A sewage phosphorus removal control system

Abstract

The present invention provides a sewage phosphorus removal control system, which relates to the technical field of sewage treatment. The sewage phosphorus removal control system includes: sewage treatment equipment, detection equipment and a control host; the sewage treatment equipment is used to perform sewage phosphorus removal treatment according to target control parameters and corresponding target parameter values; the detection equipment is used to detect sewage data parameters before and after the sewage treatment equipment performs sewage phosphorus removal treatment, and send the sewage data parameters to the control host; wherein the sewage data parameters include inlet parameters, outlet parameters, process parameters and control parameters; the control host stores a target causal prediction model. The sewage phosphorus removal control system provided by the present invention can effectively reduce the error interference caused by long-term series records to causal correlation analysis, accurately identify the contribution of key parameters in the sewage treatment process, and provide decision support for equipment control for industry personnel to reduce the phosphorus content in effluent water.

Description

Technical Field

[0001] The invention relates to the technical field of sewage treatment, and in particular to a sewage phosphorus removal control system. Background Art

[0002] In the sewage treatment process, accurate monitoring and control of multiple devices are crucial to ensure that the phosphorus content of the treated water meets the national emission standards and can be safely discharged into the river. However, due to the complexity and multi-link nature of the treatment process, the interactions between control devices are often complicated. Quickly and accurately identifying the cause-and-effect relationship between different parameters in the treatment process is the key to overcoming such decision-making and control problems. In order to reduce the phosphorus content of the treated water, it is necessary to adjust each control device by analyzing the cause-and-effect chain on the basis of maintaining the overall stability of the system to achieve treatment optimization.

[0003] The existing causal discovery of parameters in sewage treatment processes mainly relies on the work experience of industry personnel, but this method lacks certain standardization and it is difficult to adjust equipment to improve water quality. In addition, when abnormal events occur in the water inlet status, this method is not friendly to industry personnel with less work experience. Summary of the invention

[0004] In view of the shortcomings of the prior art mentioned above, the purpose of the present invention is to provide a sewage phosphorus removal control system to solve the problem that the causal discovery of parameters in the sewage treatment process in the prior art mainly relies on the work experience of industry personnel, while this method lacks certain standardization and it is difficult to adjust the equipment to improve water quality; in addition, when abnormal events occur in the water inlet state, this method is not friendly to industry personnel with less work experience.

[0005] To achieve the above-mentioned purpose and other related purposes, the present invention provides a wastewater phosphorus removal control system, including: wastewater treatment equipment, detection equipment and a control host; the wastewater treatment equipment is used to perform wastewater phosphorus removal treatment according to target control parameters and corresponding target parameter values; the detection equipment is used to detect wastewater data parameters before and after the wastewater treatment equipment performs wastewater phosphorus removal treatment, and send the wastewater data parameters to the control host; wherein the wastewater data parameters include inlet parameters, outlet parameters, process parameters and control parameters; the control host stores a target causal prediction model, and the control host can use the target causal prediction model to perform real-time state prediction, obtain a real-time state causal graph, and perform sensitivity and contribution degree on the control parameters in the real-time state causal graph. Analysis is performed to obtain target control parameters and corresponding target parameter values, and the sewage treatment equipment is controlled by the target control parameters and the corresponding target parameter values; the control host also includes a model building unit for building a target causal prediction model, and the model building unit includes: a data acquisition unit for acquiring historical sewage data parameters; an extraction unit for performing time extraction on the input parameters in the historical sewage data parameters to obtain event sequence fragments in the order of occurrence time; an association unit for associating the event sequence fragments with the outlet parameters, process parameters and control parameters to obtain an event causal graph; and a training unit for performing graph neural network training according to the mapping relationship between the event causal graph and time to obtain a target causal prediction model.

[0006] In one embodiment of the present invention, the sewage treatment equipment includes: a sewage pool, including an inlet end, an outlet end and a return pipeline, the inlet end and the outlet end are connected by the return pipeline, and when the control host detects that the real-time outlet parameters are not up to standard, the return pipeline is controlled to connect the inlet end and the outlet end, so that the sewage is discharged into the inlet end through the outlet end to perform sewage circulation treatment; at least one dosing bin, used to store dosing raw materials; at least one metering pump, the sewage pool is connected to the dosing bin through the metering pump, and is used to control the dosing raw materials to be transported to the sewage pool according to the dosing parameters corresponding to the target control parameters; and at least one oxidation ditch, used to treat the sewage in the corresponding stage of the oxidation ditch according to the aerator frequency and / or propulsion equipment frequency corresponding to the target control parameters.

[0007] In one embodiment of the present invention, the detection equipment includes: COD detection equipment for detecting chemical oxygen demand data of sewage; BOD detection equipment for detecting biological oxygen demand data of sewage; ammonia nitrogen detection equipment for detecting ammonia nitrogen content data of sewage; total nitrogen detection equipment for detecting total nitrogen content data of sewage; total phosphorus detection equipment for detecting total phosphorus content data of sewage; pH value detection equipment for detecting pH value data of sewage; dissolved oxygen detection equipment for detecting dissolved oxygen content data of sewage; backflow detection equipment for detecting backflow data of sewage; equipment operation status detection equipment for detecting equipment operation status data of sewage treatment equipment; chemical agent liquid level detection equipment for detecting chemical agent liquid level of each dosing bin of sewage treatment equipment; and sewage liquid level detection equipment for detecting the number of sewage liquid levels in sewage The control host is also used to calculate the water flow data of the sewage according to the sewage level data. The sewage data parameters include the inlet parameters corresponding to the water inlet of the sewage treatment equipment, the outlet parameters corresponding to the water outlet of the sewage treatment equipment, the process parameters and the control parameters; the inlet parameters include the first water flow data, the first chemical oxygen demand data, the first biological oxygen demand data, the first ammonia nitrogen content data, the first total nitrogen content data, the first total phosphorus content data and the first pH value data; the outlet parameters include the second water flow data, the second chemical oxygen demand data, the second biological oxygen demand data, the second ammonia nitrogen content data, the second total nitrogen content data, the second total phosphorus content data and the second pH value data; the process parameters include the dissolved oxygen data, the reflux data, the equipment operation status data and the control parameters; the control parameters include the dosing parameters, the aerator frequency and the propulsion equipment frequency.

[0008] In one embodiment of the present invention, the data acquisition unit includes: a data set acquisition subunit, which is used to acquire the data set of the entire process of the urban sewage treatment system; and a preprocessing subunit, which is used to preprocess the data set by data cleaning, normalization and data classification to obtain historical sewage data parameters.

[0009] In one embodiment of the present invention, the extraction unit includes: an event extraction subunit at a time point, which is used to obtain a standard score of each inlet parameter according to the deviation of the data point of each inlet parameter under the event corresponding to the time point from the homogeneity; according to the standard score of each inlet parameter, a water inlet state point-state combination of the data points of all inlet parameters is obtained; an event extraction subunit for a time period, which is used to obtain the self-changing rate of each inlet parameter according to the first-order derivative change of each inlet parameter under the event corresponding to the time period; according to the self-changing rate of each inlet parameter, a water inlet state time period combination of all inlet parameters is obtained; a state combination subunit, which is used to obtain a comprehensive state according to the water inlet state point-state combination and the water inlet state time period combination; a threshold detection subunit, which is used to perform threshold detection on the comprehensive state; and a segment division subunit, which is used to record the time corresponding to the comprehensive state as the starting time of the target event when the comprehensive state is greater than the set threshold, and the starting time of the next event as the ending time of the target event; and the time segment composed of the starting time and the ending time of the target event as the event sequence segment.

[0010] In one embodiment of the present invention, the association unit includes: a parameter combination subunit, which is used to sequentially form parameter combination pairs among event sequence fragments, exit parameters, process parameters and control parameters; a correlation calculation subunit, which is used to calculate the correlation of each parameter combination pair according to the formed parameter combination pairs; a first generation subunit, which is used to generate an undirected acyclic graph with different edge weights according to the correlation of each parameter combination pair; a pairwise modeling subunit, which is used to perform pairwise modeling on the parameter combination pairs to obtain the causal orientation between the parameter combination pairs; and a second generation subunit, which is used to generate a directed acyclic graph as an event causal graph according to the causal orientation and the undirected acyclic graph.

[0011] In one embodiment of the present invention, the training unit includes: a causal conversion subunit, which is used to convert the event causal graph corresponding to the event into a causal adjacency matrix; a relationship creation subunit, which is used to establish a mapping relationship between the event and the causal features; wherein the causal features include the causal adjacency matrix, the water inflow status and the timestamp status; and a model creation subunit, which is used to input the causal features for establishing the mapping relationship into the spatial domain graph neural network, and perform time series feature learning on the causal adjacency matrix, the water inflow status and the timestamp status through the attention module to obtain a causal prediction model.

[0012] In one embodiment of the present invention, the control host also includes: a sensitivity analysis unit, in which the user calculates the sewage data parameters directly affected by each control parameter and the sensitivity to the total phosphorus in the effluent according to the causal chain in the real-time causal diagram through a sensitivity coefficient analysis method; a parameter screening unit, which is used to obtain several control parameters with the greatest impact on the effluent quality among the control parameters with strong causal association according to the sewage data parameters and the sensitivity to the total phosphorus in the effluent as target control parameters; a contribution calculation unit, which is used to calculate the contribution degree of the effluent phosphorus content corresponding to each target control parameter according to the causal chain in the real-time causal diagram; a parameter sorting unit, which is used to obtain the control priority ranking of the target control parameter for the sewage treatment equipment according to the contribution degree corresponding to each target control parameter; and a testing unit, which is used to adjust and test the original parameter value of the target control parameter to obtain the test result, so as to obtain the target parameter value corresponding to the target control parameter according to the test result.

[0013] In one embodiment of the present invention, the test unit includes: a first test subunit, used to obtain the first effluent total phosphorus in the real-time outlet parameter after sewage phosphorus removal treatment according to the original parameter value of the target control parameter; a second test subunit, used to obtain the second effluent total phosphorus in the real-time outlet parameter after sewage phosphorus removal treatment according to the adjusted parameter value of the target control parameter; a comparison subunit, used to compare the first effluent total phosphorus with the second effluent total phosphorus; a first output subunit, used to, when the first effluent total phosphorus is less than the second effluent total phosphorus, the original parameter value of the target control parameter is valid, and the original parameter value of the target control parameter is used as the target parameter value corresponding to the target control parameter; and a second output subunit, used to, when the first effluent total phosphorus is greater than the second effluent total phosphorus, the adjusted parameter value of the target control parameter is valid, and the adjusted parameter value of the target control parameter is used as the target parameter value corresponding to the target control parameter.

[0014] As described above, a sewage phosphorus removal control system of the present invention has the following beneficial effects: by establishing an event causal prediction model according to the causal relationship between sewage monitoring data and based on past data to generate a real-time causal graph, the error interference caused by long-term series records to causal correlation analysis is reduced, thereby accurately identifying the contribution of key parameters in the sewage treatment process, providing industry personnel with decision support for equipment control, and reducing the phosphorus content in the effluent. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a structural block diagram of a sewage phosphorus removal control system provided in an embodiment of the present invention.

[0016] Figure 2 Shown is a schematic diagram of a directed acyclic graph provided by an embodiment of the present invention.

[0017] Figure 3Shown is a flow chart of the causal modeling process for phosphorus removal from wastewater provided by one embodiment of the present invention.

[0018] Component number description

[0019] Sewage treatment equipment 10; detection equipment 20; control host 30. DETAILED DESCRIPTION

[0020] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0021] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0022] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0023] See also Figure 1 The present invention provides a wastewater phosphorus removal control system, comprising: a wastewater treatment device 10, a detection device 20 and a control host 30; the wastewater treatment device 10 is used to perform wastewater phosphorus removal treatment according to target control parameters and corresponding target parameter values; the detection device 20 is used to detect wastewater data parameters before and after the wastewater treatment device 10 performs wastewater phosphorus removal treatment, and send the wastewater data parameters to the control host 30; wherein the wastewater data parameters include inlet parameters, outlet parameters, process parameters and control parameters; the control host 30 stores a target causal prediction model, and the control host 30 can use the target causal prediction model to perform real-time state prediction, obtain a real-time state causal graph, perform sensitivity and contribution degree analysis on the control parameters in the real-time state causal graph, obtain target control parameters and corresponding target parameter values, and control the wastewater treatment device 10 by the target control parameters and the corresponding target parameter values.

[0024] It is not difficult to find from the above content that in the sewage treatment process, accurate monitoring and control of the sewage treatment equipment 10 is crucial. Specifically, the sewage treatment equipment 10 will perform phosphorus removal treatment on the sewage according to the corresponding target control parameters and the corresponding target parameter values. And in the process of sewage treatment using the target control parameters and the corresponding target parameter values, the detection device 20 will also detect the sewage before and after the sewage phosphorus removal treatment to obtain the sewage data parameters, and send the detected sewage data parameters to the control host 30. When the detection device 20 performs sewage detection, it will monitor the water inlet and outlet separately, and then obtain the corresponding inlet parameters and outlet parameters in the sewage data parameters; at the same time, the detection device 20 will also detect the process parameters and control parameters in the sewage treatment process. The target causal prediction model is pre-stored in the control host 30, and real-time state prediction can be performed, thereby predicting the real-time state causal graph. After obtaining the real-time state causal graph, the control parameters corresponding to the real-time state causal graph will be used to perform sensitivity and contribution analysis to determine the target control parameters and the corresponding target parameter values. After obtaining the target control parameter and the corresponding target parameter value, the control host 30 controls the sewage treatment equipment 10 according to the target control parameter and the corresponding target parameter value, and further detects the corresponding real-time export parameters in the sewage data parameters after the sewage phosphorus removal treatment through the detection equipment. Through the above method, it is possible to achieve accurate decision support for the sewage treatment equipment 10 during the sewage treatment process to reduce the phosphorus content index in the effluent.

[0025] The sewage treatment equipment 10 includes: a sewage pool, including an inlet end, an outlet end and a return pipeline, the inlet end and the outlet end are connected by the return pipeline, and when the control host 30 detects that the real-time outlet parameters are not up to standard, the return pipeline is controlled to connect the inlet end and the outlet end, so that the sewage is discharged into the inlet end through the outlet end, and the sewage is circulated; at least one dosing bin, used to store dosing raw materials; at least one metering pump, the sewage pool is connected to the dosing bin through the metering pump, and is used to control the dosing raw materials to be transported to the sewage pool according to the dosing parameters corresponding to the target control parameters; and at least one oxidation ditch, used to treat the sewage in the corresponding stage of the oxidation ditch according to the aerator frequency and / or propulsion equipment frequency corresponding to the target control parameters.

[0026] In one embodiment of the present invention, during the process of phosphorus removal treatment of sewage by the sewage treatment equipment 10, the dosing raw materials in the dosing bin can be added to the sewage pool by controlling the use of a metering pump to perform phosphorus removal treatment of the sewage. When the metering pump adds the dosing raw materials in the dosing bin to the sewage pool, the metering pump performs dosing control through the corresponding dosing parameters in the target control parameters. At the same time, the oxidation ditch can also be regulated and controlled through the corresponding aerator frequency and / or propulsion equipment frequency in the target control parameters to adjust the phosphorus removal treatment effect of the sewage in the oxidation ditch stage. Specifically, there can be one dosing bin, or of course there can be multiple dosing bins. When there are multiple dosing bins, the outlets of each dosing bin are interconnected. The dosing raw materials can be accurately added to the sewage pool through the metering pump, and the metering pump can be a diaphragm pump, a plunger pump, a gear pump, etc.

[0027] The detection equipment 20 includes: COD detection equipment for detecting chemical oxygen demand data of sewage; BOD detection equipment for detecting biological oxygen demand data of sewage; ammonia nitrogen detection equipment for detecting ammonia nitrogen content data of sewage; total nitrogen detection equipment for detecting total nitrogen content data of sewage; total phosphorus detection equipment for detecting total phosphorus content data of sewage; pH value detection equipment for detecting pH value data of sewage; dissolved oxygen detection equipment for detecting dissolved oxygen data of sewage; backflow detection equipment for detecting backflow data of sewage; equipment operation status detection equipment for detecting equipment operation status data of sewage treatment equipment 10; chemical agent liquid level detection equipment for detecting chemical agent liquid level of each dosing bin of sewage treatment equipment 10; and sewage liquid level detection equipment for detecting sewage liquid level data of sewage; wherein, the control The control host 30 is also used to calculate the water flow data of the sewage according to the sewage level data. The sewage data parameters include inlet parameters corresponding to the water inlet of the sewage treatment equipment 10, outlet parameters corresponding to the water outlet of the sewage treatment equipment 10, process parameters and control parameters; the inlet parameters include first water flow data, first chemical oxygen demand data, first biological oxygen demand data, first ammonia nitrogen content data, first total nitrogen content data, first total phosphorus content data and first pH value data; the outlet parameters include second water flow data, second chemical oxygen demand data, second biological oxygen demand data, second ammonia nitrogen content data, second total nitrogen content data, second total phosphorus content data and second pH value data; the process parameters include dissolved oxygen data, reflux data, equipment operation status data and control parameters; the control parameters include dosing parameters, aerator frequency and propulsion equipment frequency.

[0028] COD testing equipment is mainly used to detect the chemical oxygen demand data of sewage. Among them, the chemical oxygen demand data (COD) refers to the oxygen equivalent consumed by the reducing inorganic and organic matter (generally organic matter) in the water sample that needs to be oxidized by chemical reaction. It is a very important indicator for judging whether the water environment is polluted. The COD testing equipment can be a COD online automatic detector.

[0029] BOD detection equipment is mainly used to detect the biological oxygen demand data of sewage. The biological oxygen demand data (BOD) refers to the dissolved oxygen content consumed by the process of organic matter in water being decomposed by microorganisms. It can be used to characterize the degree of organic pollution in water. The higher the value, the more serious the organic pollution in the water. The BOD detection equipment can be a BOD online automatic detector.

[0030] Ammonia nitrogen detection equipment is mainly used to detect the ammonia nitrogen content data of sewage, where ammonia nitrogen content data (NH) refers to the sum of all forms of ammonia nitrogen in sewage. Ammonia nitrogen detection equipment can be an online automatic ammonia nitrogen analyzer.

[0031] Total nitrogen detection equipment is mainly used to detect the total nitrogen content of sewage, where the total nitrogen content (TN) refers to the sum of all forms of nitrogen in sewage, including ammonium nitrogen, nitrite nitrogen, nitrate nitrogen and organic nitrogen. The total nitrogen detection equipment can be an online automatic total nitrogen detector.

[0032] The total phosphorus detection equipment is mainly used to detect the total phosphorus content data of sewage. The total phosphorus data (TP) refers to all forms of phosphorus in sewage, such as dissolved phosphorus (DP) and particulate phosphorus (P). The total phosphorus detection equipment can be an online automatic total phosphorus detector.

[0033] The pH value detection equipment is mainly used to detect the pH value data of sewage, where the pH value refers to the negative logarithm of the hydrogen ion concentration in sewage, which is used to quantify the acidity and alkalinity of sewage. The pH value detection equipment can be an online automatic pH value detector.

[0034] Dissolved oxygen detection equipment is mainly used to detect the dissolved oxygen data of sewage. Dissolved oxygen refers to the concentration of oxygen molecules in sewage, usually expressed in milligrams per liter (mg / L). It is one of the important indicators for evaluating water quality and the health of aquatic organisms. The dissolved oxygen detection equipment can be an online automatic dissolved oxygen detector.

[0035] The backflow detection equipment is mainly used to detect the backflow data of sewage, where the backflow refers to the water returned from the subsequent treatment equipment to the pre-treatment equipment during the sewage treatment process. The backflow detection equipment can be an online automatic backflow detector.

[0036] The equipment operation status detection device is mainly used to detect the equipment operation status data of the sewage treatment equipment 10, wherein the equipment operation status refers to the state and operation time of the equipment during normal operation, including the opening, closing, operation, and stopping of the equipment. The equipment operation status detection device can be an online automatic detector for the equipment operation status.

[0037] Sewage level detection equipment is mainly used to detect sewage level data of sewage, where sewage level refers to the water level of sewage in the sewage pool. Sewage level detection equipment can be ultrasonic level gauge, float level gauge, pressure level gauge, etc.

[0038] The control host 30 also includes a model building unit for building a target causal prediction model, and the model building unit includes: a data acquisition unit for acquiring historical sewage data parameters; an extraction unit for performing time extraction on the input parameters in the historical sewage data parameters to obtain event sequence fragments in the order of occurrence time; an association unit for associating the event sequence fragments with the outlet parameters, process parameters and control parameters to obtain an event causal graph; and a training unit for performing graph neural network training according to the mapping relationship between the event causal graph and time to obtain the target causal prediction model.

[0039] In one embodiment of the present invention, the target causal prediction model can be constructed by a model construction unit. In the process of constructing the target causal prediction model, the inlet parameters in the historical sewage data parameters obtained by the data acquisition unit are then extracted by the extraction unit to obtain event sequence fragments in the order of occurrence time. Subsequently, the association unit further establishes an event causal graph using the event sequence fragments and other historical sewage data parameters, such as outlet parameters, process parameters and control parameters. The training unit then uses a graph neural network to learn the graph structure based on the mapping relationship between the event causal graph and time to form a causal prediction model, so as to generate a real-time causal graph through the causal prediction model. Then, the sensitivity and contribution degree of the control parameters can be analyzed according to the factual causal graph to determine the target control parameters and the corresponding target parameter values. And after obtaining the target control parameters and the corresponding target parameter values, the target control parameters and the corresponding target parameter values ​​are used by the control host 30 to control the sewage treatment equipment 10, and the corresponding real-time outlet parameters in the sewage data parameters after the sewage phosphorus removal treatment are further detected by the detection equipment. The above method can effectively reduce the error interference caused by long-term series records to causal correlation analysis, accurately identify the contribution of sewage data parameters to phosphorus removal effects during sewage treatment, provide decision-making support for equipment control for industry personnel, and reduce the phosphorus content in effluent water.

[0040] The data acquisition unit includes: a data set acquisition subunit, which is used to acquire the data set of the entire process of the urban sewage treatment system; and a preprocessing subunit, which is used to preprocess the data set by data cleaning, normalization and data classification to obtain historical sewage data parameters.

[0041] In this embodiment, in the process of acquiring historical sewage data parameters through the data acquisition unit, the data set of the entire process of the urban sewage treatment system is first acquired through the data set acquisition subunit, and the data set includes the chemical oxygen demand at different stages. , Biological Oxygen Demand , water flow , ammonia nitrogen content , total nitrogen content , total phosphorus content , pH value , dissolved oxygen , Reflux , chemical agent level and equipment operation time, etc. The different stages include the water inlet stage when entering the sewage treatment equipment 10 and the water outlet stage when the sewage treatment equipment 10 is discharged. In the process of preprocessing the data set, the preprocessing subunit needs to perform data cleaning, normalization and data classification on the data set. When performing data cleaning on the data set, since the values ​​often have periodic characteristics in the sewage treatment process, it is necessary to first perform periodic decomposition on the time series.

[0042] In the process of periodic decomposition of time series, the missing data are filled using periodic characteristics, that is, the formula is: ,in, , is the periodic characteristic of the time series, is a single parameter in the dataset, represents the formula for finding the average value; A series of time series data representing a certain attribute, a single parameter in the data set can be replaced by the above, for example, chemical oxygen demand In , Biological Oxygen Demand In When cleaning the data set, Indicates that a parameter data is missing. Indicates the time point at which missing data occurred.

[0043] After cleaning the data set, the data needs to be normalized. That is to say, the units and numerical fluctuation ranges of various parameters in the sewage treatment process vary greatly. In order to reduce the interference of data dimensions on causal discovery, all parameters are normalized to the maximum and minimum values. The formula is expressed as , where the maximum and minimum values ​​are the values ​​after removing abnormal values ​​of the equipment.

[0044] Furthermore, after normalizing the data set, the preprocessing subunit also needs to classify the data to obtain historical sewage data parameters. Specifically, for the inflow flow of the sewage plant inflow data , Chemical Oxygen Demand , Biological Oxygen Demand and ammonia nitrogen content The parameters are recorded as inlet parameters, and the chemical oxygen demand of the effluent data is , Biological Oxygen Demand , ammonia nitrogen content , total nitrogen content and total phosphorus content Parameters such as the outlet parameters, dissolved oxygen , Chemical liquid level Parameters such as the operating status of the equipment are recorded as process parameters, and the controllable object values ​​in the process parameters are recorded as control parameters.

[0045] In one embodiment of the present invention, after obtaining the historical sewage data parameters through the data acquisition unit, the inlet parameters need to be further processed by the extraction unit to extract the time of each inlet parameter of the sewage, and extract the event sequence fragments from the long-term time series by combining the value and fluctuation of the time series. ,in, They are different events and are sorted in the order of their occurrence. , is the start time, For the end time, For events All parameters are in to Specifically, the value of the event sequence refers to the parameter value of the original collected data of the input parameter; fluctuation refers to the change trend of the value such as rising, falling, and stable. In the process of extracting event sequence fragments, for example, time, A sharp rise (an event represented as an attribute), Slowly rising, rise, Rise, this moment happened Event (represented as the overall event at a certain moment). Similarly, It might have happened event, It might have happened Events can constitute A fragment of the event sequence (where represents continuous time). Event It is based on the combined influence of the original data and fluctuations of an event. For example, a time series value rising from 50 to 100 and rising from 100 to 150 are represented as two different events.

[0046] Specifically, the extraction unit includes: an event extraction subunit at a time point, which is used to obtain a standard score for each inlet parameter according to the deviation of the data point of each inlet parameter under the event corresponding to the time point from the homogeneity; according to the standard score of each inlet parameter, a water inlet state point-state combination of the data points of all inlet parameters is obtained; an event extraction subunit for a time period, which is used to obtain the self-changing rate of each inlet parameter according to the first-order derivative change of each inlet parameter under the event corresponding to the time period; according to the self-changing rate of each inlet parameter, a water inlet state time period combination of all inlet parameters is obtained; a state combination subunit, which is used to obtain a comprehensive state according to the water inlet state point-state combination and the water inlet state time period combination; a threshold detection subunit, which is used to perform threshold detection on the comprehensive state; and a segment division subunit, which is used to record the moment corresponding to the comprehensive state as the starting moment of the target event when the comprehensive state is greater than the set threshold, and the starting moment of the next event as the ending moment of the target event; and the time segment composed of the starting moment and the ending moment of the target event as the event sequence segment.

[0047] In one embodiment of the present invention, when the extraction unit extracts the time of the input parameter, the event extraction subunit of the time point first extracts the event of the time point. Specifically, by calculating each parameter data point The deviation from the mean is used to obtain the standard score, and then the state combination of the water inlet state point of the data points of all inlet parameters is obtained according to the standard score of each inlet parameter. ,in, , is a user-customizable weight, It is a single parameter, and its value is For a specific category Time, parameter data points are The data mean is , users can set the parameters for each entry Set the weights, The standard score is Since the entry has multiple acquisition parameters, the point state combination For all standard scores The sum of .

[0048] Next, the event extraction subunit of the time period is used to extract the events of the time period. Specifically, the rate of change is calculated by the first-order derivative change of a single parameter. , and then according to the rate of change , get the water inlet state time period combination of all inlet parameters ,in, .

[0049] Then, the state combination subunit combines the time period and time point characteristics to determine the segmentation points between event segments and obtain the comprehensive state . And the comprehensive state is detected by the threshold detection subunit Perform threshold detection, when the comprehensive state When it is greater than the set threshold, the time corresponding to the comprehensive state can be divided into sub-units through the fragment Record the start time of the target event , and the starting time of the next event End time of the target event .

[0050] In one embodiment of the present invention, the correlation unit may be configured to detect a separate event for each water inflow state. , by cutting its corresponding starting time and end time The event sequence fragments between them are used to model the association of sewage process parameters according to the event sequence fragments, thereby constructing an event causal diagram.

[0051] Specifically, the association unit includes: a parameter combination subunit, which is used to sequentially form parameter combination pairs among event sequence fragments, exit parameters, process parameters and control parameters; a correlation calculation subunit, which is used to calculate the correlation of each parameter combination pair according to the formed parameter combination pairs; a first generation subunit, which is used to generate an undirected acyclic graph with different edge weights according to the correlation of each parameter combination pair; a pairwise modeling subunit, which is used to perform pairwise modeling on the parameter combination pairs to obtain the causal orientation between the parameter combination pairs; and a second generation subunit, which is used to generate a directed acyclic graph as an event causal graph according to the causal orientation and the undirected acyclic graph.

[0052] In this embodiment, the parameter combination subunit can sequentially form parameter combination pairs between the event sequence fragments, the outlet parameters, the process parameters and the control parameters, and the correlation calculation subunit uses Pearson to calculate the correlation of each parameter combination pair composed of the event sequence fragments of the input parameters in the historical sewage data parameters and the outlet parameters, the process parameters and the control parameters, so that the first generation subunit forms an undirected acyclic graph with different edge weights according to the correlation. Each node represents a parameter, and the edge represents the correlation between parameters. The weight of the edge can be determined according to the size of the correlation coefficient to indicate the strength of the correlation between parameters. When the absolute value of the correlation is less than the correlation threshold, the parameter and parameters Independent of each other, denoted by Among them, the parameters and parameters are different parameters in the sewage data, which are used to calculate the correlation between any two different parameters; is a node in an undirected acyclic graph, After obtaining the undirected acyclic graph, the Granger detection method can be used to perform pairwise modeling on the parameter combination pairs through the pairwise modeling subunit, and the causal relationship between the parameter combination pairs can be calculated to obtain the causal orientation between the parameter combination pairs. Afterwards, a directed acyclic graph is further generated as an event causal graph through the second generating subunit.

[0053] In one embodiment of the present invention, in the process of establishing a causal prediction model, a training unit includes establishing a mapping between a graph of a time and event causal graph, and a graph neural network is used to train and learn the graph structure, thereby forming a causal prediction model.

[0054] Specifically, the training unit includes: a causal conversion subunit, which is used to convert the event causal graph corresponding to the event into a causal adjacency matrix; a relationship creation subunit, which is used to establish a mapping relationship between events and causal features; wherein the causal features include the causal adjacency matrix, water inflow status and timestamp status; and a model creation subunit, which is used to input the causal features for establishing the mapping relationship into the spatial domain graph neural network, and perform time series feature learning on the causal adjacency matrix, water inflow status and timestamp status through the attention module to obtain a causal prediction model.

[0055] In one embodiment of the present invention, in the Event Causal Diagram When performing causal adjacency matrix transformation, the causal transformation subunit can be used to transform the Event Causal Diagram Convert to causal adjacency matrix ,in, Represents a parameter node and The strength of the correlation between When . After obtaining the causal adjacency matrix After that, create subunits through relationships according to the water inflow status and timestamp status , create an event and causal characteristics The mapping relationship between them. Among them, the causal feature Contains the causal adjacency matrix , Water inflow status and timestamp status , i.e., causal characteristics . Water inflow status , timestamp status , mainly including events Time of occurrence External characteristics, such as temperature ,precipitation and humidity Etc. Then, by creating sub-units through relationships, the causal characteristics of the mapping relationship will be established. Input into the spatial domain graph neural network of the model creation subunit, and the causal adjacency matrix , Water inflow status and timestamp status An attention module is assigned to each of them for temporal feature learning. The attention weights corresponding to the three attention modules are different to form an event causal prediction model, and the model parameters can be adjusted according to the effect of the model test set.

[0056] Specifically, during the processing, the training unit normalizes the causal features of each event. , calculation parameters Neighbor aggregation information ,in, For parameters Parameters with strong causal relationships; From the start time and end time Parameters within time period By updating the parameters Features ,in, Compared with its original information Neighbor aggregation information The weight of are the parameters that the model needs to train. is the original information of the parameter itself, For parameters Neighbor node information (that is, neighbor aggregation information), is the activation function, is the sum of the features that contain the information of itself and its neighbors after the update. Further, the features of each node are updated through loop processing. The final causal adjacency matrix Updated to , Contains each node New network features, event features Updated to .

[0057] The control host 30 also includes: a sensitivity analysis unit, in which the user calculates the sewage data parameters directly affected by each control parameter and the sensitivity to the total phosphorus in the effluent according to the causal chain in the real-time causal diagram through the sensitivity coefficient analysis method; a parameter screening unit, which is used to obtain several control parameters with the greatest impact on the effluent quality among the control parameters with strong causal association according to the sewage data parameters and the sensitivity to the total phosphorus in the effluent as target control parameters; a contribution calculation unit, which is used to calculate the contribution degree of the effluent phosphorus content corresponding to each target control parameter according to the causal chain in the real-time causal diagram; a parameter sorting unit, which is used to obtain the control priority ranking of the target control parameter for the sewage treatment equipment 10 according to the contribution degree corresponding to each target control parameter; and a testing unit, which is used to adjust and test the original parameter value of the target control parameter to obtain the test result, so as to obtain the target parameter value corresponding to the target control parameter according to the test result.

[0058] In this embodiment, according to the real-time cause-effect diagram, the sensitivity and contribution of the control parameters to the phosphorus removal effect are calculated respectively, among which the main control parameters are temperature. , frequency of propulsion equipment , frequency of aeration equipment , sodium acetate addition and polyaluminium chloride addition etc. The key effluent state is total phosphorus In combination with the causal chain in the event causal diagram, the sensitivity coefficient analysis method is used to calculate the sewage data parameters under the direct effect of each control parameter and the effect on the total phosphorus in the effluent. The influence fluctuation is sensitivity, and the three control parameters with the greatest impact on effluent quality are determined from the control parameters with strong causal association (calculated based on the influence fluctuation of effluent total phosphorus) as the target control parameters. Then, combined with the causal chain, the influence of the three control parameters on effluent phosphorus content and other related parameters is calculated to quantify the contribution of each control parameter with the greatest impact on effluent quality to wastewater phosphorus removal. , so that the parameter sorting unit can obtain the control priority of the target control parameters for the sewage treatment equipment 10 according to the contribution degree, so as to achieve the maximum water quality phosphorus removal management without affecting other process parameters. is a highly sensitive control parameter. Indicates other standards involved in the assessment of effluent water quality. Indicates the process parameter in the processing flow. Finally, the test unit uses the original parameter value of the target control parameter to perform an adjustment test to obtain a test result, and based on the test result, further obtains the target parameter value corresponding to the target control parameter.

[0059] See also Figure 2 , Figure 2 In one embodiment, in the generated directed acyclic graph, assuming that A is the control parameter, the next sewage data parameter directly affected is parameter B and parameter C. The influence of each control parameter on the next sewage data parameter directly affected is calculated, that is, when the amount of A is increased, the degree of change of B and C is calculated. The fluctuation of the influence on the total phosphorus in the effluent is calculated, that is, when the amount of A is increased, the total phosphorus in the effluent is calculated. degree of change.

[0060] Generally, the event characteristics of historical sewage data parameters are used as the input of the event causal prediction model for learning, and the event causal prediction model further performs node link prediction to generate real-time status The real-time causal graph corresponds to the water inflow status in the real-time data. and timestamp status Known.

[0061] After calculating the sensitivity and contribution of the control parameters to the phosphorus removal effect and obtaining the target control parameters, the control host 30 can intervene in the cause-and-effect diagram and link the decision-making to formulate strategies to adjust the target parameter values ​​of the key control parameters and optimize the effluent quality according to the results of the sensitivity analysis and contribution evaluation.

[0062] Specifically, the test unit includes: a first test subunit, which is used to obtain the first effluent total phosphorus in the real-time outlet parameters after sewage phosphorus removal treatment according to the original parameter value of the target control parameter; a second test subunit, which is used to obtain the second effluent total phosphorus in the real-time outlet parameters after sewage phosphorus removal treatment according to the adjusted parameter value of the target control parameter; a comparison subunit, which is used to compare the first effluent total phosphorus with the second effluent total phosphorus; a first output subunit, which is used to determine that when the first effluent total phosphorus is less than the second effluent total phosphorus, the original parameter value of the target control parameter is valid, and the original parameter value of the target control parameter is used as the target parameter value corresponding to the target control parameter; and a second output subunit, which is used to determine that when the first effluent total phosphorus is greater than the second effluent total phosphorus, the adjusted parameter value of the target control parameter is valid, and the adjusted parameter value of the target control parameter is used as the target parameter value corresponding to the target control parameter.

[0063] In one embodiment of the present invention, the first effluent total phosphorus in the real-time outlet parameter after the wastewater phosphorus removal treatment is obtained by the first test subunit according to the original parameter value of the target control parameter. ; Then, the second test subunit adjusts the parameter value of the target control parameter to obtain the second effluent total phosphorus in the real-time outlet parameter after the sewage phosphorus removal treatment. . Compare the total phosphorus of the first effluent by comparing the subunits and the second effluent total phosphorus The numerical value of can reflect the specific situation of phosphorus removal effect. When , it means that the original parameter value of the target control parameter is valid, and the original parameter value of the target control parameter is used as the target parameter value corresponding to the target control parameter; and when , it indicates that the adjustment parameter value of the target control parameter is valid, and the adjustment parameter value of the target control parameter is used as the target parameter value corresponding to the target control parameter.

[0064] Next, after determining the adjustment parameter value of the target control parameter, the original water quality data, event characteristics The test control scheme is recorded and saved. If the scheme corresponding to the adjustment parameter value of the target control parameter is valid, the control device is changed. If the scheme corresponding to the adjustment parameter value of the target control parameter is invalid, the cause is further analyzed through the real-time event cause-effect diagram, and the test scheme is adjusted to obtain the adjustment parameter value corresponding to the target control parameter that meets the test validity as the target parameter value corresponding to the target control parameter.

[0065] See also Figure 3 , Figure 3 In one embodiment given, the whole process data set of the urban sewage treatment system is obtained and pre-processed to obtain historical sewage data parameters. Then, the water inflow state event sequence in the time series is extracted from the historical sewage data parameters, and the causal relationship between the parameters is captured in combination with the data association state, and an event causal graph is established based on the causal relationship between the parameters. Through the event causal graph, a causal adjacency matrix is ​​established, and then other features of the event are used to form an event causal prediction model, so that a real-time causal graph can be obtained based on the event causal prediction model. In addition, the target control parameters can be obtained based on the attribute sensitivity and contribution, and the original parameter values ​​of the target control parameters are adjusted to pass the adjusted target control parameters and target parameter values. And the control effect of the sewage treatment equipment 10 is optimized through the target control parameters and target parameter values.

[0066] In summary, the present invention discloses a causal modeling method, system, equipment and medium for phosphorus removal in sewage. By establishing an event causal prediction model based on the causal relationship between sewage monitoring data and based on past data, a real-time causal graph is generated, thereby reducing the error interference caused by long-term series records on causal association analysis, thereby accurately identifying the contribution of key parameters in the sewage treatment process, providing decision support for equipment control for industry personnel, and reducing the phosphorus content in effluent water. Therefore, the present invention effectively overcomes various shortcomings in the prior art and has a high industrial utilization value.

[0067] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A sewage phosphorus removal control system, characterized in that: include: Sewage treatment equipment, testing equipment and control host; The sewage treatment equipment is used to perform sewage phosphorus removal treatment according to the target control parameters and the corresponding target parameter values; The detection device is used to detect the sewage data parameters before and after the sewage treatment equipment performs the phosphorus removal treatment on the sewage, and send the sewage data parameters to the control host; wherein the sewage data parameters include inlet parameters, outlet parameters, process parameters and control parameters; The control host stores a target causal prediction model, and the control host can use the target causal prediction model to perform real-time state prediction, obtain a real-time state causal graph, perform sensitivity and contribution degree analysis on the control parameters in the real-time state causal graph, obtain the target control parameters and corresponding target parameter values, and control the sewage treatment equipment through the target control parameters and the corresponding target parameter values; The control host further includes a model building unit for building the target causal prediction model, and the model building unit includes: A data acquisition unit, used to acquire historical sewage data parameters; An extraction unit, used for performing time extraction on the inlet parameters in the historical sewage data parameters to obtain event sequence fragments in the order of occurrence time; an associating unit, configured to associate the event sequence fragment with the exit parameter, the process parameter and the control parameter to obtain an event cause-effect graph; and A training unit, used to perform graph neural network training according to the mapping relationship between the event causal graph and time, to obtain the target causal prediction model; The extraction unit comprises: The event extraction subunit at a time point is used to obtain a standard score of each inlet parameter according to the deviation of the data point of each inlet parameter under the event corresponding to the time point from the homogeneity; and obtain a state combination of water inlet state points of all the data points of the inlet parameters according to the standard score of each inlet parameter; The event extraction subunit of the time period is used to obtain the self-changing rate of each inlet parameter according to the first-order derivative change of each inlet parameter under the event corresponding to the time period; and obtain the water inlet state time period combination of all the inlet parameters according to the self-changing rate of each inlet parameter; A state combination subunit, used to obtain a comprehensive state according to the state combination of the water inlet state point and the water inlet state time period combination; a threshold detection subunit, configured to perform threshold detection on the comprehensive state; and The segment division subunit is used to record the moment corresponding to the comprehensive state as the starting moment of the target event and the starting moment of the next event as the ending moment of the target event when the comprehensive state is greater than a set threshold; and to record the time segment composed of the starting moment and the ending moment of the target event as the event sequence segment.

2. The sewage phosphorus removal control system according to claim 1 is characterized in that: The sewage treatment equipment comprises: A sewage pool, comprising an inlet end, an outlet end and a return pipeline, wherein the inlet end and the outlet end are connected via the return pipeline, and when the control host detects that the real-time outlet parameter does not meet the standard, the return pipeline is controlled to connect the inlet end and the outlet end, so that the sewage is discharged into the inlet end through the outlet end, so as to perform a circulation treatment of the sewage; At least one dosing bin for storing dosing raw materials; at least one metering pump, the sewage pool is connected to the dosing bin through the metering pump, and is used to control and deliver the dosing raw material to the sewage pool according to the dosing parameter corresponding to the target control parameter; and At least one oxidation ditch is used to treat sewage at a corresponding stage of the oxidation ditch according to the aerator frequency and / or propulsion equipment frequency corresponding to the target control parameters.

3. The sewage phosphorus removal control system according to claim 1 is characterized in that: The detection equipment comprises: COD testing equipment, used to test the chemical oxygen demand data of sewage; BOD testing equipment, used to test the biological oxygen demand data of sewage; Ammonia nitrogen detection equipment, used to detect the ammonia nitrogen content data of sewage; Total nitrogen detection equipment, used to detect the total nitrogen content of sewage; Total phosphorus testing equipment, used to test the total phosphorus content of sewage; pH value detection equipment, used to detect the pH value data of sewage; Dissolved oxygen detection equipment, used to detect the dissolved oxygen data of sewage; Backflow detection equipment, used to detect the backflow data of sewage; Equipment operation status detection equipment, used to detect equipment operation status data of the sewage treatment equipment; Chemical agent liquid level detection equipment, used to detect the chemical agent liquid level of each dosing tank of the sewage treatment equipment; and Sewage level detection equipment, used to detect sewage level data of sewage; Wherein, the control host is also used to calculate the water flow data of the sewage according to the sewage liquid level data, and the sewage data parameters include inlet parameters corresponding to the water inlet of the sewage treatment equipment, outlet parameters corresponding to the water outlet of the sewage treatment equipment, process parameters and control parameters; the inlet parameters include first water flow data, first chemical oxygen demand data, first biological oxygen demand data, first ammonia nitrogen content data, first total nitrogen content data, first total phosphorus content data and first pH value data; the outlet parameters include second water flow data, second chemical oxygen demand data, second biological oxygen demand data, second ammonia nitrogen content data, second total nitrogen content data, second total phosphorus content data and second pH value data; the process parameters include dissolved oxygen data, reflux data, equipment operation status data and control parameters; the control parameters include dosing parameters, aerator frequency and propulsion equipment frequency.

4. The sewage phosphorus removal control system according to claim 1, characterized in that: The data acquisition unit comprises: A data set acquisition subunit, used to acquire data sets of the entire process of the urban sewage treatment system; and The preprocessing subunit is used to preprocess the data set by data cleaning, normalization and data classification to obtain historical sewage data parameters.

5. The sewage phosphorus removal control system according to claim 1 is characterized in that: The association unit comprises: A parameter combination subunit, used for sequentially combining the event sequence fragment, the exit parameter, the process parameter and the control parameter into a parameter combination pair; A correlation calculation subunit, used for calculating the correlation of each parameter combination pair according to the composed parameter combination pairs; A first generating subunit, used for generating an undirected acyclic graph with different edge weights according to the correlation of each parameter combination pair; a pairwise modeling subunit, configured to perform pairwise modeling on the parameter combination pairs to obtain a causal orientation between the parameter combination pairs; and The second generating subunit is used to generate a directed acyclic graph as the event causal graph according to the causal orientation and the undirected acyclic graph.

6. The sewage phosphorus removal control system according to claim 1 is characterized by: The training unit comprises: A causal conversion subunit, used to convert the event causal graph corresponding to the event into a causal adjacency matrix; a relationship creation subunit, used to establish a mapping relationship between the event and the causal characteristics; wherein the causal characteristics include a causal adjacency matrix, a water inflow state, and a timestamp state; and The model creation subunit is used to input the causal features that establish the mapping relationship into the spatial domain graph neural network, and through the attention module, perform time series feature learning on the causal adjacency matrix, the water inflow state and the timestamp state respectively to obtain the causal prediction model.

7. The sewage phosphorus removal control system according to claim 1, characterized in that: The control host also includes: Sensitivity analysis unit, the user calculates the sewage data parameters directly affected by each control parameter and the sensitivity to the effluent total phosphorus through the sensitivity coefficient analysis method according to the causal chain in the real-time causal diagram; A parameter screening unit, for obtaining, according to the sewage data parameters and the sensitivity to the effluent total phosphorus, several control parameters with the greatest impact on the effluent quality among the control parameters with strong causal association as target control parameters; A contribution calculation unit, used to calculate the contribution degree of the effluent phosphorus content corresponding to each of the target control parameters according to the cause-effect chain in the real-time cause-effect diagram; a parameter sorting unit, configured to obtain a control priority ranking of the target control parameter on the sewage treatment equipment according to the contribution degree corresponding to each target control parameter; and The testing unit is used to perform an adjustment test on the original parameter value of the target control parameter to obtain a test result, so as to obtain a target parameter value corresponding to the target control parameter according to the test result.

8. The sewage phosphorus removal control system according to claim 7, characterized in that: The test unit comprises: A first test subunit is used to obtain the first effluent total phosphorus in the real-time outlet parameter after the sewage phosphorus removal treatment according to the original parameter value of the target control parameter; A second testing subunit is used to obtain the second effluent total phosphorus in the real-time outlet parameter after the wastewater phosphorus removal treatment according to the adjustment parameter value of the target control parameter; a comparison subunit, configured to compare the total phosphorus in the first effluent water with the total phosphorus in the second effluent water; a first output subunit, configured to, when the total phosphorus in the first effluent is less than the total phosphorus in the second effluent, the original parameter value of the target control parameter is valid, and use the original parameter value of the target control parameter as the target parameter value corresponding to the target control parameter; and The second output subunit is used for, when the total phosphorus in the first effluent is greater than the total phosphorus in the second effluent, the adjustment parameter value of the target control parameter is valid, and the adjustment parameter value of the target control parameter is used as the target parameter value corresponding to the target control parameter.

Images