Sewage filtering method and system based on multi-crop cooperation

Abstract

The invention relates to the technical field of environmental engineering, and discloses a sewage filtering method and system based on multi-crop collaboration.The sewage filtering method based on multi-crop collaboration.The sewage filtering method based on multi-crop collaborationcomprises the steps that an original sewage sample is obtained, a multi-factor comprehensive evaluation method is adopted for analysis and treatment, and a multi-crop collaborationpollutant removal mechanism model is established; a multi-level system structure is constructed, and a three-dimensional hydraulic modeling method is adopted for analysis and calculation; a multi-sensor fusion monitoring network is constructed, and a data fusion algorithm is adopted for processing and analysis; system parameters are optimized and adjusted by adopting an intelligent self-adaptive control strategy; regularly adding a microbial agent to carry out system biological enhancement and dynamic regulation and control optimization, and establishing a stable microbial ecological system; all functional units are integrated into a modular system, and performance verification is carried out based on a stable microbial ecological system to obtain purified water; the system has the advantages of being high in treatment efficiency, stable in operation, high in intelligent level and low in construction cost.

Description

Technical Field

[0001] This invention relates to the field of environmental engineering technology, and more specifically, to a wastewater filtration method and system based on multi-crop synergy. Background Technology

[0002] With the rapid advancement of industrialization and increasingly stringent environmental protection requirements, wastewater treatment technology faces new challenges and development opportunities. Traditional wastewater treatment technologies mainly rely on physicochemical methods and single biological treatment technologies. While these have certain advantages in removing specific pollutants, they reveal significant shortcomings when treating industrial wastewater with complex compositions and dynamically changing concentrations. Physicochemical treatment methods are energy-intensive, have high operating costs, and are prone to generating secondary pollution; traditional biological treatment technologies have limited synergistic removal effects on multiple pollutants, making it difficult to meet the treatment needs of complex industrial wastewater.

[0003] In recent years, phytoremediation technology has received widespread attention due to its environmental friendliness and low cost. Constructed wetland systems can effectively remove multiple pollutants from wastewater through the synergistic effect of plants, microorganisms, and substrates. However, existing phytoremediation systems mainly use single crops or a few types of plants. Different plants have varying absorption capacities and treatment efficiencies for different pollutants, making it impossible to achieve synergistic and efficient removal of multiple pollutants. Furthermore, traditional systems lack intelligent monitoring and control capabilities, failing to automatically adjust operating parameters based on dynamic changes in wastewater composition, resulting in unstable treatment effects and high maintenance costs.

[0004] In decentralized wastewater treatment scenarios within industrial parks, there is an urgent need for a treatment system capable of intelligently adapting to the dynamic changes in the composition of various industrial pollutants and achieving synergistic and efficient removal of different pollutants. This system must possess not only highly efficient multi-pollutant removal capabilities and intelligent adaptive control functions, but also meet multiple requirements such as modular design, economic feasibility, and environmental friendliness. Therefore, developing intelligent wastewater filtration methods and systems based on multi-crop synergy has significant theoretical and practical value. Summary of the Invention

[0005] This invention provides a wastewater filtration method and system based on multi-crop synergy, which solves the technical problems in related technologies such as the lack of intelligent monitoring and control capabilities and the inability to automatically adjust operating parameters according to the dynamic changes in wastewater composition.

[0006] This invention provides a wastewater filtration method based on multi-crop synergy, comprising the following steps: Raw wastewater samples were obtained, and water quality parameters were tested on the raw wastewater samples. At the same time, physiological characteristic data of candidate crops were obtained. The water quality parameters and crop characteristic data were analyzed and processed using a multi-factor comprehensive evaluation method to obtain the environmental adaptability index of each crop. Based on the environmental adaptability index, a multi-crop synergistic pollutant removal mechanism model was established. Based on the multi-crop synergistic pollutant removal mechanism model, a multi-level system structure including a floating plant layer, an emergent plant layer, and a submerged plant layer was constructed. The multi-level system structure was analyzed and calculated using a three-dimensional hydraulic modeling method to obtain an accurate water flow distribution scheme. A multi-sensor fusion monitoring network is constructed based on the water flow distribution scheme. The monitoring data is processed and analyzed using a data fusion algorithm to obtain the system operation status monitoring results. Based on the system operation status monitoring results, an intelligent adaptive control strategy is adopted to optimize and adjust the system parameters, thereby obtaining optimized operating parameters; Based on optimized operating parameters, microbial agents are periodically added to conduct system bioaugmentation and dynamic regulation optimization, and a stable microbial ecosystem is established. The functional units are integrated into a modular system, and the performance is verified based on a stable microbial ecosystem to obtain purified water.

[0007] In a preferred embodiment, the analysis and processing of water quality parameters and crop characteristic data using a multi-factor comprehensive evaluation method to obtain the environmental adaptability index of each crop includes: The root morphology characteristics, pollutant absorption kinetic parameters, and growth cycle characteristics of candidate crops were obtained. The root surface area and root density parameters were determined by morphological analysis. The maximum absorption rate and half-saturation constant were obtained by fitting the pollutant absorption kinetic parameters using the Michaelis-Menten equation. Based on water quality parameters such as pH value, heavy metal concentration, and organic matter content, the adaptability of each candidate crop is calculated and analyzed using fuzzy mathematics evaluation method. The adaptability is quantified by Gaussian membership function to obtain the single-factor adaptability score of each environmental factor. The single-factor adaptability score was comprehensively calculated using a weighted comprehensive evaluation method. The corresponding weight coefficients were set according to the degree of influence of each environmental factor on crop growth, and the comprehensive environmental adaptability index of each crop was calculated.

[0008] In a preferred embodiment, the multi-crop synergistic pollutant removal mechanism model based on the environmental adaptability index includes: The top three crops in the comprehensive environmental adaptability index ranking were obtained. The pollutant absorption mechanism of the crops was analyzed using biochemical kinetic methods. Based on the pollutant characteristic data, the absorption kinetic equations of each crop for different pollutants were established. The kinetic parameters in the absorption kinetic equations were obtained through experimental measurement methods. Based on the absorption kinetic equation and kinetic parameters, the synergistic mechanism among crops is modeled and analyzed using systems biology methods. The nutrient competition relationship and pollutant transport pathway among crops are determined through the analysis of rhizosphere microenvironment interactions, and a mathematical model for multi-crop synergistic pollutant removal is established. The accuracy of the mathematical model was verified by experimental verification methods. The reliability of the model was evaluated by comparing the deviation between the predicted values ​​of the mathematical model and the actual measured values. Based on the verification results, the parameters of the mathematical model were optimized and adjusted to obtain a verified multi-crop synergistic pollutant removal mechanism model.

[0009] In a preferred embodiment, the multi-layered system structure comprising a floating plant layer, an emergent plant layer, and a submerged plant layer, constructed based on a multi-crop synergistic pollutant removal mechanism model, includes: The optimal crop configuration scheme was determined based on a validated multi-crop synergistic pollutant removal mechanism model. A surface floating plant layer was set up according to the ecological niche characteristics and pollutant removal capacity of each crop in the configuration scheme. The floating plants were selected to perform preliminary removal treatment of organic pollutants and nitrogen and phosphorus nutrients in the wastewater. Based on the water quality after surface treatment, a middle layer of emergent plants is set according to the emergent plant configuration parameters in the configuration scheme, and emergent plants are selected to further remove heavy metal ions and residual organic pollutants. Based on the water quality data after the middle layer treatment, the bottom layer of submerged plants is set according to the submerged plant configuration parameters in the configuration scheme. Submerged plants are selected to carry out deep purification treatment of sewage. Through the synergistic effect of the vertical distribution of the three layers of plants, a multi-level system structure with clear functional division is obtained.

[0010] In a preferred embodiment, the step of analyzing and calculating the multi-level system structure using a three-dimensional hydraulic modeling method to obtain an accurate water flow distribution scheme includes: Based on the geometric parameters and boundary conditions of the multi-level system structure, a three-dimensional hydraulic numerical model is established, and the flow field distribution inside the system is obtained by solving the Navier-Stokes equations. Based on the flow field distribution, an artificial neural network algorithm is used to optimize the hydraulic residence time calculation and determine the optimal hydraulic residence time allocation for each layer of vegetation. Based on the optimal hydraulic residence time allocation, the water distribution system is optimized by combining computational fluid dynamics and machine learning methods to obtain a water flow distribution scheme that can achieve uniform water distribution and optimal residence time.

[0011] In a preferred embodiment, the construction of the multi-sensor fusion monitoring network based on the water flow distribution scheme includes: Based on the spatial distribution characteristics of the multi-level system structure, water quality sensors, soil sensors, meteorological sensors and plant growth sensors are deployed in the plant areas of each level to form a multi-sensor monitoring network covering the entire system area. Based on multi-source heterogeneous data collected by a multi-sensor monitoring network, a hierarchical fusion architecture including data layer fusion, feature layer fusion, and decision layer fusion is constructed. The Kalman filter algorithm is used to preprocess and filter out noise from the sensor data. Based on a layered fusion architecture, deep learning methods are used to intelligently analyze and process the fused data to obtain system operation status monitoring results.

[0012] In a preferred embodiment, the analysis and processing of water quality parameters and crop characteristic data using a multi-factor comprehensive evaluation method to obtain the environmental adaptability index of each crop includes: Based on the water quality parameters of wastewater samples and the physiological characteristic parameters of crop data, the weight coefficients of each evaluation factor were determined by the analytic hierarchy process (AHP) and a comprehensive evaluation index system including water quality adaptability, growth adaptability and stress resistance was established. Based on a comprehensive evaluation index system, the fuzzy comprehensive evaluation method is used to quantitatively assess the adaptability of each crop under different wastewater environmental conditions, and the environmental adaptability index of each crop is obtained.

[0013] In a preferred embodiment, the multi-crop synergistic pollutant removal mechanism model includes: Based on environmental adaptability index and pollutant characteristic data, a mathematical model describing the interaction between crops is established. The mathematical model describing the interaction between crops includes the nutrient competition coefficient, root exudate interaction coefficient and pollutant removal synergy coefficient among crops. Based on a mathematical model, the optimal configuration ratio and spatial distribution scheme of each crop are solved by numerical calculation method, resulting in a multi-crop synergistic pollutant removal mechanism model that can maximize the synergistic removal effect of pollutants.

[0014] In a preferred embodiment, the deep learning method includes: Based on the multi-source sensor data processed by the hierarchical fusion architecture, a deep learning model containing a convolutional neural network and a long short-term memory network is constructed. The convolutional neural network is used to extract spatial features, and the long short-term memory network is used to capture temporal features. The system operates by training and learning fused data using a deep learning model and optimizing network parameters through backpropagation. This enables intelligent identification and classification of system operating status, resulting in system operating status monitoring results that include normal operation, anomaly warnings, and fault diagnosis.

[0015] This invention provides a wastewater filtration system based on multi-crop synergy, used to perform the aforementioned wastewater filtration method based on multi-crop synergy, comprising: The crop adaptability assessment and collaborative modeling unit is used to acquire raw sewage samples, detect water quality parameters of the raw sewage samples, acquire physiological characteristic data of candidate crops, analyze and process the water quality parameters and crop characteristic data using a multi-factor comprehensive evaluation method, obtain the environmental adaptability index of each crop, and establish a multi-crop collaborative pollutant removal mechanism model based on the environmental adaptability index. A multi-level system structure building unit is constructed based on a multi-crop synergistic pollutant removal mechanism model, which includes a floating plant layer, an emergent plant layer, and a submerged plant layer. The multi-level system structure is analyzed and calculated using a three-dimensional hydraulic modeling method to obtain an accurate water flow distribution scheme. The multi-sensor fusion monitoring unit constructs a multi-sensor fusion monitoring network based on the water flow distribution scheme, processes and analyzes the monitoring data using a data fusion algorithm, and obtains the system operation status monitoring results. The intelligent adaptive control unit optimizes and adjusts system parameters based on the system operation status monitoring results using an intelligent adaptive control strategy to obtain optimized operating parameters. The bioaugmentation and dynamic regulation unit periodically adds microbial agents according to optimized operating parameters to optimize system bioaugmentation and dynamic regulation, and establishes a stable microbial ecosystem. The system integration and performance verification unit performs performance verification based on a stable microbial ecosystem to obtain purified water.

[0016] The beneficial effects of this invention are as follows: By establishing a multi-crop synergistic pollutant removal mechanism model and scientifically and rationally configuring plant species with different functions, the synergistic and efficient removal of multiple pollutants, including heavy metals, organic pollutants, and nitrogen and phosphorus nutrients, was achieved. Compared with traditional single-crop treatment systems, the multi-crop synergistic system shows improvements in heavy metal removal rates, organic pollutant removal rates, and nutrient removal rates, and can simultaneously meet the treatment requirements of multiple pollutants. The three-layer vertically distributed system structure design fully utilizes the strengths of different plants: the surface floating plants are responsible for rapid interception and nutrient absorption, the middle emergent plants are responsible for organic matter degradation and heavy metal accumulation, and the bottom submerged plants are responsible for deep purification and ecological stability, forming a complete pollutant removal chain.

[0017] A multi-sensor fusion monitoring network and an intelligent adaptive control system were constructed, enabling real-time monitoring of wastewater composition changes, plant growth status, and system operating parameters. The system automatically adjusts its operating strategies to achieve unattended operation. Employing a hierarchical control architecture combined with fuzzy control theory and reinforcement learning algorithms, the system possesses self-learning and optimization capabilities, continuously improving its control strategies based on historical data. When wastewater composition changes or abnormal conditions occur, the system responds quickly and adjusts automatically, ensuring the stability and reliability of the treatment effect. Intelligent control reduces reliance on specialized technical personnel, lowers manual maintenance costs, and improves the system's economic efficiency and practicality. Attached Figure Description

[0018] Figure 1 This is a flowchart of a wastewater filtration method based on multi-crop synergy according to the present invention; Figure 2 This is a block diagram of a wastewater filtration system based on multi-crop synergy according to the present invention; Figure 3 This is a comparison chart of the multi-crop synergistic treatment effects of the present invention; Figure 4 This is a system operation stability analysis diagram of the present invention; Figure 5 This is a comprehensive performance evaluation diagram of the multi-crop synergistic system of the present invention. Detailed Implementation

[0019] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, some features described in the examples may be combined in other examples.

[0020] At least one embodiment of the present invention discloses a wastewater filtration method based on multi-crop synergy, such as... Figure 1 As shown, it includes: Step 1: Obtain raw sewage samples, test water quality parameters on the raw sewage samples, and simultaneously obtain physiological characteristic data of candidate crops. Analyze and process the water quality parameters and crop characteristic data using a multi-factor comprehensive evaluation method to obtain the environmental adaptability index of each crop. Based on the environmental adaptability index, establish a multi-crop synergistic pollutant removal mechanism model. Specifically, the following steps are included: Obtain water quality parameters of the target wastewater, including basic water quality indicators such as heavy metal ion concentration, organic pollutant concentration, nitrogen and phosphorus nutrient concentration, pH value, and temperature. Simultaneously, collect physiological characteristic data of candidate crops, including root morphology characteristics, pollutant absorption kinetics parameters, growth cycle characteristics, and environmental adaptability parameters.

[0021] A crop environmental adaptability assessment model was established to analyze the adaptability of various crops to water quality conditions, nutrient requirements, and growth environment. The crop adaptability assessment employed a multi-factor comprehensive evaluation method, with the following specific steps: Actual measurement data of various environmental factors were collected, including water quality parameters, soil conditions, and climate conditions, and the optimal growth range and tolerance range of each candidate crop for these environmental factors were obtained. The crop's adaptability was calculated for each environmental factor. The closer the actual value of an environmental factor is to the crop's optimal growth value, the higher the adaptability; the further the actual value deviates from the optimal value, the lower the adaptability. A Gaussian distribution model was used to calculate the adaptability, which reaches its highest level (1) at the optimal value and decreases in a bell-shaped curve as the deviation increases. The adaptability of each environmental factor was weighted and summed according to its importance to obtain the crop's comprehensive environmental adaptability index. The importance weights were determined based on the degree of influence of each environmental factor on crop growth, with key factors having higher weights and minor factors having lower weights.

[0022] Based on crop adaptability assessment, a multi-crop synergistic pollutant removal mechanism model was established. By analyzing the absorption mechanisms of heavy metals, organic pollutants, and nitrogen and phosphorus nutrients by different crops, a crop-pollutant interaction model was constructed. Taking heavy metal removal as an example, a kinetic description was established considering multiple mechanisms such as root adsorption, intracellular accumulation, and rhizosphere microbial synergy. Heavy metal removal involves three main pathways: primary crops directly adsorb and absorb heavy metal ions from water through their roots, with the removal rate directly proportional to crop biomass and heavy metal concentration; auxiliary crops synergistically remove heavy metals through different physiological mechanisms, creating a complementary effect; and rhizosphere microorganisms further remove heavy metals through biotransformation and complexation. All three pathways occur simultaneously, and the overall removal effect is the sum of the removal amounts from each pathway.

[0023] The pollutant absorption kinetics are described using the Michaelis-Menten equation, which reflects the saturation relationship between the pollutant absorption rate and its concentration: when the pollutant concentration is low, the absorption rate increases rapidly in direct proportion to the concentration; as the concentration gradually increases, the growth rate of the absorption rate gradually slows down; and when the concentration reaches a certain level, the absorption rate tends to stabilize at its maximum value. Key parameters for each crop include root surface area, root density, maximum absorption rate, and half-saturation constant. These parameters were obtained through experimental measurements and used to establish accurate predictive models for pollutant removal.

[0024] The synergistic effect assessment employed a synergistic index calculation method to quantify the enhancing effect of different crop combinations on pollutant removal efficiency. The calculation steps for the synergistic index were as follows: the pollutant removal efficiency of each individual crop was measured, the overall removal efficiency of the crop combination was measured, and finally, a comparative analysis was conducted to determine the degree of improvement in combined efficiency relative to the efficiency of individual crops. A synergistic effect was considered to exist between the two crops when the combined treatment efficiency increased by more than 20% compared to the highest efficiency of a single crop.

[0025] The synergistic pollutant removal mechanism model employs biochemical kinetics to analyze the absorption, transformation, and accumulation of pollutants within plants, combined with systems biology methods to model the synergistic effects among multiple crops. Through rhizosphere microenvironment interaction analysis, the nutrient competition relationships and pollutant transport pathways among different crops are determined. Experimental validation methods are used to evaluate the model's reliability and predictive accuracy, ensuring that the model accurately reflects actual synergistic removal effects.

[0026] Furthermore, an improved genetic algorithm can be used to optimize crop combination configurations. A fitness function is established that comprehensively considers pollutant removal efficiency, crop growth stability, and system operating costs. The optimal crop combination scheme is searched through selection, crossover, and mutation operations in the genetic algorithm. The fitness function is set as a multi-objective optimization, and a weighted summation method is used to merge multiple objectives into a single objective function.

[0027] Furthermore, deep learning neural networks can be used to optimize crop configuration strategies, replacing traditional genetic algorithm optimization methods. A crop configuration optimization model based on deep reinforcement learning is constructed, using neural networks to learn the complex nonlinear mapping relationship between wastewater components and the optimal crop configuration. The deep reinforcement learning model employs a dual deep Q-network architecture, consisting of a main network and a target network. The main network is responsible for online learning and decision-making, while the target network provides a stable training objective.

[0028] The system takes the current wastewater quality parameters as input, including key indicators such as various pollutant concentrations, pH, and temperature, and outputs the selection of crop species and density configuration for each layer. An evaluation system is established by comprehensively considering pollutant removal efficiency, system stability, and operating costs to guide the network in learning the optimal crop configuration strategy.

[0029] To ensure the comparability of evaluation indicators, data standardization is required for each indicator: the pollutant removal efficiency, system stability, and operating cost are all converted to the same numerical range, and then a comprehensive evaluation is conducted according to preset weights to obtain the final evaluation result. After standardization, the three indicators have the same dimensions and numerical range, ensuring the rationality of the comprehensive evaluation.

[0030] The network training employs an experience replay mechanism, establishing an experience pool to store four-tuple data of historical states, actions, rewards, and the next state. Random sampling of historical experience is used for network training to break the temporal correlation between data, improving learning stability. The target network uses a soft update strategy, periodically copying parameters from the main network to maintain training stability. The deep learning-optimized crop configuration strategy can automatically adapt to changes in wastewater composition without manual parameter adjustments, improving the system's adaptability. This alternative embodiment aims to further enhance the system's intelligence level, reduce reliance on technical personnel, and increase the automation of system operation.

[0031] Step 2: Based on the multi-crop synergistic pollutant removal mechanism model, a multi-level system structure including a floating plant layer, an emergent plant layer, and a submerged plant layer is constructed. The multi-level system structure is analyzed and calculated using a three-dimensional hydraulic modeling method to obtain an accurate water flow distribution scheme. Specifically, the following steps are included: Based on crop adaptability assessment results and synergistic effect analysis, a multi-level system structure was designed. The system adopts a three-layer vertical distribution design: a surface layer of floating plants, a middle layer of emergent plants, and a bottom layer of submerged plants. Each layer is separated by a porous media layer to achieve rational water flow distribution and gradient removal of pollutants.

[0032] The surface floating plant layer mainly consists of fast-growing floating plants such as water hyacinth and duckweed, with a plant coverage density controlled at 60%–80%. This layer is primarily responsible for the initial interception of large particulate pollutants and the rapid absorption of excess nutrients. The water depth is set at 0.3–0.5 meters, and the hydraulic retention time is 8–12 hours. The middle emergent plant layer consists of emergent plants with well-developed root systems such as reeds, cattails, and calamus, with a planting density of 15–20 plants / square meter. This layer has a water depth of 0.8–1.2 meters and a hydraulic retention time of 16–24 hours, primarily responsible for the biodegradation of organic pollutants and the root accumulation of heavy metals. The bottom submerged plant layer consists of submerged plants such as hornwort, hydrangea, and eelgrass, with a coverage density of 40%–60%, a water depth of 1.0–1.5 meters, and a retention time of 12–18 hours, primarily responsible for the deep purification of the water and the stable maintenance of the ecosystem.

[0033] A precise three-dimensional hydraulic model was established using computational fluid dynamics (CFD) methods to simulate the flow distribution, velocity field, and concentration field within a multi-level system. Hydraulic control was based on the fundamental principles of the laws of conservation of mass and momentum, and detailed flow field information was obtained through numerical calculations.

[0034] The 3D hydraulic modeling employs the finite volume method for numerical solution, discretizing the computational domain into a finite number of control volumes and applying conservation laws to establish discrete equations for each control volume. The finite volume method exhibits good conservation and stability, making it suitable for flow field calculations with complex geometries. The SIMPLE algorithm is used to handle pressure-velocity coupling problems in the numerical solution, ensuring the convergence and accuracy of the calculation results.

[0035] The basic idea of ​​hydraulic modeling is as follows: Establish a fluid continuity equation to describe the fluid mass conservation relationship, i.e., the mass of fluid flowing into the system equals the mass of fluid flowing out of the system plus the change in fluid mass within the system; establish a momentum equation to describe the force balance relationship of fluid motion, considering pressure gradient force, viscous force, gravity, and the resistance effect of plants on water flow. Plant resistance is described using a porous media model, treating plants as porous media with a certain porosity and permeability, and calculating the impediment effect of plants on water flow using seepage theory.

[0036] By optimizing the hydraulic retention time allocation, sufficient reaction time is ensured for each treatment unit. The hydraulic retention time for the surface floating plant layer is set to 8 to 12 hours, for the middle emergent plant layer to 16 to 24 hours, and for the bottom submerged plant layer to 12 to 18 hours. The overall system hydraulic retention time is controlled within the range of 36 to 54 hours, and is dynamically adjusted according to the influent water quality and treatment requirements.

[0037] By numerically solving the above equations, a detailed flow field distribution within the system is obtained. Based on the flow field analysis results, the design of the inlet water distribution system is optimized, adopting a multi-point decentralized inlet method to ensure uniform water flow distribution at each layer. An adjustable baffle system is installed to adjust the water flow distribution ratio between layers according to actual operational needs. The effluent collection system adopts a layered collection design, allowing for separate monitoring of the treatment effect at each layer.

[0038] Furthermore, a flow field optimization method based on artificial neural networks can be employed. By establishing a nonlinear mapping relationship between input parameters (influent flow rate, water level at each layer, guide vane angle, etc.) and output indicators (flow field uniformity, dead zone volume ratio, etc.), intelligent optimization of system structural parameters can be achieved. The neural network is trained using the backpropagation algorithm, with the hidden layer employing the ReLU activation function and the output layer employing a linear activation function.

[0039] Furthermore, a smart flow field design method that couples computational fluid dynamics and machine learning can be adopted to replace the traditional empirical design method. A flow field prediction model based on convolutional neural networks can be constructed, and by learning from a large amount of CFD simulation data, a rapid predictive relationship between geometric parameters and flow field distribution can be established. Convolutional neural networks can automatically extract the spatial features of the flow field, achieving accurate prediction of complex three-dimensional flow fields.

[0040] The network architecture employs an encoder-decoder structure. The encoder extracts feature representations of the input geometry through multi-layer convolution and pooling operations, while the decoder reconstructs the flow field distribution through deconvolution operations. The loss function is designed as a weighted combination of the flow field prediction error and physical constraint terms, ensuring that the prediction results satisfy both data fitting requirements and fundamental laws of fluid mechanics.

[0041] Training data was obtained through parametric geometric modeling and high-precision CFD simulation, covering different inflow conditions, geometric configurations, and boundary conditions. Data augmentation techniques, including geometric transformations and noise addition, were employed to improve the model's generalization ability.

[0042] Step 3: Construct a multi-sensor fusion monitoring network based on the water flow distribution scheme, process and analyze the monitoring data using a data fusion algorithm, and obtain the system operation status monitoring results; Specifically, the following steps are included: A comprehensive multi-sensor fusion monitoring network is constructed to achieve all-round real-time monitoring of the system's operating status. The monitoring network includes four types of sensor systems: water quality sensors, soil sensors, meteorological sensors, and plant growth sensors.

[0043] The water quality sensor network deploys pH sensors, dissolved oxygen sensors, conductivity sensors, turbidity sensors, online chemical oxygen demand (COD) analyzers, online ammonia nitrogen analyzers, and online total phosphorus analyzers at key locations such as the system inlet, outlets of each layer, and the final outlet. The sensor sampling frequency is set to once every 30 minutes, and data transmission is conducted wirelessly.

[0044] The soil sensor system monitors changes in the physicochemical properties of each substrate layer, including parameters such as soil pH, electrical conductivity, moisture content, and redox potential. Meteorological sensors monitor ambient conditions such as temperature, humidity, light intensity, and wind speed. Plant growth sensors utilize image recognition technology to monitor biological indicators such as plant growth status, leaf color, and canopy cover.

[0045] Multi-sensor data fusion employs a hierarchical fusion architecture, comprising three layers: data layer fusion, feature layer fusion, and decision layer fusion. Data layer fusion preprocesses and calibrates raw data from similar sensors to eliminate systematic errors between sensors. A Kalman filter algorithm is used to remove sensor noise and outliers, improving data quality. The data fusion algorithm uses a weighted fusion method, assigning weight coefficients based on the accuracy and reliability of each sensor.

[0046] The soil sensor system deploys soil temperature sensors, soil moisture sensors, soil conductivity sensors, and soil pH sensors in the rhizosphere environment of each crop planting area to monitor changes in the physicochemical state of the rhizosphere environment. The sensors are buried at three depths: 10 cm, 20 cm, and 30 cm, comprehensively reflecting the environmental conditions of the root distribution area.

[0047] Meteorological sensor stations are deployed above the system, including air temperature sensors, relative humidity sensors, wind speed and direction sensors, light intensity sensors, and rainfall sensors, to monitor changes in meteorological factors that affect plant growth and system operation.

[0048] The plant growth sensor system employs non-contact monitoring technology, using high-resolution digital cameras, multispectral cameras, and lidar to periodically collect images and three-dimensional structural data of the plant canopy. Computer vision technology is then used to analyze key growth parameters such as leaf area index, biomass changes, and plant health status.

[0049] Multi-sensor data fusion adopts a layered fusion architecture, comprising three layers: data layer fusion, feature layer fusion, and decision layer fusion. Data layer fusion primarily performs data preprocessing, including noise filtering, outlier detection, and data synchronization. A Kalman filter algorithm is used to eliminate sensor noise and improve data quality.

[0050] The basic principle of Kalman filtering is to optimally estimate the true state of a system by establishing a mathematical model of the system state and combining it with sensor observation data. The filtering process includes two steps: prediction and updating. The prediction step predicts the current state value based on the system model; the updating step uses sensor observation data to correct the predicted value, obtaining a more accurate state estimate. The system weighs the reliability of the prediction model against the reliability of the observation data. When the reliability of the prediction model is high, it relies more on the predicted value; when the reliability of the observation data is high, it relies more on the observed value.

[0051] Feature layer fusion uses principal component analysis to extract the main features from multi-sensor data, reducing data dimensionality and redundant information. Decision layer fusion uses a support vector machine classifier to intelligently identify the system's operating status, classifying it into three levels: normal operation, early warning, and abnormal operation.

[0052] Furthermore, deep learning methods can be used for data fusion processing. A time series prediction model based on a long short-term memory network can be constructed to predict the effectiveness of wastewater treatment in advance. The network input includes historical monitoring data, weather forecast data, system operating parameters, etc., and the output is a prediction of water quality change trends over a certain period of time. The deep learning model adopts a convolutional neural network architecture, which automatically learns complex patterns and regularities in sensor data through multi-layer feature extraction and nonlinear transformation. The model training uses the backpropagation algorithm to continuously adjust the network parameters, minimize prediction errors, and improve the model's prediction accuracy and generalization ability.

[0053] Furthermore, a distributed data processing architecture that combines edge computing and cloud computing can be adopted to replace traditional centralized data processing methods. Edge computing nodes are deployed on-site to achieve real-time data preprocessing, feature extraction, and preliminary analysis, reducing data transmission volume and response latency. These edge computing nodes utilize embedded AI chips, enabling deep learning inference capabilities.

[0054] The data processing flow of an edge computing node includes steps such as data acquisition, preprocessing, feature extraction, anomaly detection, and decision reasoning. The preprocessing module uses a sliding window filtering algorithm to eliminate sensor noise, the feature extraction module uses principal component analysis to reduce dimensionality, and the anomaly detection module uses the isolated forest algorithm to identify abnormal data.

[0055] The cloud server handles computationally intensive tasks such as model training, parameter updates, and global optimization. The edge-cloud collaborative architecture employs a federated learning framework, where each edge node trains its model locally, and the cloud aggregates the model parameters to achieve distributed machine learning. The parameter aggregation in federated learning uses a weighted average method, assigning weights based on the amount of data at each edge node—nodes with larger data volumes have higher weights, and nodes with smaller data volumes have lower weights—ultimately forming globally optimized model parameters.

[0056] Data security employs homomorphic encryption to ensure privacy during data transmission and computation. Communication between edge nodes and the cloud utilizes a 5G network, guaranteeing low latency and high reliability. This alternative embodiment aims to improve the real-time performance and intelligence of data processing, reduce network bandwidth requirements, and enhance system scalability and security.

[0057] Step 4: Based on the system operation status monitoring results, an intelligent adaptive control strategy is used to optimize and adjust the system parameters to obtain optimized operating parameters; Specifically, the following steps are included: Based on the fusion monitoring results of multiple sensors, an intelligent adaptive control strategy is established to achieve dynamic optimization and adjustment of system operating parameters. The control strategy adopts a hierarchical control architecture, including two levels: upper-level optimization control and lower-level execution control.

[0058] The hierarchical control strategy consists of three layers: the upper layer, the global optimization layer, responsible for formulating the overall system operation strategy and parameter settings; the middle layer, the coordination control layer, responsible for the coordination and cooperation between subsystems; and the lower layer, the local control layer, responsible for the operation and control of specific equipment. Information exchange and command transmission between these layers are achieved through standardized communication interfaces, forming a complete hierarchical control system.

[0059] The upper-level optimization control employs a multi-objective optimization algorithm, comprehensively considering multiple objectives such as pollutant removal efficiency, energy consumption cost, and system stability. It uses intelligent optimization methods such as genetic algorithms and particle swarm optimization to solve for the optimal operating parameters. The optimization objective function adopts a weighted summation form, with the weights of each objective dynamically adjusted according to actual operational needs.

[0060] The fuzzy control rules adopt the IF-THEN form, determining the output control quantity based on the fuzzy set of input variables. Input variables include influent water quality parameters, system load, and plant growth status, while output variables include water flow distribution ratio, residence time adjustment, and aeration control. The fuzzy set is defined using a trapezoidal membership function, with membership function parameters determined through expert knowledge and historical data.

[0061] Taking water flow distribution control as an example, when a high concentration of heavy metals in the influent is detected, the control rule is: IF the heavy metal concentration in the influent is "high" AND the treatment capacity of the mid-layer plants is "strong" THEN the water flow distribution ratio in the mid-layer is "increased". Fuzzy inference uses the Mamdani inference method, and defuzzification uses the centroid method to convert the fuzzy control quantity into a precise control output.

[0062] The lower-level execution control is responsible for specific equipment operation and parameter adjustment, including adjustable weir gate control, submersible mixer control, oxygenation equipment control, and chemical dosing pump control. The execution control employs a proportional-integral-derivative (PID) control algorithm to ensure control accuracy and system stability.

[0063] A PID controller regulates the system through three components: proportional, integral, and derivative. The proportional component adjusts instantly based on the current error magnitude, offering a fast response but potentially introducing steady-state error. The integral component accumulates historical errors, eliminating steady-state errors. The derivative component predicts error trends and adjusts in advance to improve system stability. Control parameters are dynamically adjusted based on system response characteristics using an adaptive tuning algorithm, ensuring good control performance under various operating conditions.

[0064] When a pollutant concentration exceeds the standard, the system automatically activates an emergency treatment mode. For example, when the heavy metal concentration exceeds a set threshold, the system automatically increases the proportion of water flow in the middle layer, prolonging the residence time in areas sensitive to heavy metals, and simultaneously activates the chemical precipitation auxiliary treatment system. When abnormal growth of a certain crop is detected, the system activates an alternative treatment plan, adjusting the treatment load of other crops to ensure that the overall treatment effect is not affected.

[0065] The system also possesses reinforcement learning capabilities, continuously optimizing the control strategy through the Q-learning algorithm. A state-action-reward mechanism is established, using pollutant removal efficiency, system stability, and operating cost as comprehensive evaluation indicators. Through interactive learning with the environment, control decisions are gradually optimized. The algorithm evaluates the long-term benefits of various control actions under different states using a value function, and combines learning rate and discount factor parameters to achieve continuous improvement and optimization of the control strategy.

[0066] Furthermore, deep reinforcement learning methods can be employed to further enhance the intelligence level of control strategies. A continuous control strategy based on a deep deterministic policy gradient algorithm can be constructed to achieve precise control of complex nonlinear systems. This method combines the advantages of value function approximation and policy gradient methods, enabling it to handle control problems in high-dimensional continuous action spaces.

[0067] Furthermore, a distributed control architecture based on multi-agent reinforcement learning can be adopted to replace the traditional centralized control method. The system is decomposed into multiple cooperating agents, each responsible for the control task of a specific subsystem. Global optimal control is achieved through communication and coordination between the agents. The multi-agent system employs an actor-critic network architecture, with each agent containing a policy network and a value network.

[0068] The coordination mechanism among agents employs a consensus algorithm to ensure that the control strategies of each agent converge to the global optimum. The consensus algorithm achieves distributed optimization through information exchange and parameter updates among agents. Each agent updates its control strategy based on the policy parameters of its neighbors and its own gradient information, achieving global consistency through iterative optimization.

[0069] The communication network is modeled using graph neural networks, which can handle dynamically changing network topologies. Graph neural networks realize information exchange between agents through message passing mechanisms, and achieve effective propagation and fusion of information in the network through node feature updates and message aggregation operations.

[0070] The purpose of this alternative embodiment is to build a more flexible and intelligent distributed control system, thereby improving the system's adaptability and robustness.

[0071] Step 5: Regularly add microbial agents according to the optimized operating parameters to carry out system bio-enhancement and dynamic regulation optimization, and establish a stable microbial ecosystem; Specifically, the following steps are included: System bioaugmentation employs targeted culture and inoculation of dominant microbial communities to enhance the system's biological treatment capacity. By analyzing the biodegradation pathways of different pollutants, microbial strains with specific degradation functions are screened, including functional groups such as nitrifying bacteria, denitrifying bacteria, polyphosphate-accumulating bacteria, and methanogens. Highly active microbial communities are cultivated using biofilm reactor technology and then periodically added to the system to establish a stable microbial ecosystem.

[0072] The microbial inoculant application strategy determines the application area and dosage based on the type and concentration of pollutants. For heavy metal pollution, heavy metal accumulating bacteria, including strains of *Pseudomonas* and *Bacillus*, are added to the mid-layer emergent plant area at a concentration of 10^6-10^7 CFU / mL. For organic pollutants, organic pollutant degrading bacteria, including strains of *Acinetobacter* and *Rhodococcus*, are added to all layers. For nitrogen and phosphorus nutrients, rhizosphere growth-promoting bacteria, such as nitrogen-fixing bacteria and phosphate-solubilizing bacteria, are added to promote the absorption and utilization of nutrients by plants.

[0073] The microbial agent is added using a combination of batch and continuous addition. During the initial system startup, large batches are added to quickly establish the microbial community; during stable operation, small, continuous additions are used to maintain microbial activity. The addition frequency is dynamically adjusted based on system load and treatment effect, generally 1-2 times per week.

[0074] Microbial community structure optimization follows niche theory, achieving spatial distribution optimization of different functional bacterial groups by regulating environmental conditions (dissolved oxygen, pH, temperature, nutrient ratio, etc.). The aerobic zone primarily cultivates nitrifying bacteria and aerobic heterotrophic bacteria, the anoxic zone cultivates denitrifying bacteria, and the anaerobic zone cultivates methanogens and sulfate-reducing bacteria. These functional zones are organically connected through rational hydraulic design, forming a complete biological treatment chain.

[0075] Dynamic regulation and optimization are based on real-time monitoring data, employing an adaptive control strategy to dynamically adjust system operating parameters. A multi-objective optimization model is established, with the objective functions of maximizing pollutant removal efficiency, minimizing operating costs, and optimizing system stability. A genetic algorithm is used to solve for the optimal combination of operating parameters. Optimized parameters include key control variables such as hydraulic retention time at each layer, dissolved oxygen concentration, pH value, and nutrient dosage.

[0076] During system operation, the system adjusts its operating strategy in real time based on changes in influent water quality and feedback on treatment effectiveness. When an increase in the concentration of a specific pollutant is detected, the system automatically enhances the corresponding biological treatment functions, such as increasing the dosage of specific functional bacteria or adjusting the operating parameters of the corresponding treatment units. Simultaneously, an early warning mechanism is established to promptly activate emergency procedures when system operating parameters deviate from the normal range, ensuring stable system operation.

[0077] Rhizosphere growth-promoting bacteria are mainly applied to areas with dense plant roots. They promote plant growth and pollutant absorption by secreting plant hormones and improving the rhizosphere environment. Strains with multiple growth-promoting functions, such as Bacillus and Pseudomonas, are selected, and the concentration of the bacterial agent is controlled within the range of 10⁶ to 10⁸ CFU per milliliter.

[0078] Heavy metal accumulating bacteria are added to the intermediate layer of the treatment system where heavy metal contamination is high, enhancing the removal efficiency through mechanisms such as biosorption and bioprecipitation. Strains with high tolerance and high accumulation capacity are selected, such as certain sulfate-reducing bacteria and iron-oxidizing bacteria.

[0079] Organic pollutant-degrading bacteria are added to treatment areas with high organic loads to accelerate the biodegradation of organic pollutants through enzymatic reactions. Strains with broad-spectrum degradation capabilities are selected, such as certain Pseudomonas and Acinetobacter species.

[0080] A system for evaluating the effectiveness of microbial agents was established, and molecular biology techniques were used to monitor the survival status and functional expression of the agents within the system. Quantitative real-time PCR was employed to detect the expression levels of specific genes and assess the activity and function of the agents. Based on the evaluation results, the agent dosing strategy was dynamically adjusted to ensure the continuous and stable bioaugmentation effect.

[0081] Implementing allelopathic enhancement technology involves regulating the composition and concentration of plant root exudates to enhance the synergistic allelopathic effects among different plants. For example, organic acids secreted by rice roots can activate phosphorus in the soil, which is beneficial for phosphorus absorption by other plants; phenolic compounds secreted by reed roots have antibacterial effects, which can reduce the harm of pathogens to other plants.

[0082] Establish a mycorrhizal symbiosis promotion mechanism by inoculating with arbuscular mycorrhizal fungi to form a plant-fungus symbiotic system. Mycorrhizal fungi can expand the effective absorption area of ​​plant roots, enhance the absorption capacity of nutrients and heavy metals, and improve the plant's stress resistance. Mycorrhizal fungi species with high tolerance to heavy metals, such as certain strains of the genera *Gloydium* and *Aphelenchus*, should be selected.

[0083] Furthermore, synthetic biology methods can be used to construct microbial strains with enhanced functions. Through genetic engineering, multiple functional genes can be integrated into a single strain to create engineered strains with multiple functions. For example, heavy metal resistance genes, organic pollutant degradation genes, and plant growth-promoting genes can be integrated into the same host bacterium to obtain superstrains with comprehensive functions.

[0084] Furthermore, CRISPR-Cas gene editing technology can be used to precisely modify microbial metabolic pathways, replacing traditional random mutation breeding methods. By designing specific guide RNAs, target gene sites can be precisely cut, and specific DNA sequences can be inserted or deleted, achieving targeted modification of microbial functions. The CRISPR-Cas system consists of two core components: Cas proteins and guide RNAs.

[0085] Guide RNAs are designed according to specific rules, with a target sequence length of 20 nucleotides, immediately adjacent to the PAM (NGG) sequence. Guided by the guide RNA, the Cas9 protein binds to the target DNA sequence, creating a double-strand break. The cell's DNA repair mechanisms then repair the break through homologous recombination or non-homologous end joining, thus achieving precise gene editing.

[0086] Gene editing efficiency was assessed using quantitative PCR and sequencing technologies. The percentage of editing efficiency was calculated by statistically analyzing the ratio of the number of successfully edited cells to the total number of cells.

[0087] Strategies for constructing multifunctional engineered strains include: enhancing heavy metal resistance by overexpressing proteins such as metallothionein and ATPase; strengthening the ability to degrade organic matter by introducing exogenous degradation enzyme genes such as monooxygenase and dioxygenase; and improving plant growth promotion by increasing the yield of plant hormones and enzymes such as IAA and ACC deaminase.

[0088] The safety assessment of engineered strains includes genetic stability testing, environmental release risk assessment, and biosafety testing. A safety switch mechanism for the genetic circuit is established to ensure that the engineered strains can only survive and function under specific conditions. The purpose of this alternative embodiment is to create highly efficient functional microorganisms through precise genetic engineering techniques, thereby enhancing bioaugmentation effects while ensuring biosafety.

[0089] Step 6: Integrate the functional units into a modular system, perform performance verification based on a stable microbial ecosystem, and obtain purified water; Specifically, the following steps are included: The system integrates multi-crop co-processing units, intelligent monitoring units, and adaptive control units into a standardized modular system. Each processing module is designed as an independent operating unit, with standardized inlet and outlet water interfaces, electrical interfaces, and communication interfaces, which can be flexibly combined and expanded according to actual processing needs.

[0090] Modular system integration follows standardized design principles, with each module having a unified interface specification and communication protocol. Standardized modules include crop collaborative modeling units, system structure design units, intelligent monitoring units, adaptive control units, bioaugmentation units, and system integration units. Data exchange and functional coordination between these units are achieved through standardized interfaces, forming a complete modular system architecture.

[0091] The interface specification comprises three layers: physical interface, data interface, and control interface. The physical interface employs standardized piping and electrical connections; the data interface uses a unified communication protocol and data format; and the control interface uses standardized control signals and instruction sets. This interface standardization enables plug-and-play functionality and flexible configuration between different modules.

[0092] Each standard module is designed to process 100 to 200 cubic meters per day, occupying an area of ​​approximately 200 to 300 square meters. Internally, each module contains a complete three-layer crop processing system, intelligent monitoring equipment, and an automatic control system. Modules are connected in series or parallel via standardized piping, allowing the total system processing capacity to be linearly expanded based on the number of modules.

[0093] Establish engineering design standards for modular systems, including module size specifications, equipment configuration standards, installation and construction specifications, and operation and maintenance procedures. Develop a quality control system to ensure the performance consistency of each module and the reliability of system integration.

[0094] The system's performance will be evaluated through pilot-scale testing. A pilot plant with a treatment capacity of 50 cubic meters per day will be constructed and operated continuously for more than 6 months. During the pilot-scale testing, simulated industrial wastewater and actual industrial wastewater will be used to verify the system's treatment effectiveness for different types of pollutants.

[0095] The main performance indicators for pilot-scale verification include: heavy metal removal rate; chemical oxygen demand (COD) removal rate; ammonia nitrogen removal rate; and total phosphorus removal rate. The system's operational stability is required to have a continuous stable operating time of over 90% and a downtime of less than 5% due to malfunctions.

[0096] A field demonstration project will be conducted, constructing a system with a processing capacity of 500 cubic meters per day in an actual industrial park to verify the system's engineering practicality and economic feasibility. The demonstration project will operate for more than one year, accumulating complete operational data and experience to provide a reliable basis for technology promotion.

[0097] Furthermore, a virtual verification platform based on digital twin technology can be established. This platform constructs precise digital models of the system, including hydraulic, biological, and control system models, and uses simulation calculations to predict the system's performance under different operating conditions. The digital twin platform can support various application scenarios such as system design optimization, operational strategy formulation, and fault diagnosis and analysis, thereby improving the system's engineering application level.

[0098] Furthermore, blockchain technology can be used to build a distributed system performance verification and certification platform, replacing traditional centralized verification methods. The immutability of blockchain ensures the authenticity and traceability of performance test data, establishing a trustworthy technical verification system. The blockchain network adopts a consortium blockchain architecture, jointly maintained by equipment manufacturers, third-party testing institutions, users, and other parties.

[0099] Smart contracts are used to automate performance verification processes, including test plan development, data collection, result evaluation, and certificate issuance. The core logic of a smart contract includes a performance metric verification function that automatically triggers the authentication process when system performance meets preset standards. Performance evaluation employs a multi-dimensional evaluation system, comprehensively considering indicators such as processing efficiency, stability, and cost-effectiveness.

[0100] The blockchain consensus mechanism employs a practical Byzantine fault-tolerant algorithm to ensure network security and consistency. The consensus process consists of three phases: pre-preparation, preparation, and commit. Digital signature technology is used to guarantee message integrity and non-repudiation during message transmission between nodes.

[0101] Before data is uploaded to the blockchain, a hash algorithm is used to generate a data fingerprint to ensure data integrity. The hash function uses the SHA-256 algorithm, which combines the original data, timestamp, and random value to generate a unique 256-bit hash value.

[0102] The implementation process of the SHA256 algorithm includes: preprocessing the input data, including message padding and length appending, to make the message length a multiple of 512 bits; decomposing the message into 512-bit data blocks, each of which is processed through 64 rounds of iterative operations; each round of operations uses different constants and logical functions, including bitwise operations, modulo addition operations, etc.; finally, a 256-bit hash digest is output, which has the avalanche effect characteristic, that is, a small change in the input data will lead to a difference in the output hash value, ensuring the reliability of data integrity verification.

[0103] The blockchain platform also integrates IoT technology to automate data collection and uploading from testing equipment. Sensor data is transmitted to blockchain nodes via an encrypted channel, ensuring data transmission security. The purpose of this alternative embodiment is to establish a trusted technology verification system, enhance the authority and credibility of system certification, and promote the standardization and industrial application of the technology.

[0104] A wastewater filtration system based on multi-crop synergy is used to perform the aforementioned wastewater filtration method based on multi-crop synergy, such as... Figure 2 As shown, it includes: The crop adaptability assessment and collaborative modeling unit is used to acquire raw sewage samples, detect water quality parameters of the raw sewage samples, acquire physiological characteristic data of candidate crops, analyze and process the water quality parameters and crop characteristic data using a multi-factor comprehensive evaluation method, obtain the environmental adaptability index of each crop, and establish a multi-crop collaborative pollutant removal mechanism model based on the environmental adaptability index. A multi-level system structure building unit is constructed based on a multi-crop synergistic pollutant removal mechanism model, which includes a floating plant layer, an emergent plant layer, and a submerged plant layer. The multi-level system structure is analyzed and calculated using a three-dimensional hydraulic modeling method to obtain an accurate water flow distribution scheme. The multi-sensor fusion monitoring unit constructs a multi-sensor fusion monitoring network based on the water flow distribution scheme, processes and analyzes the monitoring data using a data fusion algorithm, and obtains the system operation status monitoring results. The intelligent adaptive control unit optimizes and adjusts system parameters based on the system operation status monitoring results using an intelligent adaptive control strategy to obtain optimized operating parameters. The bioaugmentation and dynamic regulation unit periodically adds microbial agents according to optimized operating parameters to optimize system bioaugmentation and dynamic regulation, and establishes a stable microbial ecosystem. The system integration and performance verification unit performs performance verification based on a stable microbial ecosystem to obtain purified water.

[0105] In one embodiment of the present invention, an application example of a wastewater filtration method based on multi-crop synergy is provided: A 45-day field test was conducted at a wastewater treatment plant in an industrial park. During the test, a small-scale multi-crop co-processing wastewater filtration system was deployed between the secondary treatment unit and the final discharge unit. The test area was 200 square meters, divided into three parallel treatment units, each equipped with a complete three-layer plant structure and monitoring and control equipment. The test system had a daily treatment capacity of 300 tons, primarily targeting mixed wastewater containing heavy metals and organic pollutants. In the initial stage of the test, two sets of key data instances were acquired for initial system configuration and operational optimization.

[0106] Table 1 shows an example of the data obtained for the influent water quality parameters: Table 1: Examples of data acquisition for influent water quality parameters;

[0107] Examples of crop environmental adaptability assessment data are shown in Table 2: Table 2: Examples of crop environmental adaptability assessment data;

[0108] The obtained water quality parameter data shows that the wastewater quality in the industrial park fluctuates significantly, with the concentration of various pollutants varying between 20% and 40%, consistent with the complex and variable composition of industrial park wastewater. Crop environmental adaptability assessment data shows that different crops exhibit significant differences in their adaptability to water quality conditions. The adaptability score calculated using a multi-factor comprehensive evaluation method provides a scientific basis for system configuration. These two sets of data examples lay the foundation for subsequent synergy index calculations and the construction of a multi-level crop configuration system.

[0109] like Figure 3 As shown, the advantages of the multi-crop synergistic system over the single-crop treatment mode are demonstrated. With the increase in crop varieties and the improvement of the synergistic mechanism, the removal rate of various pollutants shows a significant upward trend. The multi-crop synergistic mode achieves optimal results in the removal of all pollutants, especially in the removal of heavy metals, where the synergistic effect is better, and the removal rate is more than 30% higher than that of the single-crop mode.

[0110] like Figure 4 As shown, after 12 months of continuous operation, the effluent quality of the system remained stable, with the COD concentration maintained between 28 and 34 mg / L and the ammonia nitrogen concentration controlled between 1.4 and 2.1 mg / L, both far below the emission standard limits. The system's operational stability index consistently remained above 0.9, indicating that the system has excellent resistance to shock loads and operational stability.

[0111] like Figure 5 As shown, this demonstrates the comprehensive advantages of multi-crop synergy technology compared to traditional technologies. While multi-crop synergy technology exhibits significant advantages in terms of construction investment, operating costs, and maintenance expenses—reducing construction investment by 30%, operating costs by 40%, and maintenance expenses by 50%—it also demonstrates outstanding performance in treatment efficiency, environmental benefits, and resource recovery. Treatment efficiency is improved by 30%, environmental benefits are somewhat improved, and resource recovery rates are increased. This technology possesses certain advantages in environmental friendliness and sustainable development, making it an important supplement and development direction for traditional wastewater treatment technologies.

[0112] The embodiments of the present invention have been described above. However, the embodiments are not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make more equivalent embodiments under the guidance of the present embodiments, and all of them are within the protection scope of the present embodiments.

Claims

1. A wastewater filtration method based on multi-crop synergy, characterized in that, Includes the following steps: Raw wastewater samples were obtained, and water quality parameters were tested on the raw wastewater samples. At the same time, physiological characteristic data of candidate crops were obtained. The water quality parameters and crop characteristic data were analyzed and processed using a multi-factor comprehensive evaluation method to obtain the environmental adaptability index of each crop. Based on the environmental adaptability index, a multi-crop synergistic pollutant removal mechanism model was established. Based on the multi-crop synergistic pollutant removal mechanism model, a multi-level system structure including a floating plant layer, an emergent plant layer, and a submerged plant layer was constructed. The multi-level system structure was analyzed and calculated using a three-dimensional hydraulic modeling method to obtain an accurate water flow distribution scheme. A multi-sensor fusion monitoring network is constructed based on the water flow distribution scheme. The monitoring data is processed and analyzed using a data fusion algorithm to obtain the system operation status monitoring results. Based on the system operation status monitoring results, an intelligent adaptive control strategy is adopted to optimize and adjust the system parameters, thereby obtaining optimized operating parameters; Based on optimized operating parameters, microbial agents are periodically added to conduct system bioaugmentation and dynamic regulation optimization, and a stable microbial ecosystem is established. The functional units are integrated into a modular system, and the performance is verified based on a stable microbial ecosystem to obtain purified water.

2. The wastewater filtration method based on multi-crop synergy according to claim 1, characterized in that, The water quality parameters and crop characteristic data were analyzed using a multi-factor comprehensive evaluation method to obtain the environmental adaptability index of each crop, including: The root morphology characteristics, pollutant absorption kinetic parameters, and growth cycle characteristics of candidate crops were obtained. The root surface area and root density parameters were determined by morphological analysis. The maximum absorption rate and half-saturation constant were obtained by fitting the pollutant absorption kinetic parameters using the Michaelis-Menten equation. Based on water quality parameters such as pH value, heavy metal concentration, and organic matter content, the adaptability of each candidate crop is calculated and analyzed using fuzzy mathematics evaluation method. The adaptability is quantified by Gaussian membership function to obtain the single-factor adaptability score of each environmental factor. The single-factor adaptability score was comprehensively calculated using a weighted comprehensive evaluation method. The corresponding weight coefficients were set according to the degree of influence of each environmental factor on crop growth, and the comprehensive environmental adaptability index of each crop was calculated.

3. The wastewater filtration method based on multi-crop synergy according to claim 1, characterized in that, The multi-crop synergistic pollutant removal mechanism model established based on the environmental adaptability index includes: The top three crops in the comprehensive environmental adaptability index ranking were obtained. The pollutant absorption mechanism of the crops was analyzed using biochemical kinetic methods. Based on the pollutant characteristic data, the absorption kinetic equations of each crop for different pollutants were established. The kinetic parameters in the absorption kinetic equations were obtained through experimental measurement methods. Based on the absorption kinetic equation and kinetic parameters, the synergistic mechanism among crops is modeled and analyzed using systems biology methods. The nutrient competition relationship and pollutant transport pathway among crops are determined through the analysis of rhizosphere microenvironment interactions, and a mathematical model for multi-crop synergistic pollutant removal is established. The accuracy of the mathematical model for the synergistic removal of pollutants by multiple crops was verified by experimental methods. The reliability of the model was evaluated by comparing the deviation between the predicted values ​​of the mathematical model and the actual measured values. Based on the verification results, the parameters of the mathematical model were optimized and adjusted to obtain a verified mechanism model for the synergistic removal of pollutants by multiple crops.

4. The wastewater filtration method based on multi-crop synergy according to claim 1, characterized in that, The multi-layered system structure, comprising a floating plant layer, an emergent plant layer, and a submerged plant layer, constructed based on the multi-crop synergistic pollutant removal mechanism model, includes: The optimal crop configuration scheme was determined based on a validated multi-crop synergistic pollutant removal mechanism model. A surface floating plant layer was set up according to the ecological niche characteristics and pollutant removal capacity of each crop in the configuration scheme. The floating plants were selected to perform preliminary removal treatment of organic pollutants and nitrogen and phosphorus nutrients in the wastewater. Based on the water quality after surface treatment, a middle layer of emergent plants is set according to the emergent plant configuration parameters in the configuration scheme, and emergent plants are selected to further remove heavy metal ions and residual organic pollutants. Based on the water quality data after the middle layer treatment, the bottom layer of submerged plants is set according to the submerged plant configuration parameters in the configuration scheme. Submerged plants are selected to carry out deep purification treatment of sewage. Through the synergistic effect of the vertical distribution of the three layers of plants, a multi-level system structure with clear functional division is obtained.

5. The wastewater filtration method based on multi-crop synergy according to claim 1, characterized in that, The analysis and calculation of the multi-level system structure using three-dimensional hydraulic modeling to obtain an accurate water flow distribution scheme includes: Based on the geometric parameters and boundary conditions of the multi-level system structure, a three-dimensional hydraulic numerical model is established, and the flow field distribution inside the system is obtained by solving the Navier-Stokes equations. Based on the flow field distribution, an artificial neural network algorithm is used to optimize the hydraulic residence time calculation and determine the optimal hydraulic residence time allocation for each layer of vegetation. Based on the optimal hydraulic residence time allocation, the water distribution system is optimized by combining computational fluid dynamics and machine learning methods to obtain a water flow distribution scheme that can achieve uniform water distribution and optimal residence time.

6. The wastewater filtration method based on multi-crop synergy according to claim 1, characterized in that, The multi-sensor fusion monitoring network constructed based on the water flow distribution scheme includes: Based on the spatial distribution characteristics of the multi-level system structure, water quality sensors, soil sensors, meteorological sensors and plant growth sensors are deployed in the plant areas of each level to form a multi-sensor monitoring network covering the entire system area. Based on multi-source heterogeneous data collected by a multi-sensor monitoring network, a hierarchical fusion architecture including data layer fusion, feature layer fusion, and decision layer fusion is constructed. The Kalman filter algorithm is used to preprocess and filter out noise from the sensor data. Based on a layered fusion architecture, deep learning methods are used to intelligently analyze and process the fused data to obtain system operation status monitoring results.

7. A wastewater filtration method based on multi-crop synergy according to claim 2, characterized in that, The water quality parameters and crop characteristic data were analyzed using a multi-factor comprehensive evaluation method to obtain the environmental adaptability index of each crop, including: Based on the water quality parameters of wastewater samples and the physiological characteristic parameters of crop data, the weight coefficients of each evaluation factor were determined by the analytic hierarchy process (AHP) and a comprehensive evaluation index system including water quality adaptability, growth adaptability and stress resistance was established. Based on a comprehensive evaluation index system, the fuzzy comprehensive evaluation method is used to quantitatively assess the adaptability of each crop under different wastewater environmental conditions, and the environmental adaptability index of each crop is obtained.

8. A wastewater filtration method based on multi-crop synergy according to claim 3, characterized in that, The multi-crop synergistic pollutant removal mechanism model includes: Based on environmental adaptability index and pollutant characteristic data, a mathematical model describing the interaction between crops is established. The mathematical model describing the interaction between crops includes the nutrient competition coefficient, root exudate interaction coefficient and pollutant removal synergy coefficient among crops. Based on a mathematical model, the optimal configuration ratio and spatial distribution scheme of each crop are solved by numerical calculation method, resulting in a multi-crop synergistic pollutant removal mechanism model that can maximize the synergistic removal effect of pollutants.

9. A wastewater filtration method based on multi-crop synergy according to claim 6, characterized in that, The deep learning methods include: Based on the multi-source sensor data processed by the hierarchical fusion architecture, a deep learning model containing a convolutional neural network and a long short-term memory network is constructed. The convolutional neural network is used to extract spatial features, and the long short-term memory network is used to capture temporal features. The system operates by training and learning fused data using a deep learning model and optimizing network parameters through backpropagation. This enables intelligent identification and classification of system operating status, resulting in system operating status monitoring results that include normal operation, anomaly warnings, and fault diagnosis.

10. A wastewater filtration system based on multi-crop synergy, characterized in that, A wastewater filtration method based on multi-crop synergy as described in any one of claims 1-9, comprising: The crop adaptability assessment and collaborative modeling unit is used to acquire raw sewage samples, detect water quality parameters of the raw sewage samples, acquire physiological characteristic data of candidate crops, analyze and process the water quality parameters and crop characteristic data using a multi-factor comprehensive evaluation method, obtain the environmental adaptability index of each crop, and establish a multi-crop collaborative pollutant removal mechanism model based on the environmental adaptability index. A multi-level system structure building unit is constructed based on a multi-crop synergistic pollutant removal mechanism model, which includes a floating plant layer, an emergent plant layer, and a submerged plant layer. The multi-level system structure is analyzed and calculated using a three-dimensional hydraulic modeling method to obtain an accurate water flow distribution scheme. The multi-sensor fusion monitoring unit constructs a multi-sensor fusion monitoring network based on the water flow distribution scheme, processes and analyzes the monitoring data using a data fusion algorithm, and obtains the system operation status monitoring results. The intelligent adaptive control unit optimizes and adjusts system parameters based on the system operation status monitoring results using an intelligent adaptive control strategy to obtain optimized operating parameters. The bioaugmentation and dynamic regulation unit periodically adds microbial agents according to optimized operating parameters to optimize system bioaugmentation and dynamic regulation, and establishes a stable microbial ecosystem. The system integration and performance verification unit performs performance verification based on a stable microbial ecosystem to obtain purified water.

Images