Method for predicting norovirus removal rate of sewage treatment plant based on machine learning

Abstract

The invention provides a method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning. The method comprises the following steps: acquiring the treatment process and scale data of the current sewage treatment plant; performing result prediction through a pre-trained prediction model based on the processing process and scale data; determining the norovirus removal rate of the current sewage treatment plant and the norovirus concentration in the effluent based on the predicted result; wherein the pre-trained prediction model is obtained by training, verifying and testing a plurality of preset machine learning models through a pre-acquired data set, and the pre-acquired data set comprises treatment processes and scale data of a plurality of known sewage treatment plants, as well as influent norovirus concentration data and effluent norovirus concentration data. The problem that an existing sewage treatment prediction model needs to firstly perform long-time sequence sampling and detection on multiple water quality indexes in inflow water and outflow water of a current sewage treatment plant, and then the model can be established for prediction is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of sewage treatment, and in particular to a method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning. Background Art

[0002] Norovirus is present in high concentrations in both urban and rural sewage, demonstrating strong persistence at varying temperatures. Norovirus is highly variable, with multiple genotypes and variants. Infection with one norovirus type provides limited protection against infection with other types. Furthermore, the protective immune response to norovirus infection is short-lived and gradually wanes over time. Therefore, reinfection is possible after infection. Norovirus infection can cause vomiting and diarrhea, followed by nausea, abdominal pain, headache, fever, chills, and muscle aches. Despite continuous improvements in wastewater treatment technology and increasingly stringent testing of reclaimed and environmental water bodies, some regions have experienced large outbreaks of norovirus infections affecting dozens or even hundreds of people in recent years. Therefore, predicting norovirus levels in wastewater treatment plant effluent based on the plant's treatment process and scale is crucial for selecting reclaimed water reuse options.

[0003] Currently, there are no models for predicting norovirus levels in sewage treatment plant effluent or their removal rates. The closest available models are those for predicting fecal coliform counts and total coliform counts. However, these models can only be used for training on a single sewage treatment plant and require a long series of influent water quality data (including pH, biochemical oxygen demand (BOD), chemical oxygen demand (COD), total suspended particulate matter, oil and grease, total Kjeldahl nitrogen, ammoniacal nitrogen, and total phosphorus). These models are not generalizable, and their use in other sewage treatment plants requires long-term monitoring of multiple influent water quality data, resulting in high costs.

[0004] Therefore, there is an urgent need to develop a method to predict the norovirus removal rate in the effluent of sewage treatment plants, which can be used to predict the norovirus removal rate in the effluent of sewage treatment plants of any known treatment process type and scale without relying on water quality data and detailed treatment processes. Summary of the Invention

[0005] The present invention provides a method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning, which is used to solve the problem that the existing sewage treatment prediction model requires first conducting long-term series sampling and detection of multiple water quality indicators in the influent and effluent of the current sewage treatment plant before establishing a model for prediction.

[0006] The present invention provides a method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning, comprising: Obtain the treatment process and scale data of current sewage treatment plants; Predicting results using a pre-trained prediction model based on the processing technology and scale data; Based on the prediction results, the current sewage treatment plant's norovirus removal rate and the norovirus concentration in the effluent are determined; wherein, the pre-trained prediction model is obtained by training and verifying multiple preset machine learning models with a pre-acquired data set, and the pre-acquired data set includes the treatment process, scale data, and influent norovirus concentration and effluent norovirus concentration data of multiple known sewage treatment plants.

[0007] According to a method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning provided by the present invention, the pre-trained prediction model is obtained by training and verifying multiple preset machine learning models using a pre-acquired data set, including: Divide the pre-acquired dataset into training and test sets in proportion; Training and verifying a plurality of preset machine learning models using the training set, testing the training results using the test set, and generating test results; Based on the test results, a corresponding machine learning model is selected as a prediction model through preset evaluation indicators.

[0008] According to a method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning provided by the present invention, the treatment processes of multiple known sewage treatment plants in the pre-acquired data set include: the primary treatment process of the sewage treatment plant, the secondary treatment process of the sewage treatment plant, and the tertiary treatment process of the sewage treatment plant; the scale data of the sewage treatment plant includes the daily treatment capacity of the sewage treatment plant, the population served by the sewage treatment plant, and the local per capita GDP.

[0009] According to a method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning provided by the present invention, the inlet norovirus concentration and the effluent norovirus concentration data in the pre-acquired data set are obtained by quantitative polymerase chain reaction. If it is determined based on the measurement results that the effluent norovirus concentration in a certain data is higher than the inlet norovirus concentration, this data is deleted.

[0010] According to a method for predicting norovirus removal rate in a sewage treatment plant based on machine learning provided by the present invention, the method includes training and validating a plurality of preset machine learning models using the training set, testing the training results using the test set, and generating test results, including: Divide the training set into five equal subsets, use four of the subsets as training sets to train the machine learning model, use the remaining subset as a validation set to validate the machine learning model, and each subset is used as a validation set to perform cross-training validation on the machine learning model, obtaining five validation results, and fine-tuning the machine learning model based on the five validation results; The fine-tuned machine learning model is tested on the test set to generate test results.

[0011] According to a method provided by the present invention for predicting the norovirus removal rate of a sewage treatment plant based on machine learning, the method selects a corresponding machine learning model as a prediction model based on the test results and a preset evaluation index, including: Evaluate the corresponding machine learning model based on the test results using the coefficient of determination, root mean square error, mean absolute error, and mean square error to generate an evaluation result; Based on the evaluation results, the machine learning model with the best prediction effect is selected as the prediction model.

[0012] The present invention also provides a system for predicting the norovirus removal rate of a sewage treatment plant based on machine learning, the system comprising: Data acquisition module, used to obtain the treatment process and scale data of the current sewage treatment plant; A prediction module, configured to predict results based on the processing technology and scale data using a pre-trained prediction model; A result output module is used to determine the norovirus removal rate of the current sewage treatment plant and the norovirus concentration in the effluent based on the predicted results; Among them, the pre-trained prediction model is obtained by training and verifying multiple preset machine learning models with pre-acquired data sets. The pre-acquired data sets include treatment processes, scale data, and influent norovirus concentration and effluent norovirus concentration data of multiple known sewage treatment plants.

[0013] The present invention also provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the computer program, it implements any of the above-described methods for predicting the norovirus removal rate of a sewage treatment plant based on machine learning.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it implements any of the above-described methods for predicting the norovirus removal rate of a sewage treatment plant based on machine learning.

[0015] The present invention also provides a computer program product, comprising a computer program, which, when executed by a processor, implements any of the above-described methods for predicting the norovirus removal rate of a sewage treatment plant based on machine learning.

[0016] The present invention provides a method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning. The method predicts the norovirus removal rate and effluent norovirus concentration of the treatment process based on the treatment process and scale data of the sewage treatment plant based on a pre-trained prediction model. The method has low reliance on water quality data, a wide range of applications, and strong generalization. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 This is a flow chart of a method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning provided by the present invention.

[0019] Figure 2 This is a system architecture diagram for predicting the norovirus removal rate of a sewage treatment plant based on machine learning provided by the present invention.

[0020] Figure 3 It is a schematic diagram of the data set distribution of the model input data and target variables provided by the present invention.

[0021] Figure 4 This is a schematic diagram of a user interface for predicting the concentration of norovirus in the effluent of a sewage treatment plant provided by the present invention.

[0022] Figure 5 This is a schematic diagram of the user interface for predicting the concentration of norovirus in the effluent of a sewage treatment plant provided by the present invention in English.

[0023] Figure 6 It is a schematic diagram of the drop-down box for selecting the "secondary processing process" in the user interface processing process provided by the present invention.

[0024] Figure 7 It is a schematic diagram of the drop-down box for selecting the "three-level processing process" in the user interface processing process provided by the present invention.

[0025] Figure 8 This is a schematic diagram of the "Norovirus Genotype" drop-down box of the Norovirus treatment effect in the user interface provided by the present invention.

[0026] Figure 9This is an example diagram of the results of the user interface provided by the present invention where the "Norovirus genotype" is "GI" and the "Norovirus concentration entering the water" is known.

[0027] Figure 10 This is an example diagram of the result of the user interface provided by the present invention where the "Norovirus genotype" is "GI" but the "Norovirus concentration entering the water" is unknown.

[0028] Figure 11 This is an example diagram of the results of the user interface provided by the present invention when the "norovirus genotype" is unknown and the "norovirus concentration entering the water" is known.

[0029] Figure 12 This is an example diagram of the results in which both the "Norovirus genotype" and the "Norovirus concentration entering the water" in the user interface provided by the present invention are unknown.

[0030] Figure 13 This is a schematic diagram of module connections of a system for predicting norovirus removal rates in sewage treatment plants based on machine learning, provided by the present invention.

[0031] Figure 14 Schematic diagram of the structure of the electronic device provided by the present invention. 110: Data acquisition module; 120: Prediction module; 130: Result output module; 1410: Processor; 1420: Communication interface; 1430: Memory; 1440: Communication bus. DETAILED DESCRIPTION

[0032] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0033] The following combination Figure 1 and Figure 2 The present invention provides a method for predicting norovirus removal rates in a sewage treatment plant based on machine learning, comprising: Step 100: Obtain the treatment process and scale data of the current sewage treatment plant.

[0034] Step 200: Predicting results based on the processing technology and scale data using a pre-trained prediction model.

[0035] Step 300: Determine the current sewage treatment plant's norovirus removal rate and the norovirus concentration in the effluent based on the prediction results.

[0036] In the present invention, the pre-trained prediction model is obtained by training and verifying multiple preset machine learning models with a pre-acquired data set. The pre-acquired data set includes the treatment process, scale data, and influent norovirus concentration and effluent norovirus concentration data of multiple known sewage treatment plants.

[0037] Specifically, the treatment process of sewage treatment plants includes the primary treatment process of sewage treatment plants, the secondary treatment process of sewage treatment plants, and the tertiary treatment process of sewage treatment plants; the scale parameters of sewage treatment plants include the daily treatment capacity of sewage treatment plants (m 3 / d), population served by the sewage treatment plant (people) and local per capita GDP (US dollars); logarithmic removal rate = log 10 (Norovirus concentration in water (GC / L) / Norovirus concentration in water outflow (GC / L)).

[0038] Data on norovirus concentrations before and after treatment at sewage treatment plants were collected from 312 articles on the web of science using (((TS=(Effluent)) AND TS=(Norovirus), where TS is Topic, WWTP is sewage treatment plant, and Norovirus is norovirus. There were 52 articles providing norovirus concentrations in the influent and effluent of sewage treatment plants before and after treatment, with a total of 1,466 valid data.

[0039] Norovirus concentrations were uniformly determined using quantitative polymerase chain reaction (qPCR). Data on norovirus concentrations determined using other methods were deleted to eliminate errors introduced by different norovirus determination methods. When the effluent norovirus concentration was higher than the influent norovirus concentration, this data item was deleted to eliminate errors in norovirus determination. After deleting the problematic data, missing items in the treatment process and scale data were filled in, such as local per capita GDP (US dollars). If the literature on the treatment process and scale of the sewage treatment plant and the concentrations of norovirus in the influent and effluent were not provided, the per capita GDP of the sewage treatment plant location in the year when the norovirus concentrations in the influent and effluent of the sewage treatment plant were collected was first found to fill in the missing items. If there was no local per capita GDP, the per capita GDP of the country in that year provided by the World Bank was used to fill in the missing items.

[0040] After filling in missing data, all data except the logarithmic removal rate of norovirus were first logarithmized, followed by standardization and normalization preprocessing. Standardization subtracted the mean from the data and then divided it by the standard deviation to transform the data into a distribution with a mean of 0 and a standard deviation of 1. This is used for training models other than deep learning models. Normalization subtracted the minimum value from the data and then divided it by the difference between the maximum and minimum values ​​to transform the data into a number between [0 and 1]. This is used for training deep learning models. This dimensionless transformation eliminates differences in the degree of variation of each variable, ensuring that each transformed variable has the same initial weight when building subsequent machine learning models.

[0041] Finally, 1052 valid data were obtained, the specific data are as follows Figure 3 There are 7 independent variables (X1-X7): X1 is the concentration of norovirus in the water before treatment (GC / L), X2 is the secondary treatment process of the sewage treatment plant, X3 is the tertiary treatment process of the sewage treatment plant, X4 is the daily treatment capacity of the sewage treatment plant (m 3 / d), X5 is the population served by the sewage treatment plant (people), X6 is the local per capita GDP (US dollars), and X7 is the type of norovirus. The dependent variable Y1 is the logarithmic removal rate of norovirus (= log 10 (Norovirus concentration in influent (GC / L) / Norovirus concentration in effluent (GC / L)), while the Norovirus concentration in effluent after Y2-treatment (GC / L) itself is not included in the training as a variable.

[0042] Specifically, (X2) secondary treatment processes can be divided into 5 categories; (X3) tertiary treatment processes can be divided into 4 categories; (X7) Norovirus can be divided into four categories.

[0043] Classification of secondary treatment processes: 1) Anaerobic / Anoxic / Oxic process (AAO (Anaerobic / Anoxic / Oxic)); 2) Anaerobic / Aerobic process (AO (Anoxic / Oxic)); 3) Biofilm process (Biofilm); 4) Membrane Bioreactors (Membrane Bioreactors); 5) Others.

[0044] Classification of the three-stage treatment process: 1) Disinfection; 2) Disinfection + Sand Filtration; 3) Sand Filtration; 4) Others.

[0045] Norovirus genotype classification: 1) GI; 2) GII; 3) GIV; 4) Other types.

[0046] Since there are many secondary treatment processes, tertiary treatment processes and norovirus species, it is necessary to encode the treatment processes and norovirus species. Using common sequential encoding may lead to incorrect numerical relationships, and the machine learning model may mistakenly think that the treatment method that is ranked later is better than the treatment method that is ranked earlier. Therefore, in the present invention, the logarithmic removal rate mean of the corresponding treatment process and norovirus species is calculated, and the logarithmic removal rate mean is logarithmized with 10 as the base, and the calculated result is assigned to the corresponding treatment process type, that is, the values ​​of the variables secondary treatment process (X2), tertiary treatment process (X3) and norovirus species (X7) are obtained. The corresponding relationship between the secondary treatment process type and the variable X2 value is shown in Table 1, the corresponding relationship between the tertiary treatment process type and the variable X3 value is shown in Table 2, and the corresponding relationship between the norovirus species and the variable X7 value is shown in Table 3.

[0047] Table 1 .

[0048] Table 2 .

[0049] Table 3 .

[0050] refer to Figure 3 Since the distribution range of influent norovirus concentration, sewage treatment plant service population, local per capita GDP, sewage treatment plant daily treatment capacity and effluent norovirus concentration is large, the influent norovirus concentration, sewage treatment plant service population, local per capita GDP, sewage treatment plant daily treatment capacity and effluent norovirus concentration are first logarithmized with base 10.

[0051] In the present invention, the pre-trained prediction model is obtained by training and verifying multiple preset machine learning models using pre-acquired data sets, including: Divide the pre-acquired dataset into training and test sets in proportion; Training and verifying a plurality of preset machine learning models using the training set, testing the training results using the test set, and generating test results; Based on the test results, a corresponding machine learning model is selected as a prediction model through preset evaluation indicators.

[0052] Specifically, the training set data was divided into five equal subsets for cross-validation. In each round of cross-validation, the model was trained using the training set (four subsets) and evaluated using the validation set (the remaining subset). These five validation runs provided the model with evaluation metrics under five different validation sets. These evaluation results were then averaged to obtain the model's average performance. The final test set was used to test the model's generalization ability to unseen data and generate test results. Specifically, a machine learning model was used to learn the norovirus removal performance of multiple wastewater treatment plants with known treatment processes and scales, as well as norovirus concentrations in influent and effluent. Six machine learning algorithms were used to train the data and evaluate the results. These algorithms were Deep Learning, Distributed Random Forest (DRF), Generalized Linear Model (GLM), Gradient Boosting Machine (GBM), Extremely Randomized Trees (XRT), and Stacked Ensemble.

[0053] Since the Deep Learning algorithm and the GLM algorithm require different data processing before training, the independent variables except the logarithmic removal rate are normalized and standardized respectively; the normalized data is trained with the Deep Learning algorithm, while the standardized data is used for training other algorithms.

[0054] The data is randomly shuffled and divided into an 80% training set and a 20% test set using stratified sampling, ensuring that the distribution of the training and test sets is consistent. Five-fold cross-validation is used to train the data, dividing the training set into five equal subsets, each containing 20% ​​of the data. Each time, one of the subsets is used as the validation set, and the other four subsets are used as training data. This process is repeated, rotating different subsets as validation sets, until all subsets have been used as validation sets. Five validation results are obtained, and the machine learning model is fine-tuned based on these five validation results. The fine-tuned machine learning model is then tested on the test set to generate test results.

[0055] Based on the test results, a corresponding machine learning model is selected as a prediction model using preset evaluation indicators, including: Evaluate the corresponding machine learning model based on the test results using the coefficient of determination, root mean square error, mean absolute error, and mean square error to generate an evaluation result; Based on the evaluation results, the machine learning model with the best prediction effect is selected as the prediction model.

[0056] Specifically, four statistical indicators were used to evaluate the model performance: coefficient of determination (R 2 ), root mean square error (RMSE), mean absolute error (MAE) and mean square error (MSE). Among them, the coefficient of determination (R 2 ), root mean square error (RMSE), mean absolute error (MAE) and mean square error (MSE) are indicators for evaluating regression prediction algorithms: in, m is the total number of test datasets, i It is different test data; y i is the true value in the test dataset; is the predicted value in the test dataset; yes y i The average value of .

[0057] In one embodiment, according to the coefficient of determination (R 2 ), root mean square error (RMSE), mean absolute error (MAE) and mean square error (MSE) were used to evaluate the prediction model. Different machine learning models were used to predict the test set, and the evaluation indicators of the obtained prediction results were compared. The results are shown in Table 4. The optimal model is the model trained by the integrated algorithm.

[0058] Table 4 .

[0059] In another embodiment, when the type of norovirus is not considered, different machine learning models are used to predict the test set, and the evaluation indicators of the obtained prediction results are compared. The results are shown in Table 5, and the optimal model is the model trained by the extreme random forest algorithm.

[0060] Table 5 .

[0061] For a new sewage treatment plant, it is only necessary to input X1-the concentration of norovirus in the water before treatment (GC / L), X2-the secondary treatment process of the sewage treatment plant, X3-the tertiary treatment process of the sewage treatment plant, X4-the daily treatment capacity of the sewage treatment plant (m 3The optimal model trained on the model can predict Y1, the logarithmic removal rate of norovirus, and Y2, the norovirus concentration in the treated effluent (GC / L). Ultimately, the model can predict the treatment effect of any sewage treatment plant with a known treatment process and scale on a known concentration of norovirus.

[0062] Alternatively, for a new sewage treatment plant, when X1 - the concentration of norovirus in the influent before treatment (GC / L) is unknown, enter X2 - the secondary treatment process of the sewage treatment plant, X3 - the tertiary treatment process of the sewage treatment plant, X4 - the daily treatment capacity of the sewage treatment plant (m 3 The optimal model trained on the data set (X5, population served by the sewage treatment plant, X6, local per capita GDP (USD), and X7, norovirus genotype) can be used to predict Y1, the logarithmic removal rate of norovirus. Ultimately, this model can predict the treatment performance of any sewage treatment plant with a known treatment process and scale for an unknown concentration of norovirus. Since the required input data, X2-X7, are readily available from public datasets such as yearbooks, the predictive model's application scope is greatly expanded.

[0063] The optimal model obtained by training, that is, the model obtained by training the extreme random forest algorithm without classifying the norovirus type, and the model obtained by training the ensemble algorithm for classifying the norovirus type are saved for calling when the user interface is running.

[0064] Specifically, the user interface of "Prediction of Norovirus Concentration in Wastewater Treatment Plant Effluent" is divided into five modules, namely language selection module, treatment level selection module, treatment process selection module, sewage treatment plant scale module, and Norovirus treatment effect module (including the "Run" button).

[0065] The five modules are arranged in a simple and intuitive manner to facilitate users to input data.

[0066] refer to Figure 4 and Figure 5 The language selection module is set as a drop-down box with two built-in options: "Chinese" and "English". Users can click the option to switch the language of the entire user interface as needed.

[0067] The treatment process selection module is set as a panel with radio buttons, and users can choose from three built-in treatment levels, including "①", "①+②", and "①+②+③"; among them, ① corresponds to "Primary Treatment Process", ①+② corresponds to "Primary Treatment Process + Secondary Treatment Process", and ①+②+③ corresponds to "Primary Treatment Process + Secondary Treatment Process + Tertiary Treatment Process".

[0068] When "①" is selected, the drop-down boxes of "Secondary Treatment Process" and "Tertiary Treatment Process" in the treatment process selection module are automatically filled with "None / NA", and cannot be pulled down for selection; when "①+②" is selected, the drop-down box of "Secondary Treatment Process" in the treatment process selection module is set to normally open, and content can be selected, and the drop-down box of "Tertiary Treatment Process" is automatically filled with "None / NA", and cannot be pulled down for selection; when "①+②+③" is selected, the drop-down boxes of "Secondary Treatment Process" and "Tertiary Treatment Process" in the treatment process selection module are set to normally open, and content can be selected.

[0069] The treatment process selection module is divided into two rows, both of which are drop-down boxes, and the default selection is "None / NA". Figure 6 As shown in the figure, users can select the first row of "Secondary Treatment Process"; Figure 7 As shown, the user can select the second row "three-level treatment process".

[0070] The sewage treatment plant scale module has three numerical input boxes, arranged horizontally in the following order: "Daily Processing Capacity (m³ / d)", "Population Served (person)", and "Local GDP Per Capita ($)". Users can enter values ​​as needed.

[0071] The Norovirus treatment effect module sets up a drop-down box, a value input box, two value output boxes and a run button; Figure 8As shown in the figure, the user can select as needed in the "Norovirus Genotypes" drop-down box, and enter a value as needed in the "Influent Concentration (GC / L)" input box; when the user enters a value in the "Influent Concentration (GC / L)" input box, the output result includes "Outlet Norovirus Concentration (GC / L)"; when the user does not enter a value in the "Influent Concentration (GC / L)" input box, the output result does not include "Outlet Norovirus Concentration (GC / L)".

[0072] The data operated and input by the user on the interface are obtained, including "Treatment Level", "Secondary Treatment Process", "Tertiary Treatment Process", "Daily Processing Capacity (m³ / d)", "Population Served", "Local Capita Per GDP ($)", "Norovirus Genotypes", and "Influent Norovirus Concentration (GC / L)".

[0073] refer to Figure 9 When the user selects "GI", "GII", "GIV", or "Other types" in the "Norovirus Genotype" and enters a value in the "Influent Norovirus Concentration (GC / L)" input box, the user clicks the "Run" button to obtain the "Treatment Grade", "Secondary Treatment Process", "Tertiary Treatment Process", "Daily Treatment Capacity (m 3 / d)", "service population (people)", "local per capita GDP (US dollars)", "norovirus genotype", and "influent norovirus concentration (GC / L)" data were input into the test set in the pre-stored dataset as a new data. The model trained by the integrated algorithm was called to predict the results, and the results of "effluent norovirus concentration (GC / L)" and "logarithmic removal rate" were obtained.

[0074] refer to Figure 10 If the user selects "GI", "GII", "GIV", or "Other Types" in the "Norovirus Genotype" but does not enter a value in the "Influent Norovirus Concentration (GC / L)" input box, the user clicks the "Run" button to obtain the "Treatment Grade", "Secondary Treatment Process", "Tertiary Treatment Process", "Daily Treatment Capacity (m 3 / d)", "Service population (people)", "Local per capita GDP (US dollars)", and "Norovirus genotype" data are input as new data into the test set in the pre-stored data set, and the model trained by the integrated algorithm is called to predict the results and obtain the "logarithmic removal rate" result.

[0075] refer to Figure 11 When the user selects "Unknown" in the "Norovirus Genotype" and enters a value in the "Influent Norovirus Concentration (GC / L)" input box, the user clicks the "Run" button to obtain the "Treatment Grade", "Secondary Treatment Process", "Tertiary Treatment Process", "Daily Treatment Capacity (m 3 / d)", "service population (people)", "local per capita GDP (US dollars)", "norovirus genotype", and "influent norovirus concentration (GC / L)" data were input into the test set in the pre-stored dataset as a new data. The model trained by the extreme random forest algorithm was called to predict the results, and the results of "effluent norovirus concentration (GC / L)" and "logarithmic removal rate" were obtained.

[0076] refer to Figure 12 When the user selects "Unknown" in the "Norovirus Genotype" but does not enter a value in the "Influent Norovirus Concentration (GC / L)" input box, the user clicks the "Run" button to obtain the "Treatment Grade", "Secondary Treatment Process", "Tertiary Treatment Process", "Daily Treatment Capacity (m 3 / d)", "Service population (people)", "Local per capita GDP (US dollars)", and "Norovirus genotype" data are input as new data into the test set in the pre-stored dataset, and the model trained by the extreme random forest algorithm is called to predict the results and obtain the "logarithmic removal rate" result.

[0077] The present invention provides a method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning. The method predicts the norovirus removal rate and effluent norovirus concentration based on the treatment process and scale data of the sewage treatment plant through a pre-trained prediction model. The method has low reliance on water quality data, a wide range of applications, and strong generalization.

[0078] refer to Figure 13 The present invention also discloses a system for predicting the norovirus removal rate of a sewage treatment plant based on machine learning, the system comprising: The data acquisition module 110 is used to obtain the treatment process and scale data of the current sewage treatment plant; A prediction module 120, configured to predict results based on the processing technology and scale data using a pre-trained prediction model; A result output module 130 is used to determine the current norovirus removal rate of the sewage treatment plant and the norovirus concentration in the effluent based on the predicted results; Among them, the pre-trained prediction model is obtained by training and verifying multiple preset machine learning models with pre-acquired data sets. The pre-acquired data sets include treatment processes, scale data, and influent norovirus concentration and effluent norovirus concentration data of multiple known sewage treatment plants.

[0079] Pre-trained prediction models are obtained by training and validating multiple machine learning models using pre-acquired datasets, including: Divide the pre-acquired dataset into training and test sets in proportion; Training and verifying a plurality of preset machine learning models using the training set, testing the training results using the test set, and generating test results; Based on the test results, a corresponding machine learning model is selected as a prediction model through preset evaluation indicators.

[0080] The treatment processes of multiple known sewage treatment plants in the pre-acquired data set include: the primary treatment process of the sewage treatment plant, the secondary treatment process of the sewage treatment plant, and the tertiary treatment process of the sewage treatment plant; the scale data of the sewage treatment plant includes the daily treatment capacity of the sewage treatment plant, the population served by the sewage treatment plant, and the local per capita GDP.

[0081] The influent norovirus concentration and effluent norovirus concentration data in the pre-acquired data set are obtained by quantitative polymerase chain reaction. If the effluent norovirus concentration in a certain data is higher than the influent norovirus concentration based on the measurement results, this data will be deleted.

[0082] The plurality of preset machine learning models are trained and verified using the training set, and the training results are tested using the test set to generate test results, including: Divide the training set into five equal subsets, use four of the subsets as training sets to train the machine learning model, use the remaining subset as a validation set to validate the machine learning model, and each subset is used as a validation set to perform cross-training validation on the machine learning model, obtaining five validation results, and fine-tuning the machine learning model based on the five validation results; The fine-tuned machine learning model is tested on the test set to generate test results.

[0083] Based on the test results, a corresponding machine learning model is selected as a prediction model using preset evaluation indicators, including: Evaluate the corresponding machine learning model based on the test results using the coefficient of determination, root mean square error, mean absolute error, and mean square error to generate an evaluation result; Based on the evaluation results, the machine learning model with the best prediction effect is selected as the prediction model.

[0084] The present invention provides a method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning. The method predicts the norovirus removal rate and effluent norovirus concentration based on the treatment process and scale data of the sewage treatment plant through a pre-trained prediction model. The method has low reliance on water quality data, a wide range of applications, and strong generalization.

[0085] Treatment process Figure 14 An example of a physical structure diagram of an electronic device is shown below. Figure 14 As shown, the electronic device may include: a processor 1410, a communications interface 1420, a memory 1430, and a communications bus 1440. The processor 1410, the communications interface 1420, and the memory 1430 communicate with each other via the communications bus 1440. The processor 1410 may invoke logic instructions in the memory 1430 to execute a method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning. The method includes: obtaining treatment process and scale data for the current sewage treatment plant; predicting results using a pre-trained prediction model based on the treatment process and scale data; and determining the current sewage treatment plant norovirus removal rate and effluent norovirus concentration based on the predicted results. The pre-trained prediction model is obtained by training and validating multiple preset machine learning models using a pre-acquired dataset, which includes treatment process and scale data, as well as influent and effluent norovirus concentration data for multiple known sewage treatment plants.

[0086] Furthermore, the logic instructions in the aforementioned memory 1430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0087] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the method provided by the above methods for predicting the norovirus removal rate of a sewage treatment plant based on machine learning, the method including: obtaining the treatment process and scale data of the current sewage treatment plant; predicting the results based on the treatment process and scale data through a pre-trained prediction model; determining the norovirus removal rate of the current sewage treatment plant and the norovirus concentration in the effluent based on the predicted results; wherein the pre-trained prediction model is obtained by training and verifying multiple preset machine learning models through a pre-acquired data set, and the pre-acquired data set includes the treatment process, scale data and influent norovirus concentration and effluent norovirus concentration data of multiple known sewage treatment plants.

[0088] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the method provided by the above-mentioned methods for predicting the norovirus removal rate of a sewage treatment plant based on machine learning, the method comprising: obtaining the treatment process and scale data of the current sewage treatment plant; predicting the results based on the treatment process and scale data through a pre-trained prediction model; determining the norovirus removal rate of the current sewage treatment plant and the norovirus concentration in the effluent based on the predicted results; wherein the pre-trained prediction model is obtained by training and verifying multiple preset machine learning models through a pre-acquired data set, and the pre-acquired data set includes the treatment process, scale data, and influent norovirus concentration and effluent norovirus concentration data of multiple known sewage treatment plants.

[0089] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0090] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for predicting norovirus removal rate in a sewage treatment plant based on machine learning, characterized in that: include: Obtain the treatment process and scale data of current sewage treatment plants; Predicting results using a pre-trained prediction model based on the processing technology and scale data; Determine the current norovirus removal rate of the sewage treatment plant and the norovirus concentration in the effluent based on the predicted results; Among them, the pre-trained prediction model is obtained by training and verifying multiple preset machine learning models with pre-acquired data sets. The pre-acquired data sets include treatment processes, scale data, and influent norovirus concentration and effluent norovirus concentration data of multiple known sewage treatment plants.

2. The method for predicting norovirus removal rate in a sewage treatment plant based on machine learning according to claim 1, characterized in that: The pre-trained prediction model is obtained by training and verifying multiple preset machine learning models using pre-acquired data sets, including: Divide the pre-acquired dataset into training and test sets in proportion; Training and verifying a plurality of preset machine learning models using the training set, testing the training results using the test set, and generating test results; Based on the test results, a corresponding machine learning model is selected as a prediction model through preset evaluation indicators.

3. The method for predicting norovirus removal rate in a sewage treatment plant based on machine learning according to claim 1, characterized in that: The treatment processes of multiple known sewage treatment plants in the pre-acquired data set include: the primary treatment process of the sewage treatment plant, the secondary treatment process of the sewage treatment plant and the tertiary treatment process of the sewage treatment plant; the scale data of the sewage treatment plant includes the daily treatment capacity of the sewage treatment plant, the population served by the sewage treatment plant and the local per capita GDP.

4. The method for predicting norovirus removal rate in a sewage treatment plant based on machine learning according to claim 1, characterized in that: The influent norovirus concentration and effluent norovirus concentration data in the pre-acquired data set are obtained by quantitative polymerase chain reaction. If it is determined based on the measurement results that the effluent norovirus concentration in a certain data is higher than the influent norovirus concentration, this data is deleted.

5. The method for predicting norovirus removal rate in a sewage treatment plant based on machine learning according to claim 2, characterized in that: The training and verification of the preset multiple machine learning models using the training set, and the testing of the training results using the test set to generate the test results include: dividing the training set into five equal subsets, using four of the subsets as training sets to train the machine learning model, and using the remaining subset as a validation set to validate the machine learning model, and each subset is used as a validation set to perform cross-training validation on the machine learning model, obtaining five validation results, and fine-tuning the machine learning model based on the five validation results; The fine-tuned machine learning model is tested on the test set to generate test results.

6. The method for predicting norovirus removal rate in a sewage treatment plant based on machine learning according to claim 2, characterized in that: The selecting a corresponding machine learning model as a prediction model based on the test results through a preset evaluation index includes: Evaluate the corresponding machine learning model based on the test results using the coefficient of determination, root mean square error, mean absolute error, and mean square error to generate an evaluation result; Based on the evaluation results, the machine learning model with the best prediction effect is selected as the prediction model.

7. A system for predicting norovirus removal rates in sewage treatment plants based on machine learning, characterized in that: The system comprises: Data acquisition module, used to obtain the treatment process and scale data of the current sewage treatment plant; A prediction module, configured to predict results based on the processing technology and scale data using a pre-trained prediction model; The result output module is used to determine the norovirus removal rate of the current sewage treatment plant and the norovirus concentration in the effluent based on the predicted results; wherein, the pre-trained prediction model is obtained by training and verifying multiple preset machine learning models with a pre-acquired data set, and the pre-acquired data set includes the treatment process, scale data, and influent norovirus concentration and effluent norovirus concentration data of multiple known sewage treatment plants.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, it implements the method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning as described in any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, it implements the method for predicting the norovirus removal rate of a sewage treatment plant based on machine learning as described in any one of claims 1 to 6.

Images