Intelligent operation management platform and method for tail water constructed wetland of large sewage plant

Abstract

The invention relates to the technical field of environmental engineering, in particular to an intelligent operation management platform and method for a tail water constructed wetland of a large sewage plant, and the operation management platform comprises a monitoring system device which is used for collecting wetland monitoring data in real time, and transmitting the collected wetland monitoring data to an analysis and storage module; the analysis storage module is used for analyzing the wetland monitoring data to obtain structured monitoring parameter data; the operation scheduling management module comprises a scheme management unit used for storing a plurality of regulation and control schemes; the operation control unit is used for regulating and controlling the regulation and control object equipment by adopting preset scheduling logic according to the current monitoring parameter data and the regulation and control scheme; the operation and maintenance management unit is used for sending an inspection work order to a mobile terminal of a corresponding operation and maintenance worker according to a preset inspection plan; and the evaluation module is also used for receiving the inspection result data reported and sent by the mobile terminal of the operation and maintenance personnel and evaluating the inspection result data by adopting a preset rule to obtain an evaluation result.

Description

Technical Field

[0001] The present application relates to the field of environmental engineering technology, and in particular to an intelligent operation and management platform and method for artificial wetlands in the tailwater of large-scale sewage treatment plants. Background Art

[0002] Against the backdrop of growing demands for water resource conservation and ecological governance, constructed wetlands, as a highly efficient, low-energy, and eco-friendly method for deep tailwater treatment, have been widely adopted in tailwater purification projects at large sewage treatment plants. In particular, in urban sewage treatment plants and centralized drainage systems in industrial parks, constructed wetlands not only purify tailwater but also, to a certain extent, serve as ecological buffers and enhance landscaping. To ensure operational effectiveness, timely regulatory responses, and ecological stability, real-time monitoring, precise scheduling, and intelligent management of constructed wetland operations have become key technical priorities for upgrading constructed wetland systems.

[0003] The traditional operation and management of artificial wetlands is still mainly based on manual inspections and regular sampling, and the monitoring methods are relatively backward. Although certain automated monitoring methods have been introduced in recent years, such as the installation of online water quality analyzers, liquid level meters and other equipment for data collection, most of this data is stored in a decentralized form and lacks unified protocol parsing and standardized structured processing capabilities, resulting in redundant data and difficulty in integration. Moreover, due to the wide variety of equipment communication protocols, the system is difficult to adapt flexibly, further limiting data interoperability and integration. In addition, at the control level, the current operation and adjustment of wetland systems rely heavily on manual experience and manual judgment of historical data, and lacks automatic control logic driven by real-time monitoring data. Once sudden water quality fluctuations or equipment abnormalities occur on site, problems such as response delays and slow control are relatively common.

[0004] On the other hand, traditional control methods mostly use static, fixed strategies and lack a library of multi-scenario, optional control solutions tailored to complex wetland structures and operating environments. Furthermore, most systems lack an automatic mapping mechanism between monitoring data and control strategies, making it impossible to intelligently match the optimal control strategy based on the current state. In terms of operations and maintenance management, while some platforms support mobile terminals receiving inspection tasks, these are mostly limited to scheduled task pushes and lack the ability to intelligently dispatch tasks based on real-time monitoring data. Furthermore, there is no automated closed-loop evaluation of inspection feedback data, leading to issues of "collection without use" or "use without accuracy."

[0005] In summary, existing constructed wetland operation and management systems commonly suffer from the following problems: inconsistent data collection protocols and weak parsing capabilities; a lack of structured, standardized data storage methods, which limits data value mining; a lack of control logic and solution matching mechanisms linked to current monitoring data, making remote automatic control impossible; delayed response times for operation and maintenance mechanisms, a disconnect between inspection work orders and equipment status, and a lack of evaluation and feedback. These issues severely restrict the advancement of the intelligent capabilities of constructed wetland systems and hinder the goal of "reduced staff, precise control, and efficient operation and maintenance." Summary of the Invention

[0006] (1) Technical issues to be resolved

[0007] In view of the above-mentioned shortcomings and deficiencies of the existing technology, the present application provides an intelligent operation and management platform and method for large-scale sewage treatment plant tailwater artificial wetlands, which solves the technical problems in the existing artificial wetland operation and management, such as inconsistent monitoring data collection protocols, chaotic data analysis and storage structures, delayed control response, inability to adaptively match control strategies, and imperfect operation and maintenance management mechanisms, and realizes a high degree of integration and intelligent improvement of real-time monitoring, structured processing, intelligent control and closed-loop operation and maintenance management of the wetland operation status.

[0008] (2) Technical solution

[0009] In order to achieve the above objectives, the main technical solutions adopted in this application include:

[0010] In a first aspect, the present invention provides an intelligent operation and management platform for large-scale sewage treatment plant tailwater artificial wetlands, including:

[0011] The monitoring system device includes a collection terminal deployed in a designated area of ​​the wetland, which is used to collect wetland monitoring data in real time and transmit the collected wetland monitoring data to an analysis and storage module via a 4G / 5G network or a dedicated line network;

[0012] The parsing and storage module is used to identify the communication protocol type used by the acquisition terminal, parse the wetland monitoring data in different protocol formats to obtain structured monitoring parameter data, and store the monitoring parameter data in a unified format;

[0013] Operation scheduling management module, including operation control unit, operation and maintenance management unit and solution management unit;

[0014] The scheme management unit is used to store multiple control schemes; wherein each adjustment scheme corresponds to monitoring parameter data at a historical moment;

[0015] An operation control unit is used to control the control target device using a preset scheduling logic based on the current monitoring parameter data in the analysis storage module and the control scheme in the scheme management unit, so as to realize remote control of the control target device;

[0016] The operation and maintenance management unit is used to send inspection work orders to the mobile terminals of corresponding operation and maintenance personnel according to the pre-set inspection plan or current monitoring parameter data; it is also used to receive the inspection result data reported by the mobile terminals of the operation and maintenance personnel in the process of executing the inspection work order, and evaluate the inspection result data using pre-set rules to obtain evaluation results.

[0017] Preferably, in some embodiments of the present application, the data collection terminals deployed in designated areas of the wetland include: water quality and water quantity monitoring equipment installed at the inlet and outlet of the wetland; dissolved oxygen sensors deployed in the front ponds and water collection channels of the wetland;

[0018] Liquid level gauges installed in vertical subsurface flow wetlands, horizontal subsurface flow wetlands, and stabilization ponds; chlorophyll monitors installed in stabilization ponds; and equipment operation data acquisition equipment installed in the wetland blower room, including equipment subject to regulation.

[0019] The equipment to be regulated includes pump stations and blowers;

[0020] The wetland monitoring data includes: water quality parameters, water volume parameters, dissolved oxygen parameters, water level parameters, chlorophyll content parameters, and operating status data of the regulated equipment; wherein the water quality parameters include COD, ammonia nitrogen, total phosphorus TP, and total nitrogen TN;

[0021] The monitoring parameter data is structured data obtained based on the analysis of wetland monitoring data, specifically including: structured water quality parameters, structured water quantity parameters, structured dissolved oxygen values, structured liquid level parameters, structured chlorophyll parameters and structured equipment operating status parameters. Each data item contains a timestamp and corresponding collection point identification information.

[0022] Preferably, in some embodiments of the present application, each control scheme is generated based on the monitoring parameter data of the corresponding historical moment and a pre-trained data-driven model according to a preset generation strategy, and the generation strategy includes the steps of:

[0023] A1. Input the preset control plan corresponding to the monitoring parameter data at any historical moment into a pre-trained data-driven model for simulation and deduction. The pre-trained data-driven model predicts the monitoring parameter data at a preset time interval after the implementation of the control plan.

[0024] Among them, the control plan includes: pump station start and stop sequence, blower operation intensity;

[0025] A2. If the predicted monitoring parameter data does not meet the set water quality standard conditions, the control plan modification instruction input by the user is received, and the control plan is adjusted according to the modification instruction to form a new control plan;

[0026] A3. Inputting the new control plan into the pre-trained data-driven model for further simulation and deduction to predict the monitoring parameter data at a preset time interval after the implementation of the new control plan;

[0027] A4. If the predicted new monitoring parameter data meets the set water quality standard conditions, the new control plan is confirmed as the control plan corresponding to the monitoring parameter data at the historical moment and stored;

[0028] If the water quality conditions are still not met, repeat steps A2 to A4 until a control plan that meets the water quality conditions is generated.

[0029] Preferably, in some embodiments of the present application, the data-driven model is a deep neural network model for simulating the effect of wetland operation regulation, the deep neural network model includes an input layer, multiple hidden layers and an output layer, the input of the data-driven model is the monitoring parameter data at a certain historical moment and the corresponding preset regulation plan, and the output of the data-driven model is the predicted value of the monitoring parameter data at a preset time interval under the corresponding regulation plan;

[0030] After the data-driven model is trained with the training data, a pre-trained data-driven model is obtained;

[0031] The training data of the deep neural network model includes monitoring parameter data collected at different historical moments under different operating conditions and the execution records of the corresponding control plans;

[0032] The deep neural network model is trained using a supervised training method, learning based on a large number of known input-output pairs, and optimizing the model's internal weight parameters to minimize the error between the predicted value and the actual observed value;

[0033] The judgment criteria for the completion of the deep neural network model training are: reaching the set error threshold on both the training set and the validation set, and the prediction error on the validation set is not greater than the preset error.

[0034] Preferably, in some embodiments of the present application, the operation control unit uses a pre-set scheduling logic to control the control target device based on the current monitoring parameter data in the parsing storage module and the control scheme in the scheme management unit to achieve remote control of the control target device; the scheduling logic specifically includes:

[0035] B1. Compare the current monitoring parameter data with the corresponding preset first threshold and second threshold to obtain a comparison result;

[0036] B2. When the comparison result shows that the current monitoring parameter data exceeds the first threshold but does not exceed the second threshold, a trend analysis operation is performed to predict whether there is a risk of exceeding the standard. If there is a risk of exceeding the standard, the corresponding preset preventive control measure information is pushed to the mobile terminal of the operation and maintenance personnel, so that the operation and maintenance personnel can perform operations based on the preventive control measure information;

[0037] B3. When the comparison result shows that the current monitoring parameter data exceeds the second threshold, the operation control unit issues an alarm message, calls a control plan corresponding to the current monitoring parameter data, and controls the control target device according to the control plan to achieve remote control of the control target device;

[0038] B4. After executing the control scheme, if the monitoring parameter data stored in the analysis storage module returns to within the set normal range within a preset time interval in the future, the control scheme is marked as valid and the data of the execution process is recorded;

[0039] B5. If the monitoring parameter data stored in the parsed storage module does not return to the set normal range within the future preset time interval, a control scheme adjustment instruction input by the user is received, the control scheme is adjusted according to the adjustment instruction to form a new control scheme, the operation control unit re-executes the new control scheme, and repeats B4-B5 until the monitoring parameter data stored in the parsed storage module returns to the normal range within the future preset time interval, thereby obtaining a final effective control scheme;

[0040] B6. Use the final effective control plan to control the control target device to achieve remote control of the control target device.

[0041] Preferably, in some embodiments of the present application, the trend analysis operation in the scheduling logic in step B2 specifically includes:

[0042] B21. Based on the current monitoring parameter data, extract historical monitoring parameter data within a preset time window before the current time point, where the preset time window includes the past several minutes;

[0043] B22. Smoothing the extracted historical monitoring parameter data to reduce the impact of occasional noise on trend judgment, wherein the smoothing method includes any one of a sliding average and an exponential smoothing method;

[0044] B23, constructing a trend analysis model based on the smoothed historical data, wherein the trend analysis model is any one of a linear regression model, an exponential fitting model, or a polynomial regression model;

[0045] B24. Predicting a change trend of the monitoring parameter within a preset prediction time period after the current time point using the trend analysis model, and obtaining a corresponding prediction value;

[0046] B25. Compare the predicted value with the second threshold value. If the predicted value will exceed the second threshold value within the predicted time period, determine that there is a risk of exceeding the standard.

[0047] Preferably, in some embodiments of the present application, the control scheme corresponding to the current monitoring parameter data called in step B3 is obtained through a preset calling strategy, and the preset calling strategy includes:

[0048] B31. When the current monitoring parameter data exceeds the second threshold, extracting historical monitoring parameter data corresponding to a plurality of stored historical control schemes from the scheme management unit;

[0049] B32. Calculating similarity between the current monitoring parameter data and the historical monitoring parameter data corresponding to each historical control scheme, wherein the similarity calculation method includes any one of Euclidean distance, cosine similarity, or Mahalanobis distance;

[0050] B33. Based on the similarity calculation result, determine the control scheme with the highest similarity to the current monitoring parameter data from the multiple historical control schemes as the control scheme corresponding to the current monitoring parameter data.

[0051] Preferably, in some embodiments of the present application, the operation management platform further includes:

[0052] An assessment and evaluation module, configured to perform periodic quantitative evaluations of wetland operation status, control effects, and equipment operation and maintenance based on the wetland monitoring data collected by the monitoring system device and the inspection data received by the operation and maintenance management unit;

[0053] The asset management module is used to uniformly manage the control object equipment and other operating equipment involved in the platform.

[0054] Preferably, in some embodiments of the present application, the assessment module includes:

[0055] A water quality assessment unit is used to automatically score and classify water quality based on the structured water quality parameters in the monitoring parameter data according to the set water quality evaluation standards;

[0056] The operation and maintenance performance unit is used to conduct statistical evaluations on indicators such as equipment maintenance frequency and response efficiency based on inspection results data, exception handling records, and equipment operating status data;

[0057] Trend analysis unit, used to perform trend modeling and change warning on assessment results within multiple cycles;

[0058] The rectification instruction unit is used to generate problem rectification tasks based on the assessment analysis results and send the corresponding task information to the mobile terminal of the operation and maintenance personnel.

[0059] Asset management module, including:

[0060] Asset entry unit, used to add, modify, import and classify the information of the controlled equipment;

[0061] Asset positioning unit, used to display the spatial distribution and basic attributes of various types of control objects in the GIS interface;

[0062] Operation and maintenance standard configuration unit, used to configure the operation, maintenance and repair standards of different types of regulated equipment;

[0063] The asset analysis unit is used to collect equipment information based on the type of equipment being regulated, operating status, and other conditions, and to export asset ledgers or operation and maintenance reports.

[0064] On the other hand, an embodiment of the present application also provides a method for intelligent operation of a large tailwater artificial wetland, which is executed by the above-mentioned intelligent operation and management platform for the tailwater artificial wetland of a large sewage treatment plant.

[0065] (3) Beneficial effects

[0066] The intelligent operation and management platform for artificial wetlands in the tail water of large-scale sewage treatment plants provided in this application fully reflects the "data-driven" intelligent operation concept in its structural design. Through the collection terminals deployed in various areas of the wetland, multi-dimensional wetland monitoring data such as water quality, water quantity, dissolved oxygen, liquid level, chlorophyll, etc. are obtained in real time, and the collected data are uploaded to the parsing and storage module through the 4G / 5G network or dedicated network. This module uniformly parses and structures the raw data in different protocol formats, and combines it with the timestamp and collection point identifier to form a complete structured monitoring parameter data system. The intelligent operation and management platform for artificial wetlands in the tail water of large-scale sewage treatment plants uses a pre-trained deep neural network model to model the response relationship between historical monitoring parameters and control plans, generate multiple control schemes, and store them in the scheme management unit, forming a data-driven control scheme library.

[0067] During operation, the operation control unit can call upon the historical control plan that best matches the current monitoring data in real time, implementing intelligent control logic that integrates empirical knowledge, model deduction, and a closed-loop data feedback loop. This approach, driven by monitoring data, generates and optimizes control plans, significantly improving the platform's adaptability and responsiveness to complex operating conditions. It effectively overcomes the problems of traditional scheduling, which relies on manual experience, has lags in control, and has a rigid operation, achieving a transition from "experience-based decision-making" to "data-based decision-making."

[0068] The operation control unit incorporated into this application's platform is its core innovation. Utilizing preset scheduling logic, it compares multiple threshold levels based on currently analyzed monitoring parameter data, enabling complex control functions such as anomaly warnings, trend prediction, and automatic selection and dynamic updating of control strategies. Specifically, when current monitoring parameter data exceeds a set threshold, this unit first predicts potential risks based on a trend analysis model. It then compares historical control solutions for similarity and automatically selects the optimal control solution. This control solution is then used to remotely adjust and control the wetland's pump stations, blowers, and other controlled equipment. Feedback from the adjusted monitoring data determines whether the results have returned to normal. If not, the solution is adjusted and iteratively optimized until an effective control solution is generated. This approach not only achieves closed-loop operational control but also forms an intelligent control process of "monitoring-judgment-control-feedback-optimization." Compared to existing technologies that rely on static thresholds or manual control, this application's intelligent operation and management platform for large-scale sewage treatment plant tailwater constructed wetlands possesses greater adaptability and intelligent decision-making capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 This is a structural diagram of an intelligent operation and management platform for tailwater artificial wetlands in a large sewage treatment plant according to one embodiment of the present application;

[0070] Figure 2 Schematic diagram of the structure of the operation scheduling management module according to one embodiment of the present application;

[0071] Figure 3 A flow chart of a generation strategy according to one embodiment of the present application;

[0072] Figure 4 The figure is a flowchart of the scheduling logic according to one embodiment of the present application. DETAILED DESCRIPTION

[0073] In order to better explain this application and facilitate understanding, the following detailed description of this application is given in conjunction with the accompanying drawings through specific implementation methods. In the related art, there are mainly three types of representative solutions for the operation monitoring and control management of large sewage plant tailwater artificial wetlands:

[0074] The first type of management approach is based on manual inspections and experience-based adjustments. This approach relies on on-site operations and maintenance personnel to regularly manually collect parameters such as water quality, water level, and dissolved oxygen in key areas of the wetland, and to adjust the start and stop of equipment such as blowers and pumping stations based on experience. This approach is labor-intensive, labor-intensive, and has delayed response times. It is difficult to promptly identify risks of water quality fluctuations, especially in the event of sudden climate change or abnormal water inflow. The optimal timing for adjustment can be missed, resulting in short-term excess tailwater quality and impacting discharge compliance.

[0075] The second category is a semi-automatic solution based on single-point automated data collection and fixed-logic control. This approach typically installs monitoring instruments at the wetland's inlet and outlet or in several areas, and uses hard-logic controllers such as PLCs to perform simple on-off operations or timing adjustments. However, this type of solution lacks overall data fusion and intelligent analysis capabilities, and the control strategy cannot dynamically adapt to the actual operating conditions on site. It also lacks an effective response mechanism to complex wetland structures and seasonal changes, and is prone to under- or over-regulation, resulting in low system efficiency, high energy consumption, and poor control accuracy.

[0076] The third category is remote monitoring solutions based on general SCADA systems and simple model pre-sets. While this solution achieves remote data transmission and centralized display to a certain extent, its data parsing capabilities are limited. It typically only supports fixed-format acquisition protocols and cannot accommodate the access needs of different brands of equipment and multiple sensor types. Furthermore, it relies on fixed control rules or manually set experience curves, making it incapable of dynamic optimization and automatic deduction based on actual operating data. Consequently, it performs poorly in areas such as sudden anomalies, water quality anomaly prediction, and iterative strategy optimization. Furthermore, such systems have low utilization rates of operational data, making it difficult to establish a comprehensive closed-loop assessment and evaluation system.

[0077] To address the above issues, this application provides a smart operation and management platform and method for large-scale sewage plant tailwater artificial wetlands. By building an integrated "monitoring-analysis-control-operation-maintenance-evaluation-management" closed-loop system, it breaks through the limitations of traditional technical solutions in data fusion, multi-source control, model optimization, and remote response. Specifically, it has the following technical advantages:

[0078] This application deploys multiple types of sensors (such as water quality, water level, DO, chlorophyll, etc.) at multiple key nodes in the wetland, and automatically identifies different protocols through the parsing storage module and uniformly converts them into structured parameters to achieve comprehensive access and standardized management of heterogeneous data.

[0079] A data-driven model constructed using deep neural networks is introduced, combining historical monitoring parameter data with control plans to automatically generate an optimized solution that meets the standards. It has the advantages of automatic deduction, high accuracy, and strong self-learning capabilities, breaking through the traditional reliance on experience-based settings. The control process supports hierarchical response and trend prediction. It not only issues an early warning before the monitoring value exceeds the standard, but also automatically matches the historical optimal strategy based on the similarity algorithm, achieving the unity of "proactive" and "targeted" control, significantly improving control accuracy and timeliness. The operation and maintenance module automatically generates work orders based on the inspection plan and current operating status, and conducts regular evaluation of the results, realizing the automatic recording, tracing, and assessment of operation and maintenance data, thereby improving operation and maintenance efficiency and quality. Through functions such as water quality scoring, performance analysis, and trend modeling, it supports periodic evaluation of control results and equipment status, and combines GIS spatial positioning and ledger analysis to achieve visual and refined management of wetland assets.

[0080] To sum up, the intelligent operation and management platform proposed in this application not only realizes full data collection, full equipment control, intelligent regulation, refined operation and maintenance, and closed-loop management, but also has strong versatility, high robustness and good adaptability. It can significantly improve the regulation efficiency and emission compliance stability of artificial wetlands in the tailwater of large sewage treatment plants, and meet the current high-standard environmental protection supervision and smart water development technical needs.

[0081] To better understand the above technical solutions, exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. Rather, these embodiments are provided to enable a clearer and more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.

[0082] As global ecological governance becomes increasingly prominent, many countries have long prioritized the ecological protection and restoration of urban water bodies and river wetlands, actively promoting wetland restoration and landscape reconstruction efforts focused on biodiversity conservation. Some countries have implemented "returning farmland to wetlands" projects in urban river wetlands, striving to enhance the self-purification capacity and environmental carrying capacity of wetland systems by restoring natural ecological processes.

[0083] In contrast, the development of artificial wetland systems in my country started relatively late. As important facilities for advanced sewage treatment and ecological regulation, artificial wetlands have been widely used in urban tailwater purification projects, playing an increasingly important role in ensuring that tailwater from large sewage treatment plants meets discharge standards. However, as the scale of artificial wetland applications continues to expand, problems in their operation and management are becoming increasingly prominent. These problems are primarily reflected in the reliance on manual experience for operational control, the use of single and extensive monitoring methods, the lack of dynamic adaptability of control strategies, and the low level of standardization in operation and maintenance management. These problems make it difficult to meet the comprehensive requirements for stable tailwater compliance, efficient and precise control, and intelligent management.

[0084] However, in the field of constructed wetlands, particularly those serving the advanced treatment of tailwater from large-scale sewage treatment plants, intelligent operations remain largely undeveloped. A comprehensive, integrated intelligent management loop integrating data monitoring, intelligent analysis, dynamic control, operation and maintenance linkage, and assessment and evaluation has yet to be established. Currently, constructed wetland operations rely primarily on manual control based on the on-site experience of operators and maintenance personnel. The startup and shutdown of key equipment, such as pumping stations and blowers, lacks data-driven and global optimization support, which can easily lead to delayed response times, high energy consumption, and unstable water quality. Furthermore, the complex ecological processes of wetland systems and their significant environmental impact on operational characteristics, coupled with the temporal and spatial heterogeneity of monitoring data and the nonlinear fluctuations in equipment status, further complicate management of traditional operation and maintenance models. Furthermore, existing engineering practices have shown that constructed wetland systems typically require two to three cycles of vegetation adaptation before achieving stable operation. A lack of systematic monitoring and meticulous management during this period can easily lead to imbalanced wetland functions and even render them useless. However, during the long-term operation stage after the wetland is built, due to the lack of standardized data analysis mechanisms, model deduction capabilities and unified regulatory logic, the operating efficiency of artificial wetlands is low, the water quality compliance rate fluctuates greatly, and it is difficult to achieve sustainable governance goals.

[0085] Therefore, to improve the operational efficiency of constructed wetlands in treating large-scale sewage plant tailwater and promote the transition from traditional "manual + experience-based" management to a "data-driven + intelligent regulation" model, it is necessary to build a comprehensive management platform that integrates sensory monitoring, structured analysis, strategic control, standardized operation and maintenance, and intelligent evaluation. By introducing IoT terminals, edge computing modules, deep learning models, and GIS spatial management systems, intelligent perception, real-time response, and remote joint control of the entire wetland process can be achieved, thus establishing a new intelligent operation and management model suitable for constructed wetland scenarios, providing a more stable, accurate, and sustainable solution for the in-depth treatment of sewage plant tailwater and ecological water replenishment. Figure 1 The figure is a schematic diagram of a large sewage plant tailwater artificial wetland intelligent operation and management platform according to one embodiment of the present application. Figure 1 As shown in the figure, the intelligent operation and management platform for the tailwater artificial wetland of the large sewage treatment plant includes:

[0086] The monitoring system device includes a collection terminal deployed in a designated area of ​​the wetland, which is used to collect wetland monitoring data in real time and transmit the collected wetland monitoring data to an analysis and storage module via a 4G / 5G network or a dedicated line network;

[0087] The data collection terminals deployed in designated wetland areas include: water quality and quantity monitoring equipment installed at the wetland inlet and outlet; dissolved oxygen sensors installed in the wetland front pond and water collection channel; liquid level gauges installed in the vertical subsurface flow wetland, horizontal subsurface flow wetland and stabilization pond; chlorophyll monitors installed in the wetland stabilization pond; equipment operation data collection equipment installed in the wetland blower room, among which the equipment installed in the wetland blower room includes the control object equipment;

[0088] The platform distributes key monitoring sensors for water quality, water quantity, liquid level, dissolved oxygen, and chlorophyll throughout multiple typical wetland process units (such as inlets and outlets, pre-ponds, subsurface wetlands, and stabilization ponds). Combined with data collected from equipment operating status, it forms a spatially comprehensive and multi-dimensional perception network. This improves the level of perceptibility of wetland operations, significantly enhancing the ability to capture anomalies, fluctuations, and changes in water quality. This ensures real-time dynamic data, ensuring the timeliness and scientific nature of regulation and operation and maintenance decisions.

[0089] The equipment to be regulated includes pump stations and blowers;

[0090] The wetland monitoring data includes: water quality parameters, water quantity parameters, dissolved oxygen parameters, water level parameters, chlorophyll content parameters, and operating status data of the regulated equipment; wherein the water quality parameters include COD (Chemical Oxygen Demand), ammonia nitrogen, total phosphorus TP, and total nitrogen TN;

[0091] The parsing and storage module is used to identify the communication protocol type used by the acquisition terminal, parse the wetland monitoring data in different protocol formats to obtain structured monitoring parameter data, and store the monitoring parameter data in a unified format;

[0092] The parsing and storage module in this embodiment can identify the data communication protocols of different acquisition devices (such as Modbus, OPC, and proprietary protocols) and perform data cleaning, formatting, and structured storage, ensuring the integration of data from multiple vendors and sources. This reduces the platform's dependence on device manufacturers and communication protocols, enables unified management of heterogeneous data, provides consistent input for control strategy formulation and model analysis, and enhances the platform's scalability and integrability.

[0093] The monitoring parameter data is structured data obtained based on the analysis of wetland monitoring data, specifically including: structured water quality parameters, structured water quantity parameters, structured dissolved oxygen values, structured liquid level parameters, structured chlorophyll parameters and structured equipment operating status parameters. Each data item contains a timestamp and corresponding collection point identification information.

[0094] See also Figure 2 , operation scheduling management module, including operation control unit, operation and maintenance management unit and solution management unit;

[0095] The scheme management unit is used to store multiple control schemes; wherein each adjustment scheme corresponds to monitoring parameter data at a historical moment;

[0096] An operation control unit is used to control the control target device using a preset scheduling logic based on the current monitoring parameter data in the analysis storage module and the control scheme in the scheme management unit, so as to realize remote control of the control target device;

[0097] In this embodiment, the operation control unit intelligently controls the start / stop, frequency, and intensity of key equipment such as blowers and pumping stations based on real-time monitoring parameters and preset strategies, establishing a closed "data-strategy-execution" loop. This enables online control and remote adjustment of the wetland's operational regulation process, avoiding manual delays and human intervention errors, and improving the dynamic adaptability of processing capacity and energy efficiency management.

[0098] The operation and maintenance management unit is used to send inspection work orders to the mobile terminals of corresponding operation and maintenance personnel according to the pre-set inspection plan or current monitoring parameter data; it is also used to receive the inspection result data reported by the mobile terminals of the operation and maintenance personnel in the process of executing the inspection work order, and evaluate the inspection result data using pre-set rules to obtain evaluation results.

[0099] In this embodiment, maintenance work orders can be dynamically generated based on real-time monitoring results and dispatched to mobile terminals via the platform, enabling on-demand dispatch. Inspection results can also be uploaded in real time, automatically evaluating inspection quality. This enables proactive, event-driven inspections, avoiding manual omissions, delays, or duplicate inspections, improving maintenance efficiency, reducing labor costs, and strengthening traceability mechanisms.

[0100] Optionally, in some embodiments of the present application, each control scheme is generated based on the monitoring parameter data of the corresponding historical moment and the pre-trained data-driven model according to a preset generation strategy, see Figure 3 , the generation strategy includes the steps of:

[0101] A1. Input the preset control plan corresponding to the monitoring parameter data at any historical moment into a pre-trained data-driven model for simulation and deduction. The pre-trained data-driven model predicts the monitoring parameter data at a preset time interval after the implementation of the control plan.

[0102] Among them, the control plan includes: pump station start and stop sequence, blower operation intensity;

[0103] A2. If the predicted monitoring parameter data does not meet the set water quality standard conditions, the control plan modification instruction input by the user is received, and the control plan is adjusted according to the modification instruction to form a new control plan;

[0104] A3. Inputting the new control plan into the pre-trained data-driven model for further simulation and deduction to predict the monitoring parameter data at a preset time interval after the implementation of the new control plan;

[0105] A4. If the predicted new monitoring parameter data meets the set water quality standard conditions, the new control plan is confirmed as the control plan corresponding to the monitoring parameter data at the historical moment and stored;

[0106] If the water quality conditions are still not met, repeat steps A2 to A4 until a control plan that meets the water quality conditions is generated.

[0107] For example, consider a large sewage treatment plant whose tailwater is treated in a constructed wetland before being discharged. At a specific historical moment (e.g., 08:00 on July 15, 2024), the following wetland monitoring parameter data were collected: COD: 42 mg / L (exceeding the limit; the limit is 40 mg / L); ammonia nitrogen: 6.3 mg / L (near the limit); total nitrogen (TN): 16 mg / L (above the limit); total phosphorus (TP): 1.0 mg / L (normal); dissolved oxygen (DO): 1.8 mg / L (low); liquid level: normal; blower operating frequency: 45 Hz; and pump station start / stop frequency: 30-minute intervals.

[0108] Step A1: The monitoring parameter data at that moment and the control plan at that time (blower 45Hz, pump station start and stop every 30 minutes) are input into a trained data-driven model (such as a deep neural network) to perform simulation and predict the "monitoring parameter value in the next hour":

[0109] The predicted results show that COD will drop to 40.5 mg / L and TN will be 15.8 mg / L, both still below the standard. This indicates that the current control plan cannot effectively treat the water quality.

[0110] Step A2: The platform notifies the user that the result does not meet the requirements, and the user attempts to enter a modification suggestion:

[0111] Adjustment 1: Increase the blower operating frequency to 55Hz (increase oxygenation intensity) and speed up the pump station start and stop frequency to every 20 minutes (accelerate water circulation) → form a new control plan.

[0112] Step A3: The model simulates the new control plan again, using "blower 55Hz + pump station 20-minute start and stop" as the new input for simulation:

[0113] Prediction results: COD = 39.2 mg / L, TN = 13.5 mg / L, DO rises to 3.2 mg / L → meets the standard.

[0114] Step A4: If the result meets the standard, the control scheme is confirmed to be effective and stored, and the "control parameter combination" is confirmed as a valid control scheme that matches the historical monitoring parameter data, and stored in the scheme library, marking the scheme as reusable when similar water quality parameters are used.

[0115] In this way, the effects of the control schemes are predicted in advance with the help of data-driven models, and a "virtual simulation" environment is constructed, which effectively avoids trial and error and waste of resources in actual operation. By simulating different control combinations, it is ensured that the selected control scheme is "evidence-based" in terms of compliance, and that key water quality indicators (COD, TN, etc.) are controlled within the regulatory red line. The data of each historical moment and its control scheme are systematically stored. When similar monitoring parameters appear next time, the historical effective schemes can be quickly retrieved directly through the scheme management unit to achieve strategy reuse and reduce decision-making time. The traditional subjective judgment based on human experience is transformed into an objective deduction based on models and data, reducing the impact of fluctuations in human quality on the control quality and adapting to more operation and maintenance personnel.

[0116] In some embodiments of the present application, the data-driven model is a deep neural network model for simulating the effect of wetland operation regulation, the deep neural network model includes an input layer, multiple hidden layers, and an output layer, the input of the data-driven model is monitoring parameter data at a certain historical moment and a preset regulation plan corresponding thereto, and the output of the data-driven model is a predicted value of the monitoring parameter data at a preset time interval under the corresponding regulation plan;

[0117] Specifically, the input layer of the data-driven model in this embodiment is used to receive input data of two dimensions: monitoring parameter data at historical moments, such as COD, ammonia nitrogen, TP, TN, DO (dissolved oxygen), liquid level, flow rate, chlorophyll, etc. at that moment; corresponding preset control plans, such as the start and stop sequence of the pump station, the blower operating frequency, the aeration duration, etc. within that period. The hidden layer is composed of multiple fully connected neuron layers. It can introduce activation functions (such as ReLU, Tanh) and normalization mechanisms according to the nonlinear characteristics of the control influence to enhance the model's ability to fit complex dynamic response relationships. The output layer outputs the predicted values ​​of the monitoring parameters predicted by the model after a preset time interval in the future (for example, COD = 28 mg / L after 60 minutes).

[0118] After the data-driven model is trained with the training data, a pre-trained data-driven model is obtained;

[0119] The training data of the deep neural network model includes monitoring parameter data collected at different historical moments under different operating conditions and the execution records of the corresponding control plans;

[0120] The deep neural network model is trained using a supervised training method, learning based on a large number of known input-output pairs, and optimizing the model's internal weight parameters to minimize the error between the predicted value and the actual observed value;

[0121] The judgment criteria for the completion of the deep neural network model training are: reaching the set error threshold on both the training set and the validation set, and the prediction error on the validation set is not greater than the preset error.

[0122] Traditional regulation relies primarily on empirical judgment or simple linear models, making it difficult to accurately predict the combined impact of various operations on wetland water quality. However, the deep neural network model employed in this application's embodiments possesses powerful nonlinear fitting and feature abstraction capabilities, effectively modeling the complex relationship between regulatory actions and environmental responses. This enables more realistic and accurate predictions of water quality responses, providing a scientific basis for program development and operational scheduling.

[0123] Preferably, in some embodiments of the present application, the operation control unit uses a preset scheduling logic to control the control object device according to the current monitoring parameter data in the parsing storage module and the control scheme in the scheme management unit, so as to realize remote control of the control object device; Figure 4 , the scheduling logic specifically includes:

[0124] B1. Compare the current monitoring parameter data with the corresponding preset first threshold and second threshold to obtain a comparison result;

[0125] B2. When the comparison result shows that the current monitoring parameter data exceeds the first threshold but does not exceed the second threshold, a trend analysis operation is performed to predict whether there is a risk of exceeding the standard. If there is a risk of exceeding the standard, the corresponding preset preventive control measure information is pushed to the mobile terminal of the operation and maintenance personnel, so that the operation and maintenance personnel can perform operations based on the preventive control measure information. In the actual application of this example, the trend analysis operation in the scheduling logic described in step B2 specifically includes:

[0126] B21. Based on the current monitoring parameter data, extract historical monitoring parameter data within a preset time window before the current time point, where the preset time window includes the past several minutes;

[0127] B22. Smoothing the extracted historical monitoring parameter data to reduce the impact of occasional noise on trend judgment, wherein the smoothing method includes any one of a sliding average and an exponential smoothing method;

[0128] B23, constructing a trend analysis model based on the smoothed historical data, wherein the trend analysis model is any one of a linear regression model, an exponential fitting model, or a polynomial regression model;

[0129] B24. Predicting a change trend of the monitoring parameter within a preset prediction time period after the current time point using the trend analysis model, and obtaining a corresponding prediction value;

[0130] B25. Compare the predicted value with the second threshold value. If the predicted value will exceed the second threshold value within the predicted time period, determine that there is a risk of exceeding the standard.

[0131] For example, first, step B21 uses the current monitoring parameter data as the time reference, and retroactively extracts the historical monitoring data within a preset time window before the current time point, such as the ammonia nitrogen concentration measurement value in the past 10 minutes, to form a set of data points with a time sequence. The selection of this data window can be flexibly set according to the dynamic change rate of different water quality parameters to ensure that trends can be captured while avoiding excessive extension of the analysis period. Then step B22, in order to reduce the interference of sudden noise data on trend judgment, a smoothing operation will be performed on the above historical data. Common methods include sliding average or exponential smoothing. For example, if the sliding average method is used, the average of every three adjacent data points is taken as the smoothing result, making the data curve smoother and more representative, thereby enhancing the stability and robustness of trend extraction. Then step B23, a trend analysis model is constructed based on the smoothed data. The model can be selected as a linear regression model, an exponential fitting model or a polynomial regression model. For example, when the short-term water quality fluctuations are approximately linear, linear regression can be used; if there is an accelerating trend in the parameter changes, quadratic or cubic polynomial regression can be selected; if there is an obvious increasing or decreasing attenuation characteristic in the changes, exponential fitting may be more reasonable. The optimal model will be automatically selected based on the data fitting error. Then in step B24, the selected trend analysis model is used to predict the trend of changes in the monitoring parameters within a predicted period of time after the current time point (such as the next 10 minutes), and the predicted value for each time point is output. For example, it is predicted that the ammonia nitrogen value may increase from the current 6.2 mg / L to 8.1 mg / L in the next 10 minutes. Finally, in step B25, the predicted value is compared with the set second threshold value (such as 8.0 mg / L). If any predicted value exceeds the threshold, it is determined that there is a risk of exceeding the standard, an early warning will be generated in time, and the preset preventive control measures (such as increasing the blower operating frequency, switching the flow path, etc.) will be triggered to intervene in the operation in advance.

[0132] On the one hand, by introducing short-term forecasting capabilities, the system no longer relies solely on static threshold judgments but instead possesses "forward-looking" decision-making capabilities, capable of identifying impending exceedance events and significantly improving the timeliness and accuracy of responses. On the other hand, through flexible data smoothing and model matching mechanisms, it enhances sensitivity to subtle trend changes in complex fluctuations and reduces false alarms and missed alarms. At the same time, this mechanism also provides a quantitative basis for decision-making for operations and maintenance personnel, making intervention measures more targeted and scientific, thereby comprehensively improving the intelligent management level of constructed wetlands and the ability to ensure water quality.

[0133] B3. When the comparison result shows that the current monitoring parameter data exceeds the second threshold, the operation control unit issues an alarm message and calls a control scheme corresponding to the current monitoring parameter data, and controls the control target device according to the control scheme to achieve remote control of the control target device; in this embodiment, the control scheme corresponding to the current monitoring parameter data called in step B3 is obtained through a pre-set calling strategy, and the pre-set calling strategy includes:

[0134] B31. When the current monitoring parameter data exceeds the second threshold, extracting historical monitoring parameter data corresponding to a plurality of stored historical control schemes from the scheme management unit;

[0135] B32. Calculating similarity between the current monitoring parameter data and the historical monitoring parameter data corresponding to each historical control scheme, wherein the similarity calculation method includes any one of Euclidean distance, cosine similarity, or Mahalanobis distance;

[0136] B33. Based on the similarity calculation result, determine the control scheme with the highest similarity to the current monitoring parameter data from the multiple historical control schemes as the control scheme corresponding to the current monitoring parameter data.

[0137] In order to ensure that the most suitable control scheme is selected quickly and accurately when the monitoring parameters exceed the standard, the operation control unit adopts a similarity matching-based strategy to call the control scheme in step B3. This strategy analyzes the similarity between the current monitoring parameter data and the parameter data in the historical scenario, and calls the most matching historical control scheme from the scheme management unit for rapid response, thereby improving the control effect and response efficiency. Specifically, when it is detected that a current monitoring parameter (such as ammonia nitrogen concentration) exceeds the second threshold, for example, the current ammonia nitrogen value is 9.2 mg / L, which is much higher than the set safety upper limit of 8.0 mg / L, step B31 in the calling strategy will be started to extract a set of stored historical control schemes from the scheme management unit. These control schemes are all associated with past historical monitoring data. For example, a historical scheme corresponds to when the ammonia nitrogen is 8.9 mg / L, the dissolved oxygen is 3.5 mg / L, the liquid level is 1.2 m, and the water volume is 800 m 3 / h. Next, in step B32, the current monitoring parameter data is similar to the monitoring parameter data corresponding to these historical control schemes. To achieve this goal, classic indicators such as Euclidean distance, cosine similarity or Mahalanobis distance can be used. For example, the Euclidean distance method is used to form a vector of the current monitoring parameters, and the distance calculation is performed item by item with the parameter vector of each historical scheme, so as to obtain a similarity score between each historical scheme and the current situation. Then, in step B33, based on the above similarity calculation results, the historical control scheme closest to the current monitoring parameters is selected. For example, if the Euclidean distance between the current parameters and a certain historical scheme is the smallest, then the historical scheme is considered to be the most suitable for the current situation, that is, this scheme is called to control equipment such as the start and stop of the pump station or the operating intensity of the blower to achieve a fast and accurate response.

[0138] This pre-set calling strategy can avoid the uncertainty brought by human experience selection and improve the scientificity and objectivity of control decisions in a data-driven manner; secondly, this strategy does not need to regenerate the control plan, but reuses historical plans, effectively reducing response delays and computing overhead, and is suitable for deployment in large-scale sewage treatment plant on-site environments with high timeliness requirements; finally, by introducing a variety of similarity index selection mechanisms, it can still maintain a strong matching ability under different data dimensions and distribution backgrounds, enhancing the versatility and robustness of control plan calling, thereby improving the level of intelligent operation and water quality compliance rate.

[0139] B4. After executing the control scheme, if the monitoring parameter data stored in the analysis storage module returns to within the set normal range within a preset time interval in the future, the control scheme is marked as valid and the data of the execution process is recorded;

[0140] B5. If the monitoring parameter data stored in the parsed storage module does not return to the set normal range within the future preset time interval, a control scheme adjustment instruction input by the user is received, the control scheme is adjusted according to the adjustment instruction to form a new control scheme, the operation control unit re-executes the new control scheme, and repeats B4-B5 until the monitoring parameter data stored in the parsed storage module returns to the normal range within the future preset time interval, thereby obtaining a final effective control scheme;

[0141] B6. Use the final effective control plan to control the control target device to achieve remote control of the control target device.

[0142] For example, for a current monitoring parameter data such as ammonia nitrogen, the first threshold is set to 5 mg / L and the second threshold is set to 8 mg / L. When the current ammonia nitrogen concentration provided by the analysis and storage module is 6 mg / L, that is, it is in the range of exceeding the first threshold but not exceeding the second threshold, the trend analysis module will be immediately triggered to predict future trends.

[0143] In this case, if the trend prediction results indicate that ammonia nitrogen levels may continue to rise and approach or exceed the second threshold, the system will automatically push pre-set preventive control measures (such as reducing the frequency of the blower or increasing the allocation of water to the wetland) to the operator's mobile device in the form of a prompt. This allows the operator to intervene in advance to prevent water quality deterioration.

[0144] If the current monitoring parameter data, such as ammonia nitrogen, exceeds the second threshold, for example, reaching 9 mg / L, it will no longer remain in the early warning stage, but will directly issue an over-standard alarm and immediately call the control plan corresponding to the abnormal situation from the plan management unit to automatically control the start and stop sequence of the pump station and the operating intensity of the blower, such as increasing the aeration volume to improve the ammonia nitrogen removal efficiency.

[0145] After the plan is executed, the recovery of relevant water quality indicators is continuously monitored within a preset time interval (such as 30 minutes). If the monitoring data shows that the ammonia nitrogen concentration has dropped to the set normal range (such as less than 5 mg / L), the current plan is judged to be valid, and its execution process and parameter change process are recorded for subsequent analysis and reuse. If the concentration still does not drop during the observation period, the control adjustment instructions input by the user will be received (such as extending the blowing time or increasing the dosage of the agent), and a new control plan will be generated and continued to be executed. This process is continuously iterated until the wetland water quality returns to normal and a final effective plan is formed.

[0146] This scheduling logic, based on a combination of multi-level threshold judgment, trend prediction, automatic control, and manual intervention, has significant benefits. First, it improves the constructed wetland's response speed and efficiency to sudden water quality anomalies, effectively preventing pollutant emissions from exceeding standards. Second, through a mechanism of trend prediction and dynamic solution iterative optimization, it implements an intelligent management and control logic based on "prevention first, response second," reducing the workload of operations and maintenance personnel and enhancing the platform's sustainable and stable operation capabilities.

[0147] Specifically, the operation management platform also includes:

[0148] An assessment and evaluation module, configured to perform periodic quantitative evaluations of wetland operation status, control effects, and equipment operation and maintenance based on the wetland monitoring data collected by the monitoring system device and the inspection data received by the operation and maintenance management unit;

[0149] The asset management module is used to uniformly manage the control object equipment and other operating equipment involved in the platform.

[0150] The assessment module includes:

[0151] A water quality assessment unit is used to automatically score and classify water quality based on the structured water quality parameters in the monitoring parameter data according to the set water quality evaluation standards;

[0152] The operation and maintenance performance unit is used to conduct statistical evaluations on indicators such as equipment maintenance frequency and response efficiency based on inspection results data, exception handling records, and equipment operating status data;

[0153] Trend analysis unit, used to perform trend modeling and change warning on assessment results within multiple cycles;

[0154] The rectification instruction unit is used to generate problem rectification tasks based on the assessment analysis results and send the corresponding task information to the mobile terminal of the operation and maintenance personnel.

[0155] Asset management module, including:

[0156] Asset entry unit, used to add, modify, import and classify the information of the controlled equipment;

[0157] Asset positioning unit, used to display the spatial distribution and basic attributes of various types of control objects in the GIS interface;

[0158] Operation and maintenance standard configuration unit, used to configure the operation, maintenance and repair standards of different types of regulated equipment;

[0159] The asset analysis unit is used to collect equipment information based on the type of equipment being regulated, operating status, and other conditions, and to export asset ledgers or operation and maintenance reports.

[0160] To achieve comprehensive management and continuous optimization of constructed wetlands, the intelligent operations management platform also includes an assessment and evaluation module and an asset management module. These two modules work together to periodically evaluate wetland performance and centrally manage and control hardware assets, such as equipment, further improving the overall platform's visualization, operational efficiency, and management accuracy.

[0161] Specifically, the assessment module integrates real-time water quality data collected by the monitoring system with inspection records collected by the operation and maintenance management unit to achieve periodic quantitative assessment of the wetland's operating status. The water quality assessment unit automatically analyzes structured water quality parameters (such as COD, ammonia nitrogen, total phosphorus, and total nitrogen) based on preset water quality scoring standards (such as the "Surface Water Environmental Quality Standard" GB 3838-2002) and calculates the scoring results. For example, after comparing the average ammonia nitrogen concentration in a wetland area over the past week with the standard, it is classified as "compliant" or "warning" to help managers quickly identify abnormal areas.

[0162] The Operations and Maintenance Performance Unit generates regular performance reports by statistically analyzing inspection work order execution, equipment exception handling records, and response times. For example, if a pump station experiences multiple anomalies within a month and these records are recorded, the unit can automatically determine that the failure frequency exceeds the standard and evaluate response efficiency based on maintenance timelines, thereby comprehensively assessing the work quality of the operation and maintenance team responsible for that equipment.

[0163] The trend analysis unit further enhances the forward-looking nature of assessments. It can construct time series models based on water quality assessment results and operational data from multiple consecutive cycles, enabling operational trend modeling and providing early warning of abnormal changes. For example, if the concentration of a particular pollutant shows an upward trend for three consecutive cycles, an alert will be issued, indicating a possible influent water quality issue.

[0164] Based on the above analysis results, the rectification instruction unit can automatically generate problem rectification tasks (such as "replace a certain model of blower impeller" or "strengthen the dosing management of the water inlet section") and send them to relevant operation and maintenance personnel through mobile terminals to ensure that the problems are handled in a timely closed-loop manner, significantly improving the closed-loop execution efficiency and standardization level of management.

[0165] At the same time, in order to achieve unified scheduling and efficient management of equipment resources, the asset management module has built an asset life cycle management system for the controlled equipment (such as pump stations, blowers, water quality instruments, etc.) and other auxiliary equipment (such as electrical cabinets, monitoring terminals, etc.).

[0166] Among them, the asset entry unit supports users to manually enter, batch import Excel data or automatically collect data through the platform to add and maintain equipment asset information, and can classify and manage it (such as by equipment type, installation location, usage status, etc.).

[0167] The asset location unit uses a GIS (Geographic Information System) interface to visualize equipment. Operations and maintenance personnel can view the real-time distribution, name, operating status, and other attributes of each device on a map, making spatial management and scheduling more intuitive. For example, clicking the blower icon for a stabilization pond in the GIS interface will pop up the device's operating status and maintenance records, improving the efficiency of equipment inspection and positioning.

[0168] The O&M Standards Configuration Unit is used to pre-set O&M cycles, inspection standards, maintenance methods, and maintenance material lists for different types of equipment. For example, a submersible pump can be scheduled for shaft seal inspection every 30 days and lubricant replacement every 90 days. Automatic reminders are generated to ensure the implementation of equipment maintenance policies.

[0169] Finally, the asset analysis unit supports comprehensive statistics of equipment assets according to multiple dimensions (such as operating status, manufacturer, installation time, etc.), and outputs equipment ledgers, failure rate analysis reports, asset distribution maps and other information to provide a basis for asset planning and scrapping assessment.

[0170] In summary, the assessment and evaluation module has established a scientific closed-loop operation management and evaluation system through water quality scoring, performance evaluation, trend warning and rectification instructions; while the asset management module has formed a full life cycle equipment asset supervision mechanism through comprehensive equipment data maintenance, visual display, standard configuration and information analysis.

[0171] In addition, the present application also provides a method for intelligent operation of a large tailwater artificial wetland, which is executed by the intelligent operation and management platform for the tailwater artificial wetland of a large sewage treatment plant described in the embodiment.

[0172] In summary, the large-scale tailwater artificial wetland intelligent operation and management platform and method of the embodiment of the present application builds a multi-dimensional monitoring system based on Internet of Things technology. It can accurately collect key indicators such as water quality, water quantity, dissolved oxygen, liquid level, chlorophyll, and equipment operating status in real time, forming a comprehensive portrait of the wetland's operating status, effectively avoiding the management blind spots caused by data lag and blind spots in traditional monitoring. Through a unified data transmission protocol and a secure and efficient Internet of Things platform, the interconnection and integrated management of multi-source heterogeneous monitoring data are achieved, ensuring the real-time and integrity of the data, and providing a solid data foundation for subsequent scheduling and operation and maintenance.

[0173] In response to the problems of complex operation and difficulty in regulating artificial wetlands due to multiple factors, this application has constructed an early warning mechanism and intelligent control process based on threshold classification, combined with operation control and periodic scheduling strategies, to achieve remote automatic control of the monitoring center and collaborative management of on-site operation and maintenance. Relying on the risk prediction and control solution library of early warning indicators, it can identify water quality fluctuation trends in advance and recommend corresponding control solutions, assisting operation and maintenance personnel to respond accurately and quickly, and significantly improving water quality compliance rate and wetland ecological safety. At the same time, by automatically generating and intelligently dispatching operation and maintenance work orders, closed-loop management and traceable execution of operation and maintenance work are achieved, ensuring the stable operation of equipment and facilities.

[0174] This application also innovatively integrates an assessment and evaluation module with an asset management module. Through GIS-based water quality cross-section monitoring and multi-dimensional performance assessment, it enables dynamic quantitative evaluation of water environment improvement effects and the operation and maintenance status of facilities and equipment, supporting scientific and transparent management decisions. The asset management module provides standardized information entry, classification management, and statistical analysis for various facilities and equipment, effectively improving asset management efficiency and maintenance levels while reducing management risks.

[0175] To sum up, the intelligent operation and management platform for large-scale sewage treatment plant tailwater artificial wetlands in this application, combined with multi-dimensional real-time monitoring, intelligent scheduling and control, scientific assessment and evaluation, and standardized asset management, has broken through the bottleneck of traditional artificial wetland management, and realized the intelligent, refined and efficient management of wetland operations, greatly improving the water quality assurance capability and operation and maintenance management level of the wetland, and providing solid technical support and guarantee for the deep treatment of sewage tailwater and ecological environment protection.

[0176] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A large-scale sewage treatment plant tailwater artificial wetland intelligent operation and management platform, characterized by: include: The monitoring system device includes a collection terminal deployed in a designated area of ​​the wetland, which is used to collect wetland monitoring data in real time and transmit the collected wetland monitoring data to an analysis and storage module via a 4G / 5G network or a dedicated line network; The parsing and storage module is used to identify the communication protocol type used by the acquisition terminal, parse the wetland monitoring data in different protocol formats to obtain structured monitoring parameter data, and store the monitoring parameter data in a unified format; Operation scheduling management module, including operation control unit, operation and maintenance management unit and solution management unit; The scheme management unit is used to store multiple control schemes; wherein each adjustment scheme corresponds to monitoring parameter data at a historical moment; An operation control unit is used to control the control target device using a preset scheduling logic based on the current monitoring parameter data in the analysis storage module and the control scheme in the scheme management unit, so as to realize remote control of the control target device; The operation and maintenance management unit is used to send inspection work orders to the mobile terminals of corresponding operation and maintenance personnel according to the pre-set inspection plan or current monitoring parameter data; it is also used to receive the inspection result data reported by the mobile terminals of the operation and maintenance personnel in the process of executing the inspection work order, and evaluate the inspection result data using pre-set rules to obtain evaluation results.

2. The intelligent operation and management platform for large-scale sewage treatment plant tailwater artificial wetlands according to claim 1, It is characterized by: in, The data collection terminals deployed in designated areas of the wetland include: water quality and water quantity monitoring equipment installed at the wetland inlet and outlet; dissolved oxygen sensors installed in the wetland front pond and water collection channel; liquid level gauges installed in the vertical subsurface flow wetland, horizontal subsurface flow wetland and stabilization pond; chlorophyll monitors installed in the wetland stabilization pond; equipment operation data collection equipment installed in the wetland blower room, among which the equipment installed in the wetland blower room includes the control object equipment; The equipment to be regulated includes pump stations and blowers; The wetland monitoring data includes: water quality parameters, water volume parameters, dissolved oxygen parameters, water level parameters, chlorophyll content parameters, and operating status data of the regulated equipment; wherein the water quality parameters include COD, ammonia nitrogen, total phosphorus TP, and total nitrogen TN; The monitoring parameter data is structured data obtained based on the analysis of wetland monitoring data, specifically including: structured water quality parameters, structured water quantity parameters, structured dissolved oxygen values, structured liquid level parameters, structured chlorophyll parameters and structured equipment operating status parameters. Each data item contains a timestamp and corresponding collection point identification information.

3. The intelligent operation and management platform for large sewage treatment plant tailwater artificial wetlands according to claim 2, It is characterized by: in, Each control scheme is generated based on the monitoring parameter data of the corresponding historical moment and the pre-trained data-driven model according to a preset generation strategy, which includes the following steps: A1. Input the preset control plan corresponding to the monitoring parameter data at any historical moment into a pre-trained data-driven model for simulation and deduction. The pre-trained data-driven model predicts the monitoring parameter data at a preset time interval after the implementation of the control plan. Among them, the control plan includes: pump station start and stop sequence, blower operation intensity; A2. If the predicted monitoring parameter data does not meet the set water quality standard conditions, the control plan modification instruction input by the user is received, and the control plan is adjusted according to the modification instruction to form a new control plan; A3. Inputting the new control plan into the pre-trained data-driven model for further simulation and deduction to predict the monitoring parameter data at a preset time interval after the implementation of the new control plan; A4. If the predicted new monitoring parameter data meets the set water quality standard conditions, the new control plan is confirmed as the control plan corresponding to the monitoring parameter data at the historical moment and stored; If the water quality conditions are still not met, repeat steps A2 to A4 until a control plan that meets the water quality conditions is generated.

4. The intelligent operation and management platform for large-scale sewage treatment plant tailwater artificial wetlands according to claim 3, It is characterized by: in, The data-driven model is a deep neural network model used to simulate the effect of wetland operation and regulation. The deep neural network model includes an input layer, multiple hidden layers, and an output layer. The input of the data-driven model is the monitoring parameter data at a certain historical moment and the corresponding preset regulation plan. The output of the data-driven model is the predicted value of the monitoring parameter data at a preset time interval under the corresponding regulation plan. After the data-driven model is trained with the training data, a pre-trained data-driven model is obtained; The training data of the deep neural network model includes monitoring parameter data collected at different historical moments under different operating conditions and the execution records of the corresponding control plans; The deep neural network model is trained using a supervised training method, learning based on a large number of known input-output pairs, and optimizing the model's internal weight parameters to minimize the error between the predicted value and the actual observed value; The judgment criteria for the completion of the deep neural network model training are: reaching the set error threshold on both the training set and the validation set, and the prediction error on the validation set is not greater than the preset error.

5. The intelligent operation and management platform for large sewage plant tailwater artificial wetlands according to claim 4 is characterized in that: The operation control unit adjusts the control target device using a preset scheduling logic based on the current monitoring parameter data in the parsing storage module and the control scheme in the scheme management unit, so as to achieve remote control of the control target device; The scheduling logic specifically includes: B1. Compare the current monitoring parameter data with the corresponding preset first threshold and second threshold to obtain a comparison result; B2. When the comparison result shows that the current monitoring parameter data exceeds the first threshold but does not exceed the second threshold, a trend analysis operation is performed to predict whether there is a risk of exceeding the standard. If there is a risk of exceeding the standard, the corresponding preset preventive control measure information is pushed to the mobile terminal of the operation and maintenance personnel, so that the operation and maintenance personnel can perform operations based on the preventive control measure information; B3. When the comparison result shows that the current monitoring parameter data exceeds the second threshold, the operation control unit issues an alarm message, calls a control plan corresponding to the current monitoring parameter data, and controls the control target device according to the control plan to achieve remote control of the control target device; B4. After executing the control scheme, if the monitoring parameter data stored in the analysis storage module returns to within the set normal range within a preset time interval in the future, the control scheme is marked as valid and the data of the execution process is recorded; B5. If the monitoring parameter data stored in the parsed storage module does not return to the set normal range within the future preset time interval, a control scheme adjustment instruction input by the user is received, the control scheme is adjusted according to the adjustment instruction to form a new control scheme, the operation control unit re-executes the new control scheme, and repeats B4-B5 until the monitoring parameter data stored in the parsed storage module returns to the normal range within the future preset time interval, thereby obtaining a final effective control scheme; B6. Use the final effective control plan to control the control target device to achieve remote control of the control target device.

6. The intelligent operation and management platform for large sewage treatment plant tailwater artificial wetlands according to claim 5 is characterized in that: in, The trend analysis operation in the scheduling logic in step B2 specifically includes: B21. Based on the current monitoring parameter data, extract historical monitoring parameter data within a preset time window before the current time point, where the preset time window includes the past several minutes; B22. Smoothing the extracted historical monitoring parameter data to reduce the impact of occasional noise on trend judgment, wherein the smoothing method includes any one of a sliding average and an exponential smoothing method; B23, constructing a trend analysis model based on the smoothed historical data, wherein the trend analysis model is any one of a linear regression model, an exponential fitting model, or a polynomial regression model; B24. Predicting a change trend of the monitoring parameter within a preset prediction time period after the current time point using the trend analysis model, and obtaining a corresponding prediction value; B25. Compare the predicted value with the second threshold value. If the predicted value will exceed the second threshold value within the predicted time period, determine that there is a risk of exceeding the standard.

7. The intelligent operation and management platform for large sewage plant tailwater artificial wetlands according to claim 6 is characterized in that: The control scheme corresponding to the current monitoring parameter data called in step B3 is obtained through a pre-set calling strategy, and the pre-set calling strategy includes: B31. When the current monitoring parameter data exceeds the second threshold, extracting historical monitoring parameter data corresponding to a plurality of stored historical control schemes from the scheme management unit; B32. Calculating similarity between the current monitoring parameter data and the historical monitoring parameter data corresponding to each historical control scheme, wherein the similarity calculation method includes any one of Euclidean distance, cosine similarity, or Mahalanobis distance; B33. Based on the similarity calculation result, determine the control scheme with the highest similarity to the current monitoring parameter data from the multiple historical control schemes as the control scheme corresponding to the current monitoring parameter data.

8. The intelligent operation and management platform for large sewage plant tailwater artificial wetlands according to claim 7 is characterized in that: The operation management platform also includes: An assessment and evaluation module, configured to perform periodic quantitative evaluations of wetland operation status, control effects, and equipment operation and maintenance based on the wetland monitoring data collected by the monitoring system device and the inspection data received by the operation and maintenance management unit; The asset management module is used to uniformly manage the control object equipment and other operating equipment involved in the platform.

9. The intelligent operation and management platform for large sewage plant tailwater artificial wetlands according to claim 8 is characterized in that: The assessment module includes: A water quality assessment unit is used to automatically score and classify water quality based on the structured water quality parameters in the monitoring parameter data according to the set water quality evaluation standards; The operation and maintenance performance unit is used to conduct statistical evaluations on indicators such as equipment maintenance frequency and response efficiency based on inspection results data, exception handling records, and equipment operating status data; Trend analysis unit, used to perform trend modeling and change warning on assessment results within multiple cycles; The rectification instruction unit is used to generate problem rectification tasks based on the assessment analysis results and send the corresponding task information to the mobile terminal of the operation and maintenance personnel. Asset management module, including: Asset entry unit, used to add, modify, import and classify the information of the controlled equipment; Asset positioning unit, used to display the spatial distribution and basic attributes of various types of control objects in the GIS interface; Operation and maintenance standard configuration unit, used to configure the operation, maintenance and repair standards of different types of regulated equipment; The asset analysis unit is used to collect equipment information based on the type of equipment being regulated, operating status, and other conditions, and to export asset ledgers or operation and maintenance reports.

10. A method for intelligent operation of a large tailwater artificial wetland, characterized in that: The method is executed by the intelligent operation and management platform for large-scale sewage treatment plant tailwater artificial wetlands described in any one of claims 1-9.

Images