CN-121998283-A - Large-model-based enabling safety responsibility list management system, equipment and medium

CN121998283ACN 121998283 ACN121998283 ACN 121998283ACN-121998283-A

Abstract

The invention provides a security responsibility list management system, equipment and medium based on large model energization, which utilizes a large language model and is based on a knowledge service base, and allocates proper responsibility people for project tasks by combining project demands and personnel capacity; the method comprises the steps of acquiring responsibility-carrying process data and a safety responsibility knowledge graph in real time, tracking, managing and early warning task progress, carrying out multidimensional analysis and generating a visual report according to task allocation and responsibility-carrying data, generating a notification based on information of each module, accurately pushing the notification through an intelligent routing strategy to promote task cooperation, and dynamically managing user rights by means of the safety responsibility knowledge graph. The knowledge service base supports knowledge index construction, update maintenance, sample dynamic assembly and retrieval enhancement generation, so that retrieval and pushing are more efficient, flexible and accurate.

Inventors

GAO TIANHAO
Guo Kunyang
ZHANG YING
LUO XI
LIU MIN
ZHANG NAN

Assignees

北京国电通网络技术有限公司
国网信息通信产业集团有限公司

Dates

Publication Date: 20260508
Application Date: 20251205

Claims (10)

1. A large model-based enabling secure inventory management system, comprising: The system comprises a responsibility list module, a knowledge service base, a safety responsibility knowledge graph, a safety responsibility list module and a safety responsibility analysis module, wherein the responsibility list module is used for distributing each task in a project to a proper responsible person by utilizing a large language model and based on the knowledge service base and combining project requirements and personnel capability; The task management module is used for tracking, managing and generating early warning on the task progress by utilizing the safety responsibility knowledge graph based on the responsibility person responsibility performance process data acquired in real time; The data analysis and report module is used for carrying out multidimensional analysis and generating a visual report based on the task allocation information of the responsibility list module and the responsibility process data of the responsibility person of the task management module; the notification cooperation module is used for generating a notification based on the information of each module and precisely pushing the notification to related responsible persons based on the intelligent routing strategy so as to realize cooperation of tasks; and the user management module is used for managing the user permission based on the dynamic business relation of each module based on the security responsibility knowledge graph.
2. The system of claim 1, wherein the responsibility list module comprises a responsibility library sub-module for constructing and storing a security responsibility knowledge-graph; The construction of the safety responsibility knowledge graph comprises the steps of pre-training a general model by utilizing a large language model based on a large-scale safety responsibility field text, and then carrying out entity identification and relation extraction by adopting simultaneous entity identification and relation through multi-task learning.
3. The system of claim 1, wherein the responsibility list module further comprises a responsibility matching sub-module comprising: the responsible person capacity portrait construction unit is used for fusing discretization data, unstructured text data and numerical data of staff to generate a unified high-dimensional capacity vector; the task demand vectorization unit is used for converting the task description text or the task description text generated from the safety responsibility knowledge graph into a task vector; And the intelligent matching algorithm unit is used for generating a responsible person recommendation list for the task by adopting a multi-objective optimization algorithm and optimizing the capacity matching degree and the load balancing degree.
4. The system of claim 1, wherein the responsibility list module further comprises a responsible person display sub-module configured to: Monitoring a relation change event related to responsibility in the safety responsibility knowledge graph; and responding to the change event, dynamically inquiring map data, and updating a display interface of responsibility allocation and task progress in real time through a visualization component from the view point of an administrator and the view point of an employee respectively.
5. The system of claim 1, wherein the responsibility list module further comprises a responsibility update sub-module comprising: a change identification unit for automatically identifying responsibility change points from the project document and the system instruction through a text difference technology and an information extraction technology; The influence analysis unit is used for carrying out path query and logic rule reasoning based on the safety responsibility knowledge graph and analyzing the influence range of the change point; And the updating unit is used for generating a knowledge graph updating scheme according to the influence analysis result and updating the security responsibility knowledge graph through an approvable workflow.
6. The system of claim 1, wherein the task management module comprises a progress tracking sub-module comprising: the progress acquisition unit is used for acquiring responsibility performing process data from a plurality of channels of text reports, space-time data and visual data, and analyzing and extracting features by using an NLP model, space-time verification rules and a computer visual model respectively; The system comprises a progress calculation unit, an automatic early warning sub-module, a control unit and a control unit, wherein the progress calculation unit is used for comprehensively judging a multisource analysis result by adopting an evidence fusion algorithm, determining the actual progress state of a task and automatically calculating the progress percentage, and the automatic early warning sub-module is configured to: when the progress calculation unit judges that the task progress is lower than a preset threshold or exceeds the time limit, an early warning message is automatically triggered and pushed to related responsible persons and management staff through the notification cooperation module.
7. The system of claim 1, wherein the data analysis and reporting module comprises a performance assessment sub-module configured to: calculating a multi-dimensional performance index value based on the related data of each module by using an assessment index system so as to obtain performance assessment on the project and/or responsible persons, wherein the multi-dimensional performance comprises the project progress and/or the task completion rate, the completion quality and the behavior of the responsible persons; And displaying the performance evaluation result in one or more visualization forms of a radar chart, a comparison histogram and a trend line chart.
8. The system of claim 1, wherein the notification collaboration module comprises: the message generating unit is used for selecting a corresponding template from a preset template library according to the type of the system event, filling event parameters through a rule or a lightweight large model and generating a personalized notification message; And the message distribution unit is used for distributing the message by adopting one or more channel combinations of intra-system message, mobile APP pushing, short message and mail according to the emergency degree of the message and the preference of the receiver.
9. The electronic equipment is characterized by comprising at least one processor and a memory, wherein the memory and the processor are connected through a bus; The memory is used for storing one or more programs; a security responsibility list management system based on big model enablement as recited in any of claims 1-8 is performed when the one or more programs are executed by the at least one processor.
10. A readable storage medium having stored thereon an execution program which, when executed, implements a security responsibility list management system based on large model enablement as claimed in any one of claims 1 to 8.

Description

Large-model-based enabling safety responsibility list management system, equipment and medium Technical Field The invention relates to artificial intelligence, algorithm and natural language processing, in particular to a security responsibility list management system, equipment and medium based on large model energization. Background The safety management is a core link of modern enterprise operation, and the safety responsibility list is a core tool for definitely and realizing the safety responsibilities of all levels of posts, and comprises five links of clear responsibilities, known responsibilities, functional responsibilities and mutual promotion, and all the links are mutually connected to form a complete chain of a safety production responsibility system. And (3) establishing responsibility systems of each unit by each unit of the company according to national relevant safety production laws and regulations and relevant regulations of the company, and defining responsibility and obligation of management staff and posts at all levels to safety production, ensuring that each responsibility is responsible for a special person and strengthening the self safety responsibility consciousness of the whole person. The responsibility is known, namely the uniformity and the standardization of the implementation of the safety production responsibility of each unit are clarified, and the safety responsibility of the self-post is practically implemented by combining training and daily learning from the aspects of inventory declaration, inventory publicity, inventory training, inventory examination, external units and the like, deepening the awareness of the safety responsibility system of the whole person, standardizing the explanation understanding of the legal laws of the regulations and the like. The responsibility performing is realized that the safety responsibility corresponding to the positions of the whole staff is put in place, and the staff at all levels are urged to fulfill the respective safety responsibility by means of safety work deployment implementation, on-duty in place, supervision and inspection, on-site disposal, communication report, assessment and evaluation and the like. And uploading and archiving the implementation result by combining the meeting summary and the work log, so that records can be checked. Supervision, namely supervising the safety responsibilities corresponding to all posts to be in place, applying means such as inspection, safety production inspection, "four-not-two-straight" supervision, special inspection, daily inspection and the like, and supervising the enterprises to establish a sound safety production responsibilities implementation supervision evaluation mechanism and supervising personnel at all levels to implement safety production responsibilities. The accountability is that each level of the company implements the security accountability pursuit system of the superior level unit to the subordinate level unit, and the principle, situation, object, procedure, mode, standard and the like of accountability are defined, and the accountability is reasonably and legally determined according to the factors of security accountability, accountability conditions and the like by combining with the security accountability system of the whole person. However, existing security responsibility list management relies mainly on manual formulation, static documents (e.g., excel, word) and offline communication procedures. There are significant limitations to this approach. The making process is low in efficiency and high in subjectivity, the making of the list is seriously dependent on personal experience of management staff, data-driven decision support is lacked, comprehensiveness, rationality and scientificity of responsibility allocation are difficult to ensure, dynamic adaptability is poor, project environments and risks are dynamically changed, traditional lists are slowly updated, automatic adjustment and optimization cannot be carried out according to real-time data (such as historical accident data and real-time monitoring information), and responsibility clauses are disjointed with actual conditions. The monitoring and tracing are difficult, the monitoring of the responsibility performance condition is usually reported manually, the efficiency is low, and the information lag or distortion is easy to occur. The full chain of "explicit, known, fun, du" is difficult to form into transparent, traceable records. Insufficient level of intelligence, lack of ability to analyze inventory data deeply, inability to intelligently predict potential risk, provide optimization advice, or generate effective decision support reports. Although some informative systems have been introduced in an attempt to solve the above problems, these systems tend to be single-functioning, lack true intelligence, fail to understand the content of responsibility deeply, automatically check the rationa