation correction based on the main flow velocity. In this embodiment, although the system uses a pressure-independent mass accumulation algorithm, the physical outflow characteristics of the solenoid valve itself are still subject to the fluid dynamics environment within the main flow channel. According to Bernoulli's principle, when the pure water flow velocity in the main flow channel increases, the dynamic pressure at the confluence node increases, causing a change in the back pressure acting on the mineralization branch outlet. This back pressure fluctuation changes the pressure difference before and after the solenoid valve, resulting in a deviation between the actual injected liquid volume and the rated value under the same opening drive duration.

[0105] To eliminate the impact of this hydraulic coupling effect on the proportioning accuracy, the injection execution module 120 is configured to calculate the real-time flow velocity of the main channel at the current moment and adjust the flow velocity parameters of the solenoid valve accordingly. The microprocessor records the time interval between two adjacent flow interruption triggers. Using the formula The real-time flow rate is obtained. Subsequently, the system introduces a pressure compensation coefficient to dynamically correct the theoretical volumetric flow rate of the solenoid valve. The correction logic follows a linear regression model:

[0106] ;

[0107] In the formula, The current theoretical volumetric flow rate of the solenoid valve after pressure compensation correction is expressed in L / ms. For the solenoid valve under standard calibration conditions (i.e., the main flow velocity is the reference flow velocity) The rated volumetric flow rate (at time), which is stored in the system firmware; The measured real-time flow rate of the mainstream channel is in L / min. The reference flow rate for calibration experiments is usually set to the system's typical operating flow rate (e.g., 1.0 L / min). This is the back pressure influence coefficient, with units of (L / min). -1 .

[0108] It should be noted that the back pressure influence coefficient This is a system characteristic constant characterizing the degree to which the pressure fluctuation of the main channel suppresses the injection capacity of the branch. Its value depends on the mechanical structure design of the tee junction (such as the pipe diameter ratio and the junction angle) and the flow resistance characteristics of the pipeline. In this embodiment, The method for obtaining this information is as follows: during the system's factory calibration phase, respectively... The actual water output of the solenoid valve is measured under maximum rated flow rate conditions, and the slope is calculated using a two-point method or multi-point fitting. For typical household water purifier piping structures, The typical value range is 0.05 to 0.15. This step ensures that the flow velocity parameters referenced by the system can accurately reflect the actual hydraulic conditions when calculating the start-up duration.

[0109] S240 executes the "accumulation-release" decision logic to actuate the solenoid valve. This step is configured to address the physical limitations between the trace mineral injection requirements and the dead zone of the solenoid valve's physical operation. After completing the flow rate correction, the injection execution module 120 utilizes the mass accumulation register updated in the aforementioned steps. and the instantaneous concentration estimation model determined by step S140. Calculate the theoretical valve opening time required to completely release the mineral mass currently accumulated in the register. :

[0110] ;

[0111] In the formula, The theoretical on-time is in milliseconds (ms). This represents the total mass of minerals to be injected currently accumulated in the register, in mg. The volumetric flow rate of the solenoid valve after correction in step S230 is expressed in L / ms. The instantaneous estimated concentration of the mineralized solution corresponding to the current cumulative effluent volume is given in mg / L.

[0112] Subsequently, the microprocessor will calculate the... With the preset minimum physical response time of the solenoid valve Perform numerical comparisons and execute the corresponding control logic based on the comparison results:

[0113] when At this point, the system determines that the current accumulated amount is insufficient to support a single effective valve action. If forced opening is attempted at this time, the solenoid valve may not be able to overcome the spring force and hysteresis effect due to the insufficient energizing time, resulting in the valve core not fully opening or unstable opening, thus introducing nonlinear errors. Therefore, the microprocessor does not output a drive signal, and the solenoid valve remains closed. (Register) The value remains unchanged and is carried over to the next control cycle to continue accumulating.

[0114] when When the system determines that the opening condition is met, the microprocessor outputs a width of [value missing] to the solenoid valve drive circuit. A high-level pulse drives the solenoid valve to open. Usually equals But Exceeding the system's maximum allowed single injection duration In special circumstances, Restricted to After the action is completed, the system deducts the released mass from the accumulation register based on the actual execution duration, and the update logic is as follows:

[0115] ;

[0116] In the formula, This is the updated residual value of the cumulative register.

[0117] It is important to note that the minimum physical response time The critical threshold is determined based on the electromechanical characteristics of the selected solenoid valve. For the direct-acting miniature solenoid valve used in this embodiment, The threshold is set between 5ms and 8ms. This threshold is set based on the principle that it is greater than the sum of the magnetic field establishment time of the solenoid valve coil and the mechanical movement delay. Through this decision logic, the system ensures that every actual injection action occurs within the linear operating range of the solenoid valve, thereby achieving precise dose control on a microscopic timescale.

[0118] See attached document Figure 4 When a complete water production cycle ends (i.e., the user turns off the tap and the system detects the cessation of flow), the control subsystem uses the process data stored in volatile memory to perform the following data alignment and quality calculation steps:

[0119] S310 performs FIFO queue-based pipeline transmission delay compensation and data alignment. In the physical pipeline structure, the first conductivity probe used to detect the baseline concentration of pure water is typically positioned downstream of the reverse osmosis membrane and upstream of the mineralization injection point, while the second conductivity probe used to detect the final effluent concentration is positioned downstream of the mineralization confluence point or even near the outlet. A physical connecting pipeline exists between the two, which has a fixed internal fluid volume, defined as the transmission delay volume. Because fluid travels a certain physical distance through this section of the pipeline, the first conductivity data and the second conductivity data collected by the microprocessor at the same moment do not correspond to the same micro-element of water. If the two are directly calculated by difference, significant timing errors will be introduced, especially when the flow velocity fluctuates greatly. This error will cause a serious phase shift in the calculated mineral increment.

[0120] To eliminate this spatiotemporal asynchrony caused by the physical layout of the sensors, in this embodiment, the feedback calibration module 130 constructs a circular data buffer based on the first-in-first-out (FIFO) principle in the microprocessor's random access memory (RAM). The depth of this buffer (i.e., the number of storage cells) is... It is not a fixed constant, but an integer value dynamically determined based on the system's hardware geometry parameters. Specifically, the number of delay steps. Determined by the following formula:

[0121] ;

[0122] In the formula, The depth of the FIFO queue is determined by taking the nearest integer. The transmission delay volume, measured in liters (L), refers to the sum of the physical volumes of all fluid channels from the center point of the mineralization injection three-way valve to the sensing center point of the second conductivity probe. In this embodiment, this parameter is calculated through 3D modeling or determined by water injection weighing during the product design phase and is stored in memory as a system constant, with a typical value range of 0.015L-0.050L.

[0123] During the water production process, whenever a flow interruption is triggered (i.e., when a flow passes through a...), When a liquid is detected, the system performs an enqueue and dequeue operation: it sets the current baseline concentration of pure water (collected and filtered) to the specified value. Write to the tail of the queue; simultaneously read from the head of the queue. Historical baseline concentration data stored from a previous cycle is used as the basis for comparing the mixed water concentration collected at the current moment. Time-domain strictly aligned corrected baseline concentration It should be noted that at the very beginning of each water production cycle... Within each cycle, since the queue is not yet full, the system reads by default. Use the current measured value or the steady-state value of the previous cycle to avoid calculation jitter during cold starts.

[0124] S320 performs full-cycle net flux integration to calculate the total mass of actual injected minerals. After completing data time-domain alignment, the feedback calibration module 130 performs discrete numerical integration on the entire process data of this water production cycle to quantify how much mineral the system actually injected into the water in the real physical world. This process aims to isolate the influence of raw water TDS fluctuations and extract only the net conductivity increment generated by the mineralization operation.

[0125] The system reads all sampling point data recorded in the memory for this water production cycle and calculates the total mass of injected minerals using the following formula. :

[0126] ;

[0127] In the formula, This represents the total mass of minerals actually dissolved in the water during this water production cycle, expressed in mg. This represents the total number of flow interruptions triggered during this water production cycle. The index of the discrete sampling point; For the first The concentration of minerals in the mixed effluent measured at each sampling point, in mg / L; The first step after alignment via step S310 The baseline concentration of pure water corresponding to each sampling point is expressed in mg / L. The formula introduces... The purpose of the function is to filter out instantaneous negative values ​​that may be generated by sensor measurement noise, ensuring that the physical meaning of mass accumulation is non-negative.

[0128] At the same time, the system retrieves the theoretical data calculated and recorded by the model prediction module 110 and the injection execution module 120 during the water production process, and calculates the theoretical target total mass for this cycle. .

[0129] ;

[0130] In the formula, The theoretical total mass of minerals that should be injected into the water, estimated based on the current control model, is expressed in mg. This represents the total number of times the solenoid valve opening action is actually performed in this cycle; The index number of the action to be performed; For the first The actual solenoid valve opening time for this action, in milliseconds; The pressure-compensated flow rate of the solenoid valve at the moment the action is performed, expressed in L / ms; The instantaneous estimated concentration of the mineralized solution at the moment the action is performed is given in mg / L.

[0131] Through the above calculations, the system obtained two key values: It represents the true result of the physical world, while This represents the theoretical expectation of the control model under current parameters. The ratio of the two directly reflects the current filter health factor. The accuracy. If Significantly smaller than This indicates that the filter element may be clogged or aged, resulting in an actual dissolution rate lower than the model's expectation. The system will adjust the filter element's health factor accordingly in subsequent steps. This allows the system to automatically increase the operating time during the next water production cycle to compensate for the shortfall.

[0132] S330: Calculate the relative execution error and update the filter health factor. This step is configured to establish a numerical mapping between physical world observations and control model predictions to correct model biases caused by filter aging, channel blockage, or surface passivation. In this embodiment, the system utilizes the actual total mass of injected minerals obtained in step S320. Total mass of theoretical target To assess the degree of deviation of the current control parameters under real operating conditions.

[0133] To suppress sporadic sensor noise that may occur in a single measurement (such as interference from tiny bubbles or reading fluctuations caused by water turbulence), this embodiment does not employ a strategy of directly resetting parameters based on a single deviation. Instead, it uses an exponentially weighted moving average (EWMA) algorithm to evaluate the filter cartridge health factor. Perform a smooth update. Filter health factors. It is a dimensionless correction factor, whose physical meaning represents the ratio of the actual effective dissolution capacity of the current filter element to the standard factory condition. The initial default value is 1.0.

[0134] Specifically, the update logic follows the iterative formula as follows:

[0135] ;

[0136] In the formula, This refers to the new filter health factor obtained after this iteration of calculation; This refers to the health factors of the old filter cartridges stored in the system before the start of this water production cycle. The convergence coefficient is a dimensionless constant between 0 and 1.

[0137] It should be noted that the iterative convergence coefficient It is a key parameter for balancing system response speed and anti-interference capability. In this embodiment, The value range is set to 0.05 to 0.20. This range is established based on the following control principles: If If the value is less than 0.05, the system's tracking lag in the actual filter element degradation is too large, and it cannot compensate in a timely manner; if If the value is greater than 0.20, the random error of a single measurement will lead to... The value fluctuates drastically, causing the taste of the water to be unstable.

[0138] Furthermore, to prevent invalid data from contaminating the control model, the system pre-sets validity determination logic before performing an update: only if the theoretical target total mass of the current cycle is valid... Greater than the minimum evaluation threshold The above update process is triggered only when the concentration is set to 0.5mg in this embodiment. This threshold is used to filter out invalid water production cycles due to insufficient sample size caused by users turning on the tap for a very short time (e.g., less than 2 seconds).

[0139] S340, non-volatile storage and cross-cycle application of execution parameters. To achieve continuous adaptive control across water production cycles, the system is configured to permanently store updated control parameters. When step S330 calculates... Then, the microprocessor first performs a boundary safety check on the value. In this embodiment, the system presets a safe range for the health factor. Typical values ​​are [0.3, 1.5]. If the calculated... If the value exceeds this range, the system determines that the sensor system is abnormal or the filter element has suffered a non-linear catastrophic failure. In this case, the update operation is not performed, and the maintenance alarm flag is set.

[0140] If the value is within a reasonable range, the microprocessor will... Write to a specific address in non-volatile memory (such as EEPROM or Flash data area) to overwrite the original data. This write operation is configured to be performed in standby mode after the system detects that the water flow has stopped, in order to prevent flash write / erase operations during periods of high water production load from affecting the real-time performance of the control.

[0141] It should be noted that, given the different media consumption rates and aging characteristics of the different mineralization branches, the system has established independent data storage areas for the four mineralization branches in the non-volatile memory.

[0142] When the system performs the above update operation, the microprocessor only checks the filter health factor corresponding to the target mineralization branch that is activated and used in this water production cycle. The health factors of other unused branches remain unchanged.

[0143] For example, if the water production is in "male mode," the system only uses the calculated relative execution error to update the health factor of the first mineralization branch. Through this discrete storage strategy, the system can accurately record and track the true lifespan of each individual filter element, avoiding confusion in aging assessments caused by frequent mode switching. This ensures that regardless of the mode the user switches to, the model prediction module 110 can retrieve the most accurate historical correction coefficients for that specific flow channel.

[0144] The next time the user turns on the tap to start the water production process, the model prediction module 110 will first read the latest data from this non-volatile memory address during the initialization phase. The value is then directly assigned to the filter health factor in the instantaneous concentration decay model of step S140. Therefore, the system calculates the instantaneous estimated concentration of the mineralized solution for the next cycle according to step S140. At that time, it will automatically include a correction gain based on the feedback results of the previous cycle, thereby calculating the required solenoid valve opening time based on the updated health status.

[0145] Through the above logic, the system can automatically sense the gradual changes in the physical properties of the filter element and automatically increase or decrease the opening time of the solenoid valve to ensure that the mineral concentration control accuracy is always maintained within the preset range throughout the entire life cycle of the filter element.

[0146] To illustrate the working principle of this invention more intuitively, a typical household morning water usage scenario will be used for detailed explanation below. Scenario: The user's home is equipped with the water purification system of this invention.

[0147] First mineralization branch (male mode / zinc enrichment): The filter has been used for 3 months, current health factor. .

[0148] Third mineralization branch (Kids mode / Low load): New filter, current health factors .

[0149] Environmental parameters: 7:00 AM, water temperature T=20℃, system has been shut down overnight. h).

[0150] Step 1: Father fetches water (Male mode)

[0151] Mode initialization: The father taps "Male Mode" on the touchscreen. The system locks the first branch and sets the target concentration. mg / L. System read .

[0152] Model Construction (S100): Due to an 8-hour downtime, the system used dissolution kinetic parameters to calculate the static peak concentration of the first branch. mg / L (high concentration due to soaking overnight). Steady-state concentration mg / L. System attenuation model construction:

[0153] ;

[0154] Dynamic injection (S200-S300):

[0155] Initial stage (first cup of water): due to The concentration is very high (approximately 7.8 mg / L), and the system calculates that no injection is needed at this point, or only a very short valve opening is required. The injection execution module controls the solenoid valve to operate with an extremely low duty cycle, utilizing the high concentration of the "stored water" itself to meet the target of 4.0 mg / L, thus avoiding the problem of "the first cup of water being too concentrated" in traditional water purifiers.

[0156] Subsequent stage (second cup of water): As the stored water flows out, The concentration drops rapidly. The system automatically increases the opening time of the solenoid valve to compensate for the concentration decay and maintain the effluent concentration at 4.0±0.2mg / L.

[0157] Feedback Update (S340): Water sampling completed. The system calculation shows that the actual average concentration is slightly lower than the target (e.g., 3.9 mg / L), indicating that the filter cartridge is aging slightly faster. Updated to 0.915 and saved for use in the next male mode.

[0158] Step 2: Switch to Kids Mode via the App

[0159] Mode switching: The mother then fetches water for the child and clicks "Child Mode." The system immediately cuts off the first branch and locks the third branch, reducing the target concentration to [value missing]. mg / L. Read .

[0160] Independent control: Although the first branch has already worked, the third branch remains in an "overnight idle" state. The system independently looks up the table and finds that the third branch... Only 2.5 mg / L (low precipitation medium characteristics).

[0161] Safety flow restriction: The system activates the flow restriction logic. Even if the model calculation requires the valve to be open for a long time, the system will forcibly limit the injection volume at one time to ensure that the concentration of the effluent does not exceed 1.0 mg / L, thus protecting children's kidneys.

[0162] Experimental verification: In order to verify the control accuracy and stability of the "control method based on attenuation model and feedback calibration" proposed in this invention under multiple operating conditions, a standard test platform was built for comparative experiments.

[0163] Test subjects: Example group: Prototypes using the system of this invention, equipped with the complete control algorithm, and with the filter health factor enabled. Dynamic updates. Comparative control group: A control prototype equipped with a traditional flow ratio dosing algorithm. The hardware structure of this prototype is consistent with the embodiment, but its control logic does not include the concentration decay model and feedback calibration module of this invention. Its control strategy is: the system opens the solenoid valve according to a fixed ratio based solely on the number of pulses fed back by the flow sensor (i.e., the set...). ,in (where is a fixed constant), thus simulating the effect of conventional active mineralization control without the adaptive algorithm of this invention.

[0164] Test conditions: Simulate the state of the filter cartridge in the middle of its service life (approximately 40% aging), and artificially introduce fluctuations in the main flow rate (randomly varying from 0.8L / min to 1.5L / min) during the test to simulate unstable water pressure in a household.

[0165] Target parameter: Set the target mineral concentration mg / L.

[0166] Analysis of experimental results: Refer to the appendix Figure 5 and attached Figure 6 Experimental data show that:

[0167] Comparison of anti-attenuation capabilities (see appendix) Figure 5 (First part): During a single water intake process, as the volume of water output increases, the concentration in the mineralization chamber naturally decreases.

[0168] The effluent concentration of the control group showed a significant "high at first, low at last" trend, with an initial concentration as high as 6 mg / L (far exceeding the target) and a final concentration that dropped to 1.5 mg / L, making it impossible to maintain a constant level.

[0169] The example set passed The model predicts the attenuation amount in real time and increases the valve opening time, so that the effluent concentration always stays close to the target line of 3.5 mg / L throughout the process, with a fluctuation range of less than ±0.3 mg / L.

[0170] Comparison of anti-interference capabilities: When the flow rate changes abruptly (between 1.5L and 2.5L in the figure), the comparative group, lacking flow rate and pressure compensation, experienced a sharp decrease in concentration as the flow rate increased. In contrast, the example group, through the pressure compensation algorithm injected into the execution module, automatically adjusted the solenoid valve's action, maintaining a stable concentration.

[0171] Long-term adaptive capability: In 50 consecutive water production cycle tests, the root mean square error of the example group showed a convergent decreasing trend with the increase of the number of cycles. This proves that the feedback calibration module 130 is working effectively and that the filter health factor is... Gradually approaching the true physical state of the filter element, the system possesses a self-learning characteristic that becomes more accurate with use.

[0172] Conclusion: Compared with traditional technologies, the technical solution proposed in this invention effectively solves the problems of excessively high initial water concentration, concentration decay during continuous water intake, and control inaccuracy caused by filter aging in intermittent mineralized water injection, and achieves precise mineral ratio throughout the entire life cycle.

[0173] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A water purification system with adjustable mineral concentration, characterized in that, include: The pure water delivery subsystem is configured to output filtered base water flow. The mineralization injection subsystem is connected to the pure water delivery subsystem and is configured to store mineral concentrate and inject the mineral concentrate into the base water flow through valve operation. The signal sensing subsystem is configured to detect fluid flow rate, temperature, and conductivity parameters. A control subsystem is connected to the pure water delivery subsystem, the mineralization injection subsystem, and the signal sensing subsystem, respectively. The control subsystem includes a model prediction module, an injection execution module, and a feedback calibration module. The model prediction module is configured to read the downtime of the water purification system and the ambient temperature when the water production cycle starts, and combine the filter health factor to construct a decay function model describing the change of the concentration of the concentrate inside the mineralization injection subsystem with the outflow volume. The injection execution module is configured to calculate the mineral quality requirement based on the flow signal and the set target concentration and store it in the accumulation register, calculate the valve opening time required to consume the value of the accumulation register according to the decay function model, and drive the mineralization injection subsystem to operate. The feedback calibration module is configured to align the flow rate data with the mixed water quality data after the water production cycle is completed, calculate the execution error between the actual total mass of injected minerals and the theoretical target total mass, and update the filter health factor based on the execution error for the model prediction module to call next time.

2. The water purification system with adjustable mineral concentration according to claim 1, characterized in that, The mineralization injection subsystem adopts a parallel topology architecture, including: A multi-channel fluid switching unit, wherein the inlet of the multi-channel fluid switching unit is connected to the pure water delivery subsystem and is configured to open a designated flow channel according to the instructions of the control subsystem; The first mineralization branch, the second mineralization branch, the third mineralization branch, and the fourth mineralization branch are connected in parallel to the outlet of the multi-fluid switching unit. The first mineralization branch, the second mineralization branch, the third mineralization branch, and the fourth mineralization branch are filled with differentiated mineralization media configured for different population patterns. A high-frequency electromagnetic regulating valve is installed at the common outlet where the first mineralization branch, the second mineralization branch, the third mineralization branch, and the fourth mineralization branch converge, and is configured to pulse-type cut-off and release of the mineral concentrate passing through it.

3. The water purification system with adjustable mineral concentration according to claim 1, characterized in that, The mineralization injection subsystem includes a mineral dissolution chamber, which is configured as a closed container filled with a solid mineralization medium for generating a mineral concentrate through diffusion at the solid-liquid interface under static conditions. The model prediction module is pre-set with dissolution characteristic parameters. The model prediction module is configured to determine the initial peak concentration and steady-state dissolution concentration by querying a pre-set data mapping table or executing a dissolution kinetics calculation model based on the downtime of the water purification system and the ambient temperature. The module then constructs a decay function model using the initial peak concentration and steady-state dissolution concentration as boundary conditions to characterize the physical process of nonlinearly decreasing concentration as the cumulative outflow volume increases.

4. The water purification system with adjustable mineral concentration according to claim 1, characterized in that, The injection execution module establishes a step-triggered mechanism based on flow pulses, which discretizes the continuous water flow into single-step injection volumes of equal volume. The injection execution module is configured to: calculate the mineral mass increment required for each single-step injection volume and accumulate it to the accumulation register when each single-step injection volume is triggered; output a drive signal only when the calculated valve opening duration is greater than the hardware dead zone threshold, and deduct the corresponding mass value from the accumulation register after the action is completed; otherwise, keep the value of the accumulation register until the next calculation cycle.

5. A water purification system with adjustable mineral concentration according to claim 1, characterized in that, The injection execution module is also configured with pressure compensation logic; The injection execution module calculates the real-time flow velocity of the main channel by monitoring the time interval between adjacent flow pulses, and corrects the theoretical volumetric flow velocity parameters of the solenoid valve based on a preset pressure coupling model. The injection execution module uses the corrected theoretical volumetric flow velocity parameters of the solenoid valve to calculate the theoretical valve opening time required to consume the current value of the accumulation register, so as to compensate for the inhibitory effect of the change in the main channel flow velocity on the injection capacity of the mineralization injection subsystem.

6. A water purification system with adjustable mineral concentration according to claim 1, characterized in that, The feedback calibration module is configured with a delay alignment algorithm based on a first-in-first-out queue; The feedback calibration module determines the transmission delay parameter based on the physical pipeline volume between the junction point of the pure water delivery subsystem and the mineralization injection subsystem and the detection point of the signal sensing subsystem. The feedback calibration module uses a first-in-first-out queue to perform phase shifting on the pure water baseline concentration data, so that the pure water baseline concentration data and the final effluent concentration data after mixing are strictly corresponding in the time domain, and calculates the total mass of the actual injected minerals accordingly.

7. A water purification system with adjustable mineral concentration according to claim 1, characterized in that, The feedback calibration module is equipped with an exponentially weighted moving average algorithm to update the filter health factor. The feedback calibration module calculates the ratio of the actual total mass of injected minerals to the theoretical target total mass, and uses a preset iterative convergence coefficient to weight and correct the current filter health factor, so that the updated filter health factor represents the true dissolution efficiency of the current mineralized medium relative to the standard state. The control subsystem is configured to disable update operations and trigger an alarm when a relative error is detected to exceed a preset safety range.

8. A water purification system with adjustable mineral concentration according to claim 2, characterized in that, The control subsystem includes a non-volatile memory, which is configured with independent data storage areas corresponding to the first mineralization branch, the second mineralization branch, the third mineralization branch and the fourth mineralization branch respectively. The feedback calibration module is configured to update only the filter health factors corresponding to the first, second, third, or fourth mineralization branches that are activated in the current water production cycle, and write the updated data into the corresponding independent data storage area to achieve independent tracking and storage of the aging status of multiple parallel filter elements.

9. A water purification system with adjustable mineral concentration according to claim 1, characterized in that, The system also includes a static mixing unit; The static mixing unit is located downstream of the physical junction of the pure water delivery subsystem and the mineralization injection subsystem. It is configured to generate turbulence using its internal baffle structure to uniformly disperse the pulsed mineral concentrate injected by the mineralization injection subsystem into the base water flow output by the pure water delivery subsystem.

10. A water purification system with adjustable mineral concentration according to claim 2, characterized in that, The system also includes a human-computer interaction unit, configured to receive user crowd mode selection instructions; The control subsystem is configured to lock the third mineralization branch and activate the safety current limiting logic in response to the child mode command. The safety current limiting logic forcibly limits the maximum single injection duration calculated by the injection execution module to prevent the effluent concentration from exceeding the physiological safety threshold for children.

Images