Constructed wetland coupling system effluent quality target control method based on sensor and machine learning

Abstract

The invention relates to the technical field of water treatment, in particular to a sensor and machine learning-based effluent quality target control method for an artificial wetland coupling system, which is applied to a control unit in an artificial wetland system, and the artificial wetland system comprises a plurality of small test systems positioned at the front end and a plurality of pilot test systems positioned at the rear end; the output ends of all the small-scale test systems are connected with the bottom of a front water distribution tank of the pilot-scale test system; and the water distribution tank is also connected with a river sewage source pipeline. According to the constructed wetland coupling system outlet water quality target control method based on sensor and machine learning, after inlet water quality data and outlet water quality data of a rear-end pilot plant test system are collected, an intelligent adjustment strategy is output through the data processing capacity of an SVM model obtained through training based on operation data of a small plant test system; the front-end small-scale test system and the rear-end pilot-scale test system are well coupled, the respective advantages are fully exerted, and the overall treatment effect is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of water treatment, and in particular to a method for controlling the effluent water quality target of an artificial wetland coupling system based on sensors and machine learning. Background Art

[0002] Traditional artificial wetland treatment systems often rely on empirical operation and management, making it difficult to achieve stable control of the effluent water quality. Especially when treating sewage with different concentrations, the removal efficiency and water quality compliance rate of the system fluctuate. Therefore, a method that can monitor water quality changes in real time and intelligently adjust treatment strategies is needed to improve the stability and compliance rate of the effluent water quality. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a method for controlling the effluent water quality target of an artificial wetland coupling system based on sensors and machine learning.

[0004] In a first aspect, an embodiment of the present invention provides a method for controlling the effluent water quality target of an artificial wetland coupling system based on sensors and machine learning, which is applied to a control unit in the artificial wetland system. The artificial wetland system includes a number of pilot-scale systems at the front end and a pilot-scale system at the back end; the output ends of all the pilot-scale systems are connected to the bottom of a water distribution tank in front of the pilot-scale system; the water distribution tank is connected to the river sewage source pipeline;

[0005] The method includes:

[0006] Obtain the influent water quality data and effluent water quality data of the pilot-scale system collected by the on-line monitoring sensor; wherein, the influent water quality data is the water quality data of the mixed sewage flowing from the water distribution tank to the pilot-scale system;

[0007] Input the influent water quality data into a pre-trained SVM model to output a predicted water quality result; wherein, the SVM model is trained based on the operation data of the pilot-scale system;

[0008] Input the predicted water quality result and the effluent water quality data into a pre-trained intelligent regulation model to output a regulation strategy;

[0009] Based on the regulation strategy, adjust the amount of river sewage flowing into the pilot-scale system, and / or, adjust the water inflow of each pilot-scale system in combination with the water quality data of the mixed liquid in the water distribution tank.

[0010] Combined with the first aspect, before the step of obtaining the influent water quality data and effluent water quality data of the pilot-scale system collected by the on-line monitoring sensor, it further includes:

[0011] Obtain the initial data set of the pilot-scale system;

[0012] Preprocess the initial data set to obtain a training data set;

[0013] Divide the training data set into a training set and a test set;

[0014] Train the initial SVM model based on a preset kernel function, regularization parameter, and γ parameter, and determine the hyperparameters based on cross-validation;

[0015] Evaluate the SVM model using the test set and optimize the initial SVM model based on the evaluation results until a target SVM model that meets the preset requirements is obtained, and output the predicted water quality results.

[0016] Combined with the first aspect, after the step of dividing the training data set into a training set and a test set, it further includes:

[0017] Obtain the initial effluent water quality data set corresponding to the initial data set to obtain a data set;

[0018] Create and train a random forest regressor;

[0019] Evaluate the random forest regressor using the test set, adjust the hyperparameters to optimize the random forest regressor until an intelligent regulation model that meets the preset requirements is obtained, and output the regulation strategy.

[0020] Combined with the first aspect, a valve is provided on the pipeline between the water distribution tank and the river sewage source, and the valve is communicatively connected to the control unit;

[0021] The step of adjusting the amount of river sewage flowing into the pilot system based on the regulation strategy includes:

[0022] Obtain the first target opening angle of the valve;

[0023] Control the valve to open at the first target opening angle.

[0024] Combined with the first aspect, a peristaltic pump is provided at the inlet end of each small-scale test system, and each peristaltic pump is communicatively connected to the control unit;

[0025] Based on the regulation strategy, adjust the amount of river sewage flowing into the pilot system, and / or, adjust the water inflow of each small-scale test system in combination with the water quality data of the mixed liquid in the water distribution tank;

[0026] Adjust the amount of river sewage flowing into the pilot system to a preset threshold;

[0027] Obtain the water quality data of the mixed liquid in the water distribution tank;

[0028] Combine the water quality data to obtain the second target opening angle of the peristaltic pump of each small-scale test system under the regulation strategy;

[0029] For each small-scale test system, control the peristaltic pump to open at the second target opening angle.

[0030] In combination with the first aspect, after the step of obtaining the influent water quality data and the effluent water quality data of the pilot system collected by the on-line monitoring sensor, the following steps are further included:

[0031] Clean, fill in missing values and detect outliers for the influent water quality data and the effluent water quality data.

[0032] In combination with the first aspect, the influent water quality data includes the water flow rate, water quality indicators and water flow temperature of the water introduced into the pilot system; among them, the water quality indicators at least include: COD index, ammonia nitrogen content, total nitrogen content, dissolved oxygen content, conductivity, pH.

[0033] In the second aspect, the present application provides an effluent water quality target control device based on sensors and machine learning for a control unit in a constructed wetland coupling system. The constructed wetland system includes a number of small-scale test systems at the front end and a pilot system at the back end; the output ends of all the small-scale test systems are connected to the bottom of a water distribution tank in front of the pilot system; the water distribution tank is also connected to a river sewage source pipeline;

[0034] The device includes:

[0035] An acquisition module, configured to acquire the influent water quality data and the effluent water quality data of the pilot system collected by the on-line monitoring sensor; among them, the influent water quality data is the water quality data of the mixed sewage flowing from the water distribution tank to the pilot system;

[0036] A water quality prediction module, configured to input the influent water quality data into a pre-trained SVM model and output a predicted water quality result; among them, the SVM model is trained based on the operation data of the small-scale test system;

[0037] A strategy output module, configured to input the predicted water quality result and the effluent water quality data into a pre-trained intelligent regulation model at the same time and output a regulation strategy;

[0038] An adjustment module, configured to adjust the amount of river sewage flowing into the pilot system based on the regulation strategy, and / or adjust the water inflow of each small-scale test system in combination with the water quality data of the mixed liquid in the water distribution tank.

[0039] In the third aspect, the present application provides an electronic device, which includes a memory and a processor. The memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the above method.

[0040] In the fourth aspect, the present application provides a readable storage medium, in which computer program instructions are stored. When the computer program instructions are read and run by a processor, the above method is executed.

[0041] The embodiments of the present invention bring the following beneficial effects: The method for controlling the effluent water quality target of the constructed wetland coupling system based on sensors and machine learning is applied to the control unit in the constructed wetland system. The constructed wetland system includes several pilot-scale systems at the front end and a pilot-scale system at the back end; the output ends of all the pilot-scale systems are connected to the bottom of the water distribution tank in front of the pilot-scale system; the water distribution tank is also connected to the river sewage source pipeline; the method includes: obtaining the influent water quality data and the effluent water quality data of the pilot-scale system collected by the on-line monitoring sensor; wherein, the influent water quality data is the water quality data of the mixed sewage flowing from the water distribution tank to the pilot-scale system; inputting the influent water quality data into the pre-trained SVM model to output the predicted water quality result; wherein, the SVM model is trained based on the operation data of the pilot-scale system; inputting the predicted water quality result and the effluent water quality data into the pre-trained intelligent regulation model at the same time to output the regulation strategy; based on the regulation strategy, adjusting the amount of river sewage flowing into the pilot-scale system, and / or, adjusting the water inflow of each pilot-scale system in combination with the water quality data of the mixed liquid in the water distribution tank.

[0042] The method for controlling the effluent water quality target of the constructed wetland coupling system based on sensors and machine learning provided by the present application, after collecting the influent water quality data and the effluent water quality data of the pilot-scale system at the back end, outputs an intelligent adjustment strategy through the data processing ability of the SVM model trained based on the operation data of the pilot-scale system, so as to couple the pilot-scale system at the front end and the pilot-scale system at the back end well, give full play to their respective advantages, and improve the overall treatment effect.

[0043] Other features and advantages of the present invention will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention are achieved and obtained by the structures specifically pointed out in the specification, the claims, and the drawings.

[0044] In order to make the above-mentioned objectives, features, and advantages of the present invention more obvious and understandable, the following specific preferred embodiments are given, and in conjunction with the accompanying drawings, the detailed description is as follows. Description of the Drawings

[0045] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those skilled in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0046] Figure 1 It is a flowchart of the method for controlling the effluent water quality target of the constructed wetland coupling system based on sensors and machine learning provided by the embodiments of the present invention;

[0047] Figure 2 Schematic diagram of the device for controlling the effluent water quality target of the constructed wetland coupling system based on sensors and machine learning provided by the embodiment of the present invention;

[0048] Figure 3 Schematic diagram of the structure of the electronic device provided by the embodiment of the present invention.

[0049] Reference numerals:

[0050] 10 - Acquisition module, 20 - Water quality prediction module, 30 - Strategy output module, 40 - Adjustment module;

[0051] 130 - Processor, 131 - Memory, 132 - Bus, 133 - Communication interface. Detailed implementation manners

[0052] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0053] To facilitate the understanding of this embodiment, the technical terms designed in this application will be briefly introduced below.

[0054] Support Vector Machine (SVM) is a supervised learning model widely used in classification and regression tasks. It finds a hyperplane to maximize the margin between different classes, thereby achieving optimal classification of data. For non - linear problems, SVM can map the data to a high - dimensional space through a kernel function, making the data that is inseparable in the low - dimensional space become separable.

[0055] Laboratory - Scale Testing refers to small - scale tests conducted under laboratory conditions. Usually, simulated or actual sewage samples are used for testing. It has the characteristics of small scale, high flexibility, and low cost, and is mainly used for pilot research and technology screening.

[0056] The pilot-scale testing refers to the intermediate test stage between laboratory-scale testing and industrial production. It is usually carried out in a specially constructed small factory or on-site, with a much larger throughput than laboratory-scale testing but smaller than actual production scale. The pilot-scale testing system can be used to simulate real environments, including factors such as water quality fluctuations and changes in influent flow rate, and can obtain more accurate data on operating parameters, energy consumption analysis, and economic evaluation, providing reliable technical support for large-scale applications.

[0057] After introducing the technical terms involved in this application, next, a brief introduction to the application scenario and design concept of the embodiments of this application will be given.

[0058] Traditional systems usually only rely on regular manual sampling and laboratory analysis, and cannot achieve real-time monitoring of water quality changes. This results in the inability to detect and respond to water quality anomalies in a timely manner, leading to large fluctuations in the effluent water quality and making it difficult to ensure stable compliance. In addition, traditional management methods are mainly based on fixed operating parameters (such as water level, water flow velocity, etc.), and it is difficult to make dynamic adjustments according to actual water quality changes, affecting the water treatment effect.

[0059] Based on this, the embodiments of this application provide a method for controlling the effluent water quality target of an artificial wetland coupling system based on sensors and machine learning.

[0060] Embodiment 1

[0061] This application provides a method for controlling the effluent water quality target of an artificial wetland coupling system based on sensors and machine learning, which is applied to the control unit in the artificial wetland system. The artificial wetland system includes several pilot-scale testing systems at the front end and a pilot-scale testing system at the back end; the output ends of all the pilot-scale testing systems are connected to the bottom of the distribution tank in front of the pilot-scale testing system; the distribution tank is also connected to the river sewage source pipeline.

[0062] Combined with Figure 1 As shown, this method includes:

[0063] S110, obtaining the influent water quality data and effluent water quality data of the pilot-scale testing system collected by the on-line monitoring sensor; among them, the influent water quality data is the water quality data of the mixed sewage flowing from the distribution tank to the pilot-scale testing system.

[0064] S120, inputting the influent water quality data into the pre-trained SVM model, and outputting the predicted water quality result; among them, the SVM model is trained based on the operation data of the pilot-scale testing system.

[0065] S130, inputting the predicted water quality result and the effluent water quality data into the pre-trained intelligent regulation model at the same time, and outputting the regulation strategy.

[0066] S140. Adjust the amount of river sewage flowing into the pilot-scale system based on the regulation strategy, and / or adjust the water inflow of each bench-scale system in combination with the water quality data of the mixed liquor in the distribution tank.

[0067] In the method for controlling the effluent water quality target of the constructed wetland coupling system based on sensors and machine learning provided in the embodiments of the present application, after collecting the influent water quality data and effluent water quality data of the pilot-scale system at the back end, first predict the effluent water quality based on the collected influent water quality data through the data processing ability of the model, and then output an intelligent adjustment strategy to achieve good coupling of the front-end bench-scale system and the back-end pilot-scale system, give full play to their respective advantages, and improve the overall treatment effect.

[0068] In this embodiment, high-precision on-line monitoring sensors are installed at both the influent and effluent ports of the pilot-scale system. These on-line monitoring sensors can monitor and record the following parameters in real time: the sewage flow rate entering the pilot-scale system, water quality indicators: including key water quality parameters such as chemical oxygen demand (COD), ammonia nitrogen, total nitrogen (TN), dissolved oxygen (DO), conductivity, and pH value, and temperature: specifically referring to the change in water temperature during the monitoring and treatment process. In particular, the on-line monitoring sensor at the influent port of the pilot-scale system is specifically used to monitor the sewage that flows from the bench-scale system into the front-end distribution tank and is fully mixed. This ensures that the water volume and water quality indicators entering the pilot-scale system are accurately recorded, providing reliable data support for the subsequent treatment process.

[0069] Among them, there are several bench-scale systems. The influent water of several bench-scale systems can be different types of sewage and can also have different concentrations. In this embodiment, there are 6 bench-scale systems, and the influent water of the 6 bench-scale systems is industrial wastewater, aquaculture wastewater, domestic sewage, medical sewage, underground sewage, and rainwater runoff respectively, and the water inflow and concentration gradient of the 6 bench-scale systems are set to increase along the first direction.

[0070] It can be understood that the effluent water of each bench-scale system flows into the distribution tank and is mixed, and after being fully mixed with the makeup water (i.e., the river sewage introduced), it is introduced into the pilot-scale system. By adjusting the makeup water volume (i.e., the amount of river sewage introduced), the concentration of the mixed liquor after the effluent water of each bench-scale system is mixed can be adjusted. Specifically, when the introduced makeup water volume increases, the mixed liquor can be diluted to reduce the concentration of various pollutants introduced into the pilot-scale system; on the contrary, when the makeup water volume is reduced, the proportion of the mixed liquor introduced into the pilot-scale system is larger.

[0071] Combined with the first aspect, before step S110, it further includes:

[0072] S010. Obtain the initial data set of the bench-scale system.

[0073] S020. Preprocess the initial data set to obtain the training data set.

[0074] S030, divide the training data set into a training set and a test set.

[0075] S040, train the initial SVM model based on a preset kernel function, regularization parameter, and γ parameter, and determine the hyperparameters based on cross-validation.

[0076] S050, evaluate the SVM model using the test set and optimize the initial SVM model based on the evaluation results until a target SVM model that meets the preset requirements is obtained, and output the predicted water quality results.

[0077] It can be understood that before actual application, the SVN model and the intelligent regulation model should be trained multiple times until the application requirements are met. The purpose of steps S010 - S050 is to train the SVN model.

[0078] Specifically, in step S010, first obtain the initial data set. At this time, the integrity and accuracy of the data should be ensured.

[0079] After that, in step S020, preprocess the initial data set, including operations such as missing value processing, outlier detection, feature selection, and standardization, and finally obtain a training data set that can be used for training.

[0080] After that, in step S030, perform data division. Specifically, randomly divide the training data set into a training set and a test set, usually according to a certain ratio (such as 80% training set, 20% test set) for division to ensure the reliability of model evaluation.

[0081] After that, in step S040, use a preset kernel function (such as RBF, linear, etc.), regularization parameter (C), and γ parameter to train the initial support vector machine (SVM) model. Among them, determine the optimal hyperparameter combination through a cross-validation method (such as k-fold cross-validation) to improve the generalization ability of the model.

[0082] Among them, the support vector machine (SVM) maps the input data to a high-dimensional space through a kernel function, thereby realizing the classification of linearly inseparable problems. Commonly used kernel functions include: linear kernel function (Linear Kernel), polynomial kernel function (Polynomial Kernel), radial basis function kernel (RBF Kernel or Gaussian Kernel), and sigmoid kernel function. In this embodiment, since there are many factors affecting the sewage treatment efficiency and the complexity is relatively high during the sewage treatment process, the radial basis function kernel is selected, which can handle complex data distributions.

[0083] An appropriate regularization parameter (C), adjusting the γ parameter can reduce the error and avoid overfitting. In this embodiment, cross-validation is used to find the optimal hyperparameter combination (regularization parameter C, γ parameter). Then, in step S050, the SVM model containing the optimal hyperparameter combination is evaluated and optimized using the test set until a target SVM model for predicting water quality is obtained.

[0084] In this embodiment, taking advantage of the characteristic that the SVM model can handle complex and high-dimensional non-linear tasks, it is applied to the water treatment field to perform real-time analysis on the data collected by the sensor and predict the water quality of the effluent.

[0085] Combined with the first aspect, after step S030, it further includes:

[0086] S041, obtaining an initial effluent water quality dataset corresponding to the initial dataset to obtain a dataset.

[0087] S051, creating and training a random forest regressor.

[0088] S061, using the test set to evaluate the random forest regressor, adjusting the hyperparameters to optimize the random forest regressor until an intelligent regulation model that meets the preset requirements is obtained, and outputting a regulation strategy.

[0089] Among them, step S041 is used to collect the effluent water quality data corresponding to the initial data in the initial dataset in step S010 to obtain a dataset.

[0090] In step S051, after having a high-quality dataset, a random forest regressor is selected for training to let the model learn how to predict the effluent water quality index based on the given features. The random forest regressor integrates multiple decision trees, reducing the overfitting problem that may occur in a single decision tree, and can capture the complex non-linear relationships between input features, can directly handle categorical variables, and is insensitive to the scale of the input data.

[0091] After the model training in step S051 is completed, it is also necessary to evaluate and optimize it to ensure that its performance meets the expected standard.

[0092] Specifically, verify the performance of the model on an independent test set, calculate key performance indicators such as mean squared error (MSE), mean absolute error (MAE), etc. Conduct hyperparameter tuning, model validation and cross-validation to further improve the model performance. Among them, hyperparameter tuning can use methods such as grid search, random search or Bayesian optimization to find the optimal hyperparameter combination.

[0093] When the model is fully trained and optimized, corresponding regulation strategies can be formulated according to actual needs. For example, based on the predicted future water quality change trend, specific management and control measures can be proposed to ensure that the water quality meets specific standards or goals.

[0094] Combined with the first aspect, a valve is provided on the pipeline between the water distribution tank and the river sewage source, and the valve is communicatively connected to the control unit.

[0095] The step of adjusting the amount of river sewage leading to the pilot system in step S140 includes:

[0096] S141, obtaining the first target opening angle of the valve in the adjustment strategy;

[0097] S142, controlling the valve to open at the first target opening angle.

[0098] It can be understood that only the adjustment strategy for the amount of river sewage leading to the pilot system is given in the regulation strategy. At this time, it means that the expected sewage purification effect can be achieved only by increasing or decreasing the amount of river sewage introduced. At this time, the valve is controlled to open at the first target opening angle to adjust the makeup water volume in the water distribution tank. Since the water quality of the river sewage is relatively stable, increasing the amount of river sewage replenished can dilute the effluent mixed liquid of each bench-scale system in the mixture for sewage treatment with the sewage purification capacity of the pilot system; reducing the amount of river sewage replenished can increase the concentration of the effluent mixed liquid of each bench-scale system to make full use of the sewage purification capacity of the pilot system.

[0099] Combined with the first aspect, a peristaltic pump is provided at the water inlet end of each bench-scale system, and each peristaltic pump is communicatively connected to the control unit. S140 includes:

[0100] S143, adjusting the amount of river sewage leading to the pilot system to be increased to a preset threshold;

[0101] S144, obtaining the water quality data of the mixed liquid in the water distribution tank;

[0102] S145, obtaining the second target opening angle of the peristaltic pump of each bench-scale system under the regulation strategy in combination with the water quality data;

[0103] S146, for each bench-scale system, controlling the peristaltic pump to open at the second target opening angle.

[0104] As another implementable approach, in the case where the regulation strategy includes both the adjustment of the amount of river sewage flowing into the pilot-scale system and the adjustment of the water inflow of the bench-scale systems, first adjust the amount of river sewage, and then adjust the opening angle of the peristaltic pump of each bench-scale system. That is, when the influent water quality of the pilot-scale system significantly exceeds the standard, at this time, first adjust the makeup water volume to observe whether the influent water quality flowing into the pilot-scale system can be adjusted by adjusting the makeup water volume. If the amount of river sewage flowing in has reached the preset threshold but still does not meet the requirements, then adjust the water inflow of the bench-scale systems. Since the sewage sources of the influent water of each bench-scale system are stable and the water quality indicators are known, the effluent mixed liquid that meets the requirements can be obtained by adjusting the water inflow of each bench-scale system, and then it is mixed with the incoming makeup water (i.e., river sewage) and sent to the pilot-scale system.

[0105] It can be understood that, as another implementable approach, in the regulation strategy, only the water quality data of the mixed liquid in the mixing tank is combined to adjust the water inflow of each bench-scale system, so as to adjust the concentration and content of the target indicators of the mixed liquid in the mixing tank by adjusting the water inflow of one of the bench-scale systems.

[0106] When receiving the instruction to adjust the influent pollutant concentration of the bench-scale system in the regulation strategy, the instruction carries the target peristaltic pump information corresponding to the specific bench-scale system to be adjusted, as well as the target opening angle information of the target peristaltic pump. Send the instruction to the target peristaltic pump to adjust the opening angle of the target peristaltic pump, so as to adjust the pollutant concentration flowing into the bench-scale system corresponding to the target peristaltic pump.

[0107] For example, if there are 8 bench-scale systems at the front end, when the output strategy includes an instruction to downward adjust the influent pollutant concentration of the 5th bench-scale system, first determine the target peristaltic pump information corresponding to the 5th bench-scale system. The target peristaltic pump information may refer to the unique identifier of the target peristaltic pump, such as the code, pre-configured name, etc. For example, the peristaltic pump code is 005. At the same time, it also includes the target opening angle of the target peristaltic pump. For example, the current opening angle is 90 degrees. At this time, in order to reduce the pollutant concentration, the opening angle of the peristaltic pump should be reduced to reduce the incoming pollutants. The example of the target opening angle at this time is 60°. Then the instruction is: peristaltic pump code - 005, opening angle 60°.

[0108] It can be understood that the regulation strategy may also include the input sewage volume, input sewage concentration, sludge concentration, sludge return ratio, aeration time, etc. of the pilot-scale system, which will not be elaborated here.

[0109] Combined with the first aspect, the pollutant concentration increases along the first direction when flowing into multiple parallel bench-scale systems.

[0110] In this embodiment, the size of the pilot system is as large as 400mm×400mm×550mm. There are a total of 8 sets of bench-scale systems, which are connected in parallel. Their effluents enter the pilot-scale system through peristaltic pumps respectively; the approximate size of the pilot-scale system is 2000mm×1200mm×1500mm. Among them, by setting an increasing gradient of the influent pollutant concentration in the first direction (for example, the total nitrogen content (TN) increases from 1mg / L in multiple gradients to 30mg / L), the change of pollutant concentration in the real environment can be more realistically simulated. In natural water bodies, the pollutant concentration is not constant, but fluctuates with time and location. In this way, the test results are more representative and predictable, and it helps to find the optimal treatment process parameters.

[0111] Among them, the distribution tank is also connected to the river sewage source. The pollutant concentration in the river sewage is relatively low, and the total nitrogen content is about 1mg / L. Among them, the influent flow rate of the pilot-scale system can be adjusted. Usually, the hydraulic load is not restricted by the design specifications, but it is not higher than 8m / d, otherwise it will affect the removal rate of TN. The larger the influent water volume, the greater the hydraulic load, but the worse the water treatment effect (that is, the removal rate decreases).

[0112] Combined with the first aspect, after step S110, it further includes:

[0113] S111, cleaning, filling missing values and detecting outliers for the influent water quality data and the effluent water quality data.

[0114] In this embodiment, through the above-mentioned pretreatment method, the collected data is preprocessed to ensure the accuracy and reliability of the data.

[0115] Combined with the first aspect, the influent water quality data includes the water flow rate, water quality indicators and water flow temperature of the water introduced into the pilot-scale system; among them, the water quality indicators at least include: COD index, ammonia nitrogen content, total nitrogen content, dissolved oxygen content, conductivity, pH.

[0116] In the second aspect, the present application provides an effluent water quality target control device based on sensors and machine learning for an artificial wetland coupling system, which is applied to the control unit in the artificial wetland system. The artificial wetland system includes a number of bench-scale systems at the front end and a pilot-scale system at the back end; the output ends of all the bench-scale systems are connected to the bottom of the distribution tank in front of the pilot-scale system; the distribution tank is connected to the river sewage source through pipelines.

[0117] Combined with Figure 2 As shown, the device includes: an acquisition module 10, a water quality prediction module 20, a strategy output module 30 and an adjustment module 40.

[0118] The acquisition module 10 is used to acquire the influent water quality data and the effluent water quality data of the pilot-scale system collected by the on-line monitoring sensors; wherein, the influent water quality data is the water quality data of the mixed sewage flowing from the water distribution tank to the pilot-scale system.

[0119] The water quality prediction module 20 is used to input the influent water quality data into a pre-trained SVM model and output a predicted water quality result; wherein, the SVM model is trained based on the operation data of the bench-scale system.

[0120] The strategy output module 30 is used to input the predicted water quality result and the effluent water quality data into a pre-trained intelligent regulation model at the same time and output a regulation strategy.

[0121] The adjustment module 40 is used to adjust the amount of river sewage flowing into the pilot-scale system based on the regulation strategy, and / or adjust the water inflow of each bench-scale system in combination with the water quality data of the mixed liquid in the water distribution tank.

[0122] In a third aspect, an embodiment of the present application provides an electronic device, as combined with Figure 3 shown, the electronic device includes a memory 131 and a processor 130. The memory 131 is used to store a computer program, and the processor 130 runs the computer program to enable the electronic device to execute the above method.

[0123] Further, as combined with Figure 3 shown, the electronic device further includes a bus 132 and a communication interface 133. The processor 130, the communication interface 133 and the memory 131 are connected through the bus 132.

[0124] Among them, the memory 131 may include a high-speed random access memory (RAM, Random Access Memory), and may also include a non-volatile memory, such as at least one disk memory. Through at least one communication interface 133 (which can be wired or wireless), a communication connection is realized between the system network element and at least one other network element, and the Internet, wide area network, local area network, metropolitan area network, etc. can be used. The bus 132 can be an ISA bus, a PCI bus or an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of representation, Figure 3 only a bidirectional arrow is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0125] The processor 130 may be an integrated circuit chip with the ability to process signals. In the implementation process, each step of the above method may be completed by the integrated logic circuit of the hardware in the processor 130 or the instructions in the form of software. The above-mentioned processor 130 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present invention may be directly embodied as being executed and completed by a hardware decoding processor, or executed and completed by a combination of hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory 131, and the processor 130 reads the information in the memory 131 and combines its hardware to complete the steps of the method in the foregoing embodiments.

[0126] In a fourth aspect, an embodiment of the present application provides a readable storage medium, in which computer program instructions are stored. When the computer program instructions are read and run by a processor, the above-mentioned method is executed.

[0127] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems and devices described above may refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again.

[0128] In addition, in the description of the embodiments of the present invention, unless otherwise clearly specified and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be a direct connection or an indirect connection through an intermediate medium, and it may be the communication inside two elements. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0129] When the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0130] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present invention. In addition, the terms "first", "second", "third" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0131] Finally, it should be noted that the above embodiments are only specific embodiments of the present invention, used to illustrate the technical solutions of the present invention, rather than limiting them. The protection scope of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art within the technical scope disclosed by the present invention can still modify the technical solutions described in the foregoing embodiments, or can easily think of changes, or perform equivalent replacements on some of the technical features; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for controlling the effluent quality of an artificial wetland coupling system based on sensors and machine learning, characterized in that: A control unit applied to an artificial wetland system, wherein the artificial wetland system comprises a plurality of small test systems at the front end and a pilot system at the rear end; the output ends of all the small test systems are connected to the bottom of a water distribution tank in front of the pilot system; the water distribution tank is also connected to a sewage source pipeline in a river; The method comprises: Acquire the inlet water quality data and outlet water quality data of the pilot system collected by the online monitoring sensor; wherein the inlet water quality data is the water quality data of the mixed sewage from the water distribution tank to the pilot system; Input the influent water quality data into a pre-trained SVM model, and output a predicted water quality result; wherein the SVM model is trained based on the operating data of the pilot system; The predicted water quality result and the effluent water quality data are simultaneously input into a pre-trained intelligent control model, and a control strategy is output; Based on the control strategy, the amount of river sewage leading to the pilot system is adjusted, and / or the water inlet of each of the pilot systems is adjusted in combination with the water quality data of the mixed liquid in the water distribution tank.

2. The method according to claim 1, characterized in that Before the step of obtaining the inlet water quality data and outlet water quality data of the pilot system collected by the online monitoring sensor, the step further includes: Obtaining an initial data set for the pilot system; Preprocessing the initial data set to obtain a training data set; Dividing the training data set into a training set and a test set; The initial SVM model is trained based on the preset kernel function, regularization parameter and γ parameter, and the hyperparameters are determined based on cross-validation; The SVM model is evaluated using a test set and the initial SVM model is optimized based on the evaluation results until a target SVM model that meets preset requirements is obtained, and a predicted water quality result is output.

3. The method according to claim 2, characterized in that After the step of dividing the training data set into a training set and a test set, the method further includes: Acquire an initial effluent water quality data set corresponding to the initial data set to obtain a data set; Create and train a random forest regressor; The random forest regressor is evaluated using the test set, and hyperparameters are adjusted to optimize the random forest regressor until the intelligent control model that meets the preset requirements is obtained, and the control strategy is output.

4. The method according to claim 1, characterized in that: A valve is provided on the pipeline between the water distribution tank and the river sewage source, and the valve is communicatively connected with the control unit; Based on the control strategy, the step of regulating the amount of river sewage leading to the pilot system includes: Obtaining a first target opening and closing angle of the valve in the control strategy; The valve is controlled to open to the first target opening and closing angle.

5. The method according to claim 1, characterized in that A peristaltic pump is provided at the water inlet end of each of the small test systems, and each of the peristaltic pumps is communicatively connected with the control unit; The steps of adjusting the amount of river sewage leading to the pilot system based on the control strategy, and / or adjusting the water inflow of each of the pilot systems in combination with the water quality data of the mixed liquid in the water distribution tank include: Regulating the amount of river sewage leading to the pilot system to a preset threshold; Acquiring water quality data of the mixed liquid in the water distribution tank; Acquire the second target opening and closing angle of the peristaltic pump of each of the small test systems under the control strategy in combination with the water quality data; For each small test system, the peristaltic pump is controlled to open the second target opening and closing angle.

6. The method according to claim 1, characterized in that After the step of obtaining the inlet water quality data and outlet water quality data of the pilot system collected by the online monitoring sensor, the method further includes: The inlet water quality data and the outlet water quality data are cleaned, missing values ​​are filled and outlier detection is performed.

7. The method according to claim 1, characterized in that The inlet water quality data includes the water flow rate, water quality indicators and water flow temperature entering the pilot system; wherein the water quality indicators include at least: COD index, ammonia nitrogen content, total nitrogen content, dissolved oxygen content, conductivity, and pH.

8. A sensor and machine learning-based effluent water quality target control device for an artificial wetland coupling system, characterized in that: A control unit applied to an artificial wetland system, wherein the artificial wetland system comprises a plurality of small test systems at the front end and a pilot system at the rear end; the output ends of all the small test systems are connected to the bottom of a water distribution tank in front of the pilot system; the water distribution tank is also connected to a sewage source pipeline in a river; The device comprises: An acquisition module, used to acquire the inlet water quality data and outlet water quality data of the pilot system collected by the online monitoring sensor; wherein the inlet water quality data is the water quality data of the mixed sewage from the water distribution tank to the pilot system; A water quality prediction module, used to input the inlet water quality data into a pre-trained SVM model and output a predicted water quality result; wherein the SVM model is trained based on the operating data of the pilot system; A strategy output module, used to input the predicted water quality result and the effluent water quality data into a pre-trained intelligent control model at the same time, and output a control strategy; The regulating module is used to regulate the amount of river sewage leading to the pilot system based on the regulation strategy, and / or to regulate the water inlet of each of the pilot systems in combination with the water quality data of the mixed liquid in the water distribution tank.

9. An electronic device, characterized in that: The electronic device comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform the method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores computer program instructions, and when the computer program instructions are read and executed by a processor, the method according to any one of claims 1 to 7 is executed.

Images