Construction method and application of industrial water quality index pool

Abstract

The invention discloses an industrial water quality index pool construction method and application thereof. The method comprises the following steps: acquiring a candidate index data set for detecting industrial pollution; classifying and marking each index in the candidate index data set; through a three-factor scoring method, each index is scored and sorted; grading and marking each index according to the scores of the three-factor scoring method from large to small; constructing an industrial water quality index pool formed by the index matrix; the industrial water quality indexes can be comprehensively, systematically and accurately screened and integrated, a scientific, reasonable and targeted index system is provided for industrial water quality detection, the industrial water quality condition can be more accurately and efficiently evaluated, and the industrial pollution monitoring and treatment level is improved.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring and water quality assessment technology, specifically to a method for constructing an industrial water quality index pool and its application. Background Technology

[0002] Industrial water quality monitoring and evaluation is crucial for water environment management. Currently, this work mainly relies on national standards (such as GB 31570, HJ / T91, etc.) and industry specifications to determine monitoring indicators.

[0003] In practical applications, existing technologies face numerous challenges. For instance, the indicator system is incomplete, with limited coverage, preventing the effective inclusion of some key pollutants in the monitoring scope. Furthermore, the sheer number of monitoring indicators significantly increases testing costs and adds to management pressure. Simultaneously, some indicators, such as polycyclic aromatic hydrocarbons, trace metals, and specific biotoxic substances, require extremely stringent detection conditions, making them difficult for conventional laboratories to handle. More critically, existing indicator systems lack a systematic approach based on process mechanisms. Therefore, there is an urgent need for a method to systematically construct and scientifically screen monitoring indicators, offering a highly operable indicator pool construction method to overcome problems such as insufficient indicator coverage, excessive redundancy, difficulty in detecting some indicators, and incomplete indicator system categories in existing technologies. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides a method for constructing an industrial water quality index pool and its application.

[0005] A method for constructing an industrial water quality index pool includes the following steps: Obtain a dataset of candidate indicators for detecting industrial pollution; Based on the chemical categories of industrial pollutants, each indicator in the candidate indicator dataset is classified and labeled. Each indicator is scored using a three-factor scoring method. The three-factor scoring method is a method of obtaining a score by weighting and summing the mechanistic relevance, detection accessibility, and redundancy penalty value of each indicator. Within the same category, each indicator is graded and labeled according to its score in descending order using the three-factor scoring method. Generate an index matrix containing classification, scoring, and grading result labels, and summarize the results to obtain an industrial water quality index pool composed of the index matrix.

[0006] Explanation: The above method clarifies the classification and labeling of indicators in terms of pollution risk and category. It utilizes a three-factor scoring method to scientifically score and rank indicators, comprehensively considering mechanism relevance, detection accessibility, and redundancy penalty values. Based on these scores, hierarchical labeling is then applied, ultimately generating an indicator matrix containing multi-dimensional labels to construct an industrial water quality indicator pool. Mechanism-driven panoramic identification avoids omissions; accessibility quantification improves on-site feasibility; collaborative scoring reduces redundancy and blind screening; a category guarantee mechanism ensures the representativeness of each category; and statistical verification enhances the credibility of the solution. This allows for comprehensive, systematic, and accurate screening and integration of industrial water quality indicators, providing a scientific, reasonable, and targeted indicator system for industrial water quality testing. This helps to more accurately and efficiently assess industrial water quality and improve the level of industrial pollution monitoring and treatment.

[0007] Furthermore, the acquisition of a dataset of candidate indicators for detecting industrial pollution includes: A pollution mechanism mapping table for industrial processes is constructed. The pollution mechanism mapping table is used to record the pollutants involved and generated in each process unit of the industrial process, the process sections affected by the pollutants, and the degree of impact of the pollutants on the process sections. The pollutants and corresponding monitoring indicators for each work section in the pollution mechanism mapping table are identified as the first indicator set. Find the monitoring indicators specified in the industrial process within the existing standards as the second set of indicators; The candidate indicator dataset is obtained by taking the union of the first indicator set and the second indicator set and removing duplicates.

[0008] Explanation: The above method, by constructing a pollution mechanism mapping table for industrial processes, can accurately grasp the generation and impact of pollutants in each stage of industrial production, providing a solid theoretical basis for indicator selection; it identifies the first indicator set for targeted monitoring of key pollutants and process stages, and finds the second indicator set that can be included in existing standard-stipulated indicators to ensure compliance; by combining the two sets and removing duplicates, a candidate indicator dataset is obtained, which comprehensively covers the indicators required for industrial pollution monitoring while avoiding duplication, making the candidate indicator dataset more scientific and complete, and laying a solid foundation for the subsequent construction of a high-quality industrial water quality indicator pool.

[0009] Furthermore, the classification and labeling of each indicator in the candidate indicator dataset based on the chemical categories of industrial pollutants includes: Based on the legal enforceability and risk priority of pollution, the candidate indicator dataset is hierarchically labeled to obtain basic layer indicators included in legal monitoring, extended layer indicators included in industry control monitoring, and mechanism layer indicators that are not included in legal and industry monitoring but have high risks. The candidate index dataset is classified and labeled according to the chemical category of the pollutants. The categories include conventional physicochemical pollutants, inorganic salts, nitrogen-containing pollutants, metal pollutants, organic pollutants, and toxic pollutants.

[0010] Explanation: The above method clearly distinguishes between basic, extended, and mechanistic indicators by stratifying them according to the legal mandatory nature and risk priority of pollution. This ensures that the monitoring foundation meets regulatory requirements, takes into account industry control needs, and highlights special indicators that are high-risk but not included in routine monitoring. At the same time, classifying and labeling indicators according to the chemical categories of pollutants helps to systematically sort out the indicators from a professional perspective, which facilitates the adoption of targeted monitoring and analysis methods for different types of pollutants. This improves the accuracy and professionalism of industrial water quality monitoring and provides a comprehensive and systematic indicator basis for the scientific assessment of industrial pollution status.

[0011] Furthermore, the formula for calculating the relevance of the mechanism is as follows: (1) In equation (1), For mechanism-related degree, To find the indicators in the pollution mechanism mapping table i The complete set of paths pointing to one or more functional endpoints. L p For path length, Contribute to the path α This is the path length attenuation coefficient.

[0012] Note: The above method helps to accurately quantify the correlation between each indicator and the functional endpoint of industrial pollution.

[0013] Furthermore, the formula for calculating the reachability of the detection is as follows: (2) In equation (2), To detect accessibility, , , To detect the weighting coefficients, To assess the maturity of the testing method, For historical data coverage, To standardize testing costs.

[0014] Note: The above method of using weighting coefficients to reasonably allocate the influence of each factor on the accessibility of detection can scientifically, comprehensively and accurately evaluate the operability and convenience of each indicator in actual detection.

[0015] Furthermore, the method for calculating the redundancy penalty value is as follows: calculate the correlation between every two indicators in the candidate indicator dataset, take the absolute value of each calculated correlation, and then average the correlations of all indicators in the candidate indicator dataset to obtain the redundancy penalty value.

[0016] Note: The above method, which calculates the correlation one by one, takes the absolute value, and then averages the results, comprehensively and objectively reflects the close relationship between the indicators, effectively avoiding information duplication and resource waste caused by highly redundant indicators.

[0017] Furthermore, the calculation formula for the three-factor scoring method is as follows: (3) In equation (3), The scores are the three-factor ratings. For mechanism-related degree, To detect accessibility, This is a redundancy penalty value. , , These are the scoring weight parameters.

[0018] Note: The above quantitatively integrates the three key dimensions of mechanism relevance, detection accessibility, and redundancy penalty value, and uses weight parameters to flexibly adjust the influence of each factor on the final score, which can comprehensively and scientifically evaluate the value of each indicator in industrial water quality monitoring.

[0019] Furthermore, among all indicators in the same category, each indicator is graded and labeled according to its score in descending order using the three-factor scoring method; including: Among all indicators in the same category, the top N indicators with the highest scores using the three-factor scoring method are designated as mandatory, the Nth to Mth highest-scoring indicators are designated as recommended, and the remaining indicators are designated as backup. The values ​​of N range from 1 to 20, and the values ​​of M range from 20 to 50.

[0020] This invention also provides an application of an industrial water quality index pool, which, based on the above-mentioned industrial water quality index pool, includes: Establish an association model for mapping the process units in industrial production to the industrial water quality index pool; The input of the association model is a certain process unit, and the output is an index matrix in the industrial water quality index pool that is associated with the certain process unit, and the index matrix associated with the certain process unit is arranged according to the score.

[0021] Note: The above correlation model achieves precise docking between industrial production process units and industrial water quality indicator pools. It can quickly and accurately screen out water quality indicators closely related to specific process units from a large indicator pool, improve monitoring efficiency and targeting, help to promptly detect water quality problems in process units, and provide strong support for water quality control and pollution prevention in industrial production processes.

[0022] The beneficial effects of this invention are: This invention overcomes the problems of insufficient indicator coverage, excessive redundancy, difficulty in detecting some indicators, and incomplete indicator system categories in existing technologies. It clarifies the classification and labeling of indicators based on pollution risk and category, and uses a three-factor scoring method to scientifically score and rank indicators by comprehensively considering mechanism relevance, detection accessibility, and redundancy penalty values. Then, it performs hierarchical labeling based on the scores, ultimately generating an indicator matrix containing multi-dimensional labels to construct an industrial water quality indicator pool. Mechanism-driven panoramic identification avoids omissions; accessibility quantification improves on-site feasibility; collaborative scoring reduces redundancy and blind screening; a class protection mechanism ensures the representativeness of each category; and statistical verification improves the credibility of the solution. Therefore, it can comprehensively, systematically, and accurately screen and integrate industrial water quality indicators, providing a scientific, reasonable, and targeted indicator system for industrial water quality testing, which helps to more accurately and efficiently assess the status of industrial water quality and improve the level of industrial pollution monitoring and treatment. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention. Detailed Implementation

[0024] To further illustrate the methods and effects of this invention, the technical solution of this invention will be clearly and completely described below in conjunction with experiments.

[0025] Example: A method for constructing an industrial water quality index pool, comprising the following steps: S101. Obtain a dataset of candidate indicators for detecting industrial pollution; including ①~④: ① Construct a pollution mechanism mapping table for industrial processes, as shown in Table 1; the pollution mechanism mapping table is used to record the pollutants involved and generated in each process unit of the industrial process, the process sections affected by the pollutants, and the degree of influence of the pollutants on the process sections (i.e., the weights in Table 1). For example, the above-mentioned pollution mechanism mapping table can ensure the comprehensive identification of potential pollutants from both process and material perspectives, avoiding omissions. Required documents: process flow diagram, list of water-using units (e.g., equipment, process water supply, wastewater treatment unit, etc.), bill of materials (raw materials, additives, intermediates), historical accident / exceedance records, national / industry standards and relevant documents; Table 1. Pollution Mechanism Mapping Table

[0026] The weighting is based on the following criteria: pollutant detection indicators that fall under national mandatory standards or explicit legal requirements are assigned a value of 1.0; pollutant detection indicators that have repeatedly exceeded standards in the company's history / in monitoring records are assigned a value of 0.9; pollutant detection indicators that are found in peer-reviewed literature or authoritative reports are assigned a value of 0.8; and pollutant detection indicators that are assessed by external experts are assigned a value of 0.7. Multiple pieces of evidence can be recorded and the maximum value can be selected, or a weighted composite method can be used depending on the scenario. ② Determine the pollutants in the pollution mechanism mapping table and the corresponding monitoring indicators for each work section, as the first indicator set; ③ Identify the monitoring indicators specified in the existing standards for industrial processes as the second set of indicators; Summarize a preliminary set of indicators based on existing technologies. Existing standards include: a preliminary set of indicators I0 (the number of examples can range from dozens to hundreds, depending on the complexity of the industry) compiled from national / industry standards, technical guidelines, open literature, factory historical monitoring databases, bills of materials, and on-site event records. ④ Take the union of the first indicator set and the second indicator set, and remove duplicates to obtain the candidate indicator dataset.

[0027] Comparing with the pollution mechanism mapping table P0: If a pollutant has a path in the mapping table but does not appear in I0, it is added to the candidate set I (to complete the potential indicators based on the mechanism). Conversely, for indicators that do not have a mechanism path, they can be included in the subsequent scoring as low-priority candidates, but this does not affect the formation of the main set.

[0028] This embodiment takes a petrochemical plant as an example. The process flow, bill of materials, 5 years of historical monitoring data and industry standards are collected step by step to construct a mapping table and initially collect 120 candidate indicators. The mapping table example and evidence weight are assigned according to the aforementioned preferred values. The results are shown in Table 4 below.

[0029] S102. Based on the chemical categories of industrial pollutants, classify and label each indicator in the candidate indicator dataset; Including 1)~2): 1) Based on the legal enforceability and risk priority of pollution, the candidate indicator dataset is hierarchically labeled to obtain basic layer indicators included in legal monitoring, extended layer indicators included in industry control monitoring, and mechanism layer indicators that are not included in legal or industry monitoring but have high risks; as shown in Table 2. This scheme is not limited to the specific indicators listed in the table. In actual application, it can be dynamically adjusted according to industry characteristics or monitoring needs. The hierarchical labeling results provide a reference for subsequent operators to screen indicators. Table 2. Criteria and Examples for Layering

[0030] 2) Based on the chemical categories of the pollutants, the candidate index dataset is classified and labeled. These categories include conventional physicochemical pollutants, inorganic salts, nitrogen-containing pollutants, metal pollutants, organic pollutants, and toxic pollutants. See Table 3 for details. Table 3: Examples of Classifications

[0031] The above classification and stratification are used to implement class protection during the screening process (to prevent a category from being completely eliminated due to low scores). Specifically, two columns, stratification and category, can be added to the candidate indicator table for easier statistics and management.

[0032] It should be understood that the aforementioned mechanistic relevance is calculated based on the strength and length of the mapping path. The reachability score is determined comprehensively by the maturity of the detection method, historical data coverage, and detection cost. The redundancy score is used to measure the relevance of the indicator to similar indicators. A collaborative scoring formula is constructed based on the above three elements to obtain the comprehensive score of each indicator. Based on the comprehensive score, indicators are divided into three categories: key retention, candidate observation, and elimination for backup. To avoid the complete elimination of an entire category of indicators, a class protection rule can be set, that is, a certain number of representative indicators are retained in each category.

[0033] S103. Each indicator is scored using a three-factor scoring method. The three-factor scoring method is a method of obtaining a score by weighting and summing the mechanism relevance, detection accessibility, and redundancy penalty value of each indicator. For example, three representative indicators, pH, COD, and PAH, can be used for calculation. The formula for calculating the mechanism relevance is as follows: (1) In equation (1), For mechanism-related degree, To find the indicators in the pollution mechanism mapping table i The complete set of paths pointing to one or more functional endpoints. L p For path length, Contribute to the path α This is the path length attenuation coefficient; The specific calculation principle is as follows: find the indicators in the pollution mechanism mapping table. i The set of all paths pointing to one or more functional endpoints For path Its path weight product is the product of the evidence weights of each edge on the path. ; path length is (Number of edges). Define path contribution as... Path length attenuation coefficient α Option 1 is preferred.

[0034] The formula for calculating the reachability of the detection is as follows: (2) In equation (2), To detect accessibility, , , To detect the weighting coefficients, To assess the maturity of the testing method, For historical data coverage, To standardize testing costs; Preferred value: = 0.5、 = 0.3、 = 0.2 Maturity of the detection method (National / Industry Standards = 1.0, Industry Methods = 0.8, Research Methods = 0.5), Historical Data Coverage (0~1) Standardized testing costs (0~1, Smaller is better); The redundancy penalty value is calculated as follows: The correlation between any two indicators in the candidate indicator dataset is calculated, and the absolute value of each correlation is taken. Then, the average correlation of all indicators in the candidate indicator dataset is calculated to obtain the redundancy penalty value; as shown in the formula: ; The calculation formula for the three-factor scoring method is as follows: (3) In equation (3), The scores are the three-factor ratings. For mechanism-related degree, To detect accessibility, This is a redundancy penalty value. , , For the scoring weight parameters, (preferred) = 0.5, = 0.35, = 0.15).

[0035] In some embodiments, when A value ≥0.75 is preferred. When 0.5≤ When <0.75, it is reserved as a critical value; when If the value is less than 0.5, discard it or keep it as a backup. In other embodiments, the threshold can be adjusted as needed for engineering purposes; S104. Among all indicators in the same category, each indicator is graded and labeled according to the three-factor scoring method from largest to smallest. Among all indicators in the same category, the top N indicators with the highest scores using the three-factor scoring method are designated as mandatory, and the Nth to Mth indicators are designated as recommended. In some embodiments, N ranges from 1 to 20, and M ranges from 20 to 50. If all scores in a category are below the retention threshold, the top N items in that category will still be retained in descending order of scores.

[0036] S105. Generate an index matrix containing hierarchical, classification, scoring, and graded result labels, and summarize it to obtain an industrial water quality index pool composed of the index matrix. For example, a data platform might use a database or spreadsheet to store mapping tables and candidate metric tables, and implement path search (breadth-first or depth-first) in the software to list... ; Generate each metric , , , Export the results as an auditable filter reasoning table; The final indicator matrix is ​​generated based on Si and the class protection rules. The structure of the indicator matrix is ​​shown in Table 4 below: The rows in Table 4 represent indicator items, the columns represent levels and categories, and the right side is the judgment column and the screening reason. Table 4. Indicator Matrix Structure

[0037] After scoring and applying class-based protection rules, the following indicator matrix is ​​obtained: 40 core items, 50 recommended items, and 30 backup items. Some examples are shown in Table 4. The system also exports a report explaining the selection rationale for each item (including...). , , (and sources of evidence), for reference in supervision, review and project implementation.

[0038] An application of an industrial water quality indicator pool includes: Establish an association model for mapping the process units in industrial production to the industrial water quality index pool; The correlation model takes a specific process unit as input and outputs a matrix of indicators associated with that process unit in the industrial water quality indicator pool. These indicators are arranged according to their scores. The correlation model can utilize existing computer servers to index the matrix against a process unit correspondence table (fast matching mechanism). After monitoring personnel input the process unit / device type, the system returns the corresponding priority monitoring indicator combination and the reasons for selection.

[0039] S106, Statistical Validation; When continuous historical monitoring data is available, the mechanism-weighted sparse regression method (such as LASSO) is preferred. Other equivalent multivariate selection methods can also be used to verify the predictive ability of candidate indicators or their correlation with key targets (such as overall pollution load and emergency event indicators). Weight design example: Setting the LASSO penalty coefficient to... , In the formula , To adjust the parameters, For mechanistic relevance, how should we configure indicators that have strong mechanistic evidence be more easily retained in regression? Validation results: Statistical consistency (e.g., the overlap rate between core indicators and regression-selected features) serves as a measure of the robustness of the scheme; if the consistency is low, the weights should be adjusted retrospectively or data sources should be supplemented to improve the mapping table or detect accessibility assessment.

[0040] Based on five years of monitoring data from the plant, a mechanism-weighted LASSO regression was used for verification. The results showed that the core indicators and the significant features selected by LASSO had an overlap rate of ≥ 90%, proving that the collaborative scoring has good consistency and feasibility in this scenario.

[0041] The above formulas, weights, and thresholds are preferred embodiments of the present invention. Specific weights and thresholds can be adjusted according to industry characteristics, regulatory requirements, or engineering needs. The above embodiments are only used to illustrate the implementation of the present invention and are not intended to limit the scope of protection of the present invention. The scope of protection of the present invention should be determined by the claims.

Claims

1. A method for constructing an industrial water quality index pool, characterized in that, Includes the following steps: Obtain a dataset of candidate indicators for detecting industrial pollution; Based on the chemical categories of industrial pollutants, each indicator in the candidate indicator dataset is classified and labeled. Each indicator after classification and labeling is scored using a three-factor scoring method. The three-factor scoring method is a method of obtaining a score by weighting and summing the mechanism relevance, detection accessibility and redundancy penalty value of each indicator. Within the same category, each indicator is graded and labeled according to its score in descending order using the three-factor scoring method. Generate an index matrix containing classification, scoring, and grading result labels, and summarize the results to obtain an industrial water quality index pool composed of the index matrix.

2. The method for constructing an industrial water quality index pool as described in claim 1, characterized in that, The acquisition of a candidate indicator dataset for detecting industrial pollution includes: A pollution mechanism mapping table for industrial processes is constructed. The pollution mechanism mapping table is used to record the pollutants involved and generated in each process unit of the industrial process, the process sections affected by the pollutants, and the degree of impact of the pollutants on the process sections. The pollutants and corresponding monitoring indicators for each work section in the pollution mechanism mapping table are identified as the first indicator set. Find the monitoring indicators specified in the industrial process within the existing standards as the second set of indicators; The candidate indicator dataset is obtained by taking the union of the first indicator set and the second indicator set and removing duplicates.

3. The method for constructing an industrial water quality index pool as described in claim 1, characterized in that, The classification and labeling of each indicator in the candidate indicator dataset based on the chemical categories of industrial pollutants includes: The candidate index dataset is classified and labeled according to the chemical category of the pollutants. The categories include conventional physicochemical pollutants, inorganic salts, nitrogen-containing pollutants, metal pollutants, organic pollutants, and toxic pollutants.

4. The method for constructing an industrial water quality index pool as described in claim 2, characterized in that, The formula for calculating the mechanism relevance is as follows: （1） In equation (1), For mechanism-related degree, To find the indicators in the pollution mechanism mapping table i The complete set of paths pointing to one or more functional endpoints. L p For path length, Contribute to the path α This is the path length attenuation coefficient.

5. The method for constructing an industrial water quality index pool as described in claim 2, characterized in that, The formula for calculating the reachability of the detection is as follows: （2） In equation (2), To detect accessibility, , , To detect the weighting coefficients, To assess the maturity of the testing method, For historical data coverage, To standardize testing costs.

6. The method for constructing an industrial water quality index pool as described in claim 2, characterized in that, The method for calculating the redundancy penalty value is as follows: calculate the correlation between every two indicators in the candidate indicator dataset, take the absolute value of each calculated correlation, and then average the correlations of all indicators in the candidate indicator dataset to obtain the redundancy penalty value.

7. The method for constructing an industrial water quality index pool as described in claim 1, characterized in that, The calculation formula for the three-factor scoring method is as follows: （3） In equation (3), The scores for the three factors are as follows: For mechanism-related degree, To detect accessibility, This is a redundancy penalty value. , , These are the scoring weight parameters.

8. The method for constructing an industrial water quality index pool as described in claim 1, characterized in that, Within the same category, each indicator is graded and labeled according to its score in descending order using the three-factor scoring method; including: Among all indicators in the same category, the top N indicators with the highest scores using the three-factor scoring method are designated as mandatory, the Nth to Mth highest scores are designated as recommended, and the remaining indicators are designated as backup. The value of N ranges from 1 to 20, and the value of M ranges from 20 to 50.

9. An application of an industrial water quality index pool, based on the industrial water quality index pool according to any one of claims 1 to 8, characterized in that, include: Establish an association model for mapping the process units in industrial production to the industrial water quality index pool; The input of the association model is a certain process unit, and the output is an index matrix in the industrial water quality index pool that is associated with the certain process unit, and the index matrix associated with the certain process unit is arranged according to the score.

Images