Water quality evaluation method and device based on quantitative grading of non-quantitative environmental pollution indicators

Abstract

This application discloses a water quality assessment method and apparatus based on the quantitative grading of non-quantitative environmental pollution indicators. The method includes: screening surface water quality indicators and classifying them into direct and indirect water quality indicators; determining the assessment section and assessment period, and obtaining raw water quality concentration data for different times and sections within the assessment period; screening a predetermined number of direct water quality indicators related to indirect water quality indicators; performing PLSR regression and calculating the regression quantitative grading standard for each indirect water quality indicator; standardizing the raw detection data, the national grading standard for direct water quality indicators, and the regression quantitative grading standard for indirect water quality indicators into dimensionless values; performing PCA analysis according to the assessment period and assessment section to determine the achieved water quality environmental functional zone category; and obtaining typical water quality indicators for the section based on the PCA analysis. This method and apparatus can comprehensively consider the relationship between quantitative and non-quantitative environmental pollution indicators and the contribution of indirect indicators, making the assessment results more comprehensive and objective.

Description

Technical Field

[0001] This invention belongs to the field of water quality assessment technology, and in particular relates to a water quality assessment method and apparatus based on the quantitative classification of non-quantitative indicators of environmental pollution. Background Technology

[0002] Surface water assessment methods include single-factor assessment and comprehensive factor assessment. Among them, the single-factor assessment method can only evaluate indicators directly related to the surface water environmental quality status. It uses the ratio of each water quality indicator to the current national surface water environmental function standard as the evaluation result. A value greater than 1 indicates that the indicator exceeds the environmental function zone standard. If any one of the many surface water environmental indicators exceeds the standard, it means that the environmental function zone of the water quality section does not meet the standard. The evaluation results of this single-factor assessment method are pessimistic and one-sided, and do not pay attention to the relationship between multiple water quality indicators. The comprehensive factor evaluation method, also known as the multi-factor evaluation method, is mainly divided into two categories: The first category uses one or more weighting coefficient algorithms to assign weight coefficients to multiple water quality indicators, and finally superimposes them to form a comprehensive evaluation index. Among these weighting coefficient algorithms are entropy method, CRITIC weighting method, independence weighting coefficient method, coefficient of variation method, and custom weighting method, etc. The rationality of the weighting coefficients needs to be verified by mathematical testing methods. The second category first standardizes each indicator, selects one or more regression algorithms to determine the intrinsic relationship and importance between multiple indicators, and finally fits the regression to form a comprehensive evaluation index. Among these multi-indicator regression algorithms are least squares method, partial least squares method, machine learning method, and clustering, etc. The effectiveness of the fitted regression model needs to be verified by mathematical testing methods. The comprehensive evaluation method completes the comprehensive evaluation of surface water by comparing the comprehensive evaluation index of multiple cross sections at different times, which can reflect the overall condition of surface water and make the evaluation results more objective and comprehensive.

[0003] Water quality indicators refer to a series of standards used to describe the state of water quality. Usually, the degree of pollution of a water body can be judged by evaluating some water quality indicators. Commonly used water quality indicators include the following categories: (1) Physical indicators, including sensory physical indicators, such as temperature, color, turbidity, transparency, flow velocity (rivers), flow rate (lake and reservoir capacity), etc., as well as other physical water quality indicators, such as total solids, suspended solids, fixed solids, conductivity (resistivity), dissolved oxygen, oxidation-reduction potential, chlorophyll a, etc.; (2) Chemical water quality indicators, including organic matter indicators, such as permanganate index (CODmn), chemical oxygen demand (CODmn), etc. CODcr), five-day biochemical oxygen demand (BOD5), total organic carbon (TOC) and total oxygen demand (TOD), as well as eutrophication indicators such as ammonia nitrogen, Kjeldahl nitrogen, nitrite, nitrate, total nitrogen and total phosphorus, etc., and inorganic non-metallic compounds such as total arsenic (As), selenium (Se), fluoride, sulfide etc., and heavy metals such as mercury (Hg), cadmium (Cd), lead (Pb), chromium (Cr), etc., and toxic and harmful organic substances such as volatile phenols, cyanide, petroleum, anionic surfactants etc.; (3) biological water quality indicators, including total bacteria count, total coliform count, fecal coliform, etc. Most chemical indicators, biological indicators and some physical indicators have clear national classification standards, and the surface water environmental quality pollutant status can be directly evaluated using water quality evaluation methods. The scheme is defined as direct water quality indicators (quantitative indicators of environmental pollution). While other physical indicators have well-established detection methods, they lack clear national classification standards. However, they remain closely related to surface water environmental quality and are defined in this scheme as indirect water quality indicators (non-quantitative indicators of environmental pollutants). In the surface water environment, direct and indirect water quality indicators are correlated and work together to affect environmental quality. Some indirect water quality indicators even cause changes in direct water quality indicators and are directly related to the state of water quality. However, existing methods, whether single-factor or comprehensive factor evaluation, only focus on direct water quality indicators and do not provide a quantitative comprehensive evaluation of the contribution of indirect indicators to the surface water environment or the relationship between direct and indirect water quality indicators. Furthermore, non-quantitative environmental pollution indicators lack systematic classification standards, making it difficult to participate in quantitative comprehensive water quality evaluation. Summary of the Invention

[0004] To address the aforementioned issues, this invention provides a water quality assessment method and apparatus based on the quantitative grading of non-quantitative environmental pollution indicators. This method comprehensively considers the relationship between quantitative and non-quantitative environmental pollution indicators, as well as the contribution of indirect indicators, to achieve the quantitative grading of non-quantitative environmental pollution indicators. By combining single-factor environmental overload assessment with comprehensive factor environmental condition assessment, the assessment results become more comprehensive and objective.

[0005] The water quality assessment method based on quantitative grading of non-quantitative environmental pollution indicators provided by this invention includes:

[0006] Surface water quality indicators were screened and classified into direct and indirect water quality indicators.

[0007] Determine the evaluation section and evaluation period, and obtain raw water quality concentration detection data for different times and different sections within the evaluation period;

[0008] Process the maximum values ​​in the indirect water quality indicators, and filter a preset number of direct water quality indicators that are related to the indirect water quality indicators;

[0009] By using the original detection data of concentrations of a single indirect water quality indicator and a pre-selected number of direct water quality indicators, the PLSR regression is performed to determine the environmental standards for evaluating the environmental functional zones of surface water bodies and the direct indicators. The regression quantitative grading standards for indirect water quality indicators are calculated one by one using the PLSR regression formula.

[0010] The original test data, the national grading standards for direct water quality indicators, and the regression quantitative grading standards for indirect water quality indicators were all standardized into dimensionless values.

[0011] The standardized direct water quality indicators and the national classification standards, the indirect water quality indicators and the regression quantification classification standards of the indirect water quality indicators are subjected to PCA analysis according to the evaluation period and evaluation section. The water quality environmental functional zone category is determined by the ranking order of the comprehensive PCA scores of different periods, different sections and classification standards.

[0012] Typical water quality indicators of the section are analyzed based on the PCA load coefficient obtained from PCA analysis.

[0013] Preferably, in the above-mentioned water quality assessment method based on the quantitative grading of non-quantitative environmental pollution indicators, the feature importance function in the decision tree regression algorithm for indirect and direct water quality indicators is used to screen a preset number of direct water quality indicators related to indirect water quality indicators.

[0014] Preferably, in the above-mentioned water quality assessment method based on the quantitative grading of non-quantitative environmental pollution indicators, the maximum value of the indirect water quality indicator is:

[0015] Calculate the coefficient of variation (CV) of indirect water quality indicators. When an indirect water quality indicator is more than three times the maximum CV of a direct water quality indicator, it is determined to be an outlier. When performing PLSR regression, the outlier is changed to the average value within a preset time range before and after it.

[0016] Preferably, in the above-mentioned water quality evaluation method based on the quantitative grading of non-quantitative environmental pollution indicators, the original detection data, the direct water quality indicator grading standard, and the indirect water quality indicator regression quantitative grading standard are all standardized into dimensionless values ​​using the environmental pollution index method or mathematical methods.

[0017] Preferably, in the above-mentioned water quality evaluation method based on the quantitative grading of non-quantitative environmental pollution indicators, the environmental pollution index method is the single-factor standard index method or the Nemerow method.

[0018] Preferably, in the above-mentioned water quality assessment method based on the quantitative grading of non-quantitative environmental pollution indicators, the selection of a preset number of direct water quality indicators related to indirect water quality indicators includes:

[0019] The selected direct water quality indicators shall be no less than three, and shall be selected in descending order of importance based on their characteristics, with a cumulative importance ratio of no less than 90%.

[0020] Preferably, in the above-mentioned water quality evaluation method based on the quantitative grading of non-quantitative indicators of environmental pollution, the preset time is two cycles.

[0021] The present invention provides a water quality assessment device based on quantitative grading of non-quantitative indicators of environmental pollution, comprising:

[0022] Screening and classification unit, used to screen surface water quality indicators and classify them into direct water quality indicators and indirect water quality indicators;

[0023] The raw detection data acquisition unit is used to determine the evaluation section and evaluation period, and to acquire raw detection data of water quality concentration at different times and different sections within the evaluation period.

[0024] A direct water quality index screening unit is used to process the maximum value in the indirect water quality index and screen a preset number of direct water quality indicators related to the indirect water quality index.

[0025] The PLSR regression unit is used to perform PLSR regression by comparing the original detection data of the concentration of a single indirect water quality indicator with the preset number of direct water quality indicators after screening. This determines the environmental standards for evaluating the environmental functional zones of surface water bodies and the environmental standards implemented for direct indicators. The PLSR regression formula is used to calculate the regression quantitative grading standards for indirect water quality indicators one by one.

[0026] The standardized unit is used to standardize the original test data, the national grading standards for direct water quality indicators, and the regression quantitative grading standards for indirect water quality indicators into dimensionless values.

[0027] The water quality environmental functional zone category determination unit is used to perform PCA analysis on the standardized direct water quality indicators and the national classification standards, the indirect water quality indicators and the indirect water quality indicator regression quantification classification standards according to the evaluation period and evaluation section, and determine the water quality environmental functional zone category achieved by ranking the PCA comprehensive scores of different periods, different sections and classification standards.

[0028] Typical water quality index analysis unit, used to analyze typical water quality indexes of the section based on the PCA load coefficient obtained from PCA analysis.

[0029] Preferably, in the above-mentioned water quality assessment device based on the quantitative grading of non-quantitative indicators of environmental pollution, the direct water quality indicator screening unit is specifically used to screen a preset number of direct water quality indicators related to indirect water quality indicators by utilizing the feature importance function in the decision tree regression algorithm of indirect and direct water quality indicators.

[0030] Preferably, in the above-mentioned water quality assessment device based on the quantitative grading of non-quantitative indicators of environmental pollution, the direct water quality indicator screening unit is also specifically used to calculate the coefficient of variation (CV) value of indirect water quality indicators. When an indirect water quality indicator that is more than three times the maximum CV value of the direct water quality indicator is determined to be an outlier, the outlier value is changed to the average value within a preset time range before and after it when performing PLSR regression.

[0031] As described above, the water quality assessment method based on the quantitative grading of non-quantitative environmental pollution indicators provided by this invention includes: screening surface water quality indicators and classifying them into direct and indirect water quality indicators; determining the assessment section and assessment period, and obtaining raw water quality concentration detection data at different times and sections within the assessment period; processing the maximum values ​​in the indirect water quality indicators, and screening a preset number of direct water quality indicators related to the indirect water quality indicators; performing PLSR regression on the raw concentration detection data of a single indirect water quality indicator and the preset number of direct water quality indicators after screening, determining the environmental functional zones of the surface water body and the environmental standards implemented for the direct indicators, and calculating the regression quantitative grading standards for the indirect water quality indicators one by one using the PLSR regression formula; and combining the raw detection data and direct water quality indicators... The national grading standards and the indirect water quality index regression quantitative grading standards are all standardized into dimensionless values. The standardized direct water quality indicators, the national grading standards, the indirect water quality indicators, and the indirect water quality index regression quantitative grading standards are then subjected to PCA analysis according to the evaluation period and evaluation section. The water quality environmental functional zone category is determined by the ranking order of the PCA comprehensive scores of different periods, different sections, and grading standards. Based on the PCA load coefficients obtained from the PCA analysis, typical water quality indicators of the sections are analyzed. Therefore, the relationship between quantitative and non-quantitative environmental pollution indicators and the contribution of indirect indicators can be comprehensively considered, achieving quantitative grading of non-quantitative environmental pollution indicators. Combining single-factor environmental overload assessment and comprehensive factor environmental actual condition assessment makes the evaluation results more comprehensive and objective. The water quality evaluation device based on the quantitative grading of non-quantitative environmental pollution indicators provided by this invention has the same advantages as the aforementioned water quality evaluation method based on the quantitative grading of non-quantitative environmental pollution indicators. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0033] Figure 1 This is a schematic diagram of an embodiment of a water quality evaluation method based on the quantitative classification of non-quantitative environmental pollution indicators provided by the present invention;

[0034] Figure 2 A schematic diagram of an embodiment of a water quality assessment device based on quantitative grading of non-quantitative indicators of environmental pollution provided by the present invention;

[0035] Figure 3 A scree plot drawn based on the degree to which each principal component explains the variation in the data;

[0036] Figure 4 This is a heatmap of the factor loading matrix. Detailed Implementation

[0037] The core of this invention is to provide a water quality assessment method and device based on the quantitative grading of non-quantitative environmental pollution indicators. This method can comprehensively consider the relationship between quantitative and non-quantitative environmental pollution indicators and the contribution of indirect indicators, thereby achieving the quantitative grading of non-quantitative environmental pollution indicators. By combining single-factor environmental overload assessment and comprehensive factor environmental actual condition assessment, the assessment results are more comprehensive and objective.

[0038] 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.

[0039] An example implementation of a water quality assessment method based on the quantitative classification of non-quantitative environmental pollution indicators provided by this invention. Figure 1 As shown, Figure 1 This is a schematic diagram of an embodiment of a water quality assessment method based on the quantitative classification of non-quantitative indicators of environmental pollution provided by the present invention. The method may include the following steps:

[0040] S1: Screen surface water quality indicators and classify them into direct water quality indicators and indirect water quality indicators;

[0041] It should be noted that the aforementioned direct water quality indicators refer to quantitative environmental pollution indicators, while indirect indicators refer to non-quantitative environmental pollution indicators. Specifically, a set of water quality indicators M can be screened and categorized based on historical surface water pollution records and drainage types. If no characteristic pollutants are found, conventional water quality testing indicators from GB3838-2002 "Surface Water Environmental Quality Standard" can be selected, along with river flow velocity, lake / reservoir flow rate Q, chlorophyll a, air temperature T, transparency SD, suspended solids SS, turbidity NUL, color, oxidation-reduction potential ORP, and conductivity.

[0042] S2: Determine the evaluation section and evaluation period, and obtain the original water quality concentration detection data of different times and different sections within the evaluation period;

[0043] It should be noted that surface water does not have an ice-free period. The criteria for judgment are: when T < 0℃, TW ≥ 0℃. The testing period is from January to December each year, and the tests are conducted 1 to 3 times per month. Here, T is the air temperature of the area where the water body is located, in ℃; and TW is the water temperature of the water body, in ℃.

[0044] S3: Process the maximum value in indirect water quality indicators and filter a preset number of direct water quality indicators related to indirect water quality indicators;

[0045] Specifically, the coefficient of variation (CV) of indirect water quality indicators can be calculated. If the CV value is more than three times the maximum CV value of direct water quality indicators, the outlier value will be the average of the two months before and after the PLSR regression. It should be noted that the maximum value of the indirect water quality indicator, i.e., the outlier, is only replaced during PLSR regression to obtain the best regression effect and minimize regression error. However, the original value is still used when calculating the water quality category in PCA. The feature importance function of the machine learning regression algorithm from indirect water quality indicators to direct water quality indicators is used to select the most important direct water quality indicators related to the indirect water quality indicators. The selected direct water quality indicators should be no less than three, and their cumulative importance should be no less than 90%. Of course, the above processing method can be adaptively adjusted according to actual needs, which is not limited here.

[0046] S4: By using the original detection data of the concentration of a single indirect water quality indicator and a pre-selected number of direct water quality indicators, the PLSR regression is performed to determine the environmental standards for evaluating the surface water environmental functional zones and the direct indicators. The regression quantitative grading standards for indirect water quality indicators are calculated one by one using the PLSR regression formula.

[0047] It should be noted that existing national, local, and industry standards do not provide an effective classification of non-quantitative environmental pollution indicators. However, if we are to participate in comprehensive water quality assessment, we need reasonable and effective classification standards. Therefore, this step can be used to dynamically screen non-quantitative environmental pollution indicators and calculate them one by one, which means quantifying the classification standards of these indicators.

[0048] S5: Standardize the original test data, the national grading standards for direct water quality indicators, and the regression quantitative grading standards for indirect water quality indicators into dimensionless values.

[0049] Specifically, the environmental pollution index method or mathematical methods can be used to standardize the original detection data, the national classification standards for direct water quality indicators, and the regression quantitative classification standards for indirect indicators into dimensionless values. The mathematical methods include, but are not limited to, maximum value, average value, weighted average method, and coefficient of variation method, etc.

[0050] S6: The standardized direct water quality indicators and national classification standards, indirect water quality indicators and indirect water quality indicators regression quantification classification standards are subjected to PCA analysis according to the evaluation period and evaluation section. The water quality environmental functional zone category is determined by the ranking order of the comprehensive PCA scores of different periods, different sections and classification standards.

[0051] It should be noted that the PCA analysis mentioned above is also known as principal component analysis.

[0052] S7: Typical water quality indicators of the section based on the PCA load coefficient obtained from PCA analysis.

[0053] As described above, the embodiments of the water quality assessment method based on the quantitative grading of non-quantitative environmental pollution indicators provided by the present invention include: screening surface water quality indicators and classifying them into direct and indirect water quality indicators; determining the assessment section and assessment period, and obtaining raw water quality concentration detection data at different times and sections within the assessment period; processing the maximum values ​​in indirect water quality indicators, and screening a preset number of direct water quality indicators related to indirect water quality indicators; performing PLSR regression on the raw concentration detection data of a single indirect water quality indicator and the preset number of direct water quality indicators after screening, determining the environmental functional zones of the surface water body and the environmental standards implemented for direct indicators, and calculating the regression quantitative grading standards for indirect water quality indicators one by one using the PLSR regression formula; and combining the raw detection data and direct water quality indicators with the indirect water quality indicators. Both the national grading standards for direct water quality indicators and the regression-quantification grading standards for indirect water quality indicators are standardized into dimensionless values. The standardized direct water quality indicators and the national grading standards, as well as the indirect water quality indicators and the regression-quantification grading standards, are then subjected to PCA analysis according to the evaluation period and evaluation section. The water quality environmental functional zone category is determined by the ranking order of the PCA comprehensive scores for different periods, different sections, and grading standards. Typical water quality indicators of the sections are analyzed based on the PCA load coefficients obtained from the PCA analysis. Therefore, it is possible to comprehensively consider the relationship between quantitative and non-quantitative environmental pollution indicators and the contribution of indirect indicators, achieving quantitative grading of non-quantitative environmental pollution indicators. Combining single-factor environmental overload assessment and comprehensive factor environmental condition assessment makes the evaluation results more comprehensive and objective.

[0054] In a specific embodiment of the water quality assessment method based on the quantitative grading of non-quantitative environmental pollution indicators, the feature importance function in the decision tree regression algorithm for indirect and direct water quality indicators is used to screen a preset number of direct water quality indicators related to indirect water quality indicators.

[0055] Specifically, the method utilizes decision tree machine learning regression to screen the importance of indirect water quality indicators with other direct water quality indicators. By performing partial least squares regression (PLSR) on the original detection data of the concentrations of a single indirect water quality indicator and the screened direct water quality indicators, the method determines the environmental functional zones of surface water bodies and the environmental standards to be implemented for direct indicators. Finally, the method uses the PLSR regression formula to calculate the comprehensive evaluation and quantitative grading standards for indirect water quality indicators one by one.

[0056] In another specific embodiment of the water quality assessment method based on the quantitative grading of non-quantitative environmental pollution indicators described above, the treatment of the maximum value in indirect water quality indicators can be specifically as follows:

[0057] Calculate the coefficient of variation (CV) of indirect water quality indicators. If an indirect water quality indicator is more than three times the maximum CV of a direct water quality indicator, it is considered an outlier. When performing PLSR regression, the outlier is changed to the average value within a preset time range before and after it.

[0058] In this embodiment, by processing the maximum value of the indirect indicator in the above manner, the error during subsequent PLSR regression can be reduced. Furthermore, the preset time can preferably be two consecutive periods. It should be noted that when monthly data is selected, if an outlier occurs, the average of the two consecutive months is used to replace the outlier. If annual or daily data is selected, the average of the two consecutive years and the two consecutive days are used to replace the outlier.

[0059] In another specific embodiment of the water quality assessment method based on the quantitative grading of non-quantitative environmental pollution indicators, the original detection data, direct water quality indicator grading standards, and indirect water quality indicator regression quantitative grading standards can all be standardized into dimensionless values ​​using the environmental pollution index method or mathematical methods. Furthermore, this environmental pollution index method can be a single-factor standardized index method or the Nemerow method.

[0060] In a preferred embodiment of the water quality assessment method based on the quantitative grading of non-quantitative environmental pollution indicators described above, the selection of a preset number of direct water quality indicators related to indirect water quality indicators may include:

[0061] No fewer than three direct water quality indicators should be selected, and they should be selected continuously from largest to smallest according to their importance ratio, with a cumulative importance ratio of no less than 90%.

[0062] Specifically, the water quality environmental functional zone category can be determined by the PCA comprehensive score ranking order of different periods, cross sections, and grading standards. The water quality change trend of different periods within the same cross section can be qualitatively compared by the PCA comprehensive score ranking order of the same cross section and the time series linear regression. Typical water quality indicators of the cross section can be analyzed based on the PCA load factor.

[0063] In summary, the above method establishes a comprehensive evaluation system for surface water quality that can comprehensively consider direct water quality indicators (quantitative indicators of environmental pollution) and indirect water quality indicators (non-quantitative indicators of environmental pollution), combine single-factor environmental overload assessment and comprehensive factor environmental condition assessment, and quantitatively analyze water quality change trends, thereby making the evaluation results more comprehensive and objective.

[0064] An example implementation of a water quality assessment device based on the quantitative grading of non-quantitative environmental pollution indicators provided by this invention. Figure 2 As shown, Figure 2 This is a schematic diagram of an embodiment of a water quality assessment device based on the quantitative grading of non-quantitative indicators of environmental pollution provided by the present invention. The device may include:

[0065] The screening and classification unit 201 is used to screen surface water quality indicators and classify them into direct and indirect water quality indicators. It should be noted that the aforementioned direct water quality indicators refer to quantitative environmental pollution indicators, while indirect indicators refer to non-quantitative environmental pollution indicators. Specifically, the set of water quality indicators M can be screened and classified based on historical surface water pollution records and drainage types. If no characteristic pollutants are found, conventional water quality testing indicators from GB3838-2002 "Surface Water Environmental Quality Standard" are selected, along with river flow velocity, lake / reservoir flow rate Q, chlorophyll a, air temperature T, transparency SD, suspended solids SS, turbidity NUL, color, oxidation-reduction potential ORP, and conductivity.

[0066] The raw detection data acquisition unit 202 is used to determine the evaluation section and evaluation period, and to acquire raw detection data of water quality concentration at different times and different sections within the evaluation period. It should be noted that surface water does not have an ice-covered period. The judgment criteria are: when T < 0℃, TW ≥ 0℃. The detection period is selected from January to December each year, and 1 to 3 tests are conducted per month. Here, T is the air temperature of the area where the water body is located, in ℃; TW is the water temperature of the water body, in ℃.

[0067] The direct water quality indicator screening unit 203 is used to process the maximum values ​​in indirect water quality indicators and screen a preset number of direct water quality indicators related to the indirect water quality indicators. Specifically, it can calculate the coefficient of variation (CV) value of the indirect water quality indicators. If the CV value is more than three times the maximum CV value of the direct water quality indicators, the outlier is taken as the average of the two months before and after. Utilizing the feature importance function of the machine learning regression algorithm from indirect water quality indicators to direct water quality indicators, it screens the most important direct water quality indicators related to the indirect water quality indicators. At least three direct water quality indicators are screened, and their cumulative importance is not less than 90%. Of course, the above processing method can be adaptively adjusted according to actual needs; this is not limited here.

[0068] The PLSR regression unit 204 is used to perform PLSR regression by comparing the original detection data of the concentrations of a single indirect water quality indicator with the preset number of direct water quality indicators after screening. This determines the environmental standards for evaluating the environmental functional zones of surface water bodies and the environmental standards to be implemented for direct indicators. The PLSR regression formula is used to calculate the regression quantitative grading standards for indirect water quality indicators one by one. It should be noted that there is no effective grading of non-quantitative environmental pollution indicators in the existing national, local, and industry standards. However, if participating in comprehensive water quality assessment, reasonable and effective grading standards are required. Therefore, this step can be used to dynamically screen non-quantitative environmental pollution indicators and calculate them one by one, that is, to quantify the grading standards of these indicators.

[0069] Standardization unit 205 is used to standardize the original test data, the national classification standards for direct water quality indicators, and the regression quantitative classification standards for indirect water quality indicators into dimensionless values. Specifically, the environmental pollution index method or mathematical methods can be used to standardize the original test data, the national classification standards for direct water quality indicators, and the regression quantitative classification standards for indirect indicators into dimensionless values. The mathematical methods include, but are not limited to, maximum value, average value, weighted average method, and coefficient of variation method, etc.

[0070] Unit 206, which is used to determine the category of water quality environmental functional zone, is used to perform PCA analysis on the standardized direct water quality indicators and national classification standards, indirect water quality indicators and indirect water quality indicator regression quantification classification standards according to the evaluation period and evaluation section. The water quality environmental functional zone category is determined by the ranking order of the comprehensive PCA scores of different periods, different sections and classification standards. It should be noted that the above PCA analysis is also principal component analysis.

[0071] Typical water quality index analysis unit 207 is used to analyze typical water quality indices of the section based on the PCA load coefficient obtained from PCA analysis.

[0072] In summary, the water quality assessment device based on the quantitative grading of non-quantitative environmental pollution indicators described above can comprehensively consider the relationship between quantitative and non-quantitative environmental pollution indicators and the contribution of indirect indicators, thereby achieving the quantitative grading of non-quantitative environmental pollution indicators. By combining single-factor environmental overload assessment and comprehensive factor environmental condition assessment, the assessment results are more comprehensive and objective.

[0073] In a specific embodiment of the water quality assessment device based on the quantitative grading of non-quantitative environmental pollution indicators, the direct water quality indicator screening unit can be used to screen a preset number of direct water quality indicators related to indirect water quality indicators by utilizing the feature importance function in the decision tree regression algorithm for indirect and direct water quality indicators. Specifically, it uses decision tree machine learning regression to screen the feature importance of indirect water quality indicators with multiple other direct water quality indicators. Then, it performs partial least squares (PLSR) regression on the original detection data of concentrations of a single indirect water quality indicator and multiple screened direct water quality indicators to determine the environmental functional zones of surface water bodies and the environmental standards implemented for direct indicators. Finally, it uses the PLSR regression formula to calculate the comprehensive evaluation quantitative grading standard for each indirect water quality indicator.

[0074] In another specific embodiment of the water quality assessment device based on the quantitative grading of non-quantitative environmental pollution indicators, the direct water quality indicator screening unit can also be used to calculate the coefficient of variation (CV) of indirect water quality indicators. When an indirect water quality indicator is more than three times the maximum CV of the direct water quality indicator, it is determined to be an outlier. During PLSR regression, the outlier is changed to the average value within a preset time range before and after it. In this embodiment, by processing the maximum value of the indirect indicator in the above manner, the error in subsequent PLSR regression can be reduced. Furthermore, the preset time can preferably be two periods before and after. It should be noted that when monthly data is selected, if an outlier occurs, the average value of the two months before and after is replaced with the outlier. If annual or daily data is selected, the average value of the two years before and after and the two days before and after are replaced with the outlier.

[0075] The following is a specific example to illustrate in detail the water quality assessment method and apparatus based on the quantitative grading of non-quantitative indicators of environmental pollution:

[0076] Five hundred and fifty-two water quality data points from January to December 2013 were collected from two monitoring sections of a lake / reservoir. Twenty-three water quality indicators were analyzed, of which 15 were detected. The tested indicators were: flow rate (Q), water temperature (TW), air temperature (T), pH, dissolved oxygen (DO), permanganate index (CODmn), chemical oxygen demand (CODcr), five-day biochemical oxygen demand (BOD5), ammonia nitrogen (NH3-N), total phosphorus (TP), total nitrogen (TN), copper (Cu), zinc (Zn), and fluoride (F). - Selenium (Se), total arsenic (As), total mercury (Hg), total cadmium (Cd), hexavalent chromium (Cr) 6+ Total lead (Pb), total cyanide (CN) - volatile phenols, fecal coliforms and chlorophyll a Chla, transparency SD.

[0077] (1) Surface water without characteristic industrial wastewater flows in, select M = {Q, TW, T, pH, DO, high CODmn, CODcr, BOD5, NH3-N, TP, TN, F} - , As, Se, Cu, Zn, Hg, Cd, Cr 6+ Pb, CN - The list of detected substances included volatile phenols, fecal coliforms, Chla, and SD, with 15 items detected, including Cu, Zn, Hg, Cd, and Cr. 6+ Pb, CN - Water quality indicators that do not show detectable volatile phenols are not included in the evaluation; it is directly determined that there is no pollution of this type of indicator.

[0078] M was initially selected 间接指标 ={Q, pH, SD, Chla, TW}; M 直接指标 ={DO, F -The data included fecal coliform bacteria, TN, As, TP, CODcr, NH3-N, BOD5, and CODmn. Simultaneously, direct indicator grading standards were collected, selecting standards I, II, III, IV, and V from the "Surface Water Environmental Quality Standard (GB3838-2002)" and standard V1 from the extended category of the surface water environmental quality standard in DB11 / T1722-2020 "Technical Specification for Water Ecological Health Assessment"; the grading standard set B = {I, II, III, IV, V, V1}.

[0079] (2) Based on the air temperature T and water temperature TW, it can be seen that there is no ice-covered period for surface water. The judgment standard is: when T < 0℃, TW ≥ 0℃. Therefore, select water quality test data from January to December of a natural year; select 2 test sections.

[0080] (3) The maximum value of Chla is 0.42, and CV = 4.3, which is more than 3 times the highest value of the direct indicator CV = 0.74. The value of 0.42 is processed into the average value of the two months before and after, which is 0.004.

[0081] We use a decision tree machine learning regression method to identify the most important direct indicators for each indirect indicator.

[0082] Q and M 直接指标 Decision tree regression was performed on {DO, TN, CODmn, TP, NH3-N, CODcr, BOD5}. The importance of the first three direct indicators and features was: DO 41%, TN 27%, NH3-N 26.4%, with a cumulative importance of 94.4%.

[0083] TW and M 直接指标 Make decisions based on {DO, TN, CODmn, TP, NH3-N, CODcr, BOD5}

[0084] For tree regression, the importance of the top four direct indicators and features is as follows: DO 48.2%, CODmn 18.3%, TP 14.2%, NH3-N 13.5%, with a cumulative importance of 94.2%.

[0085] Chla and M 直接指标 Make decisions based on {DO, TN, CODmn, TP, NH3-N, CODcr, BOD5}

[0086] For tree regression, the importance of the top 5 direct indicators and features is: DO 69.7%, CODcr 22.8%, BOD 55.8%, TN 0.9%, TP 0.8%, with a cumulative importance of 100%.

[0087] pH and M 直接指标 Make decisions based on {DO, TN, CODmn, TP, NH3-N, CODcr, BOD5}

[0088] For tree regression, the importance of the top four direct indicators and features is as follows: CODcr 46.1%, CODmn 37.3%, TP 11.1%, DO 3.4%, with a cumulative importance of 97.9%.

[0089] SD and M 直接指标 Make decisions based on {DO, TN, CODmn, TP, NH3-N, CODcr, BOD5}

[0090] For tree regression, the importance of the top four direct indicators and features is: TN 42.4%, CODcr 38.5%, CODmn 15.4%, TP 2.5%, with a cumulative importance of 98.8%.

[0091] (4) The PLSR partial least squares method is used to calculate the grading standard of indirect indicators, and the regression relationship between indirect indicators and feature importance indicators is finally determined based on the fitting error. The grading standard of indirect indicators is calculated based on the corrected error logic values.

[0092] B Chla-PLSR =0.005-0.001×M DO +0.004×M TN +0.003×M TP +0.001×M BOD5 .

[0093] B Chla-PLSR =0.004 + 0.0 × M CODcr +0.001×M BOD5 -0.001×M DO +0.003×M TN -0.012×

[0094] M TP The maximum fitting error was 0.00527 mg / L. When regressing Chla with DO, CODcr, BOD5, TN, and TP, the coefficient of CODcr was 0. Therefore, the regression of Chla with DO, BOD5, TN, and TP was repeated.

[0095] B Chla ={Ⅰ,Ⅱ,Ⅲ,Ⅳ,Ⅴ,V1}={0.00133, 0.0055, 0.0083, 0.01445,

[0096] 0.0216, 0.03425

[0097] B pH-PLSR =8.696 - 0.094 × M CODmn +0.039×M TN -0.054×M DO +0.029×M CODcr;

[0098] B pH ={Ⅰ,Ⅱ,Ⅲ,Ⅳ,Ⅴ,V1}={8.5458, 8.4505, 8.481, 8.5225, 8.416,

[0099] 8.38} Although the maximum fitting error of pH calculation is only 6%, it does not conform to the logic that the higher the classification standard, the larger the value. Therefore, it is not adopted. In the subsequent standardization of pH classification standards, the average value after standardization by CODmn, CODcr, DO, and TN classification standards will be used.

[0100] B SD-PLSR =0.336 + 0.017 × M CODmn -0.052×M TN -0.004×M CODcr +0.017×M DO .maximum

[0101] The fitting error is 14.23%. (B) SD ={Ⅰ, Ⅱ, Ⅲ, Ⅳ, Ⅴ, V1} ={0.4271, 0.42, 0.391, 0.359, 0.361, 0.2415}. Since logically, the SD value should decrease as the standard level increases, and levels IV and V do not meet this requirement, the average of the two values ​​can be taken as the classification standard for both. B TRA-Ⅳ =B TRA-Ⅴ =0.36. B SD ={Ⅰ,Ⅱ,Ⅲ,Ⅳ,Ⅴ,V1}={0.4271, 0.42, 0.391, 0.36, 0.36, 0.2415},

[0102] B TW-PLSR =41.601 - 3.569 × M DO +9.902×M NH4 +0.528×M TP -0.074×M BOD5 -0.902×

[0103] M CODmn The maximum error was 10.32℃. The large fitting error of PLSR indicates that this indirect indicator is not necessarily related to the direct indicators related to environmental pollution. Therefore, TW was not selected as an indirect indicator for PCA evaluation.

[0104] B TW ={Ⅰ,Ⅱ,Ⅲ,Ⅳ,Ⅴ,V1}={14.298, 21.572, 28.0028, 36.3622,

[0105] The calculated TW classification standard (40.1026, 67.1095) does not conform to the basic logic of water temperature.

[0106] B Q-PLSR =14018.356 + 425.643 × M DO +1570.17×M TN -6202.286×M NH4 Maximum Fit

[0107] Error 84%,

[0108] B Q ={Ⅰ,Ⅱ,Ⅲ,Ⅳ,Ⅴ,V1}={16594, 14256, 11514, 8347, 5605,

[0109] -8503}, V1=-8503, negative numbers are unreasonable, take V1=V=5605, B Q ={Ⅰ,Ⅱ,Ⅲ,Ⅳ,Ⅴ,V1}={16594, 14256, 11514, 8347, 5605, 5605}.

[0110] M was finally selected. 间接指标 = {Q, pH, SD, Chla}, where Q, SD, and Chla have been calculated with reasonable grading standards according to PLSR regression. pH does not have a reasonable PLSR standard. When standardizing the pH grading standard, the average value of the standardized grading standards of CODmn, CODcr, DO, and TN will be directly taken (CODmn, CODcr, DO, and TN are direct indicators of feature importance when regressing pH decision trees).

[0111] (5) Direct indicators and grading standards (the grading standards for direct indicators are derived from national standards GB3838-2002 and DB11 / T1722-2020), and the finally selected indirect indicators and PRSL grading standards are standardized using the single-factor standardized index method.

[0112] Since there is no reasonable PLSR standard for pH, the average value of the standardized values ​​of CODmn, CODcr, DO, and TN will be directly used when standardizing the pH classification standard.

[0113] DO standard index formula: ratio of environmental functional zone standard to graded standard; other water quality indicator standard index formula: ratio of graded standard to environmental functional zone standard.

[0114] (6) The standardized direct water quality indicators and national classification standards, indirect water quality indicators and calculated classification standards are subjected to principal component analysis (PCA) according to the evaluation period and cross section. The water quality environmental functional zone category is determined by the ranking of the PCA comprehensive scores of different periods, cross sections and classification standards.

[0115] Perform PCA algorithm analysis on the values ​​in step (5):

[0116] Variable: {S Q S pH S DO S SD S F- S chla S 粪大肠杆菌 S TN S As S TP S CODcr S NH3-N S BOD5 S CODmn}; Index item: {Water quality monitoring section, TIME}

[0117] First, perform the KMO and Bartlett's tests to determine if principal component analysis (PCA) is suitable. For KMO values: above 0.8 is very suitable for PCA, between 0.7 and 0.8 is generally suitable, between 0.6 and 0.7 is not very suitable, between 0.5 and 0.6 indicates poor performance, and below 0.5 indicates extremely unsuitable performance. For Bartlett's test, if the p-value is less than 0.05, reject the null hypothesis, indicating that PCA can be performed. If the null hypothesis is not rejected, it means that these variables may independently provide some information, and PCA is not suitable.

[0118] In this example, the KMO value is 0.833, which is very suitable for PCA analysis. Bartlett's test shows that P < 0.05, indicating significance, so principal component analysis can be performed.

[0119] By analyzing the variance explanation table and such Figure 3 The diagram of the gravel shown, Figure 3A scree plot is used to determine the number of principal components and explain the variance of the data. The variance explanation table primarily examines the contribution rate of each principal component to the explained variables. The scree plot is used to determine the number of principal components to be selected based on the slope of the eigenvalue descent. Combining these two methods can be used to confirm or adjust the number of principal components. Each principal component is represented by a point. The number of principal components is extracted based on the unknown factor of "the slope becoming gentler." Through variance explanation and scree plot analysis, it is found that when the eigenvalue is 0.761, the scree plot tends to be gentler, and the cumulative variance explained rate is 85.736%. Therefore, 3 principal components are determined. Finally, as shown in Table 1, which is a component matrix table, the PCA comprehensive score of each section for each period in the component matrix table is compared with the grading standards I, II, III, IV, V, and V1. It is concluded that the comprehensive evaluation of all detection times and detection sections meets the requirements of the Class V standard for environmental functional zones. Specifically, Section 1 in August, January, February, April, June, May, July, September, March, November, October, and December 2013, and Section 2 in June 2013 are between the standards of IV and V. Section 2 in January, April, and March 2013 are between the standards of III and IV. Section 2 in May, August, September, February, October, July, November, and December 2013 are between the standards of I and II.

[0120] Table 1. Component Matrix

[0121]

[0122]

[0123] (7) Analyze typical water quality indicators of the section based on the PCA load factor.

[0124] refer to Figure 4 , Figure 4 The factor loading matrix heatmap, after factor loading coefficient analysis, shows that principal component 2 consists of environmental overload factors, namely CODcr, pH, and F. - The environmental impact is significant; the higher the coefficient, the greater the environmental impact. The system is already overloaded or nearing overload, primarily characterized by mixed inorganic, organic, and acid-base pollution. Principal component 1 represents a factor with low environmental carrying capacity and has a relatively small environmental impact.

[0125] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A water quality assessment method based on quantitative grading of non-quantitative environmental pollution indicators, characterized in that, include: Surface water quality indicators are screened and classified into direct water quality indicators and indirect water quality indicators. Direct water quality indicators refer to indicators with clear national classification standards and which directly evaluate the pollutant status of surface water environment quality using water quality assessment methods. Indirect water quality indicators refer to indicators without clear national classification standards but which are closely related to the surface water environment quality status. Determine the evaluation section and evaluation period, and obtain raw water quality concentration detection data for different times and different sections within the evaluation period; Process the maximum values ​​in the indirect water quality indicators, and filter a preset number of direct water quality indicators that are related to the indirect water quality indicators; By using the original detection data of concentrations of a single indirect water quality indicator and a pre-selected number of direct water quality indicators, the PLSR regression is performed to determine the environmental standards for evaluating the environmental functional zones of surface water bodies and the direct water quality indicators. The regression quantitative grading standards for indirect water quality indicators are calculated one by one using the PLSR regression formula. The original test data, the national grading standards for direct water quality indicators, and the regression quantitative grading standards for indirect water quality indicators were all standardized into dimensionless values. The standardized direct water quality indicators and the national grading standards for direct water quality indicators, as well as the indirect water quality indicators and the regression quantification grading standards for indirect water quality indicators, are subjected to PCA analysis according to the evaluation period and evaluation section. The water quality environmental functional zone category is determined by the ranking order of the comprehensive PCA scores of different periods, different sections, and grading standards. Typical water quality indicators of the section are analyzed based on the PCA load coefficient obtained from PCA analysis.

2. The water quality assessment method based on quantitative grading of non-quantitative environmental pollution indicators according to claim 1, characterized in that, By utilizing the feature importance function in the decision tree regression algorithm for indirect and direct water quality indicators, a predetermined number of direct water quality indicators related to indirect water quality indicators are selected.

3. The water quality assessment method based on quantitative grading of non-quantitative environmental pollution indicators according to claim 1, characterized in that, The maximum value of the indirect water quality index is: Calculate the coefficient of variation (CV) of indirect water quality indicators. When an indirect water quality indicator is more than three times the maximum CV of a direct water quality indicator, it is determined to be an outlier. When performing PLSR regression, the outlier is changed to the average value within a preset time range before and after it.

4. The water quality assessment method based on quantitative grading of non-quantitative environmental pollution indicators according to claim 1, characterized in that, The original detection data, direct water quality index grading standards, and indirect water quality index regression quantification grading standards are all standardized into dimensionless values ​​using the environmental pollution index method or mathematical methods.

5. The water quality assessment method based on quantitative grading of non-quantitative environmental pollution indicators according to claim 4, characterized in that, The environmental pollution index method is either the single-factor standardized index method or the Nemerow method.

6. The water quality assessment method based on quantitative grading of non-quantitative environmental pollution indicators according to claim 1, characterized in that, The preset number of direct water quality indicators related to indirect water quality indicators include: The selected direct water quality indicators shall be no less than three, and shall be selected in descending order of importance based on their characteristics, with a cumulative importance ratio of no less than 90%.

7. The water quality assessment method based on quantitative grading of non-quantitative environmental pollution indicators according to claim 3, characterized in that, The preset time consists of two cycles, one before and one after.

8. A water quality assessment device based on the quantitative grading of non-quantitative environmental pollution indicators, characterized in that, include: The screening and classification unit is used to screen surface water quality indicators and classify them into direct water quality indicators and indirect water quality indicators. The direct water quality indicators refer to indicators that have clear national classification standards and are used to directly evaluate the pollutant status of surface water environmental quality using water quality assessment methods. The indirect water quality indicators refer to indicators that do not have clear national classification standards but are closely related to the surface water environmental quality status. The raw detection data acquisition unit is used to determine the evaluation section and evaluation period, and to acquire raw detection data of water quality concentration at different times and different sections within the evaluation period. A direct water quality index screening unit is used to process the maximum value in the indirect water quality index and screen a preset number of direct water quality indicators related to the indirect water quality index. The PLSR regression unit is used to perform PLSR regression by comparing the original detection data of the concentration of a single indirect water quality indicator with the preset number of direct water quality indicators after screening. This determines the environmental standards for evaluating the environmental functional zones of surface water bodies and the environmental standards implemented for direct water quality indicators. The PLSR regression formula is used to calculate the regression quantitative grading standards for indirect water quality indicators one by one. The standardized unit is used to standardize the original test data, the national grading standards for direct water quality indicators, and the regression quantitative grading standards for indirect water quality indicators into dimensionless values. The water quality environmental functional zone category determination unit is used to perform PCA analysis on the standardized direct water quality indicators and the national classification standards for direct water quality indicators, the indirect water quality indicators and the regression quantitative classification standards for indirect water quality indicators according to the evaluation period and evaluation section, and determine the water quality environmental functional zone category achieved by ranking the PCA comprehensive scores of different periods, different sections and classification standards. Typical water quality index analysis unit, used to analyze typical water quality indexes of the section based on the PCA load coefficient obtained from PCA analysis.

9. The water quality assessment device based on quantitative grading of non-quantitative environmental pollution indicators according to claim 8, characterized in that, The direct water quality index screening unit is specifically used to utilize the feature importance function in the decision tree regression algorithm for indirect and direct water quality indicators to screen a preset number of direct water quality indicators related to indirect water quality indicators.

10. The water quality assessment device based on quantitative grading of non-quantitative environmental pollution indicators according to claim 8, characterized in that, The direct water quality index screening unit is also specifically used to calculate the coefficient of variation (CV) value of indirect water quality indexes. When an indirect water quality index that is more than three times the maximum CV value of the direct water quality index is determined to be an outlier, the outlier value is changed to the average value within a preset time range before and after it when performing PLSR regression.

Images