Calibration method for water inlet on-line instrument and sewage treatment system

Through the water quality prediction model constructed by using machine learning algorithms in the sewage treatment plant, the rate of the online instrument data of the sewage plant water inlet is determined, which solves the problem of abnormal monitoring data, improves the safety and efficiency of the sewage treatment process, and reduces the operation Maintenance cost.

CN119023925APending Publication Date: 2024-11-26HANGZHOU BEISHUI FUTURE TECHNOLOGY CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411227284.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-03
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The monitoring data of the online instrument of the sewage plant water inlet is susceptible to environmental factors, resulting in abnormal data. Relying on experience to adjust process parameters has problems such as high energy consumption and safety hazards.

Method used

The water quality prediction model is constructed based on machine learning algorithms, and by obtaining online instrument data and laboratory test data in real time, data cleaning and definition of characteristic variable target variables, a random forest regression model is constructed to predict water quality parameters, as the process flow control in accordance with.

Benefits of technology

The accuracy of the incoming online instrument data is achieved, the impact of abnormal data is reduced, the safety and efficiency of the sewage treatment process is improved, and the operation and maintenance costs are reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119023925A_ABST
    Figure CN119023925A_ABST
Patent Text Reader

Abstract

The invention provides a calibration method for a water inlet on-line instrument and a sewage treatment system, and relates to the technical field of sewage treatment.The calibration method comprises the steps that water quality parameter data output by the water inlet on-line instrument are obtained in real time; according to the water quality parameter data, utilizing a pre-configured first water quality prediction model to predict corresponding water quality parameter true values to obtain first water quality parameter prediction data; feeding back and outputting the first water quality parameter prediction data as a water quality parameter calibration result; wherein the first water quality prediction model is constructed by adopting a machine learning algorithm based on historical monitoring data of an inlet water online instrument and laboratory test data of inlet water quality in a corresponding historical period. According to the technical scheme, on the basis that the real-time requirement of process control is met, the calibration result is closer to the true value, and safe and effective operation of the whole sewage treatment process is facilitated.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of sewage treatment, and in particular to a calibration method for an inlet online instrument and a sewage treatment system. Background Art

[0002] The calibration of an instrument is also called calibration. Through calibration, the error of the instrument can be corrected to make it closer to the true value. In the production and operation of a sewage treatment plant, online instruments are required to monitor the water quality of the sewage (also called influent or influent) that is to enter the sewage treatment process. The online water quality data obtained by monitoring is one of the basis for the regulation of the subsequent process control treatment process, which is of great significance.

[0003] In the process of realizing the present invention, the inventors found that in practice, the pollutant concentration of the sewage treatment plant influent is relatively high, and the content of impurities such as silt is high. The monitoring environment of the online instrument is poor, and its monitoring data is often affected by environmental factors (such as photoelectric online instruments are easily affected by factors such as silt and pollution), resulting in abnormal mutation values. Regular operation and maintenance of the instrument can reduce this situation to a certain extent. However, in actual scenarios, in some cases, for such anomalies, based on timeliness and cost considerations, the long-term operating experience of the water plant operators will be directly relied on to judge the abnormal situation, and the parameters of the sewage treatment process will be adjusted accordingly. Usually, higher power operating parameters will be selected to ensure that the effluent water quality meets the standards. This places high demands on the experience and quality of the operators, and is likely to lead to unnecessary energy consumption; and in this case, if there is no experienced and high-quality personnel to respond reasonably, production accidents may also occur, leading to adverse effects.

[0004] Therefore, how to use relevant technical means to calibrate the data processing of the monitoring data of the water inlet online instrument in this scenario to obtain data that is closer to reality, so as to effectively adjust the subsequent sewage treatment process and avoid the drawbacks of the above-mentioned existing treatment methods to the greatest extent has become a technical problem that needs to be solved urgently.

[0005] The above contents are only used to assist in understanding the technical solution of the present invention and do not constitute an admission that the above contents are prior art. Summary of the invention

[0006] In order to overcome the problems existing in the related art to at least a certain extent, the embodiments of the present application provide a calibration method for an online water inlet instrument and a sewage treatment system to help obtain water quality data that is closer to the actual water quality, so that the overall sewage treatment process can be carried out more safely and effectively.

[0007] First aspect The embodiment of the present application provides a calibration method for an inlet water online instrument, the calibration method comprising: Obtain water quality parameter data output by online inlet instrument in real time; According to the water quality parameter data, a pre-configured first water quality prediction model is used to predict the true value of the corresponding water quality parameter to obtain first water quality parameter prediction data; Feedback and output the first water quality parameter prediction data as a water quality parameter calibration result; Among them, the first water quality prediction model is constructed using a machine learning algorithm based on the historical monitoring data of the water inlet online instrument and the laboratory test data of the water inlet quality in the corresponding historical period.

[0008] In a possible implementation, the machine learning algorithm includes a random forest algorithm; and the preconfiguration process of the first water quality prediction model includes: Performing data cleaning on the historical monitoring data and the laboratory test data to obtain a processed intermediate data set; Based on the intermediate data set, a sample data set is prepared with the online monitoring parameter data as the characteristic variable and the laboratory test parameter data at the corresponding time node as the target variable; Based on the sample data set, the constructed random forest regression model is trained and optimized to obtain the first water quality prediction model.

[0009] In a possible implementation manner, the characteristic variables include chemical oxygen demand parameters, ammonia nitrogen parameters, total nitrogen parameters, and carbon-nitrogen ratio parameters measured by online instruments; The target variables include chemical oxygen demand parameters, ammonia nitrogen parameters, and total nitrogen parameters tested in the laboratory.

[0010] In a possible implementation manner, the process of performing data cleaning on the historical monitoring data and the laboratory test data includes: Perform statistical processing on the time series data of each parameter type in the laboratory test data, and determine the abnormal data in the test data of each parameter type based on the three-times standard deviation criterion; According to the time points corresponding to the abnormal data, the data at the corresponding time points in the historical monitoring data and the laboratory test data are removed.

[0011] In a possible implementation manner, the time length corresponding to the historical monitoring data is greater than or equal to two years.

[0012] Second aspect The embodiment of the present application provides a sewage treatment system, which includes a sewage treatment device; the sewage treatment device includes an online instrument arranged at a water inlet end of the device, and a first control device communicatively connected to the online instrument, wherein the first control device is used to control a process flow of the device based on monitoring data of the online instrument; and is characterized in that it also includes a second control device communicatively connected to the first control device; The first control device is configured to, when the communication connection with the second control device is successfully established, transmit the received data uploaded by the online instrument to the second control device, and use the water quality parameter calibration result fed back and output by the second control device as the basis for process control; The second control device is configured to, in response to establishment of a channel connection with the first control device, execute the steps of the above-mentioned rating method.

[0013] In a possible implementation manner, the first control device is further configured to, when the communication connection with the second control device is not successfully established, call a locally preset second water quality prediction model to predict the true value of the corresponding water quality parameter according to the received data uploaded by the online instrument, and obtain the second water quality parameter prediction data as a basis for process flow control; Among them, the second water quality prediction model is an empirical model constructed based on statistical analysis.

[0014] In a possible implementation, the empirical model is expressed based on the following expression: in, Indicates the predicted value of the chemical oxygen demand parameter at the current moment, Indicates the predicted value of ammonia nitrogen parameters at the current moment, represents the predicted value of the total nitrogen parameter at the current moment, Indicates the effective test value of the chemical oxygen demand parameter at the current moment. Indicates the effective test value of ammonia nitrogen parameter at the current moment. Indicates the effective test value of the total nitrogen parameter at the current moment. Indicates the online measurement value of the chemical oxygen demand parameter at the current moment, Indicates the online measurement value of ammonia nitrogen parameters at the current moment, Indicates the online measurement value of the total nitrogen parameter at the current moment, L1, L2, L3, H1, H2, H3, B1, B2, and B3 represent model application boundary parameters determined based on statistical analysis, and A represents model coefficient parameters determined based on statistical analysis.

[0015] In a possible implementation manner, the first control device is implemented based on a PLC controller; and the second control device is implemented based on an industrial computer.

[0016] The calibration method for the water inlet online instrument provided in the embodiment of the present application includes real-time acquisition of water quality parameter data output by the water inlet online instrument; predicting the real value of the corresponding water quality parameter using a pre-configured first water quality prediction model according to the water quality parameter data to obtain the first water quality parameter prediction data; feeding back and outputting the first water quality parameter prediction data as the water quality parameter calibration result; the first water quality prediction model is constructed using a machine learning algorithm based on the historical monitoring data of the water inlet online instrument and the laboratory test data of the water quality of the inlet water in the corresponding historical period. The technical solution of the present application, in the process control scenario of sewage treatment, for the water inlet online instrument, according to the online monitoring data of the instrument, a water quality prediction model pre-constructed based on a machine learning algorithm is used to predict the real value of the water quality parameter, and the prediction result is output as the parameter calibration result. The calibration of the water inlet online instrument achieved by this technical means, on the basis of meeting the real-time requirements of process control, the calibration result is closer to the real value, which is conducive to the safe and effective implementation of the overall sewage treatment process.

[0017] Other advantages, objectives, and features of the present application will be described in part in the following description, and in part will be apparent to those skilled in the art based on the following examination and study, or may be taught from the practice of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are used to provide a further understanding of the technical solution of the present application or the prior art, and constitute a part of the specification. Among them, the accompanying drawings expressing the embodiments of the present application are used together with the embodiments of the present application to explain the technical solution of the present application, but do not constitute a limitation on the technical solution of the present application.

[0019] Figure 1 A schematic diagram of a flow chart of a calibration method for an inlet water online instrument provided in one embodiment of the present application; Figure 2 A schematic diagram of a system block diagram of a sewage treatment system provided by an embodiment of the present application; Figure 3 for Figure 2 The diagram is a schematic illustration of the process flow involved in the calibration implementation of the sewage treatment system shown. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be described in detail below. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other implementation methods obtained by ordinary technicians in the field without making creative work belong to the scope of protection of the present application.

[0021] As described in the background technology, during the production and operation of a sewage treatment plant, online instruments need to be installed to monitor the water quality of the sewage (also called influent or influent) that is to enter the sewage treatment process device. The online water quality data obtained by monitoring is one of the basis for regulating the subsequent process control treatment process and is of great significance.

[0022] In the process of realizing the present invention, the inventors found that in practice, the pollutant concentration of the sewage treatment plant influent is relatively high, and the content of impurities such as silt is high. The monitoring environment of the online instrument is poor, and its monitoring data is often affected by environmental factors (such as photoelectric online instruments are easily affected by factors such as silt and pollution), resulting in abnormal mutation values. Regular operation and maintenance of the instrument can reduce this situation to a certain extent. However, in actual scenarios, in some cases, for such anomalies, based on timeliness and cost considerations, the long-term operating experience of the water plant operators will be directly relied on to judge the abnormal situation, and the parameters of the sewage treatment process will be adjusted accordingly. Usually, higher power operating parameters will be selected to ensure that the effluent water quality meets the standards. This places high demands on the experience and quality of the operators, and is likely to lead to unnecessary energy consumption; and in this case, if there is no experienced and high-quality personnel to respond reasonably, production accidents may also occur, leading to adverse effects.

[0023] Based on this, the present application proposes a calibration method for an inlet water online instrument to help obtain inlet water quality data that is closer to the actual situation, so that the overall sewage treatment process can be carried out more safely and effectively.

[0024] like Figure 1 As shown, in one embodiment, the method for calibrating an online instrument proposed in the present application includes: Step S110, real-time acquisition of water quality parameter data output by an online water inlet instrument. For example, the online instrument types here include a chemical oxygen demand (COD) meter, an ammonia nitrogen (NH3-N) meter, and a total nitrogen (TN) meter, and the acquired water quality parameter data include COD data, NH3-N data, and TN data; Step S120, predicting the true value of the corresponding water quality parameter using the pre-configured first water quality prediction model according to the water quality parameter data in step S110, to obtain first water quality parameter prediction data; Step S130, the first water quality parameter prediction data is fed back and output as a water quality parameter calibration result, so as to be used for regulating the overall sewage treatment process.

[0025] In the above process, the first water quality prediction model in step S120 is constructed based on the historical monitoring data of the inlet water online instruments (such as the chemical oxygen demand meter, ammonia nitrogen meter, and total nitrogen meter mentioned above) and the laboratory test data of the inlet water quality in the corresponding historical period, using a machine learning algorithm. The machine learning algorithm here includes a random forest algorithm, etc.

[0026] It is easy for those skilled in the art to understand that in the daily production of sewage treatment plants, the water quality of the influent is sampled and measured. The laboratory test data of the influent water quality in the above description refers to the data obtained by such sampling and measurement. In the present application, such data can be regarded as the true value of the influent water quality, that is, the technical solution of the present application adopts a machine learning method to obtain the relationship between the instrument measurement value and the true value, and then uses this relationship to realize the calibration of the instrument, so as to quickly obtain influent water quality parameters that are closer to the actual ones to meet the regulation needs.

[0027] The technical solution of the present application, in the process control scenario of sewage treatment, for the water inlet online instrument, according to the online monitoring data of the instrument, adopts a water quality prediction model pre-constructed based on a machine learning algorithm to predict the true value of the water quality parameter, and outputs the prediction result as the parameter calibration result. The calibration of the water inlet online instrument achieved by this technical means, on the basis of meeting the real-time requirements of process control, the calibration result is closer to the true value, which is conducive to the safe and effective implementation of the overall sewage treatment process.

[0028] In some embodiments, the machine learning algorithm used is a random forest algorithm, and the pre-configuration process of the first water quality prediction model includes: Step (a), performing data cleaning on historical monitoring data and laboratory test data to obtain a processed intermediate data set; As a specific implementation method, the above-mentioned data cleaning process includes performing statistical processing on the time series data of each parameter type in the laboratory test data, and determining the abnormal data in the test data of each parameter type based on the triple standard deviation criterion; according to the time point corresponding to each abnormal data, the data at the corresponding time point in the historical monitoring data and the laboratory test data are eliminated to obtain an intermediate data set.

[0029] For example, for a sewage treatment plant in a certain city, the online historical data of the plant's water inlet instrument for two years or more (such as 2018 to 2021) and the laboratory test data of the corresponding time are selected (that is, to ensure the needs of subsequent model construction and training, the time length corresponding to the historical monitoring data is greater than or equal to two years); for example, the types of water quality parameters measured by the online instrument here include chemical oxygen demand (COD), ammonia nitrogen (NH3-N) and total nitrogen (TN). Correspondingly, the laboratory test data selects the water quality laboratory test data of COD, NH3-N and TN in the period from 2018 to 2021; then, for each test data parameter, three times the standard deviation is used to eliminate outliers, that is, if a data of the water quality parameter Y Outside the range ( and are the arithmetic mean and standard deviation of parameter Y respectively), then eliminate The entire data of its corresponding time node, that is, if the COD parameter data at a certain time point is an abnormal value, other parameter type data at that time point will also be removed as abnormal values.

[0030] Then, step (b) is performed to prepare a sample data set based on the intermediate data set obtained in step (a), using the online monitoring parameter data as the feature variable and the laboratory test parameter data at the corresponding time node as the target variable; Specifically, in the technical solution of the present application, based on the process requirements of the sewage treatment scenario and the correlation between the influent water quality parameters, the characteristic variables include the chemical oxygen demand parameter, ammonia nitrogen parameter, total nitrogen parameter and carbon-nitrogen ratio parameter (i.e., the ratio of the chemical oxygen demand parameter to the total nitrogen parameter) measured by the online instrument. In the technical scenario of the present application, the carbon-nitrogen ratio represents the ratio of organic matter to total nitrogen in the sewage, directly affects the growth and metabolism of microorganisms in the sewage, is related to the amount of carbon source added in the sewage treatment process, and is an important parameter in the optimization of the sewage treatment process; and the target variables include the chemical oxygen demand parameter, ammonia nitrogen parameter, and total nitrogen parameter measured by the laboratory test respectively; Based on the selection of the above-mentioned feature variables and target variables, the corresponding sample data set can be produced according to the intermediate data set. It is easy to understand that since the target variables are three independent ones, three sample data sets can be produced for the subsequent construction and training of three random forest regression models. These three models can be expressed in the following expression form: (1) (2) (3) In expressions (1), (2), and (3), Characteristic variable X = [COD OL` , NH3OL` , T.N. OL` , C / N OL` ], COD OL` Indicates the chemical oxygen demand measured by the online instrument, NH3 OL` Indicates the ammonia nitrogen measured online, TN OL` Indicates total nitrogen measured online, C / N OL` represents the carbon-nitrogen ratio determined based on online measurement data, It represents the chemical oxygen demand in laboratory tests. Indicates the ammonia nitrogen tested in the laboratory. represents the total nitrogen tested in the laboratory, and F, G, and H represent the regression relationships of the corresponding models.

[0031] After step (b), step (c) is performed to train and optimize the constructed random forest regression model based on the prepared sample data set to obtain a first water quality prediction model. It is easy to understand that based on the above expressions (1), (2), and (3), there are three first water quality prediction models here, which are used for the prediction of chemical oxygen demand, ammonia nitrogen, and total nitrogen, respectively.

[0032] In the technical solution of this application, the training and tuning method of the random forest regression model is similar to that of the prior art. Here, from the perspective of implementation, the process is briefly described as follows: In this application, the machine learning library scikit-learn provided by the Python language is used to train the relationship between the feature variables and the target variables in the above expressions (1)-(3) based on the training set of the corresponding sample data set (e.g., the sample data set is divided and 80% of the sample data is randomly selected as the training set).

[0033] Since the values ​​of different parameters in the random forest regression model have a great influence on the model accuracy, it is necessary to optimize the relevant parameters to reduce the generalization error of the model and improve the accuracy. Random forest is a classifier that uses multiple decision trees to train and predict samples, and integrates multiple decision trees through the bagging framework. The tuning process can be divided into bagging framework parameter tuning and decision tree parameter tuning.

[0034] The bagging framework represents the integration method of the decision tree model. Since there is no dependency among the weak learners in the bagging framework, the difficulty of adjusting parameters is relatively low. Usually, only the maximum number of decision trees n_estimator needs to be tuned. The important parameters of the decision tree model include max_features, max_depth, min_samples_split, and min_samples_leaf. Among them, max_features refers to the maximum number of features considered when building the optimal model of the decision tree. There are three modes: auto, sqrt, and log2, which represent that the maximum number of features is N2 or log2N respectively; max_depth refers to the depth of the decision tree, which can be an integer or None. When None is selected, it means that the decision tree does not limit the depth of the subtree when building the optimal model; min_samples_split refers to the minimum number of samples that can be divided into a node, and its default value is 2; min_samples_leaf refers to the minimum number of samples contained in a leaf node. In the technical solution of this application, the above parameters are selected for optimization and adjustment.

[0035] Specifically, the above parameters can be tuned using the grid search method. For each parameter, set its parameter adjustment range and step size for debugging; randomly select 80% of the sample data in the sample data set as the training set and the remaining 20% ​​as the test set, and use the mean square error (MSE) of the prediction results of the random forest regression model obtained based on the training set for the test set and the actual test set as the accuracy indicator for accuracy evaluation (the lower the MSE, the higher the model accuracy), and select the parameters corresponding to the model with the highest accuracy as the preferred parameters.

[0036] Since the random forest model can be understood as embedding the decision tree model into the bagging framework, in order to reduce the computing power consumed by the grid search method, firstly, the n_estimator parameter of the outer bagging framework is optimized, and its value is changed in turn according to the set value range and change step size, and other parameters are set to constants (default values); then the grid search method is used to optimize the parameters of the inner decision tree model. Finally, the optimal parameter combination is selected and fixed as the model parameters of each water quality parameter prediction model.

[0037] In a specific embodiment, the optimal parameter combination of the obtained model is shown in Table 1 below: Table 1 - Optimal parameter combinations for parameter prediction models In one embodiment, if Figure 2As shown, the present application also proposes a sewage treatment system, which includes a sewage treatment device; the sewage treatment device includes an online instrument (such as Figure 2 ), and a first control device in communication with the online instrument, the first control device being used to control the process flow of the device based on the monitoring data of the online instrument ( Figure 2 The legends S1, S2, etc. indicate related control objects, such as aeration devices, sludge dosing devices, etc.), for example, the first control device here can be implemented by a PLC controller; like Figure 2 As shown, the sewage treatment system further includes a second control device that is communicatively connected to the first control device. For example, the second control device here can be implemented using an industrial control computer; In this embodiment, the first control device is configured to, when the communication connection with the second control device is successfully established, transmit the received data uploaded by the online instrument to the second control device, and use the water quality parameter calibration result fed back and output by the second control device as the basis for process flow control; Correspondingly, the second control device is configured to, in response to the establishment of the access connection with the first control device, execute the steps of the method for calibrating the online instrument in the above method embodiment; Regarding the calibration method executed by the second control device in this embodiment, its specific method has been described in detail in the embodiment of the method, and will not be elaborated here.

[0038] The sewage treatment system of the technical solution of the present application adds a second control device to the existing sewage treatment system, and makes corresponding adjustments to the software configuration. Based on the second control device, specific online water quality parameter calibration processing is performed. Since the calibration result is closer to the true value, it is beneficial to the safe and effective implementation of the overall sewage treatment process. On the other hand, the use of software to achieve online calibration of water quality monitoring data has to a certain extent solved the data anomaly caused by the lack of regular operation and maintenance of the sewage plant's online water inlet instruments, which is beneficial to reducing hardware operation and maintenance costs. In addition, the independent setting of the second control device is conducive to the low-cost upgrade and transformation of the existing system devices, and the promotion and application of the present technical solution.

[0039] Furthermore, taking into account the possibility of communication anomalies in the actual production environment, in order to ensure the normal and effective implementation of the control process, the first control device is also configured to, when the communication connection with the second control device is not successfully established, call the locally preset second water quality prediction model to predict the true value of the corresponding water quality parameter based on the data uploaded by the received online instrument, and obtain the second water quality parameter prediction data as the basis for process flow control; wherein the second water quality prediction model is an empirical model constructed based on statistical analysis.

[0040] like Figure 3 As shown, this is a schematic diagram of the overall process of implementing the rated treatment of the sewage treatment system.

[0041] Specifically, because the first control device based on the PLC controller can generally only perform relatively simple numerical calculation processing, the second water quality prediction model here can be expressed based on the following expression: (4) (5) (6) In the above expressions (4), (5), and (6), Indicates the predicted value of the chemical oxygen demand parameter at the current moment, Indicates the predicted value of ammonia nitrogen parameters at the current moment, represents the predicted value of the total nitrogen parameter at the current moment, Indicates the current effective test value of the chemical oxygen demand parameter. Indicates the current effective test value of ammonia nitrogen parameters. Indicates the current effective test value of the total nitrogen parameter. Indicates the online measurement value of the chemical oxygen demand parameter at the current moment, Indicates the online measurement value of ammonia nitrogen parameters at the current moment, Indicates the online measurement value of the total nitrogen parameter at the current moment, L1, L2, L3, H1, H2, H3, B1, B2, and B3 represent model application boundary parameters determined based on statistical analysis, and A represents model coefficient parameters determined based on statistical analysis.

[0042] For example, taking a water plant in northern my country as an example, we collected and sorted out its effective online water inlet instrument and test data in the past two years. After removing abnormal values, a total of 711 valid data were obtained. Based on the statistical analysis of these historical water quality online data, the boundary parameters (all in mg / L) that can be determined are L1=500, L2=35, L3=45, H1=70, H2=15, H3=20, B1=3.5, B2=1, B3=0.5, and the model coefficient parameter A=4; To facilitate understanding of the second water quality prediction model in this application, the following is further explained in conjunction with this example: Taking expression (4) as an example, when the current online COD parameter exceeds the empirical maximum or minimum value, the current COD effective test value is used as the predicted value; and when the current COD OL / TN OL When ≤3.5, four times the current total nitrogen online parameter value is used as the predicted value; in other cases, the current COD online parameter value is directly used as the predicted value; the cases of expressions (5) and (6) are similar; It should be noted that the laboratory data of general sewage treatment plants are sampled and measured once a day at 9 am. Accordingly, the current valid laboratory data mentioned above are the laboratory data measured on that day (that is, the current here is not the same as the current of the online instrument). It is easy to understand that if the data for the day is not released, the valid data from the previous day or the most recent time can be used.

[0043] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by any person familiar with the technology within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A calibration method for inlet water online instrumentation, It is characterized in that include: Obtain water quality parameter data output by online inlet instrument in real time; According to the water quality parameter data, a pre-configured first water quality prediction model is used to predict the true value of the corresponding water quality parameter to obtain first water quality parameter prediction data; Feedback and output the first water quality parameter prediction data as a water quality parameter calibration result; Among them, the first water quality prediction model is constructed using a machine learning algorithm based on the historical monitoring data of the water inlet online instrument and the laboratory test data of the water inlet quality in the corresponding historical period.

2. The calibration method for an inlet water online instrument according to claim 1, in, The machine learning algorithm includes a random forest algorithm; the pre-configuration process of the first water quality prediction model includes: Performing data cleaning on the historical monitoring data and the laboratory test data to obtain a processed intermediate data set; Based on the intermediate data set, a sample data set is prepared with the online monitoring parameter data as the characteristic variable and the laboratory test parameter data at the corresponding time node as the target variable; Based on the sample data set, the constructed random forest regression model is trained and optimized to obtain the first water quality prediction model.

3. The calibration method for an inlet water online instrument according to claim 2, in, The characteristic variables include chemical oxygen demand parameters, ammonia nitrogen parameters, total nitrogen parameters and carbon-nitrogen ratio parameters measured by online instruments; The target variables include chemical oxygen demand parameters, ammonia nitrogen parameters, and total nitrogen parameters tested in the laboratory.

4. The calibration method for an inlet water online instrument according to claim 2, in, The process of performing data cleaning on the historical monitoring data and the laboratory test data includes: Perform statistical processing on the time series data of each parameter type in the laboratory test data, and determine the abnormal data in the test data of each parameter type based on the three-times standard deviation criterion; According to the time points corresponding to the abnormal data, the data at the corresponding time points in the historical monitoring data and the laboratory test data are removed.

5. The calibration method for an inlet water online instrument according to claim 1, in, The time length corresponding to the historical monitoring data is greater than or equal to two years.

6. A sewage treatment system, comprising a sewage treatment device; the sewage treatment device comprises an online instrument arranged at a water inlet of the device, and a first control device in communication with the online instrument, wherein the first control device is used to control a process flow of the device based on monitoring data of the online instrument; It is characterized in that Also included is a second control device communicatively connected to the first control device; The first control device is configured to, when the communication connection with the second control device is successfully established, transmit the received data uploaded by the online instrument to the second control device, and use the water quality parameter calibration result fed back and output by the second control device as the basis for process control; The second control device is configured to, in response to establishment of a channel connection with the first control device, execute the steps of the rating method according to any one of claims 1 to 5.

7. The sewage treatment system according to claim 6, in, The first control device is further configured to, when the communication connection with the second control device is not successfully established, call a locally preset second water quality prediction model to predict the true value of the corresponding water quality parameter according to the received data uploaded by the online instrument, and obtain the second water quality parameter prediction data as the basis for process flow control; Among them, the second water quality prediction model is an empirical model constructed based on statistical analysis.

8. The sewage treatment system according to claim 7, in, The empirical model is expressed based on the following expression: in, Indicates the predicted value of the chemical oxygen demand parameter at the current moment, Indicates the predicted value of ammonia nitrogen parameters at the current moment, represents the predicted value of the total nitrogen parameter at the current moment, Indicates the effective test value of the chemical oxygen demand parameter at the current moment. Indicates the effective test value of ammonia nitrogen parameter at the current moment. Indicates the effective test value of the total nitrogen parameter at the current moment. Indicates the online measurement value of the chemical oxygen demand parameter at the current moment, Indicates the online measurement value of ammonia nitrogen parameters at the current moment, Indicates the online measurement value of the total nitrogen parameter at the current moment, L1, L2, L3, H1, H2, H3, B1, B2, and B3 represent model application boundary parameters determined based on statistical analysis, and A represents model coefficient parameters determined based on statistical analysis.

9. The sewage treatment system according to claim 6, in, The first control device is implemented based on a PLC controller; the second control device is implemented based on an industrial computer.