An environmental protection management method and system for the food industry based on a large model
By combining large language models and dynamic knowledge graphs, the problems of low efficiency and insufficient accuracy in pollutant monitoring in the food industry have been solved, enabling efficient and accurate pollutant analysis and real-time supervision by government systems.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NANJING YRD ECO DEV RI CO LTD
- Filing Date
- 2025-07-30
- Publication Date
- 2026-07-31
AI Technical Summary
Existing technologies for monitoring and analyzing pollutants in the food industry are inefficient, with insufficient accuracy in manual sampling and analysis, making continuous monitoring impossible and the analysis results inaccurate.
An environmental management approach based on a large language model is adopted. Through real-time data acquisition from sensors, dynamic knowledge graph updates, and a self-attention mechanism, accurate evaluation reports are generated and synchronized with the government system. The data acquisition cycle is adjusted to improve monitoring efficiency and accuracy.
It has achieved efficient, continuous, and accurate monitoring of pollutants in the food industry, and improved the reliability of analysis reports and the compliance of government affairs processing.
Smart Images
Figure CN120996006B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of environmental protection technology, specifically to an environmental management method and system for the food industry based on a large model. Background Technology
[0002] In the food industry, monitoring and analyzing the discharge of pollutants, including various solid wastes and wastewater, is not only a requirement of policies and regulations, but also a necessity for environmental protection, and is closely related to food safety and human health.
[0003] The solid waste includes defective products, spoiled food, waste oil, and waste organic solvents. The wastewater contains harmful substances such as animal and vegetable oils, ammonia nitrogen, pesticide residues, and heavy metals. In the existing technology, the monitoring and analysis of solid waste and wastewater mostly require the environmental protection department of the enterprise or a professional institution to conduct manual sampling and analysis with the assistance of relevant instruments.
[0004] However, the above methods have the following drawbacks: First, manual sampling is inefficient and cannot be used for effective continuous monitoring. Second, during manual analysis, the limited knowledge and retrieval capabilities of analysts, the dynamic updates of relevant regulations, and the iterative updates of technologies such as waste storage and wastewater treatment create information gaps that make it impossible to guarantee the accuracy of the analysis results. Summary of the Invention
[0005] The purpose of this invention is to provide an environmental management method and system for the food industry based on a large model, so as to solve the technical problems of low efficiency and low accuracy of analysis results when the pollutant sampling and analysis is based on manual sampling in the prior art.
[0006] To achieve the above objectives, the present invention proposes the following technical solution:
[0007] Firstly, this technical solution provides an environmental management method for the food industry based on a large-scale model, including:
[0008] Acquire various real-time data from each sensor according to a preset cycle, and fill them into a preset data template to obtain a real-time data report;
[0009] Based on the large language model, prompt words are extracted from the real-time data report to obtain a first prompt word set, and several first knowledge fragments are obtained from the dynamically updated knowledge graph based on the first prompt word set.
[0010] Based on the large language model, prompt words are extracted from each first knowledge segment to obtain a second prompt word set, and the first prompt word set and the second prompt word set are merged to form the optimal prompt word set;
[0011] The large language model uses preset evaluation items as indexes and generates the first initial evaluation text based on the optimal prompt word set. The evaluation items include data type sub-items, pollution level sub-items, and optimization scheme sub-items, each with a corresponding setting. The large language model introduces a self-attention mechanism to dynamically allocate weights to each data type sub-item according to the latest environmental regulations. The large language model is periodically iterated and updated at a preset frequency, and data samples are constructed using the latest pollution-related data from various industries during both the initial training and each iteration update.
[0012] The first initial evaluation text is segmented according to the preset evaluation report template and filled into the corresponding position of the evaluation report template to obtain a real-time evaluation report corresponding to the real-time data report.
[0013] Furthermore, after the large language model generates a first initial evaluation text based on the optimal prompt word set, using preset evaluation items as indexes, it includes:
[0014] Input the first initial evaluation text into a dynamically updated knowledge graph to obtain several second knowledge fragments;
[0015] The first initial evaluation text and the second knowledge fragment are compared based on a large language model. When it is determined that some content in the first initial evaluation text has the opposite semantics to the corresponding content in the second knowledge fragment, some content in the first initial evaluation text is replaced with the corresponding content in the second knowledge fragment to obtain an updated evaluation text.
[0016] Furthermore, the step of segmenting the first initial evaluation text according to a preset evaluation report template and filling it into the corresponding position in the evaluation report template to obtain a real-time evaluation report corresponding to the real-time data report includes:
[0017] The real-time evaluation report will be synchronized and updated to the government affairs system.
[0018] When an abnormality in emissions is detected, a penalty instruction is issued to the enterprise corresponding to the real-time evaluation report, and the acquisition cycle of various real-time data is adjusted according to the rectification cycle to acquire and analyze the emission data after rectification in real time.
[0019] Furthermore, the step of acquiring various real-time data from each sensor at a preset period and filling them into a preset data template to obtain a real-time data report includes:
[0020] Get the number of penalty instructions received by any enterprise within the first preset time period;
[0021] The acquisition cycle of various real-time data of the corresponding enterprise is dynamically adjusted based on the number of instructions; wherein, the acquisition cycle is inversely proportional to the number of instructions.
[0022] Furthermore, including:
[0023] Obtain all real-time evaluation reports for any enterprise within the second preset time period;
[0024] The real-time evaluation report is analyzed and processed based on a large language model to obtain a second initial report text.
[0025] The second initial evaluation text is segmented according to the preset summary report template and filled into the corresponding positions of the summary report template to obtain an evaluation summary report corresponding to all real-time evaluation reports.
[0026] Secondly, this technical solution provides an environmental management system for the food industry based on a large-scale model, including:
[0027] The first acquisition module is used to acquire various real-time data from each sensor according to a preset cycle and fill them into a preset data template to obtain a real-time data report.
[0028] The second acquisition module is used to extract prompt words from the real-time data report based on the large language model to obtain a first prompt word set, and to acquire several first knowledge fragments from the dynamically updated knowledge graph based on the first prompt word set.
[0029] The third acquisition module extracts prompt words from each first knowledge segment based on the large language model to obtain a second prompt word set, and merges the first prompt word set and the second prompt word set to form the optimal prompt word set;
[0030] The text generation module is used by the large language model to generate the first initial evaluation text based on the preset evaluation items and the optimal prompt word set. The evaluation items include data type sub-items, pollution level sub-items, and optimization scheme sub-items, each with a corresponding setting. The large language model introduces a self-attention mechanism to dynamically allocate weights to each data type sub-item according to the latest environmental regulations. The large language model is periodically iterated and updated at a preset frequency, and data samples are constructed using the latest pollution-related data from various industries during both the initial training and each iteration update.
[0031] The report generation module is used to segment the first initial evaluation text according to a preset evaluation report template and fill it into the corresponding position of the evaluation report template to obtain a real-time evaluation report corresponding to the real-time data report.
[0032] Furthermore, including:
[0033] The fourth acquisition module is used to input the first initial evaluation text into a dynamically updated knowledge graph to obtain several second knowledge fragments;
[0034] The text correction module is used to compare the first initial evaluation text and the second knowledge fragment based on the large language model, and when it is determined that some content in the first initial evaluation text has the opposite semantics to the corresponding content in the second knowledge fragment, it replaces some content in the first initial evaluation text with the corresponding content in the second knowledge fragment to obtain an updated evaluation text.
[0035] Furthermore, including:
[0036] The synchronization update module is used to synchronize the real-time evaluation report to the government affairs system.
[0037] The periodic update module is used to issue penalty instructions to the enterprise corresponding to the real-time evaluation report when abnormal emissions are detected, and to adjust the acquisition cycle of various real-time data according to the rectification cycle so as to acquire and analyze the emission data after rectification in real time.
[0038] Thirdly, this technical solution provides an electronic device including at least one processor coupled to a memory, the memory storing a computer program configured to be executed by the processor when it is run.
[0039] Fourthly, this technical solution provides a computer-readable storage medium, characterized in that it stores a computer program thereon, which, when executed by a computer, implements the method described thereon.
[0040] Beneficial effects:
[0041] As can be seen from the above technical solutions, the technical solution of the present invention provides an environmental management method for the food industry based on a large model to solve the technical defects of low efficiency and low accuracy of analysis results in the existing technology when pollutant sampling and analysis are based on manual sampling.
[0042] The method includes: First, acquiring various real-time data from each sensor at a preset cycle and filling them into a preset data template to obtain a real-time data report. This sensor-based real-time data reporting improves the efficiency and sustainability of data acquisition. Simultaneously, standardizing the real-time data according to the data template also improves the accuracy of subsequent processing based on a large language model. Next, using the large language model, prompt words are extracted from the real-time data report to obtain a first prompt word set, and several first knowledge fragments are acquired from a dynamically updated knowledge graph based on the first prompt word set. Prompt words are then extracted from each of the first knowledge fragments using the large language model to obtain a second prompt word set, and the first and second prompt word sets are merged to form an optimal prompt word set. Here, the introduction of a dynamically updated knowledge graph during prompt word acquisition ensures the comprehensiveness and accuracy of each prompt word in the optimal prompt word set. Then, the large language model uses preset evaluation items as an index and generates a first initial evaluation text based on the optimal prompt word set. Specifically, the evaluation items include corresponding data type sub-items, pollution level sub-items, and optimization scheme sub-items. The large language model is periodically iterated and updated at a preset frequency. During initial training and each subsequent iteration, data samples are constructed using the latest pollution discharge data from various industries. A self-attention mechanism is also introduced into the large language model to dynamically allocate weights to each data type sub-item according to the latest environmental regulations, thereby improving the accuracy of the large language model's text output. Finally, the first initial evaluation text is segmented according to a preset evaluation report template, and the data is filled into the corresponding positions in the evaluation report template to obtain a real-time evaluation report corresponding to the real-time data report. At this point, the real-time evaluation report has higher accuracy compared to manually analyzed reports.
[0043] It should be understood that all combinations of the foregoing concepts and the additional concepts described in more detail below can be considered part of the inventive subject matter of this disclosure, provided that such concepts do not contradict each other.
[0044] The foregoing and other aspects, embodiments, and features of the teachings of the present invention will be more fully understood from the following description in conjunction with the accompanying drawings. Other additional aspects of the invention, such as features and / or beneficial effects of exemplary embodiments, will become apparent from the following description or may be learned through practice of specific embodiments according to the teachings of the present invention. Attached Figure Description
[0045] The accompanying drawings are not intended to be drawn to scale. In the drawings, each identical or nearly identical component shown in the various figures may be denoted by the same reference numeral. For clarity, not every component is labeled in each figure. Embodiments of various aspects of the invention will now be described by way of example and with reference to the accompanying drawings, wherein:
[0046] Figure 1 This is a flowchart of the environmental management method for the food industry based on a large model, as described in this embodiment.
[0047] Figure 2 A flowchart for updating the initial evaluation text;
[0048] Figure 3 A flowchart for synchronizing real-time evaluation reports to the government affairs system;
[0049] Figure 4 A flowchart for adjusting the real-time data acquisition cycle;
[0050] Figure 5 A flowchart for obtaining the evaluation summary report;
[0051] Figure 6 This is a structural block diagram of the environmental management system for the food industry based on a large model, as described in this embodiment.
[0052] Figure 7 This is a structural block diagram of the electronic device described in this embodiment. Detailed Implementation
[0053] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention without creative effort are within the scope of protection of the present invention. Unless otherwise defined, the technical or scientific terms used herein should have the ordinary meaning understood by those skilled in the art to which this invention pertains.
[0054] The terms "first," "second," and similar words used in this application specification and claims do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Similarly, unless the context clearly indicates otherwise, the singular forms of "an," "a," or "the," etc., do not indicate a quantity limitation, but rather indicate the presence of at least one. Terms such as "comprising" or "including" mean that the element or object preceding "comprising" encompasses the features, integrals, steps, operations, elements, and / or components listed following "comprising" or "including," and do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or sets thereof. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0055] In the food industry, monitoring and analyzing pollutant emissions is not only a policy and regulatory requirement but also a necessity for environmental protection, closely linked to food safety and human health. However, current technologies largely rely on manual methods for these processes, leading to the following drawbacks: manual sampling is inefficient and cannot provide continuous monitoring. Furthermore, the limited knowledge and retrieval capabilities of analysts, coupled with the dynamic updates to relevant regulations and the iterative advancements in waste storage and wastewater treatment technologies, result in information gaps that compromise the accuracy of analytical results. Therefore, this embodiment aims to provide a large-scale model-based environmental management method for the food industry to simultaneously address these technical shortcomings.
[0056] The environmental management method for the food industry based on a large model, as described in this embodiment, will be described in detail below with reference to the accompanying drawings.
[0057] Combination Figure 1 As shown, the method includes the following steps:
[0058] Step S202: Acquire various real-time data from each sensor according to a preset cycle, and fill them into a preset data template to obtain a real-time data report.
[0059] In this embodiment, the real-time data includes: COD emission concentration, BOD emission concentration, SS emission concentration, animal and vegetable oil emission concentration, ammonia emission concentration, nitrogen emission concentration, etc.
[0060] Due to differences in the data types and accuracy of sensors from various companies, real-time data is preprocessed by filling a preset data template into the large language model to improve the accuracy of prompt word extraction. This method also further detects missing data, allowing companies to promptly identify and address anomalies, including sensor malfunctions and data link failures. It also helps prevent data falsification that might occur during manual data collection.
[0061] Step S204: Extract prompt words from the real-time data report based on the large language model to obtain a first prompt word set, and obtain several first knowledge fragments from the dynamically updated knowledge graph based on the first prompt word set.
[0062] Step S206: Extract prompt words from each first knowledge segment based on the large language model to obtain a second prompt word set, and merge the first prompt word set and the second prompt word set to form the optimal prompt word set.
[0063] Based on steps S204 to S206, since a dynamically updated knowledge graph is introduced during the acquisition of prompt words, the comprehensiveness and accuracy of each prompt word in the optimal prompt word set are ensured.
[0064] Step S208: The large language model uses the preset evaluation items as indexes and generates the first initial evaluation text based on the optimal prompt word set.
[0065] In this embodiment, the evaluation items include data type sub-items, pollution discharge level sub-items, and optimization scheme sub-items, each with a corresponding sub-item. The data type sub-items correspond to the types of the aforementioned real-time data.
[0066] To improve the accuracy of the large language model output, the following measures were taken during the prediction process: First, data samples were constructed using the latest pollution discharge data from various industries during model training; second, the model was periodically updated at a preset frequency. Third, a self-attention mechanism was introduced during prediction to dynamically allocate weights to various data type sub-items according to the latest environmental regulations, thereby capturing semantic relationships between different levels of abstraction.
[0067] As a preferred implementation method, combined with Figure 2 As shown, the following steps are included after step S208:
[0068] Step S20902: Input the first initial evaluation text into the dynamically updated knowledge graph to obtain several second knowledge fragments.
[0069] Step S20904: Based on the large language model, compare the first initial evaluation text and the second knowledge fragment, and when it is determined that some content in the first initial evaluation text has the opposite semantics to the corresponding content in the second knowledge fragment, replace some content in the first initial evaluation text with the corresponding content in the second knowledge fragment to obtain an updated evaluation text.
[0070] At this point, based on steps S20902 to S20904, the first initial evaluation text output by the large language model can be factually verified and corrected based on the knowledge graph, thereby further ensuring the reliability of the subsequent real-time evaluation report.
[0071] Step S210: Segment the first initial evaluation text according to the preset evaluation report template, and fill it into the corresponding position of the evaluation report template to obtain a real-time evaluation report corresponding to the real-time data report.
[0072] As a preferred implementation method, considering the regulatory needs of government environmental protection departments regarding pollution discharge, and combining... Figure 3 As shown, the following steps are included after step S210:
[0073] Step S21202: Synchronously update the real-time evaluation report to the government affairs system.
[0074] Step S21204: When it is determined that there is an abnormality in emissions, a penalty instruction is issued to the enterprise corresponding to the real-time evaluation report, and the acquisition cycle of various real-time data is adjusted according to the rectification cycle to acquire and analyze the emission data after rectification in real time.
[0075] At this point, based on steps S21202 to S21204, government departments can directly issue corresponding instructions based on the real-time evaluation report. Because the basis for issuing these instructions is more reliable, the compliance and objectivity of the instructions themselves are improved. Simultaneously, the efficiency of government processing is also improved. Furthermore, since the data acquisition cycle is adjusted in real-time according to the rectification cycle, the timeliness of tracking subsequent rectification progress is improved.
[0076] Further preferred, based on steps S21202 to S21204, combined with Figure 4 As shown, the data acquisition cycle in step S202 is also adjusted as follows:
[0077] Step S20202: Obtain the number of penalty instructions received by any enterprise within the first preset time period.
[0078] Step S20204: Dynamically adjust the acquisition cycle of various real-time data of the corresponding enterprise based on the number of instructions.
[0079] In this step, the acquisition cycle is inversely proportional to the number of instructions. Therefore, steps S20202-S20204 can improve the rationality of data acquisition frequency, avoid the consumption of computing resources, and prioritize key monitored enterprises, thus improving the rationality of resource allocation.
[0080] However, considering the need for data aggregation over a certain period of time in actual monitoring to facilitate overall control and the formulation of subsequent sewage discharge or work plans, combined with... Figure 5 As shown, the following steps are included after step S210:
[0081] Step S21222: Obtain all real-time evaluation reports of any enterprise within the second preset time period.
[0082] Step S21224: Analyze and process the real-time evaluation report based on the large language model to obtain the second initial report text.
[0083] Step S21226: Segment the second initial evaluation text according to the preset summary report template, and fill it into the corresponding position of the summary report template to obtain the evaluation summary report corresponding to all real-time evaluation reports.
[0084] At this point, based on steps S21222 to S21224, an evaluation summary report can be directly generated based on the large language model, which is not only more efficient but also more accurate. This provides a reliable basis for the planning and implementation of subsequent work.
[0085] In summary, this embodiment incorporates online monitoring, a large language model, and a dynamic knowledge graph into its environmental management method. This not only improves data monitoring efficiency but also eliminates the information gap between dynamically updated knowledge data and human knowledge reserves, effectively enhancing the reliability of the final analysis report. Furthermore, the analysis report is linked to the government system to improve the efficiency and compliance of its supervision of enterprise pollution discharge.
[0086] The aforementioned program can run in a processor or be stored in memory (or a computer-readable storage medium). Computer-readable media includes both permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media, such as modulated data signals and carrier waves.
[0087] These computer programs may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes can be implemented using different modules, and different steps can be implemented using different modules.
[0088] This embodiment also provides an environmental management system for the food industry based on a large model, combined with... Figure 6 As shown, it includes the following functional modules:
[0089] The first acquisition module is used to acquire various real-time data from each sensor according to a preset cycle and fill them into a preset data template to obtain a real-time data report.
[0090] The second acquisition module is used to extract prompt words from the real-time data report based on the large language model to obtain a first prompt word set, and to acquire several first knowledge fragments from the dynamically updated knowledge graph based on the first prompt word set.
[0091] The third acquisition module extracts prompt words from each first knowledge segment based on the large language model to obtain a second prompt word set, and merges the first prompt word set and the second prompt word set to form the optimal prompt word set.
[0092] The text generation module is used by the large language model to generate the first initial evaluation text based on the preset evaluation items and the optimal prompt word set. The evaluation items include data type sub-items, pollution level sub-items, and optimization scheme sub-items that are set one-to-one. The large language model introduces a self-attention mechanism to dynamically allocate weights to each data type sub-item according to the latest environmental regulations. The large language model is periodically iterated and updated at a preset frequency, and data samples are constructed with the latest pollution-related data of each industry during the initial training and each iteration update.
[0093] The report generation module is used to segment the first initial evaluation text according to a preset evaluation report template and fill it into the corresponding position of the evaluation report template to obtain a real-time evaluation report corresponding to the real-time data report.
[0094] Since the system is built based on the method described above, the above-described features will not be repeated here.
[0095] For example, it also includes:
[0096] The fourth acquisition module is used to input the first initial evaluation text into a dynamically updated knowledge graph to obtain several second knowledge fragments.
[0097] The text correction module is used to compare the first initial evaluation text and the second knowledge fragment based on the large language model, and when it is determined that some content in the first initial evaluation text has the opposite semantics to the corresponding content in the second knowledge fragment, it replaces some content in the first initial evaluation text with the corresponding content in the second knowledge fragment to obtain an updated evaluation text.
[0098] For example, it also includes:
[0099] The synchronization update module is used to synchronize and update the real-time evaluation report to the government affairs system.
[0100] The periodic update module is used to issue penalty instructions to the enterprise corresponding to the real-time evaluation report when abnormal emissions are detected, and to adjust the acquisition cycle of various real-time data according to the rectification cycle so as to acquire and analyze the emission data after rectification in real time.
[0101] At the same time, an electronic device is also provided, combined with Figure 7 As shown, it includes at least one processor coupled to a memory storing a computer program configured to be executed by the processor when run.
[0102] Furthermore, this embodiment also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a computer, implements the method described thereon.
[0103] Since the system, electronic devices, and storage media are all built based on the method and used to implement the method, they also have the advantages of high efficiency and high reliability of output reports in practical applications.
[0104] While the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the invention. Those skilled in the art can make various modifications and refinements without departing from the spirit and scope of the invention. Therefore, the scope of protection of the present invention shall be determined by the claims.
Claims
1. An environmental management method for the food industry based on a large-scale model, characterized in that, include: Acquire various real-time data from each sensor according to a preset cycle, and fill them into a preset data template to obtain a real-time data report; Based on the large language model, prompt words are extracted from the real-time data report to obtain a first prompt word set, and several first knowledge fragments are obtained from the dynamically updated knowledge graph based on the first prompt word set. Based on the large language model, prompt words are extracted from each first knowledge segment to obtain a second prompt word set, and the first prompt word set and the second prompt word set are merged to form the optimal prompt word set; The large language model uses preset evaluation items as indexes and generates the first initial evaluation text based on the optimal prompt word set. The evaluation items include data type sub-items, pollution level sub-items, and optimization scheme sub-items, each with a corresponding setting. The large language model introduces a self-attention mechanism to dynamically allocate weights to each data type sub-item according to the latest environmental regulations. The large language model is periodically iterated and updated at a preset frequency, and data samples are constructed using the latest pollution-related data from various industries during both the initial training and each iteration update. The first initial evaluation text is segmented according to the preset evaluation report template and filled into the corresponding position of the evaluation report template to obtain a real-time evaluation report corresponding to the real-time data report.
2. The environmental management method for the food industry based on a large model according to claim 1, characterized in that, The large language model uses preset evaluation items as indexes and generates a first initial evaluation text based on the optimal prompt word set; it includes: Input the first initial evaluation text into a dynamically updated knowledge graph to obtain several second knowledge fragments; The first initial evaluation text and the second knowledge fragment are compared based on a large language model. When it is determined that some content in the first initial evaluation text has the opposite semantics to the corresponding content in the second knowledge fragment, some content in the first initial evaluation text is replaced with the corresponding content in the second knowledge fragment to obtain an updated evaluation text.
3. The environmental management method for the food industry based on a large model according to claim 1, characterized in that, The step of segmenting the first initial evaluation text according to a preset evaluation report template and filling it into the corresponding position in the evaluation report template to obtain a real-time evaluation report corresponding to the real-time data report includes: The real-time evaluation report will be synchronized and updated to the government affairs system. When an abnormality in emissions is detected, a penalty instruction is issued to the enterprise corresponding to the real-time evaluation report, and the acquisition cycle of various real-time data is adjusted according to the rectification cycle to acquire and analyze the emission data after rectification in real time.
4. The environmental management method for the food industry based on a large model according to claim 3, characterized in that, The process of acquiring various real-time data from each sensor at a preset cycle and filling them into a preset data template to obtain a real-time data report includes: Get the number of penalty instructions received by any enterprise within the first preset time period; The acquisition cycle of various real-time data of the corresponding enterprise is dynamically adjusted based on the number of instructions; wherein, the acquisition cycle is inversely proportional to the number of instructions.
5. The environmental management method for the food industry based on a large model according to claim 1, characterized in that, include: Obtain all real-time evaluation reports for any enterprise within the second preset time period; The real-time evaluation report is analyzed and processed based on a large language model to obtain a second initial report text. The second initial evaluation text is segmented according to the preset summary report template and filled into the corresponding positions of the summary report template to obtain an evaluation summary report corresponding to all real-time evaluation reports.
6. An environmental management system for the food industry based on a large model, characterized in that, include: The first acquisition module is used to acquire various real-time data from each sensor according to a preset cycle and fill them into a preset data template to obtain a real-time data report. The second acquisition module is used to extract prompt words from the real-time data report based on the large language model to obtain a first prompt word set, and to acquire several first knowledge fragments from the dynamically updated knowledge graph based on the first prompt word set. The third acquisition module extracts prompt words from each first knowledge segment based on the large language model to obtain a second prompt word set, and merges the first prompt word set and the second prompt word set to form the optimal prompt word set; The text generation module is used by the large language model to generate the first initial evaluation text based on the preset evaluation items and the optimal prompt word set. The evaluation items include data type sub-items, pollution level sub-items, and optimization scheme sub-items, each with a corresponding setting. The large language model introduces a self-attention mechanism to dynamically allocate weights to each data type sub-item according to the latest environmental regulations. The large language model is periodically iterated and updated at a preset frequency, and data samples are constructed using the latest pollution-related data from various industries during both the initial training and each iteration update. The report generation module is used to segment the first initial evaluation text according to a preset evaluation report template and fill it into the corresponding position of the evaluation report template to obtain a real-time evaluation report corresponding to the real-time data report.
7. The environmental management system for the food industry based on a large model as described in claim 6, characterized in that, include: The fourth acquisition module is used to input the first initial evaluation text into a dynamically updated knowledge graph to obtain several second knowledge fragments; The text correction module is used to compare the first initial evaluation text and the second knowledge fragment based on the large language model, and when it is determined that some content in the first initial evaluation text has the opposite semantics to the corresponding content in the second knowledge fragment, it replaces some content in the first initial evaluation text with the corresponding content in the second knowledge fragment to obtain an updated evaluation text.
8. The environmental management system for the food industry based on a large model as described in claim 6, characterized in that, include: The synchronization update module is used to synchronize the real-time evaluation report to the government affairs system. The periodic update module is used to issue penalty instructions to the enterprise corresponding to the real-time evaluation report when abnormal emissions are detected, and to adjust the acquisition cycle of various real-time data according to the rectification cycle so as to acquire and analyze the emission data after rectification in real time.
9. An electronic device, characterized in that, It includes at least one processor coupled to a memory storing a computer program configured to be executed by the processor to perform the method of any one of claims 1-5.
10. A computer-readable storage medium, characterized in that, It stores a computer program thereon, which, when executed by a computer, implements the method described in any one of claims 1-5.