A rapid evaluation method for pollution index in sewage and application thereof

Abstract

The present application relates to a kind of fast evaluation method and application of pollution index in sewage, belong to sewage detection technical field.The present application provides a kind of detection method of pollution index in sewage, comprising (1) collecting sample and sewage to be measured, carries out near infrared spectrum analysis and pretreatment, respectively obtains near infrared transmission spectrum;(2) using moving window partial least squares method analysis sample near infrared transmission spectrum and the actual content of pollution index in sample, obtains the characteristic wave band and quantitative model of pollution index;(3) using quantitative model analysis near infrared transmission spectrum of sample to be measured, obtains the evaluation result of corresponding pollution index in sewage sample to be measured.The present application uses spectral analysis to combine MWPLS modeling means, establishes high accuracy pollution index evaluation method, simultaneously realizes the fast analysis and evaluation of multiple pollution index, significantly improves test efficiency;And prediction accuracy is high, meets reproducibility requirement, has higher application value.

Description

Technical Field

[0001] This invention relates to the field of wastewater testing technology, and in particular to a rapid evaluation method and application for pollution indicators in wastewater. Background Technology

[0002] Chemical oxygen demand (COD) is the amount of oxidant consumed when a water sample is treated with a strong oxidant under certain conditions. It reflects the degree of pollution in the water; a higher COD indicates more severe organic pollution. Ammonia nitrogen in water refers to free ammonia (NH4+) in the form of nitrogen. 3+ ) and ammonium ions (NH4+) 4+ Nitrogen exists in the form of ammonia nitrogen; the higher the ammonia nitrogen value, the more severe the eutrophication of the water body. Sulfides in water are mainly produced under anaerobic conditions by the reduction of sulfates by bacteria, and some are produced by the decomposition of sulfur-containing organic matter, primarily in the form of H₂S and H₂S. - \S 2- Sulfides exist in water; soluble metal sulfides, soluble sulfides, and unionized organic and inorganic sulfides in suspended solids are also the main forms of sulfides. Chemical oxygen demand (COD), ammonia nitrogen, and sulfides are important indicators for comprehensively assessing the degree of water pollution.

[0003] Existing water quality analysis methods mainly rely on traditional detection principles, requiring a single instrument to analyze one type of water compound. Due to differing methodologies, they lack compatibility; furthermore, maintenance and repair costs are high, and the purchase cycle for easily worn parts is long. Unsafe and substandard portable analyzers have extremely low testing efficiency, are inconvenient to operate, and are prone to liquid splashes affecting personnel. Therefore, current water quality analysis methods for chemical oxygen demand (COD), ammonia nitrogen, and sulfides involve long pretreatment times and fragmented testing, resulting in lengthy analysis times and hindering rapid water quality monitoring.

[0004] Therefore, developing a rapid and collaborative method for analyzing chemical oxygen demand, ammonia nitrogen, and sulfides in water quality, and comprehensively and rapidly evaluating water pollutant indicators, can effectively improve analysis efficiency and has high application value in the rapid monitoring of multiple wastewater indicators. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a rapid evaluation method and application for pollutant indicators in wastewater. By establishing a rapid evaluation model for pollutant indicators, rapid synergistic analysis of multiple pollutant indicators in wastewater can be achieved, avoiding the cumbersome pretreatment steps of chemical methods, significantly improving testing efficiency, and ensuring that the evaluation results meet the requirements of detection reproducibility, thus having wide applicability.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] In a first aspect, the present invention provides a rapid evaluation method for pollution indicators in wastewater, comprising the following steps:

[0008] (1) Collect samples and wastewater to be tested, perform near-infrared spectroscopy analysis and preprocessing to obtain the near-infrared transmission spectrum of the samples and the near-infrared transmission spectrum of the wastewater to be tested; the samples include polluted water bodies with the same background as the wastewater to be tested, and the number of samples N≥50.

[0009] (2) The near-infrared transmission spectrum of the sample and the actual content of the pollution index in the sample were analyzed by the moving window partial least squares method to obtain the characteristic bands and quantitative models of the pollution index;

[0010] (3) The quantitative model is used to analyze the characteristic bands of the pollution indicators in the near-infrared transmission spectrum of the wastewater to be tested, and the pollution indicator evaluation results of the wastewater to be tested are obtained.

[0011] The pollution indicators include at least one of chemical oxygen demand, ammonia nitrogen content, and sulfide content; the characteristic wavelength of chemical oxygen demand is 820-855nm; the characteristic wavelength of ammonia nitrogen is 1180-1185nm; and the characteristic wavelengths of sulfides are 1420-1430nm and 1300-1340nm.

[0012] This invention uses an infrared light source to perform near-infrared spectral analysis on samples and wastewater samples to be tested, and obtains the corresponding near-infrared transmission spectra after data preprocessing; and establishes a rapid evaluation system for pollution indicators in wastewater through the moving window partial least squares method (MWPLS).

[0013] The main pollution indicators include: chemical oxygen demand (COD), ammonia nitrogen, and sulfides. COD, under certain conditions, is the amount of oxidant consumed after treating a water sample with a strong oxidant, reflecting the degree of organic pollution in the water. Ammonia nitrogen is the concentration of free ammonia (NH3) and ammonium ions (NH4+) in the water. + Nitrogen exists in the form of ) ; sulfides in water are mainly in the reduced state as H2S or S 2- It exists; in the oxidized state, it may be converted into sulfate, and coexist with metals to form soluble or insoluble sulfates.

[0014] By combining MWPLS analysis with the near-infrared transmission spectra of at least 50 samples, the actual content of the samples was correlated with the near-infrared transmission spectra to establish a partial least squares (PLS) model. The optimal characteristic peak bands for assessing pollution indicators were selected, and a quantitative PLS model was constructed. Using the established quantitative model, the near-infrared transmission spectra of the wastewater under test were analyzed, enabling simultaneous and rapid evaluation of multiple pollutants. Within the corresponding characteristic band ranges of the aforementioned pollution indicators, the model exhibits high prediction accuracy and small error index; it can predict the content of pollution indicators in the wastewater under test and comprehensively assess the pollution level. Its predicted values ​​show minimal deviation from the measured analytical values ​​obtained by traditional chemical methods and meet the application requirements for rapid wastewater monitoring.

[0015] Furthermore, the detection method provided by this invention requires only 2-5 mL of sample per test, significantly reducing the sample volume compared to conventional chemical methods. Its detection range is 0-3000 mg / L, making it widely applicable to the determination of various types of wastewater with different background levels. In addition, wastewater generally does not contain visible oil and can be directly collected and analyzed using the above-mentioned rapid evaluation method without pretreatment. For wastewater containing visible oil, or wastewater that is black or contains impurities under abnormal conditions, only filtration or oil skimming is required for detection, avoiding a large number of complex and time-consuming chemical pretreatment processes. This enables rapid simultaneous determination of multiple samples and synergistic analysis of multiple components.

[0016] Compared with traditional chemical methods or other detection methods, the near-infrared transmission spectroscopy combined with model for rapid evaluation and analysis has the advantages of high efficiency, non-destructive nature, safety and environmental protection. It can analyze multiple samples simultaneously, has strong real-time performance, and is suitable for a variety of water quality monitoring application scenarios.

[0017] Preferably, in step (1), the wavelength range of the near-infrared spectroscopy analysis is 780-2500 nm.

[0018] Preferably, in step (1), the preprocessing includes spectral denoising, baseline correction, and spectral normalization. Spectral data is acquired using near-infrared spectroscopy and preprocessed, including but not limited to denoising, baseline correction, and spectral normalization.

[0019] Preferably, in step (2), the actual content of the pollution index in the sample is the content of the pollution index obtained by detecting the sample using chemical methods.

[0020] As a preferred option, the actual contents of the pollutants chemical oxygen demand, ammonia nitrogen, and sulfides in the wastewater samples are the measured contents obtained by using traditional chemical methods HJ / T 399-2007, HJ / T 195-2005, and HJ / T 200-2005, respectively.

[0021] Preferably, in step (2), the moving window partial least squares analysis includes the following steps:

[0022] S1. Determine the width of the movable window: With a width of W... i A subset of the near-infrared transmission spectrum of the sample is extracted within the window, and a PLS model is established with the actual content of pollutants in the sample to calculate the PLS based on the width W. i The error index of the PLS model of the window, wherein the window width corresponding to the PLS model with the smallest error index is W. opt ;

[0023] S2. Construct the initial PLS model and determine the number of latent variables and principal factors;

[0024] S3. Iteration of the moving window: with a width of W opt The window is a movable window that moves within the band of the near-infrared spectral analysis.

[0025] Each time the window moves by at least one wavelength unit, a spectral subset of the near-infrared transmission spectrum of the sample is extracted within the band range of the moving window. Based on the number of latent variables and the number of principal factors, a PLS model is established with respect to the actual content of pollution indicators in the sample.

[0026] Calculate the error index of the PLS model within the band range of the moving window. The PLS model with the smallest error index is the quantitative model. The band range corresponding to the quantitative model is the characteristic band of the pollution index.

[0027] S4. Model Validation: The quantitative model is validated using an independent test set.

[0028] Preferably, the general formula for the PLS model is Y = XP + E. Where Y is the response matrix (chemical analysis data), X is the spectral data matrix, P is the weight matrix of the PLS model, and E is the error matrix.

[0029] Preferably, in S1, W i It consists of 10-50 wavelength units.

[0030] Preferably, the error index is mean square error (MSE) and / or prediction error (RMSEP).

[0031] In MWPLS analysis, choosing an appropriate window size is crucial: the size of the moving window determines the range of spectral bands selected in each iteration; the window size needs to be large enough to capture important spectral features, but should avoid being too large to contain too much irrelevant information.

[0032] Set an initial window size from 10 to 50 wavelength units, with a step size of 1 wavelength unit, and gradually adjust the window size within this range.

[0033] As a preferred approach, for each window size, perform the following operations in sequence:

[0034] ① Select spectral subset: Extract the corresponding band range from the spectral data based on the current window size;

[0035] ②PLS modeling: A PLS model is established using the extracted spectral subset and the actual content of sample contamination indicators;

[0036] ③ Model Evaluation: Evaluate the performance of the current window model using the validation set or cross-validation, and record model error metrics (such as mean squared error (MSE), prediction error (RMSEP), etc.); compare the model performance metrics under different window sizes, and select the window size with the smallest error as the optimal window size W. opt .

[0037] The optimization process described above ensures that the selected window size provides optimal prediction performance in practical applications. If such optimization is performed in a specific experiment, this method can be used to select and record the optimal window size and its corresponding error metric.

[0038] Preferably, in step S2, the number of latent variables is determined through cross-validation.

[0039] In step S2 of MWPLS, an initial PLS model is constructed using spectral data to determine the number of latent variables; and the principal factors corresponding to the highest correlation coefficient R are selected as the principal factors of the model. Subsequent window moving processes are all based on these principal factors.

[0040] As a preferred approach, the number of latent variables is optimized using leave-one-out cross-validation methods or k-fold cross-validation to ensure the model's generalization ability. Determining the number of latent variables (LVs) is a crucial step, primarily aimed at finding the optimal model complexity—capturing the main information in the data without causing overfitting. This is achieved by varying the number of latent variables and observing the model performance under different numbers, such as mean squared error (MSE), correlation coefficient (R), and coefficient of determination (R²). 2 The optimal solution is usually chosen by selecting the number of latent variables that minimizes the error index or maximizes the correlation coefficient, thus determining the number of latent variables.

[0041] As a preferred embodiment, in step S3, the moving window iteration specifically involves moving the window within the band range of the infrared spectral analysis, moving one or more wavelength units at a time. For each window position, the following steps are performed:

[0042] ① Select spectral subset: Extract bands within the current window range from the spectral data;

[0043] ②PLS modeling: A PLS model is established using a selected spectral subset and the actual content of sample contamination indicators;

[0044] ③ Model evaluation: Use the validation set or cross-validation to evaluate the performance of the current window model and record the model error (such as mean squared error MSE or prediction error RMSEP);

[0045] After traversing all window positions, the performance metrics of each window model are compared, and the window with the smallest error is selected as the best model; a subset of bands containing the most relevant spectral information is found, thereby improving the model's predictive ability and stability.

[0046] Preferably, in S4, the verification includes evaluating the accuracy and robustness of the quantitative model.

[0047] The final PLS model within a selected window range is validated using an independent test set to evaluate the model's prediction accuracy and robustness. Evaluation metrics include correlation coefficient (R), root mean square error (RMSE), and bias.

[0048] Secondly, this invention provides the application of the rapid evaluation method for pollution indicators in wastewater in water quality monitoring.

[0049] The resulting quantitative model can be applied to the rapid assessment of wastewater, and its applicability and stability can be continuously optimized based on feedback from practical applications. The model's performance can be evaluated by comparing and summarizing the mean deviations of the reference data from the calibration and validation sets, as well as the predicted data.

[0050] During application, the system can be calibrated using newly acquired sample spectra to better evaluate the accuracy and extension of the model. When it is found that there is room for improvement in the model performance, the spectrum and data corresponding to the sample can be incorporated into the model to achieve the purpose of model calibration.

[0051] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0052] This invention provides a method for detecting pollutants in wastewater, enabling rapid and synergistic analysis of chemical oxygen demand (COD), ammonia nitrogen, and sulfides. This significantly improves the efficiency of pollutant analysis and has wide applicability. The method reduces personnel contact with wastewater samples, lowers pretreatment and detection times, and enhances work efficiency and safety. Multiple pollutant indicators can be rapidly and synergistically analyzed, eliminating the cumbersome processes of traditional high-temperature digestion and chemical reactions. The analytical efficiency is expected to increase by 98%, making it highly valuable for rapid water quality monitoring. Attached Figure Description

[0053] Figure 1 This is a diagram showing the interface for analyzing and processing the near-infrared transmission spectrum of the sample from Example 1.

[0054] Figure 2 This is a comparison chart of the chemical oxygen demand (COD) results and laboratory values ​​from the quantitative model analysis in Example 1.

[0055] Figure 3 This is a comparison chart of the ammonia nitrogen content analysis results and laboratory values ​​from Example 1 using the quantitative model.

[0056] Figure 4 This is a comparison chart of the sulfide content analysis results and laboratory values ​​from Example 1 using the quantitative model. Detailed Implementation

[0057] To better illustrate the purpose, technical solution, and advantages of the present invention, the present invention will be further described below in conjunction with specific embodiments. Unless otherwise specified, the experimental methods used in the following embodiments are conventional methods; the materials and reagents used, unless otherwise specified, are commercially available reagents and materials.

[0058] Example 1

[0059] An embodiment of the rapid evaluation method for pollution indicators in wastewater according to the present invention specifically includes the following steps:

[0060] Taking the pollutant indicator chemical oxygen demand (COD) in this embodiment as an example, the following quantitative model establishment steps (1)-(6) are performed.

[0061] (1) Data Acquisition and Preprocessing

[0062] 169 wastewater samples were collected from the inlet / outlet of the rainwater grease trap, MBR effluent, reused wastewater, A / O sedimentation tank, contact oxidation tank effluent, Szorb flue gas desulfurization wastewater, and the pearl discharge outlet of the wastewater treatment system. Based on the background levels of the wastewater samples, they were categorized into seven types as shown in Table 1. The wastewater samples in this embodiment all came from Category I wastewater.

[0063] Near-infrared spectroscopy (780-2500 nm) was used to analyze the spectral data of 169 samples. Preprocessing included noise reduction, baseline correction, and spectral normalization to obtain near-infrared transmission spectra. Simultaneously, COD values ​​were measured using chemical methods according to HJ / T 399-2007, representing the actual COD content of the pollution indicator. The processed spectra and actual content results were then input into a processing system for further processing. The interface is shown below. Figure 1 .

[0064] Table 1. Classification list of wastewater background in a wastewater treatment plant according to the present invention.

[0065]

[0066] (2) Determine the moving window

[0067] ① Set the initial window size range: Set the window width W i It consists of 10 to 50 wavelength units, with a step size of 1 wavelength unit.

[0068] ② Iterate through window sizes: Within the range mentioned above, gradually adjust the window width. For each window width, perform the following operations in sequence:

[0069] • Select spectral subset: Extract the corresponding band range from the spectral data of the sample based on the current window size.

[0070] • PLS modeling: A PLS model is built using the extracted spectral subset and the actual COD content of the sample.

[0071] • Model evaluation: Evaluate the performance of the current window model using the validation set or cross-validation, and record the model error metrics: mean squared error (MSE) and prediction error (RMSEP).

[0072] ③ Select the optimal window size: Compare the error indices of the PLS models built with different window widths, and select the window width with the smallest error as the optimal window width W. opt The results are shown in Table 2.

[0073] As can be seen, the model has the smallest mean squared error (MSE) and prediction error (RMSEP) when the window width is 35 wavelength units. Therefore, 35 wavelength units are chosen as the optimal window width W. opt .

[0074] Table 2 Error Indicators of PLS ​​Models with Different Window Widths (Wi)

[0075] Window width Wi Mean Square Error (MSE) Prediction Error RMSEP 10 5.4 7.8 15 4.8 7.1 20 4.5 6.9 25 4.3 6.7 30 4.1 6.6 35 4.0 6.5 40 4.2 6.6 45 4.5 6.9 50 4.8 7.1

[0076] (3) Constructing the initial PLS model

[0077] An initial PLS model was constructed using spectral data from 69 preprocessed samples. The number of latent variables was determined through cross-validation, and the number of principal factors was selected based on the MSE and Bias results under different principal factors.

[0078] The general formula for the PLS model used is Y = XP + E. Where Y is the response matrix (chemical analysis data), X is the spectral data matrix, P is the weight matrix of the PLS model, and E is the error matrix. As shown in Tables 3 and 4, the number of latent variables with the smallest error index (5) was selected, and the correlation coefficient R was chosen. 2 The highest principal factor number, 9, was used as the number of latent variables and principal factors in subsequent moving window iterations.

[0079] Table 3. Screening results for the number of latent variables

[0080]

[0081] Table 4. Results of principal factor selection

[0082] Number of principal factors <![CDATA[R 2 ]]> RMSEC RMSEP RPD 9 0.953 6.71 5.94 3.984 8 0.922 7.04 6.03 3.414 7 0.895 7.82 6.16 3.407 6 0.886 8.45 6.26 3.369 5 0.832 9.94 6.44 3.32 4 0.795 10.56 6.98 3.31 3 0.346 11.84 7.65 3.217 2 0.265 12.46 8.96 2.394 1 0.178 14.21 10.62 1.651

[0083] (4) Moving window iteration

[0084] With a width of W opt The window is a moving window, moving within the 780-2500nm wavelength range in the near-infrared spectral analysis of the sample, moving one or more wavelengths at a time; for each window position, the following steps are performed:

[0085] • Select spectral subset: Extract bands within the current window range from the spectral data;

[0086] • PLS modeling: A PLS model is built using a selected spectral subset and the actual COD content of the corresponding sample.

[0087] • Model evaluation: Use the validation set or cross-validation to evaluate the performance of the current window model and record the model error metrics: mean squared error (MSE) and prediction error (RMSEP).

[0088] (5) Select the best window

[0089] After traversing all window positions, the error index of each window model is compared, and the window with the smallest error is selected as the basis for the quantitative model. Based on this window, a PLS quantitative model for the corresponding pollution index is established; the range of this window is the characteristic band range.

[0090] (6) Final model validation

[0091] Fifty wastewater samples, originating from the same background as the sample but independent of it, were collected as an independent test set. The quantitative PLS model within the characteristic band range was validated using the independent test set to evaluate the model's prediction accuracy and robustness.

[0092] (7) The modeling and evaluation process of the above (1)-(6) was carried out on the pollution indicators ammonia nitrogen and sulfide content in the sample. The model verification results are shown in Table 5 below.

[0093] In determining the characteristic bands of sulfides, it was found that the combination of the two bands in the table (1420-1430nm and 1300-1340nm) can provide higher prediction accuracy and lower error index.

[0094] The verification results of the quantitative model show that the three pollution indicators obtained by the above method in specific characteristic bands have good quantitative accuracy and robustness.

[0095] Table 5. Validation results of the quantitative model for pollution indicators

[0096] Characteristic band (nm) R RMSEC RMSEP RPD Chemical oxygen demand 820-855 0.968 4.21 4.69 3.43 ammonia nitrogen 1180-1185 0.973 0.082 0.103 5.32 sulfides 1420-1430、1300-1340 0.956 0.038 0.074 4.68

[0097] (7) Application of rapid evaluation model for pollution indicators

[0098] The obtained quantitative model was applied to the determination of pollution indicators in the wastewater to be tested:

[0099] A total of 36 wastewater samples were collected from the inlet / outlet of the rainwater oil separator, the effluent of the MBR, the reused wastewater (effluent from the adsorption tank), the A / O sedimentation tank, the effluent from the contact oxidation tank, and the Szorb flue gas desulfurization wastewater. The near-infrared transmission spectrum data of the wastewater samples were obtained by the same processing as in step (1). The rapid evaluation values ​​of COD, ammonia nitrogen and sulfide were obtained by quantitative model analysis of the three pollution indicators in the corresponding characteristic bands.

[0100] Meanwhile, the COD, ammonia nitrogen, and sulfide contents of 36 wastewater samples were determined using HJ / T 399-2007, HJ / T 195-2005, and HJ / T 200-2005, respectively. The measured values ​​for the corresponding samples were obtained and compared in Table 6 below. Figure 2-4 .

[0101] From Table 6 Figure 2-4 It can be seen that the detection method provided by this invention can simultaneously analyze pollutants such as chemical oxygen demand (COD), ammonia nitrogen, and sulfides in wastewater, and the analytical results are relatively accurate. Compared with the traditional method's detection time of about 50 minutes, this invention can eliminate cumbersome processes such as high-temperature digestion and chemical reactions, improve detection efficiency, reduce labor costs, and increase analytical efficiency by approximately 98%.

[0102] Table 6 Comparison of the rapid evaluation results of pollution indicators by this invention with actual test values.

[0103]

[0104]

[0105] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the essence and scope of the technical solutions of the present invention.

Claims

1. A rapid evaluation method for pollution indicators in wastewater, characterized in that, Includes the following steps: (1) Collect samples and wastewater to be tested, perform near-infrared spectroscopy analysis and preprocessing to obtain the near-infrared transmission spectrum of the samples and the near-infrared transmission spectrum of the wastewater to be tested; the samples include polluted water bodies with the same background as the wastewater to be tested, and the number of samples N≥50. (2) The near-infrared transmission spectrum of the sample and the actual content of the pollution index in the sample were analyzed by the moving window partial least squares method to obtain the characteristic bands and quantitative models of the pollution index; (3) The quantitative model is used to analyze the characteristic bands of the pollution indicators in the near-infrared transmission spectrum of the wastewater to be tested, and the pollution indicator evaluation results of the wastewater to be tested are obtained. The pollution indicators include at least one of chemical oxygen demand, ammonia nitrogen content, and sulfide content; the characteristic wavelength of chemical oxygen demand is 820-855nm; the characteristic wavelength of ammonia nitrogen is 1180-1185nm; and the characteristic wavelengths of sulfides are 1420-1430nm and 1300-1340nm.

2. The rapid evaluation method for pollution indicators in wastewater according to claim 1, characterized in that, In step (1), the wavelength range of the near-infrared spectroscopy analysis is 780-2500nm.

3. The rapid evaluation method for pollution indicators in wastewater according to claim 1, characterized in that, In step (1), the preprocessing includes spectral denoising, baseline correction, and spectral normalization.

4. The rapid evaluation method for pollution indicators in wastewater according to claim 1, characterized in that, In step (2), the actual content of the pollution index in the sample is the content of the pollution index obtained by detecting the sample using chemical methods.

5. The rapid evaluation method for pollution indicators in wastewater according to claim 2, characterized in that, In step (2), the moving window partial least squares analysis includes the following steps: S1. Determine the width of the movable window: With a width of W... i A subset of the near-infrared transmission spectrum of the sample is extracted within the window, and a PLS model is established with the actual content of pollutants in the sample to calculate the PLS based on the width W. i The error index of the PLS model of the window, wherein the window width corresponding to the PLS model with the smallest error index is W. opt ; S2. Construct the initial PLS model and determine the number of latent variables and principal factors; S3. Iteration of the moving window: with a width of W opt The window is a movable window that moves within the band of the near-infrared spectral analysis. Each time the window moves by at least one wavelength unit, a spectral subset of the near-infrared transmission spectrum of the sample is extracted within the band range of the moving window. Based on the number of latent variables and the number of principal factors, a PLS model is established with respect to the actual content of pollution indicators in the sample. Calculate the error index of the PLS model within the band range of the moving window. The PLS model with the smallest error index is the quantitative model. The band range corresponding to the quantitative model is the characteristic band of the pollution index. S4. Model Validation: The quantitative model is validated using an independent test set.

6. The rapid evaluation method for pollution indicators in wastewater according to claim 5, characterized in that, In S1, W i It consists of 10-50 wavelength units.

7. The rapid evaluation method for pollution indicators in wastewater according to claim 5, characterized in that, The error metrics are mean square error (MSE) and / or prediction error (RMSEP).

8. The rapid evaluation method for pollution indicators in wastewater according to claim 7, characterized in that, In S2, the number of latent variables is determined through cross-validation.

9. The rapid evaluation method for pollution indicators in wastewater according to claim 5, characterized in that, In S4, verification includes evaluating the accuracy and robustness of the quantitative model.

10. The application of the rapid evaluation method for pollution indicators in wastewater according to any one of claims 1-9 in water quality monitoring.

Images