Wastewater Energy Saving Control Method and System Based on Multidimensional Variable Data Analysis

Abstract

This invention discloses a wastewater energy-saving control method and system based on multidimensional variable data analysis, belonging to the field of data analysis technology. It includes analyzing the correlations between variable data at different treatment stages. These correlations include the relationship between independent and dependent variables, as well as the interaction relationships between independent variables, accurately describing the complex coupling relationships between multiple variables and providing a reliable foundation for subsequent regulation. The method determines the regulation delay and variability at each regulation node, considering the regulation lag and variable changes in the wastewater treatment process, thereby setting thresholds and generating evaluation indicators to guide subsequent wastewater regulation. By controlling the regulation strategy at each node based on the correlations between variable data and the regulation delay, the adaptability and accuracy of wastewater node analysis and control are improved, ensuring real-time wastewater treatment effects and enhancing wastewater treatment efficiency.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, and in particular to a wastewater energy-saving control method and system based on multidimensional variable data analysis. Background Technology

[0002] The wastewater energy-saving control scheme based on multidimensional variable data analysis relies on modern sensing technology, big data analysis, and machine learning algorithms to address the problems of lagging regulation and coarse parameters in traditional wastewater treatment. Traditional methods often rely on threshold control of single indicators (such as COD and pH), which is difficult to cope with water quality fluctuations and multivariate coupling effects, resulting in low treatment efficiency and high energy consumption. Current technology deploys a high-precision sensor network to collect multidimensional data such as flow rate, dissolved oxygen, and temperature in real time. Combined with time-lag analysis and association rule mining, it reveals the dynamic relationships between variables (such as the time-lag effect between influent water quality and aeration rate). This technology significantly improves system response speed and anti-interference capability, enabling wastewater treatment to shift from experience-driven to data-driven approaches, providing core support for smart water management.

[0003] In existing technologies, threshold control of a single indicator (such as COD or pH) results in poor adaptability and accuracy of wastewater node analysis and control, and cannot guarantee the real-time treatment effect of wastewater.

[0004] Therefore, improving the adaptability and accuracy of wastewater node analysis and control is a technical problem that needs to be solved. Summary of the Invention

[0005] The purpose of this invention is to address the problem of poor adaptability and accuracy in existing wastewater node analysis and control technologies, and to propose a wastewater energy-saving control method based on multidimensional variable data analysis, which includes:

[0006] The wastewater treatment process is obtained, divided into multiple treatment stages, and monitoring and control nodes are deployed at each treatment stage to determine all variables involved in each treatment stage.

[0007] Acquire historical variable data of the wastewater treatment process and analyze the correlation between variable data at different treatment stages;

[0008] Historical variable data is analyzed to determine the control delay and variability at each control node. The variability is used to set the variable data threshold for each control node and generate evaluation indicators.

[0009] Based on the evaluation indicators, the control strategy for each control node is controlled according to the correlation between variable data and the control time delay, so as to ensure the real-time and efficient nature of sewage treatment.

[0010] In some embodiments of this application, the wastewater treatment process includes a pretreatment stage, a primary treatment stage, a secondary treatment stage, an advanced treatment stage, and a sludge disposal stage, with each stage further including multiple treatment units.

[0011] In some embodiments of this application, monitoring nodes and control nodes are deployed during the processing phase, including,

[0012] Determine the processing flow and processing requirements of all processing units in each processing stage, and determine the location of monitoring nodes and control nodes in the processing flow based on the processing flow and processing requirements.

[0013] Monitoring nodes and control nodes are deployed at the locations of monitoring nodes and control nodes, respectively. Monitoring nodes are used to collect and monitor variable data at the corresponding locations in the treatment process, while control nodes are used to control and adjust the wastewater treatment strategy at the corresponding locations in the treatment process.

[0014] In some embodiments of this application, the correlation between variable data at different processing stages is analyzed, including,

[0015] Based on the variable data between different processing stages, independent and dependent variables are distinguished, and the matching relationship between independent and dependent variables is established. The association between independent and dependent variables is described by a multinomial regression model between independent and dependent variables.

[0016] The model parameters in the multinomial regression model are divided into fixed model parameters and variable model parameters;

[0017] Fixed model parameters include polynomial order and model interaction terms;

[0018] Calculate the degree of nonlinearity between independent and dependent variables and the noise level of each independent and dependent variable under each matching relationship. Determine the overall noise level of the matching relationship between independent and dependent variables based on the noise levels of each independent and dependent variable. Combine the degree of nonlinearity and the overall noise level to determine the complexity. Map a polynomial order through the complexity.

[0019] The combinations of independent variables are determined based on the matching relationship between independent and dependent variables. Each combination of independent variables is detected by the response surface methodology, and the combinations of independent variables are selected as model interaction terms.

[0020] The initialization of variable model parameters is set based on the independent variable-dependent variable, and the multinomial regression model is trained based on the initialized variable model parameters and the multinomial order.

[0021] In some embodiments of this application, the initialization of variable model parameters is set according to the independent variable-dependent variable relationship, and a multinomial regression model is trained based on the initialized variable model parameters and the multinomial order, including:

[0022] Variable model parameters include independent variable coefficients and interaction term coefficients;

[0023] Based on the matching relationship between independent and dependent variables, the correlation between each independent variable and the dependent variable is calculated, and the correlation between independent variables in the model interaction term is calculated. The correlation between independent and dependent variables is recorded as the first correlation, and the correlation between independent variables in the model interaction term is recorded as the second correlation.

[0024] The initial values ​​of the coefficients of each independent variable are determined based on the first correlation, and the initial values ​​of the coefficients of each interaction term are determined based on the second correlation.

[0025] Obtain the data for the independent and dependent variables, and divide the data into training and test sets based on their timestamps;

[0026] The model structure is defined based on the polynomial order, model interaction terms, initial values ​​of independent variable coefficients, and initial values ​​of interaction term coefficients. Based on the training and test sets, the initial values ​​of independent variable coefficients and interaction term coefficients are continuously optimized through the loss function until the loss function converges, resulting in the final polynomial regression model.

[0027] In some embodiments of this application, historical variable data is analyzed to determine the corresponding control delay and variability at each control node, including:

[0028] The historical variable data of the control node is a record of historical variables, which records the changes in variables over time and control information.

[0029] Mark the multiple control positions on the historical variable record according to the control information, calculate the cross-correlation function between the control information and the variable data, take the time node corresponding to the maximum value of the cross-correlation function as the first time lag value of a single control position, and combine the multiple control positions to obtain the average value of the first time lag value, and take the average value as the first time lag value of the control node.

[0030] Multiple control positions and control information are used as step inputs, and response curves of variable data are plotted. The curve time characteristics at the points of significant change in variable data are determined on the response curves, and the second time lag value of the control node is determined based on the curve time characteristics.

[0031] The time lag target value is determined by combining the first and second time lag values ​​of the control node;

[0032] The variability description parameters for each variable data under the statistical control node are integrated to determine the change index of each variable data, and the variability of the variable data is described through the change index.

[0033] In some embodiments of this application, the variable data threshold for each control node is set by variability, including:

[0034] Determine the basic threshold for each variable data at each control node, map a safety margin based on the change index of each variable data, and superimpose the safety margin onto the basic threshold to form the threshold range.

[0035] In some embodiments of this application, the generation of evaluation metrics includes,

[0036] By comparing the actual values ​​of variable data with the threshold range of variable data, the deviation is obtained. The control response time is statistically analyzed, and evaluation indicators are generated based on the deviation and control response time.

[0037] Correspondingly, this application also provides a wastewater energy-saving control system based on multidimensional variable data analysis, including,

[0038] The first module is used to acquire the wastewater treatment process, divide the wastewater treatment process into multiple treatment stages, deploy monitoring nodes and control nodes in the treatment stages, and determine all variables involved in each treatment stage.

[0039] The second module is used to acquire historical variable data of the wastewater treatment process and analyze the correlation between variable data at different treatment stages.

[0040] The third module is used to analyze historical variable data to determine the corresponding control delay and variability at each control node, and to set the variable data threshold for each control node based on the variability, thereby generating evaluation indicators.

[0041] The fourth module is used to control the control strategy of each control node based on the correlation between variable data and the control time delay, based on the evaluation indicators, so as to ensure the real-time and efficient nature of sewage treatment.

[0042] Compared with the prior art, the beneficial effects of this invention are as follows:

[0043] 1. Analyze the correlations between variable data at different treatment stages. These correlations include the relationship between independent and dependent variables, as well as the interaction relationships among independent variables. This provides a comprehensive and accurate description of the complex coupling relationships between multiple variables, offering a reliable foundation for subsequent control. Determine the control time delay and variability at each control node, considering the control lag and variable changes in the wastewater treatment process. This allows for the setting of thresholds and the generation of evaluation indicators to guide subsequent wastewater control.

[0044] 2. By controlling the control strategy of each control node based on the correlation between variable data and the control time delay, the adaptability and accuracy of wastewater node analysis and control are improved, ensuring the real-time treatment effect of wastewater and improving wastewater treatment efficiency and energy saving. Attached Figure Description

[0045] Figure 1 This is a schematic diagram of the wastewater energy-saving control method based on multidimensional variable data analysis proposed in this invention;

[0046] Figure 2 This is a schematic diagram of the wastewater energy-saving control system based on multidimensional variable data analysis proposed in this invention. Detailed Implementation

[0047] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0048] Reference Figure 1 A wastewater energy-saving control method based on multidimensional variable data analysis includes the following steps:

[0049] Step S101: Obtain the wastewater treatment process, divide the wastewater treatment process into multiple treatment stages, deploy monitoring nodes and control nodes in the treatment stages, and determine all variables involved in each treatment stage.

[0050] In this embodiment, the processing stages are divided as follows:

[0051] Pretreatment stage: screen, grit chamber (to remove large particulate impurities).

[0052] Primary treatment stage: primary sedimentation tank (removal of suspended solids, SS removal rate 40-50%).

[0053] Secondary treatment stage: bioreactor (activated sludge process or MBR, COD removal rate 85-95%).

[0054] Advanced treatment stage: reverse osmosis, advanced oxidation (further removal of nitrogen, phosphorus, and micro-pollutants).

[0055] Sludge treatment stage: sludge thickening, dewatering, and incineration (resource utilization).

[0056] In some embodiments of this application, the wastewater treatment process includes a pretreatment stage, a primary treatment stage, a secondary treatment stage, an advanced treatment stage, and a sludge disposal stage, with each stage further including multiple treatment units.

[0057] In this embodiment, specifically:

[0058] Objective: To remove large suspended solids, floating matter, and some organic matter from wastewater, creating favorable conditions for subsequent treatment.

[0059] Main processing unit:

[0060] Grilles: Used to intercept larger suspended and floating objects, such as tree branches and plastics.

[0061] Grit chamber: Removes inorganic sand particles from wastewater through gravity sedimentation.

[0062] Equalization tank: Regulates the quantity and quality of wastewater to make the subsequent treatment process more stable.

[0063] Primary sedimentation tank (optional): In some cases, the pretreatment stage may also include a primary sedimentation tank to remove some of the suspended solids.

[0064] 2. First-level processing stage

[0065] Objective: To further remove suspended solids from wastewater using physical methods, thereby reducing the concentrations of BOD (biochemical oxygen demand) and SS (suspended solids).

[0066] Main processing unit:

[0067] Primary sedimentation tank (if not set up in the pretreatment stage): Removes suspended solids from wastewater by gravity sedimentation.

[0068] Other physical treatment units, such as flotation tanks, are used to remove grease and fine suspended solids from wastewater.

[0069] 3. Secondary processing stage

[0070] Objective: To remove organic matter and nutrients such as nitrogen and phosphorus from wastewater using biological treatment methods, thereby significantly reducing the pollution load of wastewater.

[0071] Main processing unit:

[0072] Activated sludge process: Removes organic matter from wastewater through the metabolic activity of microorganisms.

[0073] Biofilm method: such as biofilters, biodiscs, etc., which utilize microorganisms on the biofilm to degrade pollutants in wastewater.

[0074] Oxidation ditch: an improved activated sludge process with a longer hydraulic retention time and better nitrogen and phosphorus removal effects.

[0075] Sequencing Batch Reactor (SBR): A wastewater treatment technology using activated sludge that operates in an intermittent aeration mode.

[0076] 4. Deep processing stage

[0077] Objective: To further remove recalcitrant organic matter, microorganisms, salts, and specific pollutants from wastewater, so that the treated water meets higher discharge standards or reuse requirements.

[0078] Main processing unit:

[0079] Filtration: such as sand filtration and activated carbon filtration, is used to remove fine suspended solids and dissolved organic matter from wastewater.

[0080] Disinfection: such as ultraviolet disinfection and chlorine disinfection, to kill pathogenic microorganisms in sewage.

[0081] Membrane separation technologies, such as reverse osmosis, ultrafiltration, and nanofiltration, remove specific pollutants from wastewater through the selective permeability of membranes.

[0082] Advanced oxidation technologies, such as ozone oxidation and Fenton oxidation, are used to degrade recalcitrant organic matter in wastewater.

[0083] 5. Sludge treatment and disposal stage (accompanying the entire treatment process)

[0084] Objective: To treat and dispose of sludge generated during wastewater treatment to prevent secondary pollution.

[0085] Main processing unit:

[0086] Sludge thickening: reducing the volume of sludge.

[0087] Sludge dewatering: Further reduces the moisture content of sludge to facilitate subsequent treatment.

[0088] Sludge stabilization: Stabilizing the organic matter in sludge through biological or chemical methods.

[0089] Sludge disposal: such as landfill, incineration, composting, etc., to safely dispose of the treated sludge. In some embodiments of this application, monitoring nodes and control nodes are deployed during the treatment phase, including,

[0090] Determine the processing flow and processing requirements of all processing units in each processing stage, and determine the location of monitoring nodes and control nodes in the processing flow based on the processing flow and processing requirements.

[0091] Monitoring nodes and control nodes are deployed at the locations of monitoring nodes and control nodes, respectively. Monitoring nodes are used to collect and monitor variable data at the corresponding locations in the treatment process, while control nodes are used to control and adjust the wastewater treatment strategy at the corresponding locations in the treatment process.

[0092] In this embodiment, monitoring nodes are deployed at the inlets and outlets of key units (such as the effluent outlet of the primary sedimentation tank and the dissolved oxygen probe of the bioreactor). Control nodes are deployed at controllable equipment (such as aeration blowers and sludge return pumps). The node locations are determined based on treatment requirements (e.g., the aeration rate control node is located at the front end of the bioreactor). Monitoring nodes collect data such as flow rate, COD, and dissolved oxygen in real time (sampling frequency 1 time / minute).

[0093] It is understandable that the control nodes here are used to control and adjust the wastewater treatment strategy at the corresponding location in the treatment process. This wastewater treatment strategy can be achieved by controlling some chemical dosages or related parameters, such as adjusting the aeration rate, sludge return ratio, and chemical dosage.

[0094] Step S102: Obtain historical variable data of the wastewater treatment process and analyze the correlation between variable data of different treatment stages.

[0095] In this example, we have influent COD (independent variable) vs. secondary treated dissolved oxygen (dependent variable). Note that there can be multiple independent variables, and often a single dependent variable involves multiple independent variables. Furthermore, there may be mutual influence relationships between the independent variables.

[0096] In some embodiments of this application, the correlation between variable data at different processing stages is analyzed, including,

[0097] Based on the variable data between different processing stages, independent and dependent variables are distinguished, and the matching relationship between independent and dependent variables is established. The association between independent and dependent variables is described by a multinomial regression model between independent and dependent variables.

[0098] The model parameters in the multinomial regression model are divided into fixed model parameters and variable model parameters;

[0099] Fixed model parameters include polynomial order and model interaction terms;

[0100] Calculate the degree of nonlinearity between independent and dependent variables and the noise level of each independent and dependent variable under each matching relationship. Determine the overall noise level of the matching relationship between independent and dependent variables based on the noise levels of each independent and dependent variable. Combine the degree of nonlinearity and the overall noise level to determine the complexity. Map a polynomial order through the complexity.

[0101] The combinations of independent variables are determined based on the matching relationship between independent and dependent variables. Each combination of independent variables is detected by the response surface methodology, and the combinations of independent variables are selected as model interaction terms.

[0102] The initialization of variable model parameters is set based on the independent variable-dependent variable, and the multinomial regression model is trained based on the initialized variable model parameters and the multinomial order.

[0103] In this embodiment, a multinomial regression model is used to describe the relationship between the independent and dependent variables. Fixed model parameters are those that need to be determined in advance, while variable model parameters are continuously optimized through training with subsequent data. The degree of nonlinearity can be described using cross-correlation functions or partial correlation coefficients. The complexity is determined by combining the degree of nonlinearity with the overall noise level (calculated by normalization and weighted summation, etc.). If the data exhibits a significant nonlinear relationship, a higher polynomial order is needed to capture this nonlinearity. Noisy data may require a lower polynomial order to avoid overfitting the noise.

[0104] In this embodiment, there may be interactions (mutual influences) between independent variables. The Response Surface Method (RSM) is used to draw a three-dimensional response surface plot and observe the degree of surface curvature (such as saddle points indicating strong interactions).

[0105] In some embodiments of this application, the initialization of variable model parameters is set according to the independent variable-dependent variable relationship, and a multinomial regression model is trained based on the initialized variable model parameters and the multinomial order, including:

[0106] Variable model parameters include independent variable coefficients and interaction term coefficients;

[0107] Based on the matching relationship between independent and dependent variables, the correlation between each independent variable and the dependent variable is calculated, and the correlation between independent variables in the model interaction term is calculated. The correlation between independent and dependent variables is recorded as the first correlation, and the correlation between independent variables in the model interaction term is recorded as the second correlation.

[0108] The initial values ​​of the coefficients of each independent variable are determined based on the first correlation, and the initial values ​​of the coefficients of each interaction term are determined based on the second correlation.

[0109] Obtain the data for the independent and dependent variables, and divide the data into training and test sets based on their timestamps;

[0110] The model structure is defined based on the polynomial order, model interaction terms, initial values ​​of independent variable coefficients, and initial values ​​of interaction term coefficients. Based on the training and test sets, the initial values ​​of independent variable coefficients and interaction term coefficients are continuously optimized through the loss function until the loss function converges, resulting in the final polynomial regression model.

[0111] In this embodiment, the correlation can be described using the Pearson correlation coefficient. Training steps:

[0112] 1. Calculate the correlation between the independent variable and the dependent variable (e.g., Pearson coefficient).

[0113] 2. Initialization coefficients (high correlation corresponds to high initial values).

[0114] 3. Divide the dataset into a training set (70%) and a test set (30%), and use mean squared error (MSE) as the loss function.

[0115] 4. Iterate and optimize the coefficients until the MSE converges.

[0116] For example, the relationship between two independent variables and a dependent variable can be represented by a multinomial regression model as follows:

[0117] ;

[0118] in, As the dependent variable, It is the intercept term. , They are respectively , The coefficients of the independent variable, It is the coefficient of the interaction term. , Each of the two independent variables is a separate variable. This is the error term. It should be noted that the multinomial regression model can be adjusted based on the specific relationships between the variables, and this adjustment can be made multiple times. The intercept and error terms are continuously optimized through training.

[0119] Step S103: Analyze historical variable data to determine the corresponding control delay and variability at each control node, and use the variability to set the variable data threshold for each control node to generate evaluation indicators.

[0120] In this embodiment, due to the continuity and complexity of the wastewater treatment process, there will be a time lag in regulation; that is, the effect of a certain regulation will only be reflected and captured by the data some time after the regulation point. Cross-correlation function method: Calculate the cross-correlation function between the regulation action and the variable response; the peak corresponds to the time lag. Step response method: Plot the variable response curve after regulation to determine the time point of significant change. Variability refers to the change in variable data. Due to the fluctuations in the wastewater treatment process, fixed variable data thresholds are difficult to accurately describe the rationality of wastewater treatment regulation; therefore, variable data thresholds are set considering the change in variable data.

[0121] In some embodiments of this application, historical variable data is analyzed to determine the corresponding control delay and variability at each control node, including:

[0122] The historical variable data of the control node is a record of historical variables, which records the changes in variables over time and control information.

[0123] Mark the multiple control positions on the historical variable record according to the control information, calculate the cross-correlation function between the control information and the variable data, take the time node corresponding to the maximum value of the cross-correlation function as the first time lag value of a single control position, and combine the multiple control positions to obtain the average value of the first time lag value, and take the average value as the first time lag value of the control node.

[0124] Multiple control positions and control information are used as step inputs, and response curves of variable data are plotted. The curve time characteristics at the points of significant change in variable data are determined on the response curves, and the second time lag value of the control node is determined based on the curve time characteristics.

[0125] The time lag target value is determined by combining the first and second time lag values ​​of the control node;

[0126] The variability description parameters for each variable data under the statistical control node are integrated to determine the change index of each variable data, and the variability of the variable data is described through the change index.

[0127] In this embodiment, the cross-correlation function is used to measure the correlation between two signals under different time lags. The cross-correlation function between the control measure signal and the variable data signal is calculated, yielding a series of correlation coefficients under different time lags. The maximum value in the cross-correlation function is found; the time lag corresponding to this value is the preliminary estimated time lag. It should be noted that the cross-correlation function may have multiple peaks; it is necessary to determine which peak represents the true lag time based on the actual situation. The control measure is treated as a step input, and the response curve of the variable data is observed. The response curve typically includes characteristics such as rise time, peak time, and settling time. The time point at which the variable data begins to change significantly is measured from the start of the measure's implementation. Significant change can be defined as a data change exceeding a certain threshold or a rate of change exceeding a certain value. Based on the time characteristics of all curves, the second time lag value of the control node is determined. The target time lag value is determined by combining the first and second time lag values ​​of the control node, using the following specific calculation formula:

[0128] ;

[0129] in, For the time lag target value, , These are the lag weights for the first and second time lag values, respectively. , These are the first time lag value and the second time lag value, respectively. for , The larger of the two, As a preset constant, This indicates the correction for the average of the two values ​​based on the larger value. This is to balance the size of the correction function and more accurately combine the two methods to determine the time delay.

[0130] In this embodiment, the variability description parameters include three categories: amplitude, frequency, and trend.

[0131] Amplitude analysis:

[0132] Calculate the amount of change: Calculate the amount of change in variable data after the implementation of control measures, such as the difference between the maximum and minimum values, the change in the average value, etc.

[0133] Statistical distribution: Analyze the statistical distribution of changes, such as mean, standard deviation, quantiles, etc., to understand the overall situation of the changes.

[0134] Frequency analysis:

[0135] Frequency of fluctuations: The number of times variable data fluctuates within a unit of time to understand the frequency of changes.

[0136] Periodicity analysis: Using methods such as Fourier transform and wavelet analysis to extract the periodic components of variable data in order to identify whether there are periodic changes.

[0137] Trend Analysis:

[0138] Long-term trend: Analyze the trend of variable data over a long period of time, such as rising, falling, or remaining stable.

[0139] Short-term fluctuations: Pay attention to the fluctuations of variable data in a short period of time to understand the immediate impact of regulatory measures on the data.

[0140] A change indicator is determined by combining the three categories of amplitude, frequency, and trend.

[0141] In some embodiments of this application, the variable data threshold for each control node is set by variability, including:

[0142] Determine the basic threshold for each variable data at each control node, map a safety margin based on the change index of each variable data, and superimpose the safety margin onto the basic threshold to form the threshold range.

[0143] In this embodiment, the basic threshold for variable data can be determined by historical data, process requirements, safety redundancy, etc. The larger the change index, the larger the safety margin. The safety margin is superimposed on the basic threshold, and the upper and lower limits of the basic threshold are adjusted to form the threshold range.

[0144] In some embodiments of this application, the generation of evaluation metrics includes,

[0145] By comparing the actual values ​​of variable data with the threshold range of variable data, the deviation is obtained. The control response time is statistically analyzed, and evaluation indicators are generated based on the deviation and control response time.

[0146] In this embodiment, deviation amount: the degree of deviation between the actual value and the threshold range (e.g., dissolved oxygen 2.6 mg / L, deviation +0.1 mg / L). Control response time: the time from the control action to the variable entering the threshold range (e.g., dissolved oxygen stabilizes 12 minutes after aeration adjustment). Evaluation indicators are generated based on the deviation amount and control response time of each variable data (selected from all variable data in this step). The specific formula is as follows:

[0147] ;

[0148] in, For the first Evaluation indicators for each control node For the first The number of variable data that can be used for evaluation at each control node. For the first Evaluation weights for each variable data, For the first The first control node The deviation of the data for each variable For the first The first control node The adjustment response time of individual variable data For the first The second constant of the data of each variable. This indicates the correction of the deviation amount by the control response time. It is used to describe the evaluation index, which is the sum of the deviations of all variables that can be used for evaluation in the control process. Since the control process involves multiple variables, they are combined for evaluation.

[0149] Step S104: Based on the evaluation indicators, the control strategy of each control node is controlled according to the correlation between variable data and the control time delay, so as to ensure the real-time and efficient nature of sewage treatment.

[0150] In this embodiment, the current treatment effect is analyzed through evaluation indicators, which can be specific to a certain variable. A proportional-integral-derivative (PID) control strategy can be used to dynamically adjust the aeration rate based on the DO deviation: rapid adjustment when the deviation is large (proportional P), cumulative adjustment when the deviation is sustained (integral I), and predictive adjustment when the rate of change of the deviation is large (derivative D). Example: If the DO is lower than the set value of 0.5 mg / L, the PID output increases the aeration rate by 10%. Fuzzy control is used, fuzzifying the variable into "high / medium / low" and generating instructions based on a rule base: Rule: If DO is low and COD is high, increase the aeration rate.

[0151] Correspondingly, this application also provides a wastewater energy-saving control system based on multidimensional variable data analysis, such as... Figure 2 As shown, including,

[0152] The first module is used to acquire the wastewater treatment process, divide the wastewater treatment process into multiple treatment stages, deploy monitoring nodes and control nodes in the treatment stages, and determine all variables involved in each treatment stage.

[0153] The second module is used to acquire historical variable data of the wastewater treatment process and analyze the correlation between variable data at different treatment stages.

[0154] The third module is used to analyze historical variable data to determine the corresponding control delay and variability at each control node, and to set the variable data threshold for each control node based on the variability, thereby generating evaluation indicators.

[0155] The fourth module is used to control the control strategy of each control node based on the correlation between variable data and the control time delay, based on the evaluation indicators, so as to ensure the real-time and efficient nature of sewage treatment.

[0156] Compared with the prior art, the beneficial effects of this invention are as follows:

[0157] 1. Analyze the correlations between variable data at different treatment stages. These correlations include the relationship between independent and dependent variables, as well as the interaction relationships among independent variables. This provides a comprehensive and accurate description of the complex coupling relationships between multiple variables, offering a reliable foundation for subsequent control. Determine the control time delay and variability at each control node, considering the control lag and variable changes in the wastewater treatment process. This allows for the setting of thresholds and the generation of evaluation indicators to guide subsequent wastewater control.

[0158] 2. By controlling the control strategy of each control node based on the correlation between variable data and the control time delay, the adaptability and accuracy of wastewater node analysis and control are improved, ensuring the real-time treatment effect of wastewater and improving wastewater treatment efficiency.

[0159] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented in hardware or by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0160] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing the present invention.

[0161] Those skilled in the art will understand that the modules in the system of the implementation scenario can be distributed throughout the system of the implementation scenario as described, or they can be modified to reside in one or more systems different from this implementation scenario. The modules of the above-mentioned implementation scenario can be merged into one module, or they can be further divided into multiple sub-modules.

[0162] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A wastewater energy-saving control method based on multidimensional variable data analysis, characterized in that, include, The wastewater treatment process is obtained, divided into multiple treatment stages, and monitoring and control nodes are deployed at each treatment stage to determine all variables involved in each treatment stage. Acquire historical variable data of the wastewater treatment process and analyze the correlation between variable data at different treatment stages; Historical variable data is analyzed to determine the control delay and variability at each control node. The variability is used to set the variable data threshold for each control node and generate evaluation indicators. Based on the evaluation indicators, the control strategy of each control node is controlled according to the correlation between variable data and the control time delay, so as to ensure the real-time and high efficiency of sewage treatment. in, Analyze the relationships between variable data at different processing stages, including: Based on the variable data between different processing stages, independent and dependent variables are distinguished, and the matching relationship between independent and dependent variables is established. The association between independent and dependent variables is described by a multinomial regression model between independent and dependent variables. The model parameters in the multinomial regression model are divided into fixed model parameters and variable model parameters; Fixed model parameters include polynomial order and model interaction terms; Calculate the degree of nonlinearity between independent and dependent variables and the noise level of each independent and dependent variable under each matching relationship. Determine the overall noise level of the matching relationship between independent and dependent variables based on the noise levels of each independent and dependent variable. Combine the degree of nonlinearity and the overall noise level to determine the complexity. Map a polynomial order through the complexity. The combinations of independent variables are determined based on the matching relationship between independent and dependent variables. Each combination of independent variables is detected by the response surface methodology, and the combinations of independent variables are selected as model interaction terms. The initialization of variable model parameters is set according to the independent variable-dependent variable, and the multinomial regression model is trained based on the initialized variable model parameters and the order of the multinomial. The initialization of variable model parameters is based on the independent-dependent variable ratio. A multinomial regression model is then trained based on the initialized variable model parameters and the multinomial order, including... Variable model parameters include independent variable coefficients and interaction term coefficients; Based on the matching relationship between independent and dependent variables, the correlation between each independent variable and the dependent variable is calculated, and the correlation between independent variables in the model interaction term is calculated. The correlation between independent and dependent variables is recorded as the first correlation, and the correlation between independent variables in the model interaction term is recorded as the second correlation. The initial values ​​of the coefficients of each independent variable are determined based on the first correlation, and the initial values ​​of the coefficients of each interaction term are determined based on the second correlation. Obtain the data for the independent and dependent variables, and divide the data into training and test sets based on their timestamps; The model structure is defined based on the polynomial order, model interaction terms, initial values ​​of independent variable coefficients and interaction term coefficients. Based on the training and test sets, the initial values ​​of independent variable coefficients and interaction term coefficients are continuously optimized through the loss function until the loss function converges, thus obtaining the final polynomial regression model. Analyzing historical variable data to determine the timing and variability of regulation at each control point, including: The historical variable data of the control node is a record of historical variables, which records the changes in variables over time and control information. Mark the multiple control positions on the historical variable record according to the control information, calculate the cross-correlation function between the control information and the variable data, take the time node corresponding to the maximum value of the cross-correlation function as the first time lag value of a single control position, and combine the multiple control positions to obtain the average value of the first time lag value, and take the average value as the first time lag value of the control node. Multiple control positions and control information are used as step inputs, and response curves of variable data are plotted. The curve time characteristics at the points of significant change in variable data are determined on the response curves, and the second time lag value of the control node is determined based on the curve time characteristics. The time lag target value is determined by combining the first and second time lag values ​​of the control node; The variability description parameters for each variable data under the statistical control node are integrated to determine the change index of each variable data, and the variability of the variable data is described through the change index.

2. The wastewater energy-saving control method based on multidimensional variable data analysis according to claim 1, characterized in that, The wastewater treatment process includes pretreatment, primary treatment, secondary treatment, advanced treatment, and sludge disposal. Each treatment stage also includes multiple treatment units.

3. The wastewater energy-saving control method based on multidimensional variable data analysis according to claim 2, characterized in that, During the processing phase, monitoring nodes and control nodes are deployed, including: Determine the processing flow and processing requirements of all processing units in each processing stage, and determine the location of monitoring nodes and control nodes in the processing flow based on the processing flow and processing requirements. Monitoring nodes and control nodes are deployed at the locations of monitoring nodes and control nodes, respectively. Monitoring nodes are used to collect and monitor variable data at the corresponding locations in the treatment process, while control nodes are used to control and adjust the wastewater treatment strategy at the corresponding locations in the treatment process.

4. The wastewater energy-saving control method based on multidimensional variable data analysis according to claim 1, characterized in that, The variable data threshold for each control node is set by using variability. include, Determine the basic threshold for each variable data at each control node, map a safety margin based on the change index of each variable data, and superimpose the safety margin onto the basic threshold to form the threshold range.

5. The wastewater energy-saving control method based on multidimensional variable data analysis according to claim 4, characterized in that, Generate evaluation indicators, including: By comparing the actual values ​​of variable data with the threshold range of variable data, the deviation is obtained. The control response time is statistically analyzed, and evaluation indicators are generated based on the deviation and control response time.

6. A wastewater energy-saving control system based on multidimensional variable data analysis, characterized in that, The system is used to implement the wastewater energy-saving control method based on multidimensional variable data analysis as described in any one of claims 1-5, the system comprising: The first module is used to acquire the wastewater treatment process, divide the wastewater treatment process into multiple treatment stages, deploy monitoring nodes and control nodes in the treatment stages, and determine all variables involved in each treatment stage. The second module is used to acquire historical variable data of the wastewater treatment process and analyze the correlation between variable data at different treatment stages. The third module is used to analyze historical variable data to determine the corresponding control delay and variability at each control node, and to set the variable data threshold for each control node based on the variability, thereby generating evaluation indicators. The fourth module is used to control the control strategy of each control node based on the correlation between variable data and the control time delay, based on the evaluation indicators, so as to ensure the real-time and efficient nature of sewage treatment.

Images