Automatic regulation and control method and system for tidal flow constructed wetland based on AI

Abstract

The invention discloses an AI-based tidal flow constructed wetland automatic regulation control method and system, and the method comprises the steps: collecting multi-source sensing data of a tidal flow constructed wetland system in real time, and carrying out the preprocessing and feature extraction of the data; constructing a denitrification efficiency prediction model based on an LSTM network to determine a TN removal rate prediction value and an optimal DO threshold value; according to the TN removal rate predicted value and the optimal DO threshold value, a joint reward function is established based on a DRL algorithm, a multi-objective optimization strategy is determined, and adjustment parameters of the tidal flow constructed wetland system are dynamically generated; adjusting the water inlet / drainage frequency according to the adjustment parameters, accurately adding the carbon source, and adjusting the aeration intensity according to the gradient; and obtaining microbial activity feedback data, and optimizing the regulation and control strategy according to the microbial activity feedback data, so that the effluent TN reaches a preset standard. According to the method, a plurality of optimization targets can be considered at the same time, so that efficient, energy-saving and environment-friendly control on the constructed wetland is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial wetland environmental control, and particularly to an AI-based automated regulation and control method and system for tidal flow artificial wetlands. Background Art

[0002] In the modern water treatment process, it is often necessary to remove pollutants through the synergistic action of physics, chemistry, and biology. In particular, the removal of nitrogen depends on the balance of nitrification and denitrification processes. Traditional artificial wetland systems usually rely on manual monitoring and adjustment, which is cumbersome and inefficient. The AI-based automated regulation method can achieve real-time monitoring and intelligent control, reduce manual intervention, and improve the operation efficiency and stability of the system.

[0003] Currently, artificial intelligence (AI) technology has shown significant advantages in water quality prediction, parameter optimization, and real-time control. For example, the AI water service system of Asia Symbol achieves second-level response through high-precision sensors and dynamic algorithms. However, its application scenarios are limited to industrial aeration and chemical dosing, and do not cover the multi-dimensional regulation of wetland ecosystems. Existing research lacks the combination of AI and the bio-hydrodynamic coupling model of tidal flow wetlands, and adjusts the flooding / vacancy time according to real-time water quality data (such as NH4 + -N, NO3 - -N concentration) to balance the nitrification and denitrification requirements and achieve dynamic optimization of the tidal cycle. In addition, existing systems are difficult to adapt to complex environmental changes and lack the ability to make real-time adjustments to dynamic changes.

[0004] Therefore, it is necessary to provide an AI-based automated regulation and control method and system for tidal flow artificial wetlands. By using the LSTM network to predict the denitrification efficiency and considering the balance of energy consumption and carbon source dosage, the regulation strategy of the artificial wetland system is more comprehensive, capable of taking into account multiple optimization goals simultaneously, so as to achieve efficient, energy-saving, and environmentally friendly control of the artificial wetland. Summary of the Invention

[0005] In view of this, the present invention provides an AI-based automated regulation and control method and system for tidal flow artificial wetlands to solve the technical problem that the current artificial wetland control method is difficult to adapt to complex environmental changes and lacks the ability to make real-time adjustments to dynamic changes.

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

[0007] In the first aspect, the present invention provides an AI-based automated regulation and control method for tidal flow artificial wetlands, including:

[0008] Real-time collecting multi-source sensing data of the tidal flow artificial wetland system;

[0009] Preprocess and extract features from the multi-source sensing data to obtain key driving factors;

[0010] Build a denitrification efficiency prediction model based on the LSTM network, and determine the predicted value of TN removal rate and the optimal DO threshold according to historical operation data and key driving factors collected in real time;

[0011] According to the predicted value of TN removal rate and the optimal DO threshold, establish a combined reward function with the maximization of denitrification efficiency, the minimization of carbon source dosage, and the lowest energy consumption based on the DRL algorithm, determine the multi-objective optimization strategy, and dynamically generate the adjustment parameters of the tidal flow constructed wetland system;

[0012] Adjust the influent / effluent frequency through the electric gate controller according to the adjustment parameters, accurately add carbon source through the peristaltic pump and slow-release carbon source, and adjust the aeration intensity in gradient according to the predicted value of DO concentration;

[0013] Obtain the microbial activity feedback data, optimize the control strategy according to the microbial activity feedback data, and make the effluent TN reach the preset standard.

[0014] Further, the multi-source sensing data includes: water quality sensing data, environmental perception data, and biofilm activity monitoring data;

[0015] The water quality parameters include NH4 + -N, NO3 - -N, DO, pH, COD, temperature, flow rate, and water level data, which are used to characterize the health status of the water body and the changes of pollutants;

[0016] The environmental perception data includes light intensity, rainfall, and wind speed, which are used to characterize the impact of the external climate environment on water quality changes;

[0017] The biofilm activity monitoring data includes the metabolic activity of denitrifying bacteria and the abundance of nitrification / denitrification functional genes obtained by dynamic monitoring, which are used to characterize the impact of the microbial community in the water body on water quality.

[0018] Further, preprocessing and feature extraction of the multi-source sensing data to obtain key driving factors include:

[0019] Normalize the multi-source sensing data and extract candidate regulation and control features;

[0020] Use the principal component analysis method and the mutual information screening method to construct a denitrification efficiency driving factor matrix based on the candidate regulation and control features, and obtain the key driving factors affecting denitrification efficiency.

[0021] Furthermore, a denitrification efficiency prediction model is constructed based on the LSTM network. According to the historical operation data and the key driving factors collected in real time, the predicted value of the TN removal rate and the optimal DO threshold are determined, including:

[0022] Using water quality sensing data, environmental perception data, and biofilm activity monitoring data as the input of the LSTM network;

[0023] Construct a time series input through a preset time window, use bidirectional LSTM to capture the temporal dependencies before and after, and output the predicted value of the TN removal rate and the DO threshold for a preset future period.

[0024] Furthermore, according to the predicted value of the TN removal rate and the optimal DO threshold, a joint reward function for maximizing denitrification efficiency, minimizing carbon source dosage, and minimizing energy consumption is established based on the DRL algorithm, and a multi-objective optimization strategy is determined to dynamically generate the adjustment parameters of the tidal flow constructed wetland system, including:

[0025] Taking the predicted value of the TN removal rate and the optimal DO threshold output by the LSTM network as the extended state, and jointly constituting the multi-dimensional state space of DRL with the real-time parameters;

[0026] Taking the tidal cycle adjustment rate, aeration intensity grading, and carbon source feeding rate as the action space;

[0027] Deeply couple the multi-dimensional state space, action space, and joint reward function of DRL to form a closed-loop optimization mechanism.

[0028] Furthermore, the reward function is determined according to the joint optimization of denitrification efficiency, carbon source consumption, and aeration energy consumption, and the strategy is iteratively updated through the Q-learning algorithm.

[0029] On the other hand, the present invention also provides a tidal flow constructed wetland system, including a sensing layer, an AI decision-making layer, an execution layer, and a wetland function layer connected in sequence;

[0030] The sensing layer is used to collect multi-source sensing data of the tidal flow constructed wetland system in real time;

[0031] The AI decision-making layer is used to preprocess and extract features from the multi-source sensing data to obtain key driving factors, construct a denitrification efficiency prediction model based on the LSTM network, and determine the predicted value of the TN removal rate and the optimal DO threshold according to the historical operation data and the key driving factors collected in real time; according to the predicted value of the TN removal rate and the optimal DO threshold, establish a joint reward function for maximizing denitrification efficiency, minimizing carbon source dosage, and minimizing energy consumption based on the DRL algorithm, determine a multi-objective optimization strategy, and dynamically generate the adjustment parameters of the tidal flow constructed wetland system;

[0032] The execution layer is used to adjust the water intake / drainage frequency through an electric gate controller according to the adjustment parameters, precisely add carbon sources through a peristaltic pump and a slow-release carbon source, and adjust the aeration intensity in gradients according to the predicted DO concentration;

[0033] The wetland function layer adopts a layered filler optimization layer and a plant root oxygenation structure to construct a multi-path denitrification system.

[0034] Further, the layered filler optimization layer of the wetland function layer includes an upper aerobic zone and a lower anoxic zone; the upper aerobic zone is filled with a composite of volcanic rock and iron-carbon composite material to strengthen the synergistic effect of nitrifying bacteria enrichment and chemical oxidation; the lower anoxic zone is filled with a mixture of biochar and sulfur autotrophic denitrification substrate to promote the coupling of sulfur-driven denitrification and biochar adsorption;

[0035] The plant root bionic oxygenation structure realizes the oxygen supply function by planting emergent plants.

[0036] Further, the sensing layer includes a water quality sensing module, an environmental perception module, and a biofilm activity monitoring module;

[0037] The water quality sensing module includes multiple sensors, and the multiple sensors are arranged in the water distribution tank, the layered filler area, and the collection tank to collect water quality data in real time;

[0038] The environmental perception module is used to collect data on light intensity, rainfall, and wind speed;

[0039] The biofilm activity monitoring module is used to collect data on the abundance of microbial functional genes based on electrochemical impedance spectroscopy and qPCR technology.

[0040] Further, the execution layer includes a tidal gate controller, a gradient aeration system, and a carbon source dual-mode dosing device;

[0041] The tidal gate controller is used to adjust and control the water intake flow;

[0042] The gradient aeration system is used to control the aeration intensity in different areas to optimize the dissolved oxygen level in the water body;

[0043] The carbon source dual-mode dosing device is used to dose organic carbon sources to support the nitrogen removal process of microorganisms.

[0044] Compared with the prior art, the AI-based automated regulation and control method and system for tidal flow constructed wetlands proposed by the present invention have the following advantages:

[0045] (1) By using the LSTM network to train historical data and real-time sensing data, the temporal characteristics of water quality changes can be effectively captured, and the removal rate of TN (total nitrogen) and the DO (dissolved oxygen) threshold can be accurately predicted, making the prediction of denitrification efficiency more accurate.

[0046] (2) Through optimization based on the DRL algorithm, not only the improvement of denitrification efficiency is considered, but also the carbon source dosage and energy consumption are taken as optimization goals. While maximizing the denitrification efficiency, it can reduce the excessive use of carbon source and energy consumption, ensuring the dual optimization of environmental and economic benefits. Through the precise dosing of peristaltic pumps and slow-release carbon sources, the addition amount of carbon source can be accurately controlled, avoiding carbon source waste and reducing operating costs.

[0047] (3) By obtaining microbial activity feedback data, the control strategy can be further adjusted. This enables the system to have a powerful self-optimization ability, capable of adjusting the water treatment strategy according to the actual operation conditions to ensure that the effluent TN meets the preset standards.

[0048] In summary, through the closed-loop mechanism of "data-driven decision-making" and "feedback-driven learning", the present invention realizes the precise control of the addition amount of carbon source, avoids carbon source waste, reduces operating costs, and improves the long-term operation stability of the equipment, providing an efficient, low-carbon and reliable solution for the sewage treatment field. Brief Description of the Drawings

[0049] Figure 1 It is a schematic flow chart of the AI-based automatic regulation and control method for tidal flow constructed wetlands provided by the present invention;

[0050] Figure 2 It is a flow chart for the establishment and dynamic optimization of the prediction model provided by the present invention;

[0051] Figure 3 It is a schematic structural diagram of the AI-based automatic regulation and control system for tidal flow constructed wetlands provided by the present invention;

[0052] Figure 4 It is a schematic diagram of the connection relationship of each part of the system provided by the present invention;

[0053] In the figure, 1 - water inlet, 2 - connecting wire, 3 - electric water inlet valve, 4 - bracket, 5 - light sensor, 6 - rain gauge, 7 - anemometer, 8 - leaf photosynthesis detector, 9 - water distribution pool, 10 - water quality sensor at water inlet, 11 - carbon addition device, 12 - water inlet pump, 13 - water quality sensor in upper packing area, 14 - biological activity probe in upper packing area, 15 - water quality sensor in lower packing area, 16 - biological activity probe in lower packing area, 17 - AI decision-making layer, 18 - emergent plants, 19 - upper packing, 20 - microporous aeration pipe network, 21 - aeration hole, 22 - lower packing, 23 - solar aerator, 24 - water outlet pump, 25 - water quality sensor at water outlet, 26 - collection pool, 27 - electric water outlet valve, 28 - water outlet. Specific implementation manner

[0054] The following will specifically describe the preferred embodiments of the present invention in conjunction with the accompanying drawings. Among them, the accompanying drawings form a part of this application and are used together with the embodiments of the present invention to explain the principle of the present invention, rather than to limit the scope of the present invention.

[0055] Embodiment 1

[0056] Please refer to Figure 1 , this embodiment provides an AI-based automatic regulation and control method for tidal flow constructed wetlands, including:

[0057] Step S101: Real-time collect multi-source sensing data of the tidal flow constructed wetland system;

[0058] Step S102: Preprocess and extract features from the multi-source sensing data to obtain key driving factors;

[0059] Step S103: Build a denitrification efficiency prediction model based on the LSTM network, and determine the predicted value of TN removal rate and the optimal DO threshold according to historical operation data and key driving factors collected in real time;

[0060] Step S104: According to the predicted value of TN removal rate and the optimal DO threshold, establish a joint reward function with maximizing denitrification efficiency, minimizing carbon source dosage, and minimizing energy consumption based on the DRL algorithm, determine a multi-objective optimization strategy, and dynamically generate adjustment parameters for the tidal flow constructed wetland system;

[0061] Step S105: Adjust the inlet / drainage frequency through an electric gate controller according to the adjustment parameters, accurately add carbon source through a peristaltic pump and a slow-release carbon source, and adjust the aeration intensity in gradient according to the predicted value of DO concentration;

[0062] Step S106: Obtain microbial activity feedback data, and optimize the regulation strategy according to the microbial activity feedback data to make the effluent TN reach the preset standard.

[0063] In the method of this embodiment, the LSTM network is used to train historical data and real-time sensing data, which can effectively capture the temporal characteristics of water quality changes, accurately predict the TN (total nitrogen) removal rate and DO (dissolved oxygen) threshold, and make the prediction of denitrification efficiency more accurate. Through optimization based on the DRL algorithm, not only the improvement of denitrification efficiency is considered, but also the carbon source dosage and energy consumption are taken as optimization objectives, which can maximize the denitrification efficiency while reducing the excessive use of carbon source and energy consumption, ensuring the dual optimization of environmental and economic benefits. Through the precise dosing of the peristaltic pump and the slow-release carbon source, the addition amount of the carbon source can be accurately controlled, avoiding carbon source waste and reducing the operating cost. By obtaining the microbial activity feedback data, the regulation strategy can be further adjusted. This enables the system to have a strong self-optimization ability, capable of adjusting the water treatment strategy according to the actual operation situation to ensure that the effluent TN meets the preset standard.

[0064] As a preferred embodiment, in step S101, the multi-source sensing data includes: water quality sensing data, environmental perception data, and biofilm activity monitoring data;

[0065] The water quality parameters include NH4 + -N, NO3 - -N, DO, pH, COD, temperature, flow rate, and water level data, which are used to characterize the health status of the water body and the changes of pollutants;

[0066] The environmental perception data includes light intensity, rainfall, and wind speed, which are used to characterize the impact of the external climate environment on water quality changes;

[0067] The biofilm activity monitoring data includes the metabolic activity of denitrifying bacteria and the abundance of nitrification / denitrification functional genes obtained by dynamic monitoring, which are used to characterize the impact of the microbial community in the water body on water quality.

[0068] As a preferred embodiment, in step S102, the multi-source sensing data is preprocessed and feature extracted to obtain key driving factors, including:

[0069] The multi-source sensing data is normalized, and candidate regulation and control features are extracted;

[0070] Using the principal component analysis method and the mutual information screening method, a denitrification efficiency driving factor matrix is constructed based on the candidate regulation and control features to obtain the key driving factors affecting denitrification efficiency.

[0071] As a preferred embodiment, in step S103, a denitrification efficiency prediction model is constructed based on the LSTM network. According to the historical operation data and the key driving factors collected in real time, the predicted value of the TN removal rate and the optimal DO threshold are determined, including:

[0072] Use water quality sensing data, environmental perception data, and biofilm activity monitoring data as the input of the LSTM network;

[0073] Construct a time series input through a preset time window, use bidirectional LSTM to capture the temporal dependencies before and after, and output the predicted values of TN removal rate and DO threshold for a preset future period.

[0074] As a preferred embodiment, in step S104, according to the predicted value of TN removal rate and the optimal DO threshold, establish a joint reward function based on the DRL algorithm to maximize denitrification efficiency, minimize carbon source dosage, and minimize energy consumption, determine the multi-objective optimization strategy, and dynamically generate the adjustment parameters of the tidal flow constructed wetland system, including:

[0075] Use the predicted value of TN removal rate and the optimal DO threshold output by the LSTM network as the extended state, and jointly form the multi-dimensional state space of DRL with the real-time parameters;

[0076] Use the tidal cycle adjustment rate, aeration intensity grading, and carbon source dosing rate as the action space;

[0077] Deeply couple the multi-dimensional state space, action space, and joint reward function of DRL to form a closed-loop optimization mechanism.

[0078] As a preferred embodiment, the reward function is determined according to the joint optimization of denitrification efficiency, carbon source consumption, and aeration energy consumption, and the strategy is iteratively updated through the Q-learning algorithm.

[0079] As a specific embodiment, in the LSTM water quality prediction model: the input layer includes 12-dimensional real-time parameters (NH4 + -N, NO3 - -N, DO, pH, COD, temperature, flow rate, water level, light, rainfall, wind speed, functional gene abundance). In the optimization engine based on deep reinforcement learning (DRL), the predicted values of TN removal rate and DO threshold suggestions for the next 2 hours output by the dynamic prediction model are deeply coupled with the state space, action space, and reward function of DRL to form a closed-loop optimization mechanism.

[0080] Specifically, the future 2-hour total nitrogen (TN) removal rate and dissolved oxygen (DO) threshold recommendations output in real time by the LSTM dynamic prediction model are used as extended state variables to input into the DRL framework, and together with real-time states such as water quality parameters, meteorological data, operating parameters, and microbial gene abundance, a multi-dimensional state space is constructed; the action space is designed to cover tidal cycle adjustment (±10%), aeration intensity grading (0-5 levels), and carbon source dosing rate (0-20 mL / min). The reward function is constructed to jointly optimize the denitrification efficiency (TN removal rate weight 0.6), carbon source consumption (weight -0.3), and aeration energy consumption (weight -0.1), and the strategy is iteratively updated through the Q-learning algorithm.

[0081] In some embodiments, the method further includes strengthening the pre-set emergency plan for transfer learning. By analyzing historical data, a model for sudden high ammonia nitrogen scenarios is constructed, and an intelligent control plan is generated in advance (such as dynamically adjusting the tidal frequency to enhance the water body reoxygenation efficiency, accurately starting and stopping the emergency carbon source library to balance the carbon-nitrogen ratio), and the parameters of the plan are dynamically optimized in combination with transfer learning technology. The method triggers pre-control by predicting the water quality fluctuation trend, reduces the redundant operation of the aeration system, synchronously optimizes the emergency chemical dosing strategy, significantly reduces energy consumption and resource consumption while efficiently reducing the ammonia nitrogen impact load, shortens the abnormal condition recovery time, and finally realizes the comprehensive improvement of the shock resistance ability and refined management and control of the sewage treatment process.

[0082] Specifically, when the mutation of the influent NH4 + -N concentration exceeds the threshold (>30 mg / L), the transfer learning module calls the historical plan library to generate an emergency control strategy (such as shortening the tidal cycle to 0.5 hours and starting the pulsed dosing of slow-release carbon source). Based on the real-time biofilm activity data (EIS impedance change rate >15%), the plan parameters are adjusted through online learning to ensure that the system quickly and dynamically corrects and gradually returns to the steady state.

[0083] In addition, to improve the reliability, efficiency, and stability of the system, reduce operating costs and equipment failures, sensor calibration (standard solution calibration), aeration disk backwashing (high-pressure water gun pulse), and electric valve sensitivity inspection (based on water level change monitoring) are automatically performed monthly. The operation data is uploaded to the cloud quarterly, the global AI model is updated through learning, and then sent to the local edge terminal.

[0084] It should be noted that the slow-release carbon source mentioned in step S105 refers to a substance that can gradually release carbon source for the utilization of denitrifying bacteria, and is used to improve the denitrification effect. Specifically, common slow-release carbon sources include: acetate, polymer-based slow-release carbon sources, organic waste, propionate, and so on.

[0085] As Figure 2 shown, Figure 2 shows the flow chart of the establishment and dynamic optimization of the prediction model of the method.

[0086] Example 2

[0087] An embodiment of the present invention provides a tidal flow constructed wetland system, including a sensing layer, an AI decision-making layer, an execution layer, and a wetland function layer connected in sequence;

[0088] The sensing layer is used to collect multi-source sensing data of the tidal flow constructed wetland system in real time;

[0089] The AI decision-making layer is used to preprocess and extract features from the multi-source sensing data to obtain key driving factors, construct a denitrification efficiency prediction model based on the LSTM network, determine the predicted value of TN removal rate and the optimal DO threshold according to historical operation data and key driving factors collected in real time; according to the predicted value of TN removal rate and the optimal DO threshold, establish a joint reward function with the maximization of denitrification efficiency, the minimization of carbon source dosage, and the lowest energy consumption based on the DRL algorithm, determine the multi-objective optimization strategy, and dynamically generate adjustment parameters of the tidal flow constructed wetland system;

[0090] The execution layer is used to adjust the water inlet / drainage frequency through an electric gate controller according to the adjustment parameters, accurately add carbon source through a peristaltic pump and a slow-release carbon source, and adjust the aeration intensity in gradient according to the predicted DO concentration;

[0091] The wetland function layer adopts a layered filler optimization layer and a plant root oxygenation structure to construct a multi-path denitrification system.

[0092] The system of this embodiment constructs an "intelligent perception - dynamic decision-making - precise execution - function enhancement" integrated system, and realizes the dual improvement of denitrification efficiency and operation energy efficiency through multi-dimensional collaborative optimization.

[0093] As a preferred embodiment, as Figure 3 shown Figure 3 shows the actual structure diagram of the system of this embodiment. The layered filler optimization layer of the wetland function layer includes an upper aerobic zone and a lower anoxic zone; volcanic rock (particle size 10 - 20 mm) and iron-carbon composite material (Fe 30 / C mass ratio 1:2) are filled in the upper aerobic zone, with a thickness of 40 cm. Biochar (specific surface area ≥ 800 m 2 / g) and sulfur autotrophic denitrification substrate (sulfur powder / limestone volume ratio 3:1) are filled in the lower anoxic zone, with a thickness of 60 cm. The plant root bionic oxygenation structure achieves the function of oxygen supply by planting emergent plants.

[0094] Emergent plants are planted in the plant root oxygenation structure to provide oxygen, and cooperate with an aerator to meet the metabolic needs of denitrifying bacteria.

[0095] Through the synergistic effect of hierarchical packing optimization (volcanic rock / iron-carbon + biochar / sulfur matrix) and bionic aeration structure, a multi-path nitrogen removal system of chemical oxidation-biological adsorption-sulfur autotrophic denitrification is constructed, significantly improving the TN removal rate and effectively enhancing the contribution rate of simultaneous nitrification and denitrification.

[0096] In the upper packing area, the volcanic rock-iron-carbon composite packing strengthens the enrichment ability of nitrifying bacteria, and in the lower packing area, the biochar-sulfur matrix combination improves the resistance to water quality fluctuations, ensuring the stable operation of the system under water quality fluctuations.

[0097] As a preferred embodiment, the sensing layer includes a water quality sensing module, an environmental perception module, and a biofilm activity monitoring module;

[0098] As Figure 3 shown, the water quality sensing module includes multiple sensors, and the multiple sensors are arranged in the water distribution tank, hierarchical packing area, and collection tank of the wetland functional layer for real-time collection of water quality data;

[0099] The environmental perception module is used to collect data on light intensity, rainfall, and wind speed;

[0100] The biofilm activity monitoring module is used to collect data on the abundance of microbial functional genes based on electrochemical impedance spectroscopy and qPCR technology.

[0101] In some embodiments, the water quality sensing module real-time collects water quality and environmental data by arranging multi-parameter high-precision online sensors, and dynamically adjusts the sampling frequency (5 - 30 minutes / time). Through the high-precision sensor array, NH4 + -N, NO3 - -N, DO, pH, COD, temperature, flow rate, water level, etc. are monitored and collected.

[0102] The environmental perception module integrates sensors such as light intensity, rainfall, and wind speed, can synchronously obtain meteorological data, and constructs a meteorological-water quality coupling data set. The photosynthesis of plants and other aspects are monitored through a leaf surface photosynthesis detector.

[0103] The biofilm activity monitoring module, based on the combination of electrochemical impedance spectroscopy (EIS) and qPCR technology, dynamically monitors the metabolic activity of denitrifying bacteria and the abundance of nitrification / denitrification functional genes (such as amoA, nirS, nosZ), providing dynamic response data at the microbial level for AI decision-making.

[0104] As a specific embodiment, light intensity sensors, rain gauges and anemometers are installed around the wetland to collect meteorological data and transmit it to the edge computing terminal. A leaf photosynthesis detector is installed around the wetland to monitor the photosynthesis intensity of emergent plants in real time through spectral analysis. Combining the light intensity and water level data, it is dynamically fed back to the AI decision-making layer to optimize the aeration and tidal cycle strategies.

[0105] Electrochemical impedance spectroscopy (EIS) probes are embedded in the packing layer, and qPCR analysis is carried out in combination with periodic sampling to dynamically monitor the functional gene abundances of nitrifying bacteria (amoA gene) and denitrifying bacteria (nirS, nosZ genes).

[0106] During the actual working process, each sensor uploads data to the edge computing terminal every 10 minutes, and abnormal data triggers an encrypted retransmission mechanism.

[0107] As a specific embodiment, the AI decision-making layer includes data fusion and feature extraction, a dynamic prediction model, and an intelligent optimization engine.

[0108] Among them, data fusion and feature extraction perform normalization processing on multi-source heterogeneous data through the edge computing terminal to extract the key driving factor set of nitrogen removal efficiency (such as DO gradient, C / N ratio, hydraulic retention time, etc.). The dynamic prediction model uses an LSTM neural network to construct a nitrogen removal efficiency prediction model. The input variables include real-time water quality data, environmental factors, and historical operation parameters, and the output is the predicted value of TN removal rate and the optimal DO threshold. The intelligent optimization engine establishes a multi-objective optimization strategy based on the deep reinforcement learning (DRL) algorithm, with the maximization of nitrogen removal efficiency, the minimization of carbon source dosage, and the lowest energy consumption as the joint reward function, and dynamically generates the optimal solution set of tidal cycle, aeration intensity, and carbon source dosage.

[0109] The two-way cyclic optimization mode of the AI decision-making layer in this embodiment realizes continuous model update and equipment maintenance through the closed-loop mechanism of "data-driven decision-making" and "feedback-driven learning", ensuring long-term operation stability. This mode upgrades the traditional static control system to a dynamically adaptive intelligent ecosystem, providing an efficient, low-carbon, and reliable solution for the sewage treatment field.

[0110] Furthermore, predictions and decisions are generated based on real-time data. After execution, data is received and feedback is obtained. In the continuous execution and feedback, data is obtained for autonomous training, the model is updated and upgraded, decision optimization is achieved, and a new round of execution is carried out.

[0111] As a preferred embodiment, the execution layer includes a tidal gate controller, a gradient aeration system, and a dual-mode carbon source dosing device;

[0112] The tidal gate controller is used to adjust and control the influent flow rate;

[0113] The gradient aeration system is used to control the aeration intensity in different areas to optimize the dissolved oxygen level in the water body;

[0114] The carbon source dual-mode dosing device is used to add an organic carbon source to support the nitrogen removal process of microorganisms.

[0115] like Figure 4 As shown, Figure 4 A schematic diagram showing the connection relationship between the various parts of the system of this embodiment is shown.

[0116] Specifically, the tidal gate controller adjusts the frequency of water inflow / drainage (adjustable from 0.5 to 24 hours) through the electric gate controller to form a tidal water inflow; the outlet channel is installed with several electric valves from top to bottom (the last valve is 5 cm from the bottom of the wetland), each with an interval of 10 cm; AI is used to control the opening of each electric valve to control the height of the water outlet. An electric gate controller is installed on the wetland inlet and outlet pipes, and the time and frequency of gate opening and closing are controlled by AI, combined with a water level sensor to form a tidal water inflow (the cycle is adjustable from 0.5 to 24 hours).

[0117] The gradient aeration system uses a micro-aerator linked to an aeration network, adjusts the aeration intensity according to a gradient (0.5-6 mg / L) based on the predicted DO concentration value, designs a spatiotemporal difference control mode of "enhanced aeration in the early stage of submergence - trace oxygen supply in the late stage", and gives priority to enhancing the oxygen transfer efficiency of the nitrification zone.

[0118] The carbon source dual-mode dosing device is based on the denitrification demand, and is synergistically dosing through a peristaltic pump and a slow-release carbon source (polycaprolactone / starch composite particles) device. Polycaprolactone (PCL) provides mechanical strength, and starch composite particles (particle size 5-10mm) serve as a degradable carbon source. Microorganisms degrade PCL / starch composite particles, slowly release dissolved organic carbon (DOC), and maintain a dynamic balance of C / N ratio of 3.5-5.0.

[0119] Based on EIS electrochemical impedance spectroscopy and qPCR molecular monitoring data, the carbon source was added synergistically through a peristaltic pump and a slow-release carbon source (polycaprolactone / starch composite particles) device, and a response surface model of the dynamic C / N ratio (3.5-5.0) and the carbon source addition ratio (5%-20%) was established. Compared with traditional methods, the waste of carbon sources was reduced, and at the same time, the demand for exogenous carbon was reduced through sulfur autotrophic denitrification, thereby achieving synergistic efficiency in pollution reduction and carbon reduction.

[0120] A microporous aeration network is laid at the bottom of the aerobic zone, and a micro-aerator is used to release gas to meet basic oxygenation needs. The aeration intensity is dynamically adjusted by the AI decision-making layer according to the preset gradient (0.5-6 mg / L).

[0121] A peristaltic pump and a slow-release carbon source dosing device (polycaprolactone / starch composite particles) are configured in the cloth water tank. The peristaltic pump controls the hose compression frequency through a high-precision stepper motor to achieve precise dosing of the carbon source liquid (such as sodium acetate solution) (rate range: 0–20 mL / min). The starch composite particles are decomposed to slowly release dissolved organic carbon (DOC) to provide the basic carbon source, and the dosing amount is dynamically adjusted based on the AI model.

[0122] As a specific embodiment, the AI decision-making layer has the ability of autonomous training and generates initial control instructions according to the LSTM prediction results; the DRL engine combines historical data and real-time feedback to optimize the instruction parameters; the execution layer synchronously adjusts the tidal cycle, aeration intensity and carbon source dosing amount. When the system is running, the deviation between the predicted value and the actual value will trigger online retraining of the model. At the same time, the experience replay mechanism continuously optimizes the strategy, forming a closed-loop iteration of "data-driven prediction → prediction-guided decision-making → decision feedback correction prediction", enabling the system to have forward-looking regulation (such as pre-judging high load and adjusting parameters in advance), dynamic adaptability (such as reducing aeration when the light is sufficient) and self-optimization ability.

[0123] The system adopted in this embodiment uses advanced deep learning and deep reinforcement learning technologies, combines multi-source sensing data, and provides an intelligent, automated, efficient and energy-saving water quality management and optimization method. Through precise prediction and dynamic adjustment, it can effectively improve the denitrification efficiency, reduce the consumption of carbon source and energy, and ensure the efficient operation of the water treatment system and the achievement of environmental protection goals. Through the deep coupling of the intelligent sensing layer, decision-making optimization layer and execution reinforcement layer, a smart denitrification paradigm of "prediction-decision-making-execution-feedback" is constructed. Compared with the traditional process, on the premise of keeping the effluent TN stably better than the surface water class IV standard, the cost per ton of water treatment is reduced and the carbon footprint is reduced, providing a solution with subversive innovation value for the advanced purification of the tail water of urban sewage treatment plants, and showing significant comprehensive environmental-economic benefits.

[0124] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention.

Claims

1. An AI-based automated regulation and control method for tidal flow constructed wetlands, characterized in that, Including: Real-time collection of multi-source sensing data of tidal flow constructed wetland system; Preprocessing and feature extraction of the multi-source sensing data to obtain key driving factors; Constructing a denitrification efficiency prediction model based on the LSTM network, and determining the predicted value of TN removal rate and the optimal DO threshold according to historical operation data and key driving factors collected in real time; According to the predicted value of TN removal rate and the optimal DO threshold, establishing a joint reward function with the maximization of denitrification efficiency, the minimization of carbon source dosage, and the lowest energy consumption based on the DRL algorithm, determining the multi-objective optimization strategy, and dynamically generating the adjustment parameters of the tidal flow constructed wetland system; Adjusting the influent / effluent frequency through the electric gate controller according to the adjustment parameters, precisely dosing the carbon source through the peristaltic pump and slow-release carbon source, and adjusting the aeration intensity in gradient according to the predicted value of DO concentration; Obtaining the microbial activity feedback data, and optimizing the regulation strategy according to the microbial activity feedback data to make the effluent TN reach the preset standard.

2. The AI-based automated regulation and control method for tidal flow constructed wetland according to claim 1, wherein The multi-source sensing data includes: water quality sensing data, environmental perception data, and biofilm activity monitoring data; The water quality parameters include NH4 + -N, NO3 - -N, DO, pH, COD, temperature, flow rate, and water level data, which are used to characterize the health status of the water body and the changes in pollutants; The environmental perception data includes light intensity, rainfall, and wind speed, which are used to characterize the impact of external climate environment on water quality changes; The biofilm activity monitoring data includes the metabolic activity of denitrifying bacteria and the abundance of nitrification / denitrification functional genes obtained by dynamic monitoring, which are used to characterize the impact of microbial communities in water on water quality.

3. The AI-based automated regulation and control method for tidal flow constructed wetland according to claim 1, wherein, Preprocessing and feature extraction of the multi-source sensing data to obtain key driving factors, including: Normalizing the multi-source sensing data and extracting candidate regulation and control features; Using the principal component analysis method and the mutual information screening method to construct a denitrification efficiency driving factor matrix according to the candidate regulation and control features, and obtaining the key driving factors affecting denitrification efficiency.

4. The AI-based automated regulation and control method for tidal flow constructed wetland according to claim 2, characterized in that Constructing a denitrification efficiency prediction model based on the LSTM network, and determining the predicted value of TN removal rate and the optimal DO threshold according to historical operation data and key driving factors collected in real time, including: Taking the water quality sensing data, environmental perception data, and biofilm activity monitoring data as the input of the LSTM network; Constructing a time series input through a preset time window, using a bidirectional LSTM to capture the temporal dependencies before and after, and outputting the predicted value of TN removal rate and the DO threshold for a future preset period.

5. The AI-based automated regulation and control method for tidal flow constructed wetland according to claim 4, wherein According to the predicted value of TN removal rate and the optimal DO threshold, establishing a joint reward function with the maximization of denitrification efficiency, the minimization of carbon source dosage, and the lowest energy consumption based on the DRL algorithm, determining the multi-objective optimization strategy, and dynamically generating the adjustment parameters of the tidal flow constructed wetland system, including: Taking the predicted value of TN removal rate and the optimal DO threshold output by the LSTM network as the extended state, and jointly constituting the multi-dimensional state space of DRL with the real-time parameters; Taking the tidal cycle adjustment rate, aeration intensity grading, and carbon source dosing rate as the action space; Deeply coupling the multi-dimensional state space, action space, and joint reward function of DRL to form a closed-loop optimization mechanism.

6. The AI-based automatic regulation and control method for tidal flow constructed wetland according to claim 5, wherein, The reward function is determined according to the joint optimization of denitrification efficiency, carbon source consumption, and aeration energy consumption, and the strategy is iteratively updated through the Q-learning algorithm.

7. A tidal flow constructed wetland system, characterized in that, It includes a sensing layer, an AI decision-making layer, an execution layer, and a wetland function layer connected in sequence; The sensing layer is used to collect multi-source sensing data of the tidal flow constructed wetland system in real time; The AI decision-making layer is used to preprocess and extract features from the multi-source sensing data to obtain key driving factors, construct a denitrification efficiency prediction model based on the LSTM network, determine the predicted value of TN removal rate and the optimal DO threshold according to historical operation data and key driving factors collected in real time; according to the predicted value of TN removal rate and the optimal DO threshold, establish a joint reward function with the maximization of denitrification efficiency, the minimization of carbon source dosage, and the lowest energy consumption based on the DRL algorithm, determine the multi-objective optimization strategy, and dynamically generate the adjustment parameters of the tidal flow constructed wetland system; The execution layer is used to adjust the influent / drainage frequency through the electric gate controller according to the adjustment parameters, precisely add carbon source through the peristaltic pump and slow-release carbon source, and adjust the aeration intensity in gradient according to the predicted value of DO concentration; The wetland function layer adopts a layered filler optimization layer and a plant root oxygenation structure to construct a multi-path denitrification system.

8. The tidal flow constructed wetland system according to claim 7, characterized in that, The layered filler optimization layer of the wetland function layer includes an upper aerobic zone and a lower anoxic zone; the upper aerobic zone is filled with a composite of volcanic rock and iron-carbon composite material to strengthen the synergistic effect of nitrifying bacteria enrichment and chemical oxidation; the lower anoxic zone is filled with a mixture of biochar and sulfur autotrophic denitrification matrix to promote the coupling of sulfur-driven denitrification and biochar adsorption; The plant root bionic oxygenation structure realizes the oxygen supply function by planting emergent plants.

9. The tidal flow constructed wetland system according to claim 7, characterized in that The sensing layer includes a water quality sensing module, an environmental perception module, and a biofilm activity monitoring module; The water quality sensing module includes multiple sensors, and the multiple sensors are arranged in the inlet basin, the layered filler area, and the collection basin to collect water quality data in real time; The environmental perception module is used to collect data on light intensity, rainfall, and wind speed; The biofilm activity monitoring module is used to collect data on the abundance of microbial functional genes based on electrochemical impedance spectroscopy and qPCR technology.

10. The tidal flow constructed wetland system according to claim 7, characterized in that, The execution layer includes a tidal gate controller, a gradient aeration system, and a dual-mode carbon source dosing device; The tidal gate controller is used to adjust and control the influent flow rate; The gradient aeration system is used to control the aeration intensity in different areas to optimize the dissolved oxygen level in the water body; The dual-mode carbon source dosing device is used to dose organic carbon source to support the nitrogen removal process of microorganisms.

Images