Safety production hidden danger troubleshooting management system

Abstract

The invention relates to the technical field of safety production management and informatization, and discloses a safety production hidden danger checking management system which comprises a hidden danger description AI regulation matching module, a hidden danger checking input module, a rectification closed-loop management module, a multi-subject data collaboration module, a risk early warning analysis module and a report generation module. According to the method, a vector database constructed by combining a Tonyuan large model with a law and regulation library, a standard library and a case library is adopted, a law and regulation basis corresponding to hidden danger problem description is automatically matched based on a retrieval enhancement generation technology, the matching priority is that the case library > the law and regulation library > the standard library, and manual adjustment is supported; real-time data sharing among a security service mechanism, an insurance buying enterprise and an insurance mechanism is realized based on a unified data platform and a role authority control mechanism; generating a risk analysis chart through the data cockpit; according to the invention, the reliability of accident prevention is ensured, the continuity of production safety is enhanced, and the active protection capability of safety production is ensured.

Description

Technical Field

[0001] This invention relates to the field of safety production management and information technology, specifically a safety production hazard investigation and management system. Background Technology

[0002] Safety production management refers to the management and control of safety production work, while informatization refers to the historical process of cultivating and developing new productive forces represented by intelligent tools, mainly computers, and making them benefit society.

[0003] With the introduction of policies such as the "Supervision and Management Measures for Accident Prevention Technical Services of Safety Production Liability Insurance Underwriting Institutions of Guangdong Provincial Construction Engineering Group Co., Ltd.", which require "giving full play to the advantages of the safety technology center and promoting the entrustment of accident prevention services by the underwriting institution", the digital transformation of safety production hazard investigation, as the core link of accident prevention, is urgently needed. Traditional hazard investigation relies on manual operation, with fragmented processes and a lack of a unified management platform, resulting in low service efficiency and difficulty in tracing responsibility. In the current safety production hazard investigation work, there are common problems such as low efficiency of manual recording, reliance on experience to match regulations and standards, untimely tracking of rectification and closure, serious data silos among multiple entities, and delayed risk warning.

[0004] Therefore, a safety production hazard investigation and management system is proposed to solve the above problems. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a safety production hazard investigation and management system, which solves the problems mentioned in the background section, such as the lack of a unified management platform, low efficiency of manual recording, reliance on experience for matching regulations and standards which is prone to errors, untimely tracking of rectification and closure, serious data silos among multiple entities, and delayed risk warning.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a safety production hazard investigation and management system, comprising:

[0007] The hazard description AI regulation matching module receives hazard description data input by the user, parses semantic features through the natural language processing unit, calls the regulatory case library for matching through the vector retrieval unit, and outputs recommendation basis data through the priority sorting unit.

[0008] The hazard investigation and entry module receives the recommended basis data, acquires on-site inspection data through the mobile acquisition unit, identifies hazard attributes using the classification and grading unit, and outputs structured hazard records through the corroboration and association unit.

[0009] The rectification closed-loop management module receives the structured hidden danger records, pushes rectification instructions to the enterprise through the task dispatch unit, monitors the rectification progress through the status tracking unit, and outputs closed-loop status data through the review and verification unit.

[0010] The multi-entity data collaboration module receives the closed-loop status data, allocates data access permissions through the permission control unit, integrates data from multiple parties using the data middle platform unit, and outputs a shared data stream through the real-time synchronization unit.

[0011] The risk warning and analysis module receives the shared data stream, analyzes risk patterns through the statistical modeling unit, generates warning charts using the visualization unit, and outputs risk intervention signals through the threshold alarm unit.

[0012] The report generation module receives the risk intervention signal, integrates the inspection data and assessment results through the template rendering unit, and generates a customized report using the multi-dimensional export unit.

[0013] Preferably, the process of receiving hazard description data input by the user in the hazard description AI regulatory matching module, parsing semantic features through a natural language processing unit, calling the regulatory case library for matching through a vector retrieval unit, and outputting recommendation basis data through a priority ranking unit specifically includes:

[0014] Artificial Intelligence Large Model Connection Unit: Connects to the Tongyi large model and utilizes its algorithms and computing power for natural language processing;

[0015] Vector database building block: Converts the professional knowledge of the legal library, standards library, and case library into vector form for storage;

[0016] The retrieval enhancement generation technology application unit performs semantic recognition and scene matching on the description of potential hazards, retrieves the closest regulations and cases from the vector database, calculates the similarity between the hazard description vector and the regulation vector using a dimensionally normalized similarity calculation formula, and generates recommendations based on these similarities. The similarity calculation formula is as follows:

[0017] ;

[0018] in, This is for normalized similarity scores, with values ​​ranging from [-1, 1], and in practical applications, the absolute value is taken as [0, 1]). This is the normalized feature vector describing potential hazards, with a magnitude of 1. The normalized feature vector of the legal provisions has a magnitude of 1. Automatic matching succeeds in time;

[0019] Manual adjustment unit: When the automatic matching results do not match, technicians can manually select the appropriate criteria from the library.

[0020] Preferably, the hazard investigation and entry module receives the recommended basis data, acquires on-site inspection data through the mobile acquisition unit, identifies hazard attributes using the classification and grading unit, and outputs structured hazard records through the corroboration and association unit, specifically including:

[0021] Checklist selection unit: Before entering the hazard investigation record, technicians can select the appropriate checklist for the inspected company from their mobile devices. Multiple checklist types can be selected.

[0022] Hazard information entry unit: Enter the inspection results item by item, including hazard classification, hazard level, hazard problem description, rectification opinions and rectification deadline. Hazard classification can be customized through backend configuration.

[0023] Supporting Document Upload Unit: Supports taking photos on-site and uploading supporting documents, automatically linking location information;

[0024] Based on the recommendation unit: when inputting a description of a potential hazard, the system calls upon a large AI model to match the relevant data from a database and recommends the closest applicable regulations and standards.

[0025] Preferably, the construction and use of the regulatory library, standard library, and case library include:

[0026] The regulatory database storage unit stores national and industry-specific safety production regulations and provisions.

[0027] Standard library storage unit: stores safety production technology standards;

[0028] Case library storage unit: stores historical hazard case data;

[0029] Matching Priority Execution Unit: When matching criteria, first search for similar hidden danger data in the case library, then search the legal library, and finally search the standard library.

[0030] Preferably, the process of receiving the structured hazard record in the rectification closed-loop management module, pushing rectification instructions to the enterprise through the task dispatch unit, monitoring the rectification progress through the status tracking unit, and outputting closed-loop status data through the review and verification unit specifically includes:

[0031] Rectification Task Dispatch Unit: Safety service organizations push rectification tasks to enterprises through the organization's terminal. The tasks include hazard details, rectification requirements, and rectification deadlines.

[0032] Enterprise Feedback Unit: Enterprises upload rectification results and supporting materials through the enterprise portal;

[0033] Organizational review unit: Safety service organizations verify the rectification content against the original hazard information through the organization's platform to avoid discrepancies and delays;

[0034] Status tracking unit: Records operation logs and tracks the status from pending rectification to pending review and then to rectification completion, ensuring that every potential hazard has a complete closed-loop record.

[0035] Preferably, the process of receiving the closed-loop status data in the multi-entity data collaboration module, allocating data access permissions through the permission control unit, integrating multi-party data using the data middleware unit, and outputting a shared data stream through the real-time synchronization unit specifically includes:

[0036] Unified Data Platform Unit: Integrates hazard data, rectification status, and legal basis;

[0037] Role-based access control unit: Data access permissions are assigned according to roles. Security service agencies can view all project data, enterprises are limited to their own data, and insurance institutions are limited to data of insured projects.

[0038] Data sharing channel unit: Real-time data synchronization supports insurance institutions in risk assessment and provides a basis for enterprise traceability and inspection.

[0039] Preferably, the real-time functions of the mobile terminal include:

[0040] On-site data entry unit: Technicians can select inspection forms and fill in hazard information on-site via a mobile app;

[0041] Photo upload unit: Take and upload on-site supporting photos, automatically embedding geolocation information;

[0042] Real-time synchronization unit: Data is uploaded to the cloud instantly, avoiding manual entry afterwards.

[0043] Preferably, the process of receiving the shared data stream in the risk warning analysis module, analyzing risk patterns through the statistical modeling unit, generating warning charts using the visualization unit, and outputting risk intervention signals through the threshold alarm unit specifically includes:

[0044] Real-time data statistics unit: Automatically summarizes the distribution of hazard types, frequency of risk levels, and rectification progress rate;

[0045] Visualized cockpit unit: Generates risk analysis charts, including bar charts, pie charts, and trend lines;

[0046] High-risk identification unit: Calculates the hazard risk value using a dimensionally normalized risk scoring formula, sets a threshold alarm, and triggers an early warning notification when the risk value exceeds the threshold. The risk scoring formula is as follows: ;

[0047] in, This is a normalized risk score, with values ​​ranging from [0,1]. To normalize the hazard level coefficient, To normalize the rectification progress delay coefficient, , These are the normalized weighting coefficients. ,set up ,when A red alert was triggered at that time.

[0048] Preferably, the role permission control unit specifically includes:

[0049] Access control hierarchy: Define the roles of security service organizations as having full access to data, enterprise roles as having only the right to edit data within their own organization, and insurance institution roles as having only the right to view underwriting data;

[0050] Data isolation unit: Restricts cross-subject data leakage through access control lists;

[0051] Audit log unit: Records all data access operations to ensure accountability.

[0052] Preferably, the implementation of each terminal of the system includes:

[0053] Mobile: Safety technicians input on-site hazard investigation data, which runs on smartphones and tablets;

[0054] Enterprise H5 platform: Enterprises receive rectification tasks and provide feedback on rectification results, accessible through a web browser;

[0055] Organization PC version: Security service organizations manage inspection records, dispatch orders, and review them, running on desktop computers;

[0056] On the insurance side: Insurance institutions can view risk data for insured projects, which is integrated into the insurance management system.

[0057] Compared with the prior art, the present invention provides a safety production hazard investigation and management system, which has the following beneficial effects:

[0058] 1. In this invention, by setting up an AI-driven automatic matching mechanism for regulations, standards, and cases, when conducting safety hazard investigations, the invention utilizes a vector database constructed by combining a large-scale artificial intelligence model with a regulatory library, a standards library, and a case library to achieve intelligent semantic matching of hazard descriptions, automatically recommending the most appropriate regulatory basis, ensuring the accuracy and professionalism of the hazard investigation basis, while shortening the regulatory retrieval time, improving investigation efficiency, reducing the risk of human error, and ensuring the reliability of accident prevention.

[0059] 2. In this invention, by setting up a multi-terminal collaborative closed-loop management link for hazard rectification, a digital process is established during the investigation of safety production hazards, including dispatching orders by safety service agencies, feedback from enterprises on rectification, and review and archiving by agencies. This ensures that every hazard has a complete and traceable record from discovery to resolution, avoiding omissions and delays in rectification, improving the efficiency and closed-loop effect of hazard handling, and enhancing the sustainability of production safety.

[0060] 3. In this invention, by setting up a role-based access control unit, when conducting safety production hazard investigation, data collaboration and sharing among safety service agencies, insured enterprises, and insurance institutions are realized based on a unified data platform. Access scope is controlled according to permissions. At the same time, the data dashboard provides real-time statistics on hazard types and rectification progress, generates risk analysis charts, identifies high-risk points in advance, and intervenes in prevention and control, thereby improving data collaboration and the accuracy of risk warning, and ensuring proactive protection capabilities for safety production. Attached Figure Description

[0061] Figure 1 This is a schematic diagram of the architecture of a safety production hazard investigation and management system according to the present invention;

[0062] Figure 2 This is a schematic diagram illustrating the AI-based regulatory matching of hazard descriptions in a safety production hazard investigation and management system according to the present invention.

[0063] Figure 3 This is a schematic diagram of the data flow for hazard investigation in a safety production hazard investigation and management system according to the present invention. Detailed Implementation

[0064] 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 embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0065] Please see Figure 1-3 The specific implementation of a safety production hazard investigation and management system is as follows, including:

[0066] The hazard description AI regulation matching module receives hazard description data input by the user, parses semantic features through the natural language processing unit, calls the regulatory case library for matching through the vector retrieval unit, and outputs recommendation basis data through the priority sorting unit.

[0067] The hazard identification and entry module receives recommended data, acquires on-site inspection data through mobile acquisition units, identifies hazard attributes using classification and grading units, and outputs structured hazard records through supporting association units.

[0068] The rectification closed-loop management module receives structured hazard records, pushes rectification instructions to the enterprise through the task dispatch unit, monitors the rectification progress through the status tracking unit, and outputs closed-loop status data through the review and verification unit.

[0069] The multi-entity data collaboration module receives closed-loop status data, allocates data access permissions through the permission control unit, integrates data from multiple parties using the data middle platform unit, and outputs a shared data stream through the real-time synchronization unit.

[0070] The risk warning and analysis module receives shared data streams, analyzes risk patterns through statistical modeling units, generates warning charts using visualization units, and outputs risk intervention signals through threshold alarm units.

[0071] The report generation module receives risk intervention signals, integrates inspection data and assessment results through the template rendering unit, and generates customized reports using the multi-dimensional export unit.

[0072] The AI-powered regulatory matching module for hazard description receives hazard description data input by the user, parses semantic features through a natural language processing unit, uses a vector retrieval unit to call upon a regulatory case library for matching, and outputs recommended data through a priority ranking unit. The specific process includes:

[0073] Artificial Intelligence Large Model Connection Unit: Connects to the Tongyi large model, quantifying the large model's semantic parsing ability for hazard descriptions through a semantic understanding degree calculation formula. The calculation formula is as follows:

[0074] ;

[0075] in, The semantic understanding level, with a value range of [0,1]. The number of semantic units correctly parsed by a large model. The total number of semantic units in the input text, when If the semantic understanding is deemed satisfactory, its algorithm and computing power will be used for natural language processing.

[0076] Vector database building block: Converts the professional knowledge of the legal library, standards library, and case library into vector form for storage;

[0077] The construction and use of regulatory libraries, standards libraries, and case libraries include:

[0078] The regulatory database storage unit stores national and industry-specific safety production regulations and provisions.

[0079] Standard library storage unit: stores safety production technology standards;

[0080] Case library storage unit: stores historical hazard case data;

[0081] Priority matching execution unit: When matching criteria, it prioritizes searching similar hidden danger data in the case library, then searches the legal library, and finally searches the standards library;

[0082] The retrieval enhancement generation technology application unit performs semantic recognition and scene matching on the description of potential hazards, retrieves the closest regulations and cases from the vector database, calculates the similarity between the hazard description vector and the regulation vector using a dimensionally normalized similarity calculation formula, and generates recommendations based on these similarities. The similarity calculation formula is as follows:

[0083] ;

[0084] in, This is for normalized similarity scores, with values ​​ranging from [-1, 1], and in practical applications, the absolute value is taken as [0, 1]). This is the normalized feature vector describing potential hazards, with a magnitude of 1. The normalized feature vector of the legal provisions has a magnitude of 1. Automatic matching succeeds in time;

[0085] Manual adjustment unit: When the automatic matching results do not match, technicians can manually select the appropriate criteria from the library.

[0086] The hazard identification and entry module receives recommended data, acquires on-site inspection data through a mobile data collection unit, identifies hazard attributes using a classification and grading unit, and outputs structured hazard records through a supporting correlation unit. Specifically, these records include:

[0087] Checklist selection unit: Before entering the hazard investigation record, technicians can select the appropriate checklist for the inspected company from their mobile devices. Multiple checklist types can be selected.

[0088] Real-time mobile features include:

[0089] On-site data entry unit: Technicians can select inspection forms and fill in hazard information on-site via a mobile app;

[0090] Photo upload unit: Take and upload on-site supporting photos, automatically embedding geolocation information;

[0091] Real-time synchronization unit: Data is uploaded to the cloud instantly, avoiding manual entry afterwards;

[0092] Hazard information entry unit: Enter the inspection results item by item, including hazard classification, hazard level, hazard problem description, rectification opinions and rectification deadline. Hazard classification can be customized through backend configuration.

[0093] Supporting Document Upload Unit: Supports taking photos on-site and uploading supporting documents, automatically linking location information;

[0094] Based on the recommendation unit: When inputting a description of a potential problem, the system calls upon a large-scale AI model to match the data from its database. The reliability of the recommendations is assessed using a confidence score calculation formula, which is as follows:

[0095] ;

[0096] in, To normalize the recommendation confidence level, the value range is [0,1]. To normalize semantic similarity, For semantic understanding, , These are the normalized weighting coefficients. ,when It automatically pushes recommendations based on the most relevant laws and standards.

[0097] The process of receiving structured hazard records in the rectification closed-loop management module, pushing rectification instructions to the enterprise through the task dispatch unit, monitoring rectification progress through the status tracking unit, and outputting closed-loop status data through the verification unit specifically includes:

[0098] Rectification Task Dispatch Unit: Safety service organizations push rectification tasks to enterprises through the organization's terminal. The tasks include hazard details, rectification requirements, and rectification deadlines.

[0099] Enterprise Feedback Unit: Enterprises upload rectification results and supporting materials through the enterprise portal;

[0100] Organizational review unit: Safety service organizations verify the rectification content against the original hazard information through the organization's platform to avoid discrepancies and delays;

[0101] Status tracking unit: Records operation logs, tracks the status from pending rectification to pending review and then to rectification completion, and quantifies the rectification effect through a closed-loop efficiency index calculation formula, which is as follows:

[0102] ;

[0103] in, The normalized closed-loop efficiency index has a value range of [0,1]. To improve the rectification completion rate, This represents the average number of days of delay. The maximum tolerable delay days, , These are the normalized weighting coefficients. ,when The system automatically triggers early warnings for rectification efficiency, ensuring that every potential hazard has a complete closed-loop record.

[0104] The process of receiving closed-loop status data in the multi-entity data collaboration module, allocating data access permissions through the permission control unit, integrating multi-party data using the data middleware unit, and outputting a shared data stream through the real-time synchronization unit specifically includes:

[0105] Unified Data Platform Unit: Integrates hazard data, rectification status, and legal basis;

[0106] Role-based access control unit: Data access permissions are assigned according to roles. Security service agencies can view all project data, enterprises are limited to their own data, and insurance institutions are limited to data of insured projects.

[0107] The role-based access control unit specifically includes:

[0108] Access control hierarchy: Define the roles of security service organizations as having full access to data, enterprise roles as having only the right to edit data within their own organization, and insurance institution roles as having only the right to view underwriting data;

[0109] Data isolation unit: Restricts cross-subject data leakage through access control lists;

[0110] Audit log unit: Records all data access operations to ensure accountability;

[0111] Data sharing channel unit: Real-time data synchronization, quantifying the efficiency of multi-entity data flow through a collaborative efficiency index formula, the calculation formula of which is:

[0112] ;

[0113] in, The normalized collaborative efficiency index has a value range of [0,1]. For data integrity rate, The average synchronization delay is expressed in seconds. For maximum tolerable delay, , Normalized weighting coefficients ,when Timely triggering of collaborative optimization alarms;

[0114] Support insurance institutions in conducting risk assessments and provide a basis for enterprises to conduct retrospective inspections.

[0115] The risk warning and analysis module receives a shared data stream, analyzes risk patterns through a statistical modeling unit, generates warning charts using a visualization unit, and outputs risk intervention signals through a threshold alarm unit. The specific process includes:

[0116] Real-time data statistics unit: Automatically summarizes the distribution of hazard types, frequency of risk levels, and rectification progress rate;

[0117] Visualized cockpit unit: Generates risk analysis charts, including bar charts, pie charts, and trend lines;

[0118] High-risk identification unit: Calculates the hazard risk value using a dimensionally normalized risk scoring formula, sets a threshold alarm, and triggers an early warning notification when the risk value exceeds the threshold. The risk scoring formula is as follows: ;

[0119] in, This is a normalized risk score, with values ​​ranging from [0,1]. To normalize the hazard level coefficient, To normalize the rectification progress delay coefficient, , These are the normalized weighting coefficients. ,set up ,when A red alert was triggered at that time.

[0120] The implementation of each terminal of the system includes:

[0121] Mobile: Safety technicians input on-site hazard investigation data, which runs on smartphones and tablets;

[0122] Enterprise H5 platform: Enterprises receive rectification tasks and provide feedback on rectification results, accessible through a web browser;

[0123] Organization PC version: Security service organizations manage inspection records, dispatch orders, and review them, running on desktop computers;

[0124] On the insurance side: Insurance institutions can view risk data for insured projects, which is integrated into the insurance management system;

[0125] Cross-terminal collaborative monitoring unit: Evaluates the collaborative performance of multiple terminals and quantifies operational efficiency through a system response performance formula, the calculation formula of which is as follows:

[0126] ;

[0127] in, The normalized system overall performance index, ranging from [0,1]. For mobile response efficiency, To improve data collaboration efficiency, To improve the closed-loop efficiency of rectification. , , These are the normalized weighting coefficients. ,when System-level optimization alerts are triggered at certain times.

[0128] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0129] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A safety production hidden danger investigation management system, characterized in that: The application comprises the following: The hidden danger description AI regulation matching module receives user input hidden danger description data, analyzes semantic features through a natural language processing unit, matches by calling a regulation case library using a vector retrieval unit, and outputs recommended basis data through a priority sorting unit; The hidden danger investigation input module receives the recommended basis data, acquires on-site inspection data through a mobile collection unit, identifies hidden danger attributes using a classification and grading unit, and outputs structured hidden danger records through a supporting evidence association unit; The rectification closed-loop management module receives the structured hidden danger records, pushes rectification instructions to the enterprise end through a task assignment unit, monitors rectification progress using a state tracking unit, and outputs closed-loop state data through a review and verification unit; The multi-subject data collaboration module receives the closed-loop state data, distributes data access permissions through a permission control unit, integrates multi-party data using a data hub unit, and outputs shared data streams through a real-time synchronization unit; The risk early warning analysis module receives the shared data streams, analyzes risk patterns through a statistical modeling unit, generates early warning charts using a visualization unit, and outputs risk intervention signals through a threshold alarm unit; The report generation module receives the risk intervention signals, integrates inspection data and evaluation results through a template rendering unit, and generates customized reports using a multi-dimensional export unit.

2. The safety production hidden danger checking management system according to claim 1, characterized in that: The hidden danger description AI regulation matching module receives user input hidden danger description data, analyzes semantic features through a natural language processing unit, matches by calling a regulation case library using a vector retrieval unit, and outputs recommended basis data through a priority sorting unit. The process specifically includes: An artificial intelligence large model docking unit: docking a Tongyi large model to perform natural language processing using its algorithms and computing power; A vector database construction unit: converting professional knowledge from regulation libraries, standard libraries, and case libraries into vector form for storage; A retrieval enhancement generation technology application unit: performing semantic recognition and scene matching on hidden danger problem descriptions, retrieving the closest regulations and cases in the vector database, calculating the similarity between hidden danger description vectors and regulation vectors through a dimensionless similarity calculation formula, and generating basis recommendations, wherein the similarity calculation formula is: ； wherein, is a normalized similarity score, with a value range of [-1, 1], and in practical applications, the absolute value [0, 1] is taken, is a normalized hidden danger description feature vector, with a length of 1, is a normalized regulation article feature vector, with a length of 1, and when an automatic matching is successful. A manual adjustment unit: when the automatic matching result does not match, supporting technical personnel to manually select appropriate basis from the library.

3. The safety production hidden danger checking management system according to claim 1, characterized in that: The hidden danger investigation input module receives the recommended basis data, acquires on-site inspection data through a mobile collection unit, identifies hidden danger attributes using a classification and grading unit, and outputs structured hidden danger records through a supporting evidence association unit. Specifically, the process includes: An inspection table selection unit: before inputting hidden danger investigation records, supporting technical personnel to select inspection tables suitable for the inspected enterprise from the mobile end, and the inspection table type can be multiple selected; A hidden danger information filling unit: inputting inspection results item by item, including filling in hidden danger classification, hidden danger grading, hidden danger problem description, rectification opinions, and rectification period, wherein the hidden danger classification is customized through background configuration; An evidence material uploading unit: supporting on-site photographing and uploading of supporting evidence materials, and automatically associating location positioning information; According to the recommendation unit: when inputting the hidden danger problem description, the artificial intelligence big model matching basis library is called to recommend the closest regulations and standard basis.

4. The safety production hidden danger checking management system according to claim 2, characterized in that: The construction and use of the regulation library, standard library and case library include: The regulation library storage unit: stores national and industrial safety production regulations; The standard library storage unit: stores safety production technical standards; The case library storage unit: stores historical hidden danger case data; The matching priority execution unit: when matching the basis, the similar hidden danger data in the case library is preferentially searched, then the regulation library is searched, and finally the standard library is searched.

5. The safety production hidden danger checking management system according to claim 1, characterized in that: The process in which the rectification closed-loop management module receives the structured hidden danger record, pushes the rectification instruction to the enterprise end through the task dispatching unit, monitors the rectification progress by using the state tracking unit, and outputs the closed-loop state data through the review verification unit specifically includes: The rectification task dispatching unit: the safety service institution pushes the rectification task to the enterprise end through the institution end, and the task includes hidden danger details, rectification requirements and rectification deadline; The enterprise feedback unit: the enterprise uploads the rectification result and supporting materials through the enterprise end; The institution review unit: the safety service institution checks the rectification content and the original hidden danger information through the institution end to avoid non-correspondence and delay problems; The state tracking unit: records the operation log and tracks the state from the to-be-rectified state to the to-be-reviewed state and then to the rectification completion state, so that complete closed-loop records are ensured for each hidden danger.

6. The safety production hidden danger checking management system according to claim 1, characterized in that: The process in which the multi-subject data collaboration module receives the closed-loop state data, distributes data access permissions through the permission control unit, integrates multi-party data by using the data middle platform unit, and outputs shared data flow through the real-time synchronization unit specifically includes: The unified data middle platform unit: integrates hidden danger data, rectification situation and regulation basis; The role permission control unit: distributes data access permissions according to roles, the safety service institution end can view all project data, the enterprise end is limited to the data of the unit, and the insurance institution end is limited to the data of the insured project; The data sharing channel unit: synchronizes data in real time, supports the insurance institution to perform risk assessment, and supports the enterprise to trace check the basis.

7. The safety production hidden danger checking management system according to claim 3, characterized in that: The mobile end real-time function includes: The on-site input unit: the technical personnel select an inspection table, fill in hidden danger information through the mobile end applet at the inspection site; The photograph uploading unit: photographs and uploads on-site supporting photos, and automatically embeds geographical positioning information; The real-time synchronization unit: data is uploaded to the cloud in real time to avoid manual input after the event.

8. The safety production hidden danger checking management system according to claim 1, characterized in that: The process in which the risk early warning analysis module receives the shared data flow, analyzes the risk mode through the statistical modeling unit, generates a warning chart by using the visualization unit, and outputs a risk intervention signal through the threshold alarm unit specifically includes: The real-time data statistical unit: automatically summarizes hidden danger type distribution, risk level frequency and rectification progress rate; The visualization cockpit unit: generates a risk analysis chart, including a column chart, a pie chart and a trend line; The high-risk identification unit calculates the risk value of the hidden danger through the dimensionless risk score formula, sets a threshold value alarm, and triggers a pre-warning notification when the risk value exceeds the threshold value, wherein the risk score formula is: ; wherein, is a normalized risk score, with a value range [0, 1], is a normalized hidden danger level coefficient, is a normalized rectification progress delay coefficient, , is a normalized weight coefficient, , set a red early warning is triggered when .

9. The safety production hidden danger checking management system according to claim 6, characterized in that: The role permission control unit specifically includes: The permission grading unit: defines that the safety service institution role has full data access right, the enterprise role is limited to self-unit data editing right, and the insurance institution role is limited to insured data viewing right; The data isolation unit: limits cross-subject data leakage through an access control list; Audit log unit: record all data access operations, ensure accountability.

10. The safety production hidden danger checking management system according to claim 1, characterized in that: The implementation of each end of the system includes: Mobile end: security technicians on-site hidden trouble investigation input, running on smart phones and tablet devices; Enterprise H5 end: enterprise receives rectification tasks, feedback rectification results, through web browsers; Institution PC end: security service agency management inspection records, dispatch and review, running on desktop computers; Insurance end: insurance agencies view risk data of underwriting projects, integrated into insurance management systems.

Images