Intelligent control and cleaning method of two-stage reverse osmosis purified water system

Abstract

The invention relates to the technical field of purified water systems, and discloses an intelligent control and cleaning method of a two-stage reverse osmosis purified water system, which comprises the following steps: step 1, acquiring water temperature, pH value, conductivity and instrument data of a water inlet and a water outlet in the purified water preparation process; 2, comparing the obtained data of each instrument with a set data reference and fluctuation range, and judging whether the data of each instrument is abnormal or not; and step 3, when the instrument data is abnormal, obtaining estimated correct data of the abnormal instrument according to the normal and abnormal instrument data and the historical data. According to the method, whether part of instrument data is abnormal or not is judged through various data in the pure water preparation process, when the instrument data is abnormal, estimated correct data is obtained according to association characteristics of an abnormal instrument and other normal data and the obtained various data, and then the obtained estimated correct data is verified.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of purified water systems, in particular to an intelligent control and cleaning method for a two-stage reverse osmosis purified water system. BACKGROUND

[0002] The two-stage reverse osmosis purified water system is a water treatment device that uses a series double-stage RO membrane separation technology to deeply desalt and purify raw water. After pretreatment, the raw water is first removed about 97% of the dissolved solids by the first-stage reverse osmosis, and then further desalted by the second-stage reverse osmosis to produce high-purity water. The system is composed of pretreatment, first-stage and second-stage RO units, post-treatment, and automatic control, and has the characteristics of high water purity, stable operation, high intelligence, and is widely used in industries with strict water quality requirements such as seawater desalination, medicine, and food. When the two-stage reverse osmosis purified water system is working, if the conductivity meter, pH meter, etc. are out of calibration, it is not possible to accurately determine whether the produced water meets the standards. In this case, continuing to operate the system may cause unqualified water to enter the subsequent use link, affecting product quality or causing other losses. If the flow, pressure, or temperature sensors are out of calibration or have abnormal data, the system can continue to work, but the out-of-calibration or abnormal data is more dependent on experience and lacks a certain degree of accuracy. Therefore, the present application proposes an intelligent control and cleaning method for a two-stage reverse osmosis purified water system to solve the deficiencies in the prior art. SUMMARY

[0003] The purpose of the present application is to provide an intelligent control and cleaning method for a two-stage reverse osmosis purified water system to solve the deficiencies in the background art.

[0004] The purpose of the present application can be achieved by the following technical solutions: An intelligent control method for a two-stage reverse osmosis purified water system, comprising the following steps: Step one, obtaining the water temperature, pH value, conductivity, and instrument data of the water inlet and outlet in the current purified water preparation process; Step two, comparing the obtained instrument data with the set data reference and fluctuation range to determine whether the instrument data is abnormal; Step three, when the instrument data is abnormal, obtaining the estimated correct data of the abnormal instrument according to the normal abnormal instrument data and historical data; Step four, verifying the obtained estimated correct data, and recording the estimated correct data after verification Preferably, the method for working in step two is: Retrieve multiple sets of historical data from the recorded historical data as reference historical data for the current purified water preparation. According to the retrieved reference historical data, the data benchmark and fluctuation range of each instrument are constructed; The instrument data obtained this time is compared with the data benchmark and fluctuation range of each instrument; The instrument data exceeding or being lower than the data benchmark and fluctuation range of the instrument is marked as abnormal instrument data.

[0005] Preferably, the process of retrieving reference historical data is: The dosing amount of each chemical agent in the current purified water preparation process is obtained; The dosing amount ratio is obtained by comparing the dosing amount of each chemical agent in the current purified water preparation process with the dosing amount of each chemical agent recorded in the historical data; The historical data with the dosing amount ratio within the set dosing amount ratio interval is marked as selected historical data; The recovery rate of each set of historical data in the selected historical data is obtained, and the recovery rate of the current purified water preparation is compared with the recovery rate of each set of historical data in the selected historical data; The historical data with the recovery rate ratio within the set recovery rate ratio interval is selected as reference historical data.

[0006] Preferably, the method for working in step three is: The selected multiple sets of reference historical data are classified, the monitoring data corresponding to each instrument are obtained, and the target data of the current abnormal instrument is matched with the same target data in the reference historical data; The target data in the reference historical data and the monitoring data of other instruments are statistically analyzed to obtain the correlation characteristics of the target data in the reference historical data and the monitoring data of other instruments, and the normal change rule of the target data under different states of other data is obtained according to the correlation characteristics; The monitoring data of each normal instrument in the current purified water preparation is obtained, the correlation characteristics of the target data in the reference historical data and the monitoring data of other normal instruments are obtained, the initial estimated data of the abnormal instrument monitoring data is obtained by combining the monitoring data of each normal instrument in this time and through a pre-set correlation matching method; The target data corresponding to the multiple sets of reference historical data in the reference historical data that match the matching interval of the normal instrument data of the current purified water preparation is obtained, and the reference comparison numerical range is obtained; The numerical interval of the initial estimated data is determined according to the reference comparison numerical range, the initial estimated data obtained this time is compared with the numerical interval, and it is judged whether the initial estimated data is within the normal range; If the initial estimated data is within the numerical interval, it is determined as the estimated correct data of the abnormal instrument; If the initial estimated data is outside the numerical range, re-filter historical data and repeat the above steps until a reasonable estimated data is obtained.

[0007] Preferably, the process of obtaining the initial estimated data of abnormal instrument monitoring data through a preset association matching method is as follows: From normal instrument monitoring data, filter out normal correlation data for matching; Extract the corresponding normal correlation data and target data from each set of historical reference data; Based on the association content between the normal association data and the target data record in the reference historical data, obtain the reference records whose similarity to the current normal association data exceeds the similarity threshold, and select at least three groups of high similarity reference records based on the number of qualified records; Statistical calculations are performed on the target data in the high-similarity reference records to obtain the initial estimated values ​​of the abnormal instrument monitoring data.

[0008] Preferably, the similarity acquisition method is as follows: Based on the degree of influence of each related data on the target data, obtain the weight coefficient of each related data item; Obtain the deviation value of each normal correlation data point in this purified water preparation process from the normal correlation data point in the reference historical data; Multiply the deviation value of each normally associated data item by its corresponding weight coefficient to obtain the weighted deviation value of the normally associated data, and then calculate the weighted deviation value of all associated data items. Calculate the total weighted deviation between each set of historical reference data and the current associated normal data, compare the total weighted deviation with the preset basic deviation threshold, and select the reference data sets whose total deviation does not exceed the threshold. The similarity score is calculated based on the ratio of the total weighted deviation to the basic deviation threshold. The similarity score is then compared with the preset qualified similarity threshold to obtain a high similarity reference record.

[0009] Preferably, the method for performing step four is as follows: Compare the predicted data with the data benchmark and fluctuation range to determine whether the predicted data is within a reasonable range; If the estimated data is within a reasonable range, then based on the correspondence between the target data and other related data in the historical data, the change logic between the estimated correct data and the correct data of this purified water preparation is matched with the correspondence to determine whether the estimated data conforms to the data change correlation pattern. The average value and fluctuation range of the target data are obtained by statistically analyzing and calculating the target data in the historical reference data. The difference between the predicted correct data and the average value is calculated, and the difference is compared with the fluctuation range. Based on the comparison results, it is determined whether the deviation between the predicted correct data and the historical normal statistical level is within the target data deviation range. If the deviation between the predicted correct data and the historical normal statistical level is within the target data deviation range, the predicted correct data will be recorded according to time, abnormal instrument type, and monitoring data. If the deviation between the predicted correct data and the historical normal statistical level is outside the target data deviation range, return to step three; or if the predicted data is not within a reasonable range, obtain the predicted correct data again.

[0010] A cleaning method for a two-stage reverse osmosis purified water system includes the following steps: Step 1: Open the valve according to the second-level trial operation, and use the cleaning pump to pump clean, chlorine-free reverse osmosis product water from the cleaning tank into the pressure vessel of the membrane housing and discharge it for a few minutes. Step 2: Prepare the cleaning solution, open the secondary RO cleaning valve and start the cleaning pump, and control the inlet water pressure to be lower than the preset pressure; Step 3: Circulate the cleaning fluid in the pressure vessel for a pre-set time. Step 4: Record the pH value change and compare it with the initial value. If the pH value drops significantly, the cleaning solution should be prepared again. If the pH value does not change significantly, it proves that the cleaning is complete. Step 5: After cleaning the cleaning pump and heat exchanger with clean water, empty the cleaning tank and rinse it clean. Then fill it with clean product water for the next rinsing step. Step 6: Open the valves as normal, turn on the raw water pump manually, and use the pump to pump the clean, chlorine-free product water into the pressure vessel and discharge it. Step 7: After rinsing the reverse osmosis system, follow the valve opening and closing procedures during the secondary trial run until the product water is clean, free of foam or cleaning agent.

[0011] The beneficial effects of this invention are: 1. This invention determines whether some instrument data is abnormal by analyzing various data in the pure water preparation process. When the instrument data is abnormal, it obtains the estimated correct data based on the correlation characteristics between the abnormal instrument and other normal data, as well as the acquired data. Then, the estimated correct data is verified. If the estimated data passes the verification, it can replace the recorded abnormal data, realizing the intelligent dynamic decision-making of the system. This effectively solves the problem of relying on experience or backup instruments to obtain correct data when some instruments are abnormal.

[0012] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0013] 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. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a flowchart illustrating the steps of an intelligent control method for a two-stage reverse osmosis purified water system according to the present invention.

[0015] Figure 2 This is a flowchart illustrating the steps of a cleaning method for a two-stage reverse osmosis purified water system according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Please see Figure 1 As shown, this invention is an intelligent control method for a two-stage reverse osmosis purified water system. The two-stage reverse osmosis purified water system mainly includes a pretreatment system, two-stage reverse osmosis devices, a post-treatment system, a cleaning system, and an electrical control system. The pretreatment system mainly includes a raw water pump, a dosing device, a quartz sand filter, an activated carbon filter, and a precision filter, used to reduce the pollution index, residual chlorine, hardness, and other impurities in the raw water, and to protect the reverse osmosis membrane elements. The two-stage reverse osmosis devices mainly include a first-stage reverse osmosis device and a second-stage reverse osmosis device. The post-treatment system includes an ultraviolet sterilizer and a precision filter, used to kill microorganisms and trap particles to ensure sterile effluent. The cleaning system includes a cleaning tank, a filter, and a cleaning pump, used to periodically clean the reverse osmosis membrane and restore its performance. The electrical control system includes an integrated PLC automatic control and a touch screen, realizing equipment start-up and shutdown, parameter monitoring, fault alarm, and data recording, supporting manual or automatic modes. When using a two-stage reverse osmosis purified water system to desalinate seawater and obtain freshwater resources, monitoring instruments are used to obtain the water temperature, pH value, and conductivity at the inlet and outlet. At the same time, the two-stage reverse osmosis purified water system also records the monitoring data of various monitoring instruments in real time. The monitoring instruments mainly include flow meters, pressure meters, and product water conductivity meters, so as to comprehensively grasp the current working status of the purified water preparation process and avoid judgment errors caused by missing data.

[0018] After collecting real-time data from each instrument, the system automatically compares the recorded instrument data with the data benchmarks and fluctuation ranges set according to system characteristics, historical normal operation data, and process standards. This allows for the rapid identification of instrument data that exceeds or falls below reasonable ranges. Once abnormal instrument data is identified, the abnormal instrument and its data are marked. Compared to traditional manual inspection and judgment, this system offers advantages in automation and accuracy, significantly improving judgment efficiency. When abnormal instrument data is detected, the system obtains the expected correct data for the abnormal instrument based on the normal instrument data already confirmed in the current process and the historical normal operation data accumulated by the system. After obtaining the expected correct data, it is verified. Once the expected correct data passes verification, it is recorded, including the corresponding time node, the type of abnormal instrument, and the current operating condition parameters.

[0019] The data benchmarks and fluctuation ranges of each instrument are mainly constructed based on historical data. Historical data that is more consistent with the current preparation is selected from the recorded historical data. First, the dosage of various chemical agents in the current process is obtained. The chemical agents mainly include flocculants, scale inhibitors, pH adjusters, and defluorinators. The dosage of these agents is related to the stability of the feed water quality and the operating efficiency of the reverse osmosis membrane. For example, insufficient scale inhibitor dosage will lead to scaling of the membrane module, which will affect the measurement data of the feed water conductivity and pressure instruments. The dosage of pH adjuster will change the feed water pH value, which will affect the monitoring results of the pH instrument and conductivity instrument. After obtaining the dosage of each chemical agent in this purified water preparation, the dosage of each chemical agent in this preparation is compared one by one with the dosage of each group of chemical agents recorded in the historical database. The dosage ratio of each group of historical data to the dosage in this preparation is calculated, that is, the ratio of historical dosage to current dosage. The dosage ratio is then compared with the preset dosage ratio range. The dosage ratio range is mainly set based on the long-term operating experience of the system and is a range of ±10% of the dosage of the same type of purified water preparation. Historical data with ratios within the range are marked as candidate historical data. After obtaining the candidate historical data, the recovery rate in purified water preparation is further compared. The recovery rate is expressed as a percentage, which is the ratio of the product water flow rate to the total influent flow rate. The recovery rate of this purified water preparation is obtained based on the ratio of the product water flow rate to the influent flow rate. Then, the recovery rate corresponding to each historical data is extracted from the marked candidate historical data. By comparing the current recovery rate with the recovery rate of the candidate historical data, the recovery rate ratio is calculated. Subsequently, the obtained recovery rate ratio is compared with the preset recovery rate ratio range, which is ±5% of the recovery rate of the same type of purified water preparation. Finally, historical data with a recovery rate ratio within the recovery rate ratio range are selected as reference historical data. If the current recovery rate is 75%, candidate data with a recovery rate between 70% and 80% are selected.

[0020] After selecting multiple sets of historical reference data, the average value and fluctuation range of the same data are calculated based on multiple sets of monitoring data of the same instrument. In this purified water preparation, the data obtained by the monitoring instrument can be compared with the average value and fluctuation range to identify abnormal instruments.

[0021] After identifying the abnormal instrument, the abnormal data monitored by the corresponding abnormal instrument cannot be used. Therefore, it is necessary to estimate the target data that the abnormal instrument should have monitored based on the correlation between other normal instruments and the abnormal instrument, as well as the water temperature, pH value, and conductivity data of the inlet and outlet in this purified water preparation process. That is, to obtain the estimated correct data.

[0022] The accurate data is obtained by classifying and matching the selected sets of historical reference data with the target data. Since the instruments in the secondary reverse osmosis system are diverse and the data monitored by different instruments correspond to different process meanings, such as temperature instruments reflecting the thermal state of the water, conductivity instruments reflecting water purity, and pressure instruments reflecting the system operating load, it is necessary to classify the historical reference data according to the instrument function and the data monitored. For example, the influent temperature and effluent temperature in all reference data are classified as temperature data, the influent conductivity and product water conductivity are classified as conductivity data, the influent pressure, membrane pressure, and product water pressure are classified as pressure data, and the influent flow rate, product water flow rate, and concentrate flow rate are classified as flow rate data, etc.

[0023] After classifying multiple sets of historical reference data, the target data category for the predicted correct data is determined based on the required data category. For example, if the anomaly is in the feed water temperature instrument, the target data category is feed water temperature data within the temperature category; if the anomaly is in the product water conductivity instrument, the target data category is product water conductivity data within the conductivity category. The target data category corresponding to the anomaly instrument is matched with data of the same category in all historical reference data. For example, the target data of feed water temperature is extracted from each set of historical reference data, forming a correspondence list between historical reference data sets and target data. The correlation characteristics between the target data and other instrument monitoring data in the historical reference data are obtained. There are process correlations between various parameters in the secondary reverse osmosis system. For example, changes in feed water temperature will affect the viscosity and ion activity of water, thus affecting the measurement results of the conductivity instrument; changes in feed water pressure will change the membrane flux, thus affecting the product water flow rate and product water conductivity; changes in pH value will affect the charge state of the membrane, thus affecting the membrane's ion rejection rate. Therefore, based on the target data category, the correspondence between the target data and other types of normal instrument data in all historical reference data can be statistically analyzed. For example, if the target data is inlet water temperature, then the corresponding changes in inlet water temperature for each set of reference data when inlet water conductivity, inlet water pH, and inlet water pressure change can be statistically analyzed: the range of inlet water temperature change when inlet water conductivity increases from 200 to 250, the fluctuation range of inlet water temperature when inlet water pH increases from 6.0 to 7.0, and the trend of inlet water temperature change when inlet water pressure increases from 0.8 to 1.0. Statistical analysis of multiple sets of reference data revealed the normal variation patterns of the target data under different conditions of other data: for example, when the influent conductivity increases from 200 to 300, the influent pH value increases from 6.0 to 7.0, and the influent pressure increases from 0.8 to 1.0, the normal range of the influent temperature is 24 to 26 degrees Celsius, and it increases slightly with the increase of influent pressure. The increase is positively correlated with the increase in pressure, with the temperature increasing by 0.1 degrees Celsius for every 0.1 degree Celsius increase. The correlation characteristics between the target data and other data were summarized from the actual operating data under similar historical conditions.

[0024] After obtaining the correlation characteristics between the target data and other data, normal correlation data were screened from the normal instrument data of this purified water preparation process. Then, the corresponding normal correlation data and target data were extracted from each group of reference historical data to construct a reference correlation data and reference target data control group.

[0025] After constructing the reference correlation data and the reference target data control group, weighting coefficients were set according to the degree of influence of each normal correlation data on the target data. For example, in the correlation data of product water conductivity, the feed water conductivity directly determines the ion content of the raw water and has the greatest impact on the product water conductivity. Therefore, the weighting coefficient of feed water conductivity was set to 0.4. Feed water temperature indirectly affects product water conductivity by affecting ion activity. The weighting coefficient of feed water temperature was set to 0.25. Feed water pressure indirectly affects product water conductivity by affecting membrane flux. The weighting coefficient of feed water pressure was set to 0.2. Although the membrane rejection rate has a significant impact, the weighting coefficient of membrane rejection rate was set to 0.15 because historical average values ​​were used. Then, the deviation value between each normal correlation data point and the corresponding correlation data in each set of reference records is calculated. For example, if the current inlet water conductivity is 245, and the inlet water conductivity of a certain set of reference records is 250, the deviation value is 5; if the current inlet water temperature is 24.8, and the inlet water temperature of the reference record is 25, the deviation value is 0.2; if the current inlet water pressure is 0.88, and the inlet water pressure of the reference record is 0.9, the deviation value is 0.02. Each deviation value is multiplied by its corresponding weighting coefficient to obtain the weighted deviation value. For example, the weighted deviation value for inlet water conductivity is 5 × 0.4 = 2, and the weighted deviation value for inlet water temperature is 0.2 × 0.4 = 2. 0.25 = 0.05, the weighted deviation value of the inlet pressure is 0.02 × 0.2 = 0.004. Adding all weighted deviation values ​​together: 2 + 0.05 + 0.004 = 2.054, we obtain the total weighted deviation of this group of reference records and the normally correlated data: 2.054. This is then compared with the preset basic deviation threshold, which is uniformly set to 5. Reference data groups whose total deviation does not exceed the threshold are selected. The similarity score is calculated using the formula: similarity = 1 - (total weighted deviation / basic deviation threshold). Therefore, the similarity score for this group is... The score is 58.92%, and it is compared with a preset acceptable similarity threshold of 80%. Reference records with scores greater than or equal to 80% are selected. At least three sets of high-similarity reference records must be selected. For example, four sets of reference records with similarities of 82%, 0.85%, 0.83%, and 0.81% are ultimately selected, corresponding to target data of 0.8, 0.9, 1.0, and 0.7 respectively. After obtaining the high-similarity reference records, statistical calculations are performed on the target data in these records to obtain the initial estimated value of the abnormal instrument monitoring data. The initial estimated value is the average target data of the high-similarity reference records. For example, the average target data of the above four sets of reference records is... ,but This is the initial estimated data when the pressure gauge malfunctions.

[0026] After obtaining the initial estimated data, the initial estimated data is compared with the numerical range to determine whether the initial estimated data is within the normal range. If the initial estimated data is within the numerical range, it is determined to be the correct estimated data of the abnormal instrument. If the initial estimated data is outside the numerical range, the reference historical data is re-filtered and the above steps are repeated until reasonable estimated data is obtained. The minimum and maximum values ​​of the numerical range are the minimum and maximum values ​​of the selected reference historical data target data, respectively. For example, if the minimum and maximum values ​​of the reference historical data target data are 3.1 and 3.3, then the above initial estimated data of 3.2 is within the numerical range and is marked as the correct estimated data. After obtaining the correct estimated data, the correct estimated data is compared with the established data benchmark and fluctuation range. If the correct estimated data is within the fluctuation range, the change logic of the correct estimated data and the correct data of this purified water preparation is matched with the corresponding relationship based on the change relationship of the target data and other related data in the historical data. It is then judged whether the correct estimated data conforms to the data change correlation law. For example, if the abnormal instrument is the product water flow meter and the estimated product water flow is 2.0, based on the normal inlet water pressure and inlet water temperature this time, the corresponding relationship of the product water flow usually between 1.9 and 2.1 when the pressure is 0.9 and the temperature is 25 in the historical law is compared. If the change logic of the correct estimated data is consistent with the actual state of the related parameters, then a historical statistical deviation comparison is further performed. Statistical analysis is performed on historical reference data to calculate the average value and fluctuation range of the target data. For example, if the influent temperature values ​​of 10 sets of historical reference data are 28, 30, 31, 29, 32, 30, 27, 31, 29, and 30, the average value is 29.7, and the fluctuation range is -2.7 to +2.3. Then, the difference between the current estimated correct data and this average value is calculated. For example, if the estimated conductivity of the produced water is 30, the difference is 0.3. This difference is compared with the fluctuation range to determine whether the estimated data is within the normal historical statistical deviation range. Following the example above, if the difference of 0.3 is between -2.7 and +2.3, the current estimated correct data is used as the monitoring data and recorded according to time, abnormal instrument type, and monitoring data, directly replacing the abnormal data recorded by the abnormal instrument.

[0027] After the purified water preparation is completed, the secondary reverse osmosis purified water needs to be cleaned and the instrument data needs to be adjusted.

[0028] Please see Figure 2 The present invention also provides a cleaning method for a two-stage reverse osmosis purified water system, comprising the following steps: Step 1: Open the valve according to the second-level trial operation, and use the cleaning pump to pump clean, chlorine-free reverse osmosis product water from the cleaning tank into the pressure vessel of the membrane housing and discharge it for a few minutes. Step 2: Prepare the cleaning solution, open the secondary RO cleaning valve and start the cleaning pump, and control the inlet water pressure to be lower than the preset pressure; Step 3: Circulate the cleaning fluid in the pressure vessel for a pre-set time. Step 4: Record the pH value change and compare it with the initial value. If the pH value drops significantly, the cleaning solution should be prepared again. If the pH value does not change significantly, it proves that the cleaning is complete. Step 5: After cleaning the cleaning pump and heat exchanger with clean water, empty the cleaning tank and rinse it clean. Then fill it with clean product water for the next rinsing step. Step 6: Open the valves as normal, turn on the raw water pump manually, and use the pump to pump the clean, chlorine-free product water into the pressure vessel and discharge it. Step 7: After rinsing the reverse osmosis system, follow the valve opening and closing procedures during the secondary trial run until the product water is clean, free of foam or cleaning agent.

[0029] The above description is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in the claims, they should all fall within the protection scope of the present invention.

Claims

1. A smart control method for a two-stage reverse osmosis purified water system, characterized in that, Includes the following steps: Step 1: Obtain the water temperature, pH value, conductivity, and instrument data at the inlet and outlet of the purified water preparation process. Step two: Compare the acquired data from each instrument with the set data benchmark and fluctuation range to determine whether there are any abnormalities in the data from each instrument. Step 3: When the instrument data is abnormal, obtain the predicted correct data of the abnormal instrument based on the normal abnormal instrument data and historical data; Step 4: Verify the obtained accurate prediction data, and record the accurate prediction data after successful verification.

2. The intelligent control and cleaning method for a two-stage reverse osmosis purified water system according to claim 1, characterized in that, The method for performing step two is as follows: Multiple sets of historical data were retrieved from the recorded historical data as reference historical data for this purified water preparation; Based on the retrieved historical reference data, construct the data benchmark and fluctuation range for each instrument; The data obtained from each instrument were compared with the data benchmarks and fluctuation ranges of each instrument. Instrument data that exceeds or falls below the instrument data baseline and fluctuation range will be marked as abnormal instrument data.

3. The intelligent control method for a two-stage reverse osmosis purified water system according to claim 2, characterized in that, The process of retrieving historical reference data is as follows: Obtain the dosage of each chemical reagent in this purified water preparation process; The dosage of each chemical agent in this purified water preparation process is compared with the dosage of each chemical agent recorded in the historical data to obtain the dosage ratio. Historical data with dosage ratios within the set dosage ratio range are marked as candidate historical data; Obtain the recovery rate of each set of historical data in the candidate historical data, and compare the recovery rate of this purified water preparation with the recovery rate of each set of historical data in the candidate historical data; Historical data with recovery ratios within the set recovery ratio range were selected as reference historical data.

4. The intelligent control method for a two-stage reverse osmosis purified water system according to claim 1, characterized in that, The method for performing step three is as follows: The selected sets of historical reference data are classified, the monitoring data corresponding to each instrument is obtained, and the target data corresponding to the current abnormal instrument is matched with the same target data in the historical reference data. Statistical analysis is performed on the target data in the reference historical data and the monitoring data of other instruments to obtain the correlation characteristics between the target data in the reference historical data and the monitoring data of other instruments. Based on the correlation characteristics, the normal variation pattern of the target data under different states of other data is obtained. Acquire the monitoring data of each normal instrument in this purified water preparation. Based on the correlation characteristics of the target data obtained from the historical data and the monitoring data of other normal instruments, and combined with the monitoring data of each normal instrument in this preparation, obtain the initial estimated data of the abnormal instrument monitoring data through a preset correlation matching method. Obtain target data corresponding to multiple sets of reference historical data that match the normal instrument data of this purified water preparation in the reference historical data, and obtain the reference comparison value range; The numerical range of the initial estimated data is determined by comparing the reference numerical range. The initial estimated data obtained in this study is then compared with the numerical range to determine whether the initial estimated data is within the normal range. If the initial estimated data is within the numerical range, it is determined to be the correct estimated data of the abnormal instrument; If the initial estimated data is outside the numerical range, re-filter historical data and repeat the above steps until a reasonable estimated data is obtained.

5. The intelligent control method for a two-stage reverse osmosis purified water system according to claim 4, characterized in that, The process of obtaining the initial estimated data of abnormal instrument monitoring data through a preset correlation matching method is as follows: From normal instrument monitoring data, filter out normal correlation data for matching; Extract the corresponding normal correlation data and target data from each set of historical reference data; Based on the association content between the normal association data and the target data record in the reference historical data, obtain the reference records whose similarity to the current normal association data exceeds the similarity threshold, and select at least three groups of high similarity reference records based on the number of qualified records. Statistical calculations are performed on the target data in the high-similarity reference records to obtain the initial estimated values ​​of the abnormal instrument monitoring data.

6. The intelligent control method for a two-stage reverse osmosis purified water system according to claim 5, characterized in that, The similarity acquisition method is as follows: Based on the degree of influence of each related data on the target data, obtain the weight coefficient of each related data item; Obtain the deviation value of each normal correlation data point in this purified water preparation process from the normal correlation data point in the reference historical data; Multiply the deviation value of each normally associated data item by its corresponding weight coefficient to obtain the weighted deviation value of the normally associated data, and then calculate the weighted deviation value of all associated data items. Calculate the total weighted deviation between each set of historical reference data and the current associated normal data, compare the total weighted deviation with the preset basic deviation threshold, and select the reference data sets whose total deviation does not exceed the threshold. The similarity score is calculated based on the ratio of the total weighted deviation to the basic deviation threshold. The similarity score is then compared with the preset qualified similarity threshold to obtain a high similarity reference record.

7. The intelligent control method for a two-stage reverse osmosis purified water system according to claim 1, characterized in that, The method for performing step four is as follows: Compare the predicted data with the data benchmark and fluctuation range to determine whether the predicted data is within a reasonable range; If the estimated data is within a reasonable range, then based on the correspondence between the target data and other related data in the historical data, the change logic between the estimated correct data and the correct data of this purified water preparation is matched with the correspondence to determine whether the estimated data conforms to the data change correlation pattern. The average value and fluctuation range of the target data are obtained by statistically analyzing and calculating the target data in the historical reference data. The difference between the predicted correct data and the average value is calculated, and the difference is compared with the fluctuation range. Based on the comparison results, it is determined whether the deviation between the predicted correct data and the historical normal statistical level is within the target data deviation range. If the deviation between the predicted correct data and the historical normal statistical level is within the target data deviation range, the predicted correct data will be recorded according to time, abnormal instrument type, and monitoring data. If the deviation between the predicted correct data and the historical normal statistical level is outside the target data deviation range, return to step three; or if the predicted data is not within a reasonable range, obtain the predicted correct data again.

8. A cleaning method for a two-stage reverse osmosis purified water system, characterized in that, Includes the following steps: Step 1: Open the valve according to the second-level trial operation, and use the cleaning pump to pump clean, chlorine-free reverse osmosis product water from the cleaning tank into the pressure vessel of the membrane housing and discharge it for a few minutes. Step 2: Prepare the cleaning solution, open the secondary RO cleaning valve and start the cleaning pump, and control the inlet water pressure to be lower than the preset pressure; Step 3: Circulate the cleaning fluid in the pressure vessel for a preset time; Step 4: Record the pH value change and compare it with the initial value. If the pH value drops significantly, the cleaning solution should be prepared again. If the pH value does not change significantly, it proves that the cleaning is complete. Step 5: After cleaning the cleaning pump and heat exchanger with clean water, empty the cleaning tank and rinse it clean. Then fill it with clean product water for the next rinsing step. Step 6: Open the valves as normal, turn on the raw water pump manually, and use the pump to pump the clean, chlorine-free product water into the pressure vessel and discharge it. Step 7: After rinsing the reverse osmosis system, follow the valve opening and closing procedures during the secondary trial run until the product water is clean, free of foam or cleaning agent.

Images