An intelligent sewage environmental protection treatment regulation method and system

Abstract

The present application relates to sewage treatment and industrial intelligent control technical field, disclose a kind of sewage environmental protection treatment intelligent regulation and control method and system, including full-dimension multi-source data sensing acquisition module, data standardization pretreatment and working condition calibration module, water quality and quantity short-time prediction and working condition pre-judgment module, intelligent optimization module under multiple constraint conditions Process parameters, distributed real-time execution and dynamic correction module, regulation and control effect closed-loop evaluation and model self-iteration module, comprising the following steps: S1, full-dimension data acquisition stage;S2, data pretreatment and working condition calibration stage;S3, water trend prediction and working condition pre-judgment stage;S4, intelligent optimization stage of process parameters;S5, distributed execution and dynamic correction stage;S6, closed-loop evaluation and model iteration stage;The present application realizes the full-process intelligent management and control of sewage treatment process from water load early prediction to regulation and control execution real-time correction by constructing the closed-loop intelligent regulation and control architecture of whole process.

Description

Technical Field

[0001] This invention relates to the field of wastewater treatment and industrial intelligent control technology, and in particular to an intelligent control method and system for environmentally friendly wastewater treatment. Background Technology

[0002] Urban wastewater treatment is a core component of water environment protection and pollution control. With the continuous improvement of my country's ecological and environmental protection requirements, the effluent discharge standards of urban wastewater treatment plants are becoming increasingly stringent. At the same time, higher requirements are being placed on the energy conservation, consumption reduction, operational stability, and intelligent management and control of the wastewater treatment process.

[0003] Currently, the operation and control of municipal wastewater treatment plants mostly adopts a fixed-value regulation mode based on human experience, supplemented by conventional single-parameter automated control methods. Operation and maintenance personnel need to manually adjust the operating parameters of aeration, reflux, chemical dosing and other links according to the real-time monitoring of effluent water quality.

[0004] With the increasing complexity of wastewater treatment processes and the diversification of influent conditions, traditional control models are no longer suitable for the characteristics of wastewater treatment processes, which are characterized by multiple variables, strong coupling, and large lags. They also cannot meet the needs of refined and intelligent operation and management. Therefore, developing an intelligent control solution for wastewater environmental protection treatment that can achieve closed-loop management of the entire process has become an important development direction in the industry. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent control method and system for wastewater environmental treatment to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A smart control system for wastewater environmental protection treatment includes a multi-dimensional, multi-source data sensing and acquisition module, a data standardization preprocessing and operating condition calibration module, a short-term prediction and operating condition forecasting module for influent water quality and quantity, a smart optimization module for process parameters under multiple constraints, a distributed real-time execution and dynamic correction module, and a closed-loop evaluation and model self-iteration module for control effect. The output of the multi-dimensional, multi-source data sensing and acquisition module is connected to the input of the data standardization preprocessing and operating condition calibration module. The output of the data standardization preprocessing and operating condition calibration module is connected to both the short-term prediction and operating condition forecasting module for influent water quality and quantity and the closed-loop evaluation and model self-iteration module for control effect. The input end of the influent water quality and quantity short-term prediction and operating condition prediction module is connected to the input end of the intelligent optimization module for process parameters under multiple constraints. The output end of the intelligent optimization module for process parameters under multiple constraints is connected to the input end of the distributed real-time execution and dynamic correction module. The output end of the distributed real-time execution and dynamic correction module is connected to the input end of the full-dimensional multi-source data sensing and acquisition module and the control effect closed-loop evaluation and model self-iteration module, respectively. The output end of the control effect closed-loop evaluation and model self-iteration module is connected to the input end of the influent water quality and quantity short-term prediction and operating condition prediction module and the intelligent optimization module for process parameters under multiple constraints, respectively.

[0008] As a further improvement to this technical solution: the multi-dimensional multi-source data sensing and acquisition module includes an inlet sensing device, a pretreatment section sensing device, a biochemical section multi-corridor sensing device, a deep treatment section sensing device, an effluent sensing device, an environmental data docking device, and an equipment status acquisition device. The inlet sensing device is installed at the inlet of the wastewater treatment plant, the pretreatment section sensing device is installed in the bar screen and grit chamber unit, the biochemical section multi-corridor sensing device is installed in each process compartment of the anaerobic tank and anoxic tank, each process corridor of the aerobic tank, and the corresponding monitoring point of the secondary sedimentation tank, the deep treatment section sensing device is installed in the deep treatment unit, the effluent sensing device is installed at the wastewater treatment plant discharge outlet, the environmental data docking device establishes a data communication link with the municipal meteorological system, and the equipment status acquisition device establishes a data connection with the actuators such as fans, pumps, valves, and agitators in the entire process section.

[0009] As a further improvement to this technical solution: the data standardization preprocessing and operating condition calibration module includes a data cleaning submodule, a data standardization processing submodule, an operating condition calibration submodule, and a local time series database. The input end of the data cleaning submodule receives the collected data uploaded by the multi-dimensional multi-source data sensing and acquisition module. The output end of the data cleaning submodule is connected to the input end of the data standardization processing submodule. The output end of the data standardization processing submodule is connected to the input ends of the operating condition calibration submodule and the local time series database, respectively. The output end of the operating condition calibration submodule is connected to the input end of the local time series database.

[0010] As a further improvement to this technical solution: the short-term prediction and operating condition judgment module for influent water quality and quantity includes an influent timing prediction submodule, an operating condition judgment submodule, and a pre-regulation strategy matching submodule. The input end of the influent timing prediction submodule receives standardized data output from the data standardization preprocessing and operating condition calibration module. The output end of the influent timing prediction submodule is connected to the input end of the operating condition judgment submodule. The output end of the operating condition judgment submodule is connected to the input end of the pre-regulation strategy matching submodule. The output end of the pre-regulation strategy matching submodule is connected to the input end of the intelligent optimization module for process parameters under multiple constraints.

[0011] As a further improvement to this technical solution: the intelligent optimization module for process parameters under multiple constraints includes a multi-constraint target setting submodule, a deep reinforcement learning optimization submodule, an expert rule verification submodule, and a control command generation submodule. The output of the multi-constraint target setting submodule is connected to the input of the deep reinforcement learning optimization submodule. The input of the deep reinforcement learning optimization submodule simultaneously receives prediction data and operating condition data output by the short-term prediction and operating condition prediction module for influent water quality and quantity. The output of the deep reinforcement learning optimization submodule is connected to the input of the expert rule verification submodule. The output of the expert rule verification submodule is connected to the input of the control command generation submodule. The output of the control command generation submodule is connected to the input of the distributed real-time execution and dynamic correction module. The deep reinforcement learning optimization submodule adopts a deep reinforcement learning architecture based on DQN to construct a process parameter optimization model under multiple constraints. The input of the model is influent prediction data, real-time operating condition data, and current operating parameters. The output of the model is the optimal control parameter set for the entire process section. The optimization objective function of the model is: In the formula To comprehensively optimize the target value, The weighting coefficient for electricity consumption per ton of water. This is the weighting coefficient for the amount of medicine consumed per ton of water. The electricity consumption per unit volume of water treated by a wastewater treatment plant. The chemical consumption per unit volume of wastewater treated by the wastewater treatment plant is optimized to meet rigid constraints. The constraints are that all indicators of effluent quality meet the Class A standard requirements of the pollutant discharge standard for urban wastewater treatment plants, and the operating parameters of all equipment are within the rated operating range of the equipment.

[0012] As a further improvement to this technical solution: the distributed real-time execution and dynamic correction module includes an edge computing gateway, a distributed PLC control unit, a dynamic correction control submodule, and an anomaly alarm submodule. The input of the edge computing gateway receives control commands issued by the intelligent optimization module for process parameters under multiple constraints. The output of the edge computing gateway is connected to the input of the distributed PLC control unit, and the output of the distributed PLC control unit is connected to the field actuator. The input of the dynamic correction control submodule receives real-time operating data transmitted back from the multi-dimensional multi-source data sensing and acquisition module. The output of the dynamic correction control submodule is connected to the input of the distributed PLC control unit. The input of the anomaly alarm submodule is connected to the outputs of the edge computing gateway and the distributed PLC control unit, respectively. The dynamic correction control submodule adopts a control architecture combining fuzzy PID and feedforward compensation to construct a real-time closed-loop correction model. The control output formula of the model is: In the formula for Control output value at any time, This is the proportionality coefficient. The integral coefficient is... These are the differential coefficients. for The deviation between the actual value of the controlled parameter and the set target value at any given time. The integral term of the deviation, For the differential term of the deviation, As a feedforward compensation term based on changes in influent parameters, the model collects process parameters after execution in real time, compares them with the set target threshold, and automatically triggers closed-loop correction when the parameter deviation exceeds the preset range, thus completing real-time fine-tuning of equipment operating parameters.

[0013] As a further improvement to this technical solution: the closed-loop evaluation and model self-iteration module for the control effect includes a control effect evaluation submodule, a model incremental learning submodule, a rule base update submodule, and an operation and maintenance report generation submodule. The input end of the control effect evaluation submodule receives the running data returned by the multi-dimensional multi-source data perception and acquisition module and the distributed real-time execution and dynamic correction module, respectively. The output end of the control effect evaluation submodule is connected to the input ends of the model incremental learning submodule and the rule base update submodule, respectively. The output end of the model incremental learning submodule is connected to the input ends of the short-term prediction and operating condition prediction module for influent water quality and quantity and the intelligent optimization module for process parameters under multiple constraints, respectively. The output end of the rule base update submodule is connected to the input end of the data standardization preprocessing and operating condition calibration module. The input end of the operation and maintenance report generation submodule is connected to the output end of the control effect evaluation submodule.

[0014] A smart control method for wastewater environmental protection treatment includes the following steps:

[0015] S1. In the full-dimensional data acquisition stage, the water quality parameters, operating parameters, and equipment status parameters of the entire process link of the sewage treatment plant are collected through the full-dimensional multi-source data sensing and acquisition module. Environmental parameters are obtained by connecting with the municipal meteorological system. Data acquisition is completed and uploaded in real time according to the preset frequency.

[0016] S2. In the data preprocessing and operating condition calibration stage, the data standardization preprocessing and operating condition calibration module completes the removal of outliers and the completion of missing data in the collected data. It also standardizes multi-dimensional data to unify the units of measurement and completes the calibration and classification storage of typical operating conditions based on historical and real-time operating data.

[0017] S3, Influent Water Quality and Quantity Prediction and Operating Condition Prediction Stage: Through the influent water quality and quantity short-term prediction and operating condition prediction module, using standardized processed historical data, real-time collected data, and meteorological data as input, the module completes the prediction of influent water quality and quantity change trends in multiple time dimensions. Based on the prediction results, the module identifies operating conditions, classifies shock risk levels, matches and outputs corresponding pre-control strategies.

[0018] S4. Intelligent optimization stage of process parameters: Through the intelligent optimization module of process parameters under multiple constraints, rigid constraints and optimization objectives are set. With water inflow prediction data, real-time operating data and current operating parameters as input, the intelligent optimization of control parameters of the entire process section is completed through a deep reinforcement learning model. After compliance verification, executable control instructions are generated and issued.

[0019] S5. Distributed execution and dynamic correction stage: Through the distributed real-time execution and dynamic correction module, the control instructions are received and distributed to the corresponding field actuators to complete precise actions. The process parameters after execution are collected in real time, and the deviation is identified by comparing with the set target threshold. The parameters are fine-tuned in real time through the closed-loop correction model. Equipment abnormalities are identified simultaneously and alarms and backup plans are triggered.

[0020] S6. In the closed-loop evaluation and model iteration stage, the control effect closed-loop evaluation and model self-iteration module collects full-process operation data to complete the multi-dimensional evaluation of the control effect. Based on the evaluation results and newly added operation data, the incremental learning and iteration of the prediction model and the optimization model are completed. The operating condition database and expert rule base are updated simultaneously to generate the corresponding cycle of control effect evaluation report.

[0021] Compared with the prior art, the beneficial effects of the present invention are:

[0022] 1. This invention, by constructing a closed-loop intelligent control architecture for the entire process, realizes intelligent control of the entire wastewater treatment process, from the advance prediction of influent load to the real-time correction of control execution. It can effectively adapt to the dynamic changes in influent water quality and quantity, predict and respond to various operating condition fluctuations and impact risks in advance, significantly improve the operational stability of the wastewater treatment system, ensure that the effluent water quality continuously and stably meets environmental emission standards, effectively reduce the environmental compliance risks of wastewater treatment plants, and at the same time get rid of the excessive reliance on manual operation and maintenance experience in the traditional control mode, reduce the control deviations and errors caused by manual operation, and reduce the workload of operation and maintenance personnel.

[0023] 2. This invention, through a multi-parameter global collaborative optimization intelligent optimization model, can achieve refined control of energy and chemical consumption in the sewage treatment process while strictly ensuring that the effluent meets standards. This reduces unnecessary energy and chemical consumption, lowers the daily operating costs of sewage treatment plants, and, relying on the model's autonomous iterative optimization mechanism, can effectively adapt to changes in operating conditions, equipment performance degradation, seasonal environmental fluctuations, etc., during the long-term operation of sewage treatment plants, maintaining stable control accuracy. In addition, the entire process operation data can be completely retained and traced throughout the entire chain, fully meeting the relevant requirements of environmental protection supervision, and possessing good adaptability and promotion and application value.

[0024] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it according to the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings. Specific embodiments of the present invention are given in detail below with reference to the accompanying drawings. Attached Figure Description

[0025] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0026] Figure 1 This is a schematic diagram of the structure of an intelligent control method and system for wastewater environmental protection treatment. Detailed Implementation

[0027] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are for illustrative purposes only and are not intended to limit the scope of the invention. The invention is described more specifically in the following paragraphs by way of example with reference to the accompanying drawings. It should be noted that the drawings are in a very simplified form and use non-precise proportions, and are only used to facilitate and clarify the illustration of the embodiments of the present invention.

[0028] Please see Figure 1In this embodiment of the invention, a smart control system for wastewater environmental protection treatment includes a multi-dimensional, multi-source data sensing and acquisition module, a data standardization preprocessing and operating condition calibration module, a short-term prediction and operating condition forecasting module for influent water quality and quantity, a smart optimization module for process parameters under multiple constraints, a distributed real-time execution and dynamic correction module, and a closed-loop evaluation and model self-iteration module for control effect. The output of the multi-dimensional, multi-source data sensing and acquisition module is connected to the input of the data standardization preprocessing and operating condition calibration module. The output of the data standardization preprocessing and operating condition calibration module is connected to both the short-term prediction and operating condition forecasting module for influent water quality and quantity and the closed-loop evaluation and model self-iteration module for control effect. The input of the self-iteration module and the output of the short-term prediction and operating condition prediction module for influent water quality and quantity are connected to the input of the intelligent optimization module for process parameters under multiple constraints. The output of the intelligent optimization module for process parameters under multiple constraints is connected to the input of the distributed real-time execution and dynamic correction module. The output of the distributed real-time execution and dynamic correction module is connected to the input of the full-dimensional multi-source data sensing and acquisition module and the closed-loop evaluation and model self-iteration module for control effect, respectively. The output of the closed-loop evaluation and model self-iteration module for control effect is connected to the input of the short-term prediction and operating condition prediction module for influent water quality and quantity and the intelligent optimization module for process parameters under multiple constraints, respectively.

[0029] Specifically, the system defines six core modules and the signal connection and data transmission logic between these modules, thus constructing a complete closed-loop control architecture encompassing perception, preprocessing, prediction, optimization, execution, and iterative processes.

[0030] The multi-dimensional, multi-source data sensing and acquisition module includes an inlet sensing device, a pretreatment section sensing device, a biochemical section multi-corridor sensing device, an advanced treatment section sensing device, an effluent sensing device, an environmental data interface device, and an equipment status acquisition device. The inlet sensing device is deployed at the wastewater treatment plant inlet, the pretreatment section sensing device is deployed in the bar screen and grit chamber units, the biochemical section multi-corridor sensing device is deployed in each process compartment of the anaerobic and anoxic tanks, each process corridor of the aerobic tank, and the corresponding monitoring points of the secondary sedimentation tank, the advanced treatment section sensing device is deployed in the advanced treatment unit, the effluent sensing device is deployed at the wastewater treatment plant outlet, the environmental data interface device establishes a data communication link with the municipal meteorological system, and the equipment status acquisition device establishes a data connection with the actuators such as fans, pumps, valves, and agitators throughout the entire process section.

[0031] Specifically, the inlet sensing device collects basic data on the quality and quantity of influent water at the wastewater treatment plant inlet, providing source input basis for system regulation;

[0032] Pretreatment section sensing device: collects the operating status and effluent water quality data of the bar screen and grit chamber pretreatment unit, and monitors the treatment effect of the pretreatment unit;

[0033] Multi-corridor sensing device in the biological treatment section: collects core biological parameters such as water quality, liquid level, and sludge concentration in each process unit of the anaerobic tank, anoxic tank, aerobic tank, and secondary sedimentation tank, providing real-time data support for the regulation and control of the biological treatment section;

[0034] Advanced treatment section sensing device: collects water quality, chemical dosage, and filter operation status data of the advanced treatment unit to monitor the pollutant removal effect of the advanced treatment section;

[0035] Effluent sensing device: Collects effluent water quality data from the sewage treatment plant's discharge outlet, providing core basis for determining effluent compliance and data support for evaluating control effectiveness;

[0036] Environmental data docking device: docks with the municipal meteorological system to acquire environmental data such as ambient temperature, water temperature, and rainfall, providing environmental input data for water inflow prediction and operational condition forecasting;

[0037] Equipment status acquisition device: Collects data on the operating status, operating frequency, opening degree, power, etc. of actuators such as fans, pumps, valves, and agitators throughout the entire process section, providing a basis for equipment control and anomaly identification.

[0038] The data standardization preprocessing and operating condition calibration module includes a data cleaning submodule, a data standardization processing submodule, an operating condition calibration submodule, and a local time series database. The input end of the data cleaning submodule receives the collected data uploaded by the multi-dimensional multi-source data sensing and acquisition module. The output end of the data cleaning submodule is connected to the input end of the data standardization processing submodule. The output end of the data standardization processing submodule is connected to the input ends of the operating condition calibration submodule and the local time series database, respectively. The output end of the operating condition calibration submodule is connected to the input end of the local time series database.

[0039] Specifically, the data cleaning submodule removes outliers and completes missing data in the collected raw data to eliminate data distortion caused by sensor failures and network fluctuations, ensuring the validity of subsequent model input data.

[0040] Data standardization processing submodule: Normalizes and standardizes the cleaned multi-dimensional data, unifies the data units, and eliminates the interference of different parameter values ​​on subsequent model calculations;

[0041] Operating condition calibration submodule: Based on historical operating data and real-time standardized data, it classifies and calibrates typical operating conditions of wastewater treatment plants, providing a benchmark for subsequent operating condition prediction and parameter optimization;

[0042] Local time series database: Stores preprocessed standardized data and calibrated operating condition data, providing local data support for model training and data retrieval.

[0043] The short-term prediction and operating condition judgment module for influent water quality and quantity includes an influent time series prediction submodule, an operating condition judgment submodule, and a pre-control strategy matching submodule. The input end of the influent time series prediction submodule receives standardized data output from the data standardization preprocessing and operating condition calibration module. The output end of the influent time series prediction submodule is connected to the input end of the operating condition judgment submodule. The output end of the operating condition judgment submodule is connected to the input end of the pre-control strategy matching submodule. The output end of the pre-control strategy matching submodule is connected to the input end of the intelligent optimization module for process parameters under multiple constraints.

[0044] Specifically, the influent time series prediction submodule adopts the LSTM-Attention time series prediction architecture. It takes standardized historical operating data, real-time collected data, and meteorological data as inputs and outputs the prediction results of the influent water quality and water quantity change trends in four time dimensions: 2h, 6h, 12h, and 24h, reserving response time for subsequent parameter optimization.

[0045] Operating condition prediction submodule: Based on the water inflow timing prediction results and combined with the preset operating condition calibration threshold, it automatically identifies the operating conditions in future periods, classifies the water inflow load impact risk level, and provides a basis for matching pre-control strategies;

[0046] Pre-regulation strategy matching submodule: Based on the predicted operating conditions and risk levels, it matches the corresponding pre-regulation strategy, triggers pre-regulation actions in advance, and reduces the impact of influent load fluctuations on system operation.

[0047] The intelligent optimization module for process parameters under multiple constraints includes a multi-constraint objective setting submodule, a deep reinforcement learning optimization submodule, an expert rule verification submodule, and a control command generation submodule. The output of the multi-constraint objective setting submodule is connected to the input of the deep reinforcement learning optimization submodule. The input of the deep reinforcement learning optimization submodule simultaneously receives prediction data and operating condition data output from the short-term prediction and operating condition judgment module for influent water quality and quantity. The output of the deep reinforcement learning optimization submodule is connected to the input of the expert rule verification submodule. The output of the expert rule verification submodule is connected to the input of the control command generation submodule. The output of the control command generation submodule is connected to the input of the distributed real-time execution and dynamic correction module. The deep reinforcement learning optimization submodule adopts a deep reinforcement learning architecture based on DQN to construct a process parameter optimization model under multiple constraints. The model's input includes influent prediction data, real-time operating condition data, and current operating parameters. The model's output is the optimal set of control parameters for the entire process section. The model's optimization objective function is: In the formula To comprehensively optimize the target value, The weighting coefficient for electricity consumption per ton of water. This is the weighting coefficient for the amount of medicine consumed per ton of water. The electricity consumption per unit volume of water treated by a wastewater treatment plant. The chemical consumption per unit of treated water volume of the sewage treatment plant is optimized to meet rigid constraints. The constraints are that all indicators of effluent water quality meet the Class A standard requirements of the pollutant discharge standard for urban sewage treatment plants, and the operating parameters of each equipment are within the rated operating range of the equipment.

[0048] Specifically, the multi-constraint target setting submodule sets rigid constraints and optimization targets for system regulation, clarifies the rigid bottom line for effluent compliance, and defines the optimization directions for energy consumption and chemical consumption.

[0049] The deep reinforcement learning optimization submodule adopts a deep reinforcement learning architecture based on DQN to construct a process parameter optimization model under multiple constraints, which is the core of the entire system's control and decision-making. The input dimensions of the model cover influent prediction data, real-time operating data, current operating parameters, ambient temperature and water temperature data, and the output is the optimal control parameter set for the entire process section, including dissolved oxygen setpoint, blower frequency, reflux ratio, and sludge discharge cycle in the biological treatment section, as well as core control parameters such as reagent dosage and filter backwashing cycle in the deep treatment section.

[0050] The supporting optimization objective function serves as the core decision-making basis of the system. The formula aims to minimize the combined electricity and chemical consumption per ton of wastewater treatment plant while ensuring effluent quality meets standards and equipment operates safely, thus achieving a globally optimal solution for the control parameters. Detailed annotations of the formula parameters are provided below. : This indicates that the goal of comprehensive optimization, J, is to find the minimum value, which is the core optimization direction of the entire optimization model; The weighting coefficient of electricity consumption per ton of water can be dynamically adjusted according to the actual operating needs of the sewage treatment plant. The larger the weighting coefficient, the higher the priority of the model for electricity consumption optimization. The weighting coefficient of chemical consumption per ton of water can be dynamically adjusted according to the actual operating needs of the wastewater treatment plant. The larger the weighting coefficient, the higher the priority of the model for chemical consumption optimization. Electricity consumption per unit volume of wastewater treated by a wastewater treatment plant, expressed in kW·h / m³, covering the power consumption of equipment operation throughout the entire process of the wastewater treatment plant; Chemical consumption per unit volume of wastewater treated by a wastewater treatment plant, expressed in kg / m³, covering the consumption of chemicals throughout the entire process, including flocculants and disinfectants.

[0051] Expert rule verification submodule: Performs compliance verification on the parameter set output by the model, removes invalid parameters that exceed the rated operating range of the equipment or violate the basic laws of biochemical reactions, and ensures the executability and safety of control commands;

[0052] Control command generation submodule: Converts the verified optimal parameter set into standardized control commands that can be directly issued to the field actuators, completing the conversion from decision to execution.

[0053] The distributed real-time execution and dynamic correction module includes an edge computing gateway, a distributed PLC control unit, a dynamic correction control submodule, and an anomaly alarm submodule. The edge computing gateway receives control commands from the intelligent optimization module for process parameters under multiple constraints. Its output is connected to the input of the distributed PLC control unit, which in turn connects to the field actuators. The dynamic correction control submodule receives real-time operating data from the multi-dimensional, multi-source data sensing and acquisition module. Its output is connected to the input of the distributed PLC control unit. The anomaly alarm submodule's input is connected to the outputs of both the edge computing gateway and the distributed PLC control unit. The dynamic correction control submodule employs a fuzzy PID control architecture combined with feedforward compensation to construct a real-time closed-loop correction model. The model's control output formula is: In the formula for Control output value at any time, This is the proportionality coefficient. The integral coefficient is... These are the differential coefficients. for The deviation between the actual value of the controlled parameter and the set target value at any given time. The integral term of the deviation, For the differential term of the deviation, As a feedforward compensation term based on changes in influent parameters, the model collects process parameters after execution in real time, compares them with the set target threshold, and automatically triggers closed-loop correction when the parameter deviation exceeds the preset range, thus completing real-time fine-tuning of equipment operating parameters.

[0054] Specifically, the edge computing gateway receives control commands from the cloud, performs local parsing and distribution of the commands, and simultaneously collects and transmits local data in real time to ensure low-latency transmission of control commands.

[0055] Distributed PLC control unit: Receives control commands distributed by the edge computing gateway, directly drives the corresponding actuators on site to complete precise actions, realizes distributed control of the entire process section, and avoids system paralysis caused by single node failure;

[0056] The dynamic correction control submodule employs a control architecture combining fuzzy PID control with feedforward compensation to construct a real-time closed-loop correction model. This ensures accurate execution of control commands and prevents deviation accumulation. The core of the model is the accompanying closed-loop control output formula. This formula combines the real-time deviation of the controlled parameters with proportional, integral, and derivative terms to complete basic closed-loop regulation. Simultaneously, it superimposes a feedforward compensation term based on changes in influent parameters to achieve proactive correction, improving control accuracy and response speed, and ensuring process parameters remain stable within the set target range.

[0057] Detailed annotation of formula parameters: for The control output values ​​at specific times are directly sent to the PLC control unit to adjust the operating parameters of the field actuators, including fan frequency, pump and valve opening, and dosing pump frequency. This is the proportional gain, which functions to quickly respond to the current deviation of the controlled parameter. The larger the deviation, the larger the output amplitude of the proportional control. It is a fundamental adjustment term in closed-loop control. This is the integral coefficient, used to eliminate the static deviation of the controlled parameter by accumulating and integrating historical deviations to ensure that the controlled parameter eventually stabilizes at the set target value. The differential coefficient is used to predict the trend of deviation changes in the controlled parameter, and output adjustment actions in advance based on the rate of change of the deviation, thereby suppressing large fluctuations in the deviation and improving the stability of the system. for The deviation between the actual value and the set target value of the controlled parameter at any given time is the core input of the entire closed-loop control, including the deviation between the actual value and the set value of dissolved oxygen, the deviation between the actual value and the set value of pH, etc. The integral term for the deviation is accumulated and integrated over the historical deviation from time 0 to time t to eliminate static deviation. The differential term of the deviation is used to calculate the rate of change of the deviation at time t, which is then used to predict the trend of deviation changes and achieve proactive adjustment. As a feedforward compensation term based on changes in influent parameters, it outputs compensation and adjustment actions in advance according to real-time changes in influent water quality and quantity, thereby improving the system's resistance to shocks.

[0058] Anomaly Alarm Submodule: Identifies abnormal situations such as equipment failure and command execution failure, automatically triggers backup control schemes, and sends alarms to maintenance personnel to ensure continuous and stable system operation.

[0059] The closed-loop evaluation and model self-iteration module for regulation effect includes a regulation effect evaluation submodule, a model incremental learning submodule, a rule base update submodule, and an operation and maintenance report generation submodule. The input end of the regulation effect evaluation submodule receives the operation data returned by the multi-dimensional multi-source data perception and acquisition module and the distributed real-time execution and dynamic correction module, respectively. The output end of the regulation effect evaluation submodule is connected to the input end of the model incremental learning submodule and the rule base update submodule, respectively. The output end of the model incremental learning submodule is connected to the input end of the short-term prediction and operating condition prediction module for influent water quality and quantity and the intelligent optimization module for process parameters under multiple constraints, respectively. The output end of the rule base update submodule is connected to the input end of the data standardization preprocessing and operating condition calibration module. The input end of the operation and maintenance report generation submodule is connected to the output end of the regulation effect evaluation submodule.

[0060] Specifically, the regulation effect evaluation submodule collects full-process operation data, constructs a multi-dimensional evaluation system, completes the quantitative evaluation of regulation effect, and provides a basis for model iteration and rule base update;

[0061] Model Incremental Learning Submodule: Based on the newly added operating data and the evaluation results of the control effect, the incremental learning algorithm is used to complete the parameter update and self-iteration of the water intake prediction model and the process parameter optimization model, so as to continuously improve the control accuracy and operating condition adaptability of the model.

[0062] Rule base update submodule: Based on the evaluation results and newly added operational data, synchronously update the operating condition calibration database and expert rule base to enrich the system's operating condition coverage and anomaly handling capabilities;

[0063] Operation and maintenance report generation submodule: Based on the control effect evaluation results, it generates control effect evaluation reports and operation and maintenance reports for the corresponding period, providing data support for the operation and maintenance management of wastewater treatment plants.

[0064] A smart control method for wastewater environmental protection treatment includes the following steps:

[0065] S1. In the full-dimensional data acquisition stage, the water quality parameters, operating parameters, and equipment status parameters of the entire process link of the sewage treatment plant are collected through the full-dimensional multi-source data sensing and acquisition module. Environmental parameters are obtained by connecting with the municipal meteorological system. Data acquisition is completed and uploaded in real time according to the preset frequency.

[0066] S2. In the data preprocessing and operating condition calibration stage, the data standardization preprocessing and operating condition calibration module completes the removal of outliers and the completion of missing data in the collected data. It also standardizes multi-dimensional data to unify the units of measurement and completes the calibration and classification storage of typical operating conditions based on historical and real-time operating data.

[0067] S3, Influent Water Quality and Quantity Prediction and Operating Condition Prediction Stage: Through the influent water quality and quantity short-term prediction and operating condition prediction module, using standardized processed historical data, real-time collected data, and meteorological data as input, the module completes the prediction of influent water quality and quantity change trends in multiple time dimensions. Based on the prediction results, the module identifies operating conditions, classifies shock risk levels, matches and outputs corresponding pre-control strategies.

[0068] S4. Intelligent optimization stage of process parameters: Through the intelligent optimization module of process parameters under multiple constraints, rigid constraints and optimization objectives are set. With water inflow prediction data, real-time operating data and current operating parameters as input, the intelligent optimization of control parameters of the entire process section is completed through a deep reinforcement learning model. After compliance verification, executable control instructions are generated and issued.

[0069] S5. Distributed execution and dynamic correction stage: Through the distributed real-time execution and dynamic correction module, the control instructions are received and distributed to the corresponding field actuators to complete precise actions. The process parameters after execution are collected in real time, and the deviation is identified by comparing with the set target threshold. The parameters are fine-tuned in real time through the closed-loop correction model. Equipment abnormalities are identified simultaneously and alarms and backup plans are triggered.

[0070] S6. In the closed-loop evaluation and model iteration stage, the control effect closed-loop evaluation and model self-iteration module collects full-process operation data to complete the multi-dimensional evaluation of the control effect. Based on the evaluation results and newly added operation data, the incremental learning and iteration of the prediction model and the optimization model are completed. The operating condition database and expert rule base are updated simultaneously to generate the control effect evaluation report for the corresponding period.

[0071] Specifically, the method defines six complete steps for intelligent control of wastewater environmental protection treatment, forming a closed-loop control process that corresponds one-to-one with the system architecture. It is fully matched with the technical characteristics of the corresponding system modules, and the steps form a complete temporal logic and closed-loop link, covering the entire process of control actions from data acquisition to model iteration.

[0072] The method of use and working principle of this invention are as follows:

[0073] Usage: First, sensing devices deployed throughout the entire wastewater treatment process collect data related to water quality, operation, and equipment status. Simultaneously, external systems are connected to acquire environmental data, which is then uploaded. Next, the collected raw data undergoes outlier removal, missing data completion, and standardization preprocessing. Based on the processed data, typical operating conditions of the wastewater treatment plant are calibrated and stored. Then, based on preprocessed historical and real-time data, trends in influent water quality and quantity are predicted. Simultaneously, operating conditions and impact risks are identified and assessed based on the prediction results, and corresponding pre-control strategies are matched. Finally, compliance with effluent standards is enforced. The system sets optimization targets based on constraints, combines predicted data with real-time operating conditions to achieve global intelligent optimization of control parameters across the entire process segment, generates executable control commands after compliance verification, and issues them out. These commands are then distributed to various actuators on-site for distributed and precise execution. Simultaneously, it collects process data after execution in real time, compares it with the set targets to identify deviations, and performs real-time closed-loop dynamic correction. Finally, it collects full-process operation data to quantitatively evaluate the control effect, and completes autonomous iterative optimization of the control model based on the evaluation results and newly added operation data. It also updates the operating condition database and rule base simultaneously, forming a complete closed-loop usage process.

[0074] Working Principle: The system acquires basic operational data for the entire wastewater treatment process through a multi-source, end-to-end sensing system. After data preprocessing to eliminate data distortion and dimensional differences, a standardized and effective dataset is formed. A time-series prediction model is used to predict the trends in influent water quality and quantity, reserving a sufficient response window for control decisions. With effluent quality compliance as a rigid constraint, a deep reinforcement learning model is used to achieve global collaborative optimization of multiple control parameters across the entire process, generating the optimal control scheme adapted to current and future operating conditions. Edge computing combined with a distributed control architecture ensures low-latency and precise execution of control commands. Simultaneously, a closed-loop control model with feedforward compensation enables real-time dynamic correction of control deviations, preventing deviation accumulation. Finally, continuous evaluation of the control effect across multiple dimensions and incremental learning enable autonomous iterative updates of the control model, continuously improving the system's adaptability and control accuracy, achieving adaptive, intelligent, and refined closed-loop management of the wastewater treatment process.

[0075] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any way. Those skilled in the art can readily implement the present invention based on the description and drawings above. However, any modifications, alterations, and variations made by those skilled in the art without departing from the scope of the present invention using the disclosed technical content are equivalent embodiments of the present invention. Furthermore, any modifications, alterations, and variations made to the above embodiments based on the essential technology of the present invention are still within the protection scope of the present invention.

Claims

1. A smart control system for wastewater environmental treatment, characterized in that, This includes a multi-dimensional, multi-source data sensing and acquisition module, a data standardization and preprocessing and operating condition calibration module, a short-term prediction and operating condition forecasting module for influent water quality and quantity, a smart optimization module for process parameters under multiple constraints, a distributed real-time execution and dynamic correction module, and a closed-loop evaluation and model self-iteration module for control effect. The output of the multi-dimensional, multi-source data sensing and acquisition module is connected to the input of the data standardization and preprocessing and operating condition calibration module. The output of the data standardization and preprocessing and operating condition calibration module is connected to the inputs of both the short-term prediction and operating condition forecasting module for influent water quality and quantity and the closed-loop evaluation and model self-iteration module for control effect. The output of the short-term water quality and quantity prediction and operating condition prediction module is connected to the input of the intelligent optimization module for process parameters under multiple constraints. The output of the intelligent optimization module for process parameters under multiple constraints is connected to the input of the distributed real-time execution and dynamic correction module. The output of the distributed real-time execution and dynamic correction module is connected to the input of the full-dimensional multi-source data sensing and acquisition module and the control effect closed-loop evaluation and model self-iteration module, respectively. The output of the control effect closed-loop evaluation and model self-iteration module is connected to the input of the short-term influent water quality and quantity prediction and operating condition prediction module and the intelligent optimization module for process parameters under multiple constraints, respectively.

2. The intelligent control system for wastewater environmental treatment according to claim 1, characterized in that, The multi-dimensional, multi-source data sensing and acquisition module includes an inlet sensing device, a pretreatment section sensing device, a biochemical section multi-corridor sensing device, an advanced treatment section sensing device, an effluent sensing device, an environmental data interface device, and an equipment status acquisition device. The inlet sensing device is deployed at the wastewater treatment plant inlet; the pretreatment section sensing device is deployed in the bar screen and grit chamber units; the biochemical section multi-corridor sensing device is deployed in each process compartment of the anaerobic and anoxic tanks, each process corridor of the aerobic tank, and the corresponding monitoring points of the secondary sedimentation tank; the advanced treatment section sensing device is deployed in the advanced treatment unit; the effluent sensing device is deployed at the wastewater treatment plant outlet; the environmental data interface device establishes a data communication link with the municipal meteorological system; and the equipment status acquisition device establishes data connections with actuators such as fans, pumps, valves, and agitators throughout the entire process.

3. The intelligent control system for wastewater environmental treatment according to claim 1, characterized in that, The data standardization preprocessing and operating condition calibration module includes a data cleaning submodule, a data standardization processing submodule, an operating condition calibration submodule, and a local time series database. The input end of the data cleaning submodule receives the collected data uploaded by the multi-dimensional multi-source data sensing and acquisition module. The output end of the data cleaning submodule is connected to the input end of the data standardization processing submodule. The output end of the data standardization processing submodule is connected to the input ends of the operating condition calibration submodule and the local time series database, respectively. The output end of the operating condition calibration submodule is connected to the input end of the local time series database.

4. The intelligent control system for wastewater environmental treatment according to claim 1, characterized in that, The short-term prediction and operating condition judgment module for influent water quality and quantity includes an influent time-series prediction submodule, an operating condition judgment submodule, and a pre-regulation strategy matching submodule. The input end of the influent time-series prediction submodule receives standardized data output from the data standardization preprocessing and operating condition calibration module. The output end of the influent time-series prediction submodule is connected to the input end of the operating condition judgment submodule. The output end of the operating condition judgment submodule is connected to the input end of the pre-regulation strategy matching submodule. The output end of the pre-regulation strategy matching submodule is connected to the input end of the intelligent optimization module for process parameters under multiple constraints.

5. The intelligent control system for wastewater environmental treatment according to claim 1, characterized in that, The intelligent optimization module for process parameters under multiple constraints includes a multi-constraint target setting submodule, a deep reinforcement learning optimization submodule, an expert rule verification submodule, and a control command generation submodule. The output of the multi-constraint target setting submodule is connected to the input of the deep reinforcement learning optimization submodule. The input of the deep reinforcement learning optimization submodule simultaneously receives prediction data and operating condition data output by the short-term prediction and operating condition judgment module for influent water quality and quantity. The output of the deep reinforcement learning optimization submodule is connected to the input of the expert rule verification submodule. The output of the expert rule verification submodule is connected to the input of the control command generation submodule. The output of the control command generation submodule is connected to the input of the distributed real-time execution and dynamic correction module. The deep reinforcement learning optimization submodule adopts a deep reinforcement learning architecture based on DQN to construct a process parameter optimization model under multiple constraints. The input of the model is influent prediction data, real-time operating condition data, and current operating parameters. The output of the model is the optimal control parameter set for the entire process section. The optimization objective function of the model is: In the formula To comprehensively optimize the target value, The weighting coefficient for electricity consumption per ton of water. This is the weighting coefficient for the amount of medicine consumed per ton of water. The electricity consumption per unit volume of water treated by a wastewater treatment plant. The chemical consumption per unit volume of wastewater treated by the wastewater treatment plant is optimized to meet rigid constraints. The constraints are that all indicators of effluent quality meet the Class A standard requirements of the pollutant discharge standard for urban wastewater treatment plants, and the operating parameters of all equipment are within the rated operating range of the equipment.

6. The intelligent control system for wastewater environmental treatment according to claim 1, characterized in that, The distributed real-time execution and dynamic correction module includes an edge computing gateway, a distributed PLC control unit, a dynamic correction control submodule, and an anomaly alarm submodule. The input of the edge computing gateway receives control commands from the intelligent optimization module for process parameters under multiple constraints. The output of the edge computing gateway is connected to the input of the distributed PLC control unit, and the output of the distributed PLC control unit is connected to the field actuator. The input of the dynamic correction control submodule receives real-time operating data from the multi-dimensional multi-source data sensing and acquisition module. The output of the dynamic correction control submodule is connected to the input of the distributed PLC control unit. The input of the anomaly alarm submodule is connected to the outputs of both the edge computing gateway and the distributed PLC control unit. The dynamic correction control submodule adopts a control architecture combining fuzzy PID and feedforward compensation to construct a real-time closed-loop correction model. The control output formula of the model is: In the formula for Control output value at any time, This is the proportionality coefficient. The integral coefficient is... These are the differential coefficients. for The deviation between the actual value of the controlled parameter and the set target value at any given time. The integral term of the deviation, For the differential term of the deviation, As a feedforward compensation term based on changes in influent parameters, the model collects process parameters after execution in real time, compares them with the set target threshold, and automatically triggers closed-loop correction when the parameter deviation exceeds the preset range, thus completing real-time fine-tuning of equipment operating parameters.

7. The intelligent control system for wastewater environmental treatment according to claim 1, characterized in that, The closed-loop evaluation and model self-iteration module for the control effect includes a control effect evaluation submodule, a model incremental learning submodule, a rule base update submodule, and an operation and maintenance report generation submodule. The input end of the control effect evaluation submodule receives the operation data returned by the multi-dimensional multi-source data perception and acquisition module and the distributed real-time execution and dynamic correction module, respectively. The output end of the control effect evaluation submodule is connected to the input ends of the model incremental learning submodule and the rule base update submodule, respectively. The output end of the model incremental learning submodule is connected to the input ends of the short-term prediction and operating condition prediction module for influent water quality and quantity and the intelligent optimization module for process parameters under multiple constraints, respectively. The output end of the rule base update submodule is connected to the input end of the data standardization preprocessing and operating condition calibration module. The input end of the operation and maintenance report generation submodule is connected to the output end of the control effect evaluation submodule.

8. A method for intelligent control of wastewater environmental treatment, applied to the intelligent control system for wastewater environmental treatment as described in any one of claims 1-7, characterized in that, Includes the following steps: S1. In the full-dimensional data acquisition stage, the water quality parameters, operating parameters, and equipment status parameters of the entire process link of the sewage treatment plant are collected through the full-dimensional multi-source data sensing and acquisition module. Environmental parameters are obtained by connecting with the municipal meteorological system. Data acquisition is completed and uploaded in real time according to the preset frequency. S2. In the data preprocessing and operating condition calibration stage, the data standardization preprocessing and operating condition calibration module completes the removal of outliers and the completion of missing data in the collected data. It also standardizes multi-dimensional data to unify the units of measurement and completes the calibration and classification storage of typical operating conditions based on historical and real-time operating data. S3, Influent Water Quality and Quantity Prediction and Operating Condition Prediction Stage: Through the influent water quality and quantity short-term prediction and operating condition prediction module, using standardized processed historical data, real-time collected data, and meteorological data as input, the module completes the prediction of influent water quality and quantity change trends in multiple time dimensions. Based on the prediction results, the module identifies operating conditions, classifies shock risk levels, matches and outputs corresponding pre-control strategies. S4. Intelligent optimization stage of process parameters: Through the intelligent optimization module of process parameters under multiple constraints, rigid constraints and optimization objectives are set. With water inflow prediction data, real-time operating data and current operating parameters as input, the intelligent optimization of control parameters of the entire process section is completed through a deep reinforcement learning model. After compliance verification, executable control instructions are generated and issued. S5. Distributed execution and dynamic correction stage: Through the distributed real-time execution and dynamic correction module, the control instructions are received and distributed to the corresponding field actuators to complete precise actions. The process parameters after execution are collected in real time, and the deviation is identified by comparing with the set target threshold. The parameters are fine-tuned in real time through the closed-loop correction model. Equipment abnormalities are identified simultaneously and alarms and backup plans are triggered. S6. In the closed-loop evaluation and model iteration stage, the control effect closed-loop evaluation and model self-iteration module collects full-process operation data to complete the multi-dimensional evaluation of the control effect. Based on the evaluation results and newly added operation data, the incremental learning and iteration of the prediction model and the optimization model are completed. The operating condition database and expert rule base are updated simultaneously to generate the corresponding cycle of control effect evaluation report.

Images