Intelligent control system and method for seawater desalination ultrafiltration reverse osmosis device

Abstract

This invention discloses an intelligent control system and method for a seawater desalination ultrafiltration reverse osmosis device, belonging to the technical field of seawater desalination. It includes the following steps: collecting seawater temperature, pH value, and conductivity; collecting the output water volume of the reverse osmosis device; recording the collected data before and after membrane flushing under different backwash water pressures and chemical dosages as historical data; setting a standard data range; preprocessing the historical data; training an LSTM neural network model based on qualified scheme data to obtain a backwash control model; and setting the maximum viscosity, maximum conductivity, and minimum output water volume; when entering a jellyfish outbreak period, comparing viscosity, conductivity, and output water volume, the backwash control model generates a predicted backwash water pressure and predicted chemical dosage based on the collected real-time data; and performing membrane flushing based on the predicted backwash water pressure and predicted chemical dosage to ensure that the ultrafiltration reverse osmosis device maintains the maximum water flow rate while ensuring safe operation.

Description

Technical Field

[0001] This invention relates to the field of seawater desalination, and in particular to an intelligent control system and method for a seawater desalination ultrafiltration reverse osmosis device. Background Technology

[0002] Currently, ultrafiltration and reverse osmosis (RO) systems are the two mainstream equipment for seawater desalination, with ultrafiltration often serving as a pretreatment unit for RO. Ultrafiltration systems require lower operating pressures, only 0.1-0.3 MPa, while RO systems require higher pressures, reaching 5.5-8.0 MPa. In addition to these two types of equipment, seawater desalination ultrafiltration / RO systems generally include other pretreatment units, pressurization systems, backwashing systems, and post-treatment units. During seawater desalination, the internal membrane of the ultrafiltration / RO system gradually becomes clogged by salt and other impurities, requiring backwashing or membrane replacement. During jellyfish blooms, jellyfish secrete mucus that creates colloids in the seawater. These colloids and the jellyfish themselves are easily broken up by other pretreatment units or filters at the inlet of the ultrafiltration / RO system, reaching the inlet and causing fouling of the membrane. This not only affects the water flow efficiency of the ultrafiltration / RO system but also pollutes the filtered water.

[0003] The existing technical solutions mentioned above have the following drawbacks: during jellyfish outbreaks, ultrafiltration reverse osmosis devices will need to perform membrane flushing more frequently and require the addition of chemicals for sterilization. However, frequent membrane flushing and chemical addition will affect the membrane life and water quality. Summary of the Invention

[0004] To reduce damage to the filter membrane and water quality during the flushing process of the ultrafiltration reverse osmosis device during jellyfish outbreaks, this application provides an intelligent control system and method for a seawater desalination ultrafiltration reverse osmosis device.

[0005] On the one hand, the intelligent control method for a seawater desalination ultrafiltration reverse osmosis device provided in this application adopts the following technical solution: A smart control method for a seawater desalination ultrafiltration reverse osmosis device includes the following steps: The viscosity, temperature, pH value, and conductivity of seawater at the inlet of the ultrafiltration unit are collected; the viscosity, temperature, pH value, and conductivity of seawater at the outlet of the reverse osmosis unit are collected; and the output water volume of the reverse osmosis unit is collected. Data collected before and after backwashing of the filter membrane under different backwash water pressures and chemical dosages are recorded as historical data. Set a standard data range, preprocess historical data, and select data sets that fall within the standard data range after filter membrane rinsing as qualified solution data; Set up an LSTM neural network model, train the LSTM neural network model based on qualified scheme data, and obtain a backwash control model; Set the maximum viscosity, maximum conductivity, and minimum output flow rate; When the jellyfish bloom period begins, the viscosity and conductivity of the seawater at the outlet of the reverse osmosis unit are compared with the maximum viscosity and maximum conductivity, respectively, and the output flow rate of the reverse osmosis unit is compared with the minimum output flow rate. When the viscosity of the seawater at the outlet of the reverse osmosis unit exceeds the maximum viscosity, the conductivity exceeds the maximum conductivity, or the outflow rate is less than the minimum outflow rate, the backwash control model generates the expected backwash water pressure and the expected dosage based on the collected real-time data, and performs filter membrane flushing based on the expected backwash water pressure and the expected dosage.

[0006] By adopting the above scheme, the system generates a backwash control model based on local historical data. During jellyfish outbreaks, the backwash control system calculates the minimum backwash water pressure and chemical dosage to achieve standard backwashing results, minimizing the impact of the filter membrane flushing process on the filter membrane and water quality. Because this system generates the model based on local data, the control accuracy is higher, and it is more adaptable to complex environments.

[0007] Preferably, the following steps are also included: Adjust the pressure at the inlet of the ultrafiltration unit and the reverse osmosis unit, and record the pressure value, the data collected at the inlet of the ultrafiltration unit, the data collected at the outlet of the reverse osmosis unit, and the output water volume of the reverse osmosis unit; Set up an LSTM neural network model. The LSTM neural network model estimates the impact of each of the following data on the output water volume: viscosity, water temperature, pH value, and conductivity of seawater at the inlet of the ultrafiltration device and viscosity, water temperature, pH value, and conductivity of seawater at the outlet of the reverse osmosis device. The model then generates a data impact curve. Based on the data impact curve of each data item and the recorded data, the LSTM neural network model is retrained using the XGBoost machine learning model until the LSTM neural network model can calculate the same water output as the recorded data based on any combination of data, thus obtaining an automatic data processing model. The automated data processing model then generates a data impact curve based on the recorded data. Calculate the data at the point of highest water volume based on the data impact curve; The automatic control model is trained based on the calculated data and data influence curves at the highest water output. The automatic control model controls the pressure at the inlet of the ultrafiltration device and the reverse osmosis device based on the currently collected data on the viscosity, temperature, pH value, and conductivity of the seawater at the inlet of the ultrafiltration device and the viscosity, temperature, pH value, and conductivity of the seawater at the outlet of the reverse osmosis device, so as to maximize the water output and calculate the impact of changes in various data on the water output.

[0008] By adopting the above scheme, before the ultrafiltration reverse osmosis unit operates, the system, after simple experiments, generates a data impact curve based on the surrounding environment using an LSTM neural network model. This model then generates a dedicated automatic control model, which automatically adjusts the water pressure to control the output water flow. This allows the ultrafiltration reverse osmosis unit's output water flow to be automatically and flexibly controlled according to the surrounding seawater conditions, ensuring that the unit maintains the maximum flow rate while ensuring safe operation. Simultaneously, the system calculates the impact of various seawater data changes on the output water flow. Users can view historical output water flow data and predicted impact values ​​to determine the approximate influence of the environment on the ultrafiltration reverse osmosis unit and estimate future output water flow.

[0009] Preferably, the step of "maximizing the water output and calculating the impact of changes in various data on the water output" further includes: Based on the data influence curve and the current water output, calculate the changes in water output after the changes in the values ​​of seawater temperature, pH value, and conductivity at the inlet of the ultrafiltration unit and the seawater temperature, pH value, and conductivity at the outlet of the reverse osmosis unit, and generate a water output change curve with the current water output as the origin. To obtain future trends in the seawater surrounding the ultrafiltration reverse osmosis unit; Based on the outflow change curve of various data and the future trend of seawater change, calculate the outflow within a set time period in the future, and generate an outflow prediction curve. The pressure at the inlet of the ultrafiltration and reverse osmosis units is adjusted in advance based on the predicted water output curve.

[0010] By adopting the above solution, users can more easily predict future changes in water output based on the water output change curve. Because changing the water pressure takes time, pre-control can adjust the water pressure in time to better protect the ultrafiltration reverse osmosis unit.

[0011] Preferably, the following steps are also included: Preset maximum difference; Calculate the current required drainage volume based on the control results of the automatic control model, and compare the current required drainage volume with the actual collected current drainage volume; If the actual current drainage volume is lower than the current required drainage volume, and the difference between the current required drainage volume and the actual current drainage volume is greater than the maximum difference, then the ultrafiltration reverse osmosis unit is judged to have malfunctioned and an alarm is issued.

[0012] By adopting the above solution, the system can automatically determine whether the ultrafiltration reverse osmosis device is malfunctioning. When the ultrafiltration reverse osmosis device experiences a significant drop in water output due to reasons such as filter membrane blockage or water pressure failing to reach the expected value, the system will automatically issue an alarm to remind the user to perform backwashing or maintenance.

[0013] Preferably, the following steps are also included: Preset pressure warning limits and maximum pressure limits for the filter membrane; Compare the water pressure at the inlet of the reverse osmosis unit with the pressure warning limit and the maximum pressure limit; When the water pressure at the inlet of the reverse osmosis unit exceeds the pressure warning limit but is less than the maximum pressure limit, timing begins. When the water pressure at the inlet of the reverse osmosis unit is less than the pressure warning limit, timing stops. When the timing reaches the set time, the water pressure at the inlet of the reverse osmosis unit is forcibly reduced and a warning signal is issued, and the automatic control model is suspended. When the water pressure at the inlet of the reverse osmosis unit exceeds the maximum pressure limit, the water pressure at the inlet of the reverse osmosis unit is forcibly reduced and an alarm signal is issued, suspending the operation of the automatic control model.

[0014] By adopting the above solution, the filter membrane itself has a pressure limit. The system can automatically determine whether the filter membrane is in a state of extreme pressure for a long time or whether the water pressure exceeds the pressure limit based on the current water pressure. It can reduce the water pressure in time to avoid damage to the filter membrane caused by working under high water pressure for a long time.

[0015] On the other hand, the intelligent control system for a seawater desalination ultrafiltration reverse osmosis device provided in this application adopts the following technical solution: An intelligent control system for a seawater desalination ultrafiltration reverse osmosis device includes a data acquisition module, a data storage module, an experimental setup module, a model generation module, a water treatment module, a model training module, a model control module, and an automatic control module. The data acquisition module collects the water temperature, pH value, TDS value, and conductivity of the seawater at the inlet of the ultrafiltration device, the water temperature, pH value, TDS value, and conductivity of the seawater at the outlet of the reverse osmosis device, and the water output of the reverse osmosis device, and transmits the collected data to the data storage module. The data storage module receives and stores data; The experimental setup module calls the data storage module to collect data before and after filter membrane rinsing under different backwash water pressures and dosages, and records it as historical data. It sets a standard data range, preprocesses the historical data, and selects the data group after filter membrane rinsing that is within the standard data range as qualified scheme data. The model generation module sets up an LSTM neural network model, calls up the qualified scheme data from the experimental setup module, trains the LSTM neural network model based on the qualified scheme data, and obtains a backwashing control model. The water treatment module is preset with maximum viscosity, maximum conductivity, and minimum effluent flow rate. It calls the backwash control model of the model generation module. Upon receiving the instruction, it compares the viscosity and conductivity of the seawater at the outlet of the reverse osmosis device with the maximum viscosity and conductivity, respectively, and compares the effluent flow rate of the reverse osmosis device with the minimum effluent flow rate. When the viscosity of the seawater at the outlet of the reverse osmosis device exceeds the maximum viscosity, the conductivity exceeds the maximum conductivity, or the effluent flow rate is less than the minimum effluent flow rate, the backwash control model generates the expected backwash water pressure and expected chemical dosage based on the collected real-time data, and performs filter membrane flushing according to the expected backwash water pressure and expected chemical dosage.

[0016] By adopting the above scheme, the system generates a backwash control model based on local historical data. During jellyfish outbreaks, the backwash control system calculates the minimum backwash water pressure and chemical dosage to achieve standard backwashing results, minimizing the impact of the filter membrane flushing process on the filter membrane and water quality. Because this system generates the model based on local data, the control accuracy is higher, and it is more adaptable to complex environments.

[0017] Preferably, it also includes a model training module, a model control module, and an automatic control module; After receiving the instruction, the experimental setup module adjusts the pressure at the inlet of the ultrafiltration device and the reverse osmosis device according to the instruction, and calls the pressure value stored in the data storage module during the process, the data collected at the inlet of the ultrafiltration device, the data at the outlet of the reverse osmosis device, and the water output of the reverse osmosis device, and transmits them to the model training module. The model training module sets up an LSTM neural network model. The LSTM neural network model estimates the impact of each of the following data on the output water volume: the water temperature, pH value, and conductivity of the seawater at the inlet of the ultrafiltration device, and the water temperature, pH value, and conductivity of the seawater at the outlet of the reverse osmosis device. It then generates a data impact curve. The LSTM neural network model is retrained based on the data impact curve of each data item and the recorded data until the LSTM neural network model can calculate the same output water volume as the recorded data based on any combination of data, thus obtaining an automatic data processing model. The model control module calls the automatic data processing model trained by the model training module, and the automatic data processing model generates a data influence curve based on the recorded data. Based on the data influence curve, it calculates the data at the highest water volume, and trains the automatic control model based on the calculated data at the highest water volume and the data influence curve. The automatic control module calls upon the automatic control model trained by the model control module and the latest stored data from the data storage module, including the seawater temperature, pH value, and conductivity at the inlet of the ultrafiltration device and the seawater temperature, pH value, and conductivity at the outlet of the reverse osmosis device, to control the pressure at the inlet of the ultrafiltration device and the reverse osmosis device. The automatic control model controls the pressure at the inlet of the ultrafiltration device and the reverse osmosis device based on the currently collected data, so as to maximize the output water volume, and calculates the impact of changes in various data on the output water volume.

[0018] By adopting the above scheme, before the ultrafiltration reverse osmosis unit operates, the system, after simple experiments, generates a data impact curve based on the surrounding environment using an LSTM neural network model. This model then generates a dedicated automatic control model, which automatically adjusts the water pressure to control the output water flow. This allows the ultrafiltration reverse osmosis unit's output water flow to be automatically and flexibly controlled according to the surrounding seawater conditions, ensuring that the unit maintains the maximum flow rate while ensuring safe operation. Simultaneously, the system calculates the impact of various seawater data changes on the output water flow. Users can view historical output water flow data and predicted impact values ​​to determine the approximate influence of the environment on the ultrafiltration reverse osmosis unit and estimate future output water flow.

[0019] Preferably, it also includes a pre-adjustment module, wherein the automatic control module calculates the change in water output after the changes in the values ​​of seawater temperature, pH value, conductivity at the inlet of the ultrafiltration device and seawater temperature, pH value, and conductivity at the outlet of the reverse osmosis device, based on the data influence curve and the current water output, and generates a water output change curve with the current water output as the origin. The data acquisition module acquires the future trend of seawater changes around the ultrafiltration reverse osmosis device and transmits it to the data storage module. The pre-adjustment module calls the water output change curve of the automatic control module and the future seawater change trend stored in the data storage module. Based on the water output change curve of various data changes and the future seawater change trend, it calculates the water output within a set time period in the future and generates a water output prediction curve. Based on the water output prediction curve, it adjusts the pressure at the inlet of the ultrafiltration device and the reverse osmosis device in advance.

[0020] By adopting the above solution, users can more easily predict future changes in water output based on the water output change curve. Because changing the water pressure takes time, pre-control can adjust the water pressure in time to better protect the ultrafiltration reverse osmosis unit.

[0021] Preferably, it also includes a fault alarm module, wherein the automatic control module calculates the current required drainage volume based on the control results of the automatic control model; The fault alarm module is preset with a maximum difference. The fault alarm module calls the current required drainage volume of the automatic control module and the latest current drainage volume stored in the data storage module, compares the current required drainage volume with the actual collected current drainage volume. If the actual collected current drainage volume is lower than the current required drainage volume, and the difference between the current required drainage volume and the actual collected current drainage volume is greater than the maximum difference, then it is determined that the ultrafiltration reverse osmosis device has malfunctioned and an alarm is issued.

[0022] By adopting the above solution, the system can automatically determine whether the ultrafiltration reverse osmosis device is malfunctioning. When the ultrafiltration reverse osmosis device experiences a significant drop in water output due to reasons such as filter membrane blockage or water pressure failing to reach the expected value, the system will automatically issue an alarm to remind the user to perform backwashing or maintenance.

[0023] Preferably, the system also includes a water pressure limit module. This module presets a pressure warning limit and a maximum pressure limit for the filter membrane. The water pressure limit module retrieves the latest water pressure at the inlet of the reverse osmosis device from the data storage module and compares it with the pressure warning limit and the maximum pressure limit. When the water pressure at the inlet exceeds the pressure warning limit but is less than the maximum pressure limit, a timer is started. When the water pressure at the inlet is less than the pressure warning limit, the timer stops. When the timer reaches the set time, the water pressure at the inlet of the reverse osmosis device is forcibly reduced, a warning signal is issued, and the automatic control module is suspended. When the water pressure at the inlet of the reverse osmosis device exceeds the maximum pressure limit, the water pressure at the inlet of the reverse osmosis device is forcibly reduced, an alarm signal is issued, and the automatic control module is suspended.

[0024] By adopting the above solution, the filter membrane itself has a pressure limit. The system can automatically determine whether the filter membrane is in a state of extreme pressure for a long time or whether the water pressure exceeds the pressure limit based on the current water pressure. It can reduce the water pressure in time to avoid damage to the filter membrane caused by working under high water pressure for a long time.

[0025] In summary, the present invention has the following beneficial effects: 1. During jellyfish outbreaks, the backwash control system calculates the minimum backwash water pressure and dosage to achieve the standard backwash effect, minimizing the impact of the filter membrane flushing process on the filter membrane and water quality.

[0026] 2. The system also calculates the impact of various seawater data changes on the output water volume. By viewing historical output water volume data and the expected impact, users can determine the approximate impact of the environment on the ultrafiltration reverse osmosis device and estimate the future output water volume. Attached Figure Description

[0027] Figure 1 This is a flowchart of Embodiment 1 of this application.

[0028] Figure 2 This is an overall module block diagram of Embodiment 2 of this application.

[0029] Explanation of reference numerals in the attached figures: 1. Data acquisition module; 2. Data storage module; 3. Experiment setup module; 4. Model training module; 5. Model control module; 6. Automatic control module; 7. Pre-adjustment module; 8. Fault alarm module; 9. Water pressure limit module; 10. Model generation module; 11. Water treatment module. Detailed Implementation

[0030] Example 1: This application discloses an intelligent control method for a seawater desalination ultrafiltration reverse osmosis device, such as... Figure 1 As shown, the specific steps are as follows: S100, preset maximum difference, pressure warning limit of the filter membrane, and maximum pressure limit.

[0031] S101. Collect the seawater temperature, pH value, and conductivity at the inlet of the ultrafiltration unit. Collect the seawater temperature, pH value, and conductivity at the outlet of the reverse osmosis unit. Collect the effluent flow rate of the reverse osmosis unit. Obtain the future trend of seawater changes around the ultrafiltration and reverse osmosis units.

[0032] S200: Record the data collected before and after backwashing of the filter membrane under different backwash water pressures and dosages as historical data.

[0033] S201. Set a standard data range, preprocess the historical data, and select the data group whose data after filter membrane rinsing is within the standard data range as qualified solution data.

[0034] S202. Set up an LSTM neural network model and train the LSTM neural network model based on the qualified scheme data to obtain the backwashing control model.

[0035] S203, set the maximum viscosity, maximum conductivity and minimum output water volume.

[0036] S300. When entering the jellyfish outbreak period, compare the viscosity and conductivity of the seawater at the outlet of the reverse osmosis unit with the maximum viscosity and maximum conductivity, respectively, and compare the output of the reverse osmosis unit with the minimum output.

[0037] S301. When the viscosity of the seawater at the outlet of the reverse osmosis unit exceeds the maximum viscosity, the conductivity exceeds the maximum conductivity, or the water flow rate is less than the minimum water flow rate, the backwash control model generates the expected backwash water pressure and the expected dosage based on the collected real-time data, and performs filter membrane flushing based on the expected backwash water pressure and the expected dosage.

[0038] S400. Adjust the pressure at the inlet of the ultrafiltration unit and the reverse osmosis unit, and record the pressure value, the data collected at the inlet of the ultrafiltration unit, the data collected at the outlet of the reverse osmosis unit, and the output water volume of the reverse osmosis unit.

[0039] S500. Set up an LSTM neural network model. The LSTM neural network model estimates the impact of each of the following data on the output water volume: the water temperature, pH value, and conductivity of the seawater at the inlet of the ultrafiltration device and the water temperature, pH value, and conductivity of the seawater at the outlet of the reverse osmosis device. It then generates a data impact curve.

[0040] S501. Retrain the LSTM neural network model based on the data influence curve of each data item and the recorded data until the LSTM neural network model can calculate the same water output as the recorded data based on any combination of data.

[0041] S502. The trained LSTM neural network model generates a data impact curve again based on the recorded data.

[0042] S503. Calculate the data at the point of maximum water volume based on the data influence curve.

[0043] S600: An automatic control model is set up. Based on the calculated data and data influence curves at the peak water output, the automatic control model is trained. It controls the pressure at the inlets of the ultrafiltration and reverse osmosis units according to the currently collected seawater temperature, pH, and conductivity at the inlet of the ultrafiltration unit, and the seawater temperature, pH, and conductivity at the outlet of the reverse osmosis unit, maximizing the water output. The model also calculates the impact of changes in these data on the water output. Based on the data influence curves and the current water output, the changes in water output after variations in the seawater temperature, pH, and conductivity at the inlet of the ultrafiltration unit and the outlet of the reverse osmosis unit are calculated, and a water output change curve is generated with the current water output as the origin. Users can more easily predict future water output changes based on this curve.

[0044] S601. Calculate the outflow volume within a set time period based on the outflow volume change curve of various data and the future trend of seawater change, and generate an outflow volume prediction curve.

[0045] S602. Adjust the inlet pressure of the ultrafiltration and reverse osmosis units in advance based on the predicted water flow curve. Because changing the water pressure takes time, pre-control can adjust the water pressure in time to better protect the ultrafiltration and reverse osmosis units.

[0046] S700: Calculate the current required drainage volume based on the control results of the automatic control model, and compare the current required drainage volume with the actual collected current drainage volume.

[0047] S701. If the actual current drainage volume is lower than the required drainage volume, and the difference between the required drainage volume and the actual drainage volume is greater than the maximum difference, the ultrafiltration reverse osmosis unit is judged to have malfunctioned, and an alarm is issued. The system can automatically determine whether the ultrafiltration reverse osmosis unit has malfunctioned. When the ultrafiltration reverse osmosis unit experiences a significant drop in output water due to reasons such as membrane clogging or water pressure failing to reach the expected value, the system will automatically issue an alarm to remind the user to perform backwashing or maintenance.

[0048] S800: Compare the water pressure at the inlet of the reverse osmosis unit with the pressure warning limit and the maximum pressure limit.

[0049] S801. When the water pressure at the inlet of the reverse osmosis unit exceeds the pressure warning limit but is less than the maximum pressure limit, timing begins. When the water pressure at the inlet of the reverse osmosis unit is less than the pressure warning limit, timing stops. When the timing reaches the set time, the water pressure at the inlet of the reverse osmosis unit is forcibly reduced and a warning signal is issued, pausing the operation of the automatic control model.

[0050] S802. When the water pressure at the inlet of the reverse osmosis unit exceeds the maximum pressure limit, the system forcibly reduces the water pressure at the inlet of the reverse osmosis unit, issues an alarm signal, and suspends the operation of the automatic control model. The filter membrane itself has a pressure limit. The system can automatically determine whether the filter membrane has been in a limit state for a long time or whether the water pressure has exceeded the pressure limit based on the current water pressure, and can reduce the water pressure in time to avoid damage to the filter membrane caused by working under high water pressure for a long time.

[0051] The implementation principle of the intelligent control method for a seawater desalination ultrafiltration reverse osmosis device in this application embodiment is as follows: The system generates a backwash control model based on local historical data. During jellyfish outbreaks, the backwash control system calculates the minimum backwash water pressure and chemical dosage to achieve the standard backwash effect, minimizing the impact of the filter membrane flushing process on the filter membrane and water quality. Because this system generates the model based on local data, the control accuracy is higher and it can better adapt to complex environments.

[0052] Before the ultrafiltration reverse osmosis unit begins operation, the system, after simple experiments, uses an LSTM neural network model to generate a data impact curve based on the surrounding environment. This model then generates a dedicated automatic control model, which automatically adjusts the water pressure to control the output flow. This allows the ultrafiltration reverse osmosis unit's output flow to be automatically and flexibly controlled according to the surrounding seawater conditions, ensuring the unit maintains maximum flow rate while ensuring safe operation. Simultaneously, the system calculates the impact of various seawater data changes on the output flow. By viewing historical output flow data and projected impacts, users can assess the approximate environmental influence on the ultrafiltration reverse osmosis unit and estimate future output flow.

[0053] Example 2: This application discloses an intelligent control system for a seawater desalination ultrafiltration reverse osmosis device, such as... Figure 2 As shown, it includes a data acquisition module 1, a data storage module 2, an experimental setup module 3, a model training module 4, a model control module 5, an automatic control module 6, a pre-adjustment module 7, a fault alarm module 8, a water pressure limit module 9, a model generation module 10, and a water quality treatment module 11.

[0054] like Figure 2 As shown, data acquisition module 1 collects the seawater temperature, pH value, and conductivity at the inlet of the ultrafiltration unit, the seawater temperature, pH value, and conductivity at the outlet of the reverse osmosis unit, and the effluent flow rate of the reverse osmosis unit. The collected data is then transmitted to data storage module 2. Data acquisition module 1 also acquires future trends in the seawater surrounding the ultrafiltration and reverse osmosis units and transmits this data to data storage module 2. Data storage module 2 receives and stores the data.

[0055] like Figure 2 As shown, the experimental setup module 3 calls the data storage module 2 to collect data before and after filter membrane rinsing under different backwash water pressures and dosages and records it as historical data. It sets a standard data range, preprocesses the historical data, and selects the data group whose data after filter membrane rinsing is within the standard data range as qualified scheme data. like Figure 2 As shown, the model generation module 10 sets up the LSTM neural network model, calls the qualified scheme data of the experimental setting module (3), trains the LSTM neural network model according to the qualified scheme data, and obtains the backwashing control model. like Figure 2 As shown, the water treatment module 11 is preset with maximum viscosity, maximum conductivity, and minimum effluent flow rate. It calls the backwash control model from the model generation module 10. Upon receiving a command, it compares the viscosity and conductivity of the seawater at the reverse osmosis unit outlet with the maximum viscosity and conductivity, and compares the effluent flow rate of the reverse osmosis unit with the minimum effluent flow rate. When the viscosity or conductivity of the seawater at the reverse osmosis unit outlet exceeds the maximum viscosity or conductivity, or the effluent flow rate is less than the minimum effluent flow rate, the backwash control model generates a predicted backwash water pressure and a predicted chemical dosage based on real-time data. The filter membrane is then flushed according to the predicted backwash water pressure and predicted chemical dosage. The system generates a backwash control model based on local historical data. During jellyfish outbreaks, the backwash control system calculates the minimum backwash water pressure and chemical dosage to achieve the standard flushing effect, minimizing the impact of the filter membrane flushing process on the filter membrane and water quality. Because this system generates a model based on local data, the control accuracy is higher, and it is more adaptable to complex environments.

[0056] like Figure 2As shown, after receiving the instruction, the experimental setup module 3 adjusts the pressure at the inlet of the ultrafiltration device and the reverse osmosis device according to the instruction, and calls the pressure value stored in the data storage module 2, the data collected at the inlet of the ultrafiltration device, the data at the outlet of the reverse osmosis device, and the water output of the reverse osmosis device, and transmits them to the model training module 4.

[0057] like Figure 2 As shown, model training module 4 sets up an LSTM neural network model. The LSTM neural network model estimates the influence of each of the following data on the output water volume: the water temperature, pH value, and conductivity of the seawater at the inlet of the ultrafiltration device, and the water temperature, pH value, and conductivity of the seawater at the outlet of the reverse osmosis device. It also generates a data influence curve. The LSTM neural network model is retrained based on the data influence curve of each data item and the recorded data until the LSTM neural network model can calculate the same output water volume as the recorded data based on any combination of data.

[0058] like Figure 2 As shown, the model control module 5 calls the LSTM neural network model trained by the model training module 4, and the trained LSTM neural network model generates a data influence curve based on the recorded data. Based on the data influence curve, it calculates the data when the water volume is at its highest. Based on the calculated data when the water volume is at its highest and the data influence curve, it trains the automatic control model.

[0059] like Figure 2 As shown, the automatic control module 6 calls the automatic control model trained by the model control module 5 and the latest stored data (temperature, pH, conductivity) of seawater at the inlet of the ultrafiltration unit and the outlet of the reverse osmosis unit (temperature, pH, conductivity) in the data storage module 2 to control the pressure at the inlets of the ultrafiltration and reverse osmosis units. The automatic control model controls the pressure at the inlets of the ultrafiltration and reverse osmosis units based on the currently collected data to maximize the water output and calculates the impact of changes in various data on the water output. The automatic control module 6 calculates the change in water output after changes in the values ​​of the seawater temperature, pH, conductivity at the inlet of the ultrafiltration unit and the outlet of the reverse osmosis unit, based on the data influence curve and the current water output, and generates a water output change curve with the current water output as the origin. The automatic control module 6 calculates the current required discharge volume based on the control results of the automatic control model.

[0060] like Figure 2As shown, the pre-adjustment module 7 calls upon the water flow change curve from the automatic control module 6 and the future seawater trend stored in the data storage module 2. Based on the water flow change curve and the future seawater trend, it calculates the water flow within a set time period and generates a water flow prediction curve. The pressure at the inlet of the ultrafiltration and reverse osmosis devices is adjusted in advance based on this prediction curve. Changing the water pressure takes time; pre-control allows for timely pressure adjustments to better protect the ultrafiltration and reverse osmosis devices.

[0061] like Figure 2 As shown, the fault alarm module 8 has a preset maximum difference. The fault alarm module 8 calls the current required drainage volume from the automatic control module 6 and the latest stored current drainage volume from the data storage module 2, comparing the current required drainage volume with the actual collected current drainage volume. If the actual collected current drainage volume is lower than the current required drainage volume, and the difference between the current required drainage volume and the actual collected current drainage volume is greater than the maximum difference, then the ultrafiltration reverse osmosis unit is judged to have malfunctioned, and an alarm is issued. The system can automatically determine whether the ultrafiltration reverse osmosis unit has malfunctioned. When the ultrafiltration reverse osmosis unit experiences a significant drop in output water due to reasons such as filter membrane clogging or water pressure failing to reach the expected value, the system will automatically issue an alarm to remind the user to perform backwashing or maintenance.

[0062] like Figure 2 As shown, the water pressure limit module 9 presets the pressure warning limit and the maximum pressure limit for the filter membrane. The water pressure limit module 9 retrieves the latest water pressure at the inlet of the reverse osmosis unit stored in the data storage module 2 and compares it with the pressure warning limit and the maximum pressure limit. When the water pressure at the inlet of the reverse osmosis unit exceeds the pressure warning limit but is less than the maximum pressure limit, a timer is started. When the water pressure at the inlet of the reverse osmosis unit is less than the pressure warning limit, the timer stops. When the timer reaches the set time, the water pressure at the inlet of the reverse osmosis unit is forcibly reduced and a warning signal is issued, pausing the operation of the automatic control module 6. When the water pressure at the inlet of the reverse osmosis unit exceeds the maximum pressure limit, the water pressure at the inlet of the reverse osmosis unit is forcibly reduced and an alarm signal is issued, pausing the operation of the automatic control module 6. The filter membrane itself has a pressure limit. The system can automatically determine whether the filter membrane has been in a state of extreme pressure for a long time or whether the water pressure has exceeded the pressure limit based on the current water pressure, and can reduce the water pressure in time to avoid damage to the filter membrane caused by prolonged operation under high water pressure.

[0063] The embodiments described herein are preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape, and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A smart control method for a seawater desalination ultrafiltration reverse osmosis device, characterized in that, Includes the following steps: The viscosity, temperature, pH value, and conductivity of seawater at the inlet of the ultrafiltration unit are collected; the viscosity, temperature, pH value, and conductivity of seawater at the outlet of the reverse osmosis unit are collected; and the output water volume of the reverse osmosis unit is collected. Data collected before and after backwashing of the filter membrane under different backwash water pressures and chemical dosages are recorded as historical data. Set a standard data range, preprocess historical data, and select data sets that fall within the standard data range after filter membrane rinsing as qualified solution data; Set up an LSTM neural network model, train the LSTM neural network model based on qualified scheme data, and obtain a backwash control model; Set the maximum viscosity, maximum conductivity, and minimum output flow rate; When the jellyfish bloom period begins, the viscosity and conductivity of the seawater at the outlet of the reverse osmosis unit are compared with the maximum viscosity and maximum conductivity, respectively, and the output flow rate of the reverse osmosis unit is compared with the minimum output flow rate. When the viscosity of the seawater at the outlet of the reverse osmosis unit exceeds the maximum viscosity, the conductivity exceeds the maximum conductivity, or the outflow rate is less than the minimum outflow rate, the backwash control model generates the expected backwash water pressure and the expected dosage based on the collected real-time data, and performs filter membrane flushing based on the expected backwash water pressure and the expected dosage. Adjust the pressure at the inlet of the ultrafiltration unit and the reverse osmosis unit, and record the pressure value, the data collected at the inlet of the ultrafiltration unit, the data collected at the outlet of the reverse osmosis unit, and the output water volume of the reverse osmosis unit; Set up an LSTM neural network model. The LSTM neural network model estimates the impact of each of the following data on the output water volume: viscosity, temperature, pH value, and conductivity of seawater at the inlet of the ultrafiltration device and viscosity, temperature, pH value, and conductivity of seawater at the outlet of the reverse osmosis device. The model then generates a data impact curve. Based on the data impact curve of each data item and the recorded data, the LSTM neural network model is retrained using the XGBoost machine learning model until the LSTM neural network model can calculate the same water output as the recorded data based on any combination of data, thus obtaining an automatic data processing model. The automated data processing model then generates a data impact curve based on the recorded data. Calculate the data at the point of highest water volume based on the data impact curve; The automatic control model is trained based on the calculated data and data influence curves at the highest water output. The automatic control model controls the pressure at the inlet of the ultrafiltration device and the reverse osmosis device based on the currently collected data on the viscosity, temperature, pH value, and conductivity of the seawater at the inlet of the ultrafiltration device and the viscosity, temperature, pH value, and conductivity of the seawater at the outlet of the reverse osmosis device, so as to maximize the water output and calculate the impact of changes in various data on the water output.

2. The intelligent control method for a seawater desalination ultrafiltration reverse osmosis device according to claim 1, characterized in that, The step of "maximizing the water output and calculating the impact of changes in various data on the water output" also includes: Based on the data influence curve and the current water output, calculate the changes in water output after the changes in the values ​​of seawater temperature, pH value, and conductivity at the inlet of the ultrafiltration unit and the seawater temperature, pH value, and conductivity at the outlet of the reverse osmosis unit, and generate a water output change curve with the current water output as the origin. To obtain future trends in the seawater surrounding the ultrafiltration reverse osmosis unit; Based on the outflow change curve of various data and the future trend of seawater change, calculate the outflow within a set time period in the future, and generate an outflow prediction curve. The pressure at the inlet of the ultrafiltration and reverse osmosis units is adjusted in advance based on the predicted water output curve.

3. The intelligent control method for a seawater desalination ultrafiltration reverse osmosis device according to claim 1, characterized in that, It also includes the following steps: Preset maximum difference; Calculate the current required drainage volume based on the control results of the automatic control model, and compare the current required drainage volume with the actual collected current drainage volume; If the actual current drainage volume is lower than the current required drainage volume, and the difference between the current required drainage volume and the actual current drainage volume is greater than the maximum difference, then the ultrafiltration reverse osmosis unit is judged to have malfunctioned and an alarm is issued.

4. The intelligent control method for a seawater desalination ultrafiltration reverse osmosis device according to claim 1, characterized in that, It also includes the following steps: Preset pressure warning limits and maximum pressure limits for the filter membrane; Compare the water pressure at the inlet of the reverse osmosis unit with the pressure warning limit and the maximum pressure limit; When the water pressure at the inlet of the reverse osmosis unit exceeds the pressure warning limit but is less than the maximum pressure limit, timing begins. When the water pressure at the inlet of the reverse osmosis unit is less than the pressure warning limit, timing stops. When the timing reaches the set time, the water pressure at the inlet of the reverse osmosis unit is forcibly reduced and a warning signal is issued, and the automatic control model is suspended. When the water pressure at the inlet of the reverse osmosis unit exceeds the maximum pressure limit, the water pressure at the inlet of the reverse osmosis unit is forcibly reduced and an alarm signal is issued, suspending the operation of the automatic control model.

5. An intelligent control system for a seawater desalination ultrafiltration reverse osmosis device, characterized in that: It includes a data acquisition module (1), a data storage module (2), an experimental setup module (3), a model generation module (10), a water treatment module (11), a model training module (4), a model control module (5), and an automatic control module (6); The data acquisition module (1) collects the water temperature, pH value and conductivity of the seawater at the inlet of the ultrafiltration device, collects the water temperature, pH value and conductivity of the seawater at the outlet of the reverse osmosis device, collects the water output of the reverse osmosis device, and transmits the collected data to the data storage module (2). The data storage module (2) receives and stores data; The experimental setup module (3) calls the data storage module (2) to collect data before and after filter membrane rinsing under different backwash water pressures and dosages and records it as historical data. It sets a standard data range, preprocesses the historical data, and selects the data group whose data after filter membrane rinsing is within the standard data range as qualified scheme data. The model generation module (10) sets up an LSTM neural network model, calls up the qualified scheme data of the experimental setting module (3), trains the LSTM neural network model based on the qualified scheme data, and obtains the backwash control model. The water treatment module (11) is preset with maximum viscosity, maximum conductivity and minimum effluent. It calls the backwash control model of the model generation module (10). After receiving the instruction, it compares the viscosity and conductivity of the seawater at the outlet of the reverse osmosis device with the maximum viscosity and maximum conductivity, and compares the effluent of the reverse osmosis device with the minimum effluent. When the viscosity of the seawater at the outlet of the reverse osmosis device exceeds the maximum viscosity, the conductivity exceeds the maximum conductivity or the effluent is less than the minimum effluent, the backwash control model generates the expected backwash water pressure and expected dosage based on the collected real-time data, and performs filter membrane flushing based on the expected backwash water pressure and expected dosage. After receiving the instruction, the experimental setup module (3) adjusts the pressure of the inlet of the ultrafiltration device and the reverse osmosis device according to the instruction, and calls the pressure value stored in the data storage module (2) during the process, the data collected at the inlet of the ultrafiltration device, the data at the outlet of the reverse osmosis device, and the water output of the reverse osmosis device, and transmits them to the model training module (4). The model training module (4) sets up an LSTM neural network model. The LSTM neural network model estimates the influence of each of the following data on the output water volume: the water temperature, pH value, and conductivity of the seawater at the inlet of the ultrafiltration device and the water temperature, pH value, and conductivity of the seawater at the outlet of the reverse osmosis device, based on the recorded data, and generates a data influence curve. The LSTM neural network model is retrained based on the data influence curve of each data item and the recorded data until the LSTM neural network model can calculate the same output water volume as the recorded data based on any combination of data, thus obtaining an automatic data processing model. The model control module (5) calls the automatic data processing model trained by the model training module (4), and the automatic data processing model generates a data influence curve based on the recorded data. Based on the data influence curve, it calculates the data when the water volume is at its highest, and trains the automatic control model based on the calculated data and the data influence curve when the water volume is at its highest. The automatic control module (6) calls the automatic control model trained by the model control module (5) and the latest stored data of seawater temperature, pH value, conductivity at the inlet of the ultrafiltration device and seawater temperature, pH value, conductivity at the outlet of the reverse osmosis device to control the pressure at the inlet of the ultrafiltration device and the reverse osmosis device. The automatic control model controls the pressure at the inlet of the ultrafiltration device and the reverse osmosis device based on the currently collected seawater temperature, pH value, conductivity at the inlet of the ultrafiltration device and seawater temperature, pH value, conductivity at the outlet of the reverse osmosis device to maximize the output water volume, and calculates the impact of changes in various data on the output water volume.

6. The intelligent control system for a seawater desalination ultrafiltration reverse osmosis device according to claim 5, characterized in that: It also includes a pre-adjustment module (7), wherein the automatic control module (6) calculates the changes in water temperature, pH value, conductivity of seawater at the inlet of the ultrafiltration device and the changes in water temperature, pH value, conductivity of seawater at the outlet of the reverse osmosis device after the changes in the values ​​of these values, based on the data influence curve and the current water output, and generates a water output change curve with the current water output as the origin. The data acquisition module (1) acquires the future trend of seawater changes around the ultrafiltration reverse osmosis device and transmits it to the data storage module (2); The pre-adjustment module (7) calls the water output change curve of the automatic control module (6) and the future seawater change trend stored in the data storage module (2), calculates the water output within a set time period based on the water output change curve of various data changes and the future seawater change trend, and generates a water output prediction curve. Based on the water output prediction curve, the pressure at the inlet of the ultrafiltration device and the reverse osmosis device is adjusted in advance.

7. The intelligent control system for a seawater desalination ultrafiltration reverse osmosis device according to claim 5, characterized in that: It also includes a fault alarm module (8), and the automatic control module (6) calculates the current water discharge volume based on the control results of the automatic control model; The fault alarm module (8) is preset with a maximum difference. The fault alarm module (8) calls the current required drainage volume of the automatic control module (6) and the latest current drainage volume stored in the data storage module (2), compares the current required drainage volume with the actual collected current drainage volume. If the actual collected current drainage volume is lower than the current required drainage volume, and the difference between the current required drainage volume and the actual collected current drainage volume is greater than the maximum difference, then it is determined that the ultrafiltration reverse osmosis device has malfunctioned and an alarm is issued.

8. The intelligent control system for a seawater desalination ultrafiltration reverse osmosis device according to claim 5, characterized in that: It also includes a water pressure limit module (9), which presets the pressure warning limit and the pressure maximum limit of the filter membrane. The water pressure limit module (9) calls the latest water pressure at the inlet of the reverse osmosis device stored in the data storage module (2), compares the water pressure at the inlet of the reverse osmosis device with the pressure warning limit and the pressure maximum limit. When the water pressure at the inlet of the reverse osmosis device exceeds the pressure warning limit but is less than the pressure maximum limit, it starts timing. When the water pressure at the inlet of the reverse osmosis device is less than the pressure warning limit, it stops timing. When the timing reaches the set time, it forcibly reduces the water pressure at the inlet of the reverse osmosis device and issues a warning signal, suspending the operation of the automatic control module (6). When the water pressure at the inlet of the reverse osmosis device exceeds the pressure maximum limit, it forcibly reduces the water pressure at the inlet of the reverse osmosis device and issues an alarm signal, suspending the operation of the automatic control module (6).

Images