A safe production management method and system based on a large model and a multi-tenant architecture

Abstract

The application provides a kind of based on big model and the security production management method and system of multi-tenant architecture, it is related to production management technical field, method includes: based on enterprise organization structure information establishes multi-tenant organization model, determines each tenant identification and corresponding role permission;Under the constraint of model, obtain each sub-tenant safety production management data, and bind tenant identification for data;Each sub-tenant safety production data is processed to knowledge, and knowledge vector is generated;Under the premise of meeting multi-tenant isolation rules and role permissions, cross-tenant data access view is generated through parent company summary access mechanism;Based on data access view, the centralized intelligent analysis summary of each sub-tenant knowledge vector is carried out;Based on the summary result, the open source big model accessed is called to generate the safety production management strategy for each sub-tenant and issue;Each sub-tenant executes strategy and feeds back the execution result to parent company;Parent company analyzes feedback and updates safety production management strategy and strategy template library.

Description

Technical Field

[0001] This invention relates to the field of production management technology, and in particular to a safety production management method and system based on a large model and multi-tenant architecture. Background Technology

[0002] Safety production management is a crucial component of enterprise production and operation activities, involving multiple stages such as risk and hazard identification, control measure formulation, implementation supervision, and effectiveness evaluation. With the continuous expansion of enterprise scale and the prevalence of group-based operation models, safety production management is gradually exhibiting characteristics such as multiple organizational levels, dispersed management objects, and complex data types. To improve the standardization, digitalization, and refinement of safety production management, more and more enterprises are beginning to centrally manage and analyze safety-related data through information systems.

[0003] In recent years, large-scale modeling technology has demonstrated strong capabilities in natural language understanding, knowledge reasoning, and complex information processing, providing new technical means for scenarios such as policy text parsing, risk information analysis, and management decision support in the field of safety production. Meanwhile, multi-tenant architecture, as an information system architecture pattern that supports parallel use by multiple organizations, can achieve data isolation and access control between different organizations within the same system, and is increasingly being applied to the safety production management systems of large-scale enterprises. Therefore, how to effectively combine large-scale modeling technology with multi-tenant architecture to support large-scale safety production management has become one of the key technical directions of interest in this field.

[0004] However, existing technologies often rely on simple invocation of large-scale models, lacking effective linkage mechanisms with safety production execution feedback data. This makes it difficult to dynamically adjust model outputs based on actual execution results, easily leading to mismatches between generated safety production management strategies and specific tenant situations, thus impacting safety production management effectiveness. Furthermore, existing safety production management methods lack dynamic analysis tools for the safety production execution process and results, making it difficult to form a continuous optimization loop based on execution feedback. Safety production management strategies often depend on manual experience for adjustment, resulting in limited overall intelligence. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the purpose of this invention is to provide a safety production management method based on a large model and multi-tenant architecture. This method addresses the shortcomings of existing technologies where large model applications often remain at a simple invocation level, lacking an effective linkage mechanism with safety production execution feedback data. The model output results are difficult to dynamically adjust based on actual execution effects, easily leading to a mismatch between the generated safety production management strategy and the specific tenant's actual situation, thus affecting the effectiveness of safety production management. Furthermore, existing safety production management methods lack dynamic analysis tools for the safety production execution process and results, making it difficult to form a continuous optimization closed loop based on execution feedback. Safety production management strategies often rely on manual experience for adjustment, resulting in limited overall intelligence.

[0006] A first aspect of this invention proposes a security production management method based on a large model and multi-tenant architecture, comprising: S1: Based on the enterprise's organizational structure information, establish a multi-tenant organizational model and determine the corresponding role permissions; S2: Under the constraints of the multi-tenant organization model, obtain the safety production management data of each sub-tenant and bind the corresponding tenant identifier to the safety production management data; S3: Perform knowledge processing on the safety production data of each subtenant to generate a knowledge vector for each subtenant; S4: Under the premise of satisfying the multi-tenant data isolation rules and the aforementioned role permissions, generate a cross-tenant data access view of the parent company through the parent company's aggregated access mechanism; S5: Based on the data access view, perform centralized intelligent analysis and summarization of the knowledge vectors of each sub-tenant to generate a summary result of safety production analysis at the parent company level; S6: Based on the summarized results of the safety production analysis, generate safety production management strategies for each of the sub-tenants through the accessed open-source large model, and distribute them to the corresponding sub-tenants; S7: Each sub-tenant executes the corresponding safety production management strategy and feeds back the execution results to the parent company; S8: The parent company analyzes the execution feedback of each of the sub-tenants and updates the safety production management strategy and strategy template library based on the analysis results.

[0007] A second aspect of this invention provides a security production management system based on a large model and multi-tenant architecture, comprising: a processor and a memory; The memory stores programs or instructions that can run on the processor, which, when executed by the processor, implement the steps of the secure production management method based on a large model and multi-tenant architecture as described in the first aspect.

[0008] A third aspect of the present invention provides a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the security production management method based on a large model and multi-tenant architecture as described in the first aspect.

[0009] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this embodiment of the invention, a closed-loop optimization mechanism is constructed, centered on centralized analysis of knowledge vectors, generation of large-scale model strategies, collection of execution feedback, and updating of strategies. This enables the application of large-scale models to move beyond static invocation and achieve linkage with the safe production execution process. Through centralized intelligent analysis of cross-tenant safe production data and knowledge vectors, combined with the generation of safe production management strategies for sub-tenants by the large-scale model, and continuous feedback and analysis of strategy execution results, the safe production management strategies are dynamically adjusted and iteratively optimized based on actual execution effects. This reduces reliance on manual experience adjustments and avoids mismatches between generated strategies and the actual production conditions of different tenants, thereby significantly improving the intelligence level and actual management effectiveness of safe production management. Attached Figure Description

[0010] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0011] Figure 1 This is a flowchart illustrating a security production management method based on a large model and multi-tenant architecture provided by an embodiment of the present invention.

[0012] Figure 2 This is a schematic diagram of a safety production management system based on a large model and multi-tenant architecture provided by an embodiment of the present invention. Detailed Implementation

[0013] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0014] The following description, in conjunction with the accompanying drawings, details the security production management method based on a large model and multi-tenant architecture provided by the embodiments of the present invention through specific implementations and application scenarios.

[0015] Reference manual attached Figure 1 The diagram illustrates a process flow diagram of a security production management method based on a large model and multi-tenant architecture provided by an embodiment of the present invention.

[0016] This invention provides a secure production management method based on a large model and multi-tenant architecture, which may include the following steps: S1: Based on the enterprise's organizational structure information, establish a multi-tenant organizational model and determine the corresponding role permissions.

[0017] The multi-tenant organization model is a data and permission organization model used to describe multiple organizational entities operating in parallel within the same information system, sharing system resources but remaining isolated from each other. In this model, different tenants are logically independent of each other, each corresponding to a different organizational entity. The system achieves precise management of each tenant's data access and business operations through a unified tenant identifier and permission control mechanism.

[0018] Among them, role permissions are a set of permission rules that define which tenant data a "parent company / sub-tenant / different positions" can access, which functional modules they can use, and whether they can execute cross-tenant summary analysis and policy distribution in a multi-tenant security production system.

[0019] In one possible implementation, S1 specifically includes: S101: Parse the enterprise's organizational structure information, identify the parent company and multiple tenants, and generate a unique tenant identifier for each tenant.

[0020] S102: Utilize the organizational affiliation relationships in the enterprise's organizational structure information to establish a multi-tenant organizational model and record the control relationship types between tenants.

[0021] S103: Based on the multi-tenant organization model, configure a role identifier for each tenant and associate the role identifier with the corresponding functional permissions and data access permissions.

[0022] S104: Bind the role identifier to the specific user, and use the tenant identifier and role identifier as the permission constraints for subsequent access to and processing of secure production data.

[0023] Specifically, the system first establishes a multi-tenant organizational model based on the enterprise's organizational structure information and determines the corresponding role permissions. Specifically, the system parses the enterprise's organizational structure information, identifies the parent company and multiple subordinate organizational units, and models each organizational unit as a tenant in the system. A unique tenant identifier is generated for each tenant for subsequent data isolation and access control. Further, utilizing the organizational hierarchy reflected in the enterprise's organizational structure information, parent-child dependencies between tenants are established, and the control relationship types between tenants are recorded to clarify the management hierarchy between the parent company and each sub-tenant. After completing the tenant structure modeling, the system configures at least one role identifier for each tenant based on the multi-tenant organizational model and associates the role identifier with the system's functional permissions and data access permissions, enabling different tenants and their users to have differentiated operational capabilities and data access scopes. Finally, the role identifier is bound to a specific user, and the tenant identifier and role identifier serve as permission constraints in the subsequent collection, storage, retrieval, and analysis of safety production data. This ensures that data and configurations between different tenants do not interfere with each other when multiple tenants share system resources, and provides a foundation for the parent company's subsequent cross-tenant aggregated access.

[0024] It should be noted that implementing safe production management practices through a multi-tenant architecture allows multiple tenants to share resources while ensuring that data and configurations do not interfere with each other. To address the issues of user data selection and storage under multi-tenant isolation, this invention adopts a strategy of splitting role-based control permissions. This strategy makes the selection and storage of user data (parent company, branch office, subsidiary) more efficient under a multi-tenant architecture, avoiding data mixing.

[0025] S2: Under the constraints of the multi-tenant organization model, obtain the safety production management data of each sub-tenant and bind the corresponding tenant identifier to the safety production management data.

[0026] Specifically, each subtenant collects its own safety production business data. This data undergoes cleaning and format standardization to generate standard safety production business data that conforms to data exchange standards. Tenant identifiers, business type identifiers, and timestamp information are then bound to each standard safety production business data set.

[0027] S3: Perform knowledge processing on the safety production data of each subtenant to generate a knowledge vector for each subtenant.

[0028] It should be noted that knowledge processing means "processing" safety production ledgers / systems / records into knowledge fragments that can be retrieved and referenced by large models, and then establishing indexes or vector representations.

[0029] In one possible implementation, S3 specifically includes: S301: Parse and process the safety production data of each subtenant to extract structured intermediate data.

[0030] Structured intermediate data refers to the data extracted from unstructured text / files after parsing raw safety production data (such as PDF policy documents, hazard investigation records, training forms, emergency plans, etc.). This extracted information is then organized into directly calculable and storable data objects according to a unified data format, serving as a transitional result for subsequent vectorized modeling and intelligent analysis. This data typically contains clearly defined fields and labels, such as tenant identifier, business type, document source, risk point name, hazard description, rectification measures, responsible person, planned completion time, actual completion time, and risk level.

[0031] Specifically, text extraction or OCR recognition is performed, and the text content is segmented into chapters and clauses. Further, based on preset field templates, key fields such as risk points, hazard descriptions, rectification measures, responsible parties, planned completion time, actual completion time, and risk level are extracted from the segmented text blocks. For structured or semi-structured data such as Excel ledgers, header recognition, field mapping, format standardization, and data cleaning are performed to generate structured records consistent with preset fields.

[0032] S302: Employs the Embedding model to vectorize various structured data and generate knowledge vectors related to business semantics.

[0033] Among them, the Embedding model is a machine learning model used to convert unstructured information such as text, sentences or documents into low-dimensional dense vector representations. Its core function is to map content with similar semantics to similar positions in the vector space, so that computers can achieve semantic-level retrieval and matching through vector similarity calculation.

[0034] In this embodiment of the invention, the Embedding model is used to vectorize the structured intermediate data formed by the knowledge-based processing of the safety production data of each subtenant, generate corresponding knowledge vectors, and construct the subtenant knowledge vector space. This enables the parent company to quickly locate content semantically related to the safety production analysis request based on vector retrieval under permission constraints, providing a reliable data foundation for subsequently inputting the retrieval context into the large model for risk assessment and strategy generation.

[0035] S4: Under the premise of meeting the multi-tenant data isolation rules and role permissions, generate the parent company's cross-tenant data access view through the parent company's aggregated access mechanism.

[0036] The aggregated access mechanism refers to a controlled access mechanism by which a parent company tenant can uniformly read, organize, and analyze data from multiple sub-tenants under multi-tenant data isolation and role-based access control. Its essence is not to merge and store sub-tenant data, but to achieve controlled aggregated access across tenants through permission calculation and access view generation.

[0037] In a multi-tenant architecture, the data access view is a logical data access layer built to enable unified access to multiple data sources (such as data spaces of different subtenants). It doesn't physically merge the data of each subtenant into the same database; instead, based on tenant identifiers, role permissions, and data isolation rules, it calculates and organizes the range of data accessible to the parent company, forming a unified entry point that is "queryable, searchable, and statistically analyzeable." Through the data access view, the parent company can perform cross-tenant reading and summary analysis of security production data from multiple subtenants under permission constraints.

[0038] In one possible implementation, S4 specifically includes: S401: Based on the parent-child dependency relationship of tenants in the multi-tenant organization model, determine the range of child tenant data that the parent company role can access.

[0039] S402: Configure the parent company's role permissions separately, mapping the accessible data permissions to the corresponding sub-tenant's data set.

[0040] S403: Based on each data set, organize it according to unified data access rules to generate a data access view based on permission rules.

[0041] In this embodiment of the invention, the above-mentioned aggregated access mechanism generates a cross-tenant data access view of the parent company, which enables the parent company to achieve controlled aggregated access to the secure production data of multiple sub-tenants while ensuring multi-tenant data isolation and role-based access control. This avoids data duplication and mixing, improves cross-tenant analysis efficiency and strategy generation efficiency, and enhances the system's security, traceability, and scalability.

[0042] In one possible implementation, the process after S4 includes: Through access control mechanisms, real-time permission verification is performed on data access views, and based on the permission verification results, the parent company's access behavior to data access views is controlled, allowing cross-tenant data access operations only when the permission verification passes.

[0043] It should be noted that an access control mechanism refers to a set of control rules and execution processes that a system uses to verify a user's identity, role permissions, and tenant scope when the user accesses system functions or data resources in order to ensure data security and permission compliance, and decide whether to "allow access, restrict access, or deny access" accordingly.

[0044] In this embodiment of the invention, the access control mechanism performs real-time permission verification on the parent company's access behavior, which upgrades cross-tenant access from "static authorization" to a controlled access mode of "dynamic verification". This allows the access results to be updated in a timely manner when the parent company's user roles, tenant relationships or permission configurations change, avoiding the risk of unauthorized access or data leakage caused by delayed permission changes.

[0045] S5: Based on the data access view, perform centralized intelligent analysis and summarization of the knowledge vectors of each subtenant to generate a summary of safety production analysis results at the parent company level.

[0046] In one possible implementation, S5 specifically includes: S501: Convert the safety production analysis request initiated by the parent company into a vectorized query representation.

[0047] S502: Under the constraints of the data access view, retrieve relevant data from the knowledge vector space of each subtenant based on the vectorized query representation.

[0048] S503: Reorganize the retrieved data according to business relevance to form a unified analytical context.

[0049] S504: Input the analysis context into the safety production big model to generate risk assessment results for each tenant.

[0050] Among them, the safety production big model is an industry big model for safety production management formed on the basis of general open source big models (such as big language models), combined with professional knowledge such as policies, regulations, rules and regulations, operating procedures, risk and hazard cases, and emergency plans in the field of safety production, and adapted to safety production business scenarios.

[0051] S505: Summarize the results of each risk assessment to obtain the summary results of the safety production analysis.

[0052] Optionally, the summary results of the safety production analysis include the risk assessment results, risk level or risk score, identification results of major risk points, summary of hidden danger investigation and management (including the number of hidden dangers, rectification progress and overdue status), risk change trends and early warning prompts for each sub-tenant.

[0053] Specifically, after the parent company initiates a safety production analysis request in the system, the system first vectorizes the request to obtain a corresponding vectorized query representation. Then, under the constraints of the data access view, the system retrieves semantically relevant data from the knowledge vector space of each sub-tenant based on the vectorized query representation. This retrieval process is limited to the sub-tenant data authorized for access by the parent company, ensuring multi-tenant data isolation and access compliance. After completing the relevant data retrieval, the system further filters, sorts, and reorganizes the retrieval results according to business relevance, forming a unified analysis context. This analysis context is then input into the large-scale safety production model, enabling the model to generate risk assessment results for each sub-tenant. Finally, the system summarizes and analyzes the risk assessment results of each sub-tenant to obtain a summary safety production analysis result at the parent company level, providing data support and decision-making basis for the subsequent generation and distribution of safety production management strategies.

[0054] In this embodiment of the invention, the S5 step described above, which uses a data access view to perform centralized intelligent analysis and aggregation of the knowledge vectors of each sub-tenant, enables the parent company to achieve unified retrieval and comprehensive assessment of safety production information from multiple sub-tenants, while meeting the requirements of multi-tenant data isolation and access constraints. This avoids the analytical delays and information omissions caused by the dispersion of data across systems and departments in traditional methods. Simultaneously, by using vectorized queries and knowledge vector retrieval to locate relevant data at the semantic level, and reorganizing the retrieval results into an analytical context input into the large-scale safety production model, the accuracy and interpretability of risk assessment can be significantly improved, reducing the workload and subjective bias of manual aggregation and analysis. This, in turn, enhances the parent company's centralized identification capability and decision-making efficiency regarding safety production risks.

[0055] S6: Based on the summary results of safety production analysis, it generates safety production management policies for each sub-tenant through the accessed open-source large model and distributes them to the corresponding sub-tenants.

[0056] It should be noted that this invention provides additional information input to the safety production knowledge base by accessing an open-source large model instead of maintaining a traditional knowledge base, greatly improving the timeliness of the knowledge base.

[0057] Specifically, the system first structures the results of the safety production analysis, compiling the risk assessment results, major risk point lists, hazard management status, risk trends, and early warning information of each subtenant into policy generation input data, and establishing a "tenant-risk characteristic" mapping relationship based on tenant identifiers. Then, the system calls the large-scale model integration service, combining the policy generation input data with preset safety production policy templates, management system constraints, and policy generation rules to construct prompts. These prompts are then input into the integrated open-source large-scale model, enabling the model to generate safety production management policy content that matches the risk characteristics of each subtenant.

[0058] Optionally, a safety production management strategy should include at least the following strategic elements: rectification objectives, control measures, responsible parties, implementation cycle, and acceptance standards.

[0059] S7: Each subtenant executes the corresponding safety production management strategy and feeds back the execution results to the parent company.

[0060] In one possible implementation, S7 specifically includes: S701: Each subtenant receives the safety production management policy issued by the parent company, parses the safety production management policy, and obtains the corresponding execution requirements.

[0061] S702: Based on the execution requirements, each subtenant executes the corresponding safety production management strategy and generates execution process data corresponding to the safety production management strategy.

[0062] S703: Based on execution process data, calculate the policy applicability score for each sub-tenant's execution of the corresponding safety production management policy: in, Indicates the strategy applicability score. This indicates the weight of the execution efficiency coefficient. This represents the execution efficiency coefficient. This indicates the weight of the cost adaptation coefficient. Indicates the cost fit coefficient. This indicates that the risk reduction coefficient weighting is applied. This represents the risk reduction coefficient.

[0063] in, Represents an exponential function. This represents the attenuation coefficient, which is set by technicians based on experience. This indicates taking the maximum value. Indicates the actual completion time. Indicates the planned completion time.

[0064] in, Represents the hyperbolic tangent function. Represents absolute value. Indicates actual cost, This indicates the budget cost.

[0065] in, This indicates the risk value before implementation. This indicates the risk value after implementation. Indicates the number of significant risk points involved. This indicates the total number of risk points.

[0066] It should be noted that, .

[0067] S704: Based on the policy applicability score, determine the execution status of each subtenant in implementing the corresponding safety production management policy, and generate execution status information and execution result data.

[0068] S705: Bind the execution process data, execution status information, and execution result data with the corresponding policy identifier and tenant identifier, and feed them back to the parent company.

[0069] In this embodiment of the invention, after receiving and parsing the safety production management strategy issued by the parent company, each sub-tenant can implement the strategy according to unified execution requirements and generate traceable execution process data during the execution process. At the same time, by introducing a strategy applicability scoring mechanism based on execution efficiency, cost adaptability, and risk reduction, the effectiveness of strategy execution can be quantitatively evaluated, thereby improving the closed-loop control capability and decision-making efficiency of group-wide safety production management.

[0070] S8: The parent company analyzes the execution feedback of each sub-tenant and updates the safety production management strategy and strategy template library based on the analysis results.

[0071] The strategy template library is a data collection that stores safety production management strategy templates. The strategy templates include strategy structure fields, strategy generation rules, applicable conditions, and parameter configurations. The system matches and parameterizes the risk characteristics of different sub-tenants based on the strategy template library to generate corresponding safety production management strategies.

[0072] In one possible implementation, S8 specifically includes: S801: The parent company receives execution process data, execution status information, execution result data, and corresponding strategy applicability scores from each sub-tenant.

[0073] S802: Strategies with applicability scores greater than or equal to the first scoring threshold are classified as generalizable strategies. Strategies with applicability scores greater than or equal to the second scoring threshold but less than the first scoring threshold are classified as strategies requiring localization adjustments. Strategies with applicability scores less than the second scoring threshold are classified as ineffective strategies.

[0074] It should be noted that those skilled in the art can set the size of the first and second scoring thresholds according to actual needs, and this invention does not limit this.

[0075] S803: Store the safety production management strategy corresponding to the promotable strategy into the parent company's strategy template library and mark it as the first safety production management strategy that can be reused by similar tenants.

[0076] S804: For safety production management strategies that require localization adjustment, adjust parameters based on the execution process data and risk characteristics of the corresponding sub-tenant to generate an adjusted second safety production management strategy.

[0077] S805: For the safety production management strategy corresponding to the failure strategy, based on the execution feedback data, trigger the update processing of the open source big model to generate a third safety production management strategy.

[0078] S806: Record the first safety production management strategy, the second safety production management strategy, and the third safety production management strategy into the strategy template library.

[0079] In this embodiment of the invention, the parent company can uniformly analyze and classify the execution effect of safety production management strategies based on the feedback data and strategy applicability scores received from each sub-tenant, thereby transforming the strategy effect from manual experience-based judgment to an objective decision-making mechanism based on threshold rules. For strategies with high scores that can be promoted, they are stored in a strategy template library to achieve strategy accumulation and cross-tenant reuse, improving the standardization level of group-wide safety production management. For strategies that require localization adjustments, parameter adjustments are made in conjunction with sub-tenant execution process data and risk characteristics, so that the strategies can better fit the actual production environment of different tenants, improving strategy adaptability and execution effect. For failed strategies, new strategies are generated by triggering the update processing of the open-source large model, realizing continuous optimization and iterative evolution of strategy generation capabilities. At the same time, recording different types of strategies uniformly in the strategy template library is conducive to forming a traceable, reusable, and iterative strategy asset system, thereby enhancing the parent company's closed-loop control capability and decision-making efficiency for multi-tenant safety production management.

[0080] Reference manual attached Figure 2 The diagram shows a structural schematic of a safety production management system based on a large model and multi-tenant architecture provided by an embodiment of the present invention.

[0081] This invention provides a security production management system 20 based on a large model and multi-tenant architecture, including: a processor 201 and a memory 202; The memory 202 stores programs or instructions that can run on the processor 201. When the program or instructions are executed by the processor 201, they implement the steps of the above-mentioned security production management method based on large model and multi-tenant architecture and achieve the same technical effect. To avoid repetition, the present invention will not elaborate further.

[0082] It should be understood that the processor 201 in this embodiment of the invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0083] It should also be understood that the memory 202 in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DR RAM).

[0084] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0085] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0086] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0087] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0088] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0089] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0090] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0091] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0092] This invention provides a readable storage medium comprising: storing a program or instructions on the readable storage medium, wherein when the program or instructions are executed by a processor, the program or instructions implement the steps of the above-described secure production management method based on a large model and multi-tenant architecture, and can achieve the same technical effect. To avoid repetition, this invention will not elaborate further.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, and are not intended 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 of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the protection scope of the present invention.

Claims

1. A safety production management method based on a large model and multi-tenant architecture, characterized in that, include: S1: Based on the enterprise's organizational structure information, establish a multi-tenant organizational model and determine the corresponding role permissions; S2: Under the constraints of the multi-tenant organization model, obtain the safety production management data of each sub-tenant and bind the corresponding tenant identifier to the safety production management data; S3: Perform knowledge processing on the safety production data of each subtenant to generate a knowledge vector for each subtenant; S4: Under the premise of satisfying the multi-tenant data isolation rules and the aforementioned role permissions, generate a cross-tenant data access view of the parent company through the parent company's aggregated access mechanism; S5: Based on the data access view, perform centralized intelligent analysis and summarization of the knowledge vectors of each sub-tenant to generate a summary result of safety production analysis at the parent company level; S6: Based on the summarized results of the safety production analysis, generate safety production management strategies for each of the sub-tenants through the accessed open-source large model, and distribute them to the corresponding sub-tenants; S7: Each sub-tenant executes the corresponding safety production management strategy and feeds back the execution results to the parent company; S8: The parent company analyzes the execution feedback of each of the sub-tenants and updates the safety production management strategy and strategy template library based on the analysis results.

2. The safety production management method based on large model and multi-tenant architecture according to claim 1, characterized in that, S1 specifically includes: S101: Parse the enterprise organizational structure information, identify the parent company and multiple tenants, and generate a unique tenant identifier for each tenant; S102: Using the organizational affiliation relationships in the enterprise organizational structure information, establish the multi-tenant organizational model and record the control relationship types between tenants; S103: Based on the multi-tenant organization model, configure a role identifier for each tenant, and associate the role identifier with corresponding functional permissions and data access permissions; S104: Bind the role identifier to a specific user, and use the tenant identifier and the role identifier as permission constraints for subsequent access to and processing of secure production data.

3. The safety production management method based on a large model and multi-tenant architecture according to claim 1, characterized in that, S3 specifically includes: S301: Parse and process the safety production data of each subtenant to extract structured intermediate data; S302: Using the Embedding model, the structured data is vectorized to generate knowledge vectors related to business semantics.

4. The safety production management method based on large model and multi-tenant architecture according to claim 1, characterized in that, S4 specifically includes: S401: Based on the parent-child dependency relationship of tenants in the multi-tenant organization model, determine the range of sub-tenant data that the parent company role can access; S402: Configure the parent company's role permissions separately, mapping the accessible data permissions to the corresponding subtenant's data set; S403: Based on each of the aforementioned data sets, organize them according to unified data access rules to generate a data access view based on permission rules.

5. The safety production management method based on large model and multi-tenant architecture according to claim 1, characterized in that, Following S4, the following is also included: The access control mechanism performs real-time permission verification on the data access view and controls the parent company's access behavior to the data access view based on the permission verification results, allowing cross-tenant data access operations only when the permission verification passes.

6. The safety production management method based on large model and multi-tenant architecture according to claim 1, characterized in that, S5 specifically includes: S501: Convert the safety production analysis request initiated by the parent company into a vectorized query representation; S502: Under the constraints of the data access view, relevant data is retrieved from the knowledge vector space of each subtenant based on the vectorized query representation; S503: Reorganize the retrieved data according to business relevance to form a unified analytical context; S504: Input the analysis context into the safety production big model to generate risk assessment results for each tenant; S505: Summarize the results of each risk assessment to obtain the summary results of the safety production analysis.

7. The safety production management method based on large model and multi-tenant architecture according to claim 1, characterized in that, Specifically, S7 includes: S701: Each of the subtenants receives the safety production management policy issued by the parent company, parses the safety production management policy, and obtains the corresponding execution requirements; S702: Based on the execution requirements, each sub-tenant executes the corresponding safety production management strategy and generates execution process data corresponding to the safety production management strategy; S703: Based on the execution process data, calculate the policy applicability score of the corresponding safety production management strategy executed by each sub-tenant; S704: Based on the applicability score of the strategy, determine the execution status of each sub-tenant in implementing the corresponding safety production management strategy, and generate execution status information and execution result data; S705: Bind the execution process data, execution status information, and execution result data with the corresponding policy identifier and tenant identifier, and feed them back to the parent company.

8. The safety production management method based on large model and multi-tenant architecture according to claim 1, characterized in that, S8 specifically includes: S801: The parent company receives execution process data, execution status information, execution result data, and corresponding strategy applicability scores from each of the sub-tenants; S802: Among the various strategy applicability scores, those that are greater than or equal to the first scoring threshold are determined to be generalizable strategies; those that are greater than or equal to the second scoring threshold and less than the first scoring threshold are determined to be strategies that require localization adjustment; and those that are less than the second scoring threshold are determined to be ineffective strategies. S803: Store the safety production management strategy corresponding to the promotable strategy into the strategy template library of the parent company and mark it as the first safety production management strategy that can be reused by similar tenants. S804: For the safety production management strategy corresponding to the localized adjustment strategy, adjust the parameters based on the execution process data and risk characteristics of the corresponding sub-tenant to generate the adjusted second safety production management strategy; S805: For the safety production management strategy corresponding to the failure strategy, based on the execution feedback data, trigger the update processing of the open source big model to generate a third safety production management strategy; S806: Record the first safety production management strategy, the second safety production management strategy, and the third safety production management strategy into the strategy template library.

9. A safety production management system based on a large model and multi-tenant architecture, characterized in that, include: Processor and memory; The memory stores programs or instructions that can run on the processor, and when the programs or instructions are executed by the processor, they implement the steps of the security production management method based on a large model and multi-tenant architecture as described in any one of claims 1 to 8.

10. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions, which, when executed by a processor, implement the steps of the secure production management method based on a large model and multi-tenant architecture as described in any one of claims 1 to 8.

Images