Filter element self-adaptive backwashing system based on self-learning

Abstract

A filter element self-adaptive backwashing system based on self-learning adopts a self-learning algorithm, combines historical data and user feedback, and dynamically adjusts backwashing time. And a user feedback mechanism is introduced, so that the system better meets user requirements on the basis of real-time optimization. Mode recognition and water consumption habit modeling are carried out through an algorithm, a clustering algorithm is adopted to analyze a user water consumption mode, and a model of main water consumption time periods and water consumption distribution is generated. The adaptation capability of the system to different users can be improved, and individual requirements can be met. User interaction and feedback optimization are perfectly achieved, feedback data are collected through a user interaction interface, and self-learning algorithm parameters are dynamically optimized. The flexibility and adaptability of the system are enhanced on the whole, and continuous improvement is achieved.

Description

[Technical field]

[0001] The present invention relates to the technical field of water purification equipment, and in particular to technologies related to adaptive flushing and intelligent cleaning of filter elements in water purification equipment. [Background technology]

[0002] With the continuous improvement of living standards and the widespread popularization of household and commercial water treatment equipment, users have increasingly higher requirements for the quality of drinking water. The filter element in the drinking water filtration system will emphasize its self-cleaning ability. Generally, the industry will backwash the filter element in the filtration system to remove pollutants accumulated on the surface of the filter element and restore its filtration capacity. However, the traditional system adopts a fixed time interval backwashing method, which fails to take into account the changes in user water habits. There are the following problems: ⑴ The TDS value of the first cup of water is too high: the impurities accumulated on the surface of the filter element after long-term non-use are dissolved, resulting in the TDS value of the first cup of water exceeding the standard, affecting user experience and water quality safety. ⑵ Inefficiency: The fixed time interval cannot dynamically adapt to the actual water use habits of the user, which may lead to frequent backwashing or insufficient backwashing, thereby affecting the life of the filter element and water treatment efficiency. Therefore, there is an urgent need for an intelligent system that can adaptively adjust the backwash time according to the user's water use habits to solve the above problems. [Summary of the invention]

[0003] The present invention proposes an improvement on the existing filter element flushing or cleaning system that uses a fixed time interval flushing method, and provides a filter element adaptive backwashing system based on self-learning, which can dynamically adjust the backwashing time and optimize the filter element performance by real-time monitoring and analyzing the user's water use habits. It can reduce the high TDS value of the first cup of water by optimizing the flushing method, improve the water quality and drinking experience; extend the service life of the filter element, and improve the operating efficiency of the water treatment equipment; adapt to the water use habits of different users and provide personalized backwashing strategies.

[0004] The filter element adaptive backwashing system involved in the present invention is characterized in that the backwashing system comprises the following modules:

[0005] Data collection module: monitors user water consumption, water use time and TDS value, and generates historical data;

[0006] Data processing module: Use clustering algorithm to analyze users’ water usage habits and generate optimized backwashing time;

[0007] Control module: controls the action of the backwash execution module according to the optimized backwash time;

[0008] User interaction module: receives user satisfaction feedback data, makes adjustments after feedback calculation, and further optimizes the learning model;

[0009] Backwash execution module: responsible for the specific filter element backwash operation.

[0010] The adaptive backwash system uses a self-learning algorithm to dynamically adjust the backwash time based on historical data and user feedback. The algorithm performs pattern recognition and water use habit modeling, and uses a clustering algorithm to analyze user water use patterns to generate a model of the main water use periods and water consumption distribution.

[0011] The data acquisition module monitors and records the user's water consumption D, water use time and TDS value in real time.

[0012] The water consumption D is recorded as D = {d1, d2, ..., d n},

[0013] The TDS value is recorded as TDS = {TDS1, TDS2, ..., TDS n}

[0014] The current temperature is T current =30℃

[0015] The temperature compensation formula is: TDS actual =TDS×(1+a·(T current -T ref ))

[0016] Where a=0.02T ref =25℃

[0017] Calculate the compensated TDS value one by one through the temperature compensation formula to obtain the compensated data TDS actual

[0018] The data analysis module analyzes the user's water usage habits through a clustering algorithm (such as K-means), and uses the algorithm to perform pattern recognition on the water usage data to find out the main water usage time periods and characteristic values.

[0019] The cluster center μ represents the main water consumption distribution:

[0020] Step 1: Initialize the cluster center;

[0021] Select K = 2 cluster centers. Assuming that the user's water consumption pattern has two main periods: off-peak period and peak period, our initial cluster centers are:

[0022] Cluster center 1μ1 comes from the low-peak time period, and cluster center 2μ2 comes from the peak time period;

[0023] Step 2: Assign data points to the nearest cluster center and convert the compensated TDS actualThe data points and water consumption data points are taken as input, the Euclidean distance from each data point to the two cluster centers is calculated, and then the data point is assigned to the center with the closer distance;

[0024] Step 3: Calculate new cluster centers;

[0025] After assigning the data points to different cluster centers, calculate the new cluster center, i.e., the mean;

[0026] Step 4: Check whether the clustering has converged;

[0027] Compare the new cluster center with the previous cluster center. If the change is less than the set threshold (0.01ppm), the clustering is considered to have converged. If the clustering is not converged enough, return to step 2 and repeat steps 2 to 4 until the cluster center is converged.

[0028] The generated optimized backwash time is based on the TDS value history and target TDS value target , using self-learning algorithm to adjust the backwash time: T new =T old +a·(TDS t -TDS target ),

[0029] in:

[0030] T new : New backwash time;

[0031] T old : The last round of backwashing time;

[0032] α: learning rate, controls the adjustment speed;

[0033] By bringing in the system value and setting the learning rate α as required, a new backwash time can be calculated based on the new TDS data. Different backwash times can be calculated for cluster center 1 and cluster center 2 respectively to obtain the optimized backwash time.

[0034] The user interaction module refers to the part that collects user feedback and dynamically optimizes, where the user feedback score is F, and the full score of F is set to 5. The expected feedback score is F target , feedback adjustment coefficient β, feedback adjustment formula is: T adjusted =T new +β·(FF target )

[0035] Calculate the adjustment time T for off-peak time and peak time respectively new .

[0036] Backwash operation: the system triggers the backwash process according to the optimized time to clean the filter element and maintain stable water quality.

[0037] The adaptive flushing system involved in the present invention adopts a self-learning algorithm, combined with historical data and user feedback, to dynamically adjust the backwashing time. The user feedback mechanism is introduced to make the system more in line with user needs on the basis of real-time optimization. Pattern recognition and water use habit modeling are carried out through the algorithm, and the user's water use pattern is analyzed by a clustering algorithm to generate a model of the main water use period and water consumption distribution. It can improve the system's adaptability to different users and meet personalized needs. It perfectly realizes user interaction and feedback optimization, collects feedback data through the user interaction interface, and dynamically optimizes the self-learning algorithm parameters. Overall, it enhances the flexibility and adaptability of the system and achieves continuous improvement. [Specific implementation method]

[0038] The present invention will be described in detail below with reference to embodiments. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0039] In the description of the present invention, it is necessary to understand that the orientations or positional relationships indicated by terms such as “center”, “up”, “down”, “front”, “back”, “left”, “right”, “vertical”, “horizontal”, “top”, “bottom”, “inside” and “outside” are positional relationships, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0040] The filter element adaptive backwashing system involved in the present invention is characterized in that the backwashing system comprises the following modules:

[0041] Data collection module: monitors user water consumption, water use time and TDS value, and generates historical data;

[0042] Data processing module: Use clustering algorithm to analyze users’ water usage habits and generate optimized backwashing time;

[0043] Control module: controls the action of the backwash execution module according to the optimized backwash time;

[0044] User interaction module: receives user satisfaction feedback data, makes adjustments after feedback calculation, and further optimizes the learning model;

[0045] Backwash execution module: responsible for the specific filter element backwash operation.

[0046] The data acquisition module monitors and records the user's water consumption D, water use time and TDS value in real time.

[0047] The water consumption D is recorded as D = {d1, d2, ..., d n},

[0048] The TDS value is recorded as TDS = {TDS1, TDS2, ..., TDS n},

[0049] The water consumption data recorded in real time: D = {2, 3, 5, 7, 9} L;

[0050] TDS value data: TDS = {10, 15, 12, 20, 25} ppm;

[0051] The current temperature is T current =30℃

[0052] The temperature compensation formula is: TDS actual =TDS×(1+a·(T current -T ref ))

[0053] Where a=0.02T ref =25℃

[0054] Calculate the compensated TDS value one by one through the temperature compensation formula to obtain the compensated data TDS actual

[0055] TDS actual,1 =10×(1+0.02·(30-25))=10×1.1=11ppm

[0056] TDS actual,2 =15×1.1=16.5ppm

[0057] TDS actual,3 =12×1.1=13.2ppm

[0058] TDS actual,4 =20×1.1=22ppm

[0059] TDS actual,5 =25×1.1=27.5ppm

[0060] The compensated value is obtained through the temperature compensation formula: TDS actual ={11,16.5,13.2,22,27.5}ppm;

[0061] The data analysis module analyzes the user's water usage habits through a clustering algorithm (such as K-means), and uses the algorithm to perform pattern recognition on the water usage data to find out the main water usage time periods and characteristic values.

[0062] The cluster center μ represents the main water consumption distribution:

[0063] Step 1: Initialize the cluster center;

[0064] Select K = 2 cluster centers. Assuming that the user's water consumption pattern has two main periods: off-peak period and peak period, our initial cluster centers are:

[0065] Cluster center 1μ1 comes from the low-peak time period, and cluster center 2μ2 comes from the peak time period;

[0066] Cluster center 1μ1=13.2ppm (from low peak period)

[0067] Cluster center 2μ2 = 22ppm (from peak hours)

[0068] Step 2: Assign data points to the nearest cluster center and convert the compensated TDS actual The data point and water consumption data point D are taken as input, the Euclidean distance from each data point to the two cluster centers is calculated, and then the data point is assigned to the center with the closer distance;

[0069] TDS value after compensation: TDS actual ={11,16.5,13.2,22,27.5}ppm;

[0070] Corresponding water consumption data: D = {2, 3, 5, 7, 9} L;

[0071] Calculate the distance between each data point and the cluster center:

[0072] For TDS = 11ppm:

[0073] Distance from cluster center 1: |11-13.2|=2.2;

[0074] Distance from cluster center 2: |11-22|=11;

[0075] The nearest cluster center: cluster center 1 (13.2ppm);

[0076] For TDS = 16.5ppm:

[0077] Distance to cluster center 1: |16.5-13.2|=3.3;

[0078] Distance to cluster center 2: |16.5-22|=5.5;

[0079] The nearest cluster center: cluster center 1 (13.2ppm);

[0080] For TDS = 13.2ppm:

[0081] Distance from cluster center 1: |13.2-13.2|=0;

[0082] Distance to cluster center 2: |13.2-22|=8.8;

[0083] The nearest cluster center: cluster center 1 (13.2ppm);

[0084] For TDS = 22ppm:

[0085] Distance to cluster center 1: |22-13.2|=8.8;

[0086] Distance from cluster center 2: |22-22|=0;

[0087] The nearest cluster center: cluster center 2 (22ppm);

[0088] For TDS = 27.5ppm:

[0089] Distance to cluster center 1: |27.5-13.2|=14.3;

[0090] Distance to cluster center 2: |27.5-22|=5.5;

[0091] The nearest cluster center: cluster center 2 (22ppm);

[0092] Based on the results of the distance calculation, we assign the data points to the cluster centers as follows:

[0093] Cluster center 1 (off-peak period): TDS slow ={11,16.5,13.2};

[0094] Cluster center 2 (peak hour): TDS high ={22,27.5};

[0095] Step 3: Calculate new cluster centers;

[0096] After assigning the data points to different cluster centers, calculate the new center of each cluster, i.e., the mean;

[0097] Cluster center 1 (off-peak period):

[0098] TDS slow ={11,16.5,13.2};

[0099] Cluster center 1μ1 calculation:

[0100]

[0101] Cluster center 2 (peak hour):

[0102] TDS high ={22,27.5};

[0103] Cluster center 2μ2 calculation:

[0104]

[0105] Step 4: Check whether the clustering has converged;

[0106] Compare the new cluster center with the previous cluster center. If the change is less than the set threshold (0.01ppm), the clustering is considered to have converged. If the clustering is not converged enough, return to step 2 and repeat steps 2 to 4 until the cluster center is converged.

[0107] New cluster centers: μ1=13.57ppm, μ2=24.57ppm;

[0108] Old cluster center: μ1=13.2ppm, μ2=22ppm;

[0109] Calculate the difference:

[0110] ∣13.57-13.2∣=0.37ppm

[0111] |24.75-22|=2.75ppm

[0112] Since both differences are greater than the threshold we set, the cluster centers need to be updated. Therefore, we will return to step 2, assign the data points to the new cluster centers again, and recalculate the cluster centers.

[0113] Step 5: Repeat steps 2 to 4 until convergence.

[0114] This process is repeated until the change in the cluster center is less than the set threshold. Assuming that after several rounds of iterations, the cluster center changes little and the clustering converges, the final clustering result is as follows:

[0115] Off-peak time clustering: TDS slow ={11,16.5,13.2}

[0116] Cluster center: μ1 = 13.57 ppm

[0117] Peak Hour Clustering: TDS high ={22,27.5}

[0118] Cluster center: μ2 = 24.75ppm

[0119] The generated optimized backwash time is based on the TDS value history and target TDS value target , using self-learning algorithm to adjust the backwash time: T new =T old +a·(TDS t -TDS target ),

[0120] in:

[0121] T new : New backwash time;

[0122] T old : The last round of backwashing time;

[0123] α: learning rate, controls the adjustment speed;

[0124] By bringing in the system value and setting the learning rate α as required, a new backwash time can be calculated based on the new TDS data. Different backwash times can be calculated for cluster center 1 and cluster center 2 respectively to obtain the optimized backwash time.

[0125] For the low peak period (cluster 1, μ1=13.57ppm):

[0126] The current TDS value is TDS t =16.5ppm=16.5ppm (from the low peak period).

[0127] Backwash time formula:

[0128] T new =T old +a·(TDS t -TDS target )

[0129] T new =10+0.1·(16.5-15)=10+0.1·1.5=10.15 seconds;

[0130] For the peak period (cluster 2, μ2 = 24.75 ppm):

[0131] The current TDS value is TDS t =22ppm (from peak hours).

[0132] Backwash time formula:

[0133] T new =T old +a·(TDS t -TDS target )

[0134] Tnew =10+0.1·(22-15)=10+0.7=10.7 seconds;

[0135] The user interaction module refers to the part that collects user feedback and dynamically optimizes, where the user feedback score is F, and the full score of F is set to 5. The expected feedback score is F target , assuming that the expected target feedback score is 4, the feedback adjustment coefficient β, the feedback adjustment formula is: T adjusted =T new +β·(FF target )

[0136] Calculate the adjustment time T for off-peak time and peak time respectively new .

[0137] For off-peak hours:

[0138] T new = 10.15 seconds

[0139] T adjusted =10.15+2·(3-4)=10.15-2=8.15 seconds;

[0140] For peak hours:

[0141] T new = 10.7 seconds

[0142] T adjusted =10.7+2·(3-4)=10.7-2=8.7 seconds;

[0143] Backwash operation: the system triggers the backwash process according to the optimized time to clean the filter element and maintain stable water quality.

[0144] Off-peak period: 8.15 seconds;

[0145] Peak hour: 8.7 seconds.

[0146] The adaptive flushing system involved in the present invention adopts a self-learning algorithm, combined with historical data and user feedback, to dynamically adjust the backwashing time. The user feedback mechanism is introduced to make the system more in line with user needs on the basis of real-time optimization. Pattern recognition and water use habit modeling are carried out through the algorithm, and the user's water use pattern is analyzed by a clustering algorithm to generate a model of the main water use period and water consumption distribution. It can improve the system's adaptability to different users and meet personalized needs. It perfectly realizes user interaction and feedback optimization, collects feedback data through the user interaction interface, and dynamically optimizes the self-learning algorithm parameters. Overall, it enhances the flexibility and adaptability of the system and achieves continuous improvement.

[0147] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention is disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technician familiar with this profession can make some changes or modifications to equivalent embodiments of equivalent changes by using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent changes and modifications made to the above embodiments according to the technology of the present invention are within the scope of the technical solution of the present invention without departing from the content of the technical solution of the present invention.

Claims

1. A filter element adaptive backwashing system based on self-learning, the filter element adaptive backwashing system is characterized in that: The backwash system includes the following modules: Data collection module: monitors user water consumption, water use time and TDS value, and generates historical data; Data processing module: Use clustering algorithm to analyze users’ water usage habits and generate optimized backwashing time; Control module: controls the action of the backwash execution module according to the optimized backwash time; User interaction module: receives user satisfaction feedback data, makes adjustments after feedback calculation, and further optimizes the learning model; Backwashing execution module: responsible for the specific filter element backwashing operation; The adaptive backwash system uses a self-learning algorithm to dynamically adjust the backwash time based on historical data and user feedback. The algorithm performs pattern recognition and water use habit modeling, and uses a clustering algorithm to analyze user water use patterns to generate a model of the main water use periods and water consumption distribution.

2. The filter element adaptive reaction flushing system based on self-learning according to claim 1, characterized in that: The data acquisition module monitors and records the user's water consumption D, water use time and TDS value in real time, where the water consumption D is recorded as D={d1,d2,...,d n }, The TDS value is recorded as TDS = {TDS1, TDS2, ..., TDS n }, The current temperature is T current =30℃; The temperature compensation formula is: TDS actual =TDS×(1+a·(T current -T ref )); Where a=0.02T ref =25℃; Calculate the compensated TDS value one by one through the temperature compensation formula to obtain the compensated data TDS actual .

3. The filter element adaptive reaction flushing system based on self-learning according to claim 1, characterized in that: The data analysis module analyzes the user's water use habits through the K-means clustering algorithm, and performs pattern recognition on the water use data through the K-means algorithm to find out the main water use time periods and characteristic values.

4. The filter element adaptive reaction flushing system based on self-learning according to claim 3 is characterized in that: The cluster center μ represents the main water consumption distribution: Step 1: Initialize the cluster center; Select K = 2 cluster centers. Assuming that the user's water consumption pattern has two main periods: off-peak period and peak period, our initial cluster centers are: Cluster center 1μ1 comes from the low-peak time period, and cluster center 2μ2 comes from the peak time period; Step 2: Assign data points to the nearest cluster center and convert the compensated TDS actual The data points and water consumption data points are taken as input, the Euclidean distance from each data point to the two cluster centers is calculated, and then the data point is assigned to the center with the closer distance; Step 3: Calculate new cluster centers; After assigning the data points to different cluster centers, calculate the new cluster center, i.e., the mean; Step 4: Check whether the clustering has converged; The new cluster center is compared with the previous cluster center. If the change is less than the set threshold, which is 0.01ppm, the clustering is considered to have converged.

5. The filter element adaptive reaction flushing system based on self-learning according to claim 4 is characterized in that: If the clustering is not converged enough, return to step 2 and repeat steps 2 to 4 until the cluster centers are converged.

6. The filter element adaptive reaction flushing system based on self-learning according to claim 4, characterized in that: The generated optimized backwash time is based on the TDS value history and target TDS value target , using self-learning algorithm to adjust the backwash time: T new =T old +a·(TDS t -TDS target ), in: T new : New backwash time; T old : The last round of backwashing time; α: learning rate, controls the adjustment speed; By bringing in the system value and setting the learning rate α as required, a new backwash time can be calculated based on the new TDS data. Different backwash times can be calculated for cluster center 1 and cluster center 2 respectively to obtain the optimized backwash time.

7. The filter element adaptive reaction flushing system based on self-learning according to claim 4, characterized in that: The user interaction module refers to the part that collects user feedback and dynamically optimizes, where the user feedback score is F, and the full score of F is set to 5. The expected feedback score is F target , feedback adjustment coefficient β, feedback adjustment formula is: T adjusted =T new +β·(FF target ), calculate the adjustment time T for off-peak time and peak time respectively new .

8. The filter element adaptive reaction flushing system based on self-learning according to claim 7, characterized in that: The backwashing operation is performed, and the system triggers the backwashing process according to the optimized time to clean the filter element and maintain stable water quality.

Images