A method and system for reconstructing wastewater biological treatment process based on machine learning

Abstract

The present invention discloses a method and system for reconstructing a wastewater biological treatment process based on machine learning, the method comprising: collecting basic water quality, ecological risk and process information data of influent and effluent of the wastewater biological treatment process and performing data preprocessing; extracting features of the wastewater biological treatment process and performing multi-unit characterization; constructing a basic effluent water quality prediction model and an effluent ecological risk prediction model for different biological treatment multi-unit combination processes and performing compliance inspection; simulating and reconstructing different wastewater biological treatment short-range unit combination processes, and performing effluent water quality compliance inspection under different influent water quality scenarios; evaluating the ecological risk of the reconstructed biological treatment short-range unit combination process with effluent water quality compliance, and determining the optimal wastewater biological treatment process. The present invention can determine a new biological treatment process that meets the basic effluent water quality standards and has the lowest biological risk for different influent water quality scenarios, and solve the selection limitations of existing biological treatment process methods.

Description

Technical Field

[0001] The present invention relates to wastewater treatment technology, and in particular to a method and system for reconstructing a wastewater biological treatment process based on machine learning. Background Art

[0002] Sewage treatment plants are important places for solving the problem of purification and discharge of industrial and domestic sewage, and biological treatment is one of the most widely used wastewater treatment technologies in sewage treatment plants. Among them, the most common biological treatment processes include anaerobic-anoxic-aerobic or anoxic-anaerobic-aerobic A2O process, anoxic-aerobic or anaerobic-aerobic AO process, oxidation ditch OD process, membrane bioreactor MBR process, and circulating activated sludge CAST process.

[0003] At present, there are few studies on the reconstruction of new wastewater biological treatment processes. Most of the studies are only applicable to specified application scenarios, and the implementation conditions are relatively harsh. Chinese patent application number 202310107389.9 discloses a wastewater biological treatment process for the simultaneous removal of nitrogen, phosphorus and sulfur. This method achieves the simultaneous removal of nitrogen, sulfur and phosphorus pollutants at a low carbon-nitrogen ratio through alternating reactions in anaerobic and anoxic stages. Chinese patent application number 200810104988.0 discloses a method for rapid realization of short-term deep denitrification by SBR method. The method includes water intake, aeration, adding an equal amount of raw sewage, aeration, adding an equal amount of raw sewage again, aeration, adding an external carbon source, aeration, sedimentation and drainage, and idle process to achieve short-term deep denitrification of low carbon-nitrogen ratio wastewater by SBR method. The Chinese patent application number 200810104989.5 discloses a method for achieving short-range deep denitrification under low temperature conditions, which includes short-range deep denitrification startup, stabilization and cooling stages, including water intake, aeration, denitrification, sedimentation, drainage, sludge discharge and idle steps, and achieves stable denitrification under low temperature conditions. The above methods are all optimizations and modifications of existing biological treatment process methods, and are used to enhance the water purification capacity of one or several pollutants in specific application scenarios by adding exogenous substances, changing reaction conditions, etc.; when it cannot meet the needs of water purification and ecological risk control in different application scenarios, there are still many limitations and it cannot meet the current growing demand for wastewater treatment. Summary of the invention

[0004] Purpose of the invention: The present invention aims to provide a method and system for reconstructing wastewater biological treatment processes based on machine learning, which can determine a new biological treatment process that meets the basic effluent water quality standards and has the lowest biological risk for different influent water quality scenarios, thereby solving the selection limitations of existing biological treatment process methods.

[0005] Technical solution: The method for reconstructing a wastewater biological treatment process based on machine learning described in the present invention comprises the following steps:

[0006] (1) Collect basic water quality, ecological risk and process information data of influent and effluent in the wastewater biological treatment process and perform data preprocessing;

[0007] (2) feature extraction and multi-unit characterization of wastewater biological treatment processes;

[0008] (3) Construct basic effluent water quality prediction models and effluent ecological risk prediction models for different biological treatment multi-unit combination processes and conduct compliance tests;

[0009] (4) Simulate and reconstruct different short-range unit combination processes for wastewater biological treatment, and conduct effluent water quality compliance tests under different influent water quality scenarios;

[0010] (5) Evaluate the ecological risk of the reconstructed biological treatment short-range unit combination process to ensure that the effluent water quality meets the standards and determine the optimal wastewater biological treatment process.

[0011] Preferably, in step (1), the basic water quality of the inlet and outlet water of the wastewater biological treatment process includes chemical oxygen demand, ammonia nitrogen, total nitrogen, total phosphorus, suspended solids, biochemical oxygen demand, animal and plant oils, petroleum, nitrate nitrogen, total organic carbon, dissolved oxygen, temperature, pH value, and fecal coliform count; the ecological risk data of the inlet and outlet water include micropollutant content, biological toxicity, etc.; and the process information of the wastewater biological treatment process includes activated sludge wastewater biological treatment process information data.

[0012] Preferably, step (2) comprises:

[0013] The biological pool link of the activated sludge wastewater biological treatment process is characterized by extracting characteristics, and the biological pool link is divided into three characteristic units: aerobic pool, anoxic pool and anaerobic pool according to dissolved oxygen; the activated sludge wastewater biological treatment process is characterized by multiple units according to the extracted characteristic units.

[0014] Preferably, the multi-unit characterization method comprises a one-hot algorithm.

[0015] Preferably, the multi-unit characterization of the bioprocess using a one-hot algorithm includes:

[0016] First, the aerobic tank unit is characterized as [1,0,0], the anoxic tank unit is characterized as [0,1,0], the anaerobic tank unit is characterized as [0,0,1], and the blank unit is characterized as [0,0,0];

[0017] Secondly, the biological pool units were characterized and the activated sludge wastewater biological treatment process was uniformly characterized by n units. When the number of units for the characterized process was insufficient, blank units were used to supplement it.

[0018] Preferably, the step (3) of constructing a basic water quality prediction model for effluent from different biological treatment multi-unit combination processes includes:

[0019] Determine the input influent water quality feature set and preset the effluent water quality target; select different machine learning algorithms, and based on the input influent water quality feature set, establish effluent basic water quality prediction models for different biological treatment multi-unit combination processes, and conduct cross-validation experiments to evaluate the performance of the prediction model to obtain the optimal effluent basic water quality prediction model; screen out the effluent ecological risk prediction model that achieves the preset effluent water quality target. Among them, the effluent water quality targets include but are not limited to effluent total phosphorus, effluent total nitrogen, ammonia nitrogen, nitrate nitrogen, biochemical oxygen demand, total organic carbon, etc. One or more different effluent basic water quality prediction models can be constructed and screened for compliance inspection.

[0020] Preferably, the step (3) of constructing the effluent ecological risk prediction model of different biological treatment multi-unit combination processes includes:

[0021] Determine the input influent water quality feature set and preset ecological risk targets; select different machine learning algorithms, and establish effluent ecological risk prediction models for different biological treatment multi-unit combination processes based on the input influent water quality feature set, and conduct cross-validation experiments to evaluate the performance of the prediction model to obtain the optimal effluent ecological risk prediction model; screen out the effluent ecological risk prediction model that meets the preset ecological risk targets. The ecological risk targets include but are not limited to micropollutant content, biological toxicity, etc.

[0022] Preferably, when performing the cross-validation experiment to evaluate the performance of the prediction model, the following three evaluation indicators are used, and the calculation method of each indicator is:

[0023]

[0024] Among them, y i represents the true value, represents the predicted value, represents the mean of the sample, MAE represents the mean absolute error between the predicted value and the true value, MSE represents the mean square error between the predicted value and the true value, R 2 Represents the overall performance of the prediction model.

[0025] Preferably, step (4) comprises:

[0026] (4.1) Simulate and reconstruct all possible short-range unit combination processes for wastewater biological treatment, and uniformly characterize the reconstructed short-range unit combinations by n units;

[0027] (4.2) Under different influent water quality scenarios, the effluent water quality of each reconstructed biological treatment short-range unit combination process is predicted respectively, and based on whether the predicted results meet the preset effluent water quality target, the reconstructed biological treatment short-range unit combination process with effluent water quality that meets the standard is selected.

[0028] The present invention provides a wastewater biological treatment process reconstruction system based on machine learning, comprising:

[0029] Data collection and preprocessing module, used to collect basic water quality, ecological risk and process information data of influent and effluent in the wastewater biological treatment process and perform data preprocessing;

[0030] Feature extraction and characterization module, used to extract features and perform multi-unit characterization of wastewater biological treatment processes;

[0031] Prediction model building module, used to build effluent basic water quality prediction models and effluent ecological risk prediction models for different biological treatment multi-unit combination processes;

[0032] The wastewater biological treatment process reconstruction and compliance inspection module is used to simulate and reconstruct all possible wastewater biological treatment short-range unit combination processes, and conduct effluent water quality compliance inspections under different influent water quality scenarios to screen out the reconstructed biological treatment short-range unit combination processes with effluent water quality that meets the standards;

[0033] The wastewater biological treatment process determination module is used to evaluate the ecological risk of the reconstructed biological treatment short-range unit combination process that meets the effluent water quality standards based on the effluent ecological risk prediction model, and determine the optimal wastewater biological treatment process based on the ecological risk control effect.

[0034] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: 1. The biological treatment process is characterized by multiple units, and then a reasonable and effective biological treatment process reconstruction method is proposed. This method can determine a new biological treatment process that meets the basic effluent water quality standards and has the lowest biological risk for different influent water quality scenarios, thereby solving the selection limitations of existing biological treatment process methods; 2. A model is established using the knowledge and process data of the wastewater treatment process, which can accurately predict the effluent water quality and biological risk indicators of the biological treatment process of the sewage treatment plant for different influent water quality scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0036] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0037] like Figure 1As shown, the method for reconstructing a wastewater biological treatment process based on machine learning according to the present invention comprises the following steps:

[0038] (1) Collect basic water quality, ecological risk and process information data of the wastewater biological treatment process of the sewage treatment plant and perform data preprocessing.

[0039] The basic water quality, ecological risk and biological treatment process information data of the inlet and outlet water of the sewage treatment plant are collected separately. The basic water quality characteristics of the inlet and outlet water include chemical oxygen demand, ammonia nitrogen, total nitrogen, total phosphorus, suspended solids, biochemical oxygen demand, animal and plant oils, petroleum, nitrate nitrogen, total organic carbon, dissolved oxygen, temperature, pH value, fecal coliform count, etc.; the ecological risk data of the inlet and outlet water include the content of micropollutants, biological toxicity, etc.; the activated sludge wastewater biological treatment process information data used by the sewage treatment plant are obtained.

[0040] The basic water quality and ecological risk data of the sewage treatment plant are preprocessed. The data preprocessing methods include abnormal data processing, default data processing, outlier data processing, repeatability and consistency detection, data normalization and standardization, etc.

[0041] (2) Feature extraction and multi-unit characterization of the wastewater biological treatment process.

[0042] The biological pool of the activated sludge wastewater biological treatment process of the sewage treatment plant is characterized by extraction; specifically, the biological pool is divided into three characteristic units: aerobic pool, anoxic pool and anaerobic pool according to dissolved oxygen. Then, a multi-unit characterization combination of the activated sludge wastewater biological treatment process is performed using methods other than the one-hot algorithm.

[0043] Among them, the method of using the one-hot algorithm to characterize the multi-unit of the biological treatment process is:

[0044] First, the aerobic tank unit is characterized as [1,0,0], the anoxic tank unit is characterized as [0,1,0], the anaerobic tank unit is characterized as [0,0,1], and the blank unit is characterized as [0,0,0].

[0045] Secondly, the activated sludge wastewater biological treatment process is uniformly characterized by n units. When the number of units in the biological pool of the characterized process is less than n, blank units are used to supplement it. The characteristic dimension of the n-unit representation is 3×n. For example, the "three units" of the anaerobic-anoxic-oxic process in the wastewater biological treatment process are represented as [0,0,1,0,1,0,1,0,0]; the "two units" of the anaerobic-oxic process are represented as [0,0,1,1,0,0], and the "three units" are represented as [0,0,1,1,0,0,0,0,0].

[0046] (3) Construct basic effluent water quality prediction models and effluent ecological risk prediction models for different biological treatment n-unit combination processes and conduct compliance tests.

[0047] The specific steps include:

[0048] (3.1) Construct the basic water quality prediction model of effluent from different biological treatment n-unit combination processes, including:

[0049] From the basic water quality characteristics of the biological treatment process preprocessed with data in step (1), determine the input inlet water quality characteristic set and preset the effluent water quality target. The effluent water quality target includes ammonia nitrogen, total nitrogen, total phosphorus, nitrate nitrogen, total organic carbon, etc. Based on different effluent water qualities, one or more effluent basic water quality prediction models can be constructed.

[0050] Different machine learning algorithms are selected, and based on the input influent water quality feature set, basic effluent water quality prediction models are established for different biological treatment n-unit characterization combination processes, and cross-validation experiments are conducted to evaluate the performance of the prediction model to obtain the optimal basic effluent water quality prediction model. Based on whether the model prediction results meet the preset effluent water quality target, the basic effluent water quality prediction model that meets the effluent water quality standard is selected.

[0051] (3.2) Construct the effluent ecological risk prediction model of different biological treatment n-unit combination processes, including:

[0052] From the basic water quality and ecological risk characteristics of the biological treatment process after data preprocessing in step (1), determine the input influent water quality characteristic set and preset the ecological risk target.

[0053] Different machine learning algorithms were selected, and based on the input influent water quality feature set, ecological risk prediction models were established for different biological treatment n-unit characterization combination processes, and cross-validation experiments were conducted to evaluate the performance of the prediction models to obtain the optimal ecological risk prediction model. Based on whether the model prediction results meet the preset ecological risk targets, the effluent ecological risk prediction models that meet the ecological risk standards were selected.

[0054] Among them, when conducting cross-validation experiments to evaluate the prediction performance of the established effluent basic water quality prediction model and ecological risk prediction model, the following three evaluation indicators are used, and the calculation method of each indicator is as follows:

[0055]

[0056] Among them, y i represents the true value, Represents the predicted value, MAE represents the mean absolute error between the predicted value and the true value, and MSE represents the mean square error between the predicted value and the true value. The smaller the MAE and MSE values ​​are, the better the prediction performance of the model is. 2 Represents the overall performance of the prediction model. The larger its value, the better the prediction performance of the model. 2 The value range is 0-1; Represents the mean of the true value, and the calculation formula is:

[0057]

[0058] (4) Simulate and reconstruct different activated sludge wastewater biological treatment short-range unit combination processes, and conduct basic water quality compliance tests under different influent water quality scenarios. Specifically include:

[0059] (4.1) Simulate and reconstruct all possible activated sludge wastewater biological treatment short-range unit combination processes, and uniformly characterize the short-range unit combination into n units as the biological process information input variable of the prediction model. Among them, the short-range unit combination includes two-pass unit combination, three-pass unit combination, four-pass unit combination, etc.

[0060] (4.2) Under different influent water quality scenarios, the optimal effluent basic water quality prediction model of the biological treatment n-unit characterization combination process trained in step (3.1) is used to predict the effluent water quality of the biological treatment short-range unit combination process reconstructed in step (4.1), and determine whether the prediction results meet the preset effluent water quality target; wherein the input of the effluent basic water quality prediction model is the input characteristics determined in step (3.1) and the wastewater biological treatment short-range unit combination process simulated and reconstructed in step (4.1), and the output of the model is the effluent water quality prediction result, and the reconstructed biological treatment short-range unit combination process whose prediction results meet the effluent water quality target is screened out, that is, the reconstructed biological treatment short-range unit combination process with effluent water quality that meets the standard.

[0061] (5) Based on the effluent ecological risk prediction model, the ecological risk of the reconstructed biological treatment short-range unit combination process that meets the effluent water quality standards is evaluated, and the wastewater biological treatment process is determined based on the ecological risk control effect.

[0062] The effluent ecological risk prediction model of the biological treatment n-unit characterization combination process trained in step (3.2) is used to predict the ecological risk of the reconstructed biological treatment short-range unit combination process with effluent water quality meeting the standard in step (4.2); wherein the input of the ecological risk prediction model is the input characteristics determined in step (3.2) and the reconstructed biological treatment short-range unit combination process with effluent water quality meeting the standard selected in step (4.2), and the output of the model is the ecological risk prediction result; the reconstructed biological treatment short-range unit combination process whose prediction results meet the effluent ecological risk target is selected again, and the optimal wastewater biological treatment process is determined based on the ecological risk control effect.

[0063] In order to further illustrate the method of the present invention, the wastewater biological treatment process data of sewage treatment plants with different wastewater biological treatment processes across the country were collected. The specific steps of wastewater biological treatment process reconstruction are as follows:

[0064] (1) Collect basic water quality, ecological risks and specific activated sludge wastewater biological treatment process information data from 180 sewage treatment plants in Shanghai, Jiangsu, Guangdong, Henan, Yunnan and other provinces and cities, and perform data preprocessing.

[0065] Among them, the basic water quality characteristics of inlet and outlet water include ammonia nitrogen, total nitrogen, total phosphorus, nitrate nitrogen, total organic carbon, temperature, and pH value; the ecological risk data of inlet and outlet water include biological toxicity.

[0066] The basic water quality and ecological risk data of the sewage treatment plant are preprocessed. The data preprocessing methods include abnormal data processing, default data processing, outlier data processing, repeatability and consistency detection, data normalization and standardization, etc.

[0067] (2) Extract features of wastewater biological treatment processes and perform short-range multi-unit characterization.

[0068] Each link of the activated sludge biological treatment process of wastewater in a sewage treatment plant is unitized, and the biological pool unit is divided into three characteristic units: aerobic pool, anoxic pool and anaerobic pool according to dissolved oxygen, and feature extraction is performed; and a method other than the one-hot algorithm is used to perform short-range multi-unit characterization combination of the activated sludge biological treatment process.

[0069] Among them, the method of using the one-hot algorithm to characterize the multi-unit of the biological treatment process is:

[0070] First, the aerobic tank unit is characterized as [1,0,0], the anoxic tank unit is characterized as [0,1,0], the anaerobic tank unit is characterized as [0,0,1], and the blank unit is characterized as [0,0,0].

[0071] Secondly, the biological treatment process of the activated sludge method is uniformly characterized by "four units". When the number of biological pool units of the characterized process is less than 4, blank units are used to supplement them. The characteristic dimension of the "four units" characterization is 12. Among them, the "four units" characterization of the anaerobic-anoxic-aerobic process in the wastewater biological treatment process is [0,0,1,0,1,0,1,0,0,0,0,0]; the "four units" characterization of the anaerobic-aerobic process is [0,0,1,1,0,0,0,0,0,0,0,0].

[0072] (3) Construct a basic water quality prediction model and an effluent ecological risk prediction model for the biological treatment process and conduct compliance inspections.

[0073] The specific steps include:

[0074] (3.1) Construct a basic water quality prediction model for the effluent from the biological treatment process and conduct compliance testing, including:

[0075] Determine the input inlet water quality feature set and preset the outlet water quality target. The inlet water quality feature set is total phosphorus, total nitrogen, ammonia nitrogen, nitrate nitrogen, total organic carbon, temperature, and pH value; the outlet water quality target is set to 60% total nitrogen removal rate.

[0076] Support vector machine and random forest algorithm were selected, and based on the input influent water quality feature set, basic effluent water quality prediction models were established for different short-range four-unit characterization combination processes of biological treatment, and cross-validation experiments were conducted to evaluate the performance of the prediction model to obtain the optimal basic effluent water quality prediction model. According to whether the model prediction results reached the preset effluent total nitrogen removal rate of 60%, the basic effluent water quality prediction model with effluent water quality meeting the standard was selected.

[0077] (3.2) Construct an ecological risk prediction model for the effluent from the biological treatment process and conduct compliance inspections, including:

[0078] Determine the input water quality characteristic set and preset the ecological risk target. The input water quality characteristic set is total phosphorus, total nitrogen, ammonia nitrogen, nitrate nitrogen, total organic carbon, temperature, and pH value; the ecological risk target is set to a 60% reduction rate of biological toxicity.

[0079] Support vector machine and random forest algorithm were selected, and based on the input influent water quality feature set, effluent ecological risk prediction models were established for different short-range four-unit characterization combination processes of biological treatment, and cross-validation experiments were conducted to evaluate the performance of the prediction model to obtain the optimal effluent ecological risk prediction model. According to whether the model prediction results reached the preset biological toxicity reduction rate of 60%, the effluent ecological risk prediction model that met the effluent ecological risk standard was selected.

[0080] Among them, MAE, MSE and R were used to evaluate the prediction performance of the established effluent basic water quality prediction model and ecological risk prediction model using cross-validation experiments. 2 The three evaluation indicators are evaluated, and the results are shown in Table 1. As can be seen from Table 1, the support vector machine method has a smaller evaluation index value for the effluent basic water quality prediction model, and the random forest method has a smaller evaluation index value for the effluent ecological risk prediction model. Therefore, the support vector machine algorithm is selected to construct the optimal effluent basic water quality prediction model, and the random forest algorithm is selected to construct the optimal ecological risk prediction model.

[0081] Table 1 Evaluation index results of different prediction models

[0082]

[0083] (4) Simulate and reconstruct different activated sludge wastewater biological treatment short-range unit combination processes, and perform effluent water quality compliance inspection based on the optimal effluent basic water quality prediction model constructed in step (3); the specific steps are as follows:

[0084] (4.1) Simulate and reconstruct all possible activated sludge wastewater biological treatment second-pass, third-pass, and fourth-pass unit combination processes, and use the one-hot algorithm to uniformly characterize the above short-pass unit combination processes as "four units";

[0085] (4.2) Under different influent water quality scenarios, for example, when the total phosphorus is 2 mg / L, the total nitrogen is 15 mg / L, the ammonia nitrogen is 13.5 mg / L, the nitrate nitrogen is 1 mg / L, the total organic carbon is 30 mg / L, the temperature is 25°C, and the pH value is 6.5, the optimal effluent basic water quality prediction model of the "four-unit" characterization combination process trained in step (3.1) is used to predict the effluent water quality of different reconstructed wastewater biological treatment "four-unit" characterization combination processes. The prediction results are shown in Table 2 below, and it is judged whether the prediction results meet the preset effluent water quality target (total nitrogen removal rate of 60%), that is, whether the effluent total nitrogen is less than 6 mg / L, and the reconstructed biological treatment short-range unit combination process with the effluent water quality meeting the standard is selected.

[0086] (5) Predict the ecological risks of the reconstructed biological treatment short-range unit combination processes that meet the effluent quality standards, and determine the best wastewater biological treatment process from the reconstructed biological treatment processes that meet the standards based on the ecological risk control effect. The specific steps are as follows:

[0087] For the reconstructed biological treatment short-cut unit combination process with effluent quality meeting the standard in step (4.2), the effluent ecological risk model of the "four-unit" characterization combination process trained in step (3.2) was used to predict the ecological risk and compared with the preset ecological risk target (biological toxicity reduction rate of 60%), as shown in Table 2. The ecological risk control effects of different reconstructed biological treatment short-cut unit combination processes were evaluated, and the optimal wastewater biological treatment short-cut unit combination process was selected.

[0088] Table 2 Basic water quality and ecological risk prediction results of the effluent of the reconstructed wastewater biological treatment unit

[0089]

[0090]

[0091] It can be seen from Table 2 that under the current water quality scenario, the basic effluent water quality and ecological risk prediction results of the above-mentioned reconstructed wastewater biological treatment short-range unit combination process meet the preset goals. According to the ecological risk control effect, the wastewater biological treatment short-range unit combination process with the highest biological toxicity reduction rate is selected as the anaerobic-aerobic-anoxic-aerobic combination process, and its biological toxicity reduction rate reaches 67.3%.

Claims

1. A method for reconstructing wastewater biological treatment processes based on machine learning, It is characterized in that The following steps are involved: (1) Collect basic water quality, ecological risk and process information data of influent and effluent in the wastewater biological treatment process and perform data preprocessing; (2) Feature extraction and multi-unit characterization of wastewater biological treatment processes; (3) Construct basic effluent water quality prediction models and effluent ecological risk prediction models for different biological treatment multi-unit combination processes and conduct compliance tests; The basic water quality prediction model for effluent of different biological treatment multi-unit combination processes described in step (3) includes: Determine the input influent water quality feature set and preset the effluent water quality target; select different machine learning algorithms, and based on the input influent water quality feature set, establish effluent basic water quality prediction models for different biological treatment multi-unit combination processes, and conduct cross-validation experiments to evaluate the prediction model performance to obtain the optimal effluent basic water quality prediction model; screen out the effluent ecological risk prediction model that achieves the preset effluent water quality target; The effluent ecological risk prediction model constructed in step (3) for different biological treatment multi-unit combination processes includes: Determine the input influent water quality feature set and preset ecological risk targets; select different machine learning algorithms, and based on the input influent water quality feature set, establish effluent ecological risk prediction models for different biological treatment multi-unit combination processes, and conduct cross-validation experiments to evaluate the performance of the prediction model to obtain the optimal effluent ecological risk prediction model; screen out the effluent ecological risk prediction model that meets the preset ecological risk targets; (4) Simulate and reconstruct different short-range unit combination processes for wastewater biological treatment, and conduct effluent quality compliance tests under different influent water quality scenarios; (5) Evaluate the ecological risk of the reconstructed biological treatment short-range unit combination process to ensure that the effluent quality meets the standards and determine the optimal wastewater biological treatment process.

2. The method for reconstructing a wastewater biological treatment process based on machine learning according to claim 1, It is characterized in that In step (1), the basic water quality of the inlet and outlet water of the wastewater biological treatment process includes chemical oxygen demand, ammonia nitrogen, total nitrogen, total phosphorus, suspended solids, biochemical oxygen demand, animal and plant oils, petroleum, nitrate nitrogen, total organic carbon, dissolved oxygen, temperature, pH value, and fecal coliform count; the ecological risk data of the inlet and outlet water includes the content of micropollutants, biological toxicity, etc.; the process information of the wastewater biological treatment process includes the activated sludge wastewater biological treatment process information data.

3. The method for reconstructing a wastewater biological treatment process based on machine learning according to claim 1, It is characterized in that Step (2) includes: The characteristics of the biological pool link of the activated sludge wastewater biological treatment process are extracted, and the biological pool link is divided into three characteristic units: aerobic pool, anoxic pool and anaerobic pool according to dissolved oxygen; the activated sludge wastewater biological treatment process is characterized by multiple units according to the extracted characteristic units.

4. The method for reconstructing a wastewater biological treatment process based on machine learning according to claim 3, It is characterized in that The multi-unit characterization method includes a one-hot algorithm.

5. The method for reconstructing a wastewater biological treatment process based on machine learning according to claim 4, It is characterized in that The one-hot algorithm is used to characterize the bioprocess in multiple units including: First, the aerobic tank unit is characterized as [1,0,0], the anoxic tank unit is characterized as [0,1,0], the anaerobic tank unit is characterized as [0,0,1], and the blank unit is characterized as [0,0,0]; Secondly, the biological pool units were characterized and the activated sludge wastewater biological treatment process was uniformly characterized by n units. When the number of units for the characterized process was insufficient, blank units were used to supplement it.

6. The method for reconstructing a wastewater biological treatment process based on machine learning according to claim 1, It is characterized in that When conducting cross-validation experiments to evaluate the performance of the prediction model, the following three evaluation indicators are used, and the calculation method of each indicator is as follows: in, represents the true value, represents the predicted value, represents the mean of the sample, MAE represents the mean absolute error between the predicted value and the true value, MSE represents the mean square error between the predicted value and the true value, R 2 Represents the overall performance of the prediction model.

7. The method for reconstructing a wastewater biological treatment process based on machine learning according to claim 1, It is characterized in that Step (4) includes: (4.1) Simulate and reconstruct all possible short-range unit combination processes for wastewater biological treatment, and uniformly characterize the reconstructed short-range unit combinations by n units; (4.2) Under different influent water quality scenarios, the effluent water quality of each reconstructed biological treatment short-range unit combination process is predicted respectively, and based on whether the predicted results meet the preset effluent water quality target, the reconstructed biological treatment short-range unit combination process with effluent water quality that meets the target is selected.

8. A reconstruction system for wastewater biological treatment process based on machine learning, It is characterized in that include: Data collection and preprocessing module, used to collect basic water quality, ecological risk and process information data of influent and effluent in the wastewater biological treatment process and perform data preprocessing; Feature extraction and characterization module, used to extract features and perform multi-unit characterization of wastewater biological treatment processes; Prediction model building module, used to build effluent basic water quality prediction models and effluent ecological risk prediction models for different biological treatment multi-unit combination processes; The wastewater biological treatment process reconstruction and compliance inspection module is used to simulate and reconstruct all possible wastewater biological treatment short-range unit combination processes, and conduct effluent water quality compliance inspections under different influent water quality scenarios to screen out the reconstructed biological treatment short-range unit combination processes with effluent water quality that meets the standards; The wastewater biological treatment process determination module is used to evaluate the ecological risk of the reconstructed biological treatment short-range unit combination process that meets the effluent water quality standards based on the effluent ecological risk prediction model, and determine the optimal wastewater biological treatment process based on the ecological risk control effect.

Images