Sewage treatment aeration control method and system

Abstract

The invention relates to the technical field of sewage treatment, in particular to a sewage treatment aeration control method and system.The method comprises the steps that according to collected historical data, abnormal values are removed through intelligent calculation in combination with optimization and correction, an accurate sample data set is obtained and then classified through a clustering model, and all classifications are subjected to training and verification testing; a trained BP neural network model is obtained, a follow-up controller outputs accurate aeration quantity through the BP neural network model according to latest data to control each aeration device, and the increase / decrease of the air quantity of the biochemical pool is automatically adjusted; the blast volume of the biochemical pool is controlled and the air volume matching working condition is controlled according to the inlet water quality and the water volume, so that the control precision of the aeration rate in the sewage treatment process can be effectively improved, the energy consumption waste is reduced, the process control efficiency is improved, the air volume of the biochemical pool under different working conditions is ensured to be at the optimal value, the outlet water quality stability and the standard-reaching rate are ensured, and the power consumption cost is reduced; the method is suitable for sewage treatment plants of various types and scales, and has a wide market application prospect.

Description

Technical Field

[0001] The present invention relates to the technical field of sewage treatment, and more specifically, to a sewage treatment aeration control method and system. Background Art

[0002] In the activated sludge method, the aerobic aeration volume and dissolved oxygen control of the biochemical pool determine the nitrogen and phosphorus removal effect of the sewage treatment process, and is one of the most important links in the secondary treatment. At present, most sewage treatment plants have two major problems in actual operation: 1. The aeration volume of the biochemical pool is designed only considering the maximum aeration volume required for the removal of the design water quality standard under the most unfavorable conditions. In actual operation, the TN of the influent in the rainy season is about 15-25 mg / L, which is lower than the design value, and the TN of the influent in the dry season is about 35-40 mg / L, which is higher than the design value. There is no clear guidance in the design specifications on how to adjust the air volume under different working conditions in the rainy season / dry season. At the same time, the air volume displayed by the blower is the actual working condition, while the design calculates the standard condition. There is a lack of specific basis for how to convert the standard condition with the actual working condition. 2. The current production scheduling of sewage treatment plants relies on the subjective judgment of technical personnel. Due to different operating experiences, the scheduling judgment lacks scientific decision-making and predictive judgment. When encountering water quality fluctuations, frequent manual adjustments and serious delays occur, making it difficult to fine-tune and standardize operations, resulting in waste of electricity and increased labor costs.

[0003] Therefore, how to transform "experience control" into "fine control" to control the appropriate air volume, improve the water quality and reduce power consumption is an urgent problem to be solved. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art in that the aeration volume control of sewage treatment is inaccurate, and to provide a sewage treatment aeration control method and system, which can effectively improve the control accuracy of the aeration volume during sewage treatment.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0006] A sewage treatment aeration control method is provided, comprising the following steps:

[0007] S1. Data collection: Collect historical data within a set time period to form data samples of different water inflow conditions including rainy season and dry season;

[0008] S2. Process calculation: Establish a process mechanism calculation model, and calculate the theoretical oxygen demand under different water inlet conditions using the process mechanism calculation model based on the collected historical data to obtain a theoretical oxygen demand data set;

[0009] S3. Data optimization: remove outliers from the theoretical oxygen demand data set and the dissolved oxygen content in the biochemical pool in the historical data to obtain the optimized theoretical oxygen demand data set and dissolved oxygen sample set;

[0010] S4. Data clustering: Use clustering algorithms to classify the optimized theoretical oxygen demand data set and the dissolved oxygen sample set of the biochemical pool into different categories;

[0011] S5. Neural network training: Establish a BP neural network model, divide the data of the same category into a training set, a validation set and a test set respectively; input the training set into the BP neural network model, train the BP neural network model to obtain a trained BP neural network model, and use the validation set and the test set to verify and test the BP neural network model;

[0012] S6. Aeration volume prediction: input the data of the sewage pool to be tested into the trained BP neural network model to obtain the predicted aeration volume;

[0013] S7. Equipment control: Control the aeration equipment of the sewage pool according to the predicted aeration volume to achieve precise aeration.

[0014] The invention discloses a sewage treatment aeration control method. According to the collected historical data, an accurate sample data set is obtained through intelligent calculation combined with optimization and correction to eliminate abnormal values, and then the data set is classified through a clustering model. Each classification is trained and verified to obtain a trained BP neural network model. A subsequent controller controls each aeration device according to the accurate aeration amount output by the BP neural network model based on the latest data, and automatically adjusts the increase / decrease of the air volume of the biochemical pool. The invention controls the air volume of the biochemical pool according to the influent water quality and water volume, controls the air volume to match the working condition, can effectively improve the control accuracy of the aeration amount in the sewage treatment process, reduce energy waste, improve process control efficiency, ensure that the air volume of the biochemical pool is at the optimal value under different working conditions, ensure the stability and compliance rate of effluent water quality, and reduce electricity consumption cost. The invention is suitable for sewage treatment plants of various types and sizes, and has a broad market application prospect.

[0015] Preferably, in step S1, historical data of inlet chemical oxygen demand, inlet biochemical oxygen demand, inlet total nitrogen, effluent total nitrogen, inlet volume, dissolved oxygen content in the biochemical pool, and actual air volume are automatically collected at every even hour in a set time period through programming code to form a large number of data sample sets including different inlet conditions in rainy and dry seasons, which are sorted and saved in Excel tables for subsequent use.

[0016] As a preferred method, the process calculation system adopts the basic algorithm of actual inlet and outlet water quality (chemical oxygen demand, biochemical oxygen demand, total nitrogen) and inlet water volume calculation to calculate the theoretical oxygen demand, laying the foundation for the subsequent neural network fixed model;

[0017] In step S2, the theoretical oxygen demand is calculated as follows:

[0018] O2=0.001aQ(S o -S e )-cΔX v +b[0.001Q(N k -N ke )-0.12ΔX v ]

[0019] -0.62[0.001Q(N t -N ke -N oe )-0.12ΔX v ]

[0020] Where: O2 is the theoretical oxygen demand, kgO2 / d; ɑ is the oxygen equivalent of carbon, when carbon-containing substances are measured in biochemical oxygen demand, it is taken as 1.47; b is a constant, the amount of oxygen required for oxidase to oxidize kg of ammonia nitrogen, kgO2 / kgN, it is taken as 4.57; c is a constant, the oxygen equivalent of bacterial cells, it is taken as 1.42; Q is the amount of water treated, m 3 / d; S0 is the biochemical oxygen demand of the influent, mg / L; Se is the biochemical oxygen demand of the effluent, mg / L; Nk is the total nitrogen of the influent, mg / L; Nke is the total nitrogen of the effluent, mg / L; Nt is the total nitrogen concentration of the influent of the bioreactor, mg / L; Noe is the nitrate nitrogen concentration of the effluent of the bioreactor, which is the empirical value of the plant area, mg / L; ΔXv is the microbial biomass discharged from the bioreactor system, kg / d; 0.12ΔXv is the microbial nitrogen content discharged from the bioreactor system, kg / d;

[0021] S0=BOD / COD×COD 进

[0022] Se=BOD / COD×COD 出

[0023] Among them, BOD / COD is the ratio of biochemical oxygen demand to chemical oxygen demand in sewage, taking the monthly average value; COD 进 is the chemical oxygen demand of the influent, mg / L; COD 出 is the chemical oxygen demand of effluent, mg / L.

[0024] Preferably, in step S3, the interquartile range (IQR) criterion is used to detect outliers in each column of the data sample set, the features are the theoretical oxygen demand and the dissolved oxygen content in the biochemical pool, the label is the actual air volume, the interquartile range, the upper and lower limits of the critical value are calculated; and outliers greater than the upper limit of the critical value and less than the lower limit of the critical value are eliminated.

[0025] Preferably, in step S4, the optimized and sorted theoretical oxygen demand and biochemical pool dissolved oxygen sample sets are clustered and divided into K categories, and the K value with the most obvious change in the slope of the descending curve is selected as the optimal cluster number by applying the elbow rule. The theoretical oxygen demand + reaction pool dissolved oxygen data sample set (excluding the actual air volume label) is classified using a clustering algorithm. The data sample sets are divided into different categories.

[0026] Preferably, in step S5, the BP neural network model system trains and verifies the samples after being divided into K categories. When the accuracy of the test set of each category reaches the total accuracy after weighted average, the BP neural network model is saved as a trained BP neural network model after the total accuracy reaches the standard. Subsequently, the controller only needs to collect the latest influent water quality, influent water volume, and dissolved oxygen data of the biochemical pool of the sewage treatment plant, and automatically convert it into theoretical oxygen demand + biochemical pool DO through the programming program code background. After passing through the neural network fixed model, an accurate predicted air volume value is obtained and output to the equipment control.

[0027] Preferably, the step S5 comprises:

[0028] Set the number of training sets and validation sets, the number of input layers, hidden layers, output layers, and neurons;

[0029] Using nonlinear activation, the predicted air volume of each group is obtained through the BP neural network model;

[0030] The loss between the predicted air volume and the actual air volume is calculated through the built-in loss function, and then the gradient of the loss function relative to the model parameter weight is calculated by back propagation to update the gradient of each parameter;

[0031] Use the built-in function optimizer of the neural network and adjust the optimization weights and biases according to the obtained gradient to minimize the loss function to reduce the error;

[0032] The average error of each generation of training set and validation set is obtained through iteration according to the above steps, and finally the BP neural network model of the generation with the smallest average error is selected and the accuracy of the test set is tested;

[0033] When the accuracy of each category of the test set is obtained by weighted average of the total accuracy, the BP neural network model is saved as the trained BP neural network model after the total accuracy reaches the standard.

[0034] As a preferred embodiment, the equipment control has a switching of multiple operation modes, including manual control mode, automatic control mode and semi-automatic control mode, etc., to meet the needs of different scenarios. The automatic control mode and semi-automatic control mode can intelligently adjust the operation parameters of the biochemical pool blower and aeration valve according to the latest predicted precise air volume to achieve precise aeration of the biochemical pool.

[0035] The present invention also provides a sewage treatment aeration control system, comprising:

[0036] Data acquisition module: used to collect historical data within a set time period to form data samples of different water inlet conditions including rainy season and dry season;

[0037] Process calculation module: used to establish a process mechanism calculation model. Based on the collected historical data, the process mechanism calculation model is used to calculate the theoretical oxygen demand under different water inlet conditions to obtain a theoretical oxygen demand data set.

[0038] Data optimization module: used to remove outliers from the theoretical oxygen demand data set and the dissolved oxygen content in the biochemical pool in the historical data, and obtain the optimized theoretical oxygen demand data set and dissolved oxygen sample set;

[0039] Data clustering: clustering algorithms are used to classify the optimized theoretical oxygen demand data set and the dissolved oxygen sample set of the biochemical pool into different categories;

[0040] BP neural network model module: used to establish a BP neural network model, divide the data of the same category into training set, verification set and test set respectively; input the training set into the BP neural network model, train the BP neural network model to obtain the trained BP neural network model, and use the verification set and test set to verify and test the BP neural network model;

[0041] Aeration volume prediction module: used to input the data of the sewage pool to be tested into the trained BP neural network model to obtain the predicted aeration volume;

[0042] Equipment control module: used to control the aeration equipment of the sewage pool according to the predicted aeration volume to achieve precise aeration.

[0043] A sewage treatment aeration control system of the present invention includes a data acquisition module, a process calculation module, an optimization and correction module, a clustering algorithm module, a neural network training module, an aeration volume prediction module, and an equipment control module. The theoretical oxygen demand is calculated by collecting influent data, using classical process algorithms, introducing correction parameter correction and other technical means. After the clustering algorithm classifies the theoretical oxygen demand and its corresponding reaction tank DO historical data sample set, the historical data samples are combined with the "label" actual air volume of the same time period, and the BP neural network model module is trained and verified to dig out the hidden relationship between the data, and finally a fixed model is obtained. The subsequent controller only needs to collect the latest influent water quality, water volume, biochemical pool DO, and air volume, and convert them through a programming program to obtain an accurate predicted air volume. This system can establish a communication connection with the original aeration system of the sewage treatment plant to directly participate in the blower control.

[0044] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned sewage treatment aeration control method when executing the computer program.

[0045] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned sewage treatment aeration control method are implemented.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] A sewage treatment aeration control method and system of the present invention can effectively improve the control accuracy of aeration volume in the sewage treatment process, reduce energy waste, improve process control efficiency, ensure that the air volume of the biochemical pool is at the optimal value under different working conditions, ensure the stability and compliance rate of effluent water quality, and reduce electricity consumption costs; it is suitable for sewage treatment plants of various types and sizes, and has a wide market application prospect. The present invention effectively improves the accuracy of machine learning and improves prediction accuracy through the combined application of the process mechanism calculation model and the BP neural network model. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A schematic diagram of a method flow of a sewage treatment aeration control method;

[0049] Figure 2 A schematic diagram of data processing steps for a sewage treatment aeration control method;

[0050] Figure 3 The figure is a schematic diagram of the operation flow of a sewage treatment aeration control method. DETAILED DESCRIPTION

[0051] The present invention is further described below in conjunction with specific implementation methods. The accompanying drawings are only used for exemplary descriptions and are only schematic diagrams, not actual drawings, and cannot be understood as limiting this patent; in order to better illustrate the embodiments of the present invention, some parts of the accompanying drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the accompanying drawings may be omitted.

[0052] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right" and the like indicate directions or positional relationships based on the directions or positional relationships 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 direction, be constructed and operated in a specific direction. Therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes and cannot be understood as limitations on this patent. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0053] Embodiment 1

[0054] This embodiment is a first embodiment of a sewage treatment aeration control method. Figures 1 to 3 As shown, a sewage treatment aeration control method is provided, comprising the following steps:

[0055] Step S1. Data collection: Collect historical data within a set time period to form data samples including different water inlet conditions in rainy season and dry season.

[0056] In step S1, the historical data of inlet chemical oxygen demand, inlet biochemical oxygen demand, inlet total nitrogen, outlet total nitrogen, inlet volume, dissolved oxygen content in the biochemical pool, and actual air volume are automatically collected at every even hour in a set time period through programming code, forming a large number of data sample sets including different inlet conditions in rainy and dry seasons, which are sorted and saved in Excel tables for subsequent use.

[0057] Step S2. Process calculation: Establish a process mechanism calculation model, and calculate the theoretical oxygen demand under different water inlet conditions using the process mechanism calculation model based on the collected historical data to obtain a theoretical oxygen demand data set.

[0058] The process calculation system uses the actual inlet and outlet water quality (chemical oxygen demand, biochemical oxygen demand, total nitrogen) and the basic algorithm for calculating the water inlet volume to calculate the theoretical oxygen demand, laying the foundation for the subsequent neural network fixed model;

[0059] In step S2, the theoretical oxygen demand is calculated as follows:

[0060] O2=0.001aQ(S o -S e )-cΔX v +b[0.001Q(N k -N ke )-0.12ΔX v ]

[0061] -0.62[0.001Q(N t -Nke -N oe )-0.12ΔX v ]

[0062] Where: O2 is the theoretical oxygen demand, kgO2 / d; ɑ is the oxygen equivalent of carbon, when carbon-containing substances are measured in biochemical oxygen demand, it is taken as 1.47; b is a constant, the amount of oxygen required for oxidase to oxidize kg of ammonia nitrogen, kgO2 / kgN, it is taken as 4.57; c is a constant, the oxygen equivalent of bacterial cells, it is taken as 1.42; Q is the amount of water treated, m 3 / d; S0 is the biochemical oxygen demand of the influent, mg / L; Se is the biochemical oxygen demand of the effluent, mg / L; Nk is the total nitrogen of the influent, mg / L; Nke is the total nitrogen of the effluent, mg / L; Nt is the total nitrogen concentration of the influent of the bioreactor, mg / L; Noe is the nitrate nitrogen concentration of the effluent of the bioreactor, which is the empirical value of the plant area, mg / L; ΔXv is the microbial biomass discharged from the bioreactor system, kg / d; 0.12ΔXv is the microbial nitrogen content discharged from the bioreactor system, kg / d;

[0063] S0=BOD / COD×COD 进

[0064] Se=BOD / COD×COD 出

[0065] Among them, BOD / COD is the ratio of biochemical oxygen demand to chemical oxygen demand in sewage, taking the monthly average value;

[0066] COD 进 is the chemical oxygen demand of the influent, mg / L; COD 出 is the chemical oxygen demand of effluent, mg / L.

[0067] Step S3. Data optimization: outliers are removed from the theoretical oxygen demand data set and the dissolved oxygen content in the biochemical pool in the historical data to obtain an optimized theoretical oxygen demand data set and dissolved oxygen content sample set.

[0068] In step S3, the interquartile range (IQR) criterion is used to detect outliers in each column of the data sample set. The features are the theoretical oxygen demand and the dissolved oxygen content in the biochemical pool, and the label is the actual air volume. The interquartile range and the upper and lower limits of the critical value are calculated; outliers greater than the upper limit of the critical value and less than the lower limit of the critical value are eliminated.

[0069] Step S4. Data clustering: using a clustering algorithm to classify the optimized theoretical oxygen demand data set and the biochemical pool dissolved oxygen sample set into different categories.

[0070] In step S4, the optimized and sorted theoretical oxygen demand and biochemical pool dissolved oxygen sample sets are clustered and divided into K categories, and the K value with the most obvious change in the slope of the descending curve is selected as the optimal cluster number by applying the elbow rule. The theoretical oxygen demand + reaction pool dissolved oxygen data sample set (excluding the actual air volume label) is classified using a clustering algorithm. The data sample set is divided into different categories.

[0071] Step S5. Establish a BP neural network model, divide the data of the same category into a training set, a validation set and a test set; input the training set into the BP neural network model, train the BP neural network model to obtain a trained BP neural network model, and use the validation set and the test set to verify and test the BP neural network model.

[0072] In step S5, the BP neural network model system trains and verifies the samples after K categories. When the accuracy of the test set of each category reaches the total accuracy after weighted average, the BP neural network model is saved as the trained BP neural network model after the total accuracy reaches the standard. Subsequently, the controller only needs to collect the latest influent water quality, influent water volume, and dissolved oxygen data of the biochemical pool of the sewage treatment plant, and automatically convert it into theoretical oxygen demand + biochemical pool DO through the programming code background. After passing through the neural network fixed model, an accurate predicted air volume value is obtained and output to the equipment control.

[0073] Step S5 includes:

[0074] Set the number of training sets and validation sets, the number of input layers, hidden layers, output layers, and neurons;

[0075] Using nonlinear activation, the predicted air volume of each group is obtained through the BP neural network model;

[0076] The loss between the predicted air volume and the actual air volume is calculated through the built-in loss function, and then the gradient of the loss function relative to the model parameter weight is calculated by back propagation to update the gradient of each parameter;

[0077] Use the built-in function optimizer of the neural network and adjust the optimization weights and biases according to the obtained gradient to minimize the loss function to reduce the error;

[0078] The average error of each generation of training set and validation set is obtained through iteration according to the above steps, and finally the BP neural network model of the generation with the smallest average error is selected and the accuracy of the test set is tested;

[0079] When the accuracy of each category of the test set is weighted averaged to obtain the total accuracy, the aeration volume is predicted when the total accuracy reaches the standard.

[0080] Step S7. The device control saves the BP neural network model as a trained BP neural network model.

[0081] like Figure 2 As shown in the figure, during the BP neural network training process, the learning rate, number of iterations, hidden layers, etc. of the BP neural network model are first set. When it is less than the specified error or the number of iterations is reached, the predicted air volume is output. When the weighted average accuracy of the air volume is less than the specified error, the trained BP neural network model is obtained. If not, the number of classification clusters is readjusted.

[0082] Step S6. Aeration volume prediction: Input the data of the sewage pool to be tested into the trained BP neural network model to obtain the system: Control the aeration equipment of the sewage pool according to the predicted aeration volume to achieve precise aeration. The equipment control has a variety of operating mode switching, including manual control mode, automatic control mode and semi-automatic control mode, etc., to meet the needs of different scenarios. The automatic control mode and semi-automatic control mode can intelligently adjust the operating parameters of the biochemical pool blower and aeration valve according to the latest predicted precise air volume to achieve precise aeration of the biochemical pool.

[0083] A sewage treatment aeration control method of the present embodiment, based on the collected historical data, obtains an accurate sample data set through intelligent calculation combined with optimization and correction to eliminate abnormal values, and then classifies it through a clustering model. Each classification is trained and verified to obtain a trained BP neural network model. The subsequent controller controls each aeration device according to the accurate aeration amount output by the BP neural network model based on the latest data, and automatically adjusts the increase / decrease of the air volume in the biochemical pool; through process calculation, the parameters such as the inlet water quality associated with the air volume are converted into an output key parameter by establishing a process mechanism calculation model, and then cluster model classification and neural network operation are performed to reduce "features", reasonably allocate the weights of each "feature", improve the calculation accuracy of the machine learning algorithm model, and realize the "dual model" algorithm combination application of the process mechanism calculation model + BP neural network model. This embodiment controls the air volume of the biochemical pool according to the influent water quality and water volume, controls the air volume to match the working conditions, can effectively improve the control accuracy of the aeration volume in the sewage treatment process, reduce energy waste, improve process control efficiency, ensure that the air volume of the biochemical pool is at the optimal value under different working conditions, ensure the stability and compliance rate of the effluent water quality, and reduce electricity consumption costs; it is suitable for sewage treatment plants of various types and sizes, and has a broad market application prospect.

[0084] Embodiment 2

[0085] This embodiment is a first embodiment of a sewage treatment aeration control system, comprising:

[0086] Data acquisition module: used to collect historical data within a set time period, forming data samples including different water inlet conditions in rainy season and dry season. Through programming code, historical data of inlet chemical oxygen demand, inlet biochemical oxygen demand, inlet total nitrogen, outlet total nitrogen, inlet volume, dissolved oxygen in the biochemical pool, and actual air volume are automatically collected at every even hour in a set time period, forming a large number of data sample sets including different water inlet conditions in rainy season and dry season, which are sorted and saved in Excel tables for subsequent use.

[0087] Process calculation module: used to establish a process mechanism calculation model. Based on the collected historical data, the process mechanism calculation model is used to calculate the theoretical oxygen demand under different water inlet conditions to obtain a theoretical oxygen demand data set. The process calculation system uses the actual inlet and outlet water quality (chemical oxygen demand, biochemical oxygen demand, total nitrogen) and inlet water calculation basic algorithms to calculate the theoretical oxygen demand, laying the foundation for the subsequent neural network fixed model.

[0088] Theoretical oxygen demand is calculated as follows:

[0089] O2=0.001aQ(S o -S e )-cΔX v +b[0.001Q(N k -N ke )-0.12ΔX v ]

[0090] -0.62[0.001Q(N t -N ke -N oe )-0.12ΔX v ]

[0091] Where: O2 is the theoretical oxygen demand, kgO2 / d; ɑ is the oxygen equivalent of carbon, when carbon-containing substances are measured in biochemical oxygen demand, it is taken as 1.47; b is a constant, the amount of oxygen required for oxidase to oxidize kg of ammonia nitrogen, kgO2 / kgN, it is taken as 4.57; c is a constant, the oxygen equivalent of bacterial cells, it is taken as 1.42; Q is the amount of water treated, m 3 / d; S0 is the biochemical oxygen demand of the influent, mg / L; Se is the biochemical oxygen demand of the effluent, mg / L; Nk is the total nitrogen of the influent, mg / L; Nke is the total nitrogen of the effluent, mg / L; Nt is the total nitrogen concentration of the influent of the bioreactor, mg / L; Noe is the nitrate nitrogen concentration of the effluent of the bioreactor, which is the empirical value of the plant area, mg / L; ΔXv is the microbial biomass discharged from the bioreactor system, kg / d; 0.12ΔXv is the microbial nitrogen content discharged from the bioreactor system, kg / d;

[0092] S0=BOD / COD×COD 进

[0093] Se=BOD / COD×COD 出

[0094] Among them, BOD / COD is the ratio of biochemical oxygen demand to chemical oxygen demand in sewage, taking the monthly average value;

[0095] COD 进 is the chemical oxygen demand of the influent, mg / L; COD 出 is the chemical oxygen demand of effluent, mg / L.

[0096] Data optimization module: used to remove outliers from the theoretical oxygen demand data set and the dissolved oxygen content in the biochemical pool in the historical data, and obtain the optimized theoretical oxygen demand data set and dissolved oxygen sample set. The interquartile range (IQR) criterion is used to detect outliers in each column of the data sample set, with the features being the theoretical oxygen demand and dissolved oxygen content in the biochemical pool, and the label being the actual air volume. The interquartile range and the upper and lower limits of the critical value are calculated; outliers greater than the upper limit of the critical value and less than the lower limit of the critical value are removed.

[0097] Data clustering: Use clustering algorithms to classify the optimized theoretical oxygen demand data set and the dissolved oxygen sample set of the biochemical pool into different categories. Perform cluster analysis on the optimized and sorted theoretical oxygen demand and dissolved oxygen sample sets of the biochemical pool, divide them into K categories, and by applying the elbow rule, select the K value with the most obvious change in the slope of the descending curve as the optimal cluster number. Use clustering algorithms to classify the theoretical oxygen demand + dissolved oxygen data sample set of the reaction pool (excluding the actual air volume of the label). Divide the data sample sets into different categories.

[0098] BP neural network model module: used to establish a BP neural network model, divide the data of the same category into training set, validation set and test set; input the training set into the BP neural network model, train the BP neural network model, obtain the trained BP neural network model, and verify and test the BP neural network model with the validation set and test set. Subsequently, the controller only needs to collect the latest influent water quality, influent water volume, and dissolved oxygen data of the biochemical pool of the sewage treatment plant, and automatically convert it into theoretical oxygen demand + biochemical pool DO through the programming code background. After passing through the neural network fixed model, an accurate predicted air volume value is obtained and output to the equipment control.

[0099] The processing process includes: setting the number of generations of training sets and validation sets, the number of input layers, hidden layers, output layers, and neurons; using nonlinear activation to obtain the predicted air volume of each group through the BP neural network model; calculating the loss between the predicted air volume and the actual air volume through the built-in loss function, and then performing back propagation to calculate the gradient of the loss function relative to the model parameter weight, and updating the gradient of each parameter; using the built-in function optimizer of the neural network and adjusting the optimization weights and deviations according to the obtained gradient, so as to minimize the loss function to reduce the error; obtaining the average error of each generation of training sets and validation sets according to the above steps in an iterative manner, and finally selecting the generation of BP neural network model with the smallest average error, and using it to test the accuracy of the test set; when the accuracy of each type of test set is obtained by weighted average of the total accuracy, the BP neural network model is saved after the total accuracy meets the standard as the trained BP neural network model.

[0100] Aeration volume prediction module: used to input the data of the sewage pool to be tested into the trained BP neural network model to obtain the predicted aeration volume.

[0101] Equipment control module: used to control the aeration equipment of the sewage pool according to the predicted aeration volume to achieve precise aeration. The equipment control has multiple operation mode switching, including manual control mode, automatic control mode and semi-automatic control mode, to meet the needs of different scenarios. The automatic control mode and semi-automatic control mode can intelligently adjust the operating parameters of the biochemical pool blower and aeration valve according to the latest predicted precise air volume to achieve precise aeration of the biochemical pool.

[0102] A sewage treatment aeration control system of the present embodiment includes a data acquisition module, a process calculation module, an optimization and correction module, a clustering algorithm module, a neural network training module, an aeration volume prediction module, and an equipment control module. The theoretical oxygen demand is calculated by collecting influent data, using classical process algorithms, introducing correction parameter correction and other technical means. After the clustering algorithm classifies the theoretical oxygen demand and its corresponding reaction tank DO historical data sample set, the historical data samples are combined with the "label" actual air volume of the same time period, and the BP neural network model module is trained and verified to dig out the hidden relationship between the data, and finally a fixed model is obtained. The subsequent controller only needs to collect the latest influent water quality, water volume, biochemical pool DO, and air volume, and convert them through the programming program to obtain accurate predicted air volume. This system can establish a communication connection with the original aeration system of the sewage treatment plant to directly participate in the blower control.

[0103] Embodiment 3

[0104] This embodiment is an embodiment of a computer device, including a memory and a processor. The memory stores a computer program, and the processor implements the steps of the sewage treatment aeration control method described in the above embodiment 1 when executing the computer program.

[0105] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned sewage treatment aeration control method are implemented.

[0106] Embodiment 4

[0107] This embodiment is an embodiment of a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the sewage treatment aeration control method described in the above embodiment 1 are implemented.

[0108] In the specific contents of the above-mentioned specific implementation methods, the various technical features can be combined in any non-contradictory manner. In order to make the description concise, not all possible combinations of the above-mentioned technical features are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0109] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A sewage treatment aeration control method, characterized in that: The following steps are involved: S1. Data collection: Collect historical data within a set time period to form data samples of different water inflow conditions including rainy season and dry season; S2. Process calculation: Establish a process mechanism calculation model, and calculate the theoretical oxygen demand under different water inlet conditions using the process mechanism calculation model based on the collected historical data to obtain a theoretical oxygen demand data set; S3. Data optimization: remove outliers from the theoretical oxygen demand data set and the dissolved oxygen content in the biochemical pool in the historical data to obtain the optimized theoretical oxygen demand data set and dissolved oxygen sample set; S4. Data clustering: Use clustering algorithms to classify the optimized theoretical oxygen demand data set and the dissolved oxygen sample set of the biochemical pool into different categories; S5. Neural network training: Establish a BP neural network model, divide the data of the same category into a training set, a validation set and a test set respectively; input the training set into the BP neural network model, train the BP neural network model to obtain a trained BP neural network model, and use the validation set and the test set to verify and test the BP neural network model; S6. Aeration volume prediction: input the data of the sewage pool to be tested into the trained BP neural network model to obtain the predicted aeration volume; S7. Equipment control: Control the aeration equipment of the sewage pool according to the predicted aeration volume to achieve precise aeration.

2. The sewage treatment aeration control method according to claim 1, characterized in that: In step S1, the historical data of inlet chemical oxygen demand, inlet biochemical oxygen demand, inlet total nitrogen, outlet total nitrogen, inlet volume, dissolved oxygen content in the biochemical pool, and actual air volume are automatically collected at every even hour in a set time period through programming code, forming a large number of data sample sets including different inlet conditions in rainy and dry seasons, which are sorted and saved in Excel tables for subsequent use.

3. The sewage treatment aeration control method according to claim 1, characterized in that: In step S2, the theoretical oxygen demand is calculated as follows: O2=0.001aQ(S o -S e )-cΔX v +b[0.001Q(N k -N ke )-0.12ΔX v ] -0.62[0.001Q(N t -N ke -N oe )-0.12ΔX v ] Where: O2 is the theoretical oxygen demand; ɑ is the oxygen equivalent of carbon, when the carbon-containing substance is measured in biochemical oxygen demand, it is taken as 1.47; b is a constant, the amount of oxygen required for oxidase to oxidize kg of ammonia nitrogen, it is taken as 4.57; c is a constant, the oxygen equivalent of bacterial cells, it is taken as 1.42; Q is the treated water volume; S0 is the influent biochemical oxygen demand; S e N is the biochemical oxygen demand of effluent; k is the total nitrogen in the influent; N ke is the total nitrogen in the effluent; Nt is the total nitrogen concentration in the bioreactor inlet; Noe is the nitrate nitrogen concentration in the bioreactor effluent; ΔXv is the amount of microorganisms discharged from the bioreactor system; 0.12ΔXv is the nitrogen content of microorganisms discharged from the bioreactor system; S0=BOD / CODE×CODE 进 Se=BOD / CODE×CODE 出 Among them, BOD / COD is the ratio of biochemical oxygen demand to chemical oxygen demand in sewage, taking the monthly average value; COD 进 Chemical oxygen demand of influent; COD 出 Chemical oxygen demand of effluent.

4. The sewage treatment aeration control method according to claim 1, characterized in that: In step S3, the interquartile range (IQR) criterion is used to detect outliers in each column of the data sample set. The features are the theoretical oxygen demand and the dissolved oxygen content in the biochemical pool, and the label is the actual air volume. The interquartile range and the upper and lower limits of the critical value are calculated; outliers greater than the upper limit of the critical value and less than the lower limit of the critical value are eliminated.

5. The sewage treatment aeration control method according to claim 4, characterized in that: In step S4, cluster analysis is performed on the optimized and sorted theoretical oxygen demand and dissolved oxygen sample sets in the biochemical pool, and the sample sets are divided into K categories. By applying the elbow rule, the K value with the most obvious change in the slope of the descending curve is selected as the optimal cluster number.

6. The sewage treatment aeration control method according to claim 5, characterized in that: In step S5, the BP neural network model system trains and verifies the samples divided into K categories. When the accuracy of the test set of each category reaches the total accuracy after weighted average, the BP neural network model is saved as the trained BP neural network model after the total accuracy reaches the standard.

7. The sewage treatment aeration control method according to claim 6, characterized in that: The step S5 comprises: Set the number of training sets and validation sets, the number of input layers, hidden layers, output layers, and neurons; Using nonlinear activation, the predicted air volume of each group is obtained through the BP neural network model; The loss between the predicted air volume and the actual air volume is calculated through the built-in loss function, and then the gradient of the loss function relative to the model parameter weight is calculated by back propagation to update the gradient of each parameter; Use the built-in function optimizer of the neural network and adjust the optimization weights and biases according to the obtained gradient to minimize the loss function to reduce the error; The average error of each generation of training set and validation set is obtained through iteration according to the above steps, and finally the BP neural network model of the generation with the smallest average error is selected, and the accuracy is tested with the test set; When the accuracy of each category of the test set is obtained by weighted average of the total accuracy, the BP neural network model is saved after the total accuracy reaches the standard as the trained BP neural network model.

8. A sewage treatment aeration control system, characterized in that: include: Data acquisition module: used to collect historical data within a set time period to form data samples of different water inlet conditions including rainy season and dry season; Process calculation module: used to establish a process mechanism calculation model. Based on the collected historical data, the process mechanism calculation model is used to calculate the theoretical oxygen demand under different water inlet conditions to obtain a theoretical oxygen demand data set. Data optimization module: used to remove outliers from the theoretical oxygen demand data set and the dissolved oxygen content in the biochemical pool in the historical data, and obtain the optimized theoretical oxygen demand data set and dissolved oxygen sample set; Data clustering: clustering algorithms are used to classify the optimized theoretical oxygen demand data set and the dissolved oxygen sample set of the biochemical pool into different categories; BP neural network model module: used to establish a BP neural network model, divide the data of the same category into training set, verification set and test set respectively; input the training set into the BP neural network model, train the BP neural network model to obtain the trained BP neural network model, and use the verification set and test set to verify and test the BP neural network model; Aeration volume prediction module: used to input the data of the sewage pool to be tested into the trained BP neural network model to obtain the predicted aeration volume; Equipment control module: used to control the aeration equipment of the sewage pool according to the predicted aeration volume to achieve precise aeration.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the sewage treatment aeration control method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the sewage treatment aeration control method according to any one of claims 1 to 7 are implemented.

Images