Micro-plastic pollution treatment effectiveness detection method based on big data coupling

Abstract

The invention relates to the technical field of data detection, and discloses a micro-plastic pollution abatement effectiveness detection method based on big data coupling. In the aspect of pollution traceability, the method adopts a mode of combining multi-dimensional feature acquisition and a feature evolution model; the problem of traceability failure caused by feature drift in a traditional method is solved, the traceability effective time is prolonged, and the traceability accuracy is improved. In the aspect of migration prediction, the multi-process coupling model and physical and chemical parameter dynamic association mechanism established by the method can accurately simulate complex interaction behaviors in the micro-plastic transportation process, and the prediction error of a migration path is reduced. In the aspect of governance evaluation, the governance process conversion behavior analysis method and the multi-dimensional evaluation system established by the method can accurately distinguish true removal, pollution transfer and form conversion, and the situation that a governance scheme with appearance reaching the standard but the actual risk not reduced is misjudged as an effective scheme is avoided.

Description

Technical Field

[0001] This invention relates to the field of data detection technology, specifically to a method for detecting the effectiveness of microplastic pollution control based on big data coupling. Background Technology

[0002] Microplastic pollution control refers to the process of reducing the environmental concentration of plastic particles smaller than 5 mm using methods such as physical interception, ecological restoration, and source control. Microplastics undergo degradation, breakage, surface aging, and biofilm formation in the environment, and their concentration changes may stem from remediation reduction, pollution transfer, or form transformation. Current technologies struggle to accurately distinguish between these factors, easily misinterpreting spatial migration or form changes of pollutants as effective remediation, leading to distorted assessments and biased decision-making. Therefore, there is an urgent need to develop accurate methods for detecting effectiveness.

[0003] However, existing effectiveness detection methods still have the following problems: Pollution source tracing methods rely on comparisons of characteristic parameters such as polymer type and additive composition. However, these characteristic parameters change due to factors such as photochemical reactions, mechanical wear, and microbial activity in the environment. When the monitoring time span is long or the pollutant transport distance is long, the correlation between the detected characteristics and the initial characteristics of the pollution source decreases significantly, affecting the accuracy of source tracing. Furthermore, existing migration and transformation models calculate hydrodynamic transport, chemical degradation, and biological processes as independent units and then superimpose the results, without fully considering the coupling effects between these processes. For example, the impact of particle density changes caused by biofilm attachment on sedimentation behavior, and the accelerated effect of structural weakening due to surface aging on mechanical breakage, lead to discrepancies between model predictions and actual conditions. Moreover, the lack of systematic analysis of the dynamic transformation process of microplastics in the assessment of treatment effectiveness may lead to misjudging secondary particle breakage during treatment as pollution reduction, or attributing natural sedimentation caused by changes in environmental conditions to the effects of treatment measures, resulting in inaccurate assessments of treatment effectiveness.

[0004] Therefore, we propose a big data-coupled method for detecting the effectiveness of microplastic pollution control in order to address the aforementioned problems. Summary of the Invention

[0005] The purpose of this invention is to provide a method for detecting the effectiveness of microplastic pollution control based on big data coupling, in order to solve the problem mentioned in the background art that pollution source tracing methods rely on comparison and identification of characteristic parameters such as polymer type and additive composition. However, microplastics in the environment cause these characteristic parameters to change due to factors such as photochemical reactions, mechanical wear, and microbial action. When the monitoring time span is long or the pollutant transmission distance is long, the correlation between the detected characteristics and the initial characteristics of the pollution source is significantly reduced, affecting the accuracy of source tracing.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for detecting the effectiveness of microplastic pollution control based on big data coupling, the specific steps of which are as follows: S1. Multidimensional feature acquisition: Establish a multi-point sampling network in the target area to collect three types of characteristic parameters of microplastics: inorganic elements, organic polymers and bio-attachment. Accelerated aging experiments are conducted in the laboratory under different environmental conditions to establish a database linking initial characteristics, environmental conditions and evolutionary characteristics. S2. Evolutionary trajectory modeling: A feature evolution model is established using Bayesian networks and long short-term memory neural networks. The initial features and environmental history are used as inputs, and the predicted values ​​and confidence intervals of the current feature parameters are output. S3. Pollution Source Identification: Input the current characteristics and migration path of the monitored samples into the evolution model, use Bayesian inversion to identify the location and emission time of pollution sources, and output the probability distribution of pollution sources. S4. Parameter Dynamic Correlation: Establish quantitative relationships between physicochemical parameters such as microplastic density and particle size and characteristic parameters, and dynamically update physicochemical parameters based on changes in characteristics during migration calculations; S5. Multi-process coupling: Construct a coupled migration model that includes hydrodynamics, chemical degradation and biological processes, and realize the mutual influence of each process through a two-way feedback mechanism; S6. Adaptive Model Correction: The monitoring data is incorporated into the coupled model using ensemble Kalman filtering to achieve adaptive correction of the model parameters. S7. Analysis of the treatment process: Monitor the characteristic changes of microplastics in the treatment facility, establish a quantitative relationship between treatment intensity and breakage rate, and distinguish between actual removal and phase transformation; S8. Multidimensional evaluation of effectiveness: Set up monitoring sections at multiple interfaces of the treatment facilities, and quantify the net contribution of the treatment measures through material balance and scenario comparison. S9. Strategy Optimization Decision: Integrate the results of pollution source identification, migration prediction and effect evaluation to build a decision support system to achieve optimized adjustment of governance strategies and early warning response.

[0007] Preferably, in step S1, the specific steps for multidimensional feature acquisition are as follows: S1.1 Sampling points were set up in the monitoring area, and the inorganic element characteristics, organic polymer characteristics, and bio-attachment characteristics of microplastic samples were detected. The inorganic element characteristics were obtained by single-particle inductively coupled plasma mass spectrometry, the organic polymer characteristics were obtained by Fourier transform infrared spectroscopy and Raman spectroscopy, and the bio-attachment characteristics were obtained by high-throughput sequencing technology. Environmental parameters such as temperature, light and hydrodynamic conditions were recorded to form initial characteristic data. S1.2. The control sample is subjected to simulated light, hydrodynamics and microbial activity under experimental conditions. Data on the changes of characteristic parameters over time are obtained through accelerated aging experiments. The correspondence between the initial characteristic parameters, environmental simulation conditions and evolutionary characteristic parameters is organized into a correlation database to generate a characteristic parameter evolution benchmark.

[0008] Preferably, in step S2, the evolutionary trajectory modeling is performed in the following specific way: S2.1 Based on the associated database of step S1, temperature, light intensity, hydrodynamic conditions and microbial activity are selected as environmental factors. The rate of change of characteristic parameters is statistically analyzed. A Bayesian network is used to establish a dependency model between environmental factors and the rate of change of characteristic parameters. The change law of characteristic parameters of different polymer types under the action of environmental factors is modeled to form the evolution relationship of characteristic parameters. S2.2. The evolutionary relationship in step S2.1 is trained using a long short-term memory neural network to obtain the temporal evolution trajectory of the feature parameters under different environmental conditions. Based on transfer learning, the accelerated aging data in the laboratory is corrected to construct a feature evolution prediction model with the initial feature parameters and environmental history as input, and output the predicted value of the feature parameters and the confidence interval at the target time.

[0009] Preferably, in step S3, the specific method for identifying the pollution source is as follows: S3.1 Based on the microplastic sample data obtained from monitoring, the current characteristic parameters are integrated with the migration path calculated by the hydrological model to form inversion input data. This inversion input data is then input into the feature evolution model in step S2 to generate feature evolution results corresponding to different pollution source scenarios. S3.2. The Bayesian inversion method is used to perform parameter space calculation on the feature evolution results generated in step S3.1. The parameter space includes the initial features of the pollution source, migration path, environmental evolution process and current detection features. The likelihood value of each pollution source scenario is calculated by combining the Markov chain Monte Carlo sampling method to determine the location and emission time of the pollution source and generate probability distribution and confidence level.

[0010] Preferably, in step S4, the specific method for dynamically associating parameters is as follows: S4.1 Based on the characteristic parameter data obtained in step S1, a quantitative relationship between biofilm coverage and particle equivalent density is established through density gradient centrifugation experiment; a relationship between particle size distribution changes under aging and crushing is established through laser particle size analysis; a quantitative relationship between surface oxidation degree and hydrophobicity changes is established through contact angle measurement; and the correspondence between density, particle size, surface roughness, hydrophobicity and characteristic parameters is organized into a physicochemical parameter correlation model. S4.2. Embed the physicochemical parameter association model from step S4.1 into the migration calculation model. In each time step of the migration calculation, update the density, particle size, surface roughness, and hydrophobicity based on the feature parameter prediction results from step S2, so that the physicochemical parameters used in the migration calculation are adjusted synchronously with the changes in the feature parameters.

[0011] Preferably, in step S5, the specific steps of multi-process coupling are as follows: S5.1. A multi-process coupled migration model is constructed based on hydrodynamic transport, chemical degradation and biological processes. The operator splitting method is used to decompose each process into independent sub-modules and solve them separately. The hydrodynamic module solves the convection-diffusion equations using the finite volume method. The chemical degradation module is calculated based on surface oxidation kinetics. The biological process module is calculated based on biofilm growth kinetics and particle aggregation kinetics. S5.2 Establish a two-way feedback mechanism between sub-modules. The change in particle density caused by biofilm growth is fed back to correct the settling rate of the hydrodynamic module. The hydrodynamic shear stress is fed back to the bio-module as the driving force for biofilm peeling. The structural weakening caused by chemical aging is fed back to the mechanical breakage module. Each module performs iterative calculations in each calculation time step until the parameters converge.

[0012] Preferably, in step S6, the specific steps of model adaptive correction are as follows: S6.1 Based on the real-time data obtained from the monitoring network deployed in step S1, the concentration of microplastics and the distribution of characteristic parameters are used as inputs. The ensemble Kalman filter method is used to integrate the real-time data into the coupled migration model established in step 5, correct the model state variables and key parameters, and use variational interpolation technology to expand the point data into a spatial field to generate the input data required for adaptive correction. S6.2. Based on the updated data generated in step S6.1, detect the deviation between the model prediction result and the assimilated data. If the deviation exceeds the preset threshold, trigger sensitivity analysis, identify the master control deviation parameter and correct it to ensure that the model state is updated in subsequent calculation time steps, thereby improving prediction accuracy and reliability.

[0013] Preferably, the specific steps of step S7, the governance process analysis, are as follows: S7.1 Through on-site pilot testing and laboratory simulation, monitor the changes in particle size, morphology, mass concentration and number concentration of microplastics before and after entering treatment facilities such as interception nets and sedimentation tanks. Use high-speed photography and image analysis technology to observe the crushing process and establish a quantitative relationship between treatment intensity, crushing rate and progeny particle size distribution to provide data support for the analysis of microplastic conversion behavior. S7.2 Monitor the removal efficiency of microplastics of different particle sizes in each treatment unit, distinguish between actual removal and phase transformation, and analyze the transformation behavior of microplastics by comparing the removal effect and transformation process of each unit, so as to provide a basis for the evaluation of the treatment effect.

[0014] Preferably, in step S8, the specific steps for multidimensional evaluation of the effect are as follows: S8.1 Establish a multi-dimensional assessment system that includes mass conservation, particle size distribution change and ecological risk change. Set up monitoring sections at the inlet, outlet, sludge and sediment of the treatment facility to collect microplastic data. Calculate the input, output, removal, transformation and cumulative mass through material balance. Compare the distribution changes of each particle size range before and after treatment. Calculate the comprehensive risk index by combining the relationship between particle size and toxicity. S8.2 Extend the coupling model from step S5 to the internal environment of the treatment facility, simulate the natural decay scenario without treatment measures and the actual treatment scenario, compare the microplastic removal effect under the two scenarios, quantify the net contribution of treatment measures, and provide a basis for optimizing the treatment scheme.

[0015] Preferably, in step S9, the specific steps for strategy optimization decision-making are as follows: S9.1 Integrate the pollution source identification results obtained in step S3, the migration prediction results obtained in step S6, and the effect evaluation results obtained in step S8 to construct a decision support system. Simulate the long-term effects and costs of different governance schemes on the digital twin platform, and use a multi-objective genetic algorithm to optimize the combination of governance strategies under budget and ecological protection constraints to generate a recommended list of optimal schemes. S9.2 Based on the decision support system built in step S9.1, monitor the output of the prediction model. When the microplastic flux suddenly increases or the sensitive area exceeds the standard, trigger an early warning and recommend emergency measures. Establish a continuous monitoring mechanism to update the evaluation results regularly. When the actual effect deviates from the expectation by more than the threshold, initiate the cause diagnosis and strategy adjustment process.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. This big data-coupled method for detecting the effectiveness of microplastic pollution control addresses several challenges. In pollution source tracing, it employs a combination of multi-dimensional feature acquisition and a feature evolution model, overcoming the problem of source tracing failure caused by feature drift in traditional methods, thus extending the effective tracing time and improving accuracy. In migration prediction, the method's multi-process coupling model and dynamic correlation mechanism with physicochemical parameters accurately simulate the complex interactive behavior during microplastic transport, reducing prediction errors in migration paths. In governance assessment, the method's analysis of governance process transformation behavior and multi-dimensional evaluation system accurately distinguish between actual removal, pollution transfer, and form transformation, avoiding misjudging governance schemes that appear to meet standards but do not actually reduce risks as effective. In decision support, by integrating data from pollution source tracing, migration prediction, and effectiveness evaluation, this method achieves multi-objective optimization and dynamic adjustment of governance strategies, establishing a closed-loop management mechanism for monitoring, evaluation, optimization, and feedback.

[0017] 2. By employing Bayesian networks to quantify the impact of environmental factors on the rate of change of feature parameters, and combining this with long short-term memory neural networks to learn the temporal relationship between environmental condition sequences and feature parameters, a quantitative model capable of predicting the evolution of microplastic characteristics over time was constructed. Existing technologies often use qualitative descriptions or simplified linear assumptions when analyzing feature evolution, making it difficult to accurately describe the feature change process under fluctuating environmental conditions. The prediction model established in this step can calculate the temporal trajectory of feature parameters based on initial features and environmental history, improving the accuracy of feature evolution prediction. In terms of pollution source tracing, existing technologies typically compare currently detected features directly with features of potential pollution sources. However, after microplastics have undergone long-term environmental effects, their features have changed, and the accuracy of direct comparison for source tracing will significantly decrease, with the effective time usually limited to several weeks to several months. The feature evolution prediction model constructed in this step can infer the initial features based on the currently detected features and the environmental history along the migration path, improving the source tracing bias caused by feature drift and extending the effective time window for source tracing analysis. The confidence intervals output by the model quantify the uncertainty of the source tracing results, enabling source tracing analysis to shift from point estimation to interval estimation, thereby improving the credibility of source tracing judgments. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the overall three-dimensional structure of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1: Please refer to Figure 1 A method for detecting the effectiveness of microplastic pollution control based on big data coupling, the specific steps of which are as follows: S1. Multidimensional feature acquisition: Establish a multi-point sampling network in the target area to collect three types of characteristic parameters of microplastics: inorganic elements, organic polymers and bio-attachment. Accelerated aging experiments are conducted in the laboratory under different environmental conditions to establish a database linking initial characteristics, environmental conditions and evolutionary characteristics. S2. Evolutionary trajectory modeling: A feature evolution model is established using Bayesian networks and long short-term memory neural networks. The initial features and environmental history are used as inputs, and the predicted values ​​and confidence intervals of the current feature parameters are output. S3. Pollution Source Identification: Input the current characteristics and migration path of the monitored samples into the evolution model, use Bayesian inversion to identify the location and emission time of pollution sources, and output the probability distribution of pollution sources. S4. Parameter Dynamic Correlation: Establish quantitative relationships between physicochemical parameters such as microplastic density and particle size and characteristic parameters, and dynamically update physicochemical parameters based on changes in characteristics during migration calculations; S5. Multi-process coupling: Construct a coupled migration model that includes hydrodynamics, chemical degradation and biological processes, and realize the mutual influence of each process through a two-way feedback mechanism; S6. Adaptive Model Correction: The monitoring data is incorporated into the coupled model using ensemble Kalman filtering to achieve adaptive correction of the model parameters. S7. Analysis of the treatment process: Monitor the characteristic changes of microplastics in the treatment facility, establish a quantitative relationship between treatment intensity and breakage rate, and distinguish between actual removal and phase transformation; S8. Multidimensional evaluation of effectiveness: Set up monitoring sections at multiple interfaces of the treatment facilities, and quantify the net contribution of the treatment measures through material balance and scenario comparison. S9. Strategy Optimization Decision: Integrate the results of pollution source identification, migration prediction and effect evaluation to build a decision support system to achieve optimized adjustment of governance strategies and early warning response.

[0021] In this embodiment: Step S1 achieves comprehensive characterization of microplastic pollution characteristics by establishing a multi-point sampling network and a multi-dimensional feature parameter acquisition system. This step collects three types of feature parameters: inorganic elements, organic polymers, and bioattachment, overcoming the shortcomings of existing technologies that rely on only a single feature. When a certain type of feature changes due to environmental factors, other types of features can be used for cross-validation, improving the stability of feature identification. The correlation database established through accelerated aging experiments in the laboratory provides a quantitative benchmark for subsequent feature evolution analysis, enabling microplastic source tracing analysis to shift from qualitative judgment to quantitative calculation, thus improving the data foundation for source tracing accuracy.

[0022] The feature evolution model established in step S2 solves the technical challenge of predicting the changes in microplastic characteristics over time. This model employs Bayesian networks and long short-term memory neural networks, enabling dynamic prediction of feature parameters at the current moment based on initial features and environmental history, thus achieving time-series tracking of feature parameters. This model significantly extends the effective timeframe for pollution source tracing, expands the spatial tracing distance, and overcomes the problem of poor timeliness in source tracing caused by feature drift in traditional methods. The confidence intervals output by the model provide a quantitative basis for assessing the uncertainty of the source tracing results.

[0023] Step S3 employs a Bayesian inversion method to achieve reverse identification of pollution sources, addressing the issue of insufficient accuracy in tracing microplastic pollution sources. This step combines current detection features, migration paths, and feature evolution patterns, performing systematic calculations in the parameter space to effectively correct the impact of feature evolution on tracing accuracy. Compared to traditional direct feature comparison methods, this step significantly improves tracing accuracy. The output pollution source probability distribution clarifies the spatial uncertainty range of the tracing results, providing a technical basis for the precise allocation of governance resources and the scientific identification of responsible parties.

[0024] The dynamic correlation mechanism of physicochemical parameters established in step S4 overcomes the limitations of the fixed parameter assumptions in existing migration models. This step correlates changes in characteristic parameters with changes in physicochemical parameters such as density and particle size through quantitative relationships, enabling migration calculations to reflect the actual parameter evolution of microplastics during transport. Compared to models with fixed parameter assumptions, this mechanism significantly reduces the prediction errors of key parameters such as settling rate and buoyancy rate, improves the accuracy of migration path prediction, and provides more reliable technical support for assessing pollution diffusion range and deploying interception facilities.

[0025] The multi-process coupled migration model constructed in step S5 achieves coordinated computation of hydrodynamics, chemical degradation, and biological processes, resolving the prediction bias caused by independent computation of each process in existing models. Through a bidirectional feedback mechanism, this model can simulate complex interactive behaviors such as biofilm growth altering particle density and thus affecting sedimentation, hydrodynamic shearing leading to biofilm peeling, and chemical aging causing mechanical breakage. Compared to the traditional linear superposition method, this model significantly reduces the spatial error in migration path prediction, providing a technical means for the accurate simulation of microplastic transport processes.

[0026] The adaptive correction mechanism for model parameters implemented in step S6 addresses the issue of declining accuracy in migration prediction models over time. This step employs an ensemble Kalman filter to integrate real-time monitoring data into the coupled model, enabling the model to continuously adjust state variables and key parameters based on observational data. As monitoring data accumulates, prediction accuracy continuously improves. This mechanism provides early warnings of microplastic flux changes under extreme weather conditions, buying time for emergency response and enhancing the proactiveness of governance and management.

[0027] Step S7 reveals the morphological transformation patterns of microplastics within the remediation facility through analysis of the transformation behavior during the remediation process, addressing the issue of incomplete characterization of remediation effectiveness in existing assessment methods. The quantitative relationship between treatment intensity and breakage rate established in this step can predict the impact of different remediation processes on microplastic particle size distribution, accurately distinguishing between actual removal and phase transformation, and avoiding the misjudgment of effective removal as breaking large particles into smaller ones. This step can identify remediation schemes that appear to meet standards but whose actual ecological risks have not decreased or have even increased, providing a technical basis for the rational design of remediation facilities and the scientific control of operating parameters.

[0028] The multi-dimensional assessment system established in step S8 achieves comprehensive quantification of the remediation effect, resolving the difficulty in distinguishing between remediation contributions and natural degradation in existing assessment methods. This step, through material balance calculations and scenario comparisons, quantitatively differentiates the contributions of pollution reduction, pollution transfer, and form transformation, avoiding misjudging natural sedimentation or secondary fragmentation as remediation effectiveness. Combined with a comprehensive assessment of mass conservation, particle size distribution changes, and ecological risk changes, it comprehensively reflects the actual impact of remediation measures on microplastic pollution, providing a scientific basis for analyzing the cost-effectiveness of remediation inputs and determining the direction of process improvement.

[0029] The decision support system constructed in step S9 enables data-driven optimization and dynamic adjustment of governance strategies, addressing the lack of systematic and forward-looking approaches in governance decision-making. This system integrates pollution source identification, migration prediction, and effectiveness evaluation results, employing a multi-objective genetic algorithm to find the optimal combination of governance strategies under multiple constraints, thus improving the efficiency of governance input conversion. The early warning and response mechanism can provide warnings and recommend emergency measures before sudden increases in microplastic flux or exceedances in sensitive areas, reducing the ecological risks caused by extreme events. The continuous monitoring and dynamic adjustment mechanism forms a closed-loop management system of monitoring, evaluation, optimization, and feedback, enabling governance strategies to adaptively adjust according to changes in environmental conditions and pollution characteristics, thereby enhancing the scientific nature of governance management.

[0030] This big data-coupled method for detecting the effectiveness of microplastic pollution control overcomes the problem of source tracing failure caused by feature drift in traditional methods by combining multi-dimensional feature acquisition with a feature evolution model in pollution source tracing, thus extending the effective tracing time and improving the accuracy of source tracing. In migration prediction, the multi-process coupling model and dynamic correlation mechanism of physicochemical parameters established by this method can accurately simulate the complex interactive behavior in the microplastic transport process, reducing the prediction error of migration paths. In governance assessment, the method's analysis of the transformation behavior of the governance process and the multi-dimensional evaluation system can accurately distinguish between actual removal, pollution transfer, and form transformation, avoiding misjudging governance solutions that appear to meet standards but do not actually reduce risks as effective solutions. In decision support, by integrating data from pollution source tracing, migration prediction, and effectiveness evaluation, this method achieves multi-objective optimization and dynamic adjustment of governance strategies, establishing a closed-loop management mechanism of monitoring, evaluation, optimization, and feedback.

[0031] Example 2: Please refer to Figure 1 In step S1, the specific steps for multidimensional feature acquisition are as follows: S1.1 Sampling points were set up in the monitoring area, and the inorganic element characteristics, organic polymer characteristics, and bio-attachment characteristics of microplastic samples were detected. The inorganic element characteristics were obtained by single-particle inductively coupled plasma mass spectrometry, the organic polymer characteristics were obtained by Fourier transform infrared spectroscopy and Raman spectroscopy, and the bio-attachment characteristics were obtained by high-throughput sequencing technology. Environmental parameters such as temperature, light and hydrodynamic conditions were recorded to form initial characteristic data. S1.2. The control sample is subjected to simulated light, hydrodynamics and microbial activity under experimental conditions. Data on the changes of characteristic parameters over time are obtained through accelerated aging experiments. The correspondence between the initial characteristic parameters, environmental simulation conditions and evolutionary characteristic parameters is organized into a correlation database to generate a characteristic parameter evolution benchmark.

[0032] In this embodiment: In step S1.1, sampling points are set up in the monitoring area, and inorganic element characteristics are detected using single-particle inductively coupled plasma mass spectrometry, organic polymer characteristics are detected using Fourier transform infrared spectroscopy and Raman spectroscopy, and bio-attachment characteristics are detected using high-throughput sequencing technology, thus characterizing the microplastic sample from three different dimensions. Existing technologies often employ a single detection method, and the reliability of source tracing analysis is affected when the target characteristics change due to environmental factors. The three types of characteristic parameters obtained in this step can mutually verify each other. When one type of characteristic changes significantly, other types of characteristics can still be used to assist in the judgment, improving the stability of characteristic identification. Simultaneously recorded environmental parameters such as temperature, light intensity, and hydrodynamic conditions provide reference data for analyzing the relationship between characteristic changes and environmental factors, providing a reference basis for establishing the characteristic evolution law under environmental conditions.

[0033] In step S1.2, accelerated aging experiments are used to simulate the characteristic changes of microplastics under different environmental conditions, obtaining data on the changes of characteristic parameters over time. Existing technologies rely on long-term field monitoring to accumulate characteristic evolution data, which is time-consuming and has limited data coverage. This step simulates the effects of light, hydrodynamics, and microbial activity under controlled experimental conditions, enabling the acquisition of characteristic evolution information over a longer timescale in a shorter time, thus improving data acquisition efficiency. The correspondence between initial characteristic parameters, environmental simulation conditions, and evolutionary characteristic parameters is compiled into a relational database, providing data support for constructing a characteristic evolution prediction model. This database shifts characteristic evolution analysis from qualitative description to quantitative calculation, providing a basis for subsequent pollution source tracing through inference.

[0034] In step S1, sampling points were set up in the monitoring area to obtain three types of characteristic parameters of microplastics: inorganic elements, organic polymers, and bioattachment. Combined with accelerated aging experiments, a correlation database of initial characteristics, environmental conditions, and evolutionary characteristics was established, realizing the systematic construction of a microplastic characteristic characterization system and evolutionary laws. Existing technologies often use single indicators for characteristic characterization; when these indicators change due to environmental aging, the accuracy of source tracing analysis decreases. The three types of characteristic parameters collected in this step constitute a complementary characteristic system. When one type of characteristic changes significantly, other characteristics can still be used to assist in judgment, improving the stability of characteristic identification. Regarding the evolutionary laws of characteristics, existing technologies lack systematic evolutionary data accumulation, making it difficult to predict the changing trends of microplastic characteristics over time, resulting in a short effective time for source tracing analysis. This step obtains characteristic evolution data under different environmental conditions through accelerated aging experiments, establishing a quantitative correlation between initial characteristics, environmental conditions, and evolutionary characteristics, providing data support for the subsequent construction of a characteristic evolution prediction model. This associated database enables source tracing analysis to infer the initial characteristics based on the currently detected feature parameters, and to identify pollution sources while taking into account the changes in characteristics caused by environmental factors. This improves upon the traditional method's failure to trace sources due to ignoring feature evolution, extends the effective time window of source tracing analysis, and improves the accuracy of pollution source tracing.

[0035] Example 3: Please refer to Figure 1 In step S2, the specific method for modeling the evolutionary trajectory is as follows: S2.1 Based on the associated database of step S1, temperature, light intensity, hydrodynamic conditions and microbial activity are selected as environmental factors. The rate of change of characteristic parameters is statistically analyzed. A Bayesian network is used to establish a dependency model between environmental factors and the rate of change of characteristic parameters. The change law of characteristic parameters of different polymer types under the action of environmental factors is modeled to form the evolution relationship of characteristic parameters. S2.2. The evolutionary relationship in step S2.1 is trained using a long short-term memory neural network to obtain the temporal evolution trajectory of the feature parameters under different environmental conditions. Based on transfer learning, the accelerated aging data in the laboratory is corrected to construct a feature evolution prediction model with the initial feature parameters and environmental history as input, and output the predicted value of the feature parameters and the confidence interval at the target time.

[0036] In this embodiment: In step S2.1, temperature, light intensity, hydrodynamic conditions, and microbial activity are selected as environmental factors to statistically analyze the rate of change of characteristic parameters. A Bayesian network is used to establish a dependency model between environmental factors and the rate of change of characteristic parameters. Existing technologies often use empirical descriptions when analyzing changes in microplastic characteristics, lacking quantitative relationships between environmental factors and changes in characteristic parameters. The dependency model established in this step quantifies the influence of different environmental factors on the rate of change of characteristic parameters and clarifies the weight of each environmental factor in the characteristic evolution process. Modeling the characteristic parameter change patterns of different polymer types under the influence of environmental factors is performed separately, considering the influence of microplastic material differences on evolutionary behavior, thus giving the established evolutionary relationship a wider range of applicability. The characteristic parameter evolution relationship formed in this step provides a basic framework for the subsequent construction of time series prediction models.

[0037] In step S2.2, the evolutionary relationship established in step S2.1 is trained using a long short-term memory neural network to obtain the temporal evolution trajectory of feature parameters under different environmental condition sequences. Existing technologies for predicting changes in microplastic characteristics often rely on linear assumptions or simplified empirical formulas, making it difficult to accurately describe the feature evolution process under fluctuating environmental conditions. The long short-term memory neural network used in this step can learn the cumulative impact of environmental condition sequences on feature parameters, capturing the nonlinear evolutionary patterns in the environmental process. Transfer learning is used to correct the accelerated aging data from the laboratory, mitigating the difference between laboratory conditions and real environmental conditions. The constructed feature evolution prediction model takes the initial feature parameters and environmental history as input and outputs the predicted value and confidence interval of the feature parameters at the target time, providing a quantitative tool for feature back-calculation in pollution source tracing, enabling source tracing analysis to infer the initial characteristics of microplastics based on current detection characteristics and environmental history.

[0038] In step S2, a quantitative model capable of predicting the evolution of microplastic characteristics over time is constructed by using a Bayesian network to quantify the impact of environmental factors on the rate of change of feature parameters and combining this with a long short-term memory neural network to learn the temporal relationship between environmental condition sequences and feature parameters. Existing technologies often employ qualitative descriptions or simplified linear assumptions when analyzing feature evolution, making it difficult to accurately describe the feature change process under fluctuating environmental conditions. The prediction model established in this step can calculate the temporal trajectory of feature parameters based on initial features and environmental history, improving the accuracy of feature evolution prediction. In pollution source tracing, existing technologies typically compare currently detected features directly with features of potential pollution sources. However, after microplastics have undergone prolonged environmental exposure, their characteristics have changed, significantly reducing the accuracy of direct comparison-based source tracing and limiting the effective timeframe to weeks to months. The feature evolution prediction model constructed in this step can infer initial features from current detected features and the environmental history along the migration path, improving source tracing bias caused by feature drift and extending the effective time window for source tracing analysis. The confidence intervals output by the model quantify the uncertainty of the source tracing results, enabling source tracing analysis to shift from point estimation to interval estimation, thereby improving the credibility of source tracing judgments.

[0039] Example 4: Please refer to Figure 1 In step S3, the specific method for identifying pollution sources is as follows: S3.1 Based on the microplastic sample data obtained from monitoring, the current characteristic parameters are integrated with the migration path calculated by the hydrological model to form inversion input data. This inversion input data is then input into the feature evolution model in step S2 to generate feature evolution results corresponding to different pollution source scenarios. S3.2. The Bayesian inversion method is used to perform parameter space calculation on the feature evolution results generated in step S3.1. The parameter space includes the initial features of the pollution source, migration path, environmental evolution process and current detection features. The likelihood value of each pollution source scenario is calculated by combining the Markov chain Monte Carlo sampling method to determine the location and emission time of the pollution source and generate probability distribution and confidence level.

[0040] In this embodiment: In step S3.1, the current characteristic parameters of the microplastic sample obtained from monitoring are integrated with the migration path calculated by the hydrological model to form the input data required for inversion, and then input into the feature evolution model established in step S2 to generate feature evolution results under different pollution source scenarios. Existing technologies typically use direct feature comparison in pollution source tracing, matching the current characteristics of the monitored sample with the known characteristics of potential pollution sources based on similarity. This method does not consider the characteristic changes of microplastics during migration. When a sample undergoes long-term or long-distance transport, its characteristics will change, reducing the accuracy of direct comparison. This step combines migration path information with characteristic parameters, using the feature evolution model to deduce the characteristic state that microplastics should exhibit after migrating through the migration path under different pollution source scenarios, establishing a correlation between pollution source hypotheses and observed characteristics. The generated feature evolution results provide a computational basis for subsequent Bayesian inversion, enabling source tracing analysis to identify pollution sources while considering feature evolution.

[0041] In step S3.2, the Bayesian inversion method is used to perform parameter space calculations on the feature evolution results generated in sub-step S3.1. The parameter space encompasses the initial characteristics of the pollution source, migration path, environmental evolution process, and current detection characteristics. The likelihood values ​​for each pollution source scenario are calculated using the Markov chain Monte Carlo sampling method. Existing technologies often employ deterministic discrimination methods in pollution source identification, providing a single source tracing result. However, due to errors in feature measurement, uncertainties in environmental conditions, and biases in migration path calculations, a single result cannot reflect the inherent uncertainties of source tracing analysis. This step uses the Bayesian inversion method to evaluate the degree of agreement between different pollution source hypotheses and observational data in the parameter space. Markov chain Monte Carlo sampling explores the parameter space, calculates the likelihood values ​​for each pollution source scenario, and generates probability distributions and confidence levels while determining the pollution source location and emission time. The output probability distribution clarifies the uncertainty range of the source tracing results, and the confidence level quantifies the credibility of the results, providing a basis for resource allocation and responsibility determination in governance decisions.

[0042] In step S3, the current detection features are integrated with migration path information and input into the feature evolution model to generate feature evolution results under different pollution source scenarios. Then, a Bayesian inversion method is used to calculate the likelihood value of each pollution source scenario in the parameter space, achieving pollution source identification under spatiotemporal feature changes. Existing technologies for tracing microplastic pollution sources often employ direct feature comparison or fingerprint matching, assuming that microplastic features remain unchanged during migration. However, after prolonged environmental exposure, the inorganic elements, organic polymers, and biofouling characteristics of microplastics change, leading to a decrease in the accuracy of traditional tracing methods and limiting their effective time to weeks to months. This step uses a feature evolution model to deduce the feature change trajectory under different pollution source scenarios, transforming source identification from static feature matching to dynamic evolution matching. This corrects for the impact of feature changes on source analysis and extends the effective time for source identification. Regarding the expression of source identification results, existing technologies often provide a single pollution source location judgment without quantifying the uncertainty of the results. This step uses the Bayesian inversion method to output the probability distribution and confidence level of the pollution source, shifting the source tracing analysis from point estimation to interval estimation. This clarifies the credibility and uncertainty range of the source tracing results, and improves the accuracy of source control and responsibility determination.

[0043] Example 5: Please refer to Figure 1 In step S4, the specific method for dynamically associating parameters is as follows: S4.1 Based on the characteristic parameter data obtained in step S1, a quantitative relationship between biofilm coverage and particle equivalent density is established through density gradient centrifugation experiment; a relationship between particle size distribution changes under aging and crushing is established through laser particle size analysis; a quantitative relationship between surface oxidation degree and hydrophobicity changes is established through contact angle measurement; and the correspondence between density, particle size, surface roughness, hydrophobicity and characteristic parameters is organized into a physicochemical parameter correlation model. S4.2. Embed the physicochemical parameter association model from step S4.1 into the migration calculation model. In each time step of the migration calculation, update the density, particle size, surface roughness, and hydrophobicity based on the feature parameter prediction results from step S2, so that the physicochemical parameters used in the migration calculation are adjusted synchronously with the changes in the feature parameters.

[0044] In this embodiment: In step S4.1, based on the characteristic parameter data obtained in step S1, a quantitative relationship between biofilm coverage and particle equivalent density is established through density gradient centrifugation experiments; a relationship between particle size distribution changes under aging and fragmentation is established through laser particle size analysis; and a quantitative relationship between surface oxidation degree and hydrophobicity changes is established through contact angle measurement. These relationships are then organized into a physicochemical parameter correlation model. Existing technologies typically use fixed physicochemical parameters in microplastic migration simulations, assuming that parameters such as microplastic density and particle size remain constant during migration. In reality, microplastics in an aquatic environment undergo biofilm adhesion, photochemical aging, and mechanical wear, causing their physicochemical parameters to change over time. This step establishes a quantitative relationship between characteristic parameters and physicochemical parameters through experimental measurement, enabling changes in physicochemical parameters to be reflected through changes in characteristic parameters, providing a conversion basis for dynamic parameter updates in subsequent migration calculations.

[0045] In step S4.2, the physicochemical parameter correlation model established in step S4.1 is embedded into the migration calculation model. At each time step of the migration calculation, density, particle size, surface roughness, and hydrophobicity are updated based on the feature parameter prediction results from step S2, ensuring that the physicochemical parameters used in the migration calculation are adjusted synchronously with changes in the feature parameters. Existing migration models assume fixed parameters, leading to discrepancies between the calculated settling rate, buoyancy rate, and other key parameters and actual conditions. When simulating long-term migration processes, accumulated errors can decrease the accuracy of migration path predictions. This step couples the parameter correlation model with the migration calculation model, enabling dynamic updates of the physicochemical parameters. When the feature evolution model predicts an increase in the biofilm coverage of microplastics, the correlation model adjusts the particle density accordingly, affecting the settling calculation. When the predicted surface oxidation level increases, the hydrophobicity parameter is adjusted accordingly, affecting the simulation of adsorption and aggregation behavior. This allows the migration calculation to reflect the impact of parameter evolution on migration behavior during the transport process of microplastics.

[0046] In step S4, a quantitative correlation model between characteristic parameters and physicochemical parameters is established through density gradient centrifugation experiments, laser particle size analysis, and contact angle measurements. This model is embedded into the migration calculation process, and parameters such as density, particle size, surface roughness, and hydrophobicity are updated synchronously based on the characteristic evolution prediction results at each time step, realizing the dynamic adjustment of physicochemical parameters in the migration simulation. Existing microplastic migration models often employ fixed parameter assumptions, setting the density, particle size, surface properties, and other physicochemical parameters of microplastics as constants that remain unchanged throughout the migration calculation process, neglecting the influence of processes such as biofilm attachment, photochemical aging, and mechanical fragmentation of microplastics in the aquatic environment on these physicochemical parameters. Actual observations show that the density of microplastics can increase due to biofilm attachment, the particle size can decrease due to mechanical action, and the surface properties can change due to aging. These parameter changes affect the migration behavior of microplastics, including sedimentation, buoyancy, and adsorption. The correlation model established in this step enables migration calculations to reflect the parameter evolution of microplastics during transportation. When the feature evolution model predicts an increase in biofilm coverage, the density parameter is adjusted accordingly and affects the sedimentation calculation. When the predicted surface oxidation level increases, the hydrophobicity parameter is adjusted accordingly and affects the adsorption behavior simulation. This improves the accuracy of calculations for key parameters such as sedimentation rate and buoyancy rate, reduces the error in migration path prediction, and enhances the scientific rigor of pollution diffusion range assessment and interception facility deployment location selection.

[0047] Example 6: Please refer to Figure 1 In step S5, the specific steps of multi-process coupling are as follows: S5.1. A multi-process coupled migration model is constructed based on hydrodynamic transport, chemical degradation and biological processes. The operator splitting method is used to decompose each process into independent sub-modules and solve them separately. The hydrodynamic module solves the convection-diffusion equations using the finite volume method. The chemical degradation module is calculated based on surface oxidation kinetics. The biological process module is calculated based on biofilm growth kinetics and particle aggregation kinetics. S5.2 Establish a two-way feedback mechanism between sub-modules. The change in particle density caused by biofilm growth is fed back to correct the settling rate of the hydrodynamic module. The hydrodynamic shear stress is fed back to the bio-module as the driving force for biofilm peeling. The structural weakening caused by chemical aging is fed back to the mechanical breakage module. Each module performs iterative calculations in each calculation time step until the parameters converge.

[0048] In this embodiment: In step S5.1, a multi-process coupled migration model is constructed based on hydrodynamic transport, chemical degradation, and biological processes. The operator splitting method is used to decompose each process into independent sub-modules and solve them separately. The hydrodynamic module uses the finite volume method to solve the convection-diffusion equations, the chemical degradation module is calculated based on surface oxidation kinetics, and the biological process module is calculated using biofilm growth kinetics and particle aggregation kinetics. Existing technologies for simulating microplastic migration typically focus on hydrodynamic transport, simplifying the treatment of chemical degradation and biological processes, or calculating each process independently and then linearly superimposing them. In reality, the migration behavior of microplastics in the aquatic environment is controlled by multiple processes, and these processes interact with each other. This step incorporates hydrodynamic transport, chemical degradation, and biological processes into a unified model framework, using the operator splitting method to solve each sub-module separately. This ensures the accuracy of the calculations for each process while providing a foundation for establishing the interactions between processes in the subsequent steps.

[0049] In step S5.2, a bidirectional feedback mechanism is established between sub-modules. Changes in particle density caused by biofilm growth provide feedback to correct the settling rate of the hydrodynamic module; hydrodynamic shear stress is fed back to the biofilm module as the driving force for biofilm peeling; and structural weakening caused by chemical aging is fed back to the mechanical fragmentation module. Each module iterates within each computational time step until the parameters converge. Existing technologies often employ unidirectional coupling or linear superposition when handling multiple processes, i.e., calculating the result of one process first and then using it as the input for the next, or simply adding the effects of each process, ignoring the interrelationships between processes. In reality, biofilm growth increases particle density and promotes settling, while hydrodynamic shearing leads to biofilm peeling, and chemical aging weakens particle structure, making it more prone to fragmentation. The bidirectional feedback mechanism established in this step can simulate the interactions between these processes, achieving a dynamic equilibrium state through iterative calculations, enabling the model to capture the nonlinear behavior during microplastic migration.

[0050] In step S5, the hydrodynamic transport, chemical degradation, and biological processes are decomposed into independent sub-modules and solved separately using an operator splitting method. A bidirectional feedback mechanism is established between the sub-modules, allowing density changes caused by biofilm growth to correct sedimentation calculations, hydrodynamic shear stress to affect biofilm peeling, and chemical aging results to influence fragmentation simulations. This achieves coupled calculation of the interactions between multiple processes. Existing technologies in microplastic migration simulations often focus on hydrodynamic transport processes, simplifying the consideration of chemical degradation and biological processes, or calculating each process independently and obtaining a comprehensive result through linear superposition, neglecting the interrelationships between processes. Actual observations show that the migration behavior of microplastics is influenced by multiple processes, and these processes interact with each other. Biofilm attachment alters particle density, affecting sedimentation and buoyancy behavior; hydrodynamic shear stress leads to biofilm peeling; and chemical aging weakens particle structure, increasing the likelihood of mechanical fragmentation. The coupled model constructed in this step captures these interactive behaviors through a two-way feedback mechanism. Compared with the traditional linear superposition method, it can more accurately simulate nonlinear processes such as biofilm growth promoting sedimentation but being limited by hydrodynamic shear and chemical aging intensifying fragmentation. This improves the accuracy of predicting migration paths and concentration distribution, and enhances the rationality of pollution diffusion range assessment and interception facility site selection.

[0051] Example 7: Please refer to Figure 1 In step S6, the specific steps of model adaptive correction are as follows: S6.1 Based on the real-time data obtained from the monitoring network deployed in step S1, the concentration of microplastics and the distribution of characteristic parameters are used as inputs. The ensemble Kalman filter method is used to integrate the real-time data into the coupled migration model established in step 5, correct the model state variables and key parameters, and use variational interpolation technology to expand the point data into a spatial field to generate the input data required for adaptive correction. S6.2. Based on the updated data generated in step S6.1, detect the deviation between the model prediction result and the assimilated data. If the deviation exceeds the preset threshold, trigger sensitivity analysis, identify the master control deviation parameter and correct it to ensure that the model state is updated in subsequent calculation time steps, thereby improving prediction accuracy and reliability.

[0052] In this embodiment: In step S6.1, real-time data is acquired based on the monitoring network deployed in step S1. Microplastic concentration and characteristic parameter distribution are used as inputs. An ensemble Kalman filter method is employed to integrate the real-time data into the coupled migration model established in step S5, correcting the model's state variables and key parameters. Variational interpolation technology is used to extend the point data into a spatial field. Existing technologies for microplastic migration prediction typically employ offline calibration, where parameters are determined using historical data during model establishment and maintained unchanged in subsequent predictions. When factors such as hydrological conditions and pollution source emission intensity change, the model's prediction accuracy decreases over time. This step uses a data assimilation method to integrate real-time monitoring data into the model calculation process, enabling the model to continuously correct state variables and key parameters based on observation data. The ensemble Kalman filter method seeks the optimal balance between model prediction and observation data through state estimation and error covariance propagation. Variational interpolation technology extends discrete point monitoring data into a continuous spatial distribution field, allowing model correction to consider spatial heterogeneity.

[0053] In step S6.2, based on the updated data generated in step S6.1, the deviation between the model's prediction results and the assimilated data is detected. If the deviation exceeds a preset threshold, sensitivity analysis is triggered to identify and correct the main control deviation parameters, ensuring that the model state is updated in subsequent calculation time steps. Existing technologies typically adjust all parameters uniformly during data assimilation, or only adjust state variables without affecting model parameters. When there is a systematic deviation between model predictions and observations, it is difficult to effectively improve model performance. This step establishes a deviation detection and parameter diagnosis mechanism. By detecting the deviation between the prediction results and observed data, sensitivity analysis is triggered when the deviation exceeds a preset threshold. Sensitivity analysis identifies parameters that contribute significantly to prediction deviation and corrects these parameters in a targeted manner. This step enables the model to continuously improve based on actual observations, and the prediction accuracy gradually improves with the accumulation of monitoring data.

[0054] In step S6, the real-time monitored microplastic concentration and characteristic parameters are integrated into the coupled migration model using an ensemble Kalman filter. Variational interpolation is used to extend the point data into a spatial field, and a deviation detection mechanism is established. When the deviation between the predicted result and the observed data exceeds a threshold, sensitivity analysis is triggered, and the main control deviation parameters are identified and corrected accordingly. This achieves adaptive adjustment of the model's state variables and key parameters. Existing technologies for microplastic migration prediction often use offline calibrated static models, where parameters remain unchanged after being determined during the model establishment phase. This approach may perform well in short-term predictions, but as the prediction time increases, actual changes in environmental conditions and pollution source emissions lead to a gradual increase in the deviation between the model prediction and the actual situation, resulting in a decrease in prediction reliability. The adaptive correction mechanism established in this step enables the model to continuously correct its state and parameters based on real-time observation data, adapting to fluctuations in hydrological conditions and changes in pollution source emissions. With the accumulation of monitoring data, the model's prediction accuracy gradually improves, enhancing the reliability of medium- and long-term predictions. In the event of extreme weather conditions or sudden pollution events, this mechanism can quickly identify and correct model prediction deviations, providing early warnings for sudden increases in microplastic flux and shortening the decision-making time for emergency response.

[0055] Example 8: Please refer to Figure 1 Step S7, the specific steps of the governance process analysis are as follows: S7.1 Through on-site pilot testing and laboratory simulation, monitor the changes in particle size, morphology, mass concentration and number concentration of microplastics before and after entering treatment facilities such as interception nets and sedimentation tanks. Use high-speed photography and image analysis technology to observe the crushing process and establish a quantitative relationship between treatment intensity, crushing rate and progeny particle size distribution to provide data support for the analysis of microplastic conversion behavior. S7.2 Monitor the removal efficiency of microplastics of different particle sizes in each treatment unit, distinguish between actual removal and phase transformation, and analyze the transformation behavior of microplastics by comparing the removal effect and transformation process of each unit, so as to provide a basis for the evaluation of the treatment effect.

[0056] In this embodiment, in step S7.1, changes in particle size, morphology, mass concentration, and number concentration of microplastics before and after entering treatment facilities such as interception nets and sedimentation tanks are monitored through on-site pilot testing and laboratory simulation. High-speed photography and image analysis techniques are used to observe the crushing process and establish a quantitative relationship between treatment intensity, crushing rate, and progeny particle size distribution. Existing technologies typically only monitor changes in the mass concentration or number concentration of microplastics when evaluating the effectiveness of treatment facilities, directly considering a decrease in concentration as a removal effect. In reality, microplastics may undergo mechanical crushing in treatment facilities; large particles break down into smaller particles, resulting in a decrease in mass concentration but an increase in number concentration, while small-diameter microplastics may pose a higher ecological risk. This step, by simultaneously monitoring changes in particle size, morphology, mass concentration, and number concentration, can identify the crushing behavior of microplastics. The quantitative relationship established by directly observing the crushing process using high-speed photography and image analysis techniques, and by establishing a relationship between treatment intensity, crushing rate, and progeny particle size distribution, can predict the impact of different treatment processes on the microplastic particle size distribution, providing a basis for judging the actual effectiveness of treatment measures.

[0057] In step S7.2, the removal efficiency of microplastics of different particle sizes in each treatment unit is monitored to distinguish between actual removal and phase transformation. By comparing the removal effect and transformation process of each unit, the transformation behavior of microplastics is analyzed. Existing technologies often use the overall removal rate as an evaluation indicator when assessing the performance of remediation facilities, without distinguishing between the removal of microplastics of different particle sizes or clearly differentiating between actual removal and phase transformation. In reality, different treatment units have varying removal effects on microplastics of different particle sizes. Mechanical bar screens may trap large particles but have limited effect on small particles, while some treatment processes may break large particles into smaller ones, seemingly achieving removal but actually resulting in particle size transformation. This step, by monitoring the removal efficiency of microplastics of different particle sizes in each treatment unit, clarifies the applicable particle size range for each unit. By distinguishing between actual removal and phase transformation, remediation schemes that appear to meet standards but do not actually reduce ecological risks can be identified, providing a basis for the design and control of remediation facility parameters.

[0058] In step S7, monitoring sections are set at the inlet and outlet of treatment facilities such as interception nets and sedimentation tanks, as well as at sludge and sediment, to simultaneously measure changes in the particle size, morphology, mass concentration, and number concentration of microplastics. High-speed photography and image analysis techniques are used to directly observe the crushing process, establishing a quantitative relationship between treatment intensity, crushing rate, and progeny particle size distribution. The removal efficiency of microplastics of different sizes in each treatment unit is monitored, and the difference between actual removal and phase transformation is distinguished, achieving a quantitative analysis of microplastic transformation behavior during the treatment process. Existing technologies, when evaluating the removal effect of water treatment facilities on microplastics, often use the method of comparing inlet and outlet concentrations to calculate the removal rate, considering a decrease in concentration as effective removal. This ignores the possible changes in morphology and particle size of microplastics during treatment, especially the particle size transformation caused by mechanical crushing. Actual observations show that some treatment processes, such as mechanical stirring and high-speed pumping, can mechanically act on microplastics, causing large particles to break into smaller particles. Although the mass concentration decreases, the number concentration increases. Smaller microplastics are more easily ingested by organisms, potentially increasing ecological risks. The quantitative relationship established in this step can predict the impact of different treatment processes on the microplastic particle size distribution. By distinguishing between actual removal and phase transformation, it identifies treatment schemes that meet the apparent removal rate but whose actual ecological risks have not been reduced or have even increased. This avoids misjudging particle size transformation as effective removal and shifts the evaluation of treatment effectiveness from extensive concentration comparison to refined quality conservation and risk assessment.

[0059] Example 9: Please refer to Figure 1 In step S8, the specific steps for multidimensional evaluation of the effect are as follows: S8.1 Establish a multi-dimensional assessment system that includes mass conservation, particle size distribution change and ecological risk change. Set up monitoring sections at the inlet, outlet, sludge and sediment of the treatment facility to collect microplastic data. Calculate the input, output, removal, transformation and cumulative mass through material balance. Compare the distribution changes of each particle size range before and after treatment. Calculate the comprehensive risk index by combining the relationship between particle size and toxicity. S8.2 Extend the coupling model from step S5 to the internal environment of the treatment facility, simulate the natural decay scenario without treatment measures and the actual treatment scenario, compare the microplastic removal effect under the two scenarios, quantify the net contribution of treatment measures, and provide a basis for optimizing the treatment scheme.

[0060] In this embodiment: Step S8.1 establishes a multi-dimensional assessment system including mass conservation, particle size distribution changes, and ecological risk changes. Monitoring sections are set up at the inlet, outlet, sludge, and sediment of the treatment facility to collect microplastic data. Input, output, removal, transformation, and accumulation masses are calculated through material balance. The distribution changes of each particle size range before and after treatment are compared, and a comprehensive risk index is calculated based on the relationship between particle size and toxicity. Existing technologies typically compare concentration changes before and after treatment or calculate a single removal rate when assessing treatment effectiveness, without fully considering the fate and morphological changes of microplastics during the treatment process. Microplastics may be removed, transferred to sludge or sediment, or undergo particle size transformation in the treatment facility; simple concentration comparisons are insufficient to reflect these processes. This step, by setting up monitoring sections at multiple interfaces and using a material balance method to track the distribution of microplastics in each unit, clarifies the mass of each part—input, output, removal, transformation, and accumulation—avoiding misjudging transfer as removal. Comparing the distribution changes of each particle size range before and after treatment allows for the identification of particle size transformation behavior. By combining the relationship between particle size and toxicity to calculate a comprehensive risk index, the assessment of governance effectiveness is expanded from a single concentration indicator to an ecological risk assessment.

[0061] In step S8.2, the coupling model from step S5 is extended to the treatment facility to simulate the natural decay scenario without treatment measures and the actual treatment scenario. The microplastic removal effect under the two scenarios is compared to quantify the net contribution of the treatment measures. Existing technologies typically attribute the concentration difference before and after treatment entirely to the treatment measures when evaluating treatment effectiveness, without considering the natural decay processes such as natural sedimentation and biodegradation that may occur during the residence of microplastics in the treatment facility. Even without artificial treatment measures, microplastics will be partially removed by natural processes after a certain residence time in the water. Misjudging this natural decay as the treatment effect will overestimate the actual effect of the treatment measures. This step establishes a scenario comparison model to simulate the natural decay scenario without treatment measures under the same residence time, and compares it with the actual treatment scenario. This quantifies the additional contribution of the treatment measures relative to the natural process, and quantitatively distinguishes between the contributions of natural decay and artificial treatment.

[0062] In step S8, microplastic data is collected by setting up monitoring sections at the inlet, outlet, sludge, and sediment of the treatment facility. The mass balance method is used to calculate the input, output, removal, transformation, and cumulative mass. The changes in particle size distribution before and after treatment are compared, and a comprehensive risk index is calculated based on the relationship between particle size and toxicity. The coupled model is extended to simulate natural decay scenarios without treatment measures and actual treatment scenarios within the treatment facility, and the removal effects under the two scenarios are compared. This achieves a multi-dimensional assessment based on mass conservation, particle size distribution changes, and ecological risk changes, as well as a quantitative distinction of the net contribution of treatment measures. Existing technologies for evaluating the effectiveness of microplastic treatment often use the method of comparing inlet and outlet concentrations to calculate the removal rate as an evaluation indicator, which has two shortcomings: first, it does not track the actual destination of microplastics in the treatment facility, which may misjudge microplastics transferred to sludge or sediment as removed; second, it does not distinguish between the contribution of natural decay and artificial treatment, which may overestimate the actual effect of treatment measures. A single concentration or removal rate indicator does not consider the impact of particle size distribution changes on ecological risk. This step tracks the actual destination of microplastics through the mass balance method, avoiding the misjudgment of pollution transfer as effective removal. By comparing changes in particle size distribution before and after remediation and calculating a comprehensive risk index, it is possible to identify situations where mass concentration decreases but quantity concentration increases, leading to a rise in ecological risk. Scenario-based simulations quantitatively distinguish the contributions of natural processes such as natural sedimentation and biodegradation from artificial remediation measures, avoiding the misjudgment of natural degradation as remediation effectiveness. This shifts the assessment of remediation effectiveness from a crude concentration comparison to mass conservation tracking and risk quantification, and from isolated effect evaluation to net contribution differentiation.

[0063] Example 10: Please refer to Figure 1 In step S9, the specific steps for strategy optimization decision-making are as follows: S9.1 Integrate the pollution source identification results obtained in step S3, the migration prediction results obtained in step S6, and the effect evaluation results obtained in step S8 to construct a decision support system. Simulate the long-term effects and costs of different governance schemes on the digital twin platform, and use a multi-objective genetic algorithm to optimize the combination of governance strategies under budget and ecological protection constraints to generate a recommended list of optimal schemes. S9.2 Based on the decision support system built in step S9.1, monitor the output of the prediction model. When the microplastic flux suddenly increases or the sensitive area exceeds the standard, trigger an early warning and recommend emergency measures. Establish a continuous monitoring mechanism to update the evaluation results regularly. When the actual effect deviates from the expectation by more than the threshold, initiate the cause diagnosis and strategy adjustment process.

[0064] In this embodiment: In step S9.1, the pollution source identification results obtained in step S3, the migration prediction results obtained in step S6, and the effect evaluation results obtained in step S8 are integrated to construct a decision support system. The long-term effects and costs of different governance schemes are simulated on a digital twin platform. A multi-objective genetic algorithm is used to optimize the combination of governance strategies under budget and ecological protection constraints, generating a recommended list of optimal schemes. Existing technologies often rely on empirical judgment or single-objective optimization when formulating microplastic governance schemes, lacking a systematic integration of pollution source distribution, migration patterns, and governance effects. Microplastic governance involves multiple stages such as source control, migration interception, and end-of-pipe treatment. The input-output ratios of different governance measures vary, requiring rational resource allocation under limited budgets. This step integrates the results of pollution source identification, migration prediction, and effect evaluation, simulating the long-term effects of different governance schemes on a digital twin platform, providing quantitative scenario predictions for decision-making. The use of a multi-objective genetic algorithm to find a balance between budget constraints and ecological protection goals can screen out the strategy combination with the best overall benefits from multiple alternative schemes.

[0065] In step S9.2, based on the decision support system built in step S9.1, the output of the prediction model is monitored. When there is a sudden increase in microplastic flux or when sensitive areas exceed the limit, an early warning is triggered and emergency measures are recommended. A continuous monitoring mechanism is established to regularly update the assessment results. When the actual effect deviates from the expectation by more than a threshold, the cause diagnosis and strategy adjustment process is initiated. Existing technologies in microplastic governance management mostly adopt a post-event assessment approach, taking countermeasures only after a pollution incident occurs, lacking a forward-looking early warning mechanism and dynamic adjustment capability. Microplastic flux is affected by factors such as rainfall and emissions, exhibiting temporal fluctuations. Under extreme weather conditions or during sudden emission events, microplastic concentrations may increase in a short period. The early warning and response mechanism established in this step can provide early warnings before a sudden increase in microplastic flux or when sensitive areas exceed the limit, buying time for emergency response. The continuous monitoring and regular update mechanism enables the governance effect assessment to track changes in the actual situation. When the actual effect deviates from the expectation, cause diagnosis and strategy adjustment are initiated, forming a closed-loop management mechanism of monitoring, assessment, optimization, and feedback.

[0066] In step S9, by inputting the pollution source identification results, migration prediction results, and effect evaluation results into the decision support system, the long-term effects and costs of different governance schemes are simulated on the digital twin platform. A multi-objective genetic algorithm is used to optimize the combination of governance strategies under budget and ecological protection constraints, generating a recommended list. The output of the prediction model is monitored, and early warnings and emergency measures recommendations are triggered when microplastic flux surges or exceeds standards in sensitive areas. A continuous monitoring mechanism is established to regularly update the evaluation results and initiate cause diagnosis and strategy adjustment when the actual effect deviates from the expected result by more than a threshold. This achieves data-driven governance decision optimization and dynamic management. Existing technologies for microplastic governance decisions largely rely on empirical judgment, lacking quantitative analysis of pollution source distribution, migration patterns, and governance effects. Governance schemes are often formulated based on local information or a single objective, making it difficult to determine the optimal resource allocation scheme under limited budgets. In terms of governance management, existing technologies mostly adopt static governance schemes and post-event evaluation methods, lacking forward-looking early warning mechanisms and dynamic adjustment capabilities, resulting in delayed responses to extreme weather conditions or sudden emission events. This step integrates end-to-end data on pollution source distribution, migration paths, and treatment effects. It simulates the implementation effects of different solutions on a digital twin platform and employs a multi-objective genetic algorithm to find the optimal strategy combination with the best overall benefits under multiple constraints. This shifts governance decision-making from experience-based judgment to quantitative analysis, and from single-objective optimization to multi-objective balancing, improving resource allocation efficiency under limited budgets. The established early warning and response mechanism provides warnings and recommends emergency measures before sudden increases in microplastic flux or exceedances in sensitive areas, shortening emergency response time. The continuous monitoring and dynamic adjustment mechanism updates evaluation results regularly, diagnoses the causes and adjusts strategies when actual effects deviate from expectations, forming a closed-loop management process of monitoring, evaluation, optimization, and feedback. This improves the adaptability of governance solutions to changes in environmental conditions and fluctuations in pollution characteristics.

[0067] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

[0068] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting the effectiveness of microplastic pollution control based on big data coupling, characterized in that: The specific steps are as follows: S1. Multidimensional feature acquisition: Establish a multi-point sampling network in the target area to collect three types of characteristic parameters of microplastics: inorganic elements, organic polymers and bio-attachment. Accelerated aging experiments are conducted in the laboratory under different environmental conditions to establish a database linking initial characteristics, environmental conditions and evolutionary characteristics. S2. Evolutionary trajectory modeling: A feature evolution model is established using Bayesian networks and long short-term memory neural networks. The initial features and environmental history are used as inputs, and the predicted values ​​and confidence intervals of the current feature parameters are output. S3. Pollution Source Identification: Input the current characteristics and migration path of the monitored samples into the evolution model, use Bayesian inversion to identify the location and emission time of pollution sources, and output the probability distribution of pollution sources. S4. Parameter Dynamic Correlation: Establish quantitative relationships between physicochemical parameters such as microplastic density and particle size and characteristic parameters, and dynamically update physicochemical parameters based on changes in characteristics during migration calculations; S5. Multi-process coupling: Construct a coupled migration model that includes hydrodynamics, chemical degradation and biological processes, and realize the mutual influence of each process through a two-way feedback mechanism; S6. Adaptive Model Correction: The monitoring data is incorporated into the coupled model using ensemble Kalman filtering to achieve adaptive correction of the model parameters. S7. Analysis of the treatment process: Monitor the characteristic changes of microplastics in the treatment facility, establish a quantitative relationship between treatment intensity and breakage rate, and distinguish between actual removal and phase transformation; S8. Multidimensional evaluation of effectiveness: Set up monitoring sections at multiple interfaces of the treatment facilities, and quantify the net contribution of the treatment measures through material balance and scenario comparison. S9. Strategy Optimization Decision: Integrate the results of pollution source identification, migration prediction and effect evaluation to build a decision support system to achieve optimized adjustment of governance strategies and early warning response.

2. The method for detecting the effectiveness of microplastic pollution control based on big data coupling according to claim 1, characterized in that: In step S1, the specific steps for multidimensional feature acquisition are as follows: S1.1 Sampling points were set up in the monitoring area, and the inorganic element characteristics, organic polymer characteristics, and bio-attachment characteristics of microplastic samples were detected. The inorganic element characteristics were obtained by single-particle inductively coupled plasma mass spectrometry, the organic polymer characteristics were obtained by Fourier transform infrared spectroscopy and Raman spectroscopy, and the bio-attachment characteristics were obtained by high-throughput sequencing technology. Environmental parameters such as temperature, light and hydrodynamic conditions were recorded to form initial characteristic data. S1.

2. The control sample is subjected to simulated light, hydrodynamics and microbial activity under experimental conditions. Data on the changes of characteristic parameters over time are obtained through accelerated aging experiments. The correspondence between the initial characteristic parameters, environmental simulation conditions and evolutionary characteristic parameters is organized into a correlation database to generate a characteristic parameter evolution benchmark.

3. The method for detecting the effectiveness of microplastic pollution control based on big data coupling according to claim 2, characterized in that: In step S2, the specific method for modeling the evolutionary trajectory is as follows: S2.1 Based on the associated database of step S1, temperature, light intensity, hydrodynamic conditions and microbial activity are selected as environmental factors. The rate of change of characteristic parameters is statistically analyzed. A Bayesian network is used to establish a dependency model between environmental factors and the rate of change of characteristic parameters. The change law of characteristic parameters of different polymer types under the action of environmental factors is modeled to form the evolution relationship of characteristic parameters. S2.

2. The evolutionary relationship in step S2.1 is trained using a long short-term memory neural network to obtain the temporal evolution trajectory of the feature parameters under different environmental conditions. Based on transfer learning, the accelerated aging data in the laboratory is corrected to construct a feature evolution prediction model with the initial feature parameters and environmental history as input, and output the predicted value of the feature parameters and the confidence interval at the target time.

4. The method for detecting the effectiveness of microplastic pollution control based on big data coupling according to claim 3, characterized in that: In step S3, the specific method for identifying pollution sources is as follows: S3.1 Based on the microplastic sample data obtained from monitoring, the current characteristic parameters are integrated with the migration path calculated by the hydrological model to form inversion input data. This inversion input data is then input into the feature evolution model in step S2 to generate feature evolution results corresponding to different pollution source scenarios. S3.

2. The Bayesian inversion method is used to perform parameter space calculation on the feature evolution results generated in step S3.

1. The parameter space includes the initial features of the pollution source, migration path, environmental evolution process and current detection features. The likelihood value of each pollution source scenario is calculated by combining the Markov chain Monte Carlo sampling method to determine the location and emission time of the pollution source and generate probability distribution and confidence level.

5. The method for detecting the effectiveness of microplastic pollution control based on big data coupling according to claim 4, characterized in that: In step S4, the specific method for dynamically associating parameters is as follows: S4.1 Based on the characteristic parameter data obtained in step S1, a quantitative relationship between biofilm coverage and particle equivalent density is established through density gradient centrifugation experiment; a relationship between particle size distribution changes under aging and crushing is established through laser particle size analysis; a quantitative relationship between surface oxidation degree and hydrophobicity changes is established through contact angle measurement; and the correspondence between density, particle size, surface roughness, hydrophobicity and characteristic parameters is organized into a physicochemical parameter correlation model. S4.

2. Embed the physicochemical parameter association model from step S4.1 into the migration calculation model. In each time step of the migration calculation, update the density, particle size, surface roughness, and hydrophobicity based on the feature parameter prediction results from step S2, so that the physicochemical parameters used in the migration calculation are adjusted synchronously with the changes in the feature parameters.

6. The method for detecting the effectiveness of microplastic pollution control based on big data coupling according to claim 5, characterized in that: In step S5, the specific steps of multi-process coupling are as follows: S5.

1. A multi-process coupled migration model is constructed based on hydrodynamic transport, chemical degradation and biological processes. The operator splitting method is used to decompose each process into independent sub-modules and solve them separately. The hydrodynamic module solves the convection-diffusion equations using the finite volume method. The chemical degradation module is calculated based on surface oxidation kinetics. The biological process module is calculated based on biofilm growth kinetics and particle aggregation kinetics. S5.2 Establish a two-way feedback mechanism between sub-modules. The change in particle density caused by biofilm growth is fed back to correct the settling rate of the hydrodynamic module. The hydrodynamic shear stress is fed back to the bio-module as the driving force for biofilm peeling. The structural weakening caused by chemical aging is fed back to the mechanical breakage module. Each module performs iterative calculations in each calculation time step until the parameters converge.

7. The method for detecting the effectiveness of microplastic pollution control based on big data coupling according to claim 6, characterized in that: In step S6, the specific steps of model adaptive correction are as follows: S6.1 Based on the real-time data obtained from the monitoring network deployed in step S1, the concentration of microplastics and the distribution of characteristic parameters are used as inputs. The ensemble Kalman filter method is used to integrate the real-time data into the coupled migration model established in step 5, correct the model state variables and key parameters, and use variational interpolation technology to expand the point data into a spatial field to generate the input data required for adaptive correction. S6.

2. Based on the updated data generated in step S6.1, detect the deviation between the model prediction result and the assimilated data. If the deviation exceeds the preset threshold, trigger sensitivity analysis, identify the master control deviation parameter and correct it to ensure that the model state is updated in subsequent calculation time steps, thereby improving prediction accuracy and reliability.

8. The method for detecting the effectiveness of microplastic pollution control based on big data coupling according to claim 7, characterized in that: The specific steps of the governance process in step S7 are as follows: S7.1 Through on-site pilot testing and laboratory simulation, monitor the changes in particle size, morphology, mass concentration and number concentration of microplastics before and after entering treatment facilities such as interception nets and sedimentation tanks. Use high-speed photography and image analysis technology to observe the crushing process and establish a quantitative relationship between treatment intensity, crushing rate and progeny particle size distribution to provide data support for the analysis of microplastic conversion behavior. S7.2 Monitor the removal efficiency of microplastics of different particle sizes in each treatment unit, distinguish between actual removal and phase transformation, and analyze the transformation behavior of microplastics by comparing the removal effect and transformation process of each unit, so as to provide a basis for the evaluation of the treatment effect.

9. The method for detecting the effectiveness of microplastic pollution control based on big data coupling according to claim 8, characterized in that: In step S8, the specific steps for multidimensional evaluation of the effect are as follows: S8.1 Establish a multi-dimensional assessment system that includes mass conservation, particle size distribution change and ecological risk change. Set up monitoring sections at the inlet, outlet, sludge and sediment of the treatment facility to collect microplastic data. Calculate the input, output, removal, transformation and accumulation mass through material balance. Compare the distribution changes of each particle size range before and after treatment. Calculate the comprehensive risk index by combining the relationship between particle size and toxicity. S8.2 Extend the coupling model from step S5 to the internal environment of the treatment facility, simulate the natural decay scenario without treatment measures and the actual treatment scenario, compare the microplastic removal effect under the two scenarios, quantify the net contribution of treatment measures, and provide a basis for optimizing the treatment scheme.

10. The method for detecting the effectiveness of microplastic pollution control based on big data coupling according to claim 9, characterized in that: In step S9, the specific steps for strategy optimization decision-making are as follows: S9.1 Integrate the pollution source identification results obtained in step S3, the migration prediction results obtained in step S6, and the effect evaluation results obtained in step S8 to construct a decision support system. Simulate the long-term effects and costs of different governance schemes on the digital twin platform, and use a multi-objective genetic algorithm to optimize the combination of governance strategies under budget and ecological protection constraints to generate a recommended list of optimal schemes. S9.2 Based on the decision support system built in step S9.1, monitor the output of the prediction model. When the microplastic flux suddenly increases or the sensitive area exceeds the standard, trigger an early warning and recommend emergency measures. Establish a continuous monitoring mechanism to update the evaluation results regularly. When the actual effect deviates from the expectation by more than the threshold, initiate the cause diagnosis and strategy adjustment process.

Images