Water pollution intelligent traceability system based on three-dimensional fluorescence spectrum

Abstract

The invention particularly relates to a water pollution intelligent traceability system based on a three-dimensional fluorescence spectrum, which relates to the technical field of environmental monitoring, and comprises a water sample collection and three-dimensional fluorescence data acquisition module, a data preprocessing and PARAFAC component decomposition module, a multi-source pollution characteristic fingerprint database module and an intelligent comparison and pollution source analysis module. According to the method, standard acquisition and preprocessing are adopted, and the original data quality is ensured by means of scattering rejection, instrument correction and the like; pure fluorescence components are decomposed through a PARAFAC algorithm, and feature extraction reliability is guaranteed through triple verification. In the comparison link, spectrum similarity, concentration proportion difference and space-time hydrological constraint multi-round screening are adopted, and irrelevant pollution sources are removed; and then the contribution proportion is quantified, and Monte Carlo simulation, water quality and specific pollutant cross validation are carried out, so that the crossing from suspected tracing to quantitative responsibility investigation is realized.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring technology, and in particular to a smart water pollution source tracing system based on three-dimensional fluorescence spectroscopy. Background Technology

[0002] Current water pollution source tracing technologies face core technological bottlenecks: traditional methods rely heavily on conventional physicochemical indicators such as COD and ammonia nitrogen. These indicators have poor specificity and are difficult to distinguish between pollution sources in different industries (for example, dyeing and printing wastewater and chemical wastewater often show similar COD values). Although some technologies use three-dimensional fluorescence spectroscopy analysis, they directly compare the original full spectrum, which is easily affected by instrument noise, background fluorescence in water (such as natural humic matter fluorescence), and fluctuations in pollutant concentration. This often leads to misjudgments of different sources with the same spectrum or different spectra with the same source, resulting in insufficient reliability of source tracing results and an inability to provide accurate basis for environmental protection supervision.

[0003] Meanwhile, the existing traceability system also suffers from a lack of compatibility between the database and the implementation of results: On the one hand, pollution source feature databases are mostly statically stored, failing to consider the shift in fluorescent fingerprints caused by adjustments in enterprise production processes and upgrades in treatment facilities, and lack efficient classification indexes, resulting in low matching efficiency when faced with massive amounts of pollution source data. On the other hand, the source tracing results mostly remain at the level of qualitative judgment, failing to quantify the contribution ratio of each pollution source. Moreover, they are presented in the form of pure data tables, which are difficult for regulatory personnel to understand quickly and lack targeted enforcement and disposal suggestions. This leads to a disconnect between technical analysis and regulatory actions, making it difficult to meet the actual needs of watershed pollution investigation and emergency response to sudden pollution events.

[0004] Therefore, a smart water pollution source tracing system based on three-dimensional fluorescence spectroscopy is proposed to solve the problems mentioned above. Summary of the Invention

[0005] The purpose of this invention is to provide a smart water pollution source tracing system based on three-dimensional fluorescence spectroscopy to solve the above-mentioned problems.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A smart water pollution source tracing system based on three-dimensional fluorescence spectroscopy includes: a water sample collection and three-dimensional fluorescence data acquisition module, a data preprocessing and PARAFAC component decomposition module, a multi-source pollution feature fingerprint database module, and an intelligent comparison and pollution source analysis module, configured as follows: Obtain the fluorescence component spectrum of the water sample to be tested, including the excitation spectrum. emission spectrum and concentration ratio ; calculate , The cosine similarity and Tucker consistency coefficient of the spectra of each pollution source component in the database were used to screen out candidate pollution sources; The concentration ratio of the analyte was calculated using Euclidean distance. The difference between the concentration ratio of the candidate pollution sources and the concentration ratio of the candidate pollution sources is used to eliminate pollution sources with excessive differences from the candidate pollution sources; By combining the sampling time and location of the water sample to be tested with the emission time and geographical location of the candidate pollution sources, pollution sources that do not match in time and space are excluded; Using the component spectra of candidate pollution sources as a standard, the total fluorescence signal of the water sample to be tested is fitted, and the contribution ratio of each candidate pollution source is calculated. The rationality of the tracing results was further confirmed through multi-dimensional cross-validation.

[0007] Preferably, the water sample collection and three-dimensional fluorescence data acquisition module is configured to be responsible for the standardized collection, preprocessing, and acquisition of three-dimensional fluorescence spectral data of water samples, generating a three-dimensional data matrix containing the number of samples, the number of excitation wavelengths, and the number of emission wavelengths. .

[0008] Preferably, the data preprocessing and PARAFAC component decomposition module is configured to denoise and purify the raw fluorescence data, and to separate the complex mixed fluorescence signals, i.e., the three-dimensional data matrix. Decomposed into ; in, This is the component concentration score matrix; To excite the load matrix; This is the launch payload matrix; The residual matrix; It enables dimensionality reduction and extraction of essential features from full-spectrum data.

[0009] Preferably, the multi-source pollution feature fingerprint database module is configured to construct a structured pollution feature database with pure fluorescent components as the core, store information on various known pollution sources, and provide a reference standard for intelligent comparison of water samples to be tested.

[0010] Preferably, in the intelligent comparison and pollution source analysis module, the pretreatment of the water sample components to be tested includes: Excitation spectrum of component data obtained from the decomposition of the water sample to be tested emission spectrum and concentration ratio The preprocessing includes component effectiveness screening, spectral normalization optimization, and concentration ratio standardization to eliminate interference from invalid information.

[0011] Preferably, the screening process for the candidate pollution sources specifically includes: Calculate the analyte separately With database source components cosine similarity ; With database source components cosine similarity ; Set a comprehensive similarity threshold and calculate... , average ,like If the overall similarity threshold is greater than or equal to the threshold, then the condition is met and the source can be retained in the candidate pollution source pool. All candidate pollution sources that do not meet the requirements are removed to obtain the screened candidate pollution source pool. For the screened candidate pollution sources, obtain their weighted Euclidean distance. Calculate the concentration ratio of the analyte Ratio of candidate source concentration Differences; The weighted Euclidean distance is obtained by setting a preset difference threshold. Compare with the difference threshold, if the weighted Euclidean distance is used If the difference is less than or equal to the preset difference threshold, it is retained in the candidate pollution source pool; all candidate pollution sources that do not meet the requirements are removed from the candidate pollution source pool to obtain the candidate pollution source pool after secondary screening. By combining time, space, and hydrological three-dimensional information, pollution sources that cannot be reached in time and space are eliminated, resulting in a candidate pollution source pool after three rounds of screening.

[0012] Preferably, the step of fitting the total fluorescence signal of the water sample to be tested, using the component spectra of candidate pollution sources as a standard, and calculating the contribution ratio of each candidate pollution source specifically includes: Based on the candidate pollution source pool after three rounds of screening, the contribution of each source to the fluorescence signal of the water sample is decomposed and converted into a mass contribution ratio. The core steps are as follows: Component spectra of the core suspect , Integrate into a standard source spectral matrix Its dimension is the number of wavelength points × the number of sources; Define algorithm constraints, including non-negativity constraints, closure constraints, and iteration termination conditions; Initialize contribution ratio matrix Alternately update the source spectral matrix With contribution ratio matrix ; , This represents the fluorescence signal matrix of the water sample to be tested. ; Iterate until RSS converges to obtain the final contribution ratio matrix. Each element corresponds to the contribution ratio of a source; Through calibration experiments, the conversion coefficient k was obtained, and the final mass contribution ratio = fluorescence signal contribution × k.

[0013] Preferably, the process further includes uncertainty assessment and cross-validation: The Monte Carlo simulation method is used to evaluate the fluctuation range of the contribution ratio by inputting random perturbation parameters, and the uncertainty classification is obtained. By verifying water quality indicators, characteristic pollutants, and comparing historical data, the rationality of the source tracing results was further confirmed.

[0014] Preferably, the method further includes: The source tracing results visualization and intelligent decision support module is configured to transform complex source tracing data into intuitive and easy-to-understand visualization results, and provide functions such as pollution early warning, disposal suggestions, and report generation, so as to realize the implementation of source tracing results into regulatory actions.

[0015] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. This invention ensures the quality of raw data by adopting standardized acquisition and preprocessing, combined with scattering elimination and instrument calibration; pure fluorescent components are decomposed using the PARAFAC algorithm, and the reliability of feature extraction is guaranteed by triple verification; the comparison process uses multiple rounds of screening based on spectral similarity, concentration ratio difference, and spatiotemporal hydrological constraints to eliminate irrelevant pollution sources; then the contribution ratio is quantified, and Monte Carlo simulation is used for cross-verification with water quality and characteristic pollutants to achieve the leap from suspected source tracing to quantitative accountability.

[0016] 2. This invention transforms the source tracing results into intuitive outputs such as spectral comparison maps and GIS transmission path maps through a visualization module, and uses time series analysis to show pollution trends; it sets dynamic early warning thresholds and pushes early warning information in real time via SMS / APP; it automatically retrieves treatment plans for pollution source types, supports task assignment and result feedback from regulatory platforms and law enforcement departments; the database is classified by industry, region, and pollutant and is dynamically updated to adapt to emergency and routine monitoring scenarios, while generating a source tracing ledger to assist in the long-term management assessment of the watershed. Attached Figure Description

[0017] Further details, features, and advantages of this application are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which: Figure 1 This is a system structure diagram of the present invention. Detailed Implementation

[0018] Several embodiments of this application will now be described in more detail with reference to the accompanying drawings to enable those skilled in the art to implement this application. This application may be embodied in many different forms and for various purposes and should not be limited to the embodiments set forth herein. These embodiments are provided to make this application thorough and complete, and to fully convey the scope of this application to those skilled in the art. The embodiments described do not limit this application.

[0019] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It will be further understood that terms such as those defined in commonly used dictionaries shall be interpreted as having a meaning consistent with their meaning in the relevant field and / or the context of this specification, and shall not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0020] Example 1 Its specific implementation method is combined with the appendix Figure 1 Please provide a detailed explanation.

[0021] Appendix Figure 1 This invention provides a structural block diagram of a smart water pollution source tracing system based on three-dimensional fluorescence spectroscopy, which shows the connection relationship between the water sample collection and three-dimensional fluorescence data acquisition module and the source tracing result visualization and smart decision support module, and marks the main functional interaction flow of each module.

[0022] In this embodiment, it includes: The water sample collection and 3D fluorescence data acquisition module is configured to perform standardized collection, preprocessing, and high-precision acquisition of 3D fluorescence spectroscopy (EEM) data of water samples, generating a 3D data matrix containing the number of samples, the number of excitation wavelengths, and the number of emission wavelengths. This ensures the integrity, accuracy, and comparability of the original data, providing a high-quality data foundation for subsequent analysis. Specifically, this includes: The hardware for data acquisition includes: Fully automatic water sampler (supports timed / quantitative / trigger-based sampling, suitable for rivers, lakes, sewage outlets and other scenarios, with sampling volume accuracy of ±2%). Water sample pretreatment unit (including a 0.45μm microporous filter, a constant temperature control module, and a degassing device to avoid interference from suspended solids and temperature fluctuations on fluorescence signals); High-precision three-dimensional fluorescence spectrometer; Based on the source tracing needs (such as watershed investigation / emergency source tracing), sampling points are preset by the system or manually input (combined with GIS map positioning) and sampling frequency (such as 1 time / hour (emergency), 1 time / day (routine monitoring)). After the automatic sampler collects water samples, they are filtered and kept at a constant temperature to remove suspended solids and temperature interference, ensuring the homogeneity of the water samples. The software automatically detects whether there are any abnormalities in the spectral data (such as low signal intensity or baseline drift). After the qualified data is labeled with sample information, it is uploaded to the cloud. Abnormal data triggers equipment fault alarms (such as light source aging or pipeline blockage).

[0023] The data preprocessing and PARAFAC component decomposition module is configured to denoise and purify the raw fluorescence data, and to separate the complex mixed fluorescence signals, i.e., the three-dimensional data matrix, using the PARAFAC (parallel factor analysis) algorithm. Decomposed into independent pure fluorescent components ; in, This is the component concentration score matrix; To excite the load matrix; This is the launch payload matrix; The residual matrix; Achieving dimensionality reduction and extraction from full-spectrum data to essential features specifically includes: Preprocessing includes: Scattering Removal: Adaptive Delaunay trigonometric interpolation is used to automatically identify Rayleigh scattering (λ emission = λ excitation) and Raman scattering (such as the Raman peak at 340 nm in pure water) regions, and interference is removed by interpolation filling; Instrument calibration: Import the excitation source intensity calibration curve and detector response function of the spectrometer to eliminate systematic errors from different batches and different instruments; Unit standardization: The fluorescence intensity is converted to Raman units (RU uses the pure water Raman peak as a reference) to eliminate the influence of concentration differences on spectral comparison.

[0024] The optimal number of groups (usually 3-7, to avoid overfitting / underfitting) is determined through triple validation, including cross-validation (CV), core consistency diagnostics, and variance explained (≥95%). Calculate the residual matrix The standard deviation (≤5%) is used to ensure the reliability of the decomposition results. At the same time, the component types (such as humic acid-like components and protein-like components) are preliminarily determined by comparing the component spectra with standard fluorescent substances (such as humic acid and tyrosine).

[0025] The multi-source pollution characteristic fingerprint database module is configured to construct a structured pollution characteristic database with pure fluorescent components as the core, storing information on various known pollution sources. This provides a reference standard for intelligent comparison of water samples to be tested and is a key support for the accuracy of source tracing. Specifically, it includes: For typical pollution sources (such as printing and dyeing plants, chemical plants, domestic sewage treatment plants, pesticide plants, livestock and poultry farms, etc.), at least 30 parallel samples (covering different production conditions and emission periods) were collected. Extract the pure component spectra and concentration ratios, and record basic information about the pollution source (company name, industry category, coordinates of the discharge outlet, emission limits, characteristic pollutants (such as aniline and rhodamine)). Core fingerprint information: Each pollution source corresponds to a set of characteristic component spectra (such as anthraquinone-like components and azo-like components in dyeing and printing wastewater) and component concentration ratio patterns (such as humic acid-like and protein-like components in chemical wastewater = 3:1). A three-dimensional classification system is used to classify and store data by industry (level 1), region (level 2, such as the Yangtze River Basin and the Pearl River Basin), and pollutant type (level 3, such as organic dyes and phenols), and to create an index to improve the efficiency of subsequent matching.

[0026] Collect data on newly added pollution sources (such as newly built factories) every quarter, and re-verify the fingerprints of existing pollution sources every year (to avoid fingerprint shifts caused by changes in production processes). By analyzing historical matching results through machine learning, the weight of fingerprints that match frequently is increased, while fingerprints with high false positive rates are recalibrated.

[0027] The intelligent comparison and pollution source analysis module is configured as follows: Obtain the fluorescence component spectrum of the water sample to be tested, including the excitation spectrum. emission spectrum and concentration ratio ; calculate , Candidate pollution sources were screened by comparing the cosine similarity (≥0.9 indicates high similarity) and Tucker consistency coefficient (≥0.85 indicates structural consistency) of the spectra of each pollution source component in the database. The concentration ratio of the analyte was calculated using Euclidean distance. The difference between the concentration ratio of the candidate pollution sources and the concentration ratio of the candidate pollution sources is used to eliminate pollution sources with excessive differences from the candidate pollution sources; By combining the sampling time and location of the water sample to be tested with the emission time and geographical location (such as watershed topology) of the candidate pollution sources, pollution sources that do not match in time and space are excluded. Using the component spectra of candidate pollution sources as a standard, the total fluorescence signal of the water sample to be tested is fitted, and the contribution ratio of each candidate pollution source is calculated. The reliability of the results is quantified by calculating the confidence interval (e.g., 70% ± 5%) of the contribution ratio through Monte Carlo simulation (simulating 1000 parameter fluctuations). The rationality of the tracing results was further confirmed through multi-dimensional cross-validation. Pretreatment of water sample components includes: Excitation spectrum of component data obtained from the decomposition of the water sample to be tested emission spectrum and concentration ratio Preprocessing was performed, including component effectiveness screening, spectral normalization optimization, and concentration ratio standardization, to eliminate interference from invalid information. Component effectiveness screening: Calculate the signal-to-noise ratio (S / N) for each component, and retain only components with S / N ≥ 3 (S / N = component fluorescence peak intensity / baseline noise; components with S / N below 3 may be false positive signals caused by instrument error). Secondary optimization of spectral normalization: Peak positions are aligned for the retained component spectra; based on the characteristic peaks of similar components in the database, the peak positions of the components to be measured are adjusted by linear interpolation to avoid matching deviations caused by instrument wavelength drift. Concentration ratio standardization: Convert to a relative percentage (concentration of a component / total concentration of all effective components × 100%) to eliminate the influence of absolute concentration differences on pattern matching.

[0028] The screening process for candidate pollution sources specifically includes: Calculate the analyte separately With database source components cosine similarity ; With database source components cosine similarity ; The calculation formula is: ; in The spectral vector of the component to be measured. For the source component spectral vectors in the database, Number of wavelength points; The characteristic peak regions of the components (such as the excitation at 280 nm and the emission at 350 nm for protein-like components) are assigned a weight of 1.5 times, while the non-characteristic peak regions are assigned a weight of 0.8 times. Because characteristic peaks are unique identifiers of pollution sources, they can reduce interference from non-specific peaks (such as background fluorescence) on matching.

[0029] Set a comprehensive similarity threshold (which can be set to 0.85) and calculate... , average ,like If the overall similarity threshold is greater than or equal to the threshold, then the condition is met and the source can be retained in the candidate pollution source pool. All candidate pollution sources that do not meet the requirements are removed to obtain the screened candidate pollution source pool. For the screened candidate pollution sources, obtain their weighted Euclidean distance. Calculate the concentration ratio of the analyte Ratio of candidate source concentration The difference is expressed by the formula: ; in, For valid component numbers, For the first The weights of each component; The higher the component, The larger; A preset difference threshold (which can be set to 0.15) is used to calculate the weighted Euclidean distance. Compare with the difference threshold, if the weighted Euclidean distance is used If the difference is less than or equal to the preset difference threshold, it is retained in the candidate pollution source pool; all candidate pollution sources that do not meet the requirements are removed from the candidate pollution source pool to obtain the candidate pollution source pool after secondary screening. By combining time, space, and hydrological three-dimensional information, pollution sources that cannot reach the sampling points in time and space are eliminated, resulting in a candidate pollution source pool after three screenings, thus avoiding misjudgment due to similar spectra but no actual relevance. Spatial matching: Based on the GIS map, the straight-line distance (L) between the candidate source discharge outlet and the sampling point is calculated. Combined with watershed hydrological data (such as water flow velocity v=0.5-2m / s), the pollution transmission time t=L / (v×3600) (unit: hours) is calculated. Time matching: Compare the emission time of the candidate source with the sampling time; if the sampling time is 10:00 and the pollution transmission time t=3 hours, then the dyeing and printing plant A needs to emit before 7:00 to reach the sampling point. If it has not started production by 7:00, then the source is excluded. Hydrological constraints: If the sampling point is located upstream of the candidate source (based on the direction of watershed flow), or if there are obstructions such as dams or sluice gates (preventing pollution from being transmitted), the source is directly excluded.

[0030] Using the component spectra of candidate pollution sources as a standard, the total fluorescence signal of the water sample to be tested is fitted, and the contribution ratio of each candidate pollution source is calculated, specifically including: Based on the candidate pollution source pool after three rounds of screening, the MCR-ALS algorithm was used to decompose the contribution of each source to the fluorescence signal of the water sample and convert it into a mass contribution ratio (i.e., the percentage of pollutants emitted by that source in the total pollutants of the water sample). The core steps are as follows: Component spectra of the core suspect , Integrate into a standard source spectral matrix Its dimension is the number of wavelength points × the number of sources, which serves as the basis vector of MCR-ALS; Define algorithm constraints, including non-negativity constraints, closure constraints, and iteration termination conditions; Non-negativity constraint: Both the source contribution ratio and fluorescence intensity are non-negative (there is no negative contribution in actual emissions). Closure constraint: The sum of the contribution ratios of all sources is 100% ± 5% (allowing for small errors due to the presence of a small amount of unidentified background fluorescence). Iteration termination conditions: residual sum of squares (RSS) ≤ 1%, or the number of iterations reaches 500 (to avoid infinite iteration).

[0031] Initialize contribution ratio matrix Alternately update the source spectral matrix With contribution ratio matrix ; , This represents the fluorescence signal matrix of the water sample to be tested. ; Iterate until RSS converges to obtain the final contribution ratio matrix. Each element corresponds to the contribution ratio of a source; If the contribution rate of a certain source is less than 5%, it is judged as a trace contribution and classified as an unknown source; if it is greater than 95%, it is judged as a single source of dominant pollution. Converting fluorescence signal contribution to mass contribution: Through calibration experiments (establishing a linear relationship between fluorescence intensity and mass concentration using polluted water samples of known concentration), the conversion coefficient k is obtained, and the final mass contribution ratio = fluorescence signal contribution × k.

[0032] It also includes uncertainty assessment and cross-validation: The Monte Carlo simulation method is used to evaluate the fluctuation range of the contribution ratio by inputting random perturbation parameters. The steps are as follows: Set perturbation parameters: randomly perturb the source spectrum peak position, component concentration, and conversion coefficient k (determine the perturbation range based on instrument error and experimental repeatability); Multiple simulations: The quantitative analysis process was repeated 1000 times to obtain 1000 sets of contribution ratio results; Confidence intervals were calculated: The 1000 sets of results were statistically analyzed, and the interval range at a 95% confidence level was taken (e.g., dyeing and printing factory A: 72%±4%, which means that there is a 95% probability that its true contribution is between 68% and 76%). Uncertainty classification is obtained: Low uncertainty: confidence interval width < 8% (results are reliable and can be directly used for regulatory purposes); Medium uncertainty: Confidence interval width 8%-15% (requires 1-2 additional sampling points for verification); High uncertainty: confidence interval width > 15% (water samples need to be collected again to check whether the component decomposition is accurate).

[0033] By verifying water quality indicators, characteristic pollutants, and comparing historical data, the rationality of the source tracing results was further confirmed. Water quality index verification: Compare the water sample to be tested with the conventional water quality indicators (such as COD, ammonia nitrogen, and color) of the core suspected source; if the suspected source is a dyeing and printing factory, the water sample should show high COD (>300mg / L), high color (>50 times), and low ammonia nitrogen (<10mg / L), consistent with the characteristics of dyeing and printing wastewater; Validation of characteristic pollutants: Detect the characteristic pollutants of the suspected source in the water sample (such as azo dyes from a dyeing and printing plant, phenols from a chemical plant); if the source is traced back to dyeing and printing plant A, the concentration of azo dye in the water sample is detected to be 2.5 mg / L (the emission limit of this dyeing and printing plant is 5 mg / L, which is consistent with its emission characteristics), and the validation result is reliable; Historical data comparison: Check the pollution source tracing records of this area for the past year; if dyeing and printing factory A was penalized for exceeding emission standards 3 months ago, and the fluorescent components of the water sample at that time were consistent with those in this case, it further confirms the correctness of the results.

[0034] The source tracing results visualization and intelligent decision support module is configured to transform complex source tracing data into intuitive and easy-to-understand visualizations, and provide decision support functions such as pollution early warning, disposal suggestions, and report generation, enabling the implementation of source tracing results into regulatory actions. Specifically, it includes: Spectral comparison chart: The excitation / emission spectra of the water sample components to be tested and the best-matching pollution source components are displayed side by side, with similar peak positions marked to intuitively present the degree of matching; Contribution ratio chart: The contribution ratio of each pollution source is displayed in a pie chart / stacked bar chart, with confidence intervals marked; GIS map display: Mark the sampling points to be tested and match the location of pollution sources on the electronic map, and use arrows to mark the pollution transmission path (in conjunction with the hydrological model). It supports map zooming and layer switching (such as watershed boundaries and administrative divisions). Time series analysis graph: shows the changes in source tracing results at different times for the same sampling point, and analyzes the trend of pollution contribution (e.g., the contribution of a factory increases at night).

[0035] When the contribution rate of a pollution source exceeds the threshold (e.g., ≥50% in a single instance or ≥30% in three consecutive instances), the system will automatically trigger an early warning (sent to environmental regulators via SMS or APP) and mark key suspected pollution sources.

[0036] Based on the matched pollution source type, the corresponding measures are retrieved from the treatment plan database (e.g., if the dyeing and printing wastewater exceeds the standard, the pretreatment facilities in the dye workshop need to be checked; if the chemical wastewater is leaking, the reactor leak needs to be investigated).

[0037] It supports retrieving historical source tracing results by time, region, and pollution source type, generating pollution source tracing ledgers to assist in long-term regulatory decision-making (such as watershed pollution control effect assessment).

[0038] Generate reports that include sampling information, data processing procedures, source tracing results, and early warning suggestions, and support manual editing and modification; Environmental regulators can use the platform to assign pollution source verification tasks to local law enforcement agencies. After the law enforcement agencies provide feedback on the verification results, the system updates the source tracing results.

[0039] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0040] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0041] It should be noted that, in this document, the use of relational terms such as "first" and "second" is merely for distinguishing one entity or operation from another, and does not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0042] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0043] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0044] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0045] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0046] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0047] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0048] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A smart water pollution source tracing system based on three-dimensional fluorescence spectroscopy, comprising: The system comprises a water sample collection and three-dimensional fluorescence data acquisition module, a data preprocessing and PARAFAC component decomposition module, and a multi-source pollution characteristic fingerprint database module. Its distinguishing feature is that it also includes an intelligent comparison and pollution source analysis module, configured as follows: Obtain the fluorescence component spectrum of the water sample to be tested, including the excitation spectrum. emission spectrum and concentration ratio ; calculate , The cosine similarity and Tucker consistency coefficient of the spectra of each pollution source component in the database were used to screen out candidate pollution sources; The concentration ratio of the analyte was calculated using Euclidean distance. The difference between the concentration ratio of the candidate pollution sources and the concentration ratio of the candidate pollution sources is used to eliminate pollution sources with excessive differences from the candidate pollution sources; By combining the sampling time and location of the water sample to be tested with the emission time and geographical location of the candidate pollution sources, pollution sources that do not match in time and space are excluded; Using the component spectra of candidate pollution sources as a standard, the total fluorescence signal of the water sample to be tested is fitted, and the contribution ratio of each candidate pollution source is calculated. The rationality of the tracing results was further confirmed through multi-dimensional cross-validation.

2. The intelligent water pollution source tracing system based on three-dimensional fluorescence spectroscopy according to claim 1, characterized in that, The water sample collection and 3D fluorescence data acquisition module is configured to perform standardized collection, preprocessing, and acquisition of 3D fluorescence spectral data of water samples, generating a 3D data matrix containing the number of samples, the number of excitation wavelengths, and the number of emission wavelengths. .

3. The intelligent water pollution source tracing system based on three-dimensional fluorescence spectroscopy according to claim 1, characterized in that, The data preprocessing and PARAFAC component decomposition module is configured to denoise and purify the raw fluorescence data, and to separate complex mixed fluorescence signals, i.e., the three-dimensional data matrix. Decomposed into ; in, This is the component concentration score matrix; To excite the load matrix; This is the launch payload matrix; The residual matrix; It enables dimensionality reduction and extraction of essential features from full-spectrum data.

4. The intelligent water pollution source tracing system based on three-dimensional fluorescence spectroscopy according to claim 1, characterized in that, The multi-source pollution feature fingerprint database module is configured to build a structured pollution feature database with pure fluorescent components as the core, store information on various known pollution sources, and provide a reference standard for intelligent comparison of water samples to be tested.

5. The intelligent water pollution source tracing system based on three-dimensional fluorescence spectroscopy according to claim 1, characterized in that, In the intelligent comparison and pollution source analysis module, the pretreatment of the water sample components to be tested includes: Excitation spectrum of component data obtained from the decomposition of the water sample to be tested emission spectrum and concentration ratio The preprocessing includes component effectiveness screening, spectral normalization optimization, and concentration ratio standardization to eliminate interference from invalid information.

6. The intelligent water pollution source tracing system based on three-dimensional fluorescence spectroscopy according to claim 5, characterized in that, The screening process for candidate pollution sources specifically includes: Calculate the analyte separately With database source components cosine similarity ; With database source components cosine similarity ; Set a comprehensive similarity threshold and calculate... , average ,like If the overall similarity threshold is greater than or equal to the threshold, then the condition is met and the source can be retained in the candidate pollution source pool. All candidate pollution sources that do not meet the requirements are removed to obtain the screened candidate pollution source pool. For the screened candidate pollution sources, obtain their weighted Euclidean distance. Calculate the concentration ratio of the analyte Ratio of candidate source concentration Differences; The weighted Euclidean distance is obtained by setting a preset difference threshold. Compare with the difference threshold, if the weighted Euclidean distance is used If the difference is less than or equal to the preset difference threshold, it is retained in the candidate pollution source pool; all candidate pollution sources that do not meet the requirements are removed from the candidate pollution source pool to obtain the candidate pollution source pool after secondary screening. By combining time, space, and hydrological three-dimensional information, pollution sources that cannot be reached in time and space are eliminated, resulting in a candidate pollution source pool after three rounds of screening.

7. The intelligent water pollution source tracing system based on three-dimensional fluorescence spectroscopy according to claim 6, characterized in that, Using the component spectra of candidate pollution sources as a standard, the total fluorescence signal of the water sample to be tested is fitted, and the contribution ratio of each candidate pollution source is calculated, specifically including: Based on the candidate pollution source pool after three rounds of screening, the contribution of each source to the fluorescence signal of the water sample is decomposed and converted into a mass contribution ratio. The core steps are as follows: Component spectra of the core suspect , Integrate into a standard source spectral matrix Its dimension is the number of wavelength points × the number of sources; Define algorithm constraints, including non-negativity constraints, closure constraints, and iteration termination conditions; Initialize contribution ratio matrix Alternately update the source spectral matrix With contribution ratio matrix ; , This represents the fluorescence signal matrix of the water sample to be tested. ; Iterate until RSS converges to obtain the final contribution ratio matrix. Each element corresponds to the contribution ratio of a source; Through calibration experiments, the conversion coefficient k was obtained, and the final mass contribution ratio = fluorescence signal contribution × k.

8. The intelligent water pollution source tracing system based on three-dimensional fluorescence spectroscopy according to claim 1, characterized in that, It also includes uncertainty assessment and cross-validation: The Monte Carlo simulation method is used to evaluate the fluctuation range of the contribution ratio by inputting random perturbation parameters, and the uncertainty classification is obtained. By verifying water quality indicators, characteristic pollutants, and comparing historical data, the rationality of the source tracing results was further confirmed.

9. A smart water pollution source tracing system based on three-dimensional fluorescence spectroscopy according to claim 1, characterized in that, Also includes: The source tracing results visualization and intelligent decision support module is configured to transform complex source tracing data into intuitive and easy-to-understand visualization results, and provide functions such as pollution early warning, disposal suggestions, and report generation, so as to realize the implementation of source tracing results into regulatory actions.

Images