Sewage treatment monitoring management method and system based on digital twinning

Abstract

The invention relates to the technical field of sewage equipment management, and discloses a digital twinning-based sewage treatment monitoring management method and system, and the method comprises the following steps: 1, data collection: collecting the operation data of a physical entity in real time through sensors disposed in each process link of a sewage treatment plant; 2, constructing a digital twinborn model based on the acquired physical entity data; 3, data interaction and synchronization: establishing a real-time data interaction channel between the physical entity and the digital twin model; 4, training and optimizing a deep learning model, and training the deep learning model based on historical operation data and real-time collection data; step 5, virtual simulation and decision generation: performing virtual simulation on the sewage treatment process by using a digital twinborn model, and combining a prediction result of a deep learning model; and 6, issuing and executing an instruction, and converting the generated optimal process regulation and control decision into a control instruction.

Description

Technical Field

[0001] This invention relates to the field of wastewater treatment equipment management technology, specifically to a wastewater treatment monitoring and management method and system based on digital twins. Background Technology

[0002] Digital technologies such as the Internet of Things, big data, and artificial intelligence are profoundly reshaping the development logic of the wastewater treatment industry. At the perception layer, intelligent sensors enable real-time monitoring of key indicators such as COD, ammonia nitrogen, MLSS, and RSS, and data is automatically collected online. At the decision-making layer, machine learning algorithms can automatically optimize aeration parameters based on fluctuations in influent water quality, reducing treatment energy consumption by 20%.

[0003] Industry practice shows that digital transformation can improve the operation and maintenance efficiency of wastewater treatment plants by more than 30%, and shorten the response time to abnormal faults to within one hour. At the same time, the virtual water plant model built by digital twin technology provides a low-cost simulation testing platform for process upgrades, avoiding the resource waste of traditional trial and error models. This "digital-driven" transformation is driving the wastewater treatment industry to shift from labor-intensive to technology-intensive, and from experience-based decision-making to data-driven decision-making, breaking through the bottlenecks and limitations of traditional wastewater treatment operations.

[0004] Based on this, a search on the patent website revealed that Chinese patent application number CN202411499120.0 discloses a sewage treatment equipment operation management system and method thereof;

[0005] Existing wastewater treatment equipment management systems lack overall simulation and collaborative optimization of the entire wastewater treatment process. This results in independent control of each process unit, making it difficult to integrate and control, and thus failing to achieve global optimization. Furthermore, data utilization efficiency is low, with a large amount of monitoring data not being effectively mined, failing to provide in-depth support for process optimization.

[0006] To address the aforementioned problems, a wastewater treatment monitoring and management method and system based on digital twins is proposed. Summary of the Invention

[0007] The purpose of this invention is to provide a wastewater treatment monitoring and management method and system based on digital twins. By using this device, the existing wastewater equipment management system is able to solve the problems of lack of overall simulation and collaborative optimization of the entire wastewater treatment process, which leads to independent control between process units, making it difficult to integrate and control, and thus failing to achieve global optimization; in addition, the data utilization efficiency is low, and a large amount of monitoring data is not effectively mined, which cannot provide in-depth support for process optimization.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a wastewater treatment monitoring and management method based on digital twins, comprising the following steps:

[0009] Step 1: Data acquisition. Sensors deployed in various process stages of the wastewater treatment plant are used to collect real-time operational data of the physical entities. The operational data includes water quality parameters, equipment operating parameters, and environmental parameters.

[0010] Step 2: Digital twin model construction. Based on the collected physical entity data, a digital twin model that maps to the physical wastewater treatment plant in a 1:1 ratio is constructed. The digital twin model includes a geometric model, a physical model, a behavioral model, and a rule model.

[0011] Step 3: Data interaction and synchronization. Establish a real-time data interaction channel between the physical entity and the digital twin model, and transmit the collected operational data to the digital twin model in real time to achieve dynamic synchronization between the virtual model and the physical entity.

[0012] Step 4: Deep learning model training and optimization. Based on historical operating data and real-time collected data, a deep learning model is trained. The deep learning model includes a water quality prediction model, a process parameter optimization model, and a fault early warning model. The model is periodically iterated and optimized using new data to ensure that the model prediction error is less than 8%.

[0013] Step 5: Virtual simulation and decision generation. The wastewater treatment process is simulated using a digital twin model. Combined with the prediction results of the deep learning model, the impact of different process parameter adjustments on the treatment effect is simulated to generate the optimal process control decision.

[0014] Step Six: Instruction Issuance and Execution. The generated optimal process control decision is transformed into control instructions and issued to the execution equipment in the physical wastewater treatment plant to achieve precise control of the wastewater treatment process. At the same time, the execution effect is monitored in real time, forming a closed-loop management of "collection-simulation-decision-execution-feedback".

[0015] Preferably, in step one, the water quality parameters include COD, ammonia nitrogen, total nitrogen, total phosphorus, dissolved oxygen (DO), pH value, and sludge concentration (MLSS); the equipment operating parameters include water pump flow rate, blower frequency, dosing pump dosage, equipment vibration value, and current and voltage; the environmental parameters include water temperature, atmospheric temperature, humidity, and rainfall; the data acquisition frequency is set according to the importance of the parameters, with key water quality parameters acquired every 5 seconds, equipment operating parameters acquired every 10 seconds, and environmental parameters acquired every 30 seconds.

[0016] Preferably, in step two, the geometric model is constructed using the Three.js engine, achieving millimeter-level modeling accuracy. It replicates the core process areas such as the bioreactor, sedimentation tank, and aeration system at a 1:1 scale, while non-critical areas are treated with lightweighting. The physical model is constructed based on the activated sludge digestion model, containing 13 state variables and 20 kinetic parameters to accurately simulate biochemical reaction processes such as nitrification, denitrification, and biological phosphorus removal. The behavioral model uses Python scripts to write the equipment operation logic, defining the association rules between equipment start-up and shutdown, parameter adjustment, and process conditions. The rule model integrates constraints such as environmental standards and equipment operation and maintenance specifications, providing a basis for virtual simulation and decision generation.

[0017] Preferably, in step three, the data interaction channel adopts an "edge computing + 5G industrial Ethernet" architecture. The edge gateway preprocesses the collected data, including filtering, outlier removal and data normalization. The preprocessed data is transmitted to the digital twin platform via 5G industrial Ethernet. When the network is interrupted, the edge gateway can cache data for 2 hours and automatically retransmit it after recovery. The data loss rate is less than 0.5%.

[0018] Preferably, in step four, the water quality prediction model uses the LSTM algorithm, inputting 18 parameters such as influent flow rate, COD, ammonia nitrogen, and water temperature to predict water quality indicators for the next 1-24 hours; the process parameter optimization model uses a "CNN-LSTM" hybrid architecture to output the optimal values ​​of key parameters such as aeration rate and carbon source dosage, stabilizing dissolved oxygen at 1.5 mg / L and controlling carbon source dosage accuracy at 5%; the fault early warning model is based on autoencoder and random forest algorithms to provide early warnings of faults in key equipment such as water pumps and blowers 7 to 14 days in advance, with an early warning accuracy rate greater than 95%.

[0019] Preferably, in step five, the virtual simulation includes full-process simulation, extreme condition simulation, and equipment failure simulation. The full-process simulation presents the operating status and water quality changes of each process unit in real time; the extreme condition simulation can simulate more than 20 scenarios such as rainstorm impact and sudden changes in influent water quality, and predict the trend of treatment effect changes; the equipment failure simulation can simulate the impact of failures such as aeration head blockage and water pump failure on the process, and verify emergency treatment plans in advance; in the decision generation process, multiple objectives such as water quality compliance rate, energy consumption, and chemical consumption are comprehensively considered, and the optimal combination of process parameters is determined through a multi-objective optimization algorithm.

[0020] Preferably, in step six, the control command is sent to the PLC control system via the OPCUA protocol to drive the execution equipment to perform actions, including adjusting the blower frequency, the dosage of the dosing pump, and the sludge return ratio; the execution effect monitoring is carried out by comparing the simulation results of the digital twin model with the actual operating data of the physical entity to calculate the deviation value. If the deviation value exceeds the preset threshold, the model parameter correction or regeneration of the control decision is triggered.

[0021] A wastewater treatment monitoring and management system based on digital twins includes:

[0022] The data acquisition module consists of multiple sensors and an edge gateway. The sensors are deployed in various process stages of the wastewater treatment plant to collect water quality parameters, equipment operating parameters, and environmental parameters. The edge gateway is used to preprocess and transmit the collected data.

[0023] The digital twin platform module includes a geometric modeling module, a physical modeling module, a data interaction module, and a virtual simulation module. The geometric modeling module constructs a three-dimensional geometric model of the physical entity, the physical modeling module constructs a physical model of the biochemical reaction, the data interaction module realizes real-time data synchronization between the physical entity and the virtual model, and the virtual simulation module simulates and analyzes the wastewater treatment process.

[0024] The deep learning module includes a model training module, a model optimization module, and a prediction and decision module. The model training module trains water quality prediction, process optimization, and fault early warning models based on historical and real-time data. The model optimization module iteratively updates the model parameters periodically. The prediction and decision module generates process control suggestions based on the model output.

[0025] The control and execution module includes a PLC control system and an execution device. The PLC control system receives control commands from the digital twin platform and the deep learning layer, and drives the execution device to perform parameter adjustment actions.

[0026] The visualization monitoring module uses Vue.js + Three.js to develop a 3D visualization interface that displays the real-time operating status of the digital twin model, trend curves of key parameters, and equipment fault warning information, and supports user interaction and data query.

[0027] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0028] 1. This application achieves refined control of the entire wastewater treatment process through real-time simulation of digital twin models and accurate prediction of deep learning models. The dissolved oxygen control accuracy reaches 0.2 mg / L, the carbon source addition accuracy reaches 5%, and the effluent compliance rate is increased to over 99.5%, solving the problem of low control accuracy of traditional methods.

[0029] 2. This application constructs a closed-loop management system of "data acquisition-simulation-decision-execution-feedback", integrating functions such as data acquisition, virtual simulation, intelligent decision-making, and precise control, promoting the transformation of sewage treatment plants from experience-driven to data-driven, and providing a replicable technical solution for the development of smart water affairs.

[0030] 3. This application reduces the response time for abnormal operating conditions from 1 hour to 15 minutes by real-time synchronization of physical entities and digital twin models, combined with the rapid prediction of deep learning models, and provides early warning of equipment failures 7 to 14 days in advance, thus solving the problem of slow response speed of traditional methods and accelerating response speed and emergency response capabilities. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the overall process of the present invention;

[0032] Figure 2 This is a flowchart of the data acquisition module of the present invention;

[0033] Figure 3 This is a flowchart of the digital twin platform module of the present invention;

[0034] Figure 4 This is a flowchart of the deep learning module of the present invention;

[0035] Figure 5 This is a flowchart of the control execution module of the present invention. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] To further understand the content of this invention, a detailed description of the invention will be provided in conjunction with the accompanying drawings.

[0038] Combination Figures 1-3 A wastewater treatment monitoring and management method based on digital twins includes the following steps:

[0039] Step 1: Data Acquisition. Sensors deployed in various process stages of the wastewater treatment plant are used to collect real-time operational data of the physical entities. This operational data includes water quality parameters, equipment operating parameters, and environmental parameters.

[0040] Step 2: Digital twin model construction. Based on the collected physical entity data, a digital twin model that maps to the physical wastewater treatment plant in a 1:1 ratio is constructed. The digital twin model includes a geometric model, a physical model, a behavioral model, and a rule model.

[0041] Step 3: Data interaction and synchronization. Establish a real-time data interaction channel between the physical entity and the digital twin model, and transmit the collected operational data to the digital twin model in real time to achieve dynamic synchronization between the virtual model and the physical entity.

[0042] Step 4: Deep learning model training and optimization. Based on historical operating data and real-time collected data, train deep learning models, including water quality prediction models, process parameter optimization models, and fault early warning models. Regularly use new data to iteratively optimize the models to ensure that the model prediction error is less than 8%.

[0043] Step 5: Virtual simulation and decision generation. The wastewater treatment process is simulated using a digital twin model. Combined with the prediction results of the deep learning model, the impact of different process parameter adjustments on the treatment effect is simulated to generate the optimal process control decision.

[0044] Step Six: Instruction Issuance and Execution. The generated optimal process control decision is transformed into control instructions and issued to the execution equipment in the physical wastewater treatment plant to achieve precise control of the wastewater treatment process. At the same time, the execution effect is monitored in real time, forming a closed-loop management of "collection-simulation-decision-execution-feedback".

[0045] In step one, water quality parameters include COD, ammonia nitrogen, total nitrogen, total phosphorus, dissolved oxygen (DO), pH value, and sludge concentration (MLSS); equipment operating parameters include water pump flow rate, blower frequency, dosing pump dosage, equipment vibration value, current, and voltage; environmental parameters include water temperature, atmospheric temperature, humidity, and rainfall; data acquisition frequency is set according to the importance of the parameters, with key water quality parameters acquired every 5 seconds, equipment operating parameters acquired every 10 seconds, and environmental parameters acquired every 30 seconds.

[0046] In step two, the geometric model is built using the Three.js engine, achieving millimeter-level modeling accuracy. It replicates the core process areas such as the bioreactor, sedimentation tank, and aeration system at a 1:1 scale, while non-critical areas are treated with lightweight materials. The physical model is built based on the Activated Sludge Digestion Model (ASDM), containing 13 state variables and 20 kinetic parameters to accurately simulate biochemical reaction processes such as nitrification, denitrification, and biological phosphorus removal. The behavioral model uses Python scripts to write the equipment operation logic, defining the association rules between equipment start-up and shutdown, parameter adjustment, and process conditions. The rule model integrates constraints such as environmental standards and equipment operation and maintenance specifications, providing a basis for virtual simulation and decision generation.

[0047] In step three, the data interaction channel adopts an "edge computing + 5G industrial Ethernet" architecture. The edge gateway preprocesses the collected data, including filtering, outlier removal and data normalization. The preprocessed data is transmitted to the digital twin platform via 5G industrial Ethernet. When the network is interrupted, the edge gateway can cache data for 2 hours and automatically retransmit it after recovery. The data loss rate is less than 0.5%.

[0048] In step four, the water quality prediction model uses the LSTM algorithm, taking 18 parameters such as influent flow rate, COD, ammonia nitrogen, and water temperature as input to predict water quality indicators for the next 1-24 hours. The process parameter optimization model uses a "CNN-LSTM" hybrid architecture to output the optimal values ​​of key parameters such as aeration rate and carbon source dosage, stabilizing dissolved oxygen at 1.5 mg / L and controlling carbon source dosage accuracy at 5%. The fault early warning model is based on autoencoder and random forest algorithms to provide early warnings of faults in key equipment such as water pumps and blowers 7 to 14 days in advance, with an accuracy rate of over 95%.

[0049] In step five, virtual simulation includes full-process simulation, extreme condition simulation, and equipment failure simulation. The full-process simulation presents the operating status and water quality changes of each process unit in real time. The extreme condition simulation can simulate more than 20 scenarios such as rainstorm impact and sudden changes in influent water quality, and predict the trend of treatment effect changes. The equipment failure simulation can simulate the impact of failures such as aeration head blockage and water pump failure on the process, and verify emergency treatment plans in advance. In the decision generation process, multiple objectives such as water quality compliance rate, energy consumption, and chemical consumption are comprehensively considered, and the optimal combination of process parameters is determined through multi-objective optimization algorithms.

[0050] In step six, control commands are sent to the PLC control system via the OPCUA protocol to drive the execution of equipment actions, including adjusting the blower frequency, the dosage of the dosing pump, and the sludge return ratio. The execution effect is monitored by comparing the simulation results of the digital twin model with the actual operating data of the physical entity to calculate the deviation value. If the deviation value exceeds the preset threshold, the model parameters are corrected or the control decision is regenerated.

[0051] A wastewater treatment monitoring and management system based on digital twins includes:

[0052] The data acquisition module consists of multiple sensors and an edge gateway. The sensors are deployed in various process stages of the wastewater treatment plant to collect water quality parameters, equipment operating parameters, and environmental parameters. The edge gateway is used to preprocess and transmit the collected data.

[0053] The digital twin platform module includes a geometric modeling module, a physical modeling module, a data interaction module, and a virtual simulation module. The geometric modeling module constructs a three-dimensional geometric model of the physical entity, the physical modeling module constructs a physical model of the biochemical reaction, the data interaction module realizes real-time data synchronization between the physical entity and the virtual model, and the virtual simulation module simulates and analyzes the wastewater treatment process.

[0054] The deep learning module includes a model training module, a model optimization module, and a prediction and decision module. The model training module trains water quality prediction, process optimization, and fault early warning models based on historical and real-time data. The model optimization module iteratively updates the model parameters periodically. The prediction and decision module generates process control suggestions based on the model output.

[0055] The control and execution module includes a PLC control system and an execution device. The PLC control system receives control commands from the digital twin platform and the deep learning layer, and drives the execution device to perform parameter adjustment actions.

[0056] The visualization monitoring module uses Vue.js + Three.js to develop a 3D visualization interface that displays the real-time operating status of the digital twin model, trend curves of key parameters, and equipment fault warning information, and supports user interaction and data query.

[0057] The present invention will be further described below with reference to embodiments.

[0058] Example 1:

[0059] A new urban domestic wastewater treatment plant project with a daily capacity of 100,000 tons will be constructed.

[0060] Following the principle of "densification in the core area and simplification in the auxiliary area," an online COD analyzer, ammonia nitrogen sensor, and ultrasonic flow meter are deployed at the inlet; fluorescent dissolved oxygen sensor, ORP sensor, and optical sludge concentration meter are deployed in different areas of the bioreactor; ultrasonic level gauge and sludge interface meter are deployed in the sedimentation tank; and an online total nitrogen / total phosphorus analyzer is deployed at the outlet. All sensors are IP68 protected and suitable for humid and corrosive environments.

[0061] Data transmission and preprocessing: Each process unit is equipped with one industrial-grade edge gateway to preprocess the collected data—using IQR to remove outliers (such as sudden increases or decreases in DO), linear interpolation to fill missing values, and Min-Max standardization to the [0,1] interval; key water quality parameters (DO, ammonia nitrogen) are collected every 5 seconds, equipment operating parameters (blower frequency, chemical dosing pump dosage) every 10 seconds, and environmental parameters (water temperature, rainfall) every 30 seconds; when the network is interrupted, the gateway caches 2 hours of data and automatically retransmits it after recovery.

[0062] Digital Twin Platform Construction: Geometric Model: A 1:1 3D model is built using the Three.js engine. The modeling accuracy of core areas such as the bioreactor and aeration system reaches the millimeter level, while non-critical areas are lightweighted. The model loading time is less than 30 seconds. Physical Model: Based on the activated sludge digestion model, it integrates the Monod and Eckenfelder models mentioned in the process optimization documents. It includes 13 state variables and 20 kinetic parameters, accurately simulating nitrification, denitrification, and biological phosphorus removal processes with a simulation error of less than 5%. Virtual Simulation Function: It supports full-process simulation, extreme condition simulation, and virtual process debugging, improving debugging efficiency by 80% compared to on-site trial and error.

[0063] Deep Learning Model Applications: Influent Water Quality Prediction Model: Employing a hybrid architecture of "LSTM + EEMD-Transformer," in the short term (1-24 hours), LSTM predicts influent COD and ammonia nitrogen with 18 input parameters and a prediction error of less than 8%. In the medium to long term (24-72 hours), EEMD-Transformer integrates meteorological data (rainfall and temperature) to predict peak flow and water quality trends, providing a basis for reagent reserves and equipment maintenance. Process Parameter Optimization Model: Using a hybrid architecture of "CNN-LSTM," convolutional layers extract DO distribution characteristics in different areas of the bioreactor, while LSTM layers capture temporal patterns, outputting optimal process parameters—the aeration system stabilizes DO at 1.5 mg / L, carbon source dosing accuracy is ±5%, and sludge return ratio adjustment error is less than 3%. Equipment Fault Early Warning Model: Based on an autoencoder and random forest algorithm, it collects pump vibration values ​​and blower current data to predict impeller wear, aeration head blockage, and other faults 7-14 days in advance with an accuracy greater than 95%.

[0064] Control execution and closed-loop management: Control command issuance: Optimized parameters are issued to the PLC control system via the OPCUA protocol to drive the execution equipment—the blower uses an ABB ACS880 frequency converter, the dosing pump uses a plunger metering pump, and the electric regulating valve; Effect feedback and iteration: Real-time comparison of physical entity data and digital twin simulation results triggers model parameter correction, forming a closed loop of "acquisition-simulation-decision-execution-feedback".

[0065] Example 2:

[0066] Smart transformation projects for existing AAO process wastewater treatment plants;

[0067] Data acquisition layer upgrade: Sensor modification: Replace the old sensors and add 22 types of high-precision sensors—ORP sensors and sludge settling ratio online monitoring instruments are added to the bioreactor; vibration sensors and current and voltage monitoring modules are added to the equipment room; rainfall sensors are added to the inlet; Edge gateway deployment: Four edge gateways are added in key areas such as the bioreactor and blower room to achieve the fusion acquisition of "existing PLC data + new sensor data" and solve the traditional "information silo" problem; the gateway supports local logic judgment, such as directly triggering an emergency aeration command when DO is less than 0.5mg / L, with a response delay of less than 100ms;

[0068] Lightweight deployment of the digital twin platform; model adaptation and modification: a simplified digital twin model is built based on existing plant CAD drawings and laser scanning data to reduce computing power requirements; the physical model adopts the ASDM framework and combines historical operating data with quasi-kinetic parameters to ensure simulation errors are less than 8%; focused simulation functions: the "process optimization simulation" and "equipment maintenance simulation" modules are developed—process optimization simulation simulates the impact of different carbon source dosages on total nitrogen, and equipment maintenance simulation simulates the changes in DO distribution when aerator heads are blocked, verifying the effectiveness of replacement solutions in advance; incremental training of deep learning models. Training; Model transfer and fine-tuning: Based on the LSTM base model in the feasibility report, three years of historical data from the plant were imported for incremental training. Parameters were fine-tuned daily using newly added data, reducing the influent water quality prediction error from 12% to below 8% within one month; Targeted optimization functions: A "sludge bulking early warning sub-model" was developed, combining MLSS, SV30 data and the judgment criteria of "excessive proliferation of filamentous bacteria" in the process optimization document to provide an early warning of sludge bulking risk three days in advance; An "aeration energy consumption optimization sub-model" was developed, combining the blower power curve to dynamically adjust the frequency and avoid "over-aeration".

[0069] Control execution system integration; Protocol adaptation: Through the OPCUA protocol, the existing DCS system is integrated to achieve data interoperability between the "digital twin platform, deep learning model and existing control equipment" without replacing all hardware, thus reducing transformation costs.

[0070] Manual-Automatic Switching: Retains manual operation privileges, and automatically switches to manual control mode when the system issues an alert; sets "optimization parameter thresholds", such as triggering manual review when the carbon source dosage exceeds the design value by 120%, to ensure operational safety.

[0071] In summary, according to the figure, the wastewater treatment monitoring and management method and system based on digital twin disclosed in this invention, with "virtual-real linkage, data intelligence, and closed-loop control" as the core logic, constructs an intelligent management system covering the entire wastewater treatment process. At the methodological level, through six steps—"data acquisition, digital twin modeling, data interaction and synchronization, deep learning training, virtual simulation decision-making, and command execution feedback"—a complete closed loop is achieved, from the perception of physical entity parameters to the dynamic mapping of the virtual model, and then to the generation and precise execution of intelligent decisions. The data acquisition layer acquires multi-dimensional parameters of water quality, equipment, and environment according to their importance, ensuring data timeliness and completeness. The digital twin model recreates the physical water plant with 1:1 accuracy, and, combined with the ASDM mechanism model, accurately simulates biochemical reactions, solving the problems of "invisibility and difficulty in prediction" in traditional processes. The deep learning model cluster uses algorithms such as LSTM and CNN-LSTM to achieve water quality prediction, process optimization, and equipment fault early warning, providing scientific support for decision-making with an accuracy rate greater than 95% and 7-14 days in advance. The virtual simulation and command execution stages are driven by multi-scenario simulation and the OPCUA protocol, ensuring optimal control schemes and efficient implementation.

[0072] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0073] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A wastewater treatment monitoring and management method based on digital twins, characterized in that, Includes the following steps: Step 1: Data acquisition. Sensors deployed in various process stages of the wastewater treatment plant are used to collect real-time operational data of the physical entities. The operational data includes water quality parameters, equipment operating parameters, and environmental parameters. Step 2: Digital twin model construction. Based on the collected physical entity data, a digital twin model that maps to the physical wastewater treatment plant in a 1:1 ratio is constructed. The digital twin model includes a geometric model, a physical model, a behavioral model, and a rule model. Step 3: Data interaction and synchronization. Establish a real-time data interaction channel between the physical entity and the digital twin model, and transmit the collected operational data to the digital twin model in real time to achieve dynamic synchronization between the virtual model and the physical entity. Step 4: Deep learning model training and optimization. Based on historical operating data and real-time collected data, a deep learning model is trained. The deep learning model includes a water quality prediction model, a process parameter optimization model, and a fault early warning model. The model is periodically iterated and optimized using new data to ensure that the model prediction error is less than 8%. Step 5: Virtual simulation and decision generation. The wastewater treatment process is simulated using a digital twin model. Combined with the prediction results of the deep learning model, the impact of different process parameter adjustments on the treatment effect is simulated to generate the optimal process control decision. Step Six: Instruction Issuance and Execution. The generated optimal process control decision is transformed into control instructions and issued to the execution equipment in the physical wastewater treatment plant to achieve precise control of the wastewater treatment process. At the same time, the execution effect is monitored in real time, forming a closed-loop management of "collection-simulation-decision-execution-feedback".

2. The wastewater treatment monitoring and management method based on digital twins according to claim 1, characterized in that: In step one, the water quality parameters include COD, ammonia nitrogen, total nitrogen, total phosphorus, dissolved oxygen (DO), pH value, and sludge concentration (MLSS); the equipment operating parameters include water pump flow rate, blower frequency, dosing pump dosage, equipment vibration value, current, and voltage; the environmental parameters include water temperature, atmospheric temperature, humidity, and rainfall; the data acquisition frequency is set according to the importance of the parameters, with key water quality parameters acquired every 5 seconds, equipment operating parameters acquired every 10 seconds, and environmental parameters acquired every 30 seconds.

3. The wastewater treatment monitoring and management method based on digital twins according to claim 1, characterized in that: In step two, the geometric model is built using the Three.js engine, achieving millimeter-level modeling accuracy. It replicates the core process areas such as the bioreactor, sedimentation tank, and aeration system at a 1:1 scale, while non-critical areas are treated with lightweighting. The physical model is built based on the Activated Sludge Digestion Model (ASDM), containing 13 state variables and 20 kinetic parameters to accurately simulate biochemical reaction processes such as nitrification, denitrification, and biological phosphorus removal. The behavioral model uses Python scripts to write the equipment operation logic, defining the association rules between equipment start-up and shutdown, parameter adjustment, and process conditions. The rule model integrates constraints such as environmental standards and equipment operation and maintenance specifications, providing a basis for virtual simulation and decision generation.

4. The wastewater treatment monitoring and management method based on digital twins according to claim 1, characterized in that: In step three, the data interaction channel adopts an "edge computing + 5G industrial Ethernet" architecture. The edge gateway preprocesses the collected data, including filtering, outlier removal and data normalization. The preprocessed data is transmitted to the digital twin platform via 5G industrial Ethernet. When the network is interrupted, the edge gateway can cache data for 2 hours and automatically retransmit it after recovery. The data loss rate is less than 0.5%.

5. The wastewater treatment monitoring and management method based on digital twins according to claim 1, characterized in that: In step four, the water quality prediction model uses the LSTM algorithm, inputting 18 parameters such as influent flow rate, COD, ammonia nitrogen, and water temperature to predict water quality indicators for the next 1-24 hours; the process parameter optimization model uses a "CNN-LSTM" hybrid architecture to output the optimal values ​​of key parameters such as aeration rate and carbon source dosage, so that dissolved oxygen is stabilized at 1.5 mg / L and carbon source dosage accuracy is controlled at 5%. The fault early warning model, based on autoencoder and random forest algorithms, provides early warnings of faults in key equipment such as water pumps and blowers 7 to 14 days in advance, with an accuracy rate of over 95%.

6. The wastewater treatment monitoring and management method based on digital twins according to claim 1, characterized in that: In step five, the virtual simulation includes full-process simulation, extreme condition simulation, and equipment failure simulation. The full-process simulation presents the operating status and water quality changes of each process unit in real time. The extreme condition simulation can simulate more than 20 scenarios such as rainstorm impact and sudden changes in influent water quality, and predict the trend of treatment effect changes. The equipment failure simulation can simulate the impact of failures such as aeration head blockage and water pump failure on the process, and verify emergency treatment plans in advance. In the decision generation process, multiple objectives such as water quality compliance rate, energy consumption, and chemical consumption are comprehensively considered, and the optimal combination of process parameters is determined through multi-objective optimization algorithms.

7. The wastewater treatment monitoring and management method based on digital twins according to claim 1, characterized in that: In step six, control commands are sent to the PLC control system via the OPCUA protocol to drive the execution of equipment actions, including adjusting the blower frequency, the dosage of the dosing pump, and the sludge return ratio; the execution effect is monitored by comparing the simulation results of the digital twin model with the actual operating data of the physical entity to calculate the deviation value. If the deviation value exceeds the preset threshold, the model parameters are corrected or the control decision is regenerated.

8. A wastewater treatment monitoring and management system based on digital twins, characterized in that: include: The data acquisition module consists of multiple sensors and an edge gateway. The sensors are deployed in various process stages of the wastewater treatment plant to collect water quality parameters, equipment operating parameters, and environmental parameters. The edge gateway is used to preprocess and transmit the collected data. The digital twin platform module includes a geometric modeling module, a physical modeling module, a data interaction module, and a virtual simulation module. The geometric modeling module constructs a three-dimensional geometric model of the physical entity, the physical modeling module constructs a physical model of the biochemical reaction, the data interaction module realizes real-time data synchronization between the physical entity and the virtual model, and the virtual simulation module simulates and analyzes the wastewater treatment process. The deep learning module includes a model training module, a model optimization module, and a prediction and decision module. The model training module trains water quality prediction, process optimization, and fault early warning models based on historical and real-time data. The model optimization module iteratively updates the model parameters periodically. The prediction and decision module generates process control suggestions based on the model output. The control and execution module includes a PLC control system and an execution device. The PLC control system receives control commands from the digital twin platform and the deep learning layer, and drives the execution device to perform parameter adjustment actions. The visualization monitoring module uses Vue.js + Three.js to develop a 3D visualization interface that displays the real-time operating status of the digital twin model, trend curves of key parameters, and equipment fault warning information, and supports user interaction and data query.

Images