Method and system for screening mine soil pollution source treatment technology based on large model

Abstract

The present application proposes a large model-based mining area soil pollution source treatment technology screening method and system, relating to the field of mining area soil pollution treatment technology. In view of the low screening efficiency and poor adaptability of existing mining area soil pollution treatment technology, the present application integrates mining area multi-source data and pollution treatment case information to construct a treatment technology directory; based on search enhancement generation technology and generative large models, the present application takes the treatment technology in the treatment technology directory as the core to construct a treatment field knowledge base; a treatment technology evaluation index system is established, combined with the analytic hierarchy process and the grey comprehensive evaluation method, to construct a coupling evaluation model; the pollution characteristics of the target mining area are taken as the basis for searching, and the candidate treatment technologies are matched in the treatment field knowledge base; based on the pollution characteristics and natural geographical conditions of the target mining area, the index weight and correlation coefficient of the coupling evaluation model are dynamically adjusted, the comprehensive score of the candidate treatment technologies is calculated, and the optimal treatment technology is screened. The present application realizes accurate and efficient screening of treatment technologies.

Description

Technical Field

[0001] This invention relates to the field of soil pollution remediation technology in mining areas, and in particular to a method and system for screening remediation technologies for soil pollution sources in mining areas based on a large model. Background Technology

[0002] Soil pollution in my country's mining areas is a prominent problem. Pollution sources such as mining and metallurgical waste, tailings ponds, and riverbed sediments, which are prevalent in these areas, continuously threaten the surrounding ecological environment and arable land safety. Currently, there are numerous technologies available for the remediation of soil pollution sources in mining areas, but the applicable scenarios, remediation effects, and economic costs vary significantly among different technologies.

[0003] In existing technologies, the selection of remediation technologies mainly relies on human experience and judgment. Some solutions use a single evaluation method for technology selection, but they do not integrate multi-source remediation technology data or combine the intelligent analysis capabilities of large models, making it difficult to cope with the diverse pollution scenarios in complex mining areas. At the same time, traditional methods lack a systematic technology catalog and a scientific evaluation system, resulting in low selection efficiency, poor technology adaptability, long decision-making cycles, and unsatisfactory remediation effects.

[0004] Therefore, developing a screening method and system for the treatment of soil pollution sources in mining areas based on a large model is of great significance for improving the efficiency and accuracy of soil pollution treatment in mining areas. Summary of the Invention

[0005] To address the problems of low efficiency and poor scenario adaptability in screening existing technologies for soil pollution remediation in mining areas, this invention provides a method and system for screening remediation technologies for soil pollution sources in mining areas based on a large model, specifically including the following steps: S1. Integrate multi-source data, pollution control case information, and technical specifications of mining areas to build a pollution control technology catalog. The pollution control technology catalog covers the pollution control technologies corresponding to three types of soil pollution sources in mining areas: waste residue, tailings ponds, and bottom sediment. S2. Based on retrieval-enhanced generation technology and generative large model, with the governance technologies in the governance technology directory as the core, and integrating reference materials in the field of pollution control, a knowledge base for the governance field is constructed. S3. Establish a governance technology evaluation index system based on the governance technologies in the governance technology catalog, the governance field knowledge base, and the pollution control needs. The index system includes a target layer, a criterion layer, and an indicator layer. The target layer includes a comprehensive evaluation index for feasible pollution control technologies. The criterion layer includes technical indicators, environmental indicators, and economic indicators. The indicator layer includes multiple detailed evaluation indicators. S4. Based on the governance technology evaluation index system, and combining the analytic hierarchy process (AHP) and grey comprehensive evaluation method, a coupled evaluation model is constructed for comprehensive evaluation. Specifically, this includes: comparing the importance of each level of elements in the index system pairwise and assigning quantitative values; constructing a judgment matrix to calculate the weights of the AHP indicators; constructing an evaluation index scoring table based on technical specifications and typical cases in the governance domain knowledge base, determining the optimal reference sequence, and then calculating the grey comprehensive evaluation correlation coefficient; obtaining the comprehensive technology score based on the AHP indicator weights and the grey comprehensive evaluation correlation coefficient, and classifying the comprehensive technology score. S5. Using the pollution characteristics of the target mining area as the retrieval basis, candidate remediation technologies are matched in the remediation knowledge base. Based on the pollution characteristics and natural geographical conditions of the target mining area, the index weights and correlation coefficients of the coupled evaluation model are dynamically adjusted to calculate the comprehensive score of the candidate remediation technologies and select the optimal remediation technology.

[0006] Furthermore, in S3, several detailed evaluation indicators include: technical indicators corresponding to technology maturity, operability, public acceptance, applicable soil permeability, pollutant removal rate, stability of applicable geological conditions, various pollutants, impact on surrounding farmland, and remediation time; environmental indicators corresponding to secondary pollution risk; and economic indicators corresponding to unit treatment cost, annual operation and maintenance cost, and economic benefits of resource recovery.

[0007] Furthermore, in S4, the calculation of the weights of the hierarchical analysis indicators includes: obtaining the subjective weights of experts using the hierarchical analysis method, extracting historical engineering data through the knowledge base of the governance field, and calculating the objective weights using the entropy weight method; introducing dynamic weighting coefficients and adjusting the proportion of subjective and objective weights according to the pollution complexity of the mining area.

[0008] Furthermore, in step S4, calculating the correlation coefficient of the grey comprehensive evaluation includes: constructing the original matrix and performing dimensionless processing to determine the analysis sequence and the reference sequence; calculating the absolute difference of each indicator to determine the minimum and maximum values ​​of each indicator; and calculating the correlation coefficient based on the difference, minimum, and maximum values.

[0009] Furthermore, the method of obtaining expert subjective weights using the analytic hierarchy process includes: using a scale of 1 to 9 to quantify the importance of each level of elements in the indicator system by pairwise comparison, constructing a judgment matrix and performing a consistency test, and calculating the expert subjective weights.

[0010] Furthermore, in S4, the calculation of the correlation coefficient of the gray comprehensive evaluation also includes: retrieving the corresponding pollution control standards based on the knowledge base of the governance field, dynamically setting the optimal reference sequence of each indicator according to the pollution type; accessing real-time pollution monitoring data of the mining area, dynamically updating the correlation coefficient, and realizing dynamic evaluation in the governance process.

[0011] Furthermore, in S2, constructing a knowledge base for the governance field includes: screening reference materials in the field of pollution control to obtain source tracing and remediation technical data; performing text vectorization processing on the source tracing and remediation technical data to obtain vectorized data; storing the vectorized data in a vector database to form the foundation of the knowledge base; and combining a generative large model to optimize the input logic through prompting engineering to achieve accurate responses to user questions.

[0012] Further text vectorization processing includes: adjusting the size of data blocks and the proportion of overlap between two adjacent data blocks to a single data block.

[0013] Furthermore, the pollution characteristics of the target mining area include: soil pollution source type, characteristic pollutants and concentration, soil permeability, geological stability of the mining area, and distance from farmland.

[0014] This invention also provides a screening system for remediation technologies of soil pollution sources in mining areas based on a large model. The system is used to perform the method described in any of the preceding claims, and the system includes: The directory database construction module is used to integrate multi-source data of mining areas, pollution control case information and technical specifications to build a directory database of control technologies. The directory database of control technologies covers the control technologies corresponding to three types of soil pollution sources in mining areas: waste residue, tailings ponds and bottom sediment. The knowledge base construction module is used to build a knowledge base for the field of pollution control based on retrieval-enhanced generation technology and generative large models, with the pollution control technology catalog as the core and integrating reference materials in the field of pollution control. The evaluation system construction module is used to establish an evaluation index system for governance technologies based on the governance technologies in the governance technology catalog, the knowledge base of governance fields, and pollution control needs. The index system includes a target layer, a criterion layer, and an indicator layer. The target layer includes comprehensive evaluation indicators for feasible pollution control technologies. The criterion layer includes technical indicators, environmental indicators, and economic indicators. The indicator layer includes multiple detailed evaluation indicators. The comprehensive evaluation module is used to construct a coupled evaluation model based on the governance technology evaluation indicator system, combined with the analytic hierarchy process (AHP) and grey comprehensive evaluation method, for comprehensive evaluation. Specifically, it includes: comparing the importance of each level of elements in the indicator system pairwise and quantifying their values; constructing a judgment matrix to calculate the weights of the AHP indicators; constructing an evaluation indicator scoring table based on technical specifications and typical cases in the governance domain knowledge base, determining the optimal reference sequence, and then calculating the grey comprehensive evaluation correlation coefficient; obtaining the comprehensive technical score based on the AHP indicator weights and the grey comprehensive evaluation correlation coefficient, and classifying the comprehensive technical score. The screening module uses the pollution characteristics of the target mining area as the retrieval basis, matches candidate remediation technologies in the remediation knowledge base, dynamically adjusts the index weights and correlation coefficients of the coupled evaluation model based on the pollution characteristics and natural geographical conditions of the target mining area, calculates the comprehensive score of the candidate remediation technologies, and screens the optimal remediation technology.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: The method for screening remediation technologies for soil pollution sources in mining areas based on a large model provided by this invention uses a directory database to provide core data for the knowledge base, which in turn provides standards and case studies to support the evaluation model. The evaluation model then provides quantitative evidence for dynamic screening, with each step building upon and supporting the others. By introducing retrieval-enhanced generation technology and constructing a generative large model knowledge base in the remediation field, the uncertainty of expert decision-making is reduced, replacing the inefficient traditional method of manually reviewing literature and cases. The knowledge integration and precise retrieval capabilities of the large model significantly improve the efficiency of candidate technology matching. Simultaneously, the knowledge base can incorporate the latest field information, solving the problems of knowledge lag and technology omissions in traditional screening, and further improving screening efficiency.

[0016] Furthermore, by designing a dynamic adjustment mechanism, the weights and correlation coefficients of the evaluation model can be flexibly adjusted according to the pollution characteristics and natural geographical conditions of the target mining area, solving the problem of poor adaptability of inherent standards in traditional screening. Especially for mining areas with different pollution levels and terrain conditions, it can output personalized optimal technologies, reduce the adaptation risk after technology implementation, and help improve scenario adaptability. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is a flowchart of the screening method for the treatment of soil pollution sources in mining areas based on a large model, provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of the screening system for the treatment technology of soil pollution sources in mining areas based on a large model, provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0020] The specific embodiments of the present invention will be described below.

[0021] To address the issues of low efficiency and poor adaptability in screening existing soil pollution remediation technologies for mining areas, this invention integrates multi-source data from mining areas and pollution remediation case information to construct a remediation technology directory. Based on retrieval-enhanced generation technology and a generative large-scale model, a knowledge base for the remediation field is built, with the remediation technologies in the directory as the core. An evaluation index system for remediation technologies is established, and a coupled evaluation model is constructed by combining the analytic hierarchy process (AHP) and grey comprehensive evaluation method. Using the pollution characteristics of the target mining area as the retrieval basis, candidate remediation technologies are matched against the knowledge base. Based on the pollution characteristics and natural geographical conditions of the target mining area, the weights and correlation coefficients of the coupled evaluation model indicators are dynamically adjusted to calculate the comprehensive score of the candidate remediation technologies, thus selecting the optimal remediation technology. This invention achieves accurate and efficient screening of remediation technologies.

[0022] Example 1 This invention provides a method for screening remediation technologies for soil pollution sources in mining areas based on a large model. Figure 1 This is a flowchart of the screening method for mine area soil pollution source remediation technologies based on a large model provided in an embodiment of the present invention, such as... Figure 1 As shown, the specific steps include the following: S1. Integrate multi-source data from mining areas, pollution control case information, and technical specifications to construct a pollution control technology catalog. The catalog covers pollution control technologies corresponding to three types of soil pollution sources in mining areas: waste residue, tailings ponds, and bottom sediment.

[0023] Multi-source data for mining areas refers to a collection of data from various monitoring and survey stages within the mining area. Specifically, it includes soil monitoring data, groundwater hydrological data, pollutant composition data from waste residue, tailings ponds, and sediment, topographic and geological data of the mining area, and historical pollution remediation project data. This data forms the foundation for selecting suitable remediation technologies. Pollution remediation case studies refer to detailed records and summaries of completed or ongoing domestic and international mining area soil pollution remediation projects, providing practical verification for the remediation technology directory. Technical specifications refer to national, industry, or local standards, procedures, and technical guidelines related to mining area pollution remediation, representing the compliance and scientific principles that remediation technologies must follow. The remediation technology directory is a systematic and categorized structured database of mining area soil pollution remediation technologies, formed by integrating multi-source data, case studies, and technical specifications. Its core function is to accurately match remediation technologies according to pollution source type, serving as the core data carrier for subsequently building a knowledge base in the remediation field and matching candidate technologies.

[0024] Among them, slag refers to solid waste containing heavy metals generated during mining and processing, and is one of the main sources of pollution in mining areas. Tailings ponds are special facilities used to store tailings generated after mineral processing; their leakage and dam failure risks can cause soil and water pollution, making them a key pollution control area. Bottom sediment refers to the silt deposited at the bottom of water bodies surrounding mining areas, which easily accumulates heavy metals and organic pollutants, and is an important carrier for pollution diffusion.

[0025] S2. Based on retrieval-enhanced generation technology and generative large model, a knowledge base for the governance field is constructed, with governance technologies in the governance technology directory as the core and reference materials in the field of pollution control integrated.

[0026] Retrieval-Augmented Generation (RAG) technology refers to first retrieving precise information from specified data sources such as pollution control technology directories and reference materials in the field of pollution control, and then inputting the search results into a large-scale model to assist in generating content. Its core function is to compensate for the knowledge lag and inaccuracy issues in pure large-scale models within the specialized field of mining pollution control. The generative large-scale model refers to an artificial intelligence model with the ability to integrate and generate structured information. It is used to integrate core information from the pollution control technology directory with field reference materials to form standardized knowledge content in the pollution control field. Reference materials in the pollution control field include industry standards, typical engineering cases, academic literature, and technical manuals related to mining pollution control, used to enrich the authoritative background information of the knowledge base. The pollution control knowledge base is a structured collection of professional knowledge formed by integrating the core information from the pollution control technology directory and reference materials in the field of pollution control. It has precise retrieval capabilities and supports subsequent candidate technology matching and evaluation model operation.

[0027] Specifically, building a knowledge base for the governance field includes: screening reference materials in the field of pollution control to obtain source tracing and remediation technical data; vectorizing the source tracing and remediation technical data to obtain vectorized data; storing the vectorized data in a vector database to form the foundation of the knowledge base; and combining a generative large model to optimize the input logic through prompting engineering to achieve accurate responses to user questions.

[0028] Text vectorization processing includes: adjusting the size of data blocks and the proportion of overlap between two adjacent data blocks to a single data block.

[0029] Source tracing and remediation technical data refers to detailed information selected from reference materials in the field of pollution control, focusing on pollution source tracing and remediation. It forms the core material for building the knowledge base, including case studies of pollution source remediation in mining areas and technical specifications. Text vectorization refers to the process of converting textual information such as source tracing and remediation technical data into computer-recognizable and computable vectors. Its core principle is to preserve the core semantics of the text, laying the foundation for efficient retrieval and matching. Vectorized data refers to the output of text vectorization, i.e., a numerical array carrying the semantics of the text. It is characterized by easy storage, easy computation, and high semantic matching efficiency, and is the core data form fundamental to the knowledge base.

[0030] Reference materials in the field of pollution control are collected and selectively screened to extract relevant content on pollution source tracing and remediation, resulting in source tracing and remediation technical data that ensures accuracy and relevance. This data is then vectorized, breaking it down into text segments based on its type and semantic length. The size of each data block (text segment) and the overlap ratio of adjacent blocks are adjusted to obtain vectorized data that retains its core semantics. This vectorized data is stored in a vector database, forming the foundation of a structured knowledge base. Combined with a generative large-scale model, the input logic is optimized through prompting engineering, enabling the knowledge base to accurately respond to user questions. Simultaneously, using a pollution control technology directory as the core, the retrieved reference information is integrated through the large-scale model to generate structured knowledge units, completing the knowledge base construction and validation.

[0031] By screening source tracing and remediation technical data and avoiding interference from irrelevant information, the knowledge base content is ensured to focus on core remediation needs and match actual scenarios of mining area pollution remediation, thereby improving the accuracy of the knowledge base content in the remediation field. Through text vectorization processing and vector database storage, knowledge retrieval efficiency is significantly improved. Adjusting data block size and overlap ratios avoids the loss of key semantics, ensuring the accuracy of retrieval matching and guaranteeing both retrieval efficiency and semantic integrity.

[0032] S3. Establish a governance technology evaluation index system based on the governance technologies in the governance technology catalog, the governance field knowledge base, and pollution control needs. The index system includes a target layer, a criterion layer, and an indicator layer. The target layer includes comprehensive evaluation indicators for feasible pollution control technologies. The criterion layer includes technical indicators, environmental indicators, and economic indicators. The indicator layer includes multiple detailed evaluation indicators.

[0033] Pollution control needs address the specific remediation demands of the three types of pollution mentioned above, serving as the core guiding principle for constructing the pollution control technology evaluation indicator system. The target layer is the top layer of the pollution control technology evaluation indicator system, clarifying the final evaluation direction. Specifically, it consists of comprehensive evaluation indicators for feasible pollution control technologies, anchoring the core objective of selecting feasible remediation technologies suitable for the mining area. The criteria layer is the intermediate level of the evaluation system, a first-level breakdown of the target layer, containing three categories: technical indicators, environmental indicators, and economic indicators, corresponding to the three evaluation dimensions of technical feasibility, environmental remediation effectiveness, and economic rationality, respectively. The indicator layer is the bottom-level execution layer of the evaluation system, a detailed implementation of the criteria layer, composed of quantifiable detailed evaluation indicators, providing a basis for subsequent evaluation parameter extraction.

[0034] Specifically, the detailed assessment indicators include: technical indicators corresponding to technology maturity, operability, public acceptance, applicable soil permeability, pollutant removal rate, stability of applicable geological conditions, various pollutants, impact on surrounding farmland and remediation time; environmental indicators corresponding to secondary pollution risk; and economic indicators corresponding to unit treatment cost, annual operation and maintenance cost, and economic benefits of resource recovery.

[0035] Based on the above embodiments, the technical characteristics of the assessment object are clarified according to the governance technology catalog, and industry standards and assessment criteria are extracted from the governance field knowledge base. Simultaneously, the specific pollution control needs of the mining area are identified. With the selection of feasible governance technologies as the core, the comprehensive evaluation indicators for feasible pollution control technologies at the target layer are first determined. Then, the target layer is broken down into a criterion layer, and finally, each criterion layer is refined into quantifiable detailed assessment indicators, forming a complete three-level indicator system. The indicator system is verified to ensure its suitability for the technology types in the governance technology catalog, its compliance with industry standards in the governance field knowledge base, and its alignment with the pollution control needs of the mining area, ensuring the system's relevance and operability. The three-level indicator system covers all dimensions of technology, environment, and economy, avoiding the problems of traditional assessment dimensions being singular and one-sided, ensuring a comprehensive consideration of governance technologies and achieving a comprehensive assessment. Furthermore, based on the governance technology catalog and the pollution control needs of the mining area, the indicator system is constructed to ensure its suitability for the governance technology characteristics of the three types of pollution sources, avoiding the blindness of a general indicator system and ensuring the relevance of the assessment.

[0036] S4. Based on the governance technology evaluation index system, and combining the analytic hierarchy process (AHP) and grey comprehensive evaluation method, a coupled evaluation model is constructed for comprehensive evaluation. Specifically, this includes: comparing the importance of each level of elements in the index system pairwise and assigning quantitative values; constructing a judgment matrix to calculate the weights of the AHP indicators; constructing an evaluation index scoring table based on technical specifications and typical cases in the governance domain knowledge base, determining the optimal reference sequence, and then calculating the grey comprehensive evaluation correlation coefficient; obtaining the comprehensive technology score based on the weights of the AHP indicators and the grey comprehensive evaluation correlation coefficient, and classifying the comprehensive technology score.

[0037] The core of the Analytic Hierarchy Process (AHP) is to construct a judgment matrix by comparing the importance of each pair of factors, calculating the weights of the indicators, and incorporating expert subjective experience to adapt to the needs of ranking the importance of evaluation indicators. The Grey Comprehensive Evaluation Method is an evaluation method for uncertain systems where some information is known and some is unknown. It achieves objective quantitative evaluation by determining the optimal reference sequence and calculating correlation coefficients, making it suitable for scenarios where pollution control data in mining areas is incomplete. The Coupled Evaluation Model refers to a comprehensive evaluation tool that integrates the AHP and the Grey Comprehensive Evaluation Method, taking into account both subjective experience and objective data to address the limitations of single evaluation methods.

[0038] The calculation of the weights of the analytic hierarchy process includes: obtaining subjective weights from experts using the analytic hierarchy process, extracting historical engineering data from the governance knowledge base, and calculating objective weights using the entropy weight method; and introducing dynamic weighting coefficients to adjust the proportion of subjective and objective weights according to the complexity of pollution in the mining area.

[0039] The analytic hierarchy process (AHP) was employed, combined with a governance technology assessment indicator system. Experts compared the pairwise importance of each level of the indicator system, quantified values ​​using standardized methods such as the 1-9 scale, constructed a judgment matrix, and performed consistency checks. Finally, subjective weights reflecting expert experience were calculated. A knowledge base in the governance field was used as the data source, precisely extracting historical engineering data related to mine pollution control. Based on this objective data, the entropy weight method was used for calculation; the higher the entropy value, the greater the impact of the indicator on the assessment results. Objective weights, unaffected by subjective experience, were calculated accordingly. A dynamic weighting coefficient was introduced as an adjustment basis, flexibly adjusting the ratio of subjective to objective weights based on the pollution complexity of the target mine area. This ensured that the weight allocation was adapted to the actual pollution situation of the specific mine area, improving the relevance and accuracy of subsequent assessments.

[0040] The calculation of the correlation coefficient of grey comprehensive evaluation includes: constructing the original matrix and performing dimensionless processing, determining the analysis series and the reference series; calculating the absolute difference of each indicator, determining the minimum and maximum values ​​of each indicator; and calculating the correlation coefficient based on the difference, minimum and maximum values.

[0041] The process involves constructing an original matrix and performing dimensionless processing, while simultaneously defining the analytical and reference sequences. First, the original matrix is ​​built based on the data from the evaluation index scoring table. Rows in the matrix correspond to different governance technologies to be evaluated, and columns correspond to specific evaluation indicators. Then, the original matrix is ​​dimensionless, the core purpose of which is to eliminate interference from differences in dimensions between different indicators. Standardization and normalization methods are commonly used. After processing, the indicator data for each governance technology are determined as the analytical sequence, and the optimal reference sequence determined earlier in conjunction with the governance domain knowledge base is used as the reference sequence. The absolute differences of each indicator are calculated, and extreme values ​​are determined. For each governance technology's analytical sequence, the absolute difference between it and the corresponding indicator element in the reference sequence is calculated one by one, obtaining the difference data for each indicator. The absolute differences of all governance technologies under each indicator are traversed, and the global minimum and global maximum values ​​in the entire difference set are selected to provide basic parameters for subsequent correlation coefficient calculations. The correlation coefficient is calculated based on the differences, minimum values, and maximum values. By combining the preset discrimination coefficients and substituting them into the core formula of the grey comprehensive evaluation method, the absolute difference, global minimum value, and global maximum value of each individual indicator are used for calculation. Finally, the correlation coefficient of each governance technology under different indicators is obtained, forming a set of correlation coefficients, which lays the foundation for subsequent comprehensive scoring by combining weight calculation technology.

[0042] For example, based on the governance technology evaluation index system, a coupled evaluation model is constructed using the analytic hierarchy process (AHP) and the grey comprehensive evaluation method for comprehensive evaluation. Specifically, this includes: quantifying the importance of each level of elements in the index system using a 1-9 scale, constructing a judgment matrix and performing consistency checks to calculate expert subjective weights; extracting historical engineering data from a governance domain knowledge base and calculating objective weights using the entropy weight method; introducing dynamic weighting coefficients to adjust the proportion of subjective and objective weights according to the complexity of pollution in the mining area, with the objective weight proportion exceeding a preset value in multi-metal synergistic pollution scenarios; extracting index parameters for each governance technology based on multi-source data from the mining area and a governance technology directory to construct an index matrix; simultaneously determining the optimal reference sequence by combining technical standards and typical case data from the governance domain knowledge base, and then calculating the grey comprehensive evaluation correlation coefficient; and obtaining the comprehensive technical score Ei based on the AHP index weights and the grey comprehensive evaluation correlation coefficient, and classifying it according to the following standards: excellent level corresponds to a comprehensive score Ei ≥ 0.7, good level corresponds to a comprehensive score 0.5 ≤ Ei < 0.7, and acceptable level corresponds to a comprehensive score Ei < 0.5. For each new real-world application case, the weights of the hierarchical analysis indicators are automatically adjusted using the case data, forming an iterative closed loop.

[0043] Through multi-dimensional design, the assessment of pollution control technologies in mining areas has achieved precision, adaptability, and long-term optimization. The Analytic Hierarchy Process (AHP) integrates expert subjective experience, while the entropy weight method mines objective data from historical projects, balancing the adaptability of experience with the objectivity of data and avoiding the one-sidedness of single-weight calculations. Dynamic weighting coefficients are introduced, and the proportion of objective weights is increased for complex scenarios such as multi-metal synergistic pollution, ensuring that weight allocation accurately matches the complexity of pollution in mining areas and solving the problems of fixed weights and poor adaptability in traditional assessments. A standardized indicator matrix is ​​constructed based on multi-source data from mining areas and a directory of pollution control technologies. Reference sequences are determined by combining authoritative standards from a knowledge base in the field of pollution control, ensuring that assessment parameters are quantifiable and based on authoritative sources. A clear three-level grading standard makes the assessment results intuitive and easy to understand, quickly distinguishing the feasibility differences of different pollution control technologies and providing a clear and reliable quantitative basis for subsequent selection of the optimal technology. By adding practical application cases, the weights of the AHP indicators are automatically adjusted, enabling the assessment model to continuously absorb practical data and align with actual pollution control needs, avoiding assessment biases caused by model rigidity and adapting to dynamic changes in different mining areas and different stages of pollution control. From extracting historical engineering data and referencing technical standards to determining indicator parameters, the entire process is supported by a knowledge base in the field of governance. This ensures that the assessment process complies with industry norms and that the assessment results are authoritative and persuasive, thus providing a guarantee for the compliant implementation of governance technologies.

[0044] S5. Using the pollution characteristics of the target mining area as the retrieval basis, candidate remediation technologies are matched in the remediation knowledge base. Based on the pollution characteristics and natural geographical conditions of the target mining area, the index weights and correlation coefficients of the coupled evaluation model are dynamically adjusted to calculate the comprehensive score of the candidate remediation technologies and select the optimal remediation technology.

[0045] The pollution characteristics of a target mining area refer to its unique pollution-related attributes, including: soil pollution source type, characteristic pollutants and their concentrations, soil permeability, geological stability, and distance from farmland. Characteristic pollutants include heavy metals and other pollutants.

[0046] Natural geographical conditions refer to the geographical and natural environmental attributes of the target mining area, such as topography, climate conditions, hydrological characteristics, and soil type. These factors affect the adaptability and implementation difficulty of treatment technologies and are important considerations for dynamically adjusting the evaluation model.

[0047] Using the pollution characteristics of the target mining area as the core retrieval basis, a precise search is conducted in the knowledge base of remediation fields. Semantic matching is used to screen remediation technologies that initially match the pollution characteristics, forming a candidate remediation technology set. Combining the pollution characteristics and natural geographical conditions of the target mining area, the indicator weights and correlation coefficients of the coupled evaluation model are dynamically adjusted to ensure that the model parameters accurately adapt to the unique scenario of the target mining area. The indicator parameters of the candidate remediation technologies are substituted into the adjusted coupled evaluation model to calculate the comprehensive score of each candidate technology. Based on the comprehensive scores, the technologies are prioritized, and the remediation technology with the highest score and optimal comprehensive feasibility is selected, completing the entire technology screening process.

[0048] In the large-scale model-based method for screening remediation technologies for soil pollution sources in mining areas provided in this embodiment, the directory database provides core data for the knowledge base, the knowledge base provides standards and case support for the evaluation model, and the evaluation model provides quantitative basis for dynamic screening. Each step is progressive and mutually supportive. By introducing retrieval-enhanced generation technology and constructing a generative large-scale model to build a knowledge base in the remediation field, the inefficient mode of traditional manual literature and case review is replaced. Through the knowledge integration and precise retrieval capabilities of the large-scale model, the matching efficiency of candidate technologies is greatly improved. At the same time, the knowledge base can integrate the latest field information, solving the problems of knowledge lag and technology omission in traditional screening, and further improving screening efficiency.

[0049] Furthermore, by designing a dynamic adjustment mechanism, the weights and correlation coefficients of the evaluation model can be flexibly adjusted according to the pollution characteristics and natural geographical conditions of the target mining area, solving the problem of poor adaptability of inherent standards in traditional screening. Especially for mining areas with different pollution levels and terrain conditions, it can output personalized optimal technologies, reduce the adaptation risk after technology implementation, and help improve scenario adaptability.

[0050] Example 2 This invention also provides a screening system for remediation technologies for soil pollution sources in mining areas based on a large model. Figure 2 This is a schematic diagram of the structure of the screening system for mine soil pollution source remediation technology based on a large model provided in an embodiment of the present invention, as shown below. Figure 2 As shown, the system includes: The directory database construction module is used to integrate multi-source data of mining areas, pollution control case information and technical specifications to build a directory database of control technologies. The directory database of control technologies covers the control technologies corresponding to three types of soil pollution sources in mining areas: waste residue, tailings ponds and bottom sediment. The knowledge base construction module is used to build a knowledge base for the field of pollution control based on retrieval-enhanced generation technology and generative large models, with the pollution control technology catalog as the core and integrating reference materials in the field of pollution control. The evaluation system construction module is used to establish an evaluation index system for governance technologies based on the governance technologies in the governance technology catalog, the knowledge base of governance fields, and pollution control needs. The index system includes a target layer, a criterion layer, and an indicator layer. The target layer includes comprehensive evaluation indicators for feasible pollution control technologies. The criterion layer includes technical indicators, environmental indicators, and economic indicators. The indicator layer includes multiple detailed evaluation indicators. The comprehensive evaluation module is used to construct a coupled evaluation model based on the governance technology evaluation indicator system, combined with the analytic hierarchy process (AHP) and grey comprehensive evaluation method, for comprehensive evaluation. Specifically, it includes: comparing the importance of each level of elements in the indicator system pairwise and quantifying their values; constructing a judgment matrix to calculate the weights of the AHP indicators; constructing an evaluation indicator scoring table based on technical specifications and typical cases in the governance domain knowledge base, determining the optimal reference sequence, and then calculating the grey comprehensive evaluation correlation coefficient; obtaining the comprehensive technical score based on the AHP indicator weights and the grey comprehensive evaluation correlation coefficient, and classifying the comprehensive technical score. The screening module uses the pollution characteristics of the target mining area as the retrieval basis, matches candidate remediation technologies in the remediation knowledge base, dynamically adjusts the index weights and correlation coefficients of the coupled evaluation model based on the pollution characteristics and natural geographical conditions of the target mining area, calculates the comprehensive score of the candidate remediation technologies, and screens the optimal remediation technology.

[0051] The large-model-based technology screening system for mine soil pollution source control provided in this embodiment executes the large-model-based technology screening method for mine soil pollution source control described in any of the above embodiments, and has the beneficial effects of any of the above embodiments, which will not be repeated here.

[0052] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A method for screening remediation technologies for soil pollution sources in mining areas based on a large-scale model, characterized in that, include: S1. Integrate multi-source data, pollution control case information, and technical specifications of mining areas to build a pollution control technology catalog. The pollution control technology catalog covers the pollution control technologies corresponding to three types of soil pollution sources in mining areas: waste residue, tailings ponds, and bottom sediment. S2. Based on retrieval-enhanced generation technology and generative large model, with the governance technologies in the governance technology directory as the core, and integrating reference materials in the field of pollution control, a knowledge base for the governance field is constructed. S3. Establish a governance technology evaluation index system based on the governance technologies in the governance technology catalog, the governance field knowledge base, and the pollution control needs. The index system includes a target layer, a criterion layer, and an indicator layer. The target layer includes a comprehensive evaluation index for feasible pollution control technologies. The criterion layer includes technical indicators, environmental indicators, and economic indicators. The indicator layer includes multiple detailed evaluation indicators. S4. Based on the governance technology evaluation index system, combined with the analytic hierarchy process (AHP) and grey comprehensive evaluation method, a coupled evaluation model is constructed for comprehensive evaluation. Specifically, this includes: comparing the importance of each level of elements in the index system pairwise and assigning quantitative values, and constructing a judgment matrix to calculate the weights of the AHP indicators. Based on technical specifications and typical cases in the governance knowledge base, an evaluation index scoring table is constructed, and the optimal reference sequence is determined. Then, the grey comprehensive evaluation correlation coefficient is calculated. According to the weight of the hierarchical analysis index and the grey comprehensive evaluation correlation coefficient, the technical comprehensive score is obtained, and the technical comprehensive score is classified. S5. Using the pollution characteristics of the target mining area as the retrieval basis, candidate remediation technologies are matched in the remediation knowledge base. Based on the pollution characteristics and natural geographical conditions of the target mining area, the index weights and correlation coefficients of the coupled evaluation model are dynamically adjusted to calculate the comprehensive score of the candidate remediation technologies and select the optimal remediation technology.

2. The method for screening remediation technologies for soil pollution sources in mining areas based on a large model, as described in claim 1, is characterized in that... The S3 includes several detailed evaluation indicators, such as: technical indicators corresponding to technology maturity, operability, public acceptance, applicable soil permeability, pollutant removal rate, applicable geological condition stability, various pollutants, impact on surrounding farmland, and remediation time; environmental indicators corresponding to secondary pollution risk; and economic indicators corresponding to unit treatment cost, annual operation and maintenance cost, and economic benefits of resource recovery.

3. The method for screening remediation technologies for soil pollution sources in mining areas based on a large model according to claim 1, characterized in that, In step S4, calculating the weights of the analytic hierarchy process indicators includes: The analytic hierarchy process was used to obtain the subjective weights of experts, and historical engineering data was extracted from the governance domain knowledge base. The entropy weight method was used to calculate the objective weights. A dynamic weighting coefficient is introduced to adjust the proportion of subjective and objective weights based on the complexity of pollution in the mining area.

4. The method for screening remediation technologies for soil pollution sources in mining areas based on a large model, as described in claim 1, is characterized in that... In step S4, calculating the correlation coefficient of the grey comprehensive evaluation includes: constructing the original matrix and performing dimensionless processing to determine the analysis sequence and the reference sequence; calculating the absolute difference of each indicator to determine the minimum and maximum values ​​of each indicator; and calculating the correlation coefficient based on the difference, minimum and maximum values.

5. The method for screening remediation technologies for soil pollution sources in mining areas based on a large model according to claim 3, characterized in that, The analytic hierarchy process (AHP) is used to obtain expert subjective weights, including: The importance of each level of elements in the indicator system is compared and quantified using a scale of 1 to 9. A judgment matrix is ​​constructed and a consistency test is performed to calculate the subjective weight of the experts.

6. The method according to claim 1, characterized in that, In step S4, the calculation of the correlation coefficient for grey comprehensive evaluation also includes: Based on the knowledge base of the governance field, the corresponding pollution control standards are retrieved, and the optimal reference sequence of each indicator is dynamically set according to the type of pollution. By accessing real-time pollution monitoring data from the mining area and dynamically updating correlation coefficients, dynamic evaluation of the remediation process can be achieved.

7. The method for screening remediation technologies for soil pollution sources in mining areas based on a large model according to claim 1, characterized in that, In S2, constructing a governance domain knowledge base includes: Reference materials in the field of pollution control were screened to obtain source tracing and remediation technology data; The technical data on source tracing and remediation are processed into vectorized text to obtain vectorized data; Vectorized data is stored in a vector database, forming the foundation of the knowledge base; By combining generative large models and optimizing input logic through prompting engineering, accurate responses to user questions can be achieved.

8. The method for screening remediation technologies for soil pollution sources in mining areas based on a large model, as described in claim 7, is characterized in that... Text vectorization processing includes: Adjust the size of the data blocks and the proportion of the overlap between two adjacent data blocks to a single data block.

9. The method for screening remediation technologies for soil pollution sources in mining areas based on a large model according to claim 1, characterized in that, The pollution characteristics of the target mining area include: soil pollution source type, characteristic pollutants and concentration, soil permeability, geological stability of the mining area, and distance from farmland.

10. A screening system for remediation technologies of soil pollution sources in mining areas based on a large model, characterized in that, The system executes the screening method for mine area soil pollution source remediation technologies based on a large model as described in any one of claims 1-9, and the system comprises: The directory database construction module is used to integrate multi-source data of mining areas, pollution control case information and technical specifications to build a directory database of control technologies. The directory database of control technologies covers the control technologies corresponding to three types of soil pollution sources in mining areas: waste residue, tailings ponds and bottom sediment. The knowledge base construction module is used to build a knowledge base for the field of pollution control based on retrieval-enhanced generation technology and generative large models, with the pollution control technology catalog as the core and integrating reference materials in the field of pollution control. The evaluation system construction module is used to establish an evaluation index system for governance technologies based on the governance technologies in the governance technology catalog, the knowledge base of governance fields, and pollution control needs. The index system includes a target layer, a criterion layer, and an indicator layer. The target layer includes comprehensive evaluation indicators for feasible pollution control technologies. The criterion layer includes technical indicators, environmental indicators, and economic indicators. The indicator layer includes multiple detailed evaluation indicators. The comprehensive evaluation module is used to construct a coupled evaluation model based on the governance technology evaluation indicator system, combined with the analytic hierarchy process (AHP) and grey comprehensive evaluation method, for comprehensive evaluation. Specifically, it includes: comparing the importance of each level of elements in the indicator system pairwise and quantifying their values; constructing a judgment matrix to calculate the weights of the AHP indicators; constructing an evaluation indicator scoring table based on technical specifications and typical cases in the governance domain knowledge base, determining the optimal reference sequence, and then calculating the grey comprehensive evaluation correlation coefficient; obtaining the comprehensive technical score based on the AHP indicator weights and the grey comprehensive evaluation correlation coefficient, and classifying the comprehensive technical score. The screening module uses the pollution characteristics of the target mining area as the retrieval basis, matches candidate remediation technologies in the remediation knowledge base, dynamically adjusts the index weights and correlation coefficients of the coupled evaluation model based on the pollution characteristics and natural geographical conditions of the target mining area, calculates the comprehensive score of the candidate remediation technologies, and screens the optimal remediation technology.

Images