Sewage treatment process closed-loop control method based on machine learning

Abstract

The invention discloses a sewage treatment process closed-loop control method based on machine learning, and relates to the technical field of intelligent control in sewage treatment. The method comprises the following steps: collecting water inlet parameters, water outlet parameters and corresponding water outlet COD (Chemical Oxygen Demand) values of the sequencing batch sewage treatment equipment in a continuous operation cycle, and normalizing the water inlet parameters, the water outlet parameters and the corresponding water outlet COD values; constructing an effluent quality COD prediction model based on machine learning to obtain an effluent quality COD prediction value; and setting an effluent quality COD (Chemical Oxygen Demand) threshold value, establishing a dynamic feedback mechanism, and performing closed-loop control on the sewage treatment process by utilizing the dynamic feedback mechanism based on the effluent quality COD predicted value and the effluent quality COD threshold value. The intelligent sewage treatment closed-loop control method is constructed through deep coupling of machine learning prediction and process control, the equipment operation efficiency is remarkably improved on the premise of ensuring the effluent quality, and the method has outstanding industrial application value.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent control in sewage treatment, and particularly relates to a closed-loop control method for sewage treatment processes based on machine learning. Background Art

[0002] With the continuous increase in the number and wide distribution of sequencing batch sewage treatment equipment, the supervision levels of the effluent water quality vary unevenly, posing challenges to environmental protection work. And the sewage treatment control methods for sequencing batch sewage treatment equipment often rely on experience and chemical detection methods, and it is difficult to effectively meet the requirements of modern sewage treatment. Therefore, there is an urgent need in this field for an efficient and low-cost control method to improve the operating efficiency of sequencing batch sewage treatment equipment. Summary of the Invention

[0003] In view of the above deficiencies in the prior art, the present invention provides a closed-loop control method for sewage treatment processes based on machine learning.

[0004] In order to achieve the above-mentioned invention objective, the technical solution adopted by the present invention is as follows: A closed-loop control method for sewage treatment processes based on machine learning, comprising the following steps: S1. Collect the influent parameters, effluent parameters, and corresponding effluent COD values (Chemical Oxygen Demand) of the sequencing batch sewage treatment equipment during consecutive operating cycles, and perform normalization processing on the influent parameters and effluent parameters of the sequencing batch sewage treatment equipment during consecutive operating cycles; S2. Construct a COD prediction model for the effluent water quality based on machine learning, and use the influent parameters, effluent parameters, and corresponding effluent COD values of the sequencing batch sewage treatment equipment during consecutive operating cycles after normalization processing to train the COD prediction model for the effluent water quality, and deploy the trained and evaluated COD prediction model for the effluent water quality to the sequencing batch sewage treatment equipment to obtain the predicted value of the effluent water quality COD; S3. Set a threshold for the effluent water quality COD, and establish a dynamic feedback mechanism. Based on the predicted value of the effluent water quality COD and the threshold for the effluent water quality COD, and use the dynamic feedback mechanism to perform closed-loop control on the sewage treatment process; The dynamic feedback mechanism is specifically as follows: Determine whether the predicted value of the effluent water quality COD is less than the threshold for the effluent water quality COD; if so, directly perform the sewage treatment compliance operation; otherwise, continue to perform the biochemical reaction, start the delay monitoring to obtain the delayed effluent parameters, and perform the delayed prediction decision based on the delayed effluent parameters until the delayed predicted value of the effluent water quality COD is less than the threshold for the effluent water quality COD to perform the sewage treatment compliance operation.

[0005] Further, in step S1, the influent parameters of the sequencing batch sewage treatment equipment during the continuous operation cycle include the oxidation-reduction potential, conductivity, and water temperature of the influent water quality; the effluent parameters of the sequencing batch sewage treatment equipment during the continuous operation cycle include the oxidation-reduction potential, conductivity, and water temperature of the effluent water quality.

[0006] Further, in step S2, a prediction model for the COD of the effluent water quality is constructed based on machine learning, specifically: a prediction model for the COD of the effluent water quality is constructed based on the BP neural network and support vector machine regression.

[0007] Further, the prediction model for the COD of the effluent water quality includes an input layer, a BP network hidden layer, a BP output layer, and a support vector machine regression layer connected in sequence; the input layer is used to input the influent parameters and effluent parameters of the sequencing batch sewage treatment equipment during the continuous operation cycle after normalization processing; the BP network hidden layer is used to extract the non-linear features in the influent parameters and effluent parameters of the sequencing batch sewage treatment equipment during the continuous operation cycle after normalization processing to obtain high-level features; the BP output layer is used to output the high-level features; the support vector machine regression layer is used to perform data processing on the high-level features according to the kernel function to obtain the predicted value of the COD of the effluent water quality.

[0008] Further, in the hidden layer, the number of layers is set to 6, the number of neurons in each layer is set to 24, the activation function is set to Sigmoid, the learning rate is set to 0.01, and the optimization algorithm is set to the conjugate gradient method.

[0009] Further, in the support vector machine regression layer, the kernel function is set to the radial basis kernel function, the kernel function parameter gamma is set to 0.1, the regularization parameter is set to 1, the tolerance parameter is set to 1e-3, and the maximum number of iterations is set to 100.

[0010] Further, in step S3, the COD threshold of the effluent water quality is set to 50 mg / L.

[0011] Further, in step S3, delayed monitoring is started to obtain delayed effluent parameters, and a delayed prediction decision is made based on the delayed effluent parameters until the delayed predicted value of the COD of the effluent water quality is less than the COD threshold of the effluent water quality to perform the sewage treatment compliance operation, including the following steps: B1. Collect the effluent parameters of the sequencing batch sewage treatment equipment after the delay period to obtain the delayed effluent parameters; B2. Perform normalization processing on the delayed effluent parameters, and input the influent parameters and delayed effluent parameters of the sequencing batch sewage treatment equipment during the continuous operation cycle after normalization processing into the evaluated prediction model for the COD of the effluent water quality to obtain the delayed predicted value of the COD of the effluent water quality; B3. Determine whether the delayed predicted value of the COD of the effluent water quality is less than the COD threshold of the effluent water quality. If so, perform the operation of meeting the sewage treatment standards; otherwise, return to step B1.

[0012] The present invention has the following beneficial effects: (1) By deeply coupling machine learning prediction and process control, the present invention constructs an intelligent closed-loop control method for sewage treatment, significantly improving the operation efficiency of equipment while ensuring the effluent water quality, and having outstanding industrial application value; (2) Based on the BP neural network and support vector machine regression, the present invention constructs a prediction model for the COD of the effluent water quality. By combining the feature extraction ability of the BP neural network and the regression performance of the support vector machine regression, the prediction accuracy of the COD concentration of the effluent water quality can be significantly improved; (3) By collecting the characteristic parameters of the influent water quality and the effluent water quality of the real-time sequencing batch sewage treatment equipment, and using the influent parameters, effluent parameters, corresponding effluent COD values and the prediction model of the effluent water quality COD of the continuous operation cycle of the sequencing batch sewage treatment equipment after normalization, the predicted value of the effluent water quality COD is obtained, which can reduce the chemical detection frequency, save the monitoring cost, and can achieve dynamic prediction to support the intelligent regulation of the sewage treatment equipment. Description of the Drawings

[0013] Figure 1 It is a schematic flow chart of a closed-loop control method for a sewage treatment process based on machine learning; Figure 2 It is a schematic structural diagram of a prediction model for the COD of the effluent water quality; Figure 3 It is a comparison chart of the true value and the predicted value of COD in the present invention; Wherein: ORP1 is the oxidation-reduction potential of the influent water quality, t1 is the water temperature of the influent water quality, EC1 is the conductivity of the influent water quality, ORP2 is the oxidation-reduction potential of the effluent water quality, t2 is the water temperature of the effluent water quality, and EC2 is the conductivity of the effluent water quality. Detailed Embodiments

[0014] The following describes the detailed embodiments of the present invention to facilitate those skilled in the art of the present technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the detailed embodiments. For those of ordinary skill in the art of the present technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions and creations using the concept of the present invention are within the scope of protection.

[0015] As Figure 1 shown, a closed-loop control method for a sewage treatment process based on machine learning includes steps S1 - S3, specifically as follows: S1. Collect the influent parameters, effluent parameters, and the corresponding effluent COD values of the continuous operation cycle of the sequencing batch sewage treatment equipment, and normalize the influent parameters and effluent parameters of the continuous operation cycle of the sequencing batch sewage treatment equipment.

[0016] In an alternative embodiment of the present invention, the influent parameters of the continuous operation cycle of the sequencing batch sewage treatment equipment include the oxidation-reduction potential (ORP1), conductivity (EC1), and water temperature (t1) of the influent water quality; the effluent parameters of the continuous operation cycle of the sequencing batch sewage treatment equipment include the oxidation-reduction potential (ORP2), conductivity (EC2), and water temperature (t2) of the effluent water quality. Each operation cycle of the sequencing batch sewage treatment equipment is 12 hours, and the equipment operates normally during this period.

[0017] The present invention normalizes the influent parameters and effluent parameters of the continuous operation cycle of the sequencing batch sewage treatment equipment, which is expressed as:

[0018] Where: is the original water quality characteristic data of the sequencing batch sewage treatment equipment, is the normalized water quality characteristic data of the sequencing batch sewage treatment equipment, and are the maximum and minimum values of the original water quality characteristic data of the sequencing batch sewage treatment equipment, respectively.

[0019] S2. Build a prediction model for the effluent water quality COD based on machine learning, and use the influent parameters, effluent parameters, and the corresponding effluent COD values of the continuous operation cycle of the sequencing batch sewage treatment equipment after normalization to train the prediction model for the effluent water quality COD. Deploy the trained and evaluated prediction model for the effluent water quality COD to the sequencing batch sewage treatment equipment to obtain the predicted value of the effluent water quality COD.

[0020] In an alternative embodiment of the present invention, the present invention builds a prediction model for the effluent water quality COD based on machine learning, specifically: build a prediction model for the effluent water quality COD based on the BP neural network and support vector machine regression.

[0021] As Figure 2 shown, the prediction model for the effluent water quality COD includes an input layer, a BP network hidden layer, a BP output layer, and a support vector machine regression layer connected in sequence.

[0022] The input layer is used to input the influent parameters and effluent parameters of the continuous operation cycle of the sequencing batch sewage treatment equipment after normalization; the BP network hidden layer is used to extract the non-linear features in the influent parameters and effluent parameters of the continuous operation cycle of the sequencing batch sewage treatment equipment after normalization, and obtain high-level features; the BP output layer is used to output high-level features; the support vector machine regression layer is used to process the high-level features according to the kernel function to obtain the predicted value of the COD of the effluent quality.

[0023] In the hidden layer of the present invention, the number of layers is set to 6, the number of neurons in each layer is set to 24, the activation function is set to Sigmoid, the learning rate is set to 0.01, and the optimization algorithm is set to the conjugate gradient method.

[0024] The data processing processes of the input layer, the BP network hidden layer and the BP output layer are as follows: The influent parameters and effluent parameters of the continuous operation cycle of the sequencing batch sewage treatment equipment after normalization are input into the input layer, and the output value of the hidden layer is obtained by using the input layer and the BP network hidden layer, which is expressed as:

[0025] Where: is the output value of the hidden layer, is the Sigmoid activation function of the hidden layer neuron of, is the total number of neurons in the input layer, is the weight between the input layer and the hidden layer, is the input value of the th neuron, is the threshold between the input layer and the hidden layer, is the neuron serial number in the hidden layer, is the total number of neurons in the hidden layer; According to the output value of the hidden layer, high-level features are obtained by using the BP network hidden layer and the BP output layer, which is expressed as:

[0026] Where: is the high-level feature, is the weight between the BP network hidden layer and the BP output layer, is the threshold between the BP network hidden layer and the BP output layer, is the neuron serial number in the output layer, is the total number of neurons in the output layer.

[0027] In the support vector machine regression layer of the present invention, the kernel function is set to the radial basis kernel function, the kernel function parameter gamma is set to 0.1, the regularization parameter is set to 1, the tolerance parameter is set to 1e-3, and the maximum number of iterations is set to 100.

[0028] The present invention uses the influent parameters, effluent parameters and the corresponding effluent COD values of the continuous operation cycle of the normalized sequencing batch sewage treatment equipment to train the effluent water quality COD prediction model, including the following steps: A1. Determine the influent parameters and effluent parameters of the continuous operation cycle of the normalized sequencing batch sewage treatment equipment as input features, and determine the effluent COD value corresponding to the continuous operation cycle of the sequencing batch sewage treatment equipment as the label of the output feature, so as to make a training set for predicting the effluent water quality COD.

[0029] Specifically, the training set includes 100 samples of the operation cycle of the sequencing batch sewage treatment equipment.

[0030] A2. Use the training set for predicting the effluent water quality COD to train the effluent water quality COD prediction model.

[0031] The present invention uses the mean square error and the coefficient of determination to evaluate the performance of the trained effluent water quality COD prediction model, and deploys the evaluated effluent water quality COD prediction model to the sequencing batch sewage treatment equipment for intelligent prediction of the effluent water quality COD.

[0032] The present invention uses the mean square error to evaluate the performance of the trained effluent water quality COD prediction model, which is expressed as:

[0033] Among them, is the predicted value of the effluent water quality COD of the sequencing batch sewage treatment equipment, is the true value of the effluent water quality COD of the sequencing batch sewage treatment equipment, is the sample number, is the number of samples.

[0034] The present invention uses the coefficient of determination to evaluate the performance of the trained effluent water quality COD prediction model, which is expressed as:

[0035] Among them, is the true value of the effluent water quality COD of the sequencing batch sewage treatment equipment, is the predicted value of the effluent water quality COD of the sequencing batch sewage treatment equipment, is the mean value of the true values of the effluent water quality COD of the sequencing batch sewage treatment equipment, is the number of samples.

[0036] Specifically, the present invention tests the effluent water quality COD prediction model using the test set for predicting the effluent water quality COD. As Figure 3 shown, the present invention calculates that the R² (coefficient of determination) of the effluent water quality COD prediction model on the test set reaches 0.9037, and the MSE is 18.15, indicating that the effluent water quality COD prediction model has good prediction ability for the ammonia nitrogen in the effluent water quality of the sequencing batch sewage treatment equipment.

[0037] S3. Set the effluent water quality COD threshold, and establish a dynamic feedback mechanism. Based on the predicted value of the effluent water quality COD and the effluent water quality COD threshold, and use the dynamic feedback mechanism to perform closed-loop control on the sewage treatment process.

[0038] The dynamic feedback mechanism is specifically as follows: Determine whether the predicted value of the effluent water quality COD is less than the effluent water quality COD threshold; if so, directly perform the sewage treatment compliance operation; otherwise, continue to perform the biochemical reaction, start the delayed monitoring to obtain the delayed effluent parameters, and perform the delayed prediction decision based on the delayed effluent parameters until the delayed predicted value of the effluent water quality COD is less than the effluent water quality COD threshold to perform the sewage treatment compliance operation.

[0039] In an optional embodiment of the present invention, the present invention sets the effluent water quality COD threshold to 50 mg / L.

[0040] The present invention starts the delayed monitoring to obtain the delayed effluent parameters, and performs the delayed prediction decision based on the delayed effluent parameters until the delayed predicted value of the effluent water quality COD is less than the effluent water quality COD threshold to perform the sewage treatment compliance operation, including the following steps: B1. Collect the effluent parameters of the sequencing batch sewage treatment equipment after the delayed period to obtain the delayed effluent parameters; B2. Normalize the delayed effluent parameters, and input the normalized influent parameters and delayed effluent parameters of the continuous operation period of the sequencing batch sewage treatment equipment into the evaluated effluent water quality COD prediction model to obtain the delayed predicted value of the effluent water quality COD; B3. Determine whether the delayed predicted value of the effluent water quality COD is less than the effluent water quality COD threshold. If so, perform the sewage treatment compliance operation; otherwise, return to step B1.

[0041] Specifically, the sewage treatment compliance operation is specifically as follows: First, stop the biochemical reaction, then start the precipitation process, and finally execute the drainage program.

[0042] In sewage treatment, the biochemical reaction refers to the process of using microorganisms to degrade organic pollutants in water. Specifically, the organic matter in the sewage is decomposed into harmless substances such as carbon dioxide, water, and inorganic substances through a series of biochemical reactions under the action of microorganisms. In the dynamic feedback mechanism of the present invention, if the predicted value of the COD of the effluent quality does not meet the standard, the biochemical reaction is continued to ensure that the predicted value of the COD of the final effluent quality meets the standard. This process depends on the continuous progress of the biochemical reaction to further degrade the organic matter until the water quality reaches the discharge standard.

[0043] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0044] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0045] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0046] Specific embodiments are applied in the present invention to elaborate on the principles and implementation manners of the present invention. The descriptions of the above embodiments are only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

[0047] Those of ordinary skill in the art will realize that the embodiments described herein are provided to assist the reader in understanding the principles of the present invention, and it should be understood that the scope of protection of the present invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific deformations and combinations that do not depart from the essence of the present invention based on these technical revelations disclosed in the present invention, and these deformations and combinations are still within the scope of protection of the present invention.

Claims

1. A closed-loop control method for sewage treatment process based on machine learning, characterized in that, It includes the following steps: S1. Collect the influent parameters, effluent parameters and the corresponding effluent COD values of the continuous operation cycle of the sequencing batch sewage treatment equipment, and normalize the influent parameters and effluent parameters of the continuous operation cycle of the sequencing batch sewage treatment equipment; S2. Build a prediction model for the effluent water quality COD based on machine learning, and use the influent parameters, effluent parameters and the corresponding effluent COD values of the continuous operation cycle of the sequencing batch sewage treatment equipment after normalization to train the prediction model for the effluent water quality COD. Deploy the trained and evaluated prediction model for the effluent water quality COD to the sequencing batch sewage treatment equipment to obtain the predicted value of the effluent water quality COD; S3. Set the threshold value of the effluent water quality COD, and establish a dynamic feedback mechanism. Based on the predicted value of the effluent water quality COD and the threshold value of the effluent water quality COD, use the dynamic feedback mechanism to perform closed-loop control on the sewage treatment process; The dynamic feedback mechanism is specifically as follows: Judge whether the predicted value of the effluent water quality COD is less than the threshold value of the effluent water quality COD; if so, directly perform the sewage treatment compliance operation; otherwise, continue to perform the biochemical reaction, start the delay monitoring to obtain the delayed effluent parameters, and perform the delayed prediction decision based on the delayed effluent parameters until the delayed predicted value of the effluent water quality COD is less than the threshold value of the effluent water quality COD to perform the sewage treatment compliance operation.

2. The closed-loop control method for sewage treatment process based on machine learning according to claim 1, wherein In step S1, the influent parameters of the continuous operation cycle of the sequencing batch sewage treatment equipment include the oxidation-reduction potential, conductivity and water temperature of the influent water quality; the effluent parameters of the continuous operation cycle of the sequencing batch sewage treatment equipment include the oxidation-reduction potential, conductivity and water temperature of the effluent water quality.

3. The closed-loop control method for sewage treatment process based on machine learning according to claim 1, characterized in that In step S2, building a prediction model for the effluent water quality COD based on machine learning is specifically as follows: Build a prediction model for the effluent water quality COD based on the BP neural network and support vector machine regression.

4. The closed-loop control method for sewage treatment process based on machine learning according to claim 3, wherein, The prediction model for the effluent water quality COD includes an input layer, a BP network hidden layer, a BP output layer and a support vector machine regression layer connected in sequence; the input layer is used to input the influent parameters and effluent parameters of the continuous operation cycle of the sequencing batch sewage treatment equipment after normalization; The BP network hidden layer is used to extract the non-linear features in the influent parameters and effluent parameters of the continuous operation cycle of the sequencing batch sewage treatment equipment after normalization to obtain high-level features; The BP output layer is used to output the high-level features; The support vector machine regression layer is used to perform data processing on the high-level features according to the kernel function to obtain the predicted value of the effluent water quality COD.

5. The closed-loop control method for sewage treatment process based on machine learning according to claim 4, wherein In the hidden layer, set the number of layers to 6, the number of neurons in each layer to 24, the activation function to Sigmoid, the learning rate to 0.01, and the optimization algorithm to the conjugate gradient method.

6. The closed-loop control method for sewage treatment process based on machine learning according to claim 4, characterized in that, In the support vector machine regression layer, set the kernel function to the radial basis kernel function, set the kernel function parameter gamma to 0.1, set the regularization parameter to 1, set the tolerance parameter to 1e-3, and set the maximum number of iterations to 100.

7. The closed-loop control method for sewage treatment process based on machine learning according to claim 1, characterized in that In step S3, set the threshold value of the effluent water quality COD to 50 mg / L.

8. The closed-loop control method for sewage treatment process based on machine learning according to claim 1, characterized in that, In step S3, start delay monitoring to obtain delay effluent parameters, and perform a delay prediction decision based on the delay effluent parameters until the predicted value of the COD in the effluent quality is less than the COD threshold of the effluent quality, so as to perform the operation of meeting the sewage treatment standards, including the following steps: B1. Collect the effluent parameters of the sequencing batch sewage treatment equipment after the delay period to obtain the delay effluent parameters; B2. Perform normalization processing on the delay effluent parameters, and input the influent parameters and delay effluent parameters of the continuous operation period of the sequencing batch sewage treatment equipment after normalization processing into the evaluated COD prediction model of the effluent quality to obtain the predicted value of the COD in the effluent quality; B3. Judge whether the predicted value of the COD in the effluent quality is less than the COD threshold of the effluent quality. If so, perform the operation of meeting the sewage treatment standards; otherwise, return to step B1.

Images