Hidden danger investigation and control system and its methods

Abstract

The present invention relates to the technical field of hidden danger investigation and management, and discloses a hidden danger investigation and management system and an investigation and management method. The hidden danger investigation and management system includes: a terminal layer, an application layer and a database layer. The technical solution provided by one or more embodiments disclosed in the present invention can determine the inspection area, inspection unit and inspection point of the target enterprise by pre-associating the basic information and regional information obtained by the terminal layer with the enterprise knowledge base, without manually analyzing and determining the inspection point each time the hidden danger investigation and management is performed, thereby improving the efficiency and accuracy of the hidden danger investigation and management.

Description

Technical Field

[0001] The present invention relates to the technical field of hidden danger investigation and management, and in particular to a hidden danger investigation and management system and an investigation and management method. Background Art

[0002] With the development of industry, the hidden danger detection work of enterprise production units in various industries has become a necessary process in the process of safe production management; the existing hidden danger detection and management system can provide enterprises with a hidden danger detection work list, but there is a problem that enterprises need to re-analyze and determine the inspection points every time they conduct detection and management, which affects the efficiency of the detection and management tasks and requires a better solution. Summary of the invention

[0003] In view of this, the present invention provides a hidden danger investigation and management system and an investigation and management method to solve the problem of low efficiency in enterprises when conducting hidden danger investigation and management.

[0004] On one hand, the present invention provides a hidden danger investigation and management system, including: a terminal layer, an application layer and a database layer;

[0005] Through the enterprise knowledge base in the database layer, obtain and store the inspection units, inspection points and control measures corresponding to all industry fields;

[0006] According to the basic information and regional information of the target enterprise obtained by the terminal layer, the basic information, the regional information and the enterprise knowledge base are pre-associated in the application layer, and the inspection unit and the inspection point whose similarity in the enterprise knowledge base is greater than a preset threshold are associated with the target enterprise;

[0007] According to the inspection unit and the inspection point, a list of inspection points corresponding to the target enterprise is determined in the application layer; the list of inspection points includes but is not limited to: an inspection area, the inspection unit and the inspection point; wherein the inspection area is determined according to the area information;

[0008] Determine the optimal investigation path according to the checkpoint list, determine the control measures corresponding to the checkpoint list from the enterprise knowledge base, and determine the investigation and governance guidelines corresponding to the target enterprise according to the optimal investigation path and the control measures;

[0009] The checkpoint list and the investigation and management guidelines are displayed to the target enterprise at the terminal layer.

[0010] Another aspect of the present invention further provides a method for troubleshooting and managing hidden dangers, which is used in the above hidden danger troubleshooting and managing system, and the method comprises:

[0011] Obtain inspection units, inspection points and control measures corresponding to all industry fields through the enterprise knowledge base;

[0012] Receiving the area information of the target enterprise uploaded by the user according to the preset form template or area image information, and determining the inspection area according to the area information;

[0013] According to the hierarchical relationship, the inspection units and inspection nodes corresponding to the inspection areas are respectively associated and determined, and the inspection node list is determined in the form of inspection area-inspection unit-inspection node;

[0014] Receive the user's troubleshooting and governance plan for the inspection node list, determine the troubleshooting and governance guidance text according to the control measures corresponding to the inspection node list, and generate the optimal troubleshooting path according to the inspection node list;

[0015] Determine the troubleshooting and management guidelines corresponding to the inspection node list according to the troubleshooting and management guide text and the optimal troubleshooting path;

[0016] Determine the department and department responsible person corresponding to the investigation and governance plan, issue the investigation and governance task to the department, and make the department responsible person execute and accept the investigation and governance task of the department.

[0017] In this process, the basic information and regional information obtained at the terminal layer can be pre-associated with the enterprise knowledge base to determine the inspection area, inspection unit and inspection point of the target enterprise. There is no need to manually analyze and determine the inspection points every time a hidden danger investigation and management is carried out, thereby improving the efficiency and accuracy of hidden danger investigation and management. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 It is a structural schematic diagram of a hidden danger detection and management system provided by an embodiment of the present invention;

[0020] Figure 2 It is a structural schematic diagram of another hidden danger detection and management system provided by an embodiment of the present invention;

[0021] Figure 3 It is a flowchart of a troubleshooting and management method provided by an embodiment of the present invention;

[0022] Figure 4It is a flowchart of another troubleshooting and management method provided by an embodiment of the present invention;

[0023] Figure 5 This is a schematic diagram of the process of building an enterprise knowledge base of a troubleshooting and management method provided by an embodiment of the present invention;

[0024] Figure 6 It is a schematic diagram of a regional information collection process of a troubleshooting and management method provided by an embodiment of the present invention;

[0025] Figure 7 It is a structural diagram of a troubleshooting and management device provided in an embodiment of the present invention.

[0026] Description of reference numerals:

[0027] Terminal layer 1; application layer 2; database layer 3; desktop terminal 11; mobile terminal 12; regional information collection module 21; investigation and management module 22; safety management module 23; regional information entry submodule 211; inspection unit association submodule 212; inspection point association submodule 213; regional inspection point list generation submodule 214; regional information three-dimensional model generation submodule 215; inspection point QR code generation submodule 216; investigation and management plan creation submodule 221; investigation and management guide generation submodule 222; investigation and management path submodule 2221; investigation and management text generation submodule 2222; investigation and management task acceptance Collection submodule 223; investigation and management task implementation submodule 224; investigation and management record storage submodule 225; overview display submodule 231; number statistics submodule 232; regional information three-dimensional model display submodule 233; enterprise knowledge base 31; enterprise information base 32; industry type library 311; public inspection unit library 312; specific inspection unit library 313; general inspection point library 314; specific inspection point library 315; public inspection point library 316; control measures library 317; information collection template library 318; enterprise regional risk knowledge map 319; enterprise basic information library 321; enterprise hidden danger information library 322. DETAILED DESCRIPTION

[0028] Among the existing hidden danger inspection and management methods and / or systems, there are: a dual prevention management system based on the hierarchical division method of region-main unit-sub-unit-job unit or the introduction of a risk management and hidden danger inspection list for industry fields, the construction of a risk management basic knowledge base for enterprise risk regions, risk points, risk units, and hidden danger inspection points, the formation of a risk classification management form, and the implementation of risk management work. However, these methods have the following problems:

[0029] First, although the above method includes volume inspection areas, inspection units and inspection points, it does not mention how to use associations or mapping relationships to carry out hidden danger inspections. Each time hidden danger inspections are carried out, the inspection points are manually analyzed and determined, which affects the inspection efficiency.

[0030] Second, although the above method can generate a risk classification control list and a hidden danger investigation and management work form, it does not provide users with exclusive investigation and management guidance, resulting in users not knowing who will check the hidden dangers, what to check, how to check, and how to correct them after they are found.

[0031] To solve at least one of the above problems, various embodiments of the present invention provide a hidden danger inspection and management system, including: a terminal layer, an application layer and a database layer; through the enterprise knowledge base in the database layer, obtain and store the inspection units, inspection points and control measures corresponding to all industry fields; according to the basic information and regional information of the target enterprise obtained by the terminal layer, the basic information, regional information and the enterprise knowledge base are pre-associated in the application layer, and the inspection units and inspection points in the enterprise knowledge base with a similarity greater than a preset threshold are associated with the target enterprise; according to the inspection units and inspection points, a list of inspection points corresponding to the target enterprise is determined in the application layer; the list of inspection points includes but is not limited to: inspection areas, inspection units and inspection points; wherein the inspection area is determined according to the area information; the optimal inspection path is determined according to the inspection point list, and the control measures corresponding to the inspection point list are determined from the enterprise knowledge base, and the inspection and management guidelines corresponding to the target enterprise are determined according to the optimal inspection path and the control measures; the inspection point list and the inspection and management guidelines are displayed to the target enterprise at the terminal layer.

[0032] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments set forth herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0033] In the description of the embodiments of the present disclosure, the term "including" and similar terms should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The term "some embodiments" should be understood as "at least some embodiments". Other explicit and implicit definitions may also be included below.

[0034] According to an embodiment of the present invention, a hidden danger detection and management system is provided. Figure 1is a schematic diagram of the structure of a hidden danger detection and management system provided by an embodiment of the present invention, such as Figure 1 As shown, the system includes a terminal layer 1, an application layer 2 and a database layer 3.

[0035] Among them, the inspection units, inspection points and control measures corresponding to all industry fields are obtained and stored through the enterprise knowledge base 31 in the database layer 3; according to the basic information and regional information of the target enterprise obtained by the terminal layer 1, the basic information, regional information and the enterprise knowledge base 31 are pre-associated in the application layer 2, and the inspection units and inspection points in the enterprise knowledge base 31 with a similarity greater than a preset threshold are associated with the target enterprise; according to the inspection units and inspection points, the inspection point list corresponding to the target enterprise is determined in the application layer 2; the inspection point list includes but is not limited to: inspection area, inspection unit and inspection point; wherein the inspection area is determined according to the regional information; the optimal inspection path is determined according to the inspection point list, and the control measures corresponding to the inspection point list are determined from the enterprise knowledge base 31, and the inspection and governance guidelines corresponding to the target enterprise are determined according to the optimal inspection path and the control measures; the inspection point list and the inspection and governance guidelines are displayed to the target enterprise at the terminal layer 1.

[0036] In one possible implementation, terminal layer 1 may refer to a user interface, which is used to assist users in interacting with the system; for example, terminal layer 1 may include but is not limited to: computers, tablet computers, smart phones, etc.; application layer 2 may refer to a logic processing layer, which is responsible for the implementation of business logic and data processing. Application layer 2 processes user requests sent by users through terminal layer 1 and interacts with database layer 3 for data; database layer 3 may refer to a data storage and management layer, which is used to store all data related to hidden danger investigation and management.

[0037] The basic information and regional information of the target enterprise uploaded by the user are obtained through the terminal layer 1. The basic information may refer to the information used to describe the target enterprise, and the basic information may include but is not limited to: the national economic classification, industry field, number of employees, organizational structure, and personnel information of the target enterprise; the regional information may refer to the information used for the corresponding area of ​​the target enterprise, and the regional information may include but is not limited to: regional name, regional description, regional scope, and regional area.

[0038] In a possible implementation, based on the basic information and regional information of the target enterprise acquired by the terminal layer 1, the basic information, regional information and the enterprise knowledge base 31 are pre-associated in the application layer 2, and the inspection units and inspection points in the enterprise knowledge base 31 whose similarity is greater than a preset threshold are associated with the target enterprise, which may include:

[0039] The database layer 3 receives the basic information and regional information uploaded by the terminal layer 1, and stores it in the enterprise information base 32 in the database layer 3; the application layer 2 interacts with the database layer 3 through data, and pre-associates the basic information and regional information in the enterprise information base 32 with the inspection units, inspection points and control measures stored in the enterprise knowledge base 31, and determines the inspection units and inspection points whose similarity is greater than a preset threshold, as well as the control measures corresponding to the inspection units and inspection points, respectively.

[0040] Here, the inspection unit may refer to a set of relatively independent and clearly defined inspection points divided according to physical partitions or professional sectors; inspection points may refer to equipment, facilities, work activities, and locations with independent functions and properties under the inspection unit; for example, the inspection area may refer to the warehouse storage area of ​​a garment factory, the inspection unit may refer to the fire safety unit within the warehouse storage area, and the inspection points may refer to the fire hydrants, fire extinguishers, fire alarms and other points contained in the fire safety unit.

[0041] In a possible implementation, the optimal investigation path is determined according to the checkpoint list, and the control measures corresponding to the checkpoint list are determined from the enterprise knowledge base 31. According to the optimal investigation path and the control measures, the investigation and governance guidelines corresponding to the target enterprise are determined, which may include:

[0042] By using application layer 2, the optimal inspection path is generated according to the inspection points included in the inspection point list, and the inspection and management guidelines corresponding to the inspection area of ​​the target enterprise are determined according to the control measures corresponding to the optimal inspection path and the inspection points.

[0043] Furthermore, at the terminal layer 1, a checkpoint list and troubleshooting and management guidelines are displayed to the target enterprise to assist the users of the target enterprise in formulating hidden danger troubleshooting and management tasks.

[0044] Through the above method, the basic information and regional information obtained by the terminal layer 1 can be pre-associated with the enterprise knowledge base to determine the inspection area, inspection unit and inspection point of the target enterprise, without the need to manually analyze and determine the inspection point every time a hidden danger investigation and management is carried out, thereby improving the efficiency and accuracy of hidden danger investigation and management; by generating an investigation and management list corresponding to the target enterprise, accurate and customized suggestions can be provided for the target enterprise to conduct investigation and management, thereby improving the investigation efficiency of the target enterprise.

[0045] In some embodiments, the application layer 2 includes: a regional information collection module 21, a troubleshooting and governance module 22 and a security management module 23; the regional information of the target enterprise is determined through the regional information collection module 21, the regional information and basic information are pre-associated, and the inspection area, inspection unit and inspection point corresponding to the target enterprise are determined according to the industry and field characteristics and actual production conditions of the target enterprise, and a list of inspection points corresponding to the target enterprise is generated; the basic information includes but is not limited to: national economic classification, industry field, number of employees and personnel information; the troubleshooting and governance module 22 determines the inspection area, inspection unit and inspection point to be checked by the target enterprise according to the list of inspection points determined by the regional information collection module 21, and determines and issues the troubleshooting and governance tasks and troubleshooting and governance guidelines; through the security management module 23, the risk situation and troubleshooting and governance situation of the target enterprise for the list of inspection points are counted and displayed.

[0046] In a possible implementation, the regional information uploaded by the user at the terminal layer 1 is collected through the regional information collection module 21, and the enterprise knowledge base is pre-associated based on the regional information and basic information to determine the inspection area, inspection unit and inspection point corresponding to the target enterprise; the inspection and governance plan determined by the user based on the inspection point list is received through the inspection and governance module 22, and the corresponding inspection and governance tasks and inspection and governance guidelines are determined based on the inspection areas, inspection units and inspection points to be inspected and governed in the inspection and governance plan; after the user inspects and governs the target enterprise according to the inspection and governance tasks and inspection and governance guidelines, the security management module 23 collects statistics on the inspection and governance tasks and inspection and governance guidelines corresponding to the target enterprise, and displays them to the user at the terminal layer 1.

[0047] Through the above method, according to the basic information and regional information of the target enterprise uploaded by the user, the corresponding checkpoint list and investigation and management guidelines are associated and determined, and the user's investigation and management situation is counted and displayed, so that the user can understand the risk situation and investigation and management situation of the target enterprise after management, thereby improving the accuracy of the user's investigation and management.

[0048] In some embodiments, the regional information acquisition module 21 includes: a regional information input submodule 211, an inspection unit association submodule 212, a checkpoint association submodule 213, a regional checkpoint list generation submodule 214, a regional information three-dimensional model generation submodule 215 and a checkpoint QR code generation submodule 216; the regional information input submodule 211 uses a preset computer vision model to perform image recognition semantic conversion based on the received regional image information of the target enterprise to determine the regional information of the target information; the inspection unit association submodule 212 and the checkpoint association submodule 213 pre-associate the regional information, basic information and the enterprise knowledge base 31, and determine the inspection units and checkpoints with similarities greater than a preset threshold to associate with the target enterprise based on the similarity analysis of the preset vectors; wherein, the inspection The unit includes at least one of the following: a public inspection unit or a specific inspection unit, and the inspection points include at least one of the following: a general inspection point, a public inspection point or a specific inspection point; the preset vectors include but are not limited to: a semantic feature embedding vector, an industry feature vector and a process feature vector; the regional inspection point list generation submodule 214 generates a structured inspection point list in the form of inspection area-inspection unit-inspection point by calling the regional information, the inspection unit and the inspection point; through the regional information three-dimensional model generation submodule 215, according to the inspection point list, a three-dimensional model including the inspection area, the inspection unit and the inspection point is generated; through the inspection point two-dimensional code generation submodule 216, the two-dimensional code information corresponding to each inspection point in the inspection point list is generated; the two-dimensional code information is used to mark the real-time status of the inspection point.

[0049] In a possible implementation, the regional image information may include, but is not limited to: a Building Information Modeling (BIM) diagram, a Computer-Aided Design (CAD) diagram, or a regional plan design diagram.

[0050] Among them, the BIM drawing may include a three-dimensional digital representation of the target enterprise's buildings and their components and related information, the CAD drawing may represent the engineering drawings corresponding to the target enterprise in two-dimensional form, and the regional plan design drawing may refer to the plan corresponding to the target enterprise.

[0051] In a possible implementation, the regional information input submodule 211 uses a preset computer vision model to perform image recognition semantic conversion based on the received regional image information of the target enterprise to determine the regional information of the target information, which may include:

[0052] The area information input submodule 211 uses an embedded computer vision model to convert the uploaded BIM diagram, CAD diagram or area plan into area information by performing image semantic recognition conversion on the BIM diagram, CAD diagram or area plan.

[0053] Specifically, the area information entry submodule 211 receives BIM drawings, CAD drawings or area plans uploaded by users, and identifies various objects and features in the drawings, such as buildings, equipment, pipelines, etc., through a computer vision model based on deep learning. The computer vision model extracts meaningful information by identifying specific features in the image, converts the information into a machine-understandable format, organizes and summarizes the information, and generates area information corresponding to the target enterprise. The area information may include but is not limited to: area name, area description, area range, and area area.

[0054] In a possible implementation, the inspection unit association submodule 212 and the inspection point association submodule 213 pre-associate the regional information, the basic information and the enterprise knowledge base 31, and determine, based on the similarity analysis of the preset vectors, that the inspection units and the inspection points with similarities greater than a preset threshold are associated with the target enterprise, which may include:

[0055] The inspection unit association submodule 212 and the inspection point association submodule 213 are embedded in an association model based on comprehensive embedded features, and automatically recommend inspection units and inspection points that can be associated according to the area information corresponding to the target enterprise, thereby realizing pre-association processing of inspection units and inspection points.

[0056] Here, pre-association processing may refer to extracting features from the area name, inspection unit name, and inspection point name in the enterprise knowledge base 31, representing them in the form of preset vectors, calculating the similarity of the preset vectors between the area name and the inspection unit name, and between the area name and the inspection point name, and pre-associating the inspection areas, inspection units, and inspection points whose similarities are greater than a preset threshold; the preset vectors may include, but are not limited to: semantic feature embedding vectors, industry field feature vectors, and process feature vectors.

[0057] Among them, the association model based on comprehensive embedded features can be a model that uses multiple features for embedding representation and establishes and identifies the association relationship between different entities through these embedded representations; the features can include but are not limited to: spatial features, attribute features, semantic features and relationship features.

[0058] In a possible implementation, the inspection unit includes: a public inspection unit or a specific inspection unit, and the inspection point includes: a universal inspection point, a public inspection point or a specific inspection point.

[0059] Specifically, a public inspection unit may refer to a risk inspection unit that all enterprises have, such as a power distribution room, a parking lot, a fire control room, etc.; a specific inspection unit may refer to a risk inspection unit that enterprises in a specific industry have, such as a molding section, a surface treatment section, a painting workshop, etc. that are unique to enterprises in the metal products segment; a public inspection point may refer to an inspection point that belongs to a public risk inspection unit, such as the indoor layout of the power distribution room inspection point, power safety tools and equipment; a parking lot inspection point, a vehicle entrance and exit, a vehicle parking area, etc.; a specific inspection point may refer to an inspection point that belongs to a specific risk inspection unit, such as a lathe, milling machine, and boring machine in the molding section; a sandblasting machine and a shot blasting machine in the surface treatment section; a general inspection point may refer to an inspection point that each risk inspection unit has, which does not belong to any inspection unit and is divided according to preset risk categories; preset risk categories may include, but are not limited to: general inspection points for fire protection, general inspection points for electricity consumption, general inspection points for special equipment, general inspection points for dangerous operations, general inspection points for hazardous chemicals, and general inspection points for gas.

[0060] In a possible implementation, the regional inspection point list generation submodule 214 generates a structured inspection point list in the form of inspection area-inspection unit-inspection point by calling the regional information of the target enterprise and associating the determined inspection units with the inspection points.

[0061] Furthermore, the regional information three-dimensional model generation submodule 215 performs three-dimensional modeling of the risk information and the investigation and management situation of the target enterprise based on the checkpoint list generated by the regional checkpoint list generation submodule 214, based on the secondary development of BIM. In particular, the regional information three-dimensional model generation submodule 215 presets common graphic elements such as buildings, facilities and equipment, and structural components possessed by general enterprises, performs overall modeling of the actual structure of the target enterprise, and generates a three-dimensional model including the inspection area, inspection unit, and inspection points to assist in determining the optimal path planning.

[0062] In a possible implementation, the checkpoint QR code generation submodule 216 generates QR code information corresponding to each checkpoint in the checkpoint list; the QR code information is used to mark the real-time status of the checkpoint, and may include:

[0063] According to the enterprise knowledge base 31 and the enterprise information base 32 in the database layer 3, a scannable QR code information including basic enterprise information, investigation and management plan, investigation and management record, hidden danger record, responsible personnel is generated, which can be used to mark the real-time status of each checkpoint.

[0064] Through the above method, by determining the public inspection units, specific inspection units corresponding to the target enterprise and the general inspection points, public inspection points and specific inspection points corresponding to each inspection unit, the inspection point list can be made more in line with the industry characteristics and actual production conditions of the target enterprise, improve the accuracy and pertinence of the inspection and management guidelines, and ensure that the hidden danger inspection and management is more efficient, scientific and systematic.

[0065] In some embodiments, the troubleshooting and governance module 22 includes: a troubleshooting and governance plan creation submodule 221, a troubleshooting and governance guideline generation submodule 222, a troubleshooting and governance task reception submodule 223, a troubleshooting and governance task implementation submodule 224, and a troubleshooting and governance record storage submodule 225;

[0066] The investigation and governance plan creation submodule 221 accesses the regional information of the target enterprise, and prioritizes all inspection points according to the representation value based on the intelligent recommendation algorithm based on the representation value, determines the inspection area, inspection unit and inspection point to be investigated of the target enterprise, and distributes the investigation and governance tasks of each inspection unit to different departments of the target enterprise; the investigation and governance guidance generation submodule 222 includes an investigation and governance path submodule 2221 and an investigation and governance text generation submodule 2222; according to the investigation and governance path submodule 2221, a preset path planning model is used to automatically plan the optimal investigation path; according to the investigation and governance text generation submodule 2222, a preset text generation model is used to retrieve the corresponding control measures in the enterprise knowledge base according to all the inspection points included in the optimal investigation path, and generate an investigation and governance guidance text;

[0067] The investigation and governance task receiving submodule 223 is connected to the investigation and governance plan creation submodule 221, and receives the investigation and governance tasks corresponding to the department and the governance acceptance tasks corresponding to the investigation and governance tasks issued by the investigation and governance plan creation submodule 221; the investigation and governance task implementation submodule 224 is connected to the investigation and governance plan creation submodule 221 and the investigation and governance guidance generation submodule 222, and determines the hidden danger conditions of the checkpoints, the description information of the hidden danger conditions, and the image information of the hidden danger conditions; wherein the hidden danger conditions include one of the following: there are hidden dangers, there are no hidden dangers, or none of this item; the investigation and governance record storage submodule 225 disassembles the historical investigation and governance plans and historical investigation and governance tasks, and stores the disassembled structured data in the enterprise information base 32 of the database layer.

[0068] In a possible implementation, the investigation and management plan creation submodule 221 accesses the regional information of the target enterprise, and prioritizes all inspection points according to the representation values ​​according to the intelligent recommendation algorithm based on the representation values, determines the inspection areas, inspection units and inspection points to be investigated in the target enterprise, and allocates the investigation and management tasks of each inspection unit to different departments of the target enterprise, which may include:

[0069] Receive the inspection area and inspection unit determined by the user, and prioritize all inspection points under the inspection unit selected by the user according to the characterization value; recommend inspection points whose priorities meet the preset conditions to the user, and create an inspection and governance plan based on the inspection points determined by the user.

[0070] In one possible implementation, the investigation and management path submodule 2221 can automatically plan and determine the optimal investigation path and call the regional information three-dimensional model generation submodule 215 for visual display according to all checkpoints that need to be investigated and managed in the investigation and management plan through a path planning model based on the Dijkstra algorithm and the A* algorithm.

[0071] Among them, the Dijkstra algorithm can be used to find the shortest path from a starting point to all other points, and the A* algorithm can be a heuristic search algorithm specifically used to find the shortest path from the starting point to the target point; the path planning model based on the Dijkstra algorithm and the A* algorithm determines the optimal inspection path according to the shortest path among all checkpoints.

[0072] In one possible implementation, the investigation and management text generation submodule 2222 can adopt a text generation model based on a deep neural network, retrieve corresponding control measures from the enterprise knowledge base 31 according to all checkpoints included in the optimal investigation path, and automatically generate an investigation and management guidance text.

[0073] Among them, Deep Neural Networks (DNN) can be a neural network structure including multiple hidden layers, and the text generation model can refer to a model that uses a neural network to generate natural language text. By pre-training the DNN-based text generation model, the investigation and governance guidance text is generated.

[0074] In one possible implementation, the investigation and management task implementation submodule 224 receives the hidden danger conditions of the checkpoints determined by the user through checking when the user performs investigation and management; through the preference intelligent analysis model, for the checkpoints whose frequency of checking "none" is greater than the preset checking frequency, they are automatically set to the default value the next time they are checked.

[0075] Specifically, the preference intelligent analysis model can be used to understand and predict user preferences or behaviors, and make predictions based on the user's current selections. If it is found that the user has a tendency to not have this item, it will be automatically set to the default value the next time the item is checked; the default value may refer to the value, state or option that the system or program uses by default in the absence of explicit input or an unclear value.

[0076] Furthermore, the hidden dangers of the checkpoints can be uploaded to the enterprise information base of the database layer 3 for storage.

[0077] In a possible implementation, the troubleshooting and governance record storage submodule 225 disassembles and converts the form data of the historical troubleshooting and governance plans and the historical troubleshooting and governance tasks, and stores the disassembled structured data in the enterprise information library 32 of the database layer.

[0078] Here, the enterprise information library 32 may include an enterprise basic information library 321 and an enterprise hidden danger information library 322 , and the disassembled structured data is stored in the enterprise hidden danger information library 322 in the enterprise information library 32 .

[0079] Specifically, in the form data of historical investigation and governance plans and historical investigation and governance tasks, the required information is extracted through text parsing, table data extraction, etc., the required information is cleansed, and the cleaned required information is converted into structured data; among them, structured data can refer to a data form that is organized and stored in a clearly defined format to facilitate processing and storage by computer programs.

[0080] Through the above method, when a user creates a troubleshooting and governance plan, the recommended priorities corresponding to the checkpoints are determined and sorted, and high-priority checkpoints are recommended to the user. This can help the user create a troubleshooting and governance plan more quickly, and allow the user to give priority to checkpoints with higher priorities, thereby maximizing the governance effect and solving the risks and hidden dangers faced by the target enterprise.

[0081] In some embodiments, the security management module 23 includes: an overview display submodule 231, a number statistics submodule 232, and a regional information three-dimensional model display submodule 233; the overview display submodule 231 is used to display the preset investigation and governance information corresponding to the target enterprise at the terminal layer 1; the preset investigation and governance information is used to describe the investigation and governance situation of the enterprise; the number statistics submodule 232 determines the investigation and governance change trend of the target enterprise by accessing the investigation and governance situation;

[0082] The regional information three-dimensional model display submodule 233 is connected to the regional information three-dimensional model generation submodule 215 and the inspection and control record storage submodule 225, and displays the specific locations of the inspection areas, inspection units and inspection points of the target enterprise in the form of a three-dimensional model, and uses red, yellow and green to indicate the inspection and control status of the inspection points; among which red indicates that the hidden dangers of the last inspection have not been completely controlled within the preset time, yellow indicates that the hidden dangers of the last inspection are being controlled, and green indicates that the hidden dangers of the last inspection have been completely controlled or no hidden dangers were found in the last inspection.

[0083] In a possible implementation, the overview display submodule 231 is used to display the preset investigation and management information corresponding to the target enterprise at the terminal layer 1, which may include:

[0084] The preset inspection and governance information of the target enterprise is displayed in an overview form through the overview display submodule 231; the preset inspection and governance information includes but is not limited to: the currently divided inspection areas, the number of inspection areas, the number of inspection units, the number of inspection points, the number of high-risk inspection points, the number of current inspection and governance plans, the number of current hidden dangers, the number of hidden dangers that have been rectified by time, the number of hidden dangers being rectified, the number of hidden dangers that have been rectified overdue, and the number of hidden dangers that have not been rectified overdue.

[0085] In a possible implementation, the number statistics submodule 232 determines the inspection and control trends by accessing the inspection and control situation determined by the overview display submodule 231, and determines the inspection and control change trends according to a certain time range; the inspection and control change trends include but are not limited to: the number change trends of inspection units and inspection points, the number of hidden dangers found, the number of hidden dangers rectified, the inspection points where hidden dangers occur most concentratedly, and the inspection points that are often overdue for rectification.

[0086] In a possible implementation, the optimal investigation path determined in the selected investigation and control plan can be displayed through the regional information three-dimensional model display submodule 233.

[0087] Through the above method, the number statistics submodule 232 of the target enterprise is determined according to the number statistics submodule 232, and the inspection and management information of the target enterprise is displayed in an overview form through the overview display submodule 231, so that the user does not need to check the inspection and management information of the checkpoints one by one, which saves the user's time in viewing the information and can help the user have a comprehensive understanding of the inspection and management situation of the target enterprise.

[0088] In some embodiments, the database layer 3 includes: an enterprise knowledge base 31 and an enterprise information base 32; the enterprise knowledge base 31 includes an industry type library 311 corresponding to all industry fields, a public inspection unit library 312, a specific inspection unit library 313, a general inspection point library 314, a specific inspection point library 315, a public inspection point library 316, a control measures library 317, an information collection template library 318 and an enterprise regional risk knowledge graph 319; the enterprise information base 32 includes an enterprise basic information library 321 and an enterprise hidden danger information library 322, the enterprise basic information 321 library is used to store the basic information corresponding to the target enterprise, and the enterprise hidden danger information library 322 is used to store the hidden danger information corresponding to the target enterprise.

[0089] In one possible implementation, the enterprise knowledge base 31 stores industry types, public inspection units, specific inspection units, general inspection points, specific inspection points, public inspection points, control measures, information collection templates, and enterprise regional risk knowledge maps corresponding to all industry fields.

[0090] Among them, the enterprise regional risk knowledge graph 319 can be a data structure built based on knowledge graph technology, which is used to describe the risk situation of various regions of the enterprise and the correlation and influence between them; the knowledge graph can be a semantic network used to represent entities in the real world (such as people, places, things, etc.) and the relationships between them.

[0091] Furthermore, according to the basic information of the target enterprise, a corresponding preset form template is constructed and stored in the information collection template library 318 of the enterprise knowledge base 31.

[0092] In a possible implementation, the national economic classification, industry field, number of employees, organizational structure, and personnel information of the target enterprise are stored in the enterprise basic information database 321; the hidden danger information self-inspected and self-reported by the target enterprise is stored in the enterprise hidden danger information database 322, including hidden danger locations, hidden danger descriptions, control measures, hidden danger discovery time, deadline for rectification, hidden danger rectification time, rectification instructions, and rectification records.

[0093] Through the above method, the basic information and hidden danger information of the target enterprise are stored in the enterprise information database, which can make data management clearer and more specific, improve the organization and management efficiency of data, and store the hidden danger information separately in the enterprise hidden danger information database, which helps to improve the security of the hidden danger information.

[0094] In some embodiments, the terminal layer 1 includes: a desktop terminal 11 and a mobile terminal 12; receiving regional information uploaded by the target enterprise according to a preset form template or regional image information through the desktop terminal 11; viewing the regional information QR code and three-dimensional model determined according to the regional information through the desktop terminal 11, setting an investigation and governance plan, and automatically generating corresponding inspection point lists and investigation and governance guidelines through the investigation and governance plan; viewing the preset investigation and governance information on the desktop terminal 11 according to the risk situation and investigation and governance situation of the target enterprise for the inspection point list; receiving the investigation and governance task through the mobile terminal 12, and viewing the investigation and governance guidelines corresponding to the inspection point list.

[0095] In a possible implementation, the desktop terminal 11 is the main terminal of the system, and the enterprise knowledge base 31 and the enterprise information base 32 are managed through the desktop terminal 11; the basic information and regional information of the target enterprise are collected through the desktop terminal 11.

[0096] Furthermore, the mobile terminal 12 is an auxiliary terminal of the system, which receives inspection and management tasks, checks checkpoint lists and inspection and management guidelines through the mobile terminal 12, enables users to check and fill in hidden danger information through the mobile terminal 12, and check rectified hidden dangers through the mobile terminal 12, thereby completing the closed-loop management of hidden danger inspection and management.

[0097] Through the above method, according to the combined use of the desktop terminal 11 and the mobile terminal 12, application access and management on different devices can be realized, thereby improving the flexibility and applicability of the system; through the mobile terminal 12, users can view task information and instructions anytime and anywhere, thereby facilitating the distribution and processing of tasks; through the mobile terminal 12, hidden danger information can be checked and filled in in real time, thereby realizing real-time entry and feedback of hidden danger information, thereby improving the timeliness and accuracy of information.

[0098] In some embodiments, Figure 3 is a flow chart of a troubleshooting and management method provided by an embodiment of the present invention, such as Figure 3 As shown, the process includes the following steps:

[0099] Step S301, obtaining inspection units, inspection points and control measures corresponding to all industry fields through the enterprise knowledge base;

[0100] Step S302, receiving the area information of the target enterprise uploaded by the user according to the preset form template or area image information, and determining the inspection area according to the area information;

[0101] Step S303, respectively associating and determining the inspection units and inspection nodes corresponding to the inspection area according to the hierarchical relationship, and determining the inspection node list in the form of inspection area-inspection unit-inspection node, including:

[0102] Input the name corresponding to the inspection area, the name corresponding to the inspection unit, and the name corresponding to the inspection point into a preset language representation model to determine a semantic feature embedding vector;

[0103] According to the correlation between the inspection area, the inspection unit, the inspection point and at least one national economic classification, a preset coding method is used to determine the industry field feature vector;

[0104] According to the inspection area, inspection unit, inspection point and corresponding specific process, a preset encoding method is used to determine the process feature vector;

[0105] Determine the inspection unit and inspection point corresponding to the inspection area according to the semantic feature embedding vector, industry feature vector and process feature vector;

[0106] Step S304: all checkpoints are prioritized according to the representation values ​​according to the intelligent recommendation algorithm based on the representation values, and the checkpoints whose priority rankings meet the preset conditions are recommended to the user, so as to assist the user in determining the inspection and management plan;

[0107] Step S305, determining the troubleshooting and governance guidance text according to the control measures corresponding to the inspection nodes in the troubleshooting and governance plan, and generating the optimal troubleshooting path according to the inspection nodes;

[0108] Step S306, determining the troubleshooting and governance guidelines corresponding to the troubleshooting and governance plan according to the troubleshooting and governance guidelines text and the optimal troubleshooting path;

[0109] Step S307, determine the department, department operator and department responsible person corresponding to the investigation and governance plan, issue the investigation and governance task to the department, make the department operator perform the investigation and governance task, and the department responsible person accepts the investigation and governance task.

[0110] In a possible implementation, receiving the region information of the target enterprise uploaded by the user according to a preset form template or region image information, and determining the inspection region according to the region information may include:

[0111] Receive basic information and regional information uploaded by users according to a preset form template, or automatically generate regional information of the target enterprise according to the BIM diagram, CAD diagram or regional plan uploaded by the user;

[0112] If there are changes in the automatically generated regional information, the regional information can be modified;

[0113] The region and / or sub-region is determined according to the region information; wherein the sub-region is subordinate to the region.

[0114] In a possible implementation, associating and determining the inspection units and inspection nodes corresponding to the inspection areas according to the hierarchical relationship may include:

[0115] According to the industry characteristics and actual production situation of the target enterprise, associate the public inspection unit corresponding to the uploaded area;

[0116] Specific inspection units corresponding to the associated regions according to the industry and sub-sectors to which the target enterprise belongs;

[0117] Modify and supplement the inspection units pre-associated with the system, and add, change and delete inspection units according to the actual production situation of the target enterprise;

[0118] Associate the departments and department responsible persons corresponding to each inspection unit.

[0119] Furthermore, the public inspection points, specific inspection points and universal inspection points corresponding to each inspection unit are associated, and addition, modification and deletion operations are performed on each inspection point.

[0120] In a possible implementation, inputting the name corresponding to the inspection area, the name corresponding to the inspection unit, and the name corresponding to the inspection point into a preset language representation model to determine the semantic feature embedding vector may include:

[0121] Input the names corresponding to the inspection area, the inspection unit, and the inspection point into a deep learning (Bidirectional Encoder Representations from Transformers, BERT) model based on pre-trained natural language processing tasks, and determine the output of the last layer of the model to determine the initial embedding representation with a tensor size of [v, e, h];

[0122] Perform feature fusion on the tensor [v, e, h] through convolutional neural network;

[0123] The tensor after feature fusion is input into the Gaussian Error Linear Unit (GELU) function for activation, and the activated tensor is determined;

[0124] Split the activated tensor into v matrices of size [100,e], corresponding to the input inspection area, inspection unit and inspection point name phrases one by one;

[0125] Extract the 1st to [w+2]th columns of the matrix for pooling operation, compress the matrix of size [100,e] into a 100-dimensional column vector as the semantic feature embedding vector, normalize the vector, and determine the semantic feature embedding vector s.

[0126] The name corresponding to the inspection area, the name corresponding to the inspection unit, and the name corresponding to the inspection point are input into the pre-trained BERT model, and the output of the last layer of the model is determined to determine the initial embedding representation with a tensor size of [v, e, h], which may include:

[0127] Determine [v,e,h] where v refers to the number of name phrases input into the model at one time, and is 1 when only one phrase is input each time; e refers to the maximum preset length of the embedded string, and is 128; h refers to the dimension used to represent the phrase, and is 768.

[0128] Exemplarily, the input is [indoor layout, power safety tools, equipment and facilities residual current protection device, building lightning protection device, distribution line]. After being embedded by the BERT model, a [5, 128, 768] tensor is obtained from the hidden layer as the initial embedding representation of the input text; wherein, the tensor has a total of 5 layers, each of which is a [128, 768] matrix, the rows of which represent that the input phrase has a total of 128 characters, and characters less than 128 are filled with 0; the columns of the matrix represent the embedding representation of a character in the phrase, with a total of 768 dimensions; the first column of the matrix is ​​the overall embedding representation of the phrase.

[0129] In one possible implementation, feature fusion of the tensor [v, e, h] through a convolutional neural network may include:

[0130] Transform the tensor [v, e, h] to determine the new tensor [v, h, e] to meet the input requirements of the convolutional layer;

[0131] Set 100 convolution kernels for the convolution layer, the convolution kernel size is 3, the step size and padding size are 1, and the tensor output size after the convolution layer is [v, 100, e].

[0132] For example, the size of the output tensor of the tensor [5, 128, 768] after the convolution layer is [5, 100, 128], where the embedding representation of each character is compressed from the original 768 dimensions to 100 dimensions.

[0133] In a possible implementation, the feature-fused tensor is input into a Gaussian error linear unit function for activation, and determining the activated tensor may include: using the following approximate calculation formula to determine the activated tensor:

[0134]

[0135] Among them, GELU(x) can represent the Gaussian error linear unit function GELU; GELU can be an activation function that can perform a nonlinear transformation on the input tensor, thereby introducing nonlinear features.

[0136] In a possible implementation, extracting the first column to the [w+2]th column of the matrix for pooling operation, compressing the matrix of size [100, e] into a 100-dimensional column vector as a semantic feature embedding vector, normalizing the vector, and determining the semantic feature embedding vector s may include: compressing the matrix into a 100-dimensional column vector using the following formula:

[0137]

[0138] Among them, x i,j are the elements in the matrix before pooling; x i-pool Each element in the 100-dimensional column vector is compressed after pooling;

[0139] After the operation, the vector needs to be normalized. The following L2 normalization formula can be used for normalization:

[0140]

[0141] Among them, x i-pool are the elements in the normalized vector;

[0142] The normalized vector is determined as the semantic feature embedding vector s.

[0143] Here, L2 normalization may be a technique of scaling a vector to unit length (ie, making its L2 norm 1).

[0144] In a possible implementation, according to the correlation between the inspection area, the inspection unit, the inspection point and at least one national economic classification, a preset coding method is used to determine the industry field feature vector, which may include:

[0145] The TF-IDF method is used to determine the relationship between each inspection area, inspection unit, and inspection point and the industry from the preset documents; the m-dimensional zero vector is determined through multi-hot encoding, and the zero vector is replaced with the elements of the original vector according to the calculated TF-IDF value in the preset industry order, and re-assigned to form a new m-dimensional vector; the m-dimensional vector is normalized. The L2 normalization method is used for normalization, and the m-dimensional vector formed after normalization is used as the semantic feature embedding vector i.

[0146] Specifically, the TF-IDF method is used to determine the relationship between each inspection area, inspection unit, inspection point and industry from the preset documents, which may include:

[0147] According to the major industry categories specified in the preset documents, determine the laws, regulations, rules, standards and specifications corresponding to the industries in each major category; use the TF-IDF method to calculate the relationship between each inspection area, inspection unit, inspection point and the industry from the preset documents.

[0148] Among them, the Term Frequency-Inverse Document Frequency (TF-IDF) method can refer to a statistical method used for text mining and information retrieval, which is used to evaluate the importance of a word in a group of documents; the term frequency TF calculation can represent the frequency of occurrence of phrases of each inspection area, inspection unit, and inspection point in specific laws, regulations, standards and specifications; the inverse document frequency IDF calculation can represent the distribution of phrases of each inspection area, inspection unit, and inspection point in all corresponding laws, regulations, standards and specifications of a single industry.

[0149] In a possible implementation, the TF calculation can be expressed as follows:

[0150]

[0151] Among them, N w It can refer to the number of times the phrases of each inspection area, inspection unit, and inspection point appear in the text; N refers to the total number of words in the text.

[0152] In a possible implementation, the following expression can be used to represent IDF calculation:

[0153]

[0154] Among them, Y can refer to the total number of texts of laws, regulations, standards and specifications of a single industry. W Refers to the total number of texts containing the current phrase.

[0155] In a possible implementation, the TF-IDF calculation can be expressed as follows:

[0156] TF-IDF=TF W ×IDF W

[0157] The product of word frequency and inverse document frequency is used to represent the importance of the phrases of each inspection area, inspection unit, and inspection point in all corresponding laws, regulations, standards and specifications of a single industry.

[0158] In a possible implementation, the m-dimensional zero vector is determined through multi-hot encoding, and the elements of the original vector are replaced with the calculated TF-IDF values ​​of the zero vector in turn according to the preset industry order, and re-assigned to form a new m-dimensional vector, which may include:

[0159] Through multi-hot encoding, an m-dimensional zero vector is constructed, where m is the number of industry categories in the specification. According to the order of the categories in the specification, the elements of the original vector are replaced with the calculated TF-IDF values ​​and re-assigned to form a new m-dimensional vector.

[0160] Here, multi-hot encoding can be an encoding method for representing discrete features and is used to handle multiple classification labels or multi-label classification problems.

[0161] In a possible implementation, according to the inspection area, the inspection unit, the inspection point and the corresponding specific process, a preset encoding method is used to determine the process feature vector, which may include:

[0162] Extract the process names that may be involved in various industries and enterprises from the laws, regulations, rules, standards and specifications corresponding to each industry field, and arrange the extracted names in alphabetical order; the process names may include but are not limited to: surface treatment and coating, cleaning and neutralization; calculate the co-occurrence of the inspection area, inspection unit and inspection point names and the process names in the text through the Pointwise Mutual Information (PMI) method; construct an n-dimensional zero vector through multi-hot encoding, and assign values ​​to the elements in the zero vector in turn according to the calculated PMI value in a pre-arranged order for each inspection area, inspection unit and inspection point to form a new n-dimensional vector; where n is the number of all processes; normalize the n-dimensional vector and determine the normalized n-dimensional vector as the semantic feature embedding vector p.

[0163] Specifically, the following formula is used to express PMI calculation:

[0164]

[0165] Among them, P(x, y) can represent the probability of co-occurrence in all texts; P(x) and P(y) can represent the probability of the inspection area, inspection unit, inspection point name and process name appearing alone in the text respectively.

[0166] Through the above method, the system can improve the efficiency of users filling in inspection areas and selecting inspection points by reducing the steps of manual input and selection; by gradually carrying out hidden danger investigation and control work according to the guidance, the subjectivity and complexity of the investigation and control process can be reduced, and the accuracy and efficiency of hidden danger investigation can be improved.

[0167] In a possible implementation, all checkpoints are prioritized according to the characterization values ​​according to an intelligent recommendation algorithm based on the characterization values, and the checkpoints whose priority order meets the preset conditions are recommended to the user to assist the user in determining the troubleshooting and governance plan, which may include:

[0168] Through an intelligent recommendation algorithm based on characterization values, characterization values ​​are defined according to at least one preset dimension; weights corresponding to each preset dimension are set respectively; preset adjustment coefficients are set according to the type of investigation and governance plan; recommended values ​​corresponding to each inspection node are determined according to each preset dimension, weight and preset adjustment coefficient; priority is sorted for each inspection node according to the recommended value, and inspection points with a priority greater than a preset priority threshold are recommended to users.

[0169] Specifically, defining the characterization value according to at least one preset dimension through the intelligent recommendation algorithm based on the characterization value may include:

[0170] By using an intelligent recommendation algorithm based on characterization values, characterization values ​​are defined according to at least one preset dimension, which may include: defining characterization values ​​according to four preset dimensions, which may include: severity dimension, frequency dimension, time sensitivity dimension, and seasonality dimension. Among them, the severity dimension can characterize the severity of an accident after any inspection node occurs; if it may cause a particularly serious accident, the characterization value e is 4; if it causes a major accident, the characterization value e is 3; if it causes a major accident, the characterization value e is 2; if it causes a general accident, the characterization value e is 1; the definition and division of particularly serious accidents, major accidents, major accidents, and general accidents can be determined based on the legal norms stored in the enterprise knowledge base.

[0171] The frequency dimension can represent the frequency of hidden dangers detected by any inspection node; the frequency dimension representation value can be calculated using the following formula:

[0172]

[0173] Among them, f can represent the frequency dimension; p can represent the number of times the hidden dangers have been detected at the checkpoint in history; μ can represent the average number of times hidden dangers have been detected at all spread points under the inspection unit corresponding to the checkpoint; σ can represent the standard deviation of the number of times hidden dangers have been detected at all checkpoints under the inspection unit corresponding to the checkpoint.

[0174] The time sensitivity dimension can represent the difference between the last time a hidden danger was detected at any inspection node and the current date. The time sensitivity dimension representation value can be calculated using the following formula:

[0175] g tmp =d max -(t n -t p )

[0176] where d max It can represent the difference in days between the current date and the earliest hidden danger discovery date of the checkpoint; t n Can be the current date; p This may be the date when the checkpoint was last checked.

[0177] Considering all the checkpoints under the same inspection unit and performing normalization, the determined time sensitivity dimension representation value can be expressed as follows:

[0178]

[0179] Among them, g max It refers to the maximum time sensitivity of all inspection points in the same inspection unit without normalization; g minIt refers to the minimum time sensitivity of all inspection points under the same inspection unit without normalization; normalization can refer to a data preprocessing technology that converts data in different numerical ranges to the same range.

[0180] The seasonal dimension can refer to the relationship between the season of hidden danger distribution at any checkpoint and the current season. The higher the representation value, the stronger the relationship.

[0181] Define the months corresponding to spring as March, April and May, the months corresponding to summer as June, July and August, the months corresponding to autumn as September, October and November, and the months corresponding to winter as December, January and February; the seasonal dimension representation value can be calculated using the following formula:

[0182]

[0183] Among them, F s (k) is the number of times the hidden dangers are detected by the inspection node k in season s, It represents the number of times the inspection node k detects hidden dangers in the current season. Considering all the inspection points under the same inspection unit, normalization is performed and the seasonal dimension representation value can be expressed as follows:

[0184]

[0185] Here, Converted to a normalized value h, which is between 0 and 1.

[0186] In a possible implementation, respectively setting the weights corresponding to the preset dimensions may include:

[0187] Determine the severity weight W corresponding to the severity dimension e is 0.35, and the frequency weight W corresponding to the frequency dimension f The time sensitivity weight W corresponding to the time sensitivity dimension is 0.25. g The seasonal weight W corresponding to the seasonal dimension is 0.25. h The weights can be adjusted dynamically according to the actual needs of users.

[0188] In a possible implementation, setting a preset adjustment coefficient according to the type of the troubleshooting and governance plan may include:

[0189] If the department to which the checkpoint belongs is a key department, the preset adjustment coefficient C is set to 1.2; if a general work-related accident has occurred at the checkpoint, the preset adjustment coefficient C is set to 1.5; if a fatal accident has occurred at the checkpoint, the preset adjustment coefficient C is set to 2.

[0190] Further, determining the recommended value corresponding to each check node according to each preset dimension, weight and preset adjustment coefficient may include: calculating the recommended value using the following formula:

[0191] r=eW e ×W f × g × h ×C

[0192] The checkpoints are prioritized according to the recommended value r corresponding to each checkpoint.

[0193] In a possible implementation, the department, department operators and department responsible persons corresponding to the investigation and governance plan are determined, and the investigation and governance tasks are issued to the departments so that the department operators can perform the investigation and governance tasks, and the department responsible persons can accept the investigation and governance tasks, which may include:

[0194] Receive the departments, department operators and department heads determined by the user in the investigation and management plan, and send the investigation and management tasks to the department operators, so that the department operators can make on-the-spot rectifications or set rectification deadlines for the inspection nodes in the investigation and management tasks according to the control measures, and complete the rectification within the rectification deadline; send the investigation and management tasks to the department responsible persons, so that the department responsible persons can inspect and accept the investigation and management tasks, and receive the hidden danger images after treatment uploaded by the department responsible persons.

[0195] Through the above method, the target enterprise can divide the regions and sub-regions according to the spatial area logic, link the inspection units and inspection points under each area step by step, automatically match the corresponding control measures, form a target enterprise-inspection point list-investigation and governance guidelines, and distribute the investigation and governance tasks to relevant departments and department responsible persons; by decomposing the hidden danger investigation locations step by step and implementing the investigation and governance tasks layer by layer, the scientific, refined and precise hidden danger investigation and governance can be achieved.

[0196] In some embodiments, determining the inspection unit and the inspection point corresponding to the inspection area according to the semantic feature embedding vector, the industry feature vector and the process feature vector includes:

[0197] According to the semantic feature embedding vector, the industry feature vector and the process feature vector, the comprehensive embedded feature vector corresponding to each inspection area, inspection unit and inspection point is determined; according to the comprehensive embedded feature vector, the relative distance between the inspection area and the inspection unit, and the inspection area and the inspection point is determined by using the cosine similarity calculation method; according to the relative distance, the similarity between the inspection area and the inspection unit, and the inspection area and the inspection point is determined; wherein the similarity is inversely proportional to the relative distance; if the similarity between the inspection unit, the inspection point and the inspection interval is respectively greater than the first preset threshold, the inspection unit, the inspection point and the inspection interval are automatically associated; if the similarity between the inspection unit, the inspection point and the inspection interval is respectively greater than the second preset threshold, the inspection unit and the inspection point are highlighted; wherein the first preset threshold is greater than the second preset threshold; according to the inspection area and the associated determined inspection unit and inspection point, a list of inspection nodes is determined.

[0198] Specifically, according to the semantic feature embedding vector, the industry feature vector and the process feature vector, determining the comprehensive embedding feature vector corresponding to each inspection area, inspection unit and inspection point respectively may include: using the following formula to calculate the comprehensive embedding feature vector v of each inspection area, inspection unit and inspection point name:

[0199]

[0200] in, Can represent vector concatenation operations.

[0201] In a possible implementation, determining the similarity between the inspection area and the inspection unit, or between the inspection area and the inspection point according to the relative distance may include: calculating the similarity using the following formula:

[0202]

[0203] Among them, v a ,v b It can represent the inspection area and inspection unit or the inspection area and inspection point.

[0204] In a possible implementation, the inspection units and inspection points with a similarity greater than 0.8 to the inspection area are recommended as priority and automatically associated when the user fills in the area information; the inspection units and inspection points with a similarity greater than 0.6 to the inspection area are recommended as alternatives and highlighted when the user fills in the area information. For example, the first preset threshold may include but is not limited to 0.8, and the second preset threshold may include but is not limited to 0.6.

[0205] Through the above method, by calculating the embedded feature vector and cosine similarity, the system can provide users with accurate recommendations; making the selection of inspection items more scientific and reasonable, and reducing the influence of subjective judgment.

[0206] In some embodiments, the inspection units, inspection points and control measures corresponding to all industry fields are obtained through the enterprise knowledge base, including:

[0207] Determine the information collection target, and collect preset safety information from the information collection target according to the preset information collection frequency; wherein the preset safety information includes but is not limited to: laws and regulations, standards and specifications, investigation and governance procedures, hidden danger investigation records and governance measures text; use named entity recognition, text classification and cluster analysis to determine the inspection units, inspection points and control measures corresponding to the preset safety information; according to the preset database form structure, import the inspection units, the inspection points and the control measures into the enterprise information database.

[0208] In one possible implementation, the information collection targets may include, but are not limited to: national and local laws and regulations, standard specification websites, and security management platforms such as security supervision, double prevention mechanisms, etc. currently in operation in various provinces and cities.

[0209] In a possible implementation, named entity recognition, text classification and cluster analysis are used to determine the inspection unit, inspection point and control measures corresponding to the preset security information, which may include:

[0210] Through the pre-trained named entity recognition neural network model, inspection units, checkpoint entities and control measures entities are extracted from the captured text data according to the preset extraction labels; the extracted inspection unit entities are automatically classified into public inspection units and specific inspection units through the text classification algorithm based on deep learning; the extracted checkpoint entities are automatically classified into general checkpoints, public checkpoints and specific checkpoints; the extracted control measures text is automatically classified according to three indicators: technical and engineering measures, personnel quality and system management measures, and individual protection and emergency management measures;

[0211] Through entity alignment and cluster analysis algorithms, the extracted and classified inspection units, checkpoints, and control measures are semantically merged, and combined with expert judgment to retain the optimal representation of the same semantic entities to form structured knowledge data.

[0212] Specifically, the pre-trained named entity recognition neural network model may include but is not limited to: a named entity recognition model based on BERT-BiLSTM-CRF, that is, a named entity recognition model based on transformer bidirectional encoder representation technology-improved recurrent neural network-sequence labeling model, and the model can be pre-trained and fine-tuned through security management big data; the extracted labels are identified using the BIO labeling method, and the BIO labeling method can be a labeling method for sequence labeling tasks.

[0213] The text classification algorithm can be a BERT-based text classification model, which is fine-tuned through the labeled control measures text. The clustering analysis algorithm can be a short text clustering method based on the Dirichlet Multinomial Mixture Model (DMM), which automatically infers the number of clusters; wherein, DMM can be a text clustering algorithm based on a probability model.

[0214] In a possible implementation, according to a preset database form structure, importing the inspection unit, the inspection point, and the control measure into an enterprise information database may include:

[0215] According to the preset database form structure, the structured data are imported into the relational database respectively to form a public inspection unit library, a specific inspection unit library, a general inspection point library, a public inspection point library, a specific inspection point library, and a control measures library; wherein the relational database may be a database system based on a relational model, which may include but is not limited to a MySQL relational database.

[0216] In one possible implementation, the correspondence between inspection units and inspection points, and between inspection points and control measures is determined, and an inspection unit-inspection point-control measure relationship network with nodes and edges as basic elements is constructed; the corresponding data in the relational database is imported into a graph database according to the relational network, and the visualization of the relational network is realized through secondary development of the graph database; wherein the graph database may include but is not limited to the Neo4j graph database.

[0217] In one possible implementation, the captured laws, regulations, standards and specifications can be analyzed, the industry types of each enterprise can be analyzed, the industry type name, definition, and scope can be determined, and the structured data can be imported into the relational database; according to the characteristics of the enterprise industry field, the enterprise-level preset form templates can be customized, and the structured form data can be imported into the relational database; the update frequency of the enterprise knowledge base can be set, and text can be captured from the information collection target at regular intervals, compared and analyzed with the existing content in the library, and the content in the library can be supplemented, deleted, replaced and updated.

[0218] Through the above method, the whole process from text collection to enterprise knowledge base construction is automated, the efficiency and quality of data processing are improved, and reliable risk management and decision-making support are provided for enterprises to detect hidden dangers; the enterprise knowledge base constructed and updated by advanced methods is comprehensive and accurate, providing factual and theoretical support for hidden danger detection, and can effectively improve the digitalization and informatization level of enterprise safety production management.

[0219] In a specific embodiment, Figure 2 This is a schematic diagram of the structure of another hidden danger investigation and management system provided by an embodiment of the present invention, as shown in the figure:

[0220] The hidden danger investigation and management system includes: a terminal layer 1, an application layer 2 and a database layer 3; the terminal layer 1 includes: a desktop terminal 11 and a mobile terminal 12; the application layer 2 includes a regional information collection module 21, an investigation and management module 22 and a safety management module 23;

[0221] The regional information acquisition module 21 includes: a regional information input submodule 211, an inspection unit association submodule 212, an inspection point association submodule 213, a regional inspection point list generation submodule 214, a regional information three-dimensional model generation submodule 215 and an inspection point QR code generation submodule 216;

[0222] The investigation and governance module 22 includes: an investigation and governance plan creation submodule 221, an investigation and governance guideline generation submodule 222, an investigation and governance task reception submodule 223, an investigation and governance task implementation submodule 224 and an investigation and governance record storage submodule 225; wherein the investigation and governance guideline generation submodule 222 further includes an investigation and governance path submodule 2221 and an investigation and governance text generation submodule 2222;

[0223] The safety management module 23 includes: an overview display submodule 231, a number statistics submodule 232 and a regional information three-dimensional model display submodule 233; the database layer 3 includes: an enterprise knowledge base 31 and an enterprise information base 32; the enterprise knowledge base 31 includes: an industry type library 311, a public inspection unit library 312, a specific inspection unit library 313, a general inspection point library 314, a specific inspection point library 315, a public inspection point library 316, a control measures library 317, an information collection template library 318 and an enterprise regional risk knowledge map 319; the enterprise information library 32 includes: an enterprise basic information library 321 and an enterprise hidden danger information library 322.

[0224] In a specific embodiment, Figure 4 FIG. 1 is a flow chart of another troubleshooting and management method provided by an embodiment of the present invention; Figure 4 As shown, the process includes the following steps:

[0225] Step S401, determine whether it is the first time to conduct an investigation; if so, go to step S402, if not, go to step S403; wherein, the preset form template uploaded by the user is determined through the desktop terminal 11 to determine whether the target enterprise is conducting an investigation for the first time.

[0226] Step S402, determining the inspection area, associating the inspection units and inspection points in the inspection area; wherein, through the preset form template uploaded by the user, determining the area information and basic information of the target enterprise, determining the inspection area corresponding to the target enterprise, recommending the inspection units and inspection points in the inspection area to the user, and determining the inspection point list;

[0227] Step S403, determining whether there are changes in the inspection area, inspection unit and inspection point; wherein, checking whether there are changes in the inspection area, inspection unit and inspection point; if there are no changes, specifying the investigation and treatment plan corresponding to the target enterprise; if there are changes, re-determining the inspection area, inspection unit and inspection point;

[0228] Step S404, formulate an inspection and management plan; wherein, the inspection area and inspection unit to be inspected determined by the user are received, inspection nodes are recommended to the user, and the user is assisted in determining the department operator and department responsible person;

[0229] Step S405, executing the investigation and management task; wherein the investigation and management task is issued to the department operator for investigation and management;

[0230] Step S406, conduct investigation, governance and acceptance; wherein the investigation and governance task is assigned to the responsible person of the department for investigation, governance and acceptance;

[0231] Step S407, determine whether the hidden danger has been eliminated; if not, go to step S404, if so, end the process.

[0232] In a specific embodiment, Figure 5 Schematic diagram of the process of constructing an enterprise knowledge base of a troubleshooting and management method provided by an embodiment of the present invention; Figure 5 As shown, the process includes the following steps:

[0233] Step S501, collecting laws and regulations, standards and specifications, troubleshooting and management procedures, hidden danger troubleshooting records and management measures text;

[0234] Here, we collect the national and local laws and regulations, standards and specifications websites, as well as the safety management platforms such as safety supervision and double prevention mechanisms currently in operation in various provinces and cities, corresponding laws and regulations, standards and specifications, investigation and governance procedures, hidden danger investigation records and governance measures texts;

[0235] Step S502, analyzing the industry type and sub-sectors of each enterprise;

[0236] Step S503, customizing the preset form template corresponding to the enterprise; wherein the preset form template filled in and uploaded by the user is customized;

[0237] Step S504, extracting inspection unit entities, inspection point entities and control measure entities based on the pre-trained named entity recognition neural network model;

[0238] Here, by pre-training the named entity recognition neural network model, the inspection unit, checkpoint entity, and control measure entity are extracted from the captured text data according to the preset extraction labels;

[0239] Step S505: Classify the inspection units, inspection points and control measures according to preset labels based on a deep learning text classification algorithm;

[0240] Here, through the text classification algorithm based on deep learning, the extracted inspection unit entities are automatically classified into public inspection units and specific inspection units; the extracted checkpoint entities are automatically classified into general checkpoints, public checkpoints, and specific checkpoints; the extracted control measures texts are automatically classified according to three indicators: technical and engineering measures, personnel quality and system management measures, and individual protection and emergency management measures;

[0241] Step S506, importing the relational database;

[0242] Step S507, importing the graph database; wherein the corresponding relationships between the inspection unit and the inspection point, and between the inspection point and the control measure are determined, and a relationship network of the inspection unit, the inspection point, and the control measure is constructed with nodes and edges as basic elements; in the relational database, the corresponding data is imported into the graph database according to the relational network, and the visualization display of the relational network is realized through secondary development of the graph database;

[0243] Step S508, determine whether the enterprise knowledge base needs to be updated; if so, go to step S501, if not, end the process.

[0244] In a specific embodiment, Figure 6 Schematic diagram of a regional information collection process of a troubleshooting and management method provided by an embodiment of the present invention; Figure 6 As shown, the process includes the following steps:

[0245] Step S601, dividing the area and sub-area, wherein the area and sub-area divided by the user are received;

[0246] Step S602, associating a public inspection unit; wherein, according to the actual production and management situation of the enterprise, associating the public inspection unit of the entered area;

[0247] Step S603, associating with specific inspection units; wherein, associating with specific inspection units of the entered area according to the industry and sub-sectors to which the enterprise belongs;

[0248] Step S604, associating the department, department operator and department responsible person corresponding to the inspection unit;

[0249] Step S605, associating public inspection points; wherein, according to the actual production and management situation of the enterprise, associating public inspection points under the public inspection unit;

[0250] Step S606, associating with a specific checkpoint;

[0251] Step S607, associating with the general checkpoint position;

[0252] Here, common inspection points that may be available in both the public inspection unit and the specific inspection unit are associated.

[0253] It should be noted that: the hidden danger investigation and treatment system provided in the above embodiment only uses the division of the above program modules as an example to illustrate when implementing the corresponding investigation and treatment method. In actual application, the above processing can be assigned to different program modules as needed, that is, the internal structure of the device can be divided into different program modules to complete all or part of the above-described processing. Figure 3 The embodiments of the method shown belong to the same concept, and the specific implementation process is detailed in the method embodiments, which will not be repeated here.

[0254] The embodiment of the present invention also provides a computer device having the above Figure 2 The troubleshooting and management device shown.

[0255] See also Figure 7 , Figure 7 is a schematic diagram of the structure of another test case verification device provided by an embodiment of the present invention, such as Figure 7 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 7 A processor 10 is taken as an example.

[0256] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0257] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0258] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0259] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0260] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0261] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0262] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A hidden danger investigation and management system, characterized in that: include: Terminal layer, application layer and database layer; Through the enterprise knowledge base in the database layer, obtain and store the inspection units, inspection points and control measures corresponding to all industry fields; According to the basic information and regional information of the target enterprise obtained by the terminal layer, the basic information, the regional information and the enterprise knowledge base are pre-associated in the application layer, and the inspection unit and the inspection point in the enterprise knowledge base whose similarity is greater than a preset threshold are pre-associated with the target enterprise; wherein the basic information of the target enterprise represents information used to describe the target enterprise, and the basic information includes at least one of the following: the national economic classification, industry field, number of employees, organizational structure, and personnel information of the target enterprise; the regional information includes at least one of the following: regional name, regional description, regional scope, and regional area; According to the inspection unit and the inspection point, a list of inspection points corresponding to the target enterprise is determined in the application layer; the list of inspection points includes but is not limited to: an inspection area, the inspection unit and the inspection point; wherein the inspection area is determined according to the area information; Determine the optimal investigation path according to the checkpoint list, determine the control measures corresponding to the checkpoint list from the enterprise knowledge base, and determine the investigation and governance guidelines corresponding to the target enterprise according to the optimal investigation path and the control measures; Displaying the checkpoint list and the investigation and management guidelines to the target enterprise at the terminal layer; The pre-association processing of the basic information, the regional information and the enterprise knowledge base in the application layer includes: According to the hierarchical relationship, the inspection units and inspection points corresponding to the inspection area are respectively associated and determined, and the inspection point list is determined in the form of inspection area-inspection unit-inspection point; wherein, Input the name corresponding to the inspection area, the name corresponding to the inspection unit, and the name corresponding to the inspection point into a preset language representation model to determine a semantic feature embedding vector; According to the correlation between the inspection area, the inspection unit, the inspection point and at least one national economic classification, a preset coding method is used to determine an industry field feature vector; According to the inspection area, inspection unit, inspection point and corresponding specific process, a preset encoding method is used to determine the process feature vector; The inspection unit and the inspection point corresponding to the inspection area are determined according to the semantic feature embedding vector, the industry field feature vector and the process feature vector.

2. The system according to claim 1, characterized in that The application layer includes: regional information collection module, investigation and management module and security management module; Determine the regional information of the target enterprise through the regional information acquisition module, pre-associate the regional information with the basic information, determine the inspection area, inspection unit and inspection point corresponding to the target enterprise according to the industry field characteristics and actual production situation of the target enterprise, and generate a list of inspection points corresponding to the target enterprise; the basic information includes but is not limited to: national economic classification, industry field, number of employees and personnel information; The inspection and management module determines the inspection area, inspection unit and inspection point of the target enterprise to be inspected according to the inspection point list determined by the area information collection module, and determines and issues the inspection and management tasks and inspection and management guidelines; Through the safety management module, the risk situation and investigation and control situation of the target enterprise for the checkpoint list are counted and displayed.

3. The system according to claim 2, characterized in that The regional information acquisition module includes: a regional information input submodule, an inspection unit association submodule, a checkpoint association submodule, a regional checkpoint list generation submodule, a regional information three-dimensional model generation submodule and a checkpoint QR code generation submodule; The regional information input submodule uses a preset computer vision model to perform image recognition semantic conversion based on the received regional image information of the target enterprise to determine the regional information of the target information; The inspection unit association submodule and the inspection point association submodule pre-associate the regional information, basic information and enterprise knowledge base, and determine the inspection units and inspection points with similarity greater than a preset threshold to pre-associate with the target enterprise according to the similarity analysis of the preset vectors; wherein the inspection unit includes at least one of the following: a public inspection unit or a specific inspection unit, and the inspection point includes at least one of the following: a general inspection point, a public inspection point or a specific inspection point; the preset vector includes but is not limited to: a semantic feature embedding vector, an industry feature vector and a process feature vector; The regional inspection point list generation submodule generates a structured inspection point list in the form of inspection area-inspection unit-inspection point by calling the regional information, the inspection unit and the inspection point; Generate a three-dimensional model including the inspection area, the inspection unit and the inspection points according to the inspection point list through the area information three-dimensional model generation submodule; The two-dimensional code information corresponding to each checkpoint in the checkpoint list is generated by the checkpoint two-dimensional code generation submodule; the two-dimensional code information is used to mark the real-time status of the checkpoint.

4. The system according to claim 3, characterized in that The investigation and governance module includes: an investigation and governance plan creation submodule, an investigation and governance guide generation submodule, an investigation and governance task reception submodule, an investigation and governance task implementation submodule, and an investigation and governance record storage submodule; The investigation and governance plan creation submodule accesses the regional information of the target enterprise, and prioritizes all inspection points according to the characterization value based on the intelligent recommendation algorithm based on the characterization value, determines the inspection area, inspection unit and inspection point to be investigated of the target enterprise, and allocates the investigation and governance tasks of each inspection unit to different departments of the target enterprise; The investigation and management guidance generation submodule includes an investigation and management path submodule and an investigation and management text generation submodule; According to the investigation and management path submodule, a preset path planning model is used to automatically plan the optimal investigation path; According to the investigation and management text generation submodule, a preset text generation model is used to retrieve corresponding control measures from the enterprise knowledge base according to all checkpoints included in the optimal investigation path, and generate an investigation and management guidance text; The investigation and governance task receiving submodule is connected to the investigation and governance plan creation submodule to receive the investigation and governance task corresponding to the department and the governance acceptance task corresponding to the investigation and governance task issued by the investigation and governance plan creation submodule; The inspection and treatment task implementation submodule is connected to the inspection and treatment plan creation submodule and the inspection and treatment guide generation submodule to determine the hidden danger situation of the checkpoint, the description information of the hidden danger situation and the image information of the hidden danger situation; wherein the hidden danger situation includes one of the following: hidden danger, no hidden danger or no hidden danger; The investigation and management record storage submodule disassembles the historical investigation and management plans and historical investigation and management tasks, and stores the disassembled structured data in the enterprise information library of the database layer.

5. The system according to claim 4, characterized in that The security management module includes: an overview display submodule, a number statistics submodule and an area information three-dimensional model display submodule; The overview display submodule is used to display the preset investigation and management information corresponding to the target enterprise at the terminal layer; the preset investigation and management information is used to describe the investigation and management status of the enterprise; The number statistics submodule accesses the investigation and management situation to determine the investigation and management change trend of the target enterprise; The regional information three-dimensional model display submodule is connected to the regional information three-dimensional model generation submodule and the inspection and management record storage submodule, and displays the specific locations of the inspection area, inspection unit and inspection point of the target enterprise in the form of a three-dimensional model, and uses red, yellow and green to represent the inspection and management status of the inspection point; wherein, the red color indicates that the hidden dangers of the last inspection have not been completely managed within the preset specified time, the yellow color indicates that the hidden dangers of the last inspection are being managed, and the green color indicates that the hidden dangers of the last inspection have been completely managed or no hidden dangers were found in the last inspection.

6. The system according to claim 5, characterized in that The database layer includes: enterprise knowledge base and enterprise information base; The enterprise knowledge base includes industry type libraries, public inspection unit libraries, specific inspection unit libraries, general inspection point libraries, specific inspection point libraries, public inspection point libraries, control measures libraries, information collection template libraries, and enterprise regional risk knowledge maps corresponding to all industry fields; The enterprise information database includes an enterprise basic information database and an enterprise hidden danger information database. The enterprise basic information database is used to store basic information corresponding to the target enterprise, and the enterprise hidden danger information database is used to store hidden danger information corresponding to the target enterprise.

7. The system according to claim 6, characterized in that The terminal layer includes: a desktop terminal and a mobile terminal; Receiving, through the desktop terminal, the region information uploaded by the target enterprise according to a preset form template or region image information; View the regional information QR code and the three-dimensional model determined according to the regional information through the desktop terminal, set an investigation and management plan, and automatically generate a corresponding checkpoint list and investigation and management guidelines through the investigation and management plan; According to the risk situation and investigation and management situation of the target enterprise for the checkpoint list, viewing the preset investigation and management information on the desktop; Receive the inspection and management task through the mobile terminal, and check the inspection and management guidance corresponding to the checkpoint list.

8. A method for troubleshooting and managing hidden dangers applied to the hidden danger troubleshooting and managing system according to any one of claims 1 to 7, characterized in that: The method comprises: Obtain inspection units, inspection points and control measures corresponding to all industry fields through the enterprise knowledge base; Receiving the area information of the target enterprise uploaded by the user according to the preset form template or area image information, and determining the inspection area according to the area information; The inspection units and inspection points corresponding to the inspection area are respectively associated and determined according to the hierarchical relationship, and the inspection point list is determined in the form of inspection area-inspection unit-inspection point, including: Input the name corresponding to the inspection area, the name corresponding to the inspection unit, and the name corresponding to the inspection point into a preset language representation model to determine a semantic feature embedding vector; According to the correlation between the inspection area, the inspection unit, the inspection point and at least one national economic classification, a preset coding method is used to determine an industry field feature vector; According to the inspection area, the inspection unit, the inspection point and the corresponding specific process, the preset encoding method is used to determine the process feature vector; Determine the inspection unit and the inspection point corresponding to the inspection area according to the semantic feature embedding vector, the industry field feature vector and the process feature vector; According to the intelligent recommendation algorithm based on the characterization value, all the checkpoints are prioritized according to the characterization value, and the checkpoints whose priority rankings meet the preset conditions are recommended to the user, so as to assist the user in determining the inspection and management plan; Determine the inspection and management guidance text according to the control measures corresponding to the checkpoints in the inspection and management plan, and generate the optimal inspection path according to the checkpoints; Determine the troubleshooting and governance guidelines corresponding to the troubleshooting and governance plan according to the troubleshooting and governance guidelines text and the optimal troubleshooting path; Determine the department, department operators and department responsible persons corresponding to the investigation and governance plan, issue the investigation and governance tasks to the departments, make the department operators perform the investigation and governance tasks, and the department responsible persons inspect and accept the investigation and governance tasks.

9. The method according to claim 8, characterized in that The determining the inspection unit and the inspection point corresponding to the inspection area according to the semantic feature embedding vector, the industry field feature vector and the process feature vector includes: According to the semantic feature embedding vector, industry feature vector and process feature vector, the comprehensive embedding feature vector corresponding to each inspection area, inspection unit and inspection point is determined; According to the comprehensive embedded feature vector, the relative distance between the inspection area and the inspection unit, and the relative distance between the inspection area and the inspection point are determined by using a cosine similarity calculation method; According to the relative distance, determining the similarity between the inspection area and the inspection unit, and between the inspection area and the inspection point; wherein the similarity is inversely proportional to the relative distance; If the similarities between the inspection unit, the inspection point and the inspection area are respectively greater than a first preset threshold, the inspection unit, the inspection point and the inspection area are automatically pre-associated; If the similarity between the inspection unit, the inspection point and the inspection area is respectively greater than a second preset threshold, the inspection unit and the inspection point are highlighted; wherein the first preset threshold is greater than the second preset threshold; The inspection point list is determined according to the inspection area and the inspection unit and the inspection point determined in association.

10. The method according to claim 9, characterized in that The inspection units, inspection points and control measures corresponding to all industry fields are obtained through the enterprise knowledge base, including: Determine an information collection target, and collect preset safety information from the information collection target according to a preset information collection frequency; wherein the preset safety information includes but is not limited to: laws and regulations, standards and specifications, investigation and management procedures, hidden danger investigation records and management measures text; Using named entity recognition, text classification and cluster analysis, determine the inspection units, inspection points and control measures corresponding to the preset security information; According to the preset database form structure, the inspection unit, the inspection point and the control measures are imported into the enterprise information database.

Images