Water purifier pre-flushing control method and system and water purifier

Abstract

The invention relates to the technical field of water treatment equipment, in particular to a water purifier pre-flushing control method and system and a water purifier. According to the method, before restarting, a digital twin model of the water purifier is constructed on the basis of the shutdown duration, the current TDS of raw water, the water temperature and the running parameters of the reverse osmosis membrane component, and the change condition of the TDS of the effluent on the pure water side in the preset evaluation stage is predicted; and the obtained first water quality predicted value and a preset first cup water target TDS threshold value are used as input of a fuzzy control algorithm, the pre-flushing duration of the starting is generated in a self-adaptive mode, and the pre-flushing process is controlled according to the pre-flushing duration. Whether pre-flushing is executed or not and the pre-flushing duration can be determined in a targeted mode under different shutdown histories and water quality working conditions, the situation that the pre-flushing time is too long or too short is avoided, the TDS control consistency of the first cup of water is improved, and meanwhile water resource waste and unnecessary user waiting time caused by invalid pre-flushing are reduced.

Description

Technical Field

[0001] This invention relates to the field of water treatment equipment technology, specifically to a pre-rinse control method, system, and pure water machine for a pure water machine. Background Technology

[0002] Reverse osmosis water purifiers are widely used for residential drinking water purification. Among them, tankless reverse osmosis water purifiers have gradually become the mainstream due to their small size, compact structure, and avoidance of the secondary pollution risk associated with storage tanks. These purifiers typically use a pump to drive raw water through a reverse osmosis membrane module to produce pure water, which is then directly output to users for drinking via pipelines. Because concentration polarization exists on the inlet side and within the membrane chamber of the reverse osmosis membrane, and because dissolved salts on the raw water side slowly seep back to the pure water side during shutdown, a certain amount of residual water with high conductivity remains in the membrane chamber and pure water pipelines. When the water purifier restarts, the initial permeate water will carry these residual salts, resulting in a significantly higher TDS (Total Dissolved Solids) in the so-called "first cup of water" compared to the TDS during the stable water production phase.

[0003] To improve the quality of the "first cup of water," existing tankless reverse osmosis water purifiers typically perform a pre-rinse upon restarting, discharging some high-TDS wastewater to the wastewater side or drain before switching back to normal water supply. However, existing pre-rinse control schemes mainly suffer from the following two problems: First, many products use a fixed pre-rinse duration or roughly set the pre-rinse time based on several levels according to the current shutdown time, without fully considering the combined effects of raw water TDS, water temperature changes, membrane flux decay, etc. This results in pre-rinse times being too long under some operating conditions, causing water waste and increased user waiting time, while pre-rinse times are too short, and the TDS of the first cup of water may still exceed the expected range, leading to poor water quality consistency. Second, some solutions attempt to install TDS sensors on the pure water side to control the start and stop of pre-rinse based on the detection results. However, such solutions usually require additional sensors and pipe interfaces to be installed inside the pure water path, making it difficult to make compatible modifications to existing models. At the same time, the sensors are mostly installed near the outlet, and in order to wait for the detected value to drop to the threshold, it is often necessary to continuously discharge pure water, making it difficult to accurately control the actual flushing volume and easily causing additional water waste. Furthermore, these sensors only provide the instantaneous TDS value at a certain location, making it difficult to accurately characterize the overall water state inside the reverse osmosis membrane chamber. When the membrane module is fouled, aged, or has local abnormalities, the detection results may also be delayed or biased, making it difficult to guarantee the stability of the control effect.

[0004] Therefore, existing technologies still struggle to balance improving the quality of the first cup of water, reducing water consumption during the pre-rinse process, and shortening user waiting time, leaving room for improvement. Summary of the Invention

[0005] (I) The technical problem to be solved by the present invention is that in a reverse osmosis pure water machine without a storage tank, when restarting, the existing technology is difficult to reflect in a timely and accurate manner the impact of the state of the water in the membrane chamber and pure water pipeline during the shutdown period on the TDS of the first cup of water. The pre-rinse control is mostly set in a rough way, such as setting a fixed duration or dividing the gear only according to the shutdown time. It is difficult to take into account the water quality of the first cup of water, the water consumption of pre-rinse, and the user's waiting time under different operating conditions.

[0006] (II) Technical Solution To address the aforementioned technical problems, this invention provides a pre-rinsing control method for a reverse osmosis water purifier, applicable to a tankless reverse osmosis water purifier that connects the raw water side and the outlet water side. The water purifier includes a reverse osmosis membrane assembly, a water pump, and a valve for switching between the pre-rinsing passage and the water supply passage. The method includes: S1. When the pure water machine switches from water production operation to shutdown state, record the shutdown start time. Before restarting, determine the shutdown duration according to the real-time clock, and collect the current raw water TDS, water temperature and the operating parameters of the reverse osmosis membrane module to form the operating condition data for this startup. S2, input the downtime and the operating condition data into the pre-established digital twin model of the pure water machine, obtain the effluent water quality prediction result based on the digital twin model to characterize the change of TDS in the pure water side of the reverse osmosis membrane module during the preset evaluation stage after this start-up, and determine the predicted characteristic value of the pure water side effluent TDS during the preset evaluation stage after this start-up from the effluent water quality prediction result as the first water quality prediction value; S3, using the first water quality prediction value and the preset first cup water target TDS threshold as input parameters, a fuzzy control algorithm is used to generate the pre-rinse control strategy for this start-up, the pre-rinse control strategy including at least the pre-rinse duration; S4, when the water purifier enters the water production start-up stage, the water pump and the valve are controlled according to the pre-rinse control strategy to make the water purifier run pre-rinse through the pre-rinse passage. After the pre-rinse is completed, the valve is switched to the water supply passage to drive the water pump to supply water to the outlet side.

[0007] This invention records the start time of shutdown when the water purifier switches from water production to shutdown. Before restarting, it combines the shutdown duration, current raw water TDS, water temperature, and reverse osmosis membrane module operating parameters to form operating condition data. Using a pre-established digital twin model of the water purifier, it predicts the change in TDS of the pure water effluent during a preset evaluation period. The obtained first water quality prediction value and the preset first cup water target TDS threshold are input into a fuzzy control algorithm to adaptively generate the pre-rinse duration for this startup and control the water pump and valves to perform pre-rinse and switch water supply accordingly. This makes the pre-rinse process match different shutdown histories and water quality conditions. Compared with the control method of installing a TDS sensor at the end of the pure water side and relying on real-time detection results to wait for the value to drop before ending the pre-rinse, it can avoid the waste of water resources and extra waiting time caused by continuous discharge to observe changes in detection values. At the same time, it reduces the insufficient or excessive pre-rinse caused by fixed pre-rinse duration or only classifying according to shutdown time, and improves the consistency of the first cup water TDS control under different operating conditions.

[0008] According to an embodiment of the present invention, in step S2, the digital twin model discretizes the preset evaluation stage at preset time intervals based on the current downtime and the operating condition data, and calculates the predicted TDS value of the pure water effluent from the reverse osmosis membrane module at each discrete time point. The effluent water quality prediction result includes the predicted TDS value of the pure water effluent at each discrete time point, and the first water quality prediction value is an effluent TDS prediction curve composed of the predicted TDS values ​​of the pure water effluent at each discrete time point in chronological order.

[0009] By discretizing the preset evaluation stages over time, the predicted TDS values ​​of the pure water effluent at each discrete moment are obtained, and these values ​​are organized into an effluent TDS prediction curve. This makes the water quality evolution process from the initial runoff to the effluent entering a stable phase visible on the timeline. In engineering practice, this curve can be correlated with the time window or water volume window corresponding to the first sample, and the TDS change trend and peak position within the window can be compared to determine whether pre-rinsing is necessary and to what extent it is needed to meet the water quality requirements of the first sample. Compared to estimating the overall water quality based solely on a single prediction point or empirical formula, this prediction curve-based approach can distinguish the differences in water quality during the first sample stage under different shutdown histories and operating conditions, providing a more detailed and evidence-based quantitative evaluation foundation for setting the subsequent pre-rinsing duration and intensity.

[0010] According to one embodiment of the present invention, the operating parameters of the reverse osmosis membrane module include a model parameter reflecting the type and desalination capacity of the reverse osmosis membrane module, and a membrane cavity residual salt state quantity R used to characterize the residual salt state in the membrane cavity of the reverse osmosis membrane module; The residual salt state quantity R in the membrane cavity is determined according to the start-stop cycle during the multiple start-stop operation of the pure water machine. The operation data of the start-stop cycle includes at least one or more of the following: raw water TDS, water production volume, pre-rinse volume, and shutdown duration. The residual salt state quantity R in the membrane cavity is recursively updated based on the operation data of adjacent start-stop cycles to reflect the residual salt state of the membrane cavity corresponding to each start-stop cycle.

[0011] By using the model parameter reflecting membrane type and desalination capacity, along with the membrane residual salt state quantity R characterizing the residual salt state within the membrane chamber, as input operating parameters, the digital twin model can distinguish between different types and desalination levels of reverse osmosis membrane modules when making water quality predictions. It can also incorporate "residual salt memory" across start-up and shutdown cycles. Specifically, R is recursively updated based on operating data such as raw water TDS, produced water volume, pre-flushing volume, and shutdown duration according to the start-up and shutdown cycles. This is equivalent to numerically characterizing the process of salt being carried out of the membrane chamber and accumulating during the shutdown period, rather than relying solely on the single dimension of the current shutdown duration to indirectly infer the residual water situation. Through this recursive approach, the model can gradually converge to a state quantity that matches the actual residual salt level in the membrane chamber over long-term operation. This allows for different prediction results for the same shutdown duration in two scenarios: a "membrane chamber that has just been fully flushed" and a "membrane chamber with high residual salt accumulation from multiple short flushes." This provides a basis for differentiated pre-flushing strategies for different historical operating conditions.

[0012] According to an embodiment of the present invention, the pre-rinsing control strategy generated in step S3 includes a pre-rinsing duration parameter and a pre-rinsing pressure parameter. The pre-rinsing duration parameter includes a reference pre-rinsing duration determined based on the first water quality prediction value under a preset reference pre-rinsing pressure and a target pre-rinsing duration. The pre-rinsing pressure parameter includes the reference pre-rinsing pressure and a first pre-rinsing pressure and a second pre-rinsing pressure for segmented control, wherein the first pre-rinsing pressure is greater than the second pre-rinsing pressure. When the reference pre-rinse duration is not greater than the preset duration threshold, the target pre-rinse duration is set as the reference pre-rinse duration. In step S4, the water pump is controlled to run pre-rinse according to the reference pre-rinse pressure within the target pre-rinse duration. When the reference pre-rinse duration is greater than the preset duration threshold, in step S3, the target pre-rinse duration is set to be less than the reference pre-rinse duration, and the target pre-rinse duration is divided into a pre-rinse start period and a pre-rinse follow-up period. The first pre-rinse pressure is used during the pre-rinse start period, and the second pre-rinse pressure is used during the pre-rinse follow-up period. This is to ensure that the TDS of the first cup of water after this start meets the target TDS threshold while shortening the pre-rinse duration. In step S4, the water pump is controlled to run the pre-rinse operation according to the target pre-rinse duration, the first pre-rinse pressure, and the second pre-rinse pressure.

[0013] By simultaneously providing pre-rinse duration and pre-rinse pressure parameters in the fuzzy control output, this scheme does not simply provide a fixed duration. Instead, it first obtains the reference pre-rinse duration required to meet the water quality requirements of the first cup of water under a preset reference pre-rinse pressure. Then, based on the relationship between the reference duration and the preset duration threshold, it determines whether to operate in a single segment according to the reference pressure or to adopt a "time-reduction and segmented pressure" approach. When the reference pre-rinse duration is short, the wastewater replacement can be completed directly using the reference pressure and reference duration without additional pressure adjustment, making control simple and the water volume predictable. When the reference pre-rinse duration is long, the target pre-rinse duration is divided into an initial period and a subsequent period. The initial period is increased to the first pre-rinse pressure to enhance the flushing efficiency of residual salt and high TDS wastewater in the membrane cavity. The subsequent period is reduced to the second pre-rinse pressure, so that the effluent conditions gradually approach normal water production. Under the premise of shortening the overall pre-rinse time, the effluent TDS corresponding to the first cup of water stage still falls within the target threshold. In this way, on the one hand, the high-pressure section is used to complete the rapid replacement of the "dirty water section" in a concentrated manner, reducing the user's perceived waiting time. On the other hand, the low-pressure section is used for a smooth transition, avoiding the energy consumption and additional impact on the membrane module caused by long-term high-pressure operation, and achieving a synergistic balance between the pre-rinsing time, rinsing intensity and the quality of the first cup of water.

[0014] According to one embodiment of the present invention, during the long-term operation of the water purifier, the user's water usage time, duration of a single water usage, and number of water usages at different times are recorded, and the user's water usage behavior characteristics are determined based on the recorded data; based on the user's water usage behavior characteristics, a time window for water usage requests within a preset time range is predicted; when the probability of a water usage request within a certain time window is predicted to be greater than a preset probability threshold and no water usage request is currently detected, the water purifier is placed in advance into the water production start-up phase at the pre-start time before the start of the time window, and steps S1 to S4 are executed.

[0015] By continuously recording water usage times, durations of individual water usage sessions, and frequency of water usage at different times during long-term operation, user behavior characteristics can be statistically extracted, such as high-frequency water usage periods in the morning and evening, differences in water usage between weekends and weekdays, and typical single water usage durations. Based on this, the time windows for potential water requests within a preset time range can be predicted. When the probability of water usage within a certain time window exceeds a preset threshold and no actual water request has been detected yet, the water purifier is put into the current water production start-up phase and pre-rinsing is completed. This is equivalent to moving the pre-rinsing process, which should normally be performed when the user uses water, to before the user is likely to use water. In this way, when the user makes a water request within the predicted time window, they can directly obtain water that has already undergone pre-rinsing, significantly shortening the perceived waiting time. At the same time, it avoids the disconnect between actual water demand and scheduled flushing, thus reducing unnecessary empty flushing and water waste.

[0016] According to one embodiment of the present invention, after the water purifier is placed in the current water production start-up phase and pre-rinsing is completed at the pre-start time before the start of the time window predicted according to the user's water use behavior characteristics, when a user's water use request is detected within the time window, the water purifier is controlled to directly supply water to the outlet side through the water supply passage without performing pre-rinsing again. At the end of the current water production start-up phase, the corresponding water use duration and water intake volume are recorded. The recorded water use duration and water intake volume are used as the operating data of the corresponding start-up and shutdown cycle to update the membrane cavity residual salt state quantity R in the operating parameters of the reverse osmosis membrane module.

[0017] With this setting, when pre-flushing is completed before the predicted time window and water is actually used within the window, the water used is directly output through the water supply channel without repeating the pre-flushing, avoiding the extra drainage and waiting caused by secondary flushing within the same start-stop cycle. At the same time, the duration and volume of water used are used as the operating data for this start-stop cycle to update the residual salt state quantity R in the membrane cavity. This makes the recursive process of R reflect not only "how much was pre-flushed" but also "how much was actually taken away". Thus, even with complex usage habits such as pre-start and intermittent water intake, it can still accurately depict the real changes in residual salt in the membrane cavity, providing a more reliable state basis for subsequent water quality prediction and pre-flushing decisions based on R.

[0018] According to one embodiment of the present invention, during the pre-rinsing process, the actual effluent TDS of the pre-rinsing discharge water is collected by a TDS sensor installed on the pre-rinsing discharge path on the pure water side. The actual effluent TDS is compared with the corresponding predicted value in the effluent water quality prediction result in step S2 to obtain the prediction error. Based on the prediction error, at least one of the model parameters of the digital twin model and the update coefficient of the membrane cavity residual salt state quantity R is adjusted online to improve the prediction accuracy of the digital twin model for the change of TDS of the pure water side effluent of the reverse osmosis membrane module under the pre-rinsing condition.

[0019] In this way, during the pre-rinsing process, the actual effluent TDS on the pure water side pre-rinsing discharge path is selected as the observation, and compared with the corresponding predicted value obtained in step S2 point by point. This can reflect the deviation of the digital twin model and the membrane cavity residual salt state quantity R in the current water quality and membrane state in real time. By feeding back this prediction error to adjust the model parameters and the update coefficient of R online, the model has the ability to self-correct changes in operating conditions such as membrane fouling, membrane aging, and seasonal fluctuations in raw water TDS. This avoids the prediction curve from gradually deviating from the actual effluent TDS after long-term operation. Thus, in subsequent start-up and shutdown cycles, the first water quality prediction value can still be reliably calculated based on the updated model and R, providing a stable and reliable prediction basis for pre-rinsing control.

[0020] According to an embodiment of the present invention, a short stop duration threshold is preset in the method. When the current stop duration determined in step S1 is less than the short stop duration threshold, the pre-rinsing duration of the current start is set to zero in step S3, and in step S4, the water purifier is controlled to directly switch the valve to the water supply passage without pre-rinsing during the current water production start-up phase.

[0021] By introducing a short-stop duration threshold, the pre-rinsing requirement is distinguished from conditions where the downtime is extremely short and the change in membrane cavity water status is negligible. When the downtime is less than the threshold, the pre-rinsing duration is set to zero and the system switches to the water supply path. This avoids unnecessary drainage and additional waiting time caused by mechanical pre-rinsing in short-stop scenarios. At the same time, the prediction-based pre-rinsing process is only triggered in conditions where the downtime is long and the risk of water accumulation is high. This makes the pre-rinsing control more closely reflect actual usage scenarios while ensuring the quality of the first water flow.

[0022] This invention also provides a pre-rinsing control system for a pure water machine, applicable to a tankless reverse osmosis pure water machine that connects the raw water side and the outlet water side, for executing the above-described method. The pure water machine includes a reverse osmosis membrane assembly, a water pump, and valves for switching between the pre-rinsing passage and the water supply passage, all connected in series between the raw water side and the outlet water side. The system includes: The raw water quality testing unit is used to detect the current TDS and temperature of the raw water. The time acquisition unit is used to acquire the time information when the water purifier switches from water production operation to shutdown state, and to determine the duration of this shutdown. The memory is used to store the digital twin model of the water purifier, the target TDS threshold of the first cup of water, and data related to the operating parameters of the reverse osmosis membrane module; The controller is signal-connected to the raw water quality detection unit, the time acquisition unit, the memory, the water pump, and the valve. The controller is configured as follows: When the pure water machine switches from water production operation to shutdown state, the shutdown start time is recorded based on the time acquisition unit. Before restarting, the shutdown duration is determined based on the time acquisition unit. The current raw water TDS and water temperature are obtained from the raw water quality detection unit, and the operating parameters of the reverse osmosis membrane module are read from the memory to form the operating condition data for this startup. The downtime and operating data are input into the digital twin model to obtain the effluent water quality prediction result, which characterizes the change of TDS in the pure water side of the reverse osmosis membrane module during the preset evaluation phase after the start-up. The first water quality prediction value after the start-up is determined from the effluent water quality prediction result. Based on the first water quality prediction value and the first cup water target TDS threshold, a pre-rinse control strategy for this activation is generated, and the pre-rinse control strategy includes at least the pre-rinse duration. During the water production start-up phase, the water pump and valves are controlled according to the pre-rinse control strategy, so that the pure water machine runs pre-rinse through the pre-rinse passage. After the pre-rinse is completed, the valves are controlled to switch to the water supply passage, and the water pump is driven to supply water to the outlet side.

[0023] Through the coordination of the aforementioned hardware units and control logic, the water purifier can automatically acquire operating data, predict effluent water quality, and execute corresponding pre-rinsing controls, ensuring that the method has a clear physical basis and engineering feasibility. This system architecture, based on traditional tankless reverse osmosis water purifiers, mainly adds detection and control modules, facilitating integration or upgrades on existing product platforms.

[0024] The present invention also provides a tankless reverse osmosis water purifier, including a raw water side and an outlet water side. A reverse osmosis membrane assembly, a water pump, and a valve for switching between a pre-rinsing passage and a water supply passage are sequentially arranged between the raw water side and the outlet water side, so that the water purifier discharges pre-rinsing water through the pre-rinsing passage in the pre-rinsing condition and supplies water to the outlet water side through the water supply passage in the water purification condition. It also includes the aforementioned pre-rinse control system for the pure water machine, wherein the controller of the pre-rinse control system is electrically connected to the water pump and the valve.

[0025] (III) Beneficial Effects of the Invention: Before restarting, the present invention uses a digital twin model of the pure water system, constructed based on the current shutdown duration, current raw water TDS, water temperature, and reverse osmosis membrane module operating parameters, to predict the changes in TDS of the pure water effluent during a preset evaluation phase. The obtained first water quality prediction value and the preset first cup water target TDS threshold are used as inputs to a fuzzy control algorithm to adaptively generate the pre-rinsing duration for this startup and control the pre-rinsing process accordingly. This invention can specifically determine whether to perform pre-rinsing and the pre-rinsing duration under different shutdown histories and water quality conditions, avoiding excessively long or short pre-rinsing times, improving the consistency of first cup water TDS control, and reducing water waste and unnecessary user waiting time caused by ineffective pre-rinsing. Attached Figure Description

[0026] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0027] Figure 1 This is a schematic flowchart of a pre-rinsing control method for a pure water machine provided in one embodiment of the present invention; Figure 2 This is a schematic diagram of a first water quality prediction process provided in one embodiment of the present invention; Figure 3 This is a schematic diagram of the pre-rinse control parameter generation process provided in one embodiment of the present invention. Detailed Implementation

[0028] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. 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. Specific implementation examples: This embodiment provides a method, system, and water purifier for pre-rinsing control of a water purifier.

[0030] In a specific application scenario, this embodiment describes a tankless reverse osmosis water purifier. The water purifier includes a reverse osmosis membrane assembly, a water pump, and valves for switching between a pre-rinsing passage and a water supply passage, as well as a pre-rinsing control system. It includes a raw water side and an outlet water side, with the reverse osmosis membrane assembly, water pump, and valves for switching between the pre-rinsing passage and the water supply passage sequentially arranged between the raw water side and the outlet water side. This allows the water purifier to discharge pre-rinsing water through the pre-rinsing passage during pre-rinsing operation and to switch between the pre-rinsing and water supply passages during water production operation. Water is supplied to the outlet side through the aforementioned water supply passage. The raw water side is usually connected to the tap water pipe through the inlet connector. The water pump is installed before the raw water enters the reverse osmosis membrane module to provide a stable inlet water pressure to the reverse osmosis membrane module under pre-rinsing and water production conditions. The pure water outlet and concentrated water outlet of the reverse osmosis membrane module are connected to the water supply passage and the pre-rinsing passage through internal pipelines, respectively. Valves are installed at the junction of the above pipelines to selectively guide the pure water flow to the drain end or the outlet end under the action of control signals. In this embodiment, conventional inlet, outlet, and drain interfaces can be arranged inside the water purifier casing. The pre-rinsing passage is preferentially connected to the drain, and the water supply passage is preferentially connected to the user's faucet or water outlet. This allows the pre-rinsing stage to prioritize the discharge of the stagnant water with high TDS in the membrane chamber and pure water pipeline through the pre-rinsing passage. After switching to water purification mode, the permeate generated by the reverse osmosis membrane module is output along the water supply passage for the user to drink. This provides the basic flow path switching conditions for the execution of the subsequent pre-rinsing control method from the overall machine structure.

[0031] The pre-rinse control system of the pure water machine includes: The raw water quality testing unit is used to detect the current raw water TDS and water temperature. The real-time clock module provides time information to determine the duration of the downtime after the water purifier switches from water purification to shutdown. The memory is used to store the digital twin model of the water purifier, fuzzy control rules, the target TDS threshold of the first cup of water, and data related to the operating parameters of the reverse osmosis membrane component; The pure water side TDS detection unit is installed on the pure water side pre-rinse discharge path of the pure water machine body, and is used to collect the actual effluent TDS of the pre-rinse discharge water during the pre-rinse process; And a controller electrically connected to the raw water quality detection unit, real-time clock module, memory, pure water side TDS detection unit, water pump and valves.

[0032] In this embodiment, the pre-rinse control system of the pure water machine is preferably arranged entirely within the control cavity inside the pure water machine casing. The controller, memory, and real-time clock module can be integrated on the same control circuit board. The raw water quality detection unit and the pure water side TDS detection unit are electrically connected to the control circuit board via cables. The raw water quality detection unit can be a combination of a conductivity sensor and a temperature sensor installed on the raw water inlet pipe. Its installation position is preferably set in the straight pipe section before the raw water enters the water pump or reverse osmosis membrane module, so as to obtain TDS and water temperature signals representing the current tap water quality, which facilitates the modification of existing pure water machines. The real-time clock module can be a real-time clock chip with a backup power supply, or it can be implemented by the controller's internal time base in conjunction with a power-off retention circuit, so as to maintain time continuity even when the pure water machine is powered off or in standby mode. The memory can be an on-chip non-volatile memory unit of the controller, or it can be an external flash memory or EEPROM, used to store digital twin model parameters, fuzzy control rules, the target TDS threshold of the first cup of water, and the reverse osmosis membrane module operating parameters corresponding to each start-stop cycle for a long time. The pure water side TDS detection unit is preferably installed on a section of the pipeline near the drain outlet in the pure water side pre-rinse discharge path. A small online TDS sensor can be used, with its probe inserted into the pure water discharge branch and connected to the controller via a cable to collect the TDS of the pure water discharged water in real time during the pre-rinse stage. The controller can be a single-chip microcomputer, microcontroller, or a dedicated control chip with corresponding computing capabilities. Its input terminals are connected to the detection unit, real-time clock module, and memory signal, and its output terminals are connected to the water pump drive circuit and valve drive circuit, respectively, to send start / stop and pressure regulation control signals to the water pump, and on / off and path switching control signals to the valves.

[0033] The controller is configured as follows: When the pure water machine switches from water production to shutdown, the shutdown start time is recorded based on the real-time clock module. Before restarting, the shutdown duration is determined according to the real-time clock module, and the current raw water TDS and water temperature are obtained from the raw water quality detection unit. At the same time, the operating parameters of the reverse osmosis membrane module are read from the memory to form the operating condition data for this startup. The downtime and operating data are input into the pre-established digital twin model of the pure water machine in the memory. Based on the digital twin model, the water quality prediction result is obtained to characterize the change of TDS in the pure water side of the reverse osmosis membrane module during the preset evaluation stage after this start-up. The predicted characteristic value of the pure water side TDS in the preset evaluation stage after this start-up is determined from the water quality prediction result and used as the first water quality prediction value. Using the first water quality prediction value and the first cup water target TDS threshold as input, the fuzzy control algorithm and fuzzy rule base stored in the memory are called to generate the pre-rinse control strategy for this start. The pre-rinse control strategy includes at least the pre-rinse duration. During the water production start-up phase, the water pump and valves are controlled according to the pre-rinse control strategy, so that the pure water machine runs pre-rinse through the pre-rinse passage. After the pre-rinse is completed, the valves are controlled to switch to the water supply passage, and the water pump is driven to supply water to the outlet side. Furthermore, during the pre-rinsing process, based on the prediction error between the actual effluent TDS of the pre-rinsing discharge water collected by the pure water side TDS detection unit and the corresponding predicted value in the effluent water quality prediction result, the model parameters of the digital twin model and / or the update coefficients related to the operating parameters of the reverse osmosis membrane module are adjusted online to improve the prediction accuracy of the digital twin model for changes in pure water side effluent TDS.

[0034] In practical implementation, the controller records the timestamp provided by the real-time clock module along with the operating status of the pure water machine. This is used to accurately mark the start time of shutdown when switching to the shutdown state and to calculate the shutdown duration before restarting. Simultaneously, it periodically collects the TDS and water temperature signals output by the raw water quality detection unit. Combined with operating parameters such as the reverse osmosis membrane module model parameters and membrane cavity residual salt state quantity R stored in the memory, it constructs the operating condition data vector corresponding to this startup. This operating condition data is input into the digital twin model software module stored in the memory. The controller executes the corresponding calculation program to obtain the predicted curve of the effluent TDS within the preset evaluation stage and the corresponding first water quality prediction value. Subsequently, the controller uses the first water quality prediction value and the target TDS threshold of the first cup of water as input to the fuzzy control algorithm. It calls the fuzzy rule base and membership function stored in the memory to calculate the control quantities regarding the pre-rinse duration and, if necessary, the pre-rinse pressure allocation in the pre-rinse control strategy for this startup. Based on this, it outputs control signals to the water pump and valves during the pre-rinse stage to realize the opening of the pre-rinse passage and the pre-rinse operation.

[0035] During the pre-rinse process, the controller can also periodically read the actual effluent TDS output from the pure water side TDS detection unit and compare it with the predicted TDS given by the digital twin model at the corresponding time points according to the preset sampling interval, thereby calculating the prediction error sequence. Based on these errors, the controller can make minor adjustments to the model parameters in the memory and the update coefficients related to the operating parameters of the reverse osmosis membrane module according to the preset self-calibration strategy. This makes the digital twin model more closely reflect the current installation environment, raw water quality, and membrane operating status in subsequent start-up and shutdown cycles, thereby improving the reliability of the first water quality prediction value and providing a more stable prediction basis for subsequent pre-rinse control.

[0036] In an optional implementation, to achieve the aforementioned pre-rinsing control on the existing tankless reverse osmosis water purifier hardware, the method described in this embodiment can be executed by the aforementioned controller as a software program. The controller can have a built-in non-volatile storage area for continuously recording the water purifier's operating status and water usage events by date and time period, and periodically calling the calculation program during standby or idle periods to perform statistical analysis and predictive calculations on historical data. Water usage events can be detected using conventional methods such as flow sensors, flow switches, or outlet pressure differential detection, so as to obtain the water usage time point and duration without changing the basic structure of the pure water flow path.

[0037] The method includes: During the long-term operation of the water purifier, the user's water usage time, duration of each water usage, and number of water usages at different times are recorded. Based on the recorded data, the user's water usage behavior characteristics are determined. Based on the user's water usage behavior characteristics, a time window for water usage requests within a preset time range is predicted. When the probability of a water usage request within a certain time window is predicted to be greater than a preset probability threshold and no water usage request is currently detected, the water purifier is placed in the water production start-up phase in advance at the pre-start time before the start of the time window, and steps S1 to S4 are executed.

[0038] In practical implementation, the controller can divide water usage times within a certain period into preset time slots, such as by hour, half-hour, or other suitable time granularity to statistically analyze the frequency of water usage throughout the day, and further analyze the distribution of the duration of a single water usage within each time slot. Based on these statistical results, the controller can calculate a value representing the "probability of a water request occurring within that time slot" for each time slot, which is used as part of the user's water usage behavior characteristics. Then, within each rolling preset time range (e.g., several minutes or several time slots in the future), the controller determines the time window where a water request may occur based on the water usage probability of each time slot within that time range and the corresponding weight. When the overall probability of a certain window is greater than a preset probability threshold and the current flow detection result shows no water request, the controller can trigger a pre-start operation at a certain time point before the start of that window, placing the water purifier in the current water production start-up phase, and performing pre-rinsing and water supply control according to steps S1 to S4 described later.

[0039] Before the start of the time window predicted based on the user's water usage behavior characteristics, the water purifier is placed in the water production start-up phase and pre-rinsing is completed. When a user's water usage request is detected within the time window, the water purifier is controlled to directly supply water to the outlet side through the water supply path without performing pre-rinsing again. At the end of the water production start-up phase, the corresponding water usage duration and water intake volume are recorded. The recorded water usage duration and water intake volume are used as the operating data for the corresponding start-up and shutdown cycle to update the membrane cavity residual salt state quantity R in the operating parameters of the reverse osmosis membrane module.

[0040] During this process, after the pre-rinse is completed during pre-start, the controller can keep the water purifier in a short-term standby state. For example, it can keep the valves in the water supply path connected and maintain a certain low flow circulation, or keep the system in a standby state that can be quickly started without affecting water quality. When the flow detection unit or the outlet switch detects a user's water request within the predicted time window, the controller will no longer trigger the pre-rinse logic, but will directly allow water to be supplied to the user through the water supply path. Subsequently, at the end of this water production start-up phase, the outlet water volume and water usage duration are counted. These data, along with the pre-rinse volume and shutdown duration of this cycle, are stored in the memory for the next update of the membrane cavity residual salt state quantity R, to reflect the amount of salt removed through water production and intake during this cycle.

[0041] During a normal start-stop cycle, the controller performs pre-flushing control according to steps S1 to S4.

[0042] S1. When the pure water machine switches from water production operation to shutdown state, record the shutdown start time. Before restarting, determine the shutdown duration according to the real-time clock, and collect the current raw water TDS, water temperature and the operating parameters of the reverse osmosis membrane module to form the operating condition data for this startup. In the specific control logic, when the controller detects the end of a user's water usage request and determines that the water purifier has switched from water purification to shutdown, it writes the timestamp provided by the real-time clock module at this time as the shutdown start time into the memory, recording it as the starting point of a new start-stop cycle. The next time a water usage request is detected or the system triggers a pre-start, the controller calculates the shutdown duration by reading the difference between the current time and the shutdown start time. The current raw water TDS can be collected in real-time by the raw water quality detection unit, and the water temperature can be provided by the same sensor or a separate temperature sensor. The operating parameters of the reverse osmosis membrane module can be directly read from the memory, including static parameters such as the membrane module model and nominal desalination rate written during factory configuration, as well as dynamic parameters that are continuously updated during multiple start-stop cycles.

[0043] The operating parameters of the reverse osmosis membrane module include model parameters reflecting the type and desalination capacity of the reverse osmosis membrane module, and membrane cavity residual salt state quantity R used to characterize the residual salt state in the membrane cavity of the reverse osmosis membrane module. The membrane cavity residual salt state quantity R is determined according to the start-stop cycle during the multiple start-stop operation of the pure water machine. The operating data of the start-stop cycle includes at least one or more of the following: raw water TDS, water production volume, pre-rinse volume, and shutdown duration. The membrane cavity residual salt state quantity R is recursively updated based on the operating data of adjacent start-stop cycles to reflect the membrane cavity residual salt state corresponding to each start-stop cycle.

[0044] In one specific implementation, the residual salinity state quantity R in the membrane cavity can be defined as a dimensionless state variable related to the amount of salt within the membrane housing, such as representing the proportion of residual salt relative to a nominal clean state. At the end of each start-up / shutdown cycle, the controller updates R using a preset recursive formula based on the raw water TDS, cumulative produced water volume, pre-flushing discharge volume, and shutdown duration for that cycle. The recursive formula comprehensively considers the "flushing effect" (decreasing R) on membrane cavity salinity during the water production and pre-flushing stages, as well as the permeate backflow effect (increasing R) from the raw water side to the pure water side during shutdown, thus numerically simulating the net impact of each cycle on the residual salinity level in the membrane cavity. Those skilled in the art can set the coefficients of the above recursive relationship through experimental calibration or empirical formulas, based on the type and volume of the membrane module and typical water quality conditions, so that R remains within a preset value range during long-term operation and corresponds to the actual trend of residual salinity in the membrane cavity.

[0045] S2, input the current shutdown duration and the operating condition data into a pre-established digital twin model of the pure water machine, obtain the effluent water quality prediction result based on the digital twin model to characterize the change of TDS in the pure water side of the reverse osmosis membrane module during the preset evaluation stage after the current startup, and determine the predicted characteristic value of the pure water side effluent TDS during the preset evaluation stage after the current startup from the effluent water quality prediction result as the first water quality prediction value; the digital twin model discretizes the preset evaluation stage according to the current shutdown duration and the operating condition data at preset time intervals, calculates the predicted value of the pure water side effluent TDS of the reverse osmosis membrane module at each discrete time point, the effluent water quality prediction result includes the pure water side effluent TDS prediction value at each discrete time point, and the first water quality prediction value is the effluent TDS prediction curve composed of the pure water side effluent TDS prediction values ​​at each discrete time point in chronological order.

[0046] In constructing a digital twin model, the process of water quality change during the startup phase of a pure water system can be abstracted into a holistic model composed of multiple coupled sub-models. Typically, a digital twin model can include a membrane performance sub-model reflecting the hydraulic flux and desalination performance of the reverse osmosis membrane, a hydraulic volume sub-model characterizing the effective volume of the raw water and pure water side pipelines and the membrane housing, and a salt mass balance sub-model describing the migration and mixing process of salts on both sides of the membrane during shutdown and startup phases. The membrane performance sub-model can be based on reverse osmosis theory or empirical formulas, taking into input parameters such as raw water TDS, water temperature, transmembrane pressure difference, membrane age, and R, and outputting the instantaneous flow rate and desalination rate of permeate water at different operating stages; the hydraulic volume sub-model can divide the raw water side, pure water side, and pure water pipeline within the membrane housing into several volume units according to their spatial location, recording the TDS state of each unit at the end of the previous start-up and shutdown cycle; the salt mass balance sub-model updates the salt concentration of each volume unit at each discrete time step based on the permeate flux, flushing flow rate, and distribution volume.

[0047] When executing step S2, the controller can first divide the evaluation phase into several discrete time points based on the preset evaluation phase length (e.g., covering the time period from startup to the end of the first cup of water consumption) and the preset time interval. Then, at each discrete time point, the digital twin model is invoked, and operating condition data such as the current shutdown duration, current raw water TDS, water temperature, membrane module model parameters, and residual salt state quantity R are input. The predicted value of the pure water effluent TDS at the corresponding time point is calculated iteratively. The predicted values ​​of all discrete time points are combined in chronological order to form the effluent TDS prediction curve, which serves as part of the effluent water quality prediction result. Those skilled in the art can select different preset evaluation phase lengths and discrete time intervals according to specific product requirements to balance computational complexity and prediction accuracy.

[0048] When determining the first water quality prediction value, the controller can extract features representing the water quality risk of the first cup of water from the effluent TDS prediction curve according to design requirements. For example, the weighted average, peak, or other statistical quantities of the predicted TDS within the time interval corresponding to the standard first cup of water volume can be selected as the prediction feature value, thereby realizing the mapping from the entire curve to a single feature value, so that subsequent fuzzy control algorithms can process it.

[0049] S3, using the first water quality prediction value and the preset first cup water target TDS threshold as input parameters, a fuzzy control algorithm is used to generate the pre-rinse control strategy for this activation. The pre-rinse control strategy includes at least a pre-rinse duration. The pre-rinse control strategy generated in step S3 includes a pre-rinse duration parameter and a pre-rinse pressure parameter. The pre-rinse duration parameter includes a reference pre-rinse duration determined based on the first water quality prediction value at a preset reference pre-rinse pressure and a target pre-rinse duration. The pre-rinse pressure parameter includes the reference pre-rinse pressure and a first pre-rinse pressure and a second pre-rinse pressure used for segmented control. The first pre-rinse pressure is greater than the second pre-rinse pressure. When the reference pre-rinse duration is not greater than the preset duration threshold, the target pre-rinse duration is set as the reference pre-rinse duration. In step S4, the water pump is controlled to run pre-rinse at the reference pre-rinse pressure within the target pre-rinse duration. When the reference pre-rinse duration is longer than the preset duration threshold, in step S3, the target pre-rinse duration is set to be shorter than the reference pre-rinse duration. The target pre-rinse duration is divided into a pre-rinse start period and a pre-rinse follow-up period. The first pre-rinse pressure is used during the pre-rinse start period, and the second pre-rinse pressure is used during the pre-rinse follow-up period. This is to shorten the pre-rinse duration while ensuring that the TDS of the first cup of water after this start meets the target TDS threshold of the first cup of water. In step S4, the water pump is controlled to run pre-rinse according to the target pre-rinse duration, the first pre-rinse pressure, and the second pre-rinse pressure.

[0050] In practical implementation, the fuzzy control algorithm can be implemented by the controller's software program. The controller can first construct an input linguistic variable representing the "risk level of the first cup of water" based on the relative relationship between the first predicted water quality value and the target TDS threshold of the first cup of water. For example, the deviation between the predicted value and the threshold can be divided into multiple membership intervals, corresponding to fuzzy levels such as "low risk," "medium risk," and "high risk." Depending on the actual product requirements, quantities such as downtime and the TDS of the raw water during this downtime can be used as auxiliary inputs after simple normalization, forming a combination of quantities related to water quality risk. When designing the fuzzy control rules, a set of rules can be pre-set. For example, when the first predicted water quality value is significantly lower than the target threshold and the downtime is short, the rule outputs a lower pre-rinse requirement level, corresponding to a shorter reference pre-rinse duration; when the first predicted water quality value is close to or exceeds the target threshold and the downtime is long, the rule outputs a higher pre-rinse requirement level, corresponding to a longer reference pre-rinse duration.

[0051] When executing the fuzzy control algorithm, the controller can calculate the membership degrees of the aforementioned input variables, call the fuzzy rule base for inference, and use common defuzzification methods such as the centroid method to obtain a continuous reference pre-rinse duration. Subsequently, based on the comparison between this reference pre-rinse duration and a preset duration threshold, the target pre-rinse duration and pressure allocation strategy are determined: when the reference pre-rinse duration is not greater than the preset duration threshold, the pre-rinse length is considered appropriate, and the target pre-rinse duration is directly set as the reference pre-rinse duration, maintaining the pump at the reference pre-rinse pressure; when the reference pre-rinse duration is too long, the target pre-rinse duration is divided into two stages: a start stage and a follow-up stage. In the start stage, the pump output pressure is increased to the first pre-rinse pressure to enhance the flushing speed of residual salt in the membrane cavity per unit time; in the follow-up stage, the pump output pressure is reduced to the second pre-rinse pressure, gradually transitioning the effluent to normal water production. This two-stage pressure control can be achieved by adjusting the pump speed, pulse width modulation duty cycle, or the opening of the driving valves. Those skilled in the art can select the appropriate execution method based on the specific pump type and system structure.

[0052] S4, when the water purifier enters the water production start-up stage, the water pump and the valve are controlled according to the pre-rinse control strategy to make the water purifier run pre-rinse through the pre-rinse passage. After the pre-rinse is completed, the valve is switched to the water supply passage to drive the water pump to supply water to the outlet side.

[0053] In step S4, the controller first sets the running time and pressure level of the pre-rinse stage according to the pre-rinse control strategy, sends a start signal to the water pump drive circuit and adjusts the drive parameters to make the water pump output reach the required pre-rinse pressure. Simultaneously, the controller controls the valve drive circuit to switch the valve to the pre-rinse passage connection state, allowing the pure water side outlet water to be discharged through the pre-rinse discharge passage. When the timer reaches the target pre-rinse duration, or when the pre-rinse stage is determined to be over, the controller sends a switching command to the valve to switch the flow path to the water supply passage. Then, the water pump drive parameters are adjusted as needed to put the pure water machine into normal water production and supply operation, continuously supplying permeate water to the outlet side. Those skilled in the art can further ensure the stability of the water output at the user end by combining this with actual product configurations, such as setting flow limiters and pressure regulating valves in the water supply passage. In this way, the entire method can be fully executed with the support of the aforementioned hardware system, realizing pre-rinse management based on digital twin prediction and fuzzy control.

[0054] In one alternative implementation, the digital twin model of the pure water machine can be constructed using a discrete-time salt mass balance model. The reverse osmosis membrane module and its pure water side piping can be divided into several series-connected equivalent volume units during a preset evaluation phase. The pure water side salt concentration of each unit at the k-th discrete time step is denoted as C_p(k), the corresponding equivalent salt concentration on the raw water side is denoted as C_f(k), and the residual salt state quantity in the membrane chamber is denoted as R(k). Within each discrete time step Δt, the controller calculates the instantaneous desalination rate η(k) and permeate flux J(k) of the membrane module based on the current feed water flow rate Q_f(k), pure water side flow rate Q_p(k), raw water TDS, and water temperature. This can be expressed in a form similar to C_p(k) = [1−η(k)]·C_f(k) + K_R·R(k), where K_R is a coefficient related to the membrane housing volume and structure, used to reflect the flow of residual salt from the membrane chamber to the permeate. The influence of water dilution; while C_f(k) and R(k) can be recursively updated in the form of C_f(k+1)=C_f(k)+Δt·f_1(Q_f(k),C_in(k),C_f(k)) and R(k+1)=R(k)+Δt·f_2(Q_f(k),Q_p(k),R(k),τ_off). Functions f_1 and f_2 are used to simulate the influent dilution, water production flushing, and shutdown re-infiltration process. Those skilled in the art can collect effluent TDS data through experiments under different raw water TDS, temperature, and membrane aging conditions, identify the parameters in η(k), K_R, and f_1 and f_2, and store the finally determined parameter table or interpolation relationship in memory. When the controller executes step S2, it only needs to start from the current shutdown duration, the operating condition data collected before the current startup, and the initial values ​​of C_f and R at the end of the previous start-stop cycle, and execute the above recursive formula in a discrete time step to obtain C_p(k) at each time point of the preset evaluation stage in sequence, thereby forming the pure water side effluent TDS prediction curve in the effluent water quality prediction result. The weighted average or maximum value of C_p(k) within the time window corresponding to the first cup of water is used as the first water quality prediction value.

[0055] In another alternative implementation, the digital twin model can also be implemented using an empirically parameterized discrete-time linear model. For example, the first water quality prediction value C_pred can be obtained by the following formula: C_pred=a_0+a_1·C_raw+a_2·T_raw+a_3·R_last+a_4·τ_off, where C_raw is the current raw water TDS, T_raw is the current water temperature, R_last is the residual salt state of the membrane cavity at the end of the previous start-up and shutdown cycle, τ_off is the current shutdown duration, and a_0~a_4 are coefficients obtained by regression from historical operating data. To obtain the prediction curve in the effluent water quality prediction results, a recursive relationship C_out(k+1)=b_0+b_1·C_out(k)+b_2·C_raw+b_3·R_last+b_4·τ_off can be established at time step k during the preset evaluation stage. The initial time C_out(0) is set as the estimated TDS value of the pure water side effluent at the end of the previous start-stop cycle, and C_out(k) is obtained iteratively. Each time point C_out(k) is used as the predicted TDS value of the pure water side effluent. The coefficients a_i and b_i can be obtained through extensive prototype testing and data fitting during the product development stage and are pre-installed in the factory firmware. Those skilled in the art can choose either a physical mechanism-driven model or an empirical regression model as the specific implementation method of the digital twin model. Both can meet the functional requirement of this invention to "provide the effluent TDS prediction curve and the first water quality prediction value within the preset evaluation stage".

[0056] In terms of fuzzy control algorithms, one specific implementation can select the deviation between the first predicted water quality value and the target TDS threshold of the first cup of water, and the current downtime, as the two input variables for fuzzy control. The controller can first calculate the normalized deviation e = (C_pred − C_tar) / C_tar, where C_pred is the first predicted water quality value and C_tar is the target TDS threshold of the first cup of water. Then, the current downtime τ_off is normalized to d according to preset upper and lower limits. Then, several triangular or trapezoidal membership functions are defined for e and d respectively. For example, e can be divided into three fuzzy subsets: "too low", "close to", and "too high", and d can be divided into three fuzzy subsets: "short down", "medium down", and "long down". The pre-rinse time requirement can be defined as the output fuzzy variable L, corresponding to the subsets such as "short time", "medium time", and "long time". A set of fuzzy control rules is pre-defined based on product requirements. For example, when e is "too high" and d is "long stop", L is "long time"; when e is "close" and d is "medium stop", L is "medium time"; when e is "too low" and d is "short stop", L is "short time", etc. During operation, the controller infers from the rule base based on the membership degrees of e and d, and uses common defuzzification methods such as the centroid method to obtain a continuous reference pre-flushing duration T_ref. Multiplying this by a pre-defined time scale coefficient yields the physical reference pre-flushing duration. This is then compared with a preset duration threshold, and combined with the first pre-flushing pressure and the second pre-flushing pressure, a target pre-flushing duration and segmented pressure setpoints are generated.

[0057] In another optional fuzzy control implementation, the residual salinity state variable R in the membrane cavity can be directly used as a third input variable in the decision-making process, jointly determining the pre-flushing level with the first water quality prediction value and the downtime. In this case, the controller can define subsets of R such as "low residual salinity," "medium residual salinity," and "high residual salinity," and expand the fuzzy rules. For example, when e is "close," d is "medium downtime," and R is "high residual salinity," the controller still outputs a pre-flushing level of "medium-to-long duration" to compensate for the risks that may be caused by prediction deviations in the short term. Those skilled in the art can select a fuzzy control structure with two or three input variables based on the computational resources and control accuracy requirements of the target product, and through experimental tuning of the membership function and rule table, obtain pre-flushing time and pressure parameters that meet the usage requirements within the framework provided by this invention, thereby ensuring the feasibility of the technical feature of "generating a pre-flushing control strategy using a fuzzy control algorithm" in this invention.

[0058] In practical implementation, the controller can periodically read the TDS sensor output set on the pre-rinse discharge path on the pure water side at sampling intervals that are the same as or integer multiples of the preset evaluation stage. The actual effluent TDS measured at the k-th sampling time is recorded as C_meas(k), and the predicted value C_pred(k) at the corresponding time is retrieved from the effluent water quality prediction result obtained in step S2. The prediction error ε(k) = C_meas(k) − C_pred(k) is calculated. Subsequently, the controller can make minor adjustments to the update coefficients of a certain model parameter p and / or membrane cavity residual salt state quantity R of the digital twin model using a simple recursive correction method. For example, it can update the model parameters and update coefficients in the form of p_new = p_old + K_p·ε(k) or R_coef,new = R_coef,old + K_R·ε(k), where K_p and K_R are small correction coefficients that are pre-tuned to control the magnitude of online adjustment, so that the model parameters and update coefficients gradually converge in the direction that can better fit the actual effluent TDS changes during multiple pre-rinse processes. Those skilled in the art can select to perform the above-mentioned online fine-tuning on a single key parameter or a small number of parameters according to the computing resources and stability requirements of the specific product, thereby realizing the adaptive correction of the digital twin model under the pre-rinsing condition.

[0059] In practical implementation, the short-stop duration threshold can be stored in the controller or memory. After executing step S1, the controller calculates the current shutdown duration based on the real-time clock and compares it with the short-stop duration threshold. If the current shutdown duration is determined to be less than the threshold, the pre-flushing duration field is directly set to zero in the pre-flushing control strategy, and the pre-flushing path operation stage is skipped when executing step S4. The control valves are directly switched from the shutdown state to the water supply path, and the water pump supplies water to the outlet side according to the water production conditions. The short-stop duration threshold can be determined through experimental calibration based on factors such as typical raw water TDS, effective membrane cavity volume, and salt reabsorption rate during shutdown. This avoids pre-flushing in scenarios with short shutdown times and minimal deterioration of the membrane cavity water condition, reducing unnecessary drainage and waiting. Meanwhile, pre-flushing control based on prediction is still performed only in scenarios with long shutdown times and a high risk of water deterioration.

[0060] The above are preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made to the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A pre-rinsing control method for a pure water machine, applied to a tankless reverse osmosis pure water machine that connects the raw water side and the outlet water side, the pure water machine comprising a reverse osmosis membrane assembly, a water pump, and a valve for switching between a pre-rinsing passage and a water supply passage, characterized in that, The method includes: S1. When the pure water machine switches from water production operation to shutdown state, record the shutdown start time. Before restarting, determine the shutdown duration according to the real-time clock, and collect the current raw water TDS, water temperature and the operating parameters of the reverse osmosis membrane module to form the operating condition data for this startup. S2, input the downtime and the operating condition data into the pre-established digital twin model of the pure water machine, obtain the effluent water quality prediction result based on the digital twin model to characterize the change of TDS in the pure water side of the reverse osmosis membrane module during the preset evaluation stage after this start-up, and determine the predicted characteristic value of the pure water side effluent TDS during the preset evaluation stage after this start-up from the effluent water quality prediction result as the first water quality prediction value; S3, using the first water quality prediction value and the preset first cup water target TDS threshold as input parameters, a fuzzy control algorithm is used to generate the pre-rinse control strategy for this start-up, the pre-rinse control strategy including at least the pre-rinse duration; S4, when the water purifier enters the water production start-up stage, the water pump and the valve are controlled according to the pre-rinse control strategy to make the water purifier run pre-rinse through the pre-rinse passage. After the pre-rinse is completed, the valve is switched to the water supply passage to drive the water pump to supply water to the outlet side.

2. The pre-rinsing control method for a pure water machine according to claim 1, characterized in that, In step S2, the digital twin model discretizes the preset evaluation stage at preset time intervals based on the current downtime and the operating condition data, and calculates the predicted TDS value of the pure water effluent from the reverse osmosis membrane module at each discrete time point. The effluent water quality prediction result includes the predicted TDS value of the pure water effluent at each discrete time point, and the first water quality prediction value is the effluent TDS prediction curve composed of the predicted TDS values ​​of the pure water effluent at each discrete time point in chronological order.

3. The pre-rinsing control method for a pure water machine according to claim 2, characterized in that, The operating parameters of the reverse osmosis membrane module include model parameters reflecting the type and desalination capacity of the reverse osmosis membrane module, and membrane cavity residual salt state quantity R used to characterize the residual salt state in the membrane cavity of the reverse osmosis membrane module. The residual salt state quantity R in the membrane cavity is determined according to the start-stop cycle during the multiple start-stop operation of the pure water machine. The operation data of the start-stop cycle includes at least one or more of the following: raw water TDS, water production volume, pre-rinse volume, and shutdown duration. The residual salt state quantity R in the membrane cavity is recursively updated based on the operation data of adjacent start-stop cycles to reflect the residual salt state of the membrane cavity corresponding to each start-stop cycle.

4. The pre-rinsing control method for a pure water machine according to claim 3, characterized in that, The pre-rinsing control strategy generated in step S3 includes a pre-rinsing duration parameter and a pre-rinsing pressure parameter. The pre-rinsing duration parameter includes a reference pre-rinsing duration and a target pre-rinsing duration determined based on the first water quality prediction value under a preset reference pre-rinsing pressure. The pre-rinsing pressure parameter includes the reference pre-rinsing pressure and a first pre-rinsing pressure and a second pre-rinsing pressure used for segmented control. The first pre-rinsing pressure is greater than the second pre-rinsing pressure. When the reference pre-rinse duration is not greater than the preset duration threshold, the target pre-rinse duration is set as the reference pre-rinse duration. In step S4, the water pump is controlled to run pre-rinse according to the reference pre-rinse pressure within the target pre-rinse duration. When the reference pre-rinse duration is greater than the preset duration threshold, in step S3, the target pre-rinse duration is set to be less than the reference pre-rinse duration, and the target pre-rinse duration is divided into a pre-rinse start period and a pre-rinse follow-up period. The first pre-rinse pressure is used during the pre-rinse start period, and the second pre-rinse pressure is used during the pre-rinse follow-up period. This is to ensure that the TDS of the first cup of water after this start meets the target TDS threshold while shortening the pre-rinse duration. In step S4, the water pump is controlled to run the pre-rinse operation according to the target pre-rinse duration, the first pre-rinse pressure, and the second pre-rinse pressure.

5. The pre-rinsing control method for a pure water machine according to claim 3, characterized in that, During the long-term operation of the water purifier, the user's water usage time, duration of each water usage, and number of water usages at different times are recorded. Based on the recorded data, the user's water usage behavior characteristics are determined. Based on the user's water usage behavior characteristics, a time window for water usage requests within a preset time range is predicted. When the probability of a water usage request within a certain time window is predicted to be greater than a preset probability threshold and no water usage request is currently detected, the water purifier is placed in the water production start-up phase in advance at the pre-start time before the start of the time window, and steps S1 to S4 are executed.

6. The pre-rinsing control method for a pure water machine according to claim 5, characterized in that, Before the start of the time window predicted based on the user's water usage behavior characteristics, the water purifier is placed in the water production start-up phase and pre-rinsing is completed. When a user's water usage request is detected within the time window, the water purifier is controlled to directly supply water to the outlet side through the water supply path without performing pre-rinsing again. At the end of the water production start-up phase, the corresponding water usage duration and water intake volume are recorded. The recorded water usage duration and water intake volume are used as the operating data for the corresponding start-up and shutdown cycle to update the membrane cavity residual salt state quantity R in the operating parameters of the reverse osmosis membrane module.

7. The pre-rinsing control method for a pure water machine according to claim 3, characterized in that, During the pre-rinsing process, the actual effluent TDS of the pre-rinsing water is collected by a TDS sensor installed on the pre-rinsing discharge path on the pure water side. The actual effluent TDS is compared with the corresponding predicted value in the effluent water quality prediction result in step S2 to obtain the prediction error. Based on the prediction error, at least one of the model parameters of the digital twin model and the update coefficient of the membrane cavity residual salt state quantity R is adjusted online to improve the prediction accuracy of the digital twin model for the change of TDS of the pure water side effluent of the reverse osmosis membrane module under pre-rinsing conditions.

8. The method for controlling the pre-rinse of a pure water machine according to any one of claims 1 to 7, characterized in that, In the method, a short stop duration threshold is preset. When the current stop duration determined in step S1 is less than the short stop duration threshold, the pre-rinse duration of the current start is set to zero in step S3, and in step S4, the water purifier is controlled to directly switch the valve to the water supply path without pre-rinse during the current water production start-up phase.

9. A pre-rinsing control system for a pure water machine, characterized in that, A tankless reverse osmosis pure water machine is applicable to connecting the raw water side and the outlet water side. The pure water machine includes a reverse osmosis membrane assembly, a water pump, and a valve for switching between the pre-rinse passage and the water supply passage, which are connected in series between the raw water side and the outlet water side. The system includes: The raw water quality testing unit is used to detect the current TDS and temperature of the raw water. The time acquisition unit is used to acquire the time information when the water purifier switches from water production operation to shutdown state, and to determine the duration of this shutdown. The memory is used to store the digital twin model of the water purifier, the target TDS threshold of the first cup of water, and data related to the operating parameters of the reverse osmosis membrane module; The controller is signal-connected to the raw water quality detection unit, the time acquisition unit, the memory, the water pump, and the valve. The controller is configured as follows: When the pure water machine switches from water production operation to shutdown state, the shutdown start time is recorded based on the time acquisition unit. Before restarting, the shutdown duration is determined based on the time acquisition unit. The current raw water TDS and water temperature are obtained from the raw water quality detection unit, and the operating parameters of the reverse osmosis membrane module are read from the memory to form the operating condition data for this startup. The downtime and operating data are input into the digital twin model to obtain the effluent water quality prediction result, which characterizes the change of TDS in the pure water side of the reverse osmosis membrane module during the preset evaluation phase after the start-up. The first water quality prediction value after the start-up is determined from the effluent water quality prediction result. Based on the first water quality prediction value and the first cup water target TDS threshold, a pre-rinse control strategy for this activation is generated, and the pre-rinse control strategy includes at least the pre-rinse duration. During the water production start-up phase, the water pump and valves are controlled according to the pre-rinse control strategy, so that the pure water machine runs pre-rinse through the pre-rinse passage. After the pre-rinse is completed, the valves are controlled to switch to the water supply passage, and the water pump is driven to supply water to the outlet side.

10. A tankless reverse osmosis pure water machine, characterized in that, It includes a raw water side and an outlet water side. A reverse osmosis membrane module, a water pump, and a valve for switching between a pre-rinsing passage and a water supply passage are sequentially provided between the raw water side and the outlet water side, so that the pure water machine discharges pre-rinsing water through the pre-rinsing passage in the pre-rinsing condition and supplies water to the outlet water side through the water supply passage in the water production condition. It also includes the pre-rinsing control system for the pure water machine as described in claim 9, wherein the controller of the pre-rinsing control system for the pure water machine is electrically connected to the water pump and the valve.

Images