A canister-type direct drinking water filtration system

Abstract

The application discloses a tank type direct drinking water filtering system, and relates to the direct drinking water filtering field.The tank type direct drinking water filtering system comprises a raw water tank, a pretreatment filter, a digital high-pressure pump, a nanofiltration membrane assembly, an ozone ultraviolet sterilization device, a digital water tank, a user direct drinking faucet, a backwater filter and a backwater adjusting mechanism arranged on a backwater pipeline.A fluid dynamic pressure balance control center is integrated in the digital high-pressure pump.The original pressure data from the pump inlet is extracted to represent the characteristics of the backwater and water game state, a fluid dynamics fingerprint is constructed, the fluid dynamics fingerprint is decoded to quantize the physical invasion degree of the backwater jet, a back pressure resistance index is obtained, a control law optimization calculation is performed on the back pressure resistance index, a pump speed dynamic compensation vector and a backwater damping adjustment coefficient are calculated, and the pump speed dynamic compensation vector and the backwater damping adjustment coefficient are converted into the entity action of the fluid machine, so that the pipeline howling and the flow field separation risk caused by the water hammer are completely eliminated.

Description

Technical Field

[0001] This invention relates to the field of direct drinking water filtration, specifically a tank-type direct drinking water filtration system. Background Technology

[0002] In existing tank-type direct drinking water filtration systems, a closed-loop design is typically used to maintain the freshness of water at the end of the pipe network and recover residual pressure. This involves returning unused drinking water from the user directly to the suction inlet of the digital high-pressure pump via a return water filter, where it merges with pre-treated raw water at a T-junction before the pump. This design works adequately under full-load water production conditions, but a fluid dynamics flaw emerges when the system is in a keep-alive cycle mode—that is, when there is no water demand at the user end and only the pipe network water quality is maintained. Because the return water pipeline is a closed flow path directly driven by the residual pressure of the high-pressure pump, the return water stream often retains a high residual dynamic pressure and velocity when it reaches the pump inlet confluence node. At this point, the raw water, after multi-stage pretreatment, is typically in a low-velocity laminar flow state, resulting in a mismatch in the physical momentum of the two fluids.

[0003] The high-energy return water jet forms a physical jet barrier upon entering the pump inlet, significantly hindering the low-energy intake water with back pressure. This dynamic pressure conflict prevents the intake check valve from maintaining a stable open state. Instead, it is forced into high-frequency mechanical flutter under the intermittent support of the return water jet, causing periodic blocking and shut-off of the intake channel. Existing constant-pressure variable frequency control strategies typically adjust only based on the pump outlet pressure, failing to detect this microsecond-level dynamic pressure game and flow field separation at the pump inlet. Long-term operation can lead to micro-cavitation inside the pump chamber and accelerate fatigue failure of the intake check valve assembly.

[0004] To address the aforementioned shortcomings, a canister-type direct drinking water filtration system is provided. Summary of the Invention

[0005] To address the technical problems mentioned in the background section, the present invention provides a tank-type direct drinking water filtration system.

[0006] A tank-type direct drinking water filtration system includes a raw water tank, a pretreatment filter, a digital high-pressure pump, a nanofiltration membrane module, an ozone ultraviolet sterilization device, a digital water tank, a user direct drinking faucet, a return water filter, and a return water regulating mechanism installed on the return water pipeline. The digital high-pressure pump integrates a fluid dynamic pressure balance control center, which includes:

[0007] The fluid pulsation sensing module is used to extract features representing the game state between return water and inlet water from the raw pressure data at the pump inlet and construct a fluid dynamic fingerprint map.

[0008] The jet countermeasure analysis module is used to decode the fluid dynamic fingerprint spectrum to quantify the physical intrusion degree of the return water jet and obtain the back pressure resistance index.

[0009] The game-theoretic optimization decision module is used to perform control law optimization calculations on the back pressure stagnation index and solve for the pump speed dynamic compensation vector and the return water damping adjustment coefficient.

[0010] The flexible coupling drive module is used to convert the pump speed dynamic compensation vector and the return water damping adjustment coefficient into the physical action of the fluid machinery, and perform closed-loop self-healing correction based on the execution effect.

[0011] Furthermore, the step of constructing the fluid dynamics fingerprint spectrum includes:

[0012] The collected instantaneous pressure waveform data is mapped to a one-dimensional equivalent pipeline space domain to construct a staged spatial envelope map containing the main channel, the confluence shear and the impact restricted area. Boolean clipping and consistency check based on physical envelope are performed on the staged spatial envelope map, and the valve disc and jet impact event sequence and rigid impact envelope segment are output.

[0013] The physical quantities of the rigid impact envelope segment are integrated to calculate the pressure-time integral area of ​​a single impact event within the fluid micro-element, and the return water pulse momentum value is obtained in combination with the pipeline cross-sectional parameters.

[0014] Furthermore, the step of constructing the fluid dynamics fingerprint map also includes:

[0015] The distribution density of the valve disc and jet impact event sequence per unit time is statistically analyzed to generate the fluid intermittent stagnation duty cycle. The return water pulse momentum value and the fluid intermittent stagnation duty cycle are used as the vertical and horizontal axes, respectively, to construct a fluid dynamic fingerprint spectrum under multi-source confluence state.

[0016] Furthermore, the steps for constructing a phased spatial envelope diagram that includes the main flow corridor, confluence shear, and impact restricted areas are as follows:

[0017] Acquire instantaneous pressure waveform data at the inlet of a digital high-pressure pump, construct a one-dimensional equivalent pipeline coordinate system based on the length of the physical pipeline and the propagation speed of the pressure wave, project the time dimension features of the instantaneous pressure waveform data into the physical position features of the pipeline, and output a pressure echo spatial mapping map.

[0018] The system receives the pressure echo spatial mapping map, divides the mainstream steady-state corridor region, the confluence shear envelope region, and the valve seat impact prohibition region according to the physical structure of the pump inlet confluence node, assigns corresponding geometric boundary constraints to each region, and outputs a staged spatial envelope map with physical criterion boundaries.

[0019] Furthermore, the steps for the output valve disc and the jet impact event sequence and rigid impact envelope segment are as follows:

[0020] Spatial Boolean operation is performed on the staged spatial envelope map to extract waveform segments that fall into the valve seat impact forbidden zone and simultaneously fall into the bus shear envelope zone. Background noise interference is removed using a sudden change amplitude threshold, and a geometric boundary candidate impact set is output.

[0021] Furthermore, the steps of the output valve disc and the jet impact event sequence and rigid impact envelope segment also include:

[0022] The acoustic reflection path is verified on the geometric boundary candidate impact set to verify whether the round-trip propagation delay of the candidate impact at the pipeline physical location conforms to the physical law, eliminate artifact interference, and output the valve disc and jet impact event sequence and rigid impact envelope segment.

[0023] Furthermore, the step of obtaining the back pressure resistance index includes:

[0024] The vertical axis features of the fluid dynamic fingerprint spectrum are analyzed, and the equivalent jet core velocity at the outlet of the return water pipeline is calculated in reverse from the momentum value of the return water pulse based on the inverse transformation of the momentum law.

[0025] The horizontal axis features of the fluid dynamic fingerprint spectrum are analyzed, and the theoretical geometric flow section of the raw water inlet is corrected by the fluid intermittent retardation duty cycle. The dynamic effective influent flux under high-frequency interference of the return water jet is then calculated.

[0026] Furthermore, the step of obtaining the back pressure resistance index also includes:

[0027] The momentum ratio of the equivalent jet core velocity to the main flow velocity of the raw water is calculated, and the momentum ratio is weighted and coupled with the dynamic effective influent flux to output the back pressure retardation index.

[0028] Furthermore, the steps for calculating the pump speed dynamic compensation vector and the return water damping adjustment coefficient are as follows:

[0029] A non-cooperative game model of hydrodynamic pressure between the inlet and return water flow paths is constructed. The back pressure resistance index is used as the conflict penalty term in the game, and the Nash equilibrium point is calculated with the goal of minimizing the total turbulent kinetic energy at the pump inlet.

[0030] The numerical coordinates of the Nash equilibrium point are compared and verified with the preset hardware security threshold range to determine the optimal operating point.

[0031] Based on the deviation between the current operating conditions and the optimal operating point, a dynamic compensation vector for pump speed and a return water damping adjustment coefficient are generated.

[0032] Furthermore, the steps for performing closed-loop self-healing correction based on the execution effect are as follows:

[0033] Signal modulation and command mapping are performed on the pump speed dynamic compensation vector and the return water damping adjustment coefficient respectively to generate an anti-phase torque pulsation signal to drive the digital high-pressure pump motor, and to generate a virtual valve opening adjustment command to drive the return water adjustment mechanism.

[0034] The pump inlet pressure residual after physical action is monitored in real time. When the residual exceeds the preset quiescent threshold, the correction parameters are iteratively calculated based on the PID self-tuning algorithm, and the anti-phase torque pulsation signal and the virtual valve opening adjustment command are dynamically updated until the fluid pulsation characteristics disappear.

[0035] Furthermore, the digital high-pressure pump adopts an electromechanical integrated structure, and the fluid dynamic pressure balance control center is integrated into the motor drive end of the digital high-pressure pump, uploading the operating data to the cloud platform through the Internet of Things module; the fluid dynamic pressure balance control center also includes a user water consumption prediction module, which is based on a user behavior learning algorithm and automatically predicts the user's water consumption for the next period based on historical data fed back from the cloud platform.

[0036] Furthermore, the digital water tank is a fully enclosed pressure tank structure, with a flexible diaphragm or air bladder inside to physically isolate the stored purified water from the outside air; the return pipeline where the return water filter is located uses the outlet residual pressure generated by the digital high-pressure pump during water production or circulation to drive the water flow to complete the circulation and keep it alive.

[0037] Furthermore, the system is configured with a pump-tank joint scheduling mode: when the user's drinking faucet is in peak water usage, the digital water tank releases the internally stored pressurized water and operates in parallel with the digital high-pressure pump to jointly supply water to the user.

[0038] Compared with the prior art, the beneficial effects of the present invention are:

[0039] This invention constructs a control architecture based on a fluid dynamic pressure non-cooperative game model, overcoming the limitation of traditional constant pressure control in handling momentum conflicts at multi-source confluence nodes. Through quantitative analysis of fluid dynamic fingerprinting and back pressure resistance index, the system can accurately identify the jet barrier strength formed by the return water jet at the pump inlet and use the Nash equilibrium algorithm to calculate the strategy combination that minimizes the total kinetic energy of the turbulent flow at the pump suction inlet. This mechanism eliminates the backwater effect of high-kinetic-energy return water on low-kinetic-energy inlet water from a physical source, avoids the inlet check valve from falling into high-frequency mechanical flutter due to dynamic pressure imbalance, and achieves orderly fusion and smooth transition of multiple fluid streams at the confluence node.

[0040] This invention transforms the pump speed dynamic compensation vector and the return water damping adjustment coefficient into physical actions of fluid machinery, and performs closed-loop self-healing correction based on the execution effect. It employs a flexible coupling drive technology based on phase lead offsetting, solving the control synchronization problem caused by fluid machinery response lag. The system does not rely on passive physical buffer containers, but instead generates a pump speed dynamic compensation vector containing microsecond-level phase prediction. This drives the motor to produce an anti-phase torque pulsation the instant the return water pulse arrives, actively counteracting water hammer energy through dynamic pressure offsetting. Combined with virtual impedance shaping of the return water flow path, the system maintains the pipeline's liveness circulation while completely eliminating the risks of pipeline whistling and flow field separation, extending the fatigue life of the valve seat sealing components. Attached Figure Description

[0041] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. The following drawings are not drawn to scale according to the actual size, but are intended to show the main idea of ​​the present invention.

[0042] Figure 1 This is a physical architecture diagram of a tank-type direct drinking water filtration system.

[0043] Figure 2 A schematic diagram of the equivalent spatial coordinate mapping and pressure echo spatial mapping process;

[0044] Figure 3 A schematic diagram of the phased spatial envelope construction and physical criterion partitioning;

[0045] Figure 4 This is a system block diagram of the fluid dynamic pressure balance control center.

[0046] Reference numerals: 1. Pretreatment filter; 2. Digital high-pressure pump; 3. Nanofiltration membrane module; 4. Ozone ultraviolet sterilization device; 5. Digital water tank; 6. User drinking water faucet; 7. Return water filter; 8. Raw water tank. Detailed Implementation

[0047] 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 also within the scope of protection of the present invention.

[0048] like Figure 1As shown, this embodiment provides a tank-type direct drinking water filtration system. Its physical architecture mainly includes: a raw water tank 8, a pretreatment filter 1, a digital high-pressure pump 2, a nanofiltration membrane module 3, an ozone ultraviolet sterilization device 4, a digital water tank 5, a user direct drinking faucet 6, a return water filter 7, and a return water regulating mechanism installed on the return water pipeline. The return water regulating mechanism is installed on the return water pipeline section between the return water filter 7 and the inlet of the digital high-pressure pump 2. It employs a proportional solenoid valve to receive valve opening adjustment commands output by the fluid dynamic pressure balance control center. By changing the flow cross-sectional area of ​​the return water pipeline, it adjusts the fluid resistance of the return water pipeline, thereby achieving continuous and adjustable control of the return water flow rate and velocity.

[0049] The system's fluid circulation path is described as follows: Tap water, as the main water source, first enters the raw water tank 8 for initial pressure stabilization and buffering; then, it passes through the pretreatment filter 1 to remove large particulate impurities and suspended solids; the pretreated water flows into the suction port of the digital high-pressure pump 2, where it is pressurized to the working pressure, such as 0.6MPa-1.0MPa; the high-pressure water flows into the nanofiltration membrane module 3 for core deep purification, removing heavy metals, viruses, and organic pollutants while retaining beneficial minerals; the produced purified water undergoes double disinfection at the end of the process via an ozone and ultraviolet sterilization device 4 to ensure safety in microbiological indicators; finally, high-quality drinking water is stored in the digital water tank 5, ready to respond to the user's drinking water tap 6 at any time. To ensure the freshness of the water at the end of the pipe network, the drinking water that is not consumed by the user is used as return water. After passing through the return water filter 7 to remove secondary impurities in the return pipeline, it flows back to the inlet of the digital high-pressure pump 2 for recirculation;

[0050] The digital high-pressure pump adopts an electromechanical integrated structure. The fluid dynamic pressure balance control center is integrated into the motor drive end of the digital high-pressure pump and uploads the operating data to the cloud platform through the Internet of Things module. The fluid dynamic pressure balance control center also includes a user water consumption prediction module. The user water consumption prediction module is based on a user behavior learning algorithm and automatically predicts the user's water consumption for the next period based on historical data fed back from the cloud platform.

[0051] Specifically, in terms of physical construction, the digital high-pressure pump adopts a compact mechatronics design. The internal space of the junction box or heat sink at the rear of the motor is specifically designed to accommodate a highly integrated variable frequency drive circuit board. This circuit board serves as the physical hardware carrier of the fluid dynamic pressure balance control center, meaning that the various logic modules of the fluid dynamic pressure balance control center share the same microprocessor hardware as the underlying drive circuit of the motor. To enable remote data interaction, this physical carrier also integrates an IoT communication module, such as a 4G / 5G or Wi-Fi module. This module is responsible for protocol encapsulating the pressure, flow, speed, and power data collected in real time by the control center and establishing an encrypted channel to upload the data to a remote cloud server in real time. The cloud server acts as a data warehouse, storing the device's historical operating data indexed by timestamps.

[0052] In terms of functional logic, to achieve intelligent energy saving and response control, in addition to integrating the core dynamic pressure balance modules of this invention, the fluid dynamic pressure balance control center also includes a user water consumption prediction module. This module is a logic unit based on embedded code, which runs a user behavior learning algorithm. The specific workflow of the user water consumption prediction module is as follows: the system discretizes a 24-hour day into several fixed-length statistical time periods, such as 15 minutes or 1 hour as a time window. Before the start of each time period, the user water consumption prediction module retrieves the flow characteristic data of the device in the same historical time period from the cloud server through the Internet of Things interface, such as the same time window in the past two weeks. The user water consumption prediction module processes the data using a weighted moving average statistical algorithm. This algorithm assigns higher weight to recent data to reflect short-term trends and lower weight to long-term average data to reflect long-term habits, thereby calculating the corrected user water consumption prediction value for the next time period, which can automatically predict user water consumption.

[0053] The following provides a detailed description of the specific hardware architecture of the fluid dynamic pressure balance control center, its hardware integration with the digital high-pressure pump, data interaction path, and control signal transmission mechanism.

[0054] I. Overall Hardware Architecture:

[0055] As mentioned earlier, the physical hardware carrier of the fluid dynamic pressure balance control center is a highly integrated variable frequency drive circuit board installed in the junction box or heat sink space at the rear of the motor. This circuit board is a multi-layer printed circuit board that integrates the following core hardware components:

[0056] (1) Main control microprocessor: An industrial-grade digital signal processor chip is used. This microprocessor integrates a floating-point arithmetic unit, a multi-channel high-precision analog-to-digital converter, a multi-channel enhanced pulse width modulation peripheral, and a digital-to-analog converter peripheral. As mentioned above, the various logic modules of the fluid dynamic pressure balance control center share the same set of microprocessor hardware with the underlying drive circuit of the motor. That is, the main control microprocessor also serves as the unified computing hardware carrier for the fluid pulsation sensing module, the jet countermeasure analysis module, the game optimization decision module, and the flexible coupling drive module. The above functional modules are stored in the on-chip Flash memory of the microprocessor in the form of embedded firmware programs and run on its CPU core.

[0057] (2) Analog signal conditioning circuit: including a pre-amplifier, a low-pass anti-aliasing filter, and a voltage protection clamping circuit. The pre-amplifier is used to receive weak analog voltage signals or 4-20mA current loop signals from pressure and flow sensors and amplify them to the full-scale input range of the on-chip analog-to-digital converter of the microprocessor; the cutoff frequency of the low-pass anti-aliasing filter is set to less than half of the sampling rate of the analog-to-digital converter to filter out high-frequency noise before analog-to-digital conversion and prevent spectral aliasing; the voltage protection clamping circuit consists of a bidirectional TVS diode and a current-limiting resistor to protect the input terminal of the analog-to-digital converter from damage caused by field surges or electrostatic discharge.

[0058] (3) Power drive stage circuit: This includes two independent channels: a three-phase inverter bridge drive circuit and a valve drive circuit. The three-phase inverter bridge drive circuit consists of a gate driver and six IGBT or MOSFET power switches. The gate driver receives six complementary PWM logic signals from the microprocessor's enhanced pulse width modulation peripheral output. After optocoupler isolation and level conversion, the signals drive the power switches to turn on and off, thereby inverting the DC bus voltage into a three-phase AC voltage to supply the motor windings of the digital high-voltage pump. The valve drive circuit consists of a PWM or DAC output connected to a voltage-to-current converter chip. This chip converts the digital or PWM signal output by the microprocessor into a 4-20mA standard industrial current signal, which is used to drive the return water regulating mechanism on the return water pipeline, such as a proportional solenoid valve.

[0059] (4) Power Management Module: Converts electrical energy from the DC bus inside the pump body into multiple independent regulated power rails, including 3.3V and 1.2V low-voltage DC power supplies for the microprocessor and digital circuits, isolated and regulated analog power supplies for the analog signal conditioning circuits, and gate drive power supplies for the power drive stage. Each power rail is electrically isolated from the others by magnetic or optical isolation to prevent power stage switching noise from coupling to the signal stage circuits.

[0060] (5) IoT Communication Module: As mentioned above, the physical carrier integrates an IoT communication module, such as a 4G / 5G or Wi-Fi module. This communication module is an independent communication daughter card soldered or plugged into the circuit board via pin headers. It has a built-in TCP / IP protocol stack and supports 4G / 5G LTE or Wi-Fi wireless communication protocols. This module interacts with the main control microprocessor through a UART or SPI serial bus interface. It is responsible for encapsulating the running data frames packaged by the microprocessor into protocols, establishing an encrypted channel to upload the data to the remote cloud server in real time, and receiving configuration parameters or historical data query results from the cloud.

[0061] II. Hardware integration method with digital high-pressure pumps:

[0062] The variable frequency drive circuit board is installed entirely within the junction box or heat sink at the rear end of the digital high-pressure pump motor. This space consists of a cast aluminum housing and a sealed end cap. The outer wall of the housing is integrally formed with the motor housing or connected via flange bolts, achieving an IP65 protection rating. The circuit board is fixed to the mounting boss inside the cavity by copper pillars at intervals, maintaining a certain air gap between it and the motor stator core to meet heat dissipation and electromagnetic compatibility requirements.

[0063] Electrical connections between the circuit board and external sensors and actuators are achieved via aviation plugs or industrial-grade waterproof connectors mounted on the housing wall. Specific connection ports include: a pressure sensor input port for connecting a pressure sensor installed at the inlet of the digital high-pressure pump, transmitting analog voltage signals or 4-20mA current signals; a flow sensor input port for connecting an electromagnetic or ultrasonic flow meter installed at the inlet of the raw water tank or upstream of the digital high-pressure pump, transmitting pulse frequency signals or 4-20mA current signals; a motor power output port for connecting to the three-phase winding terminals of the motor via three power cables, transmitting AC drive power after three-phase inversion; a valve control output port for connecting to the coil terminals of the return water regulating mechanism (i.e., the proportional solenoid valve) on the return water pipeline via a two-core shielded cable, transmitting 4-20mA control current signals or PWM signals; a DC bus input port for connecting to a power supply; and an IoT antenna port for connecting an external 4G / 5G or Wi-Fi antenna.

[0064] The power drive stage circuit on the circuit board is electrically isolated from the microprocessor digital logic circuit through a high-speed optocoupler isolation device. This ensures that the high-voltage, high-current power stage switching noise will not be coupled to the low-voltage signal stage circuit through a common ground path, thus guaranteeing the signal integrity when the microprocessor performs high-precision analog-to-digital conversion sampling and floating-point operations.

[0065] III. Complete Data Interaction Path:

[0066] The system's data interaction path forms a complete closed-loop link from the physical sensing end through the digital computing end to the physical execution end, as detailed below:

[0067] Forward sensing data path: A pressure sensor installed at the pump inlet manifold converts the instantaneous fluid pressure into a 4-20mA current signal or analog voltage signal, which is transmitted via cable to an aviation connector on the junction box housing. The signal then enters the analog signal conditioning circuit on the circuit board, undergoing differential amplification, low-pass anti-aliasing filtering, and clamping protection processing before being input to the on-chip analog-to-digital converter (ADC) module of the microprocessor for analog-to-digital conversion. The continuous analog quantity is discretized into a digital quantity at a sampling rate of at least 10kHz. The microprocessor CPU core reads the digital quantity from the ADC and sends it to the fluid pulsation sensing module firmware program for waveform discretization, momentum flux quantization, and spatiotemporal fingerprint construction calculations. The calculation result, i.e., the coordinate data of the fluid dynamics fingerprint spectrum, is transmitted to the jet countermeasure analysis module firmware program for jet velocity inversion, effective flux calculation, and hindrance index output calculation. The calculation result, i.e., the back pressure hindrance index value, is transmitted to the game optimization decision-making module firmware program for Nash equilibrium solution, safety boundary verification, and compensation vector generation calculations. Data transfer between the above firmware modules is achieved in the microprocessor's on-chip RAM through a shared data buffer, without the need for an external bus.

[0068] Meanwhile, flow sensors such as electromagnetic flow meters or ultrasonic flow meters installed at the inlet of the raw water tank or the inlet of the digital high-pressure pump convert the fluid velocity into a pulse frequency signal or a 4-20mA current signal, which is transmitted to the junction box via cable, enters the signal conditioning circuit, and is input to the microprocessor on-chip analog-to-digital converter or timer to capture peripheral devices for digital conversion, which is then read and used by the effective flux calculation unit and the hindrance index output unit in the jet countermeasure analysis module.

[0069] Reverse control data path - pump speed control channel: The game-theoretic optimization decision module outputs a dynamic compensation vector for pump speed, including the amplitude component ΔN and the phase lead angle. The firmware of the co-execution unit of the flexible coupling drive module receives the vector parameter and performs calculations using the space vector pulse width modulation (SVPWM) algorithm integrated within the microprocessor CPU. The calculated duty cycle parameter is written into the period register and comparison register of the enhanced pulse width modulation peripheral on the microprocessor chip. The enhanced pulse width modulation peripheral automatically generates six complementary PWM logic square wave signals with dead time according to the register configuration. These six PWM signals are electrically isolated and level-converted by optocoupler isolation devices on the circuit board and sent to the gate driver chip. The gate driver outputs a high-drive-capability gate pulse to drive the six IGBT or MOSFET power switches to alternately turn on and off. The three-phase inverter bridge inverts the DC bus voltage into a three-phase AC voltage with adjustable amplitude and frequency, which is directly injected into the motor drive circuit of the digital high-voltage pump via a power cable, i.e., sent to the three-phase windings of the motor. The motor generates corresponding electromagnetic torque to drive the pump impeller to rotate, realizing dynamic adjustment of the pump speed.

[0070] Reverse control data path - valve control channel: Game-theoretic optimization decision module outputs return water damping adjustment coefficient The firmware of the co-execution unit maps the dimensionless coefficient to a specific physical driving quantity based on the opening-flow resistance characteristic lookup table pre-stored in the controller ROM. Specifically, it sets the current setting value or the corresponding duty cycle PWM signal setting value within the 4-20mA range. This setting value is written to the data register of the microprocessor's on-chip DAC peripheral or the enhanced pulse width modulation peripheral is configured to output a square wave signal with the corresponding duty cycle. The analog voltage or PWM square wave output by the DAC is converted into a 4-20mA standard industrial current signal or a PWM signal with the corresponding duty cycle by the voltage-to-current conversion chip on the circuit board. This control signal is transmitted to the coil of the proportional solenoid valve, the return water regulating mechanism on the return water pipeline, via a shielded cable. The proportional solenoid valve generates a corresponding displacement of its valve core according to the magnitude of the received driving signal, continuously changing the flow cross-sectional area of ​​the valve port, thereby realizing real-time physical adjustment of the fluid resistance of the return water pipeline.

[0071] Feedback closed-loop data path: After each control cycle, the firmware program of the residual feedback self-healing unit of the flexible coupling drive module calls the fluid pulsation sensing module again through the above-mentioned forward sensing data path to obtain the latest pressure sensor data at the pump inlet, calculates the pressure residual, compares the residual with the silent threshold, and if the residual exceeds the preset silent threshold, iteratively calculates the correction parameters based on the PID self-tuning algorithm, updates the intensity of the anti-phase torque pulsation signal in the pump speed control channel and the parameters of the virtual valve opening adjustment command in the valve control channel, forming a complete closed-loop negative feedback control loop until the fluid pulsation characteristics disappear.

[0072] Cloud-based data interaction pathway: In each data reporting cycle, the microprocessor sends current operating data such as pressure, flow rate, rotation speed, and power to the IoT communication module via UART or SPI bus. The communication module encapsulates the data using protocols and establishes an encrypted channel to upload the data in real time to a remote cloud server via a 4G / 5G or Wi-Fi wireless link. The cloud server acts as a data warehouse, storing the device's historical operating data indexed by timestamps. Before the start of each prediction cycle, the user water consumption prediction module firmware sends a data query request to the communication module via the IoT interface, i.e., through the main control microprocessor. The communication module retrieves the device's flow characteristic data for the same historical period from the cloud server and transmits it back to the microprocessor via UART or SPI bus. The user water consumption prediction module firmware processes the data using a weighted moving average statistical algorithm and outputs the predicted user water consumption value for the next period for control strategy reference.

[0073] The digital water tank is a fully enclosed pressure tank structure, with a food-grade flexible diaphragm or air bladder inside to physically isolate the stored purified water from the outside air; the return pipeline where the return water filter is located uses the outlet residual pressure generated by the digital high-pressure pump during water production or circulation to drive the water flow to complete the circulation and keep it alive.

[0074] Specifically, in the construction of the digital water tank, this component is designed as a fully enclosed pressure vessel, its main body consisting of a high-strength metal shell and an internal elastic isolation component. To completely eliminate secondary pollution during water storage, the tank's interior is physically divided into two independent spaces by a food-grade flexible diaphragm or a highly elastic air bladder: an air chamber and a water chamber. The air chamber, located on one side of the diaphragm or air bladder, is pre-filled with inert gas or compressed air at a certain pressure, serving as an energy storage medium; the water chamber, located on the other side, is specifically used to store drinking water that has undergone multi-stage filtration. The physical significance of this structural design lies in its creation of a dynamic breathing system: when purified water enters, the diaphragm expands, compressing the air chamber to store energy; when purified water flows out, the air chamber expands, releasing pressure and propelling the water flow. Throughout the process, the stored purified water is always encased in food-grade flexible material, completely physically isolated from the external atmospheric environment. This fundamentally cuts off the path for dust particles and bacterial spores in the air to enter the water, allowing the system to maintain the biological stability of the water quality without relying on chemical dosing or continuous ultraviolet sterilization.

[0075] Specifically, regarding the fluid drive logic of the return water pipeline, the return water filter is installed in series on the return pipe section between the end of the user's direct drinking water network and the inlet of the digital high-pressure pump. When the system is in water production or recirculation mode, the digital high-pressure pump, as the sole power source for the entire system, generates high-pressure water at its outlet. Although the pressure decreases after passing through the nanofiltration membrane module and distribution network, it still retains significant residual hydrostatic pressure at the end of the network, i.e., outlet residual pressure. This embodiment directly utilizes this residual pressure as the driving force to drive the unconsumed drinking water to overcome the friction resistance along the return water pipeline and return water filter, automatically returning to the low-pressure inlet side of the digital high-pressure pump. This design not only simplifies the electrical control logic of the pipeline system and reduces equipment costs and energy consumption caused by adding a circulation pump, but more importantly, it achieves an adaptive live water circulation mode: as long as the digital high-pressure pump is running, the water in the network continuously flows and renews under the drive of residual pressure, ensuring that the water quality of the entire network, especially in the terminal areas, remains fresh.

[0076] The system has a pump-tank joint scheduling mode: when the user's drinking faucet is in peak water usage and the instantaneous water demand is greater than the rated flow of the digital high-pressure pump, the digital water tank releases the internally stored pressurized water and operates in parallel with the digital high-pressure pump to jointly supply water to the user.

[0077] Specifically, regarding the pump-tank joint scheduling mode of the system, the determination of whether a user's drinking water tap is in peak water usage mode is based on a complex logical verification based on time and statistical patterns. The intelligent control system internally has a pre-set high-frequency water usage time schedule generated through historical data learning, such as 07:00-09:00 AM and 11:00-13:00 PM. When the system clock enters the preset time window, and the flow sensor in the water supply network detects a flow rate change greater than a set threshold, the system marks the current operating condition as peak water usage. The threshold is lower than the rated flow of the digital high-pressure pump, typically set to 1.2 to 1.5 times the system's design average flow. Once peak water usage is determined, the digital high-pressure pump maintains its high-efficiency operating range, continuously outputting the basic flow, while the digital water tank acts as a flow supplement source, injecting additional pressurized water into the network. In terms of fluid topology, the digital high-pressure pump and digital water tank transition from alternating operation or energy storage to parallel water delivery. The two high-pressure streams converge at the main water supply pipe, jointly supplying water to the user's drinking water tap. The physical significance of this model is that digital water tanks can compensate for a portion of water consumption, enabling digital high-pressure pumps and digital water tanks to work together to supply water to users, reducing water pressure during peak periods.

[0078] like Figure 4 As shown, based on this, the digital high-pressure pump integrates a fluid dynamic pressure balance control center, which includes:

[0079] The fluid pulsation sensing module, as the system's sensing front end, is used to extract features representing the game state between return water and inlet water from the raw pressure data at the pump inlet, and to construct a fluid dynamic fingerprint map. Internally, it integrates:

[0080] The waveform discretization unit maps the collected instantaneous pressure waveform data to a one-dimensional equivalent pipeline space domain, constructs a staged spatial envelope map including the main channel corridor, confluence shear and impact restricted area, performs Boolean clipping and consistency check based on physical envelope on the staged spatial envelope map, and outputs the valve disc and jet impact event sequence and rigid impact envelope segment.

[0081] Step S1: Obtain the instantaneous pressure waveform data at the inlet of the digital high-pressure pump, construct a one-dimensional equivalent pipeline coordinate system based on the length of the physical pipeline and the propagation speed of the pressure wave, project the time dimension features of the instantaneous pressure waveform data into the physical position features of the pipeline, and output the pressure echo spatial mapping map.

[0082] Reference Figure 2 This diagram visually illustrates the complete process from the physical pipeline model to the mathematical coordinate system, and then to the spatiotemporal transformation of the signal. The upper part of the diagram shows a one-dimensional equivalent pipeline coordinate system constructed based on the physical pipeline, clearly defining the acoustic coordinates of key physical nodes: coordinates The corresponding pressure acquisition point is the sensor location, coordinates The corresponding confluence node is the return water injection point, with coordinates... The corresponding inlet check valve seat is the potential impact source. The block diagram in the middle of the figure illustrates the spatiotemporal transformation projection algorithm based on the principle of reflection. The lower part of the figure uses dashed arrows to compare and display the instantaneous pressure waveform. That is, the left figure and the pressure echo spatial mapping diagram. That is, the correspondence between the right figures, especially how the echo peak is accurately mapped and located to the valve seat impact source.

[0083] In step S1, refer to Figure 2 The core of this step is to construct a mathematical bridge connecting the time domain signal and the physical domain location, namely a one-dimensional equivalent pipeline coordinate system, and project the abstract time series data into this coordinate system to achieve physical locking of the ghost water hammer impact source.

[0084] The one-dimensional equivalent pipeline coordinate system is constructed as follows: Based on the fluid acoustic transmission line theory, the complex three-dimensional solid pipeline at the inlet of the digital high-pressure pump is simplified into a one-dimensional linear coordinate axis along the fluid flow direction. To accurately quantify the dynamic conflict between the return water jet and the inlet water, the system does not use the sensor location as the origin, but rather the pump inlet confluence node as the geometric zero point of the coordinate system. The coordinates are: The pump inlet confluence node is the physical junction of the return water pipe, the raw water inlet pipe, and the pump suction inlet. The selection of this zero point establishes the physical benchmark for fluid dynamics conflicts: on this benchmark, the direction from the confluence node to the inlet check valve seat is defined as the positive direction of the coordinate axis, and the inlet check valve seat is the potential source of water hammer impact; the direction from the confluence node to the pressure acquisition point is defined as the negative direction of the coordinate axis, and the pressure acquisition point is the sensor mounting position.

[0085] The system performs parametric calibration of the pipeline's physical space, the key to which lies in calculating the accurate acoustic location of each critical node. Specifically, it measures the physical length of the pipeline from the pressure acquisition point to the junction node, and incorporates the equivalent length of the local resistance of elbows and reducers for acoustic correction, thus obtaining the acoustic equivalent coordinates of the pressure acquisition point. Similarly, by measuring the physical length of the pipeline from the manifold to the sealing surface of the inlet check valve seat and the equivalent correction amount, the acoustic equivalent coordinates of the valve seat are obtained. The specific implementation logic and source of the correction amount for the acoustic correction using the equivalent length of the aforementioned local resistance are as follows: In actual engineering pipelines, when pressure waves, i.e., sound waves, flow through 90-degree bends, tees, or diameter changes, the abrupt change in the flow channel geometry causes wave reflection and local turbulence, resulting in microscopic attenuation of wave velocity or travel delay. If only the geometrical physical length of the pipeline centerline is measured, the calculated theoretical arrival time will deviate from the actual arrival time, leading to coordinate positioning errors. Therefore, this embodiment uses the equivalent straight pipe length method for correction. The source and calculation logic for the correction amount are as follows: Based on a pre-set fluid acoustics engineering manual or standard industrial data sheet, such as the CRANE fluid technology manual, the system obtains the equivalent length coefficient corresponding to each type of non-straight pipe component along the path. For each type of non-straight pipe component, such as elbows, tees, and valves, this coefficient represents how many times the resistance generated by the component to the fluid is equivalent to that of a straight pipe of the same diameter. For each specific component along the path, its corresponding equivalent length coefficient is multiplied by the nominal diameter of the component to calculate the equivalent straight pipe length of that single component. The equivalent straight pipe lengths calculated for all local components along the path are summed to obtain the total acoustic equivalent correction amount for that pipe segment. The final acoustic equivalent coordinates are equal to the sum of the measured length of the physical centerline of that pipe segment and the acoustic equivalent correction amount. Acoustic equivalent coordinates are like valve seat coordinates. Through this correction, complex non-straight pipe components are mathematically straightened into a long straight pipe with the same acoustic transmission characteristics, ensuring that the constructed coordinate system highly matches the actual propagation characteristics of pressure waves in the medium. Determine the pressure wave propagation velocity calibration value Since the speed of sound wave propagation in water is affected by the elastic modulus of the pipe, the pipe wall thickness, and the water temperature, in this embodiment... The velocity of sound can be obtained by consulting the standard sound velocity table for the pipe.

[0086] The time-dimensional feature projection of the instantaneous pressure waveform data is the pipeline entity location feature. That is, the spatiotemporal transformation projection process of the instantaneous pressure waveform data is as follows: Based on establishing a precisely corrected coordinate system, the system utilizes spatiotemporal transformation projection technology to transform the instantaneous pressure waveform data collected by the sensor... Mapped to the location features of the pipeline entity. For For each sampling point, especially the peak characterizing the energy abrupt change, the system performs projection calculations based on the principle of round-trip reflection. Specifically, let... The reference time is the time when the initial disturbance occurs at the corresponding confluence node. This represents the moment the wave crest reaches the sensor. The difference between the two values ​​is... This represents the total flight time of the pressure wave within the pipe. In the physical scenario of ghost water hammer, the pressure wave originates from the confluence node, which is the coordinate... Propagation forward to the inlet check valve seat, coordinates After impact and reflection, the wave travels back to the sensor in the opposite direction. This constitutes a complete round trip. Therefore, in order to obtain the unidirectional physical position corresponding to the wave crest, i.e., the equivalent spatial position... The system calculates based on the following formula:

[0087]

[0088] like Figure 2 As shown, the echo peak that appears later on the time axis in the left figure is located at time [time missing]. After halving the round-trip time in the above formula and transforming the dashed arrows by projection, it is accurately mapped to the equivalent position on the spatial axis in the right figure. Due to Includes round-trip travel, calculated as follows The value is exactly equal to ,Right now Figure 2 The coordinates marked on the horizontal axis in the lower right figure: The final output pressure echo spatial mapping diagram. ,Right now Figure 2 The bottom right corner shows a distribution map with the location of the pipeline entity on the horizontal axis and the pressure amplitude on the vertical axis. In this map, peaks that were originally lagging in the time domain are now clearly displayed on the coordinates. That is, at the valve seat of the inlet check valve.

[0089] Step S1 is the physical foundation of the entire ghost water hammer identification algorithm. In the original time-domain waveform without spatial mapping, the high-frequency rotational noise of the digital high-pressure pump motor, the fluid turbulence noise, and the weak water hammer impact signal often highly overlap in spectral characteristics and time-domain amplitude, making it impossible to distinguish between normal pump vibration and abnormal water hammer impact based solely on amplitude thresholds. Through the digital-physical mapping in step S1, this invention introduces the strong physical constraint of spatial position, that is, using the axiom of "same source, different location," transforming the complex signal identification problem into an intuitive geometric position matching problem. Only by first establishing an accurate coordinate scale through local resistance equivalent length correction, and decompressing the time-dimensional waveform to restore the spatial-dimensional entity position distribution, can we ensure that the target signal can accurately fall within the preset coordinate interval when using the valve seat impact restricted area for trimming in subsequent steps, thereby significantly reducing the false alarm rate. This is a necessary prerequisite step for achieving accurate ghost water hammer locking.

[0090] Step S2: Receive the pressure echo spatial mapping map, divide the mainstream steady-state corridor region, the confluence shear envelope region and the valve seat impact prohibition region according to the physical structure of the pump inlet confluence node, and assign corresponding geometric boundary constraints to each region, and output a staged spatial envelope map with physical criterion boundaries.

[0091] like Figure 3 The diagram shown illustrates the phased spatial envelope construction and physical criterion partitioning.

[0092] Reference Figure 3 The diagram is built on the one-dimensional equivalent pipeline coordinate system described in step S1, where the horizontal axis... Represents the physical location of the pipeline, vertical axis This represents the signal amplitude. The diagram visually illustrates that the physical pipeline is discretized into three rectangular envelope regions with different physical properties and control strategies: the green area on the left. This is the mainstream steady-state corridor region, corresponding to the negative half-axis region far from the interference source; the middle yellow area. This is the confluence shear envelope region, corresponding to the turbulent mixing section near the coordinate zero point, whose top boundary corresponds to a high noise suppression threshold. The red area on the right Valve seat impact restricted area, corresponding coordinates The near valve seat location, with its bottom boundary corresponding to the high-sensitivity capture threshold. .

[0093] In step S2, refer to Figure 3 Based on the one-dimensional equivalent pipeline coordinate system established in step S1, the physical structural features of the pipeline are mapped into a mathematically staged spatial envelope, and different signal access thresholds are assigned to different pipeline segments, thereby constructing a digital monitoring map with a zonal control mechanism.

[0094] The method for dividing the spatial region based on the physical structure of the pump inlet manifold is as follows: using a pressure echo spatial mapping diagram, the calibrated coordinate zero point, i.e., the manifold and the coordinate... That is, the valve seat location is used to discretize the one-dimensional linear space of the entire pipeline into three independent regions with distinct physical properties. This defines the valve seat impact prohibition zone. ,like Figure 3 The red area on the right corresponds to the physical inlet check valve seat and its surrounding area, defined by the acoustic equivalent coordinates calibrated in step S1. Using the geometric center as the reference point, and considering the dispersion effect of the pressure wave during propagation and the time drift error of the sensor measurement, a spatial tolerance radius is introduced. The spatial tolerance radius δ is obtained by multiplying the pressure wave propagation velocity calibration value determined in step S1 by the sampling period of the system analog-to-digital converter. This physically means converting the minimum time quantization error caused by discrete sampling into the corresponding spatial positioning uncertainty in a one-dimensional equivalent pipeline coordinate system, i.e., the maximum distance the pressure wave can propagate within a sampling interval. This distance is the minimum spatial resolution limit of the coordinate mapping. For example, when the system sampling rate is 10kHz (i.e., the sampling period is 0.1ms) and the pressure wave propagation velocity calibration value is 1400m / s, the spatial tolerance radius δ is equal to 1400 multiplied by 0.0001, resulting in a calculated δ of 0.14m. For example, the coordinate interval... Defined as the valve seat impact restricted area, this area is the only legitimate physical location for ghost water hammer. Based on the principle of source location uniqueness, any pressure surge falling within this coordinate range is considered an extremely high-risk impact event and is logically defined as a red warning zone. The confluence shear envelope region is also delineated. ,like Figure 3 The yellow area in the middle corresponds to the physical return water injection point and the resulting jet mixing section. Using the coordinate zero point determined in step S1 as a reference, the effective turbulent diffusion length of the return water jet mixing with the main flow of the original water is calculated based on the free jet diffusion theory. ,about The specific calculation implementation method is as follows: a free jet diffusion model based on momentum flux ratio is used for solution, specifically by obtaining the inner diameter of the return water pipe in the inlet manifold structure of the digital high-pressure pump. and the inner diameter of the raw water inlet pipe And collect the initial velocity of the jet injected with return water in real time. Background flow velocity of the main raw water flow Calculate the momentum flux ratio between the return water jet and the main raw water flow. The calculation formula is: This ratio reflects the penetrating power of the jet relative to the mainstream, based on the linear diffusion law of free jets and empirical engineering formulas. Calculate the diffusion length, where For a typical tee junction structure, the jet structure coefficient is... The range of values ​​is This calculation process quantifies the penetration depth and shear mixing range of the high-velocity return jet in the mainstream. The system then calculates the coordinate interval... Defined as the confluence shear envelope region, this area is the source of high-energy noise within the system and is logically defined as the yellow tolerance zone. The remaining effective pipe segment on the coordinate axis, excluding the two specific regions mentioned above, i.e., the region on the negative half of the coordinate axis far from the origin, is defined as the mainstream steady-state corridor region. ,like Figure 3 As shown in the green area on the left, this area corresponds to the laminar flow pipe section before the raw water enters the pump. Theoretically, it should be in a stable state, and is defined as the green steady-state zone in the monitoring logic.

[0095] The process of assigning corresponding geometric boundary constraints to each region and outputting a staged spatial envelope map is as follows: After completing the above spatial division, differentiated amplitude criterion thresholds are further bound to each region to form a non-uniform signal filtering template. The thresholds are dynamically obtained using statistical methods based on historical data to ensure the adaptability of the physical criteria: for valve seat impact restricted areas... This gives it a high sensitivity capture threshold. ,like Figure 3 As shown by the low-level threshold line on the right, the process of obtaining this threshold is as follows: During the quiet period when the digital high-pressure pump stops working and there is no return water injection, a period of time of static noise data from the sensor is collected, and the mean and standard deviation of this dataset are calculated. Statistical guidelines set sensitivity thresholds, for example, such as The physical purpose of setting this threshold is to detect weak early signs of water hammer or slight valve flutter, reflecting an extremely high-risk management strategy that prioritizes minimizing false alarms over omitting any. This applies to the busbar shear envelope region. This gives it a high noise suppression threshold. ,like Figure 3 The high threshold line in the middle is shown. The process of obtaining this threshold is as follows: under the condition of full-load operation of the digital high-pressure pump and full-speed injection of return water, pressure fluctuation data of the confluence area are collected, histogram statistics are performed on the data, and historical data are selected. The quantile or maximum peak envelope is used as a benchmark and multiplied by a safety factor, for example, as shown below. The physical purpose of setting this threshold is to shield against the significant turbulence noise at the confluence node; only when the signal amplitude exceeds this extremely high upper limit is it considered abnormal. This applies to the mainstream steady-state corridor region. Assign it a background stability threshold ,like Figure 3 As shown by the median threshold line on the left, this threshold uses the aforementioned high-sensitivity capture threshold. With high noise suppression threshold The weighted average is calculated by taking the weighting coefficients. Give Weighting coefficients Give ,in Pick , Pick The physical basis for biasing the weights towards the lower threshold side is that the mainstream steady-state corridor region is located in the negative half-axis region of the coordinate system, far from the confluence interference source. The normal background pressure fluctuation level in this region should physically be closer to the system's static noise floor level than the turbulence peak level in the confluence region. Therefore, the threshold setting needs to be biased towards the quieter side to maintain sensitivity to external abnormal interference. For example, when... for , for hour, equal Calculations yielded for This threshold is used to monitor sensor drift or external interference. Signals falling into this area and exceeding the threshold are identified as non-water hammer type external interference and are removed.

[0096] Step S2 achieves a logical leap from homogeneous monitoring across the entire area to differentiated control by region, completely resolving the signal-to-noise ratio inversion contradiction existing in direct drinking water systems. In physical pipelines, the normal turbulent noise amplitude at the junction nodes is often much greater than the weak water hammer signal at the valve seat. If a uniform detection threshold is used throughout the pipeline, it leads to a dilemma: suppressing turbulence will miss water hammer, while capturing water hammer will introduce false alarms. By using a coordinate system to physically isolate noise and signal sources in space through step S2, the system can apply completely different statistical thresholds to different areas, achieving adaptive detection by using a high threshold to suppress interference in high-noise areas and a low threshold for high-sensitivity capture in weak-signal areas. In addition, the clearly defined red, yellow, and green zones and their threshold boundaries in step S2 provide the necessary logical operators and operands for the spatial Boolean pruning in the subsequent step S3, serving as a key logical hub connecting physical coordinate mapping and accurate signal extraction.

[0097] Step S3: Perform spatial Boolean operation on the staged spatial envelope map to extract waveform segments that fall into the valve seat impact forbidden zone and simultaneously fall into the bus shear envelope zone, and use the abrupt amplitude threshold to remove background noise interference, and output the geometric boundary candidate impact set;

[0098] In step S3, after completing the digital partitioning of the pipeline physical space and assigning it a traffic light-style threshold attribute, the core task of step S3 is to perform spatial Boolean clipping, that is, to use mathematical set operation logic to accurately extract physically meaningful effective waveform segments from the complex global waveform and digitally remove background noise and irrelevant fluctuations.

[0099] The implementation method for performing spatial Boolean operations on a staged spatial envelope graph is as follows: the system treats the staged spatial envelope graph as the object to be operated on, and defines a preset physical region as a Boolean mask. The spatial Boolean operation is specifically manifested as a dual logical AND operation of position and amplitude. Specifically, the system generates a binary function of the spatial domain. For valve seat impact restricted area ,set up For the confluence shear envelope region and mainstream steady-state corridor region In the specific logic for finding impact candidates, its mask value is set to ; Spatial mapping of pressure echo With geometric position mask A multiplication operation is performed, which forces all pressure fluctuations located in the main channel and confluence nodes to zero, regardless of their amplitude. This mathematically eliminates the interference of these areas on water hammer determination. The main channel is the green zone, and the confluence node is the yellow zone. Only waveform segments falling within the valve seat impact forbidden zone are retained. These retained segments physically correspond to pressure abrupt changes occurring at the correct locations. The valve seat impact forbidden zone is the red zone, and the correct location is the valve seat.

[0100] The process of removing background noise interference and outputting a geometric boundary crossing candidate impact set using a mutation amplitude threshold is as follows: After completing the initial screening of the spatial locations, the remaining... The fragment performs amplitude logic verification. The segment represents a waveform segment falling into the valve seat impact restricted area. Specifically, for For each waveform segment, determine whether its peak intensity exceeds [the specified value]. If the peak value of a certain waveform segment If the signal is not only legally positioned (located at the valve seat) but also energetically effective (exceeding the static noise floor), it is marked as a geometrically out-of-bounds candidate impulse. If a waveform segment is located in the red zone, but its amplitude... If the signal is not found to be electrical drift or weak thermal noise from the sensor itself, it is determined to be rejected. The system defines the set of all waveform segments that have passed both spatial position verification and amplitude energy verification as the geometric out-of-bounds candidate impact set. Geometric out-of-bounds occurs because these signals appear in the valve seat restricted area where high-pressure fluctuations should not occur; candidate because acoustic path reflection verification has not yet been performed, and currently only their geometric position and amplitude can be confirmed to meet the characteristics.

[0101] Step S3 is the core component connecting the static map construction in S2 and the dynamic event locking in S4. Its indispensability lies in the implementation of the spatial filtering mechanism. In traditional time-domain analysis, extracting weak signals from a noisy background typically requires complex frequency-domain filtering, which often destroys the signal's temporal characteristics and incurs enormous computational costs. Step S3 utilizes the strong spatial constraints established in steps S1 and S2, employing extremely simple Boolean operations—multiplication of 0 and 1—to instantly achieve physical isolation from massive background noise, such as the huge turbulent noise in the yellow zone and the pipe network fluctuations in the green zone. This method is based on physical exclusivity: true ghost water hammer must and can only occur at the valve seat location, i.e., the red zone. Any signal outside the red zone, no matter how strong, is interference; any signal within the red zone, exceeding the noise floor, is a suspect. Step S3, through this spatial location gating analysis, reduces the computational load of subsequent algorithms while improving the signal-to-noise ratio and accuracy of identification, making it a crucial step in achieving real-time online monitoring of the system.

[0102] Step S4: Perform acoustic reflection path verification on the geometric boundary candidate impact set to verify whether the round-trip propagation delay of the candidate impact at the pipeline physical location conforms to physical laws, eliminate artifact interference, and output the valve disc and jet impact event sequence and rigid impact envelope segment.

[0103] In step S4, after initially screening out the candidate impact set with valid geometric positions, the core task of this step is to introduce time-physical causality verification analysis. Using the deterministic delay law in acoustic transmission line theory, artifact interference that falls into the red zone but is not caused by the actual impact of the valve flap is eliminated, and the envelope features characterizing the impact energy are accurately extracted.

[0104] The implementation method for acoustic reflection path verification of the geometric boundary crossing candidate impact set is as follows: the system traverses the geometric boundary crossing candidate impact set output in step S3. For each candidate impact in the set Extract the time of its occurrence. , that is, the observation time, in which This represents the total number of candidate impact waveform segments captured within the current detection period that satisfy both geometric position and amplitude energy constraints. Represents the first in the set The specific candidate impact waveform segments are the objects of the current algorithm's acoustic path verification. A theoretical acoustic model is constructed based on the physical configuration of the pipeline, specifically determining the coordinates from the digital high-pressure pump manifold to the inlet check valve seat. acoustic path length And the coordinates of the valve seat reflected back pressure acquisition point. acoustic path length Digital high-pressure pump manifold, i.e., coordinates The acoustic path length and The calculation method used in all cases employs the same local resistance equivalent length acoustic correction algorithm as in step S1, which involves superimposing the equivalent straight pipe length of elbows and reducing fittings onto the physical pipe length. This is based on the pressure wave propagation velocity calibration value at the current water temperature. The theoretical round-trip propagation delay required for a pressure wave to complete the entire closed-loop physical process of disturbance triggering, propagation, impact, reflection, and return is calculated. The calculation formula is:

[0105]

[0106] System calculates candidate impact Relative to the trigger reference time Actual observation delay Define an acoustic phase tolerance window. ,For example This window is used to accommodate minute dispersion errors caused by changes in fluid velocity or pipe wall viscoelasticity. The judgment logic is: if If the candidate impact conforms to the physical laws of acoustic reflection, it is confirmed as a real valve disc and jet impact event; if If the signal falls into the forbidden zone in the spatial mapping, it is determined that it does not have the continuity of physical propagation in terms of temporal causality. It belongs to the artifact interference caused by sensor electrical mutation, random electromagnetic interference or pump mechanical resonance, and the system removes it from the set.

[0107] The process of removing artifact interference and outputting the valve disc and jet impact event sequence and rigid impact envelope fragment is as follows: After rigorous verification in the aforementioned time dimension, the system obtains a set of valid impact events that are highly self-consistent in both the spatial and temporal domains. The system sorts these events in chronological order to generate the valve disc and jet impact event sequence. Furthermore, for each real impact event in the sequence, the system performs rigid impact envelope extraction: It locates the impact peak point, which is the sampling point where the pressure amplitude reaches its maximum within the candidate impact waveform segment, representing the maximum energy intensity released at the moment of impact; using this impact peak point as the center, it performs a infinitesimal search to the left and right of the time axis. The search continues to the left until the inflection point where the waveform curvature reverses, which is the point where the sign of the second derivative of the pressure curve changes abruptly from flat to steep, physically corresponding to the moment when the valve disc just contacts the valve seat and begins to generate a rapid pressure rise; this is defined as the impact point. The search continues to the right until the pressure value decays to a certain proportion of the peak value, defined as the relaxation point. Waveform data from the initial impact point to the relaxation point is extracted. This data segment fully records the pressure surge and damped oscillation process at the moment of violent impact between the valve disc's metal sealing surface and the rigid valve seat, and is defined as the rigid impact envelope segment. This segment is the basic physical unit for subsequent calculation of impact momentum and assessment of system damage.

[0108] Step S4 is a crucial verification step in realizing the identification of phantom water hammer from possibility to certainty. Although the spatial Boolean operation in step S3 significantly filters background noise through coordinate constraints, there is a risk of coincidental misjudgment. That is, some high-amplitude random noise may happen to fall into the valve seat coordinate interval (red zone) in the mathematical mapping. Random noise such as electromagnetic spikes or cavitation collapse noise from the frequency converter. Without the verification of step S4, the control system may make incorrect compensation responses to these false signals, leading to system oscillation. Step S4 introduces the physical uniqueness constraint of the acoustic path: the real water hammer signal must not only be in the correct position but also in the correct time. Wave propagation in the medium must follow a strict velocity-distance relationship, which is a physical law that cannot be simulated by random noise. By comparing the theoretical delay with the actual delay, step S4 effectively removes all non-causal artifacts, ensuring that every impact event in the output is a real physical impact. At the same time, the accurate extraction of the rigid impact envelope segment provides a clean sample for the subsequent quantification of the momentum of the return water jet, avoiding the interference of background turbulence on the accuracy of momentum integration, which is a necessary prerequisite for achieving precise control closed loop.

[0109] The momentum flux quantization unit, connected to the waveform discretization unit, is used to perform physical quantity integration on the rigid impact envelope segment, calculate the pressure-time integral area of ​​a single impact event within the fluid micro-element, and derive the return water pulse momentum value in combination with the pipeline cross-sectional parameters.

[0110] The momentum flux quantization unit receives the rigid impact envelope segment from the waveform discretization unit. The rigid impact envelope segment refers to the core time period during the entire water hammer wave process, after removing subsequent elastic aftershocks and damped tail waves, retaining only the data from the moment the valve disc's metal sealing surface and rigid valve seat physically come into contact, resulting in a violent momentum exchange. Mathematically, the system processes this segment as a discrete time-pressure sequence. ,in, This represents the total number of sampling points contained within the envelope segment. This is the temporal index of the sampling point within the segment. Indicates the first Each sampling time, For a moment The corresponding instantaneous pressure amplitude. The core task of this unit is to transform the scalar pressure sequence, which characterizes signal strength, into vector momentum parameters, which characterize fluid dynamic energy.

[0111] The process of calculating the pressure-time integral area of ​​a single impact event within a fluid element is as follows: The system first determines the integration time window. . This corresponds to the identified starting point, i.e. the moment when the second derivative of the pressure curve abruptly changes; This corresponds to the defined relaxation point, i.e., the moment when the pressure decays to a specific proportion of its peak value. A trapezoidal numerical integration algorithm is used to integrate the waveform within this time window. The calculation formula is:

[0112]

[0113] in, The system sampling time interval, The steady-state reference pressure before the impact is obtained by taking the start time of the time window. The corresponding pressure value, or the average static background pressure of the pipeline network collected during the quiet period in step S2, is used to eliminate the system static pressure, ensuring that the integral result only reflects the dynamic increment caused by the impact of the return water jet. The calculated integral result... Defined as the pressure-time integral area, this physical quantity represents the impulse density exerted by the valve seat on the fluid element per unit pipe cross-sectional area, reflecting the cumulative effect of pressure wave energy in the time dimension during the core impact process of initiation-relaxation.

[0114] The process of deriving the return water pulse momentum value based on the pipe cross-sectional parameters is as follows: The system converts the aforementioned impulse density into a change in fluid momentum according to the impulse-momentum theorem. In the return water confluence scenario of a direct drinking water system, the valve seat is rigid and stationary. The velocity of the return water jet element is instantaneously stopped after impacting the valve seat. Therefore, the total impulse exerted by the valve seat on the fluid is numerically equal to the total momentum of the return water jet element before impact. The system retrieves preset pipe cross-sectional parameters, specifically the inner cross-sectional area of ​​the return water pipe at the confluence node. The area is calculated by obtaining the nominal inner diameter of the return water pipe. This parameter is obtained by reading the system initialization configuration file or the pre-stored hardware parameter database. Specifically, based on the return water pipe diameter specifications entered during equipment installation, the corresponding standard inner diameter value is retrieved from the built-in industrial pipe standard table. This parameter is consistent with the parameter used to calculate the jet diffusion length in step S2; it is calculated based on the circle area formula. ; Calculate the return water pulse momentum value using the formula :

[0115]

[0116] The momentum value of the return water pulse directly quantifies how much momentum the return water fluid element carried when it collided head-on with the valve seat in this ghost water hammer incident. The larger the value, the faster the return water jet velocity, the denser the mass flow rate, and the stronger its ability to invade the water intake system.

[0117] The momentum flux quantization unit realizes the key dimension transformation from signal domain features to physical domain energy. Its indispensability is primarily reflected in the accuracy of energy quantization: although step S4 extracts the impact waveform, the simple pressure peak is easily affected by bubble collapse or sensor transient response, resulting in fluctuations that cannot fully represent the true impact force of the fluid. For example, an extremely high but short spike may have less destructive force than an impact with a slightly lower amplitude but longer duration. Through pressure-time integration, this unit captures the waveform's "fat and thin" characteristics, i.e., the area of ​​the rigid impact envelope, accurately reconstructing the total impact energy. It serves as the fundamental data source for subsequent construction of fluid dynamics fingerprints and Nash equilibrium modeling. Subsequent game theory control requires knowledge of the momentum of the return jet to calculate the necessary pump speed compensation, i.e., the influent momentum, to neutralize it. Without this unit's... Without precise calculations, subsequent control algorithms cannot be solved due to a lack of physical input, resulting in the control system's inability to achieve accurate dynamic pressure balance. Therefore, this unit serves as a physical quantization bridge connecting waveform recognition and feedback control.

[0118] The spatiotemporal fingerprint construction unit, connected to the waveform discretization unit and the momentum flux quantization unit, is used to statistically analyze the distribution density of the valve disc and jet impact event sequence within a unit time to generate the fluid intermittent stagnation duty cycle. The return water pulse momentum value and the fluid intermittent stagnation duty cycle are used as the vertical and horizontal axes, respectively, to construct a fluid dynamic fingerprint spectrum under multi-source confluence state.

[0119] The spatiotemporal fingerprinting unit serves as the logical hub connecting the sensing and decision-making ends, linking the front-end waveform discretization unit and momentum flux quantization unit. Its core function is to transform the time-discrete single impact events output from the previous stage into continuous variables reflecting the macroscopic fluid dynamics state of the system over a period of time. The system constructs a two-dimensional feature space for subsequent control modules to evaluate the state by statistically aggregating all real valve flap and jet impact events captured within a sliding statistical time window that progresses along the time axis.

[0120] The process of generating the fluid intermittent stagnation duty cycle is as follows: The system uses the logic of cumulative occupancy ratio in the time domain for analysis. The specific steps are as follows: The system establishes a fixed statistical duration as the observation window, such as selecting several past pumping cycles or a fixed time length; within this observation window, the system retrieves all valve disc and jet impact event sequences confirmed by the waveform discrete unit, and extracts the rigid impact envelope segment corresponding to each event. The system obtains the physical duration of each rigid impact envelope segment, that is, the time span from the starting point to the relaxation point of the segment; the physical durations of all effective impact events within the window are summed to obtain the total time length during which the return water jet suppresses the valve seat within the current window; this total suppression time length is compared with the total duration of the observation window, and the percentage of the former to the latter is calculated. This percentage is defined as the fluid intermittent stagnation duty cycle. The physical significance of this physical quantity is that it intuitively quantifies the time density during which the inlet check valve seat is strongly suppressed by the return water jet or experiences violent momentum conflict within the current statistical period. The higher the ratio, the more frequent and continuous the attack of the return jet, resulting in a more severe degree of intermittent cut-off or blockage of the inlet water, reflecting the degree of congestion of fluid conflict in the time dimension.

[0121] The process of constructing a fluid dynamic fingerprint map under multi-source confluence conditions is as follows: The system uses the return water pulse momentum value output by the momentum flux quantization unit as the vertical axis (Y-axis) of the map to characterize the intensity of fluid conflict in the energy dimension. The return water pulse momentum value is a statistically representative value, such as the average or maximum value, for multiple events within a window. The system uses the fluid intermittent stagnation duty cycle calculated through the above steps as the horizontal axis (X-axis) of the map to characterize the density of fluid conflict in the time dimension. The system maps the calculated horizontal and vertical coordinate values ​​at each moment to a two-dimensional Cartesian coordinate system, forming a dynamic state point. This state point and its trajectory evolving over time constitute the fluid dynamic fingerprint map. This map accurately maps the complex physical state of multi-source confluence into mathematical coordinate points, providing numerical inputs that can be directly calculated without prior prediction for subsequent control algorithms.

[0122] The spatiotemporal fingerprinting unit achieves a crucial dimensionality-up mapping from microscopic discrete events to macroscopic continuous states. Its indispensability lies in providing observable global state variables for subsequent control modules. While previous steps accurately identified individual water hammer events, for fluid balance control, a single impact is merely a transient pulse and cannot be directly used as a basis for adjusting pump speed or damping. This unit, by constructing a fingerprint spectrum, decouples the scattered impact signal into two orthogonal dimensions: frequency and intensity, creating a panoramic system state observer. Frequency is represented by duty cycle, and intensity by momentum. This not only smooths out the random fluctuations of a single event but, more importantly, provides the necessary input variables for the jet countermeasure analysis module and Nash equilibrium modeling unit mentioned later in the specification. The jet countermeasure analysis module quantifies the back pressure stagnation exponent, and the Nash equilibrium modeling unit constructs the game theory model. Without this fingerprint spectrum, subsequent modules would fail to operate due to a lack of state parameters describing the situation, leading to a break in the entire closed-loop control logic.

[0123] The jet countermeasure analysis module, connected to the fluid pulsation sensing module, serves as a disturbance observer for the system. It decodes the fluid dynamic fingerprint to quantify the physical intrusion degree of the return water jet and obtains the back pressure resistance index. Specifically, it includes the following units:

[0124] The jet velocity inversion unit is used to analyze the vertical axis features of the fluid dynamic fingerprint spectrum and, based on the inverse transformation of the momentum conservation law, calculate the equivalent jet core velocity at the outlet of the return water pipeline from the momentum value of the return water pulse.

[0125] The jet velocity inversion unit, as the core of the system's soft-sensor calculations, connects to the previously constructed fluid dynamics fingerprint map. Its task is to decode the abstract energy coordinates, i.e., the vertical axis features, in the map into concrete fluid kinematic parameters. At the physical level, since the inlet confluence area of ​​the digital high-pressure pump is in a high-pressure, highly turbulent, closed environment, this unit adopts an inversion algorithm based on a physical model to calculate the unknown jet source velocity, i.e., the equivalent jet core velocity, using the known impulse response, i.e., the return water pulse momentum value.

[0126] The process of calculating the equivalent jet core velocity at the outlet of the return water pipeline based on the inverse transformation of the momentum conservation law is as follows: The system analyzes the vertical axis data of the fluid dynamic fingerprint spectrum and extracts the representative return water pulse momentum values ​​within the current statistical window. The system retrieves the corresponding average impact duration from the associated rigid impact envelope segment data. This refers to the time span from the initial impulse point to the relaxation point. An impulse-velocity inversion model is established. Based on the law of conservation of momentum, the impulse exerted by the fluid on the valve seat is equal to the change in fluid momentum. Assuming the return jet impacts the valve seat perpendicularly and stagnates, i.e., its velocity becomes zero after the impact, then:

[0127]

[0128] in This refers to the instantaneous force exerted on the sealing surface of the valve seat when the return water jet impacts the valve seat of the inlet check valve. This represents the average impact force of the jet on the valve seat. Based on the jet impact force formula in fluid dynamics... Combining the above two equations, we can obtain the inverse transformation equation for velocity:

[0129]

[0130] The system, based on this equation, utilizes the known fluid density and the cross-sectional area of ​​the return water pipe Fluid density is the density constant of water at the current temperature, such as... The equivalent jet core velocity was calculated in reverse. :

[0131]

[0132] Calculated Defined as the equivalent jet core velocity. The physical meaning of this quantity is that it eliminates the effects of pipe friction resistance and turbulent dissipation in the mixing zone, restoring the initial velocity of the return water jet just after leaving the return water pipe outlet and before it has violently mixed with the main influent flow. This is the most direct kinematic indicator for quantifying the attack force of the return water jet, directly determining its penetration depth and dynamic pressure blocking capability at the confluence node.

[0133] The jet velocity inversion unit analysis is based on the deterministic physical correlation of the impulse-momentum theorem. The impulse generated by the jet impacting the valve seat is not a random variable, but a physical result strictly determined by the mass and velocity of the fluid. This unit establishes analytical equations based on fluid mechanics laws, logically restoring the mechanical response collected by the pressure sensor to the kinematic state of the fluid. Its indispensability lies in providing the system with the necessary fluid kinematic input parameters. The core logic of subsequent back pressure hindrance index calculation and dynamic pressure balance control lies in comparing the momentum ratio of the return water and the inlet water, and velocity is the core element in calculating momentum. Given that installing an invasive velocity sensor in the high-pressure turbulent zone at the pump inlet would disrupt flow field stability and be costly, the mechanism inversion method used in this unit is the only effective way to obtain the core jet velocity without increasing hardware costs or interfering with the flow field environment. Without this unit, the control system would be unable to quantify the true intensity of the return water jet due to the lack of key velocity data, causing subsequent control strategies to lose their physical basis.

[0134] The effective flux calculation unit is used to analyze the horizontal axis features of the fluid dynamic fingerprint spectrum, correct the theoretical geometric flow section of the raw water inlet using the fluid intermittent retardation duty cycle, and calculate the dynamic effective influent flux under the high-frequency interference of the return water jet.

[0135] The theoretical geometric flow cross-section of the raw water inlet is determined. This parameter is a static benchmark value based on the physical pipe hardware specifications. Specifically, the system reads the equipment initialization configuration file or the underlying hardware parameter database and extracts the pre-entered nominal inner diameter of the raw water inlet pipe. This nominal inner diameter is a fixed parameter entered and stored during the equipment installation and commissioning phase, based on the standard industrial specifications of the water supply pipes connected on-site. Using the circle area calculation rule, the system calculates the physical cross-sectional area of ​​the pipe using this nominal inner diameter, defining it as the theoretical geometric flow cross-section. This cross-section represents the maximum physical flow channel area that the inlet pipe can provide under ideal static conditions without any external fluid interference.

[0136] By analyzing the horizontal axis features of the fluid dynamic fingerprint spectrum, the intermittent fluid congestion duty cycle within the current statistical window is extracted. The analysis of the physical meaning of this parameter follows a fluid dynamic causal chain: when the return water jet impacts the valve seat, it generates a local high-pressure zone higher than the inlet water pressure, causing the inlet check valve to be forced to close under pressure differential or the inlet flow to stagnate under momentum conflict. Therefore, within the macroscopic statistical time window, this duty cycle value directly reflects the proportion of time the inlet channel is in a state of physical open / closed or momentum blockage. Based on this physical fact, the intermittent fluid congestion duty cycle essentially characterizes the temporal congestion density of the hydraulic channel.

[0137] The theoretical geometric flow cross-section is dynamically corrected based on the intermittent fluid hindrance duty cycle to calculate the dynamic effective flow cross-section. The physical basis of the correction is the time-domain average equivalent method: since the raw water cannot pass through during the time period corresponding to the duty cycle, statistically speaking, the effective flow capacity of the pipeline is reduced. The calculation process is as follows: calculate the non-hindrance time ratio, that is, the numerical... Subtract the duty cycle of intermittent fluid stagnation The ratio This represents the percentage of time during which the influent can flow normally; multiplying the theoretical geometric flow cross-section by this non-obstruction time ratio yields the dynamic effective flow cross-section. This step mathematically equates the discrete, high-frequency valve opening and closing behavior to a continuous reduction in the physical cross-sectional area of ​​the pipe, thereby quantifying the actual bottleneck effect of backflow disturbance on the water supply channel.

[0138] The system calculates the dynamic effective influent flux by combining the mainstream flow velocity of the raw water. The mainstream flow velocity is obtained by reading real-time sampling data from flow rate sensors installed at the inlet of the raw water tank or upstream of the digital high-pressure pump. These sensors, such as electromagnetic or ultrasonic flow meters, reflect the fluid velocity within the current water supply network context. The system then multiplies the calculated dynamic effective flow cross-section with the collected mainstream flow velocity of the raw water; the product is defined as the dynamic effective influent flux. This physical quantity differs from the cumulative flow displayed by the flow meter. It represents the upper limit of the maximum mass flow rate that the water supply network can actually draw from after overcoming the interference of the return water jet under the current transient operating conditions.

[0139] The effective flux calculation unit is a physical prerequisite for establishing a dynamic pressure balance between supply and demand, thereby eliminating cavitation. In multi-source confluence systems, the ghost water hammer effect generated by the return jet acts like a virtual valve that switches on and off at an extremely high frequency, constantly cutting off the water inlet path. If the control system ignores this physical fact and estimates the water supply capacity solely based on the physical diameter of the pipe, i.e., the theoretical geometric flow cross-section, it will inevitably lead to a misjudgment of sufficient water supply. The indispensability of this unit lies in its ability to identify the hydraulic blockage hidden inside the physical pipe by introducing the key variable of the intermittent fluid hindrance duty cycle, thus accurately calculating the discounted true water supply capacity, i.e., the dynamic effective inlet flux. This provides an insurmountable safety boundary for subsequent game theory optimization, namely, the pump's suction demand must never exceed this dynamic effective influent flux. Without this unit's calculation, the subsequent control module will lack a real flow limit reference, causing the pump to continue pumping at full speed even when the water supply is insufficient. This can lead to fluid separation, vacuum cavitation, and mechanical vibration, severely compromising the system's stability.

[0140] The retardation index output unit is used to calculate the momentum ratio between the equivalent jet core velocity and the main flow velocity of the raw water, and to weight-couple the momentum ratio with the dynamic effective influent flux to output the back pressure retardation index.

[0141] Calculate the momentum ratio of the equivalent jet core velocity to the raw water mainstream velocity. Based on the principles of fluid dynamics, the magnitude of momentum depends on the product of mass and velocity. Under the premise that the density of an incompressible fluid (water) is constant and that the volume of the infinitesimal elements is the same, the ratio of velocities directly reflects the difference in momentum density per unit volume of fluid. The calculation formula is as follows:

[0142]

[0143] in This represents the main flow velocity of the raw water from the flow sensor. The square of the velocity ratio is used here because the dynamic pressure of the fluid, i.e., the kinetic energy density, is proportional to the square of the velocity. The physical significance lies in quantifying the energy ratio between the attacker and the defender: when When the dynamic pressure of the return water jet is higher than that of the inlet water, the return water jet has the ability to penetrate and block the inlet water channel, forming a hard blockage; when When the inlet dynamic pressure is dominant, it can disperse and entrain the return water jet downstream. At this time, the return water no longer forms a physical blockade, but is transformed into a high-energy turbulence source, which increases the frictional resistance of the mainstream through fluid shearing and eddy dissipation, forming soft stagnation.

[0144] Momentum ratio With dynamic effective influent flux Weighted coupling is performed to output the back pressure hysteresis index. In this step, the system first defines and calculates the normalization factor, which is the rated maximum flow rate of the inlet pipe. . The calculation method is the product of the theoretical geometric flow cross-section and the standard water supply flow velocity. The standard water supply flow velocity is obtained by the system reading engineering design parameters stored in the equipment's underlying database. These parameters are standard economic flow velocity values ​​retrieved and set based on outdoor water supply design standards or local water supply network technical specifications, and are typically within a certain range. to This value represents the baseline flow velocity of the inlet pipe under ideal delivery conditions at the design pressure. Based on this, a penalty function based on the water supply bottleneck degree is constructed, and the weighted coupled calculation model is as follows:

[0145]

[0146] in, and The system uses pre-set weighting coefficients α and β to balance the contributions of momentum conflict and flow congestion to system impedance. The values ​​of these weighting coefficients must satisfy a normalization constraint, meaning their sum equals 1. The specific values ​​are determined as follows: In the return water confluence scenario of a direct drinking water system, the physical threat posed by the return water jet to the inlet channel is mainly manifested as direct momentum conflict caused by dynamic pressure, while flow congestion is an indirect consequence of momentum conflict. Therefore, the weight of the dynamic pressure resistance term should be higher than the weight of the channel narrowing resistance term. Based on this physical judgment, the system sets α to a dominant weight greater than 0.5 and β to a subordinate weight less than 0.5. For example, α is set to 0.6 and β to 0.4. The first term in the formula... This characterizes the dynamic pressure against resistance, that is, the additional energy cost the pump needs to overcome the backwater jet's upward pressure, whether through hard blocking or soft shearing; the second term in the formula... Characterizes the resistance to channel narrowing, when When the return water duty cycle calculated in the previous steps decreases significantly due to being too high, this value increases sharply, reflecting the increase in equivalent flow resistance caused by the shortened inlet window period. The calculated value... It is the back pressure resistance index, which is a comprehensive dimensionless scalar.

[0147] The back pressure hindrance index output unit follows the principle of cost function construction in control theory, abstracting the multivariable coupled nonlinear fluid physics field into a single mathematical evaluation value, which is a necessary input interface for the subsequent Nash equilibrium modeling unit. In the fluid dynamic pressure game model, the payoffs and costs for both sides must be clearly defined. For the inlet flow path, the presence of the return jet constitutes a significant environmental impedance. Without the back pressure hindrance index calculated by this unit, the Nash equilibrium model lacks a penalty term representing the conflict cost. The subsequent game optimization decision module is based on this... The amplitude is used to dynamically adjust the strategy: when When the temperature is low, the environment is considered relaxed, and a low-energy consumption strategy is adopted; when... During the increase in pressure, the system identifies the high-damping counter-condition and calculates a precise pump speed compensation vector to offset this resistance. Therefore, this unit is a crucial step in transforming physical fluid dynamics phenomena into mathematical game theory model parameters. The absence of this step would prevent the control algorithm from perceiving the severity of the environment, leading to control failure.

[0148] The game-theoretic optimization decision-making module, connected to the jet countermeasure analysis module, serves as the system's control law calculation unit. It performs control law optimization calculations on the back pressure hysteresis index, solving for the pump speed dynamic compensation vector and the return water damping adjustment coefficient. Specifically, it includes the following units:

[0149] The Nash equilibrium modeling unit is used to construct a non-cooperative game model of fluid dynamic pressure between the inlet flow path and the return flow path. The back pressure resistance index is used as the conflict penalty term of the game, and the Nash equilibrium point is calculated with the goal of minimizing the total kinetic energy of the turbulent flow at the pump inlet.

[0150] The fluid dynamic pressure non-cooperative game model specifically includes the following three core components: 1) Defining the players: The system abstracts the physical interaction at the confluence node into two participants with independent physical wills. Player A is the inlet flow path / pump, representing the power system of the inlet water and the digital high-pressure pump. Its physical goal is to maintain a stable inlet water flow to meet downstream demand. Player B is the return flow path / jet, representing the return water jet carrying high kinetic energy. Its physical goal is to overcome the back pressure at the confluence node to achieve a smooth loop. This game is defined as a non-cooperative game, and its physical essence lies in the fact that within the limited physical space of the confluence node, the inlet and return water jets have a natural exclusive competition in terms of flow channel cross-sectional occupancy and dynamic pressure potential energy distribution. The faster the return water velocity, the stronger the resistance to the inlet water; the higher the inlet water pressure, the greater the resistance to the return water discharge. The two sides cannot achieve coexistence through simple linear superposition and must find a balance through game theory. 2) Defining the game strategy space: that is, the set of control variables that each player can adjust. Player A's strategy variable: defined as the rotational speed component in the pump speed dynamic compensation vector. This variable corresponds to the rotational speed of the digital high-pressure pump motor and is the active control method used by Player A to change the momentum of the incoming water. Player B's strategy variable is defined as the return water damping adjustment coefficient. This variable is a variable whose value range is The dimensionless control parameter has a physical meaning that corresponds to the opening state of the virtual valve on the return water pipeline. The specific meaning is as follows: When When the virtual valve is fully open, the return water system is in an undamped natural discharge state, and the jet attack is at its strongest; when At this time, it indicates that the virtual valve is completely closed and the return water path is cut off. This parameter is used by player B to adjust the intensity of their jet. 3) Define the game payoff objective function: that is, the mathematical standard used to evaluate the merits of the current strategy. The model sets the global payoff objective. Total kinetic energy of turbulent flow at the pump inlet Minimize the value of the drinking water demand and introduce a penalty term. The specific construction is as follows: Determine the boundary constraints of the model, namely the user's direct drinking water demand. . The data is obtained in real time by the system reading the flow sensor data installed at the front end of the user's drinking faucet, or by directly reading the real-time flow setpoint in the constant pressure variable frequency control loop of the digital high-pressure pump. This parameter represents the rigid water supply task that must be met during the game. Calculate the total turbulent kinetic energy at the pump inlet. The physical meaning of this value is the energy dissipated into disordered turbulence when two fluids collide at their confluence. According to the fluid mixing zone shear theory, its value is proportional to the square of the difference between the inflow velocity vector and the return velocity vector. The calculation formula is:

[0151]

[0152] in, The characteristic mixing volume constant of the busbar node is determined by reading fixed geometric parameters stored in the device database. These parameters are physical volume values ​​of the internal fluid mixing cavity calculated based on the mechanical design drawings or three-dimensional digital models of the inlet busbar tee of the digital high-pressure pump. The term reflects the damping effect of the damping coefficient on the jet velocity. A back pressure retardation index is introduced. As a conflict penalty term, the final objective function formula is constructed as follows:

[0153]

[0154] in The global benefit is represented by and A bivariate function jointly determined; This indicates that the dynamic effective influent flux is processed in terms of pump speed. A single-variable function is used to predict the theoretical water supply capacity at different rotational speeds. This formula clarifies the game rule of the model: the target state sought by the system must maximize the total turbulent kinetic energy generated by fluid collisions. The minimum, and must pass the penalty item. Ensure that the flow rate deviation is within the allowable range. Specifically, The existence of this factor means that the harsher the environment, i.e., the higher the resistance index, the greater the penalty the system imposes on flow deviations, thus forcing the strategy to converge in a more robust direction. The penalty weighting coefficient γ is a fixed constant obtained through calibration using historical operating data. During system initialization, the control unit extracts sample data from when the system is in its historical optimal operating state and calculates the average amplitude of the total turbulent kinetic energy and the average amplitude of the flow deviation term at that time. The system uses the ratio of these two amplitudes as the value of γ to eliminate the numerical magnitude difference caused by the different physical dimensions of energy consumption and flow deviation, ensuring that these two indicators are of the same order of magnitude in the objective function. This guarantees that the Nash equilibrium model can fairly and effectively find the optimal solution.

[0155] Based on the aforementioned non-cooperative hydrodynamic game model, the specific process for solving the Nash equilibrium point is as follows: The system employs a numerical iterative optimization algorithm, such as particle swarm optimization or gradient descent, in a two-dimensional policy space. The system performs a micro-element search within the system. Different elements will be... and Substitute the above The formula is used to calculate the corresponding objective function value. The system searches for the objective function value that... The coordinates of the point that reaches the global minimum value At this point, the definition of Nash equilibrium is satisfied: the inlet flow path cannot be altered solely by changing the rotational speed component. To further reduce the system's turbulent energy or flow deviation, the return water flow path cannot be improved simply by changing the return water damping adjustment coefficient. To optimize the system state, the inlet water flow path is represented by player A, and the return water flow path by player B. The system will use this set of calculated optimal coordinates. The Nash equilibrium point is locked and used as the optimal pump speed command and optimal damping adjustment command output, respectively.

[0156] The Nash equilibrium modeling unit is a mathematical decoupling method for resolving strongly coupled nonlinear fluid conflicts. The constructed fluid dynamic pressure non-cooperative game model is a complete mathematical entity that includes players, strategy space, and payoff function. Traditional PID control often treats the return water as a single external disturbance, while this model acknowledges the competitive status of the return water flow path as an independent physical entity. Its indispensability lies in the fact that without this model and its constituent elements... and Without a penalty structure, the control system cannot predict the dynamic pressure conflict costs caused by forced acceleration. By calculating the Nash equilibrium point, the system finds a compromise steady state: at this point, the suction negative pressure provided by the pump is just enough to deflect the return water jet and allow it to flow downstream into the impeller, without causing either return water backlash or forced cutoff. This mathematically guarantees the global stability of the control strategy, fundamentally eliminating the risk of system divergence caused by control logic opposing physical laws.

[0157] A security boundary verification unit, connected to the Nash equilibrium modeling unit, is used to receive the Nash equilibrium point and compare and verify the numerical coordinates of the Nash equilibrium point with a preset hardware security threshold range to determine the optimal operating point.

[0158] The safety boundary check unit is the final hard safety lock connecting the algorithm domain and the execution domain, designed to resolve potential conflicts between the mathematically optimal solution and the physically feasible solution. Although the preceding unit calculates the Nash equilibrium point... and The mathematical model maximizes the benefit, but this value may exceed the rated operating range of the physical hardware. Therefore, a deterministic safety boundary check is performed. The specific execution process is as follows: The system reads the preset hardware safety threshold range, which is defined by the physical limit parameters of the device and stored in the controller memory. Specifically, this includes the safe pump speed range that defines the operating range of the digital high-pressure pump motor and the effective range of the damping coefficient corresponding to the physical opening limit of the virtual valve. The upper limit of the safe pump speed range is the rated maximum speed of the motor; exceeding this value will cause the motor to overheat or the bearings to be damaged. The lower limit is the minimum pressure-holding speed of the pump; falling below this value may result in insufficient outlet water pressure. A value of zero in the effective range of the damping coefficient represents that the virtual valve is fully open, and a value of one represents that the virtual valve is fully closed. Based on this, the system performs numerical comparison and boundary clamping operations, adjusting the speed component in the Nash equilibrium point. and return water damping adjustment coefficient Logical comparisons are performed with the physical intervals defined above. For speed verification, the speed component at the Nash equilibrium point... When the speed exceeds the motor's rated maximum speed, the system determines that the command poses an overspeed risk and forcibly sets the speed to the rated maximum speed; when it is lower than the minimum pressure holding speed, it is forcibly set to the minimum pressure holding speed, thus ensuring that the speed command issued to the motor is always limited to the safe range allowed by the nameplate. Regarding damping verification, when the return water damping adjustment coefficient at the Nash equilibrium point... When the value is greater than one, the system determines the command is invalid and forcibly sets it to one; when the value is less than zero, it is forcibly set to zero, thus ensuring that the valve opening command is always within the range of the physical mechanism's executable stroke. The speed coordinates and damping coordinates after the above clamping correction are defined as follows: and The system ultimately determines this coordinate combination as the optimal operating point and outputs it. This unit uses threshold gating to intercept any aggressive algorithmic instructions that might cause system failure before they reach the execution layer, ensuring the system's engineering safety and feasibility.

[0159] The compensation vector generation unit generates a dynamic compensation vector for pump speed and a return water damping adjustment coefficient based on the deviation between the current operating conditions and the optimal operating point.

[0160] The compensation vector generation unit transforms the optimal operating point into dynamic control commands that the actuator can recognize. The specific execution process is as follows: the system constructs a real-time deviation space and reads the current actual speed of the digital high-pressure pump in real time. and the actual damping state of the virtual valve It also receives the optimal operating point coordinates output by the safety boundary verification unit. Calculate the speed control deviation The system acquires the mechanical response delay time. This parameter directly calls the motor dynamic characteristic parameters stored in the underlying database. It belongs to the existing technology used to characterize the time constant required for a specific motor model to respond to a torque command and achieve the desired physical speed. Based on this, the system generates a pump speed dynamic compensation vector, constructing a vector signal containing amplitude and phase for the speed deviation. Among them, the amplitude component This refers to the speed control deviation, used to define the target intensity of the adjustment; phase lead angle. The calculation involves multiplying the predicted arrival time of the return water pulse from the previous steps, combined with the current real-time electrical angular velocity of the motor and the mechanical response delay time obtained above, to obtain the phase lead angle. The physical meaning of this parameter is: to compensate for... The resulting lag requires the controller to advance the spatial phase. The system issues action commands from various angles to ensure that the pump speed adjustment is completed precisely at the moment the return water pulse arrives. The system generates a return water damping adjustment coefficient, which is directly extracted. This is then converted into virtual valve opening commands. The system synchronously sends the pump speed dynamic compensation vector and return water damping adjustment coefficient to the collaborative execution unit. The indispensability of the compensation vector generation unit lies in its ability to solve the control time delay problem. Generate phase lead angle This was then applied to hydrodynamic game control. This step ensured precise alignment of the control action with the water hammer impact on the time axis, achieving hydrodynamic counterbalancing rather than passive following, thereby effectively eliminating high-frequency ghost water hammer.

[0161] A flexible coupling drive module, connected to the game-theoretic optimization decision module, serves as the physical execution end of the system. It converts the pump speed dynamic compensation vector and the return water damping adjustment coefficient into physical fluid mechanical actions, and performs closed-loop self-healing correction based on the execution effect. It includes the following units:

[0162] The collaborative execution unit is used to receive the pump speed dynamic compensation vector and the return water damping adjustment coefficient, respectively perform signal modulation and command mapping on them, generate an anti-phase torque pulsation signal to drive the digital high-pressure pump motor, and generate a virtual valve opening adjustment command to drive the return water adjustment mechanism.

[0163] The collaborative execution unit is the physical interface for converting soft policies into hard actions, employing a dual-channel parallel drive mode. The specific execution steps are as follows: Channel 1: Inverting modulation of the pump speed vector. The system receives data containing amplitude components. Phase Lead Angle The system utilizes the space vector pulse width modulation (SVPWM) algorithm integrated within the motor vector control chip to dynamically compensate for pump speed. For the target amplitude, with For the target phase, an inverse torque pulsation signal is generated that is out of phase in the time domain compared to the predicted return water pulse waveform. It should be noted that generating the pulse width modulation signal for the drive motor using the SVPWM algorithm based on given vector parameters is a common existing technology in the field of variable frequency motor vector control; its underlying algorithm principle will not be elaborated here. This signal is directly injected into the motor drive circuit of the digital high-pressure pump. The drive motor generates a precise transient acceleration torque one microsecond before the return water jet impacts the valve seat, artificially creating a pressure trough to counteract the pressure peak of the return water in a "dynamic braking" manner. Channel Two: Impedance Mapping of Damping Coefficient. The system receives the return water damping adjustment coefficient. The system calls the opening-flow resistance characteristic lookup table stored in the controller ROM. This lookup table is preset and is obtained by performing a full-stroke test on the return water regulating mechanism, such as a proportional solenoid valve, during the equipment's factory calibration phase. The physical flow resistance values ​​under different drive signals are recorded, and an opening-flow resistance characteristic curve is generated and digitally stored. The system then uses this lookup table to select the dimensionless... Mapped to specific physical driving quantities, dimensionless The value ranges from 0 to 1. The mapping logic is as follows: when... When the maximum flow resistance in the lookup table corresponds to the physical opening being fully closed, the system output duty cycle is... PWM signal or Current signal; when When the minimum flow resistance is reached, the physical opening is fully open, and the system output duty cycle is... PWM signal or The current signal is used to calculate the corresponding virtual valve opening adjustment command, i.e., the specific voltage duty cycle or current value, using linear interpolation. The specific calculation steps are as follows: First, obtain the system's preset maximum flow resistance value. Calculate the difference between this maximum flow resistance value and the current return water damping adjustment coefficient, and divide this difference by the maximum flow resistance value to obtain the opening ratio coefficient. If a PWM signal is used for driving, multiply this opening ratio coefficient by 100% to obtain the specific voltage duty cycle. If a current signal is used for driving, multiply the opening ratio coefficient by the effective adjustment amount of 16 mA, and add a base of 4 mA to obtain the specific current value. For example, if the current return water damping adjustment coefficient is exactly half of the maximum flow resistance value, the calculated opening ratio coefficient is 50%. At this time, the system will output a PWM signal with a 50% duty cycle or a 12 mA current signal. This command drives the physical valve core to move, thereby physically reshaping the fluid resistance characteristics of the return water pipeline.

[0164] The residual feedback self-healing unit is used to monitor the pump inlet pressure residual after the execution of physical actions in real time. When the residual exceeds the preset quiescent threshold, the correction parameters are iteratively calculated based on the PID self-tuning algorithm, and the anti-phase torque pulsation signal and the virtual valve opening adjustment command are dynamically updated until the fluid pulsation characteristics disappear.

[0165] To acquire the pressure residual amplitude, within one detection cycle after the collaborative execution unit completes its action, the system calls the preceding fluid pulsation sensing module to acquire instantaneous pressure waveform data at the pump inlet. The system extracts the peak value from this waveform data and calculates the absolute value of the difference between it and the steady-state reference pressure. This difference is defined as the pressure residual. The physical meaning of this value lies in representing the remaining pressure fluctuation energy in the pipeline that was not completely eliminated after a dynamic pressure counter-pressure. Threshold judgment and PID self-tuning will use the calculated pressure residual... The noise level is compared with a preset silence threshold. The silence threshold is obtained by reading the standard deviation of the static background noise collected in step S2 when the system is stationary, and defining it as the minimum allowable physical noise floor of the system. Upon reaching the quiescent threshold, control is deemed successful, and the system has entered a Nash equilibrium steady state without water hammer. The current control parameters are then maintained unchanged. >The silence threshold is used to determine if there is residual ghost water hammer or overcompensated oscillation. The PID self-tuning algorithm is activated, and the system calculates the amplitude gain correction parameter based on the magnitude and sign of the pressure residual. With opening gain correction parameter The specific acquisition steps are as follows: directly read the current pressure residual obtained in the previous step as the instantaneous value, and retrieve the historical residual data recorded in the controller memory for several past adjustment cycles for accumulation, thereby obtaining the historical cumulative value; directly extract the positive and negative sign and absolute value of the current residual through mathematical judgment. In order to obtain two different correction parameters, two sets of independent control coefficients are pre-written into the controller memory. These two sets of preset proportional coefficients and integral coefficients are obtained and stored in advance during the equipment's factory commissioning stage through offline experimental calibration and step response testing of the pipeline system. When calculating the amplitude gain correction parameter, the system multiplies the instantaneous residual value by a preset amplitude scaling factor and the historical cumulative value by a preset amplitude integral factor. The sum of these two products is the specific amplitude gain correction parameter. Similarly, when calculating the opening gain correction parameter, the system multiplies the instantaneous residual value by a preset opening scaling factor and the historical cumulative value by a preset opening integral factor. The sum of these two products is the specific opening gain correction parameter. In this calculation process, the absolute value of the residual directly determines the adjustment range of the final calculated correction parameter, while the sign of the residual determines whether the correction parameter is increased positively or decreased negatively. The system performs closed-loop iterative updates. Channel one, which feeds back to the collaborative execution unit, is used to fine-tune the intensity of torque ripple; Channel two, which feeds back to the collaborative execution unit, is used to fine-tune the valve opening. This process is repeated until the monitored pressure residual is reached. When the system converges to the quiescent threshold range, it achieves adaptive healing, i.e., the fluid pulsation characteristics disappear.

[0166] The flexible coupling drive module solves the nonlinear error problem between the theoretical model and physical reality. The collaborative execution unit utilizes existing SVPWM technology and pre-calibrated Cv curves to ensure the executability and accuracy of control commands at the physical level. The residual feedback self-healing unit constructs a closed-loop negative feedback mechanism by real-time reading of the pressure waveform data from the fluid pulsation sensing module and comparing it with the silent threshold determined based on physical noise floor. Without this step, the system would be unable to perceive the execution effect, and parameter drift caused by equipment aging would render all previous complex game theory calculations invalid; only by introducing residual feedback can the system be ensured to remain locked at the true water hammer-free operating point throughout its entire lifecycle.

[0167] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0168] The foregoing description is illustrative of the invention and should not be construed as limiting it. Although several exemplary embodiments of the invention have been described, those skilled in the art will readily understand that many modifications can be made to the exemplary embodiments without departing from the novel teachings and advantages of the invention. Therefore, all such modifications are intended to be included within the scope of the invention as defined in the claims. It should be understood that the foregoing description is illustrative of the invention and should not be construed as limiting it to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The invention is defined by the claims and their equivalents.

Claims

1. A tank-type direct drinking water filtration system, comprising a raw water tank, a pretreatment filter, a digital high-pressure pump, a nanofiltration membrane module, an ozone ultraviolet sterilization device, a digital water tank, a user direct drinking faucet, a return water filter, and a return water regulating mechanism installed on the return water pipeline, characterized in that, The digital high-pressure pump integrates a fluid dynamic pressure balance control center, which includes: The fluid pulsation sensing module is used to extract features representing the game state between return water and inlet water from the raw pressure data at the pump inlet and construct a fluid dynamic fingerprint map. The jet countermeasure analysis module is used to decode the fluid dynamic fingerprint spectrum to quantify the physical intrusion degree of the return water jet and obtain the back pressure resistance index. The game-theoretic optimization decision module is used to perform control law optimization calculations on the back pressure stagnation index and solve for the pump speed dynamic compensation vector and the return water damping adjustment coefficient. The flexible coupling drive module is used to convert the pump speed dynamic compensation vector and the return water damping adjustment coefficient into the physical action of the fluid machinery, and perform closed-loop self-healing correction based on the execution effect. The step of constructing the fluid dynamics fingerprint spectrum includes: The collected instantaneous pressure waveform data is mapped to a one-dimensional equivalent pipeline space domain to construct a staged spatial envelope map containing the valve seat impact forbidden zone. Boolean clipping and consistency check based on physical envelope are performed on the staged spatial envelope map, and the valve disc and jet impact event sequence and rigid impact envelope segment are output. The physical quantities of the rigid impact envelope segment are integrated to calculate the pressure-time integral area of ​​a single impact event within the fluid micro-element, and the return water pulse momentum value is obtained in combination with the pipeline cross-sectional parameters. The distribution density of the valve disc and jet impact event sequence per unit time is statistically analyzed to generate the fluid intermittent stagnation duty cycle. The return water pulse momentum value and the fluid intermittent stagnation duty cycle are used as the vertical axis and horizontal axis, respectively, to construct a fluid dynamic fingerprint spectrum under multi-source confluence state. The steps for obtaining the back pressure retardation index include: The vertical axis features of the fluid dynamic fingerprint spectrum are analyzed, and the equivalent jet core velocity at the outlet of the return water pipeline is calculated in reverse from the momentum value of the return water pulse based on the inverse transformation of the momentum law. The horizontal axis features of the fluid dynamic fingerprint spectrum are analyzed, and the theoretical geometric flow section of the raw water inlet is corrected by the fluid intermittent retardation duty cycle. The dynamic effective influent flux under high-frequency interference of the return water jet is calculated. Calculate the momentum ratio between the equivalent jet core velocity and the main flow velocity of the raw water, and weight couple the momentum ratio with the dynamic effective influent flux to output the back pressure retardation index; The steps for calculating the pump speed dynamic compensation vector and the return water damping adjustment coefficient are as follows: A non-cooperative game model of hydrodynamic pressure between the inlet and return water flow paths is constructed. The back pressure resistance index is used as the conflict penalty term in the game, and the Nash equilibrium point is calculated with the goal of minimizing the total turbulent kinetic energy at the pump inlet. The numerical coordinates of the Nash equilibrium point are compared and verified with the preset hardware security threshold range to determine the optimal operating point. Based on the deviation between the current operating conditions and the optimal operating point, a dynamic compensation vector for pump speed and a return water damping adjustment coefficient are generated. The steps for closed-loop self-healing correction based on execution results are as follows: Signal modulation and command mapping are performed on the pump speed dynamic compensation vector and the return water damping adjustment coefficient respectively to generate an anti-phase torque pulsation signal to drive the digital high-pressure pump motor, and to generate a virtual valve opening adjustment command to drive the return water adjustment mechanism. The pump inlet pressure residual after physical action is monitored in real time. When the residual exceeds the preset quiescent threshold, the correction parameters are iteratively calculated based on the PID self-tuning algorithm, and the anti-phase torque pulsation signal and the virtual valve opening adjustment command are dynamically updated until the fluid pulsation characteristics disappear.

2. The tank-type direct drinking water filtration system according to claim 1, characterized in that, The steps for constructing a staged spatial envelope map that includes the valve seat impact no-entry zone are as follows: Acquire instantaneous pressure waveform data at the inlet of a digital high-pressure pump, construct a one-dimensional equivalent pipeline coordinate system based on the length of the physical pipeline and the propagation speed of the pressure wave, project the time dimension features of the instantaneous pressure waveform data into the physical position features of the pipeline, and output a pressure echo spatial mapping map. The pressure echo spatial mapping is received, and the valve seat impact forbidden zone is divided according to the physical structure of the pump inlet confluence node. The valve seat impact forbidden zone is given corresponding geometric boundary constraints, and a staged spatial envelope map with physical criterion boundaries is output.

3. The tank-type direct drinking water filtration system according to claim 2, characterized in that, The steps of the output valve disc, jet impact event sequence, and rigid impact envelope segment are as follows: Spatial Boolean operation is performed on the staged spatial envelope graph to extract waveform segments that fall into the valve seat impact forbidden zone, and background noise interference is removed using a sudden change amplitude threshold to output a geometric boundary candidate impact set.

4. The tank-type direct drinking water filtration system according to claim 3, characterized in that, The steps of the output valve disc and the jet impact event sequence and rigid impact envelope segment also include: The acoustic reflection path is verified on the geometric boundary candidate impact set to verify whether the round-trip propagation delay of the candidate impact at the pipeline physical location conforms to the physical law, eliminate artifact interference, and output the valve disc and jet impact event sequence and rigid impact envelope segment.

5. A tank-type direct drinking water filtration system according to claim 1, characterized in that, The digital high-pressure pump adopts an electromechanical integrated structure. The fluid dynamic pressure balance control center is integrated into the motor drive end of the digital high-pressure pump and uploads the operating data to the cloud platform through the Internet of Things module. The fluid dynamic pressure balance control center also includes a user water consumption prediction module. The user water consumption prediction module is based on a user behavior learning algorithm and automatically predicts the user's water consumption for the next period based on historical data fed back from the cloud platform.

6. A tank-type direct drinking water filtration system according to claim 1, characterized in that, The digital water tank is a fully enclosed pressure tank structure, with a flexible diaphragm or air bladder inside to physically isolate the stored purified water from the outside air; the return pipeline where the return water filter is located uses the outlet residual pressure generated by the digital high-pressure pump during water production or circulation to drive the water flow to complete the circulation and keep it alive.

7. A tank-type direct drinking water filtration system according to claim 1, characterized in that, The system is configured with a pump-tank joint scheduling mode: when the user's drinking water tap is in peak water usage, the digital water tank releases the internally stored pressurized water and operates in parallel with the digital high-pressure pump to jointly supply water to the user.

Images