Whole-process three-stage linkage control intelligent flocculation dosing method

Abstract

The invention belongs to the technical field of water treatment, and particularly relates to a whole-process three-stage linkage control intelligent flocculation dosing method, which comprises the following steps: collection: collecting water quality and water quantity parameters of inlet water and outlet water in real time in the water treatment process; collecting a floc image in the reaction process, analyzing the particle size distribution and the breakage rate of the floc image, and collecting a torque parameter of a stirring device at the same time; according to the invention, multiple types of sensors are deployed in the water inlet area of the flocculation tank, the reaction area and the water outlet area of the sedimentation tank, water quality, water quantity and reaction process parameters are collected in real time, and edge calculation and cloud optimization technologies are combined, so that full-flow intelligent control from agent pre-feeding, reaction process regulation and control to water outlet effect refinement is realized; the method is suitable for scenes such as municipal sewage treatment plants and industrial wastewater treatment stations requiring precise control of flocculant addition, and is especially suitable for water treatment processes with frequent water quality fluctuation and sensitive agent cost, so that the treatment efficiency can be remarkably improved, and the operation cost can be reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of water treatment, and in particular relates to an intelligent flocculation and dosing method with three-stage linkage control throughout the entire process. Background Art

[0002] As a core step in the water treatment process, flocculation and sedimentation efficiency directly impacts effluent quality and treatment costs. Existing dosing methods primarily rely on manual experience or feedback adjustment based on a single water quality indicator, resulting in three prominent issues: First, the timeliness of control is insufficient. Traditional manual inspections and fixed-ratio dosing are unable to cope with instantaneous water quality fluctuations. Second, precision is lacking. The single-parameter feedback mechanism is susceptible to interference from factors such as pH mutations and temperature changes, resulting in low chemical utilization. Third, the level of intelligence is limited. Existing automation systems often use preset mathematical models and lack the ability to dynamically adapt to complex water quality relationships. Studies have shown that a 10% fluctuation in coagulant dosage can lead to excessive effluent turbidity or a surge in sludge production. Traditional control methods generally use chemical waste rates exceeding 25%. With the rapid development of the Internet of Things and artificial intelligence technologies, there is an urgent need to build a new intelligent dosing system that can integrate multi-source data and implement real-time dynamic control.

[0003] The rapid development of technologies such as the Internet of Things, big data, and artificial intelligence is fueling innovation in the water treatment sector. As a key enabler of these technologies, smart water treatment systems integrate advanced equipment and technologies such as high-precision sensors, intelligent controllers, and data analysis platforms. These systems enable real-time dynamic monitoring of water quality parameters, multi-dimensional data analysis, and intelligent control of the entire process, significantly enhancing the automation and intelligence of water treatment processes.

[0004] Existing intelligent dosing technologies for wastewater treatment mainly rely on two types of methods: the first is the empirical method based on chemical mechanisms, which determines the dosage through preset mathematical expressions. However, this method lacks flexibility and is prone to overdosing due to reliance on experience; The second is the modeling method based on machine learning. Although it can adapt to changes in water quality, the model effect is limited by the quality of sample data.

[0005] Currently, sample collection is mostly done manually, which is inefficient, time-consuming, and prone to model overfitting. Although IoT sampling has improved data acquisition efficiency, it is still difficult to dynamically characterize the continuity of the dosing mechanism.

[0006] The above problems make it difficult for the existing dosing system to achieve the full-link collaboration of "multi-source perception-decision-making-optimization-execution" to achieve precise dosing. There is an urgent need for a dynamic dosing control method for intelligent flocculation sedimentation tanks based on multi-source perception. Summary of the Invention

[0007] The purpose of the present invention is to address the above-mentioned technical problems and provide an intelligent flocculation and dosing method with three-level linkage control throughout the entire process. By deploying multiple types of sensors in the water inlet area, reaction area and water outlet area of ​​the flocculation tank, water quality, water quantity and reaction process parameters are collected in real time. By combining edge computing and cloud optimization technology, intelligent control of the entire process from pre-dosing of reagents, reaction process regulation to fine-tuning of effluent effects is achieved. This method is suitable for scenarios such as municipal sewage treatment plants and industrial wastewater treatment stations that require precise control of flocculant dosage, especially for water treatment processes where water quality fluctuates frequently and reagent costs are sensitive, and can significantly improve treatment efficiency and reduce operating costs.

[0008] In view of this, the present invention provides an intelligent flocculation and dosing method with three-level linkage control throughout the entire process, comprising the following steps: Collection: Real-time collection of water quality and quantity parameters of influent and effluent during water treatment; collection of flocculent images during the reaction process and analysis of their particle size distribution and breakage rate; and collection of torque parameters of the stirring device; Feedforward control: Based on the parameters collected from the inlet water, the theoretical dosage of the reagent is calculated through the feedforward control model, and the initial dosage of the reagent is completed; Process control: Based on the floc image characteristic parameters collected during the reaction process, the process control model generates instructions to dynamically adjust the dosage of the reagent and the stirring speed; Feedback control: Based on the parameters collected from the effluent, the cumulative error correction amount is calculated through the PID algorithm to complete the fine adjustment of the reagent dosage.

[0009] Preferably, the water quality parameters of the influent include at least one of raw water turbidity, pH, water temperature, and suspended matter concentration, and the water quality parameters of the effluent include turbidity.

[0010] Preferably, the feedforward control model calculates the theoretical dosage of the agent based on the inlet flow rate, raw water turbidity, suspended solids concentration, deviation of pH value from the neutral value, and deviation of water temperature from the reference temperature, wherein: Each water quality parameter is weighted by a weight coefficient determined by historical data training; The weighted result is multiplied by the inlet flow rate to form the basic dosage; The basic dosage is corrected by the process correction coefficient that is dynamically adjusted according to the water quality type and season to obtain the final theoretical dosage of the agent.

[0011] Preferably, the process control model is , the control target is the ideal floc median particle size =100~150μm and optimal turbulence intensity =20~50s -1 , e 1、 f1 is the adjustment factor; D50 is obtained by analyzing the image.

[0012] Preferably, the cumulative error correction calculation formula is: , where K2 integral coefficient, S 出 is the effluent turbidity, S 设定 Set the value for outlet water turbidity.

[0013] An intelligent flocculation and dosing system with three-level linkage control throughout the entire process is applied to an intelligent flocculation and dosing method with three-level linkage control throughout the entire process, including: The full-process monitoring module is configured to collect water quality, water quantity and reaction process parameters in real time. The full-process monitoring module includes: A first electromagnetic flowmeter and a first multi-parameter water quality sensor located in the water inlet area of ​​the flocculation tank, wherein the first electromagnetic flowmeter is configured to monitor the inlet flow in real time, and the first multi-parameter water quality sensor is configured to simultaneously collect raw water turbidity, pH, water temperature, and suspended solids concentration; A floc recognition camera and a torque sensor are located in the reaction zone of the flocculation tank. The floc recognition camera is integrated with an edge computing module and is configured to capture floc images online, analyze particle size distribution and breakage rate, and output digital parameters. The torque sensor is configured to provide real-time feedback on the torque of the agitator paddle. A turbidity meter and a second electromagnetic flowmeter located in the outlet area of ​​the sedimentation tank, wherein the turbidity meter is configured to monitor the turbidity of the outlet water in real time, and the second electromagnetic flowmeter is configured to monitor the outlet water flow in real time; The three-level linkage control module is configured to achieve intelligent control of the entire process from pre-dosing of reagents, reaction process regulation to fine-tuning of water output effects. The three-level linkage control module includes: A feedforward control unit configured to calculate a theoretical dosage based on parameters collected from the inlet area of ​​the flocculation tank through a feedforward control model, and drive the electromagnetic flowmeter to complete the initial dosage of the reagent within 10 seconds; A process control unit is configured to generate fine-tuning instructions based on the floc image characteristic parameters collected in the reaction zone of the flocculation tank through a process control model to dynamically adjust the dosage and stirring speed; A feedback control unit is configured to calculate a cumulative error correction amount using a PID algorithm based on parameters collected from the effluent area of ​​the sedimentation tank, and complete a fine adjustment of the dosage within 20 seconds; The edge computing layer is configured to pre-process and extract features from sensor data, and run the three-level linkage control module to generate a control instruction once per second; The cloud-based optimization layer is configured to train long-term prediction models based on historical data to predict raw water quality trends, analyze nonlinear relationships between process parameters to establish a process knowledge graph, and use digital twin technology to simulate extreme operating conditions on the SimscapeWater platform to verify control strategies. The intelligent execution layer is configured to accurately execute control instructions. The intelligent execution layer includes a reagent dosing unit, a stirring control unit, and a visual monitoring interface for real-time display of the control process and equipment status.

[0014] Preferably, the edge computing layer adopts the 3σ principle to eliminate abnormal data and eliminates high-frequency noise through moving average filtering. The dynamic parameters extracted by the feature extraction include: turbidity change rate, pH fluctuation amplitude, and floc growth rate.

[0015] Preferably, the long-term time series prediction model of the cloud-based optimization layer adopts the Transformer-XL algorithm, which is trained based on historical data to predict the trend of raw water quality parameters.

[0016] Preferably, the drug dosing unit of the intelligent execution layer adopts a dual electromagnetic flowmeter redundant design and is equipped with an ultrasonic atomizer. The ultrasonic atomizer is equipped with a multi-speed particle size adjustment function to adjust the atomization particle size according to water quality fluctuations.

[0017] Preferably, the stirring control unit adopts a variable frequency motor and is equipped with a torque sensor to provide real-time feedback of the stirring torque.

[0018] The beneficial effects of the present invention are: 1. During the feedforward control process, pre-dosing is achieved based on the raw water characteristics, process control relies on dynamic fine-tuning of reaction intermediate parameters, and feedback control fine-tunes the dosage according to the water output effect. The three form a full-chain coverage of "pre-dosing - in-process regulation - post-adjustment" in the time dimension. The dosage error is smaller than that of traditional single-loop control, and the response speed is improved to seconds, which significantly improves the system's ability to respond to sudden changes in water quality.

[0019] 2. Through cameras and torque sensors, the floc morphology and stirring effect that traditionally rely on manual observation are converted into quantifiable parameters such as particle size distribution, breakage rate, and G value. A dynamic mapping relationship of "dosage-stirring intensity-floc characteristics" is established, which can provide an early warning of flocculation abnormalities 15 minutes in advance, preventing problems from being transmitted to the sedimentation stage and improving flocculation efficiency.

[0020] 3. The cloud platform uses digital twin technology to generate massive amounts of simulated operating condition data, automatically optimizing feedforward model weights, process control thresholds, and PID parameters. For example, it automatically reduces the temperature compensation coefficient during high temperatures in summer and increases the target stirring intensity during low temperatures in winter. This improves the system's adaptability and allows it to adapt to seasonal and water quality changes without human intervention, enabling continuous evolution of control strategies.

[0021] 4. Execution layer equipment (such as electromagnetic flowmeters and ultrasonic atomizers) are closely linked with the control algorithm. For example, the dynamic matching of atomized particle size and floc particle size can improve the diffusion efficiency of the agent, shorten the flocculation reaction time, and increase the equipment processing load, thereby achieving the synergistic maximization of hardware performance and algorithm accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic flow diagram of the present invention; Figure 2 It is a schematic diagram of the collected floc image of the present invention. DETAILED DESCRIPTION

[0023] The following will be combined with the accompanying drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of this application.

[0024] It should be noted that all terms used in the present invention to indicate direction and position, such as "up", "down", "left", "right", "front", "back", "vertical", "horizontal", "inside", "outside", "top", "low", "lateral", "longitudinal", "center", etc., are only used to explain the relative position relationship and connection status between the components in a certain specific state (as shown in the accompanying drawings). They are only for the convenience of describing the present invention, and do not require that the present invention must be constructed and operated in a specific orientation. Therefore, they cannot be understood as limiting the present invention. In addition, the descriptions of "first", "second", etc. in the present invention are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features.

[0025] In the description of the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood broadly. For example, they may refer to fixed, detachable, or integral connections; mechanical connections; direct connections or indirect connections through an intermediary; and internal communication between two components. Those skilled in the art will understand the specific meanings of these terms in the present invention based on the specific circumstances.

[0026] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative uses of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0027] like Figure 1 and Figure 2As shown, an intelligent flocculation dosing method of whole-process three-level linkage control includes the following steps: Collection: Real-time collection of water quality and water quantity parameters of influent and effluent in the water treatment process; collection of flocculation images in the reaction process and analysis of particle size distribution and breakage rate, and collection of torque parameters of the stirring device; Feedforward control: According to the parameters collected from the influent, the theoretical dosing amount of the reagent is calculated through the feedforward control model, and the initial dosing of the reagent is completed; Process control: According to the characteristic parameters of the flocculation images collected in the reaction process, the instruction is generated through the process control model to dynamically adjust the reagent dosing amount and stirring speed; Feedback control: Based on the parameters collected from the effluent, the cumulative error correction amount is calculated through the PID algorithm to complete the fine adjustment of the reagent dosing amount.

[0028] As a preferred example of the present application, the water quality parameters of the influent include at least one of raw water turbidity, pH, water temperature and suspended solids concentration, and the water quality parameters of the effluent include turbidity; The system installs an electromagnetic flowmeter and a multi-parameter water quality sensor (integrating turbidity, pH, water temperature, SS, COD detection modules, wherein the turbidity range is 0-2000 NTU, the SS range is 0-5000 mg / L) in the influent pipeline or channel of the flocculation tank to collect the influent flow (Q 进 ), raw water turbidity (T 原 ), pH value, water temperature (t) and suspended solids concentration (C SS ) at a frequency of 5 times per second; And a turbidimeter (range 0-1000 NTU, detection lower limit 0.1 NTU) and an electromagnetic flowmeter are installed at the effluent pipeline of the sedimentation tank to monitor the effluent turbidity (S 出 ) and flow (Q 出 ) in real time, and calculate the flow difference ΔQ (ΔQ=Q 进 -Q 出 ). The system takes the set value S 设定 of the effluent turbidity (such as ≤5 NTU) as the target, and calculates the cumulative error correction amount based on the integral term of the PID algorithm: The control accuracy reaches ±0.5% within 20 seconds. If ΔQ>5%, the system determines that there may be pipeline leakage or accumulation, triggers an early warning and freezes the dosing amount adjustment until manual investigation and confirmation. This module directly feeds back the final treatment effect to ensure that the effluent water quality is stable and meets the standards, and forms a closed loop with the feedforward and process control modules to realize iterative optimization of the control accuracy of the whole process; The specific deployment mode is shown in the following table: The above sensor data is transmitted to the edge computing layer through dual channels, and after preprocessing and feature extraction, it is deeply integrated with the control model to form a complete link of "data collection-analysis and decision-making-execution feedback".

[0029] As a preferred example of the present application, the feedforward control model calculates the theoretical dosage of the agent based on the inlet flow rate, raw water turbidity, suspended solids concentration, deviation of pH value from the neutral value, and deviation of water temperature from the reference temperature, wherein: Each water quality parameter is weighted by a weight coefficient determined by historical data training; The weighted result is multiplied by the inlet flow rate to form the basic dosage; The basic dosage is corrected by the process correction coefficient that is dynamically adjusted according to the water quality type and season to obtain the final theoretical dosage of the reagent; The details are as follows: The feedforward control model is , where K1 is the process correction coefficient, which is dynamically adjusted according to the water quality type and season, and a1−d1 is the parameter weight coefficient determined by historical data training.

[0030] As a preferred example of this application, the process control model is: ; The control target is the ideal floc median particle size =100~150μm and optimal turbulence intensity =20~50s -1 , e 1、 f1 is the adjustment coefficient; D50 is obtained by analyzing the image; Deploy underwater cameras (with integrated edge computing modules and a measurement range of 1-2000 μm, see Figure 2 ) and torque sensor (accuracy ±2%), the camera collects floc images at a maximum of 60 frames / second, and analyzes parameters such as particle size distribution (D10, D50, D90), breakage rate (η) and aspect ratio in real time through the built-in edge computing module, monitors floc particle size distribution (D10, D50, D90), breakage rate (η) and stirring paddle torque (M) in real time, and calculates the value of the floc image through the formula Calculate the liquid turbulence intensity G value (where P is the stirring power, μ is the liquid viscosity, and V is the reaction zone volume). The edge computing layer is based on the process control model Generate fine-tuning instructions, and the control target is the ideal floc median particle size =100~150μm and optimal turbulence intensity =20~50s -1If the D50 value is less than the target value and the G value is higher than the upper limit, the system determines that the dosage of the agent is insufficient and automatically triggers the dosing instruction to increase the dosage by 5%-10%. If the D50 value exceeds the target value and the floc breakage rate η is greater than 15%, the system determines that the dosage is excessive and reduces the dosage by 3%-5%. The frequency conversion motor is used to reduce the stirring speed by 10%-15% until the parameters return to the target range. This module captures the reaction process parameters in real time to achieve dynamic coordinated control of the dosage and stirring intensity, preventing the problem of insufficient flocculation or excessive breakage from being transmitted to the sedimentation stage.

[0031] As a preferred example of this application, the cumulative error correction amount calculation formula is: , where K2 integral coefficient, S 出 is the effluent turbidity, S 设定 Set the value for outlet water turbidity.

[0032] An intelligent flocculation and dosing system with three-level linkage control throughout the entire process is applied to an intelligent flocculation and dosing method with three-level linkage control throughout the entire process, including: The full-process monitoring module is configured to collect water quality, water quantity and reaction process parameters in real time. The full-process monitoring module includes: A first electromagnetic flowmeter and a first multi-parameter water quality sensor located in the water inlet area of ​​the flocculation tank, wherein the first electromagnetic flowmeter is configured to monitor the inlet flow in real time, and the first multi-parameter water quality sensor is configured to simultaneously collect raw water turbidity, pH, water temperature, and suspended solids concentration; A floc recognition camera and a torque sensor are located in the reaction zone of the flocculation tank. The floc recognition camera is integrated with an edge computing module and is configured to capture floc images online, analyze particle size distribution and breakage rate, and output digital parameters. The torque sensor is configured to provide real-time feedback on the torque of the agitator paddle. A turbidity meter and a second electromagnetic flowmeter located in the outlet area of ​​the sedimentation tank, wherein the turbidity meter is configured to monitor the turbidity of the outlet water in real time, and the second electromagnetic flowmeter is configured to monitor the outlet water flow in real time; The three-level linkage control module is configured to achieve intelligent control of the entire process from pre-dosing of reagents, reaction process regulation to fine-tuning of water output effects. The three-level linkage control module includes: A feedforward control unit configured to calculate a theoretical dosage based on parameters collected from the inlet area of ​​the flocculation tank through a feedforward control model, and drive the electromagnetic flowmeter to complete the initial dosage of the reagent within 10 seconds; A process control unit is configured to generate fine-tuning instructions based on the floc image characteristic parameters collected in the reaction zone of the flocculation tank through a process control model to dynamically adjust the dosage and stirring speed; A feedback control unit is configured to calculate a cumulative error correction amount using a PID algorithm based on parameters collected from the effluent area of ​​the sedimentation tank, and complete a fine adjustment of the dosage within 20 seconds; The edge computing layer is configured to pre-process and extract features from sensor data, and run the three-level linkage control module to generate a control instruction once per second; The edge computing layer utilizes industrial-grade servers (CPU ≥ 8 cores, memory ≥ 16GB), integrating data preprocessing, feature extraction, and a three-level control algorithm to support millisecond-level response. The data preprocessing phase uses the 3σ principle to eliminate abnormal data with a mutation rate exceeding 20% ​​(e.g., a sudden jump in SS concentration > 500 mg / L). Moving average filtering (with a 3-minute time window) eliminates high-frequency noise and preserves data trends. Parameters such as particle size distribution and breakage rate transmitted by the alum floc wiper camera are directly used by the edge layer for process control model calculations without the need for additional image processing. The feature extraction module calculates dynamic parameters such as turbidity change rate (ΔNTU / s), pH fluctuation amplitude (ΔpH / 10 minutes), and floc growth rate (μm / min), providing more representative input features for the control model. The control algorithm module runs feedforward, process, and feedback control models in parallel, generating control instructions once per second to drive the execution layer to adjust the dosage and agitation parameters. For example, when a sudden increase of 500 NTU in raw water turbidity is detected, the feedforward model immediately calculates the α value increment, the process model simultaneously monitors the floc particle size growth rate, and the feedback model predicts the effluent turbidity trend. The three-level linkage completes the dosage adjustment within 10 seconds, ensuring that the dosage of the chemical responds synchronously to the change in water quality. Edge computing is based on a pre-dosing model: The theoretical dosage is calculated, where K1 is the process correction factor (ranging from 0.7 to 1.3), which is dynamically adjusted based on water quality type (e.g., municipal sewage / industrial wastewater) and season. For example, in winter, low temperatures automatically increase K1 to 1.1 to compensate for decreased agent activity. a1, b1, and d1 are parameter weighting coefficients (e.g., a1 = 0.008, b1 = 0.002, C1 = 0.2, d1 = 0.1), determined through historical data training to reflect the impact of different parameters on agent demand. After the calculation is complete, the system drives the electromagnetic flowmeter through open-loop control to complete the initial agent dosage within 10 seconds, ensuring full contact between the agent and the raw water in the mixing zone and covering approximately 80% of the theoretical agent demand, effectively addressing sudden changes in raw water quality (e.g., high turbidity influent caused by heavy rain).

[0033] The cloud-based optimization layer is configured to train long-term prediction models based on historical data to predict raw water quality trends, analyze nonlinear relationships between process parameters to establish a process knowledge graph, and use digital twin technology to simulate extreme operating conditions on the SimscapeWater platform to verify control strategies. Specifically, the cloud optimization layer is based on a cloud computing platform, integrating a long-term prediction model, a process knowledge graph, and a digital twin verification module to realize offline training and online optimization of control strategies. The long-term prediction model uses the Transformer-XL algorithm to train historical data from 2002 to 2025 (including multiple dimensions such as raw water quality, weather, and season), which can predict the trends of parameters such as turbidity and SS concentration up to 96 hours in advance, with a prediction error of less than 3%. For example, when heavy rain is predicted, the weight coefficient of the SS parameter in the feedforward model is automatically increased to increase the dosage reserve in advance. The process knowledge graph uses graph neural networks to analyze the non-linear relationships between 1800+ process parameters (such as "water temperature ↓ 2°C → drug hydrolysis rate ↓ 5% → need to increase dosage by 3%"), establishing a dynamic correlation rule base to automatically update the edge layer control model parameters every morning, improving the system's adaptability to complex water quality conditions. The digital twin verification module builds a 1:1 virtual model of the water plant on the SimscapeWater platform, simulates 2000+ extreme working conditions such as "low temperature and high turbidity" and "high pH and low SS", automatically generates a three-level control strategy package and verifies its effectiveness, with a strategy accuracy of ≥99%. For example, by simulating the influence of different temperature compensation coefficients on flocculation effect, the dosage strategy in winter conditions is optimized to improve the utilization rate of chemicals.

[0034] The intelligent execution layer is configured to accurately execute control instructions, and the intelligent execution layer includes a chemical dosage unit, a stirring control unit, and a visual monitoring interface for real-time display of control processes and equipment states. Specifically, the intelligent execution layer is composed of a chemical dosage unit, a stirring control unit, and a visual monitoring interface: Chemical dosage unit: dual electromagnetic flowmeter redundancy design (primary + backup), adjustment accuracy ±0.3%, supporting milligram-level control at 0.1L / h level; ultrasonic atomizer equipped with multi-gear particle size adjustment function (50-300μm, resolution 5μm), which can automatically adjust the atomization particle size according to water quality fluctuations, such as reducing the atomization particle size to 80μm in low temperature conditions to improve the contact efficiency of chemicals and suspended solids.

[0035] Stirring control unit: variable frequency motor drives stirring paddle, power adjustment range 20%-100%, torque sensor real-time feedback stirring torque, G value control error ±5%; when the flocculation breakage rate η > 20%, automatically trigger the "speed reduction and drug reduction" linkage mechanism, first reduce the stirring speed by 15%, then reduce the dosage by 8%, to avoid excessive breakage.

[0036] Visual monitoring interface: Displays real-time three-level control curves (α value pre-dosing, γ value fine-tuning, and β value refinement), a dynamic graph of floc particle size distribution, a cloud diagram of chemical diffusion concentration, and energy consumption trends (e.g., kW・h / m³), supporting 7-day historical data traceability. An integrated mobile early warning system automatically sends an alarm to operations and maintenance personnel when effluent turbidity exceeds the standard for 10 consecutive minutes or the chemical tank inventory falls below 10%, enabling transparent monitoring of the entire process and enabling abnormal response.

[0037] In summary, the intelligent flocculation and dosing system and system with three-level linkage control throughout the entire process proposed by the present invention achieves: 1. Collaborative optimization of the three-level linkage control architecture: Feedforward control implements pre-dosing based on raw water characteristics, process control relies on dynamic fine-tuning of reaction intermediate parameters, and feedback control fine-tunes the dosage according to the water output effect. The three form a full chain coverage of "pre-dosing - in-process regulation - post-adjustment" in the time dimension. Compared with traditional single-loop control, the dosage error is effectively reduced, the response speed is improved to seconds, and the system's ability to respond to sudden changes in water quality is significantly improved.

[0038] 2. Quantitative monitoring and dynamic control of the reaction process: Through underwater cameras and torque sensors, the floc morphology and stirring effect that traditionally rely on manual observation are converted into quantifiable parameters such as particle size distribution, breakage rate, and G value. A dynamic mapping relationship of "dosage-stirring intensity-floc characteristics" is established, and flocculation abnormalities are warned 15 minutes in advance to prevent problems from being transmitted to the sedimentation stage, thereby improving flocculation efficiency.

[0039] 3. Data-driven system self-evolution capability: The cloud platform uses digital twin technology to generate massive amounts of simulated operating condition data, automatically optimizing feedforward model weights, process control thresholds, and PID parameters. For example, it automatically reduces the temperature compensation coefficient during high temperatures in summer and increases the target stirring intensity during low temperatures in winter. This improves the system's adaptability and allows it to adapt to seasonal and water quality changes without human intervention, enabling continuous evolution of control strategies.

[0040] 4. Deep coupling design of hardware and algorithms: Execution layer devices (such as electromagnetic flowmeters and ultrasonic atomizers) are closely linked with control algorithms. For example, the dynamic matching of atomized particle size and floc particle size can improve the efficiency of drug diffusion, shorten the flocculation reaction time, and increase the equipment processing load, thereby achieving the synergistic maximization of hardware performance and algorithm accuracy.

[0041] The embodiments of the present application are described above in conjunction with the accompanying drawings. Unless there is a conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of this application, ordinary technicians in this field can also make many forms without departing from the purpose of this application and the scope of protection of the claims, all of which are within the protection of this application.

Claims

1. An intelligent flocculation and dosing method with three-level linkage control throughout the entire process, characterized by: The following steps are involved: Collection: Real-time collection of water quality and quantity parameters of influent and effluent during water treatment; collection of flocculent images during the reaction process and analysis of their particle size distribution and breakage rate; and collection of torque parameters of the stirring device; Feedforward control: Based on the parameters collected from the inlet water, the theoretical dosage of the reagent is calculated through the feedforward control model, and the initial dosage of the reagent is completed; Process control: Based on the floc image characteristic parameters collected during the reaction process, the process control model generates instructions to dynamically adjust the dosage of the reagent and the stirring speed; Feedback control: Based on the parameters collected from the effluent, the cumulative error correction amount is calculated through the PID algorithm to complete the fine adjustment of the reagent dosage.

2. The intelligent flocculation and dosing method with three-level linkage control throughout the entire process according to claim 1 is characterized by: The water quality parameters of the influent include at least one of raw water turbidity, pH, water temperature, and suspended matter concentration, and the water quality parameters of the effluent include turbidity.

3. The intelligent flocculation and dosing method with three-level linkage control throughout the entire process according to claim 2 is characterized by: The feedforward control model calculates the theoretical dosage of the reagent based on the inlet flow rate, raw water turbidity, suspended solids concentration, deviation of pH value from the neutral value, and deviation of water temperature from the reference temperature, where: Each water quality parameter is weighted by a weight coefficient determined by historical data training; The weighted result is multiplied by the inlet flow rate to form the basic dosage; The basic dosage is corrected by the process correction coefficient that is dynamically adjusted according to the water quality type and season to obtain the final theoretical dosage of the agent.

4. The intelligent flocculation and dosing method with three-level linkage control throughout the entire process according to claim 3 is characterized by: The process control model is , the control target is the ideal floc median particle size =100~150μm and optimal turbulence intensity =20~50s -1 , e 1、 f1 is the adjustment factor; D50 is obtained by analyzing the image.

5. The intelligent flocculation and dosing method with full-process three-level linkage control according to claim 4 is characterized in that: The cumulative error correction calculation formula is: , where K2 integral coefficient, S 出 is the effluent turbidity, S 设定 Set the value for outlet water turbidity.

6. An intelligent flocculation and dosing system with full-process three-stage linkage control, applied to the intelligent flocculation and dosing method with full-process three-stage linkage control according to any one of claims 1 to 5, characterized in that: include: The full-process monitoring module is configured to collect water quality, water quantity and reaction process parameters in real time. The full-process monitoring module includes: A first electromagnetic flowmeter and a first multi-parameter water quality sensor located in the water inlet area of ​​the flocculation tank, wherein the first electromagnetic flowmeter is configured to monitor the inlet flow in real time, and the first multi-parameter water quality sensor is configured to simultaneously collect raw water turbidity, pH, water temperature, and suspended solids concentration; A floc recognition camera and a torque sensor are located in the reaction zone of the flocculation tank. The floc recognition camera is integrated with an edge computing module and is configured to capture floc images online, analyze particle size distribution and breakage rate, and output digital parameters. The torque sensor is configured to provide real-time feedback on the torque of the agitator paddle. A turbidity meter and a second electromagnetic flowmeter located in the outlet area of ​​the sedimentation tank, wherein the turbidity meter is configured to monitor the turbidity of the outlet water in real time, and the second electromagnetic flowmeter is configured to monitor the outlet water flow in real time; The three-level linkage control module is configured to achieve intelligent control of the entire process from pre-dosing of reagents, reaction process regulation to fine-tuning of water output effects. The three-level linkage control module includes: A feedforward control unit configured to calculate a theoretical dosage based on parameters collected from the inlet area of ​​the flocculation tank through a feedforward control model, and drive the electromagnetic flowmeter to complete the initial dosage of the reagent within 10 seconds; A process control unit is configured to generate fine-tuning instructions based on the floc image characteristic parameters collected in the reaction zone of the flocculation tank through a process control model to dynamically adjust the dosage and stirring speed; A feedback control unit is configured to calculate a cumulative error correction amount using a PID algorithm based on parameters collected from the effluent area of ​​the sedimentation tank, and complete a fine adjustment of the dosage within 20 seconds; The edge computing layer is configured to pre-process and extract features from sensor data, and run the three-level linkage control module to generate a control instruction once per second; The cloud-based optimization layer is configured to train long-term prediction models based on historical data to predict raw water quality trends, analyze nonlinear relationships between process parameters to establish a process knowledge graph, and use digital twin technology to simulate extreme operating conditions on the SimscapeWater platform to verify control strategies. The intelligent execution layer is configured to accurately execute control instructions. The intelligent execution layer includes a reagent dosing unit, a stirring control unit, and a visual monitoring interface for real-time display of the control process and equipment status.

7. The intelligent flocculation and dosing system with three-level linkage control throughout the entire process according to claim 6 is characterized by: The edge computing layer uses the 3σ principle to eliminate abnormal data and eliminates high-frequency noise through moving average filtering. The dynamic parameters extracted by the feature extraction include: turbidity change rate, pH fluctuation amplitude, and floc growth rate.

8. The intelligent flocculation and dosing system with three-level linkage control throughout the entire process according to claim 7 is characterized by: The long-term time series prediction model of the cloud-based optimization layer adopts the Transformer-XL algorithm and is trained based on historical data to predict the trend of raw water quality parameters.

9. The intelligent flocculation and dosing system with three-level linkage control throughout the entire process according to claim 8 is characterized by: The drug dosing unit of the intelligent execution layer adopts a dual electromagnetic flowmeter redundant design and is equipped with an ultrasonic atomizer. The ultrasonic atomizer is equipped with a multi-speed particle size adjustment function to adjust the atomization particle size according to water quality fluctuations.

10. The intelligent flocculation and dosing system with three-level linkage control throughout the entire process according to claim 9, characterized in that: The stirring control unit adopts a variable frequency motor and is equipped with a torque sensor to provide real-time feedback of the stirring torque.

Images