Intelligent control method and system for denitrification filter

Abstract

The present application relates to the technical field of sewage treatment, in particular to a denitrification filter tank intelligent control method and system. Through collecting filter tank data information of multiple target filter tanks, the filter tank data information is preprocessed to obtain preprocessed information, then the preprocessed information is input into a control model, a standard filter tank is created based on the preprocessed information through the control model, and the standard reagent amount of each standard filter tank is obtained in combination with the standard filter tank, to obtain a trained control model, finally real-time filter tank information of a real-time filter tank is collected, the corresponding standard filter tank and standard reagent amount of the real-time filter tank are obtained through the trained control model, and the real-time reagent amount of the real-time filter tank is adjusted in combination with the standard reagent amount. The standard reagent amount corresponding to the standard filter tank is used as the reagent feeding standard of each standard filter tank, so that the user can quickly determine the feeding standard when feeding the real-time filter tank, thereby improving the adjustment efficiency of the user for the target filter tank.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of sewage treatment, in particular to an intelligent control method and system for a denitrification filter. BACKGROUND

[0002] The denitrification filter is a sewage treatment structure using quartz sand, volcanic rock, ceramsite and other filter materials as biofilm medium. Nitrate nitrogen is reduced to nitrogen gas under anoxic conditions by denitrifying bacteria to achieve denitrification, and suspended solids are intercepted by deep bed filtration. The depth of the filter bed is generally more than 1.5 meters.

[0003] A denitrification filter carbon source precise dosing system is disclosed in Chinese Patent No. CN112279371B. The inlet of the carbon source mixing pool is connected to the water inlet pipe, and the outlet is connected to the inlet of the denitrification filter water inlet pipe channel. The outlet of the denitrification filter is connected to the denitrification filter water outlet pipe channel. An online flowmeter and a first online nitrate meter are provided on the water inlet pipe. The carbon source mixing pool is connected to a carbon source dosing pump. An online dissolved oxygen meter test probe is provided in the denitrification filter. An online nitrite meter and a second online nitrate meter are provided on the denitrification filter water outlet pipe channel. The carbon source precise dosing controller is connected to the online flowmeter, the first online nitrate meter, the online nitrite meter, the second online nitrate meter, the online dissolved oxygen meter and the carbon source dosing pump. However, the existing technology does not establish a standard for adding chemicals to different specifications of filter pools, making it difficult for users to establish management standards when managing different specifications of filter pools, resulting in low management efficiency of the denitrification filter and difficulty in quickly determining the amount of chemicals to be added to the filter pool. SUMMARY

[0004] The present application aims to solve the problems in the background art by providing an intelligent control method and system for a denitrification filter.

[0005] The technical solution of the present application is as follows:

[0006] On the one hand, the present application provides an intelligent control method for a denitrification filter, comprising:

[0007] Collecting filter pool data information of a plurality of target filter pools, preprocessing the filter pool data information to obtain preprocessed information;

[0008] Creating a control model;

[0009] Inputting the preprocessed information into the control model, creating a standard filter pool based on the preprocessed information through the control model, and obtaining the standard chemical amount of each standard filter pool in combination with the standard filter pool to obtain a trained control model;

[0010] Collect real-time filter data of a real-time filter, obtain a standard filter and a standard dosage of a standard medicament corresponding to the real-time filter through the trained control model, and adjust the real-time dosage of the real-time filter in combination with the standard dosage of the standard medicament.

[0011] Preferably, filter data information of a plurality of target filters is collected, and the filter data information is preprocessed to obtain preprocessed information, including:

[0012] A filter database is created.

[0013] Filter data information of a plurality of target filters is collected respectively, and all collected filter data information is input to the filter database; the filter data information includes filter information and medicament information.

[0014] A filter data information is randomly selected from the filter database.

[0015] It is judged whether there is duplicate data in the filter data information.

[0016] If there is duplicate data in the filter data information, the duplicate data is removed.

[0017] A filter data information is randomly selected from the filter database until all filter data in the filter database is selected to obtain a plurality of preprocessed information; the preprocessed information includes preprocessed filter information and preprocessed medicament information.

[0018] Preferably, the preprocessed information is input to the control model, and a standard filter is created based on the preprocessed information through the control model, and a standard dosage of each standard filter is obtained in combination with the standard filter to obtain a trained control model, including:

[0019] All preprocessed information is divided into a training set and a test set according to a random ratio.

[0020] The training set is input to the control model, and clustering analysis is performed on all target filters through the control model to screen a plurality of standard filters, and a standard dosage is established based on medicament information of the standard filter to obtain a trained control model.

[0021] The test set is input to the trained control model to verify whether the trained control model is trained.

[0022] Preferably, the training set is input to the control model, and clustering analysis is performed on all target filters through the control model to screen a plurality of standard filters, and a standard dosage is established based on medicament information of the standard filter, including:

[0023] A preprocessed information of a target filter is randomly selected from the training set.

[0024] A filter model is created based on the pretreated filter information, and a coupling relationship between the filter model and the reagent information is established by combining the filter model.

[0025] Return the preprocessing information of a target filter randomly selected from the training set, until all target filters in the training set have been selected, and obtain the drug dosage for each target filter.

[0026] K target filters are randomly selected and designated as central filters. The remaining target filters in the training set are randomly assigned to a central filter to obtain K filter clusters.

[0027] Set an iteration count threshold;

[0028] For each filter cluster, calculate the distance from any point within the cluster to the central filter, and record the point corresponding to the average of all distances as the new central filter.

[0029] Determine if the number of iterations is greater than or equal to the iteration count threshold;

[0030] If the number of iterations is greater than or equal to the iteration threshold, the iteration is stopped, and the central filter obtained in the last iteration is recorded as the standard filter.

[0031] Obtain the dosage of the chemical corresponding to the standard filter, and record the dosage of the chemical corresponding to the standard filter as the standard dosage.

[0032] Preferably, real-time filter data is collected from the real-time filter, and the corresponding standard filter and standard dosage are obtained through a trained control model. The real-time dosage of the real-time filter is then adjusted based on the standard dosage, including:

[0033] Collect real-time filter data; the real-time filter data includes real-time filter information and real-time reagent information;

[0034] The real-time filter data is input into the trained control model to obtain the standard filter and labeled drug dosage corresponding to the real-time filter.

[0035] The dosage of chemicals in the real-time filter is adjusted by combining the real-time dosage with the standard dosage.

[0036] Preferably, real-time filter data is input into the trained control model to obtain the standard filter and labeled dosage corresponding to the real-time filter, including:

[0037] The real-time filter beds are clustered by combining K-means clustering analysis with real-time filter bed information to obtain the standard filter beds corresponding to the real-time filter beds.

[0038] Obtain the standard dosage of the reagent corresponding to the standard filter.

[0039] Preferably, the real-time filter dose is adjusted according to the real-time dose and the standard dose, including:

[0040] an error threshold is set;

[0041] it is determined whether the error between the real-time dose and the standard dose is greater than or equal to the error threshold;

[0042] if the error between the real-time dose and the standard dose is greater than or equal to the error threshold, the real-time dose is adjusted until the error between the real-time dose and the standard dose is less than the error threshold;

[0043] if the error between the real-time dose and the standard dose is greater than or equal to the error threshold, the current state of the real-time filter is maintained.

[0044] In another aspect, the present application also provides an intelligent control system for a denitrification filter, including a collection component and a control component, the filter data information of a plurality of target filters is collected by the collection component, the intelligent control method for the denitrification filter according to any one of the preceding aspects is executed by the control component, the standard filter and the corresponding standard dose are established by the control component in combination with the filter data information, so that the real-time filter is adjusted based on the standard filter.

[0045] Preferably, the collection component includes a filter collection module and a dose collection module, the filter information is collected by the filter collection module, and the dose information is collected by the dose collection module.

[0046] Compared with the prior art, the above technical solution of the present application has the following beneficial technical effects:

[0047] By collecting the filter data information of a plurality of target filters, the filter data information is preprocessed to obtain preprocessed information, and then the preprocessed information is input into a control model, the standard filter is created based on the preprocessed information by the control model, and the standard dose of each standard filter is obtained in combination with the standard filter, the trained control model is obtained, finally the real-time filter information of the real-time filter is collected, the corresponding standard filter and the standard dose of the real-time filter are obtained by the trained control model, and the real-time dose of the real-time filter is adjusted in combination with the standard dose, the present application clusters a plurality of target filters to obtain a standard filter, and uses the standard dose corresponding to the standard filter as the dose standard of each standard filter, so that the user can quickly determine the standard when performing dose feeding on the real-time filter, thereby improving the adjustment efficiency of the user for the target filter, making the dose more accurate, and ensuring that the effluent water quality is stable and meets the discharge standard. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 a flowchart of an intelligent control method for a denitrification filter according to the present application is provided;

[0049] Figure 2 A principle block diagram of an intelligent control system of a denitrification filter tank is provided in the present application;

[0050] BRIEF DESCRIPTION OF DRAWINGS: 100, acquisition component; 101, filter tank acquisition module; 102, reagent acquisition module;

[0051] 200, control component. DETAILED DESCRIPTION

[0052] In an embodiment, as shown in the accompanying drawings, an intelligent control method of a denitrification filter tank is provided in the present application, which comprises: Figure 1

[0053] S100, acquiring filter tank data information of a plurality of target filter tanks, pre-processing the filter tank data information to obtain pre-processed information;

[0054] S200, creating a control model;

[0055] S300, inputting the pre-processed information into the control model, creating a standard filter tank based on the pre-processed information through the control model, and obtaining a standard reagent amount of each standard filter tank in combination with the standard filter tank to obtain a trained control model;

[0056] S400, acquiring real-time filter tank information of a real-time filter tank, obtaining a standard filter tank corresponding to the real-time filter tank and a standard reagent amount through the trained control model, and adjusting a real-time reagent amount of the real-time filter tank in combination with the standard reagent amount.

[0057] In the present application, the filter tank data information of a plurality of target filter tanks is acquired, the filter tank data information is pre-processed to obtain pre-processed information, the pre-processed information is then input into a control model, a standard filter tank is created based on the pre-processed information through the control model, a standard reagent amount of each standard filter tank is obtained in combination with the standard filter tank to obtain a trained control model, finally, real-time filter tank information of a real-time filter tank is acquired, a standard filter tank corresponding to the real-time filter tank and a standard reagent amount are obtained through the trained control model, and a real-time reagent amount of the real-time filter tank is adjusted in combination with the standard reagent amount. The present application clusters a plurality of target filter tanks to obtain a standard filter tank, and uses a standard reagent amount corresponding to the standard filter tank as a reagent feeding standard of each standard filter tank, so that the user can quickly determine the feeding standard when feeding reagents to a real-time filter tank, thereby improving the adjustment efficiency of the user for the target filter tank, making the reagent amount more accurate, and ensuring that the effluent water quality is stable and meets the discharge standard.

[0058] In an optional embodiment, the S100 comprises:

[0059] S110, creating a filter tank database;

[0060] ​S120, filter pool data information of the plurality of target filter pools is collected respectively, and all the collected filter pool data information is input to a filter pool database; the filter pool data information comprises filter pool information and medicament information;

[0061] S130, one filter pool data information is randomly selected from the filter pool database;

[0062] S140, it is judged whether there is repeated data in the filter pool data information;

[0063] S150, if there is repeated data in the filter pool data information, the repeated data is removed;

[0064] S160, one filter pool data information is randomly selected from the filter pool database until all the filter pool data in the filter pool database is selected to obtain a plurality of preprocessed information; the preprocessed information comprises preprocessed filter pool information and preprocessed medicament information;

[0065] Specifically, the filter pool data mainly comprises length, width and height of the target filter pool, and the medicament information comprises types of medicaments put in different processing stages of the target filter pool and the amount of each type of medicament.

[0066] It should be noted that, by preprocessing all the collected filter pool data information, invalid data contained in the filter pool data information is removed, on the one hand, the integrity of the data information is ensured, and on the other hand, the reliability of the control model trained by the preprocessed data is ensured.

[0067] In an optional embodiment, the S300 comprises:

[0068] S310, all the preprocessed information is divided into a training set and a test set according to a random ratio;

[0069] S320, the training set is input to the control model, and all the target filter pools are analyzed by clustering to screen a plurality of standard filter pools, and a standard medicament amount is established based on medicament information of the standard filter pools to obtain a trained control model;

[0070] S330, the test set is input to the trained control model to verify whether the trained control model is trained.

[0071] It should be noted that, when the training set and the test set are divided, it is necessary to ensure that the division ratio of the training set is greater than that of the test set, so as to increase the number of training samples contained in the training set, so that the trained control model is more reliable.

[0072] The application trains a control model to perform cluster analysis on multiple target filter tanks of different types based on filter tank data information, thereby screening out the most representative filter tanks from the multiple target filter tanks, and recording these screened target filter tanks as standard filter tanks. Then, a reference standard for the drug delivery of different types of target filter tanks can be established according to the drug information corresponding to the standard filter tanks, thereby improving the management efficiency of the target filter tanks. When verifying whether the trained control model is trained, the response speed and / or accuracy of the trained control model can be used as a judgment standard.

[0073] In an optional embodiment, the S320 comprises:

[0074] S321, randomly selecting pre-processing information of a target filter tank from the training set;

[0075] S322, creating a filter tank model based on the pre-processing filter tank information, and creating a coupling relationship between the filter tank model and the drug information;

[0076] Specifically, the filter tank model can be established by a BIM model;

[0077] S323, returning to randomly selecting pre-processing information of a target filter tank from the training set until all target filter tanks in the training set are selected, and obtaining the drug amount of each target filter tank;

[0078] S324, randomly selecting K target filter tanks, recording the selected K target filter tanks as center filter tanks, and randomly assigning the remaining target filter tanks in the training set to a center filter tank to obtain K filter tank clusters;

[0079] S325, setting an iteration number threshold;

[0080] S326, for each filter tank cluster, calculating the distance from any point in the cluster to the center filter tank, and recording the point corresponding to the average of all distances as a new center filter tank;

[0081] S327, determining whether the iteration number is greater than or equal to the iteration number threshold;

[0082] S328, if the iteration number is greater than or equal to the iteration number threshold, stopping iteration, and recording the center filter tank obtained in the last iteration as a standard filter tank;

[0083] S329, obtaining the drug amount corresponding to the standard filter tank, and recording the drug amount corresponding to the standard filter tank as a standard drug amount.

[0084] It should be noted that the filter model corresponding to different target filter is obtained based on the pre-processed filter pool information, so that the real-time filter can be directly divided by the filter model subsequently, to quickly obtain the standard filter corresponding to the real-time filter, and then the standard medicament amount corresponding to the standard filter can be used to quickly adjust the medicament amount in the real-time filter, thereby improving the adjustment efficiency

[0085] Since the cluster center obtained by K-means clustering has the ability to represent the characteristics of all points in the cluster, the standard filter obtained by K-means clustering can represent different types of filters, and only in this case can the standard medicament amount obtained have reference value, so that the user can quickly complete the medicament feeding for the target filter according to the medicament amount.

[0086] In an optional embodiment, the S400 comprises:

[0087] S410, collecting real-time filter data of a real-time filter; the real-time filter data comprises real-time filter information and real-time medicament information;

[0088] S420, inputting the real-time filter data into the trained control model to obtain a standard filter corresponding to the real-time filter and a labeled medicament amount;

[0089] S430, adjusting the medicament amount in the real-time filter in combination with the real-time medicament amount and the standard medicament amount.

[0090] It should be noted that after obtaining the trained control model, only the real-time filter data needs to be input into the trained control model, and the real-time filter can be analyzed by clustering through the trained control model in combination with the real-time filter information of the real-time filter, so that the corresponding standard filter is selected from a plurality of standard filters, and after obtaining the standard filter corresponding to the real-time filter, the medicament amount in the real-time filter can be adjusted in combination with the standard medicament amount of the standard filter and the real-time medicament information of the real-time filter.

[0091] In an optional embodiment, the S420 comprises:

[0092] S421, clustering the real-time filter by K-means clustering analysis in combination with the real-time filter information, to obtain a standard filter corresponding to the real-time filter;

[0093] S422, obtaining a standard medicament amount corresponding to the standard filter;

[0094] It should be noted that the K-means clustering analysis method is used for clustering of the real-time filter pool in the present application. The K-means clustering analysis is a widely used clustering algorithm, which is used to divide a data set into k groups, and each group is represented by a center. The goal of the K-means algorithm is to minimize the sum of the squared distances between the data points in the group and their group centers.

[0095] Since the final cluster center obtained by the K-means clustering analysis can represent the characteristics of all points in the cluster, the standard filter pool obtained by the K-means clustering analysis in the real-time filter pool has the ability to represent the real-time filter pool, thereby ensuring the referenceability of the standard filter pool.

[0096] In an optional embodiment, the S430 comprises:

[0097] S431, setting an error threshold;

[0098] S432, determining whether the error between the real-time drug dose and the standard drug dose is greater than or equal to the error threshold;

[0099] S433, if the error between the real-time drug dose and the standard drug dose is greater than or equal to the error threshold, adjusting the real-time drug dose until the error between the real-time drug dose and the standard drug dose is less than the error threshold;

[0100] S434, if the error between the real-time drug dose and the standard drug dose is greater than or equal to the error threshold, maintaining the current state of the real-time filter pool.

[0101] It should be noted that after obtaining the standard drug dose, the standard drug dose can be used as a standard for measuring the drug dose of the real-time filter pool, so as to determine whether the drug dose in the real-time filter pool needs to be adjusted. When the real-time drug dose is less than the standard drug dose, the drug needs to be added to the real-time filter pool, and when the real-time drug dose is greater than the standard drug dose, water needs to be added to the target filter pool until the error between the real-time drug dose and the standard drug dose is greater than or equal to the error threshold.

[0102] As shown in Figure 2 The present application also provides an intelligent control system for a denitrification filter pool, which comprises a collection component 100 and a control component 200. The filter pool data information of a plurality of target filter pools is collected by the collection component 100, and the intelligent control method for the denitrification filter pool described in any one of the preceding embodiments is executed by the control component 200. The control component 200 establishes a standard filter pool and a corresponding standard drug dose based on the filter pool data information, so as to adjust the real-time filter pool based on the standard filter pool.

[0103] Specifically, the collecting assembly 100 in the present application comprises a plurality of different types of sensors, such as a dissolved oxygen sensor, a nitrate sensor and a DO sensor, and is respectively installed at different positions of the target filter tank, wherein the dissolved oxygen sensor is installed at the water inlet of the target filter tank.

[0104] It should be noted that the present application collects the filter tank data information at each position in the target filter tank through the collecting assembly 100, and inputs all the collected filter tank data information into the control assembly 200, and the control assembly 200 screens the standard filter tank in combination with the filter tank data information, and adjusts the dosage of the medicament in the target filter tank based on the standard dosage of the standard filter tank, until the dosage of the medicament in the target filter tank is the same as the standard dosage.

[0105] In an optional embodiment, the collecting assembly 100 comprises a filter tank collecting module 101 and a medicament collecting module 102, the filter tank information is collected through the filter tank collecting module 101, and the medicament information is collected through the medicament collecting module 102.

[0106] It should be noted that the filter tank information in the present application mainly comprises the type and specification information of the target filter tank, such as the type, length, width, depth, filter material type, filling thickness, etc. of the filter tank, and the water inflow, temperature, suspended solids interception data, etc. of the target filter tank, and the medicament information mainly comprises the type of the medicament added in the target filter tank and the corresponding addition amount of each type of medicament.

[0107] The embodiments of the present application are described in detail above in combination with the drawings, but the present application is not limited thereto, and various changes can be made within the knowledge range of those skilled in the art without departing from the purpose of the present application.

Claims

1. A method for intelligent control of a denitrification filter, characterized in that, The method comprises the following steps: Collecting filter data information of a plurality of target filters, preprocessing the filter data information to obtain preprocessed information; The preprocessed information includes preprocessed filter information and preprocessed reagent information; Creating a control model; Inputting the preprocessed information into the control model, creating a standard filter based on the preprocessed information through the control model, and obtaining the standard reagent amount of each standard filter in combination with the standard filter to obtain a trained control model; Collecting real-time filter data of a real-time filter, obtaining the standard filter corresponding to the real-time filter and the standard reagent amount through the trained control model, and adjusting the real-time reagent amount of the real-time filter in combination with the standard reagent amount; The preprocessed information is input into the control model, the standard filter is created based on the preprocessed information through the control model, and the standard reagent amount of each standard filter is obtained in combination with the standard filter to obtain a trained control model, which comprises: Dividing all preprocessed information into a training set and a test set according to a random ratio; Inputting the training set into the control model, performing cluster analysis on all target filters through the control model to screen a plurality of standard filters, and establishing a standard reagent amount based on the reagent information of the standard filter to obtain a trained control model; Inputting the test set into the trained control model to verify whether the trained control model is trained; The training set is input into the control model, cluster analysis is performed on all target filters through the control model to screen a plurality of standard filters, and a standard reagent amount is established based on the reagent information of the standard filter to obtain a trained control model, which comprises: Randomly selecting preprocessed information of a target filter from the training set; Creating a filter model based on the preprocessed filter information, and creating a coupling relationship between the filter model and the reagent information of the filter model; Returning to the preprocessed information of a target filter randomly selected from the training set until all target filters in the training set are selected to obtain the reagent amount of each target filter; Randomly selecting K target filters, taking the selected K target filters as center filters, and randomly assigning the remaining target filters in the training set to a center filter to obtain K filter clusters; Setting an iteration number threshold; For each filter cluster, the distance from any point in the cluster to the center filter is calculated, and the point corresponding to the average of all distances is taken as a new center filter; Determine whether the iteration number is greater than or equal to the iteration number threshold; If the iteration number is greater than or equal to the iteration number threshold, stop iteration, and take the center filter obtained in the last iteration as a standard filter; Obtaining the reagent amount corresponding to the standard filter, and taking the reagent amount corresponding to the standard filter as a standard reagent amount.

2. The intelligent control method of a denitrification filter according to claim 1, wherein Collecting filter data information of a plurality of target filters, preprocessing the filter data information to obtain preprocessed information, comprising: Creating a filter database; Collecting filter data information of a plurality of target filters respectively, and inputting all collected filter data information into the filter database; the filter data information includes filter information and reagent information; Randomly selecting a filter data information from the filter database; Determine whether there is duplicate data in the filter data information; If there is duplicate data in the filter data information, the duplicate data is removed; A filter data information is randomly selected from the filter database until all filter data in the filter database are selected, and a plurality of pre-processing information is obtained.

3. The intelligent control method of a denitrification filter according to claim 2, wherein Real-time filter data of a real-time filter is collected, a standard filter and a standard dosage corresponding to the real-time filter are obtained through the trained control model, and the real-time dosage of the real-time filter is adjusted in combination with the standard dosage, including: Real-time filter data of a real-time filter is collected; the real-time filter data includes real-time filter information and real-time dosage information; The real-time filter data is input into the trained control model to obtain a standard filter and a standard dosage corresponding to the real-time filter; The dosage of the real-time filter is adjusted in combination with the real-time dosage and the standard dosage.

4. The intelligent control method of a denitrification filter according to claim 3, wherein The real-time filter data is input into the trained control model to obtain a standard filter and a standard dosage corresponding to the real-time filter, including: The real-time filter is clustered through K-means clustering analysis in combination with the real-time filter information, so as to obtain a standard filter corresponding to the real-time filter; A standard dosage corresponding to the standard filter is obtained.

5. The intelligent control method of a denitrification filter according to claim 4, wherein The dosage of the real-time filter is adjusted in combination with the real-time dosage and the standard dosage, including: An error threshold is set; It is judged whether the error of the real-time dosage and the standard dosage is greater than or equal to the error threshold; If the error of the real-time dosage and the standard dosage is greater than or equal to the error threshold, the real-time dosage is adjusted until the error of the real-time dosage and the standard dosage is less than the error threshold; If the error of the real-time dosage and the standard dosage is less than the error threshold, the current state of the real-time filter is maintained.

6. An intelligent control system for a denitrification filter, characterized by, It includes: A collection component; Filter data information of a plurality of target filters is collected through the collection component; A control component; The control component executes the intelligent control method of the denitrification filter according to any one of claims 1 to 5, and establishes a standard filter and a corresponding standard dosage in combination with the filter data information through the control component, so as to adjust the real-time filter based on the standard filter.

7. The intelligent control system for a denitrification filter according to claim 6, wherein The collection component includes a filter collection module and a dosage collection module, and the filter information is collected through the filter collection module, and the dosage information is collected through the dosage collection module.

Images