Safety production risk identification method and system based on multi-dimensional risk factor quantification

Abstract

The application belongs to the technical field of enterprise safety production management, and provides a safety production risk identification method and system based on multi-dimensional risk factor quantification, determines risk type factors, single selection factors, multiple selection factors, equipment service life factors and next inspection date factors under risk points; determines comprehensive risk scores according to accumulation of the risk type factors, the single selection factors, the multiple selection factors, the equipment service life factors and the next inspection date factors; determines risk levels according to comparison of the comprehensive risk scores and threshold values; constructs a multi-dimensional risk factor quantification scoring mode based on severity and possibility, realizes automatic recommendation of the risk levels; respectively determines risk factors of different selection types, especially introduces objective time indexes such as equipment service life and inspection dates into risk quantification, realizes determination of comprehensive risk scores of multi-dimensional factor quantification, solves the shortcomings of subjective qualitative scoring modes in determining risk levels, and improves risk identification precision.

Description

Technical Field

[0001] This invention belongs to the field of enterprise safety production management technology, and in particular relates to a method and system for identifying safety production risks based on multi-dimensional risk factor quantification. Background Technology

[0002] In the field of enterprise safety production management, risk identification and hazard investigation mainly rely on manual experience. Safety inspectors judge risk points and hazard categories based on their personal experience, lacking standardized intelligent assistance methods. When faced with complex on-site situations, inspectors need to manually consult a large number of regulations, standards, and checklists, which is inefficient and prone to omissions.

[0003] Currently, risk level determination mostly relies on subjective qualitative scoring methods, lacking quantitative calculation models based on multi-dimensional risk factors. Specifically, this manifests in: failure to decompose risks into independent dimensions for separate quantification; failure to differentiate calculation logic for factors of different selection types; and the absence of a dual-track verification mechanism between self-assessed risk levels and system-calculated risk levels, resulting in poor effectiveness in identifying safety production risks. Summary of the Invention

[0004] To address the aforementioned problems, this invention proposes a method and system for identifying safety production risks based on multidimensional risk factor quantification. This invention constructs a multidimensional risk factor quantification scoring method based on severity and probability, enabling automatic recommendation of risk levels. It identifies risk factors for different selection types, particularly incorporating objective time indicators such as equipment service life and inspection date into risk quantification, thus achieving comprehensive risk score determination based on multidimensional factor quantification. This overcomes the shortcomings of subjective qualitative scoring methods in determining risk levels and improves the accuracy of risk identification.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solution: In a first aspect, the present invention provides a method for identifying safety production risks based on multidimensional risk factor quantification, comprising: Determine whether to perform quantitative identification based on the acquired identification type; If so, determine the risk type factor, single-choice factor, multiple-choice factor, equipment service life factor, and next inspection date factor under the risk point; otherwise, proceed to standardized identification. Among them, the risk type factor, single-choice factor, and multiple-choice factor are determined by quantifying the risk into two dimensions: severity and probability. The equipment service life factor is determined based on the relationship between the current year and the year of commencement of use. The next inspection date factor is determined based on the relationship between the current date and the next inspection date. The comprehensive risk score is determined by summing the risk type factor, single-choice factor, multiple-choice factor, equipment service life factor, and next inspection date factor. The risk level is determined by comparing the comprehensive risk score with the threshold.

[0006] Furthermore, the risk type factor adopts a combined determination mode of initial preset value and user manually selected value, and the contribution value of each part is based on the product of severity and probability.

[0007] Furthermore, the determination of the initial preset value includes: traversing all initialized preset data items; for each item, obtaining the corresponding severity value and probability value; if both exist, the product of the severity value and probability value is added to the total score; the determination of the user's manually selected value includes: splitting the user-input string into multiple independent selection values ​​by commas, traversing the multiple independent selection values, for each selection value, obtaining the corresponding severity value and probability value; if both exist, the product of the severity value and probability value is added to the total score.

[0008] Furthermore, the determination of the single-selection factor includes: retrieving the corresponding sub-item indicator data from the database according to the specified sub-item type and data value, and storing the results in a key-value pair set; extracting two numerical fields representing severity and probability from the obtained data set; converting the extracted severity and probability into numerical types and performing multiplication to obtain the current single-selection factor score.

[0009] Furthermore, the determination of the multiple selection factors includes: splitting the input string by commas to obtain a string array; sequentially extracting each substring from the array and assigning it to a temporary variable for processing; for the current temporary variable, retrieving and returning a mapping object containing relevant information from the database; obtaining the values ​​with the keys of severity and probability from the mapping object; converting these two values ​​to decimal numbers and determining the individual score by multiplying severity by probability; and accumulating the individual scores.

[0010] Furthermore, the equipment service life factor = ((current year - year of start of use) × severity parameter) / current year; the next inspection date factor = 1 / number of days; the number of days is equal to the difference between the current date and the next inspection date.

[0011] Furthermore, the determination of the risk level includes: if the comprehensive risk score is less than a first threshold, the risk level is low risk; if the comprehensive risk score is greater than or equal to the first threshold and less than a second threshold, the risk level is moderate risk; if the comprehensive risk score is greater than or equal to the second threshold and less than a third threshold, the risk level is relatively high risk; if the comprehensive risk score is greater than or equal to the third threshold, the risk level is significant risk; the first threshold is less than the second threshold, and the second threshold is less than the third threshold.

[0012] Furthermore, the standardized identification includes: directly reading the pre-set standard risk level in the pre-assessed risk point template; and automatically mapping the risk level value corresponding to the standard risk level to the corresponding control level.

[0013] Furthermore, the method also includes: after determining the risk level, checking all risk units bound to the current risk point one by one, and comparing the existing level of the risk unit with the level of the risk point; if the current level of the risk unit is lower than the level of the risk point, directly updating the level of the risk unit to the level of the risk point; if the risk unit has never been set with any level before, directly using the current level of the risk point to complete the initialization assignment.

[0014] Secondly, the present invention also provides a safety production risk identification system based on multidimensional risk factor quantification, comprising: The data acquisition module is configured to determine whether to perform quantitative identification based on the acquired identification type. The judgment module is configured to: if yes, determine the risk type factor, single-choice factor, multiple-choice factor, equipment service life factor, and next inspection date factor under the risk point; otherwise, proceed to standardized identification; among them, the risk type factor, single-choice factor, and multiple-choice factor are determined by quantifying the risk into two dimensions: severity and probability; the equipment service life factor is determined based on the relationship between the current year and the year of commencement of use; the next inspection date factor is determined based on the relationship between the current date and the next inspection date; The comprehensive risk score determination module is configured to determine the comprehensive risk score by accumulating the risk type factor, single-choice factor, multiple-choice factor, equipment service life factor, and next inspection date factor. The risk level determination module is configured to determine the risk level based on a comparison between the comprehensive risk score and a threshold.

[0015] Thirdly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the safety production risk identification method based on multidimensional risk factor quantification described in the first aspect.

[0016] Fourthly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the program to implement the steps of the safety production risk identification method based on multidimensional risk factor quantification described in the first aspect.

[0017] Fifthly, the present invention also provides a computer program product, the computer program product comprising a computer program, which, when executed by a processor, implements the steps of the safety production risk identification method based on multidimensional risk factor quantification described in the first aspect.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention, under a quantitative identification mode, determines risk type factors, single-choice factors, multiple-choice factors, equipment service life factors, and next inspection date factors for each risk point. Then, it determines a comprehensive risk score by accumulating these factors. Finally, it determines the risk level by comparing the comprehensive risk score with a threshold. This invention constructs a multi-dimensional risk factor quantitative scoring method based on severity and probability, enabling automatic risk level recommendation. It identifies risk factors for different selection types, particularly incorporating objective time indicators such as equipment service life and inspection date into risk quantification, achieving comprehensive risk score determination through multi-dimensional factor quantification. This overcomes the shortcomings of subjective qualitative scoring methods in determining risk levels and improves the accuracy of risk identification. Attached Figure Description

[0019] The accompanying drawings, which form part of this embodiment, are used to provide a further understanding of this embodiment. The illustrative embodiments and their descriptions are used to explain this embodiment and do not constitute an improper limitation of this embodiment.

[0020] Figure 1 This is a flowchart of the method in Embodiment 1 of the present invention; Figure 2 This is a flowchart of the standardized identification pattern of Embodiment 1 of the present invention; Figure 3 This is a flowchart of the NLP hazard intelligent identification algorithm of Embodiment 1 of the present invention; Figure 4 This is a flowchart illustrating the inspection and hidden danger linkage process of Embodiment 1 of the present invention; Figure 5 This is a flowchart illustrating the automatic propagation of risk levels across organizational hierarchies in Embodiment 1 of the present invention. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0022] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0023] Example 1: Currently, in the field of enterprise safety production management, risk identification and hazard investigation mainly rely on the following methods: The system relies heavily on human experience: safety inspectors judge risk points and hazard categories based on personal experience, lacking standardized intelligent assistance methods. When faced with complex on-site situations, inspectors need to manually consult numerous regulations, standards, and checklists, which is inefficient and prone to omissions. Risk identification lacks intelligent semantic matching: most existing systems use simple keyword-based exact matching. When the hazard description entered by the inspector does not match the wording in the standard inspection items, the system cannot automatically identify the corresponding inspection items and regulatory basis, leading to a disconnect between hazard records and standard inspection items.

[0024] Risk level assessment lacks a multi-dimensional quantitative calculation model: Existing technologies mostly use subjective qualitative scoring methods (such as high, medium, and low levels) to determine risk levels, lacking a quantitative calculation model based on multi-dimensional risk factors (such as equipment service life, inspection cycle, working environment, hazardous substances, etc.). Specifically, this manifests as: failure to decompose risk into two independent dimensions, severity (S) and likelihood (L), for separate quantification; failure to differentiate the calculation logic for different selection types of factors (single choice, multiple choice, fill-in-the-blank date / year); and failure to establish a dual-track verification mechanism between self-assessed risk levels and system-calculated risk levels.

[0025] Risk levels do not propagate automatically upwards along organizational levels: once the risk level of a risk point is assessed, the risk levels of its associated risk unit, department, and even superior unit (parent company / group) cannot be automatically updated. They need to be manually summarized level by level, resulting in the regional risk level distribution map in the management dashboard being out of sync with the actual data.

[0026] There is an information gap in the inspection-hazard-rectification process: there is a lack of automated data flow mechanism between the problems found in safety inspections and the records of hazards and rectification tasks. Inspection results cannot be automatically generated into a hazard ledger, and the status of hazard rectification cannot be traced.

[0027] Lack of an adaptive screening and statistical mechanism for production cycles: The existing system adopts a uniform screening and statistical cycle for all enterprises, without taking into account the differences in production cycles of different enterprises, resulting in incorrect statistics of "not screened" on non-production days.

[0028] To solve at least one of the above problems, such as Figure 1As shown, this embodiment provides a safety production risk identification method based on multi-dimensional risk factor quantification. It upgrades the risk identification process from subjective qualitative analysis to a quantitative model driven by L·S (severity × probability). Simultaneously, it introduces Natural Language Processing (NLP) technology to achieve intelligent semantic matching of hazard descriptions, ultimately forming an intelligent management method covering the entire safety production chain, including pre-event risk identification, in-event hazard identification, and post-event investigation and management. S1. Construct a standardized risk factor knowledge base: The system has a pre-built risk factor knowledge base based on industry categories, including major risk categories, minor risk categories, risk factor type data structures, risk factor values, and the relationship between risk factors and minor risk categories.

[0029] The risk point categories define risk point types such as major equipment and facilities, key locations, special equipment, operational activities, confined spaces, and hazardous substances / materials within the industry. The risk point subcategories are specific risk point templates under each major category, with preset default risk levels (e.g., levels ranging from 1 to 4, corresponding to major, significant, moderate, and low risk, respectively), for use in the standardized identification model. The risk factor types define the name, factor type code, and selection type of each risk factor.

[0030] The selection types include single selection, multiple selection, fill-in-the-blank, year selection, image upload, year-month-day selection, drop-down multiple selection, and pop-up single selection (select risk unit), corresponding to codes 1 to 8 respectively. Among them, single selection means selecting one from multiple preset options; multiple selection means selecting multiple from multiple preset options; fill-in-the-blank means free text input; year selection is used to record year information such as the start time of device use; year-month-day selection is used to record time information such as the next inspection date; drop-down multiple selection represents a tree-structured multiple selection.

[0031] The risk factor values ​​are optional values ​​for each risk factor type, including two core quantitative fields: severity and probability. The severity weight value typically ranges from 0 to 30; the probability weight value typically ranges from 0 to 30.

[0032] The relationship between risk factors and risk point subcategories is used to define which risk factor data needs to be collected for each risk point subcategory, and is maintained through a relationship table.

[0033] S2. Risk Point Initialization and Identification (Risk Factor Data Collection): The system retrieves the industry category of the enterprise from the enterprise information table, and automatically loads the hierarchical tree structure of the risk point major category, risk point minor category and risk factor corresponding to the industry.

[0034] Load the list of major and minor risk points based on industry category; save the selected minor risk point data as a specific risk point instance, and initialize all risk factor associated records under that risk point. Each record includes the record primary key, the associated risk point ID, the associated risk factor ID, the risk factor value selected / entered by the user, the display name corresponding to the risk factor value selected / entered by the user, and the initial preset value, etc.

[0035] When a risk point is initialized, its status is set to "Unidentified," and its hierarchical control status is set to "Pending Configuration." Within the same risk unit, duplicate risk point subcategories are not allowed; if duplicates exist, the specific conflicting risk point name will be returned.

[0036] S3. Multidimensional Risk Factor L·S Quantitative Scoring Model and Risk Level Determination (Quantitative Identification Mode): A quantitative score is generated after the user fills in all the risk factor values ​​for the risk point.

[0037] S3.1 Preprocessing: The system first reads the identification type configured by the enterprise. If the value is "1", it enters the quantitative identification mode; otherwise, it enters the standardized identification mode in step S4.

[0038] S3.2, Traverse all risk factors and calculate them by category: Iterate through the list of all risk factor records under this risk point, and apply different score accumulation strategies to each risk factor based on its selection type and factor type: S3.2.1, Treatment of Risk Type Factors: The risk type factor is determined by a combination of initial preset values ​​and user-selected values. The contribution value of each part is based on the product of severity and probability, and is accumulated with four decimal places.

[0039] The determination of the initial preset values ​​includes: traversing all initialized preset data items; for each item, obtaining its corresponding severity value and probability value; if both exist, calculating their product, rounding the result to four decimal places, and adding it to the total score. The determination of the user-selected values ​​includes: splitting the user-input string into multiple independent selection values ​​by commas, traversing these selection values, and for each value, obtaining its corresponding severity value and probability value; if both exist, calculating their product, rounding the result to four decimal places, and adding it to the total score.

[0040] In this embodiment, both the initial preset value and the user-added selection value of the risk type are included in the L·S score accumulation, ensuring that both the inherent risk types recommended by the system and the risk types added by the enterprise are included in the comprehensive score.

[0041] S3.2.2, Processing of Single-Choice Factors: Optionally, the system, based on the specified sub-item type and data value, calls the mapper to retrieve the corresponding sub-item indicator data from the database and stores the results in a key-value pair set. From the retrieved dataset, two numerical fields representing severity and probability are extracted. The extracted severity and probability values ​​are converted to high-precision numerical types and multiplied to obtain the current single-selection factor score. During the calculation, the result is rounded to four decimal places to ensure numerical accuracy.

[0042] The calculated individual risk scores are added to the existing total risk level score, and the updated total risk level score is the new risk level benchmark value.

[0043] For single-choice factors, the severity multiplied by the probability of the user's selected option is directly used as the risk score contribution for that factor.

[0044] S3.2.3, Handling of multiple selection factors: Optionally, the input string can be split by commas to obtain a string array. The program sequentially extracts each substring from the array and assigns it to a temporary variable for processing. For the current temporary variable, a mapping object containing relevant information is retrieved from the database based on the type parameter. The values ​​for severity and probability (with the keys) are obtained from the mapping object. These two values ​​are then converted to high-precision decimal numbers. The individual score is calculated as severity × probability. The product is rounded to four decimal places. The calculated individual scores are then added to the total variable. The total variable should initially be an accumulator, which continuously accumulates the risk scores of all split items as the loop progresses.

[0045] The risk contribution of multiple-choice factors is the sum of the severity and probability of each selected option. This means that the risk contribution of multiple-choice factors is the sum of the scores of each option, reflecting the principle that "the more risk factors there are, the higher the cumulative risk".

[0046] S3.2.4, Handling of Equipment Service Life Factor: Equipment lifespan factor = ((current year - year of start of use) × severity parameter) / current year.

[0047] The longer the equipment has been in use, the larger the numerator and the higher the risk score; at the same time, normalization is achieved by dividing by the current year.

[0048] S3.2.5, Handling of the next inspection date factor: Next inspection date factor = 1 / number of days; the number of days is equal to the difference between the current date and the next inspection date.

[0049] The closer to the next inspection date (the smaller the number of days), the higher the score, reflecting the characteristic that the risk of inspection delinquency increases as time approaches.

[0050] S3.3, Determination of Risk Level and Control Level: After all risk factors are scored and summed, a comprehensive risk score is obtained. The comprehensive risk score is then used to determine the risk level, and the corresponding risk level and control level are shown in Table 1. Table 1. Comparison of Risk Levels and Control Levels

[0051] The first threshold of 30, the second threshold of 60, and the third threshold of 90 are default configuration values, which can be dynamically adjusted according to industry regulatory requirements (global adjustment can be achieved by modifying static constants).

[0052] S3.4 Automatic upward propagation of risk levels along organizational hierarchies: After updating the risk level of the risk point, conduct organizational-level risk propagation: Check each risk unit bound to the current risk point one by one, and compare the current level of the risk unit with the level of the risk point. If the current level of the risk unit is lower than the level of the risk point, update the level of the risk unit to the level of the risk point. If the risk unit has never been set with any level before, use the current level of the risk point to complete the initialization assignment.

[0053] For the current department to which a risk point belongs, the rule of "the smaller the level value, the higher the risk level" is followed for judgment. If the current department's existing level value is greater than the level value of the risk point, it means that the actual risk level of the current department is lower than that of the risk point, and the department's risk level needs to be updated to the level of the risk point.

[0054] Starting from the current department, trace back to the parent organization and continuously perform hierarchical synchronization operations until the root organization node (parent department ID is "0") is reached and then stops.

[0055] If a parent department has not previously set a risk level, or if the risk level of the current risk point is higher than the existing risk level of the parent department, then the risk level of the parent department will be updated to the current risk point level, and then the process will continue to trace back to the next higher level.

[0056] This propagation mechanism ensures that when any major risk point exists within an enterprise, the risk level of that enterprise and all its superior management units (group headquarters, regional companies, etc.) will be automatically updated to major risk, so that the regional risk level distribution map in the management dashboard always reflects the latest status.

[0057] S3.5 Determining the maximum / minimum value range: To determine the maximum and minimum risk scores for all possible configuration options under a specific risk subcategory: For single-choice factors, take the maximum and minimum values ​​of severity × probability among all options. For multi-choice factors, the maximum value is the sum of severity × probability for all options (all-select scenario), and the minimum value is the minimum severity × probability among all options (only one optimal option selected scenario). Risk type factors are not included in the maximum / minimum calculation (their score is 0 because risk type is a classification of risk nature rather than a quantification of risk magnitude).

[0058] Through the above methods, the system provides a risk score reference range to help assess the relative position of the current score within the possible range.

[0059] S4. Standardized Identification Pattern (Alternative Path): like Figure 2 As shown, when the enterprise identification type is not "1", the system enters the standardized identification mode: directly reads the standard risk level preset in the risk point sub-category template pre-assessed by industry experts; automatically maps the risk level value (1-4) to the corresponding control level, skipping the L·S quantitative scoring process.

[0060] This model is suitable for scenarios with existing clear industry standards, ensuring consistency with external standards for enterprises.

[0061] S5. NLP-based intelligent hazard identification (real-time hazard identification engine): like Figure 3 As shown, when safety inspectors discover a hazard and enter a description of it, the system performs the following intelligent identification: S5.1, Chinese word segmentation: The HanLP natural language processing tool was used to perform Chinese word segmentation on the input hazard description text, dividing the continuous text into independent word units. HanLP uses a CRF (Conditional Random Field) word segmentation model, which can more accurately identify out-of-vocabulary words and technical terms compared to the traditional dictionary-based maximum matching method.

[0062] S5.2, Two-level matching strategy: S5.2.1, Level 1 (Full Text Exact Match): The complete hazard description text (unsegmented) entered by the user is used as the query condition. A LIKE fuzzy search is performed in the "Default Non-compliant Description" field of the inspection item knowledge base to match the candidate results with the most complete semantics.

[0063] S5.2.2, Second Level (Word Segmentation-Level Exact Matching): Each word segmentation result is processed with filtering rules, searching, and merging: words with a length of less than 2 are not included in the matching (excluding punctuation marks and single characters that cannot form valid semantics), each valid word is searched separately in the check item knowledge base with LIKE, and all matching results are merged and added to the global result set.

[0064] S5.3 Result Merging, Deduplication, and Sorting: The LinkedHashSet data structure is used. The HashSet property ensures the uniqueness of the result, and the Linked property ensures the insertion order of the result (the full text matching result takes precedence, and the word segmentation matching result comes later).

[0065] S5.4 Keyword highlighting: For each matching check item in the final result set, keywords in the description text that do not match the word segmentation results are wrapped in HTML red highlight tags, so that inspectors can intuitively see the matching basis of each item on the interface, enhancing the interpretability of the system.

[0066] S5.5, Deduplication Mechanism: By overriding the equals and hashCode methods, and using the combination of check item description and non-compliance description as the unique identifier of the object, we can ensure that the same check items generated by different matching methods (full text and word segmentation) are automatically deduplicated.

[0067] S6. Automatic linkage between inspection results and hazard records: like Figure 4 As shown, when an inspector marks an inspection item as "non-compliant" during the execution of an inspection task, the system performs the following linked operations: S6.1 Automatically create hazard records: The system automatically extracts the legal basis and its description from the inspection items, and then splices them together to fill the hazard basis field of the hazard record.

[0068] S6.2 Automatically identify the source of potential hazards: Automatically set the source of potential hazards according to the task context - set ordinary safety inspection to 3 (safety inspection), and safety liability insurance scenario to 9 (safety liability insurance - safety inspection). The source of potential hazards ID (yhlyId) is associated with the inspection item result ID to achieve data traceability.

[0069] S6.3 Automatically assign safety administrators for confirmation: Query the list of safety administrators of the inspected unit, automatically set the person responsible for confirming the hidden danger, and initially set the hidden danger status to pending confirmation (yhzt=1).

[0070] S6.4, Safety Liability Insurance Scenarios Differentiation: Skip the confirmation process directly and set the hazard status to pending closure.

[0071] S6.5 Automatic Synchronization of Result Changes: When inspectors change the result from "non-compliant" to "compliant", the associated hazard records are automatically deleted.

[0072] S7. Closed-loop hazard investigation and management workflow: like Figure 5 As shown, the hazard record includes a complete status flow: Pending Confirmation → Unrectified → Rectified → Closed. Abnormal statuses include deletion and overdue rectification. It supports six hazard source channels: job-specific inspection, job-specific spot checks, safety inspections, Snapshots - Job-specific, Snapshots - Other, and safety manager registration.

[0073] S8, a production cycle-adaptive troubleshooting and statistics engine: Obtain enterprise production cycle plans and iterate through each configured enterprise; determine whether the current statistical date is within the production cycle, and exclude non-production days from the statistics; for uninvestigated risk points, determine the investigation cycle of control measures, and exclude non-investigated risks from the last day of the cycle or those with investigation records within the cycle; three-level progressive statistical dimensions: job level → risk point level → control measure level (five types of measures); cumulative trend statistics: daily cumulative statistics are used for managing the dashboard trend chart.

[0074] This embodiment constructs a quantitative scoring model based on multidimensional risk factors (severity S × probability L), realizing automatic calculation of risk levels and automatic recommendation of control levels. Differentiated scoring methods are designed for risk factors with different selection types (single choice, multiple choice, fill-in-the-blank date / year), particularly regarding how to incorporate objective time indicators such as equipment service life and inspection date into risk quantification calculations. Intelligent semantic matching of hazard descriptions and standard inspection items based on HanLP natural language processing is implemented, supporting a two-level matching strategy of full-text fuzzy matching and word-segmentation-level precise matching. An automatic upward propagation and update mechanism for risk levels along organizational levels (risk point → risk unit → department → superior unit) is implemented. An automatic association and linkage mechanism between safety inspection results and hazard records is established, realizing automated data flow throughout the entire chain from inspection to hazard generation and rectification closure.

[0075] This embodiment upgrades risk identification from subjective qualitative assessment to data-driven quantitative calculation through a multi-dimensional risk factor L·S (severity × probability) cumulative scoring model. The severity and probability values ​​of each risk factor are pre-set and continuously optimized by industry experts, ensuring consistency and comparability in the assessment results of the same type of risk point across different companies. Compared to the traditional three-tier qualitative scoring of high / medium / low, this embodiment's four-level risk classification offers finer granularity and higher discriminative power.

[0076] This embodiment designs independent calculation formulas for different selection types (single choice, multiple choice, equipment usage year fill-in, inspection date fill-in)—single choice is a single value, multiple choice is a cumulative sum, usage time is normalized by (years used × severity) / current year, and inspection date is accelerated by 1 / the reciprocal of the difference in days—ensuring that the contribution of each data type to the risk score is mathematically reasonable and conforms to actual safety laws.

[0077] This embodiment uses a step-by-step upward propagation algorithm from risk point to risk unit to department to superior unit to achieve automatic synchronization of risk levels across the entire organizational chain. The regional risk level distribution map in the management dashboard always reflects the latest real situation, solving the problems of lag and inaccuracy caused by manual step-by-step summarization in the traditional method.

[0078] This embodiment utilizes HanLP natural language segmentation combined with a two-level matching strategy (full-text matching + segmentation-level matching) to enable the system to understand non-standardized natural language descriptions of potential hazards and automatically match corresponding standard inspection items and regulatory bases. Keyword highlighting in red enhances user experience and the interpretability of the results.

[0079] This embodiment supports both quantitative identification mode (based on L·S calculation) and standardized identification mode (referencing industry standard levels), allowing enterprises to flexibly choose according to their own management granularity needs.

[0080] This embodiment automates the entire data process, from marking inspection results to automatically generating hidden dangers, automatically filling in legal basis, automatically assigning confirmers, and tracking status flow, eliminating the manual transcription step.

[0081] This embodiment accurately reflects the true completion rate of investigations for non-continuous production enterprises through a dual verification mechanism of production cycle planning and investigation cycle.

[0082] Example 2: This embodiment provides a safety production risk identification system based on multi-dimensional risk factor quantification, including: The data acquisition module is configured to determine whether to perform quantitative identification based on the acquired identification type. The judgment module is configured to: if yes, determine the risk type factor, single-choice factor, multiple-choice factor, equipment service life factor, and next inspection date factor under the risk point; otherwise, proceed to standardized identification; among them, the risk type factor, single-choice factor, and multiple-choice factor are determined by quantifying the risk into two dimensions: severity and probability; the equipment service life factor is determined based on the relationship between the current year and the year of commencement of use; the next inspection date factor is determined based on the relationship between the current date and the next inspection date; The comprehensive risk score determination module is configured to determine the comprehensive risk score by accumulating the risk type factor, single-choice factor, multiple-choice factor, equipment service life factor, and next inspection date factor. The risk level determination module is configured to determine the risk level based on a comparison between the comprehensive risk score and a threshold.

[0083] The working method of the system is the same as that of the safety production risk identification method based on multidimensional risk factor quantification in Example 1, and will not be repeated here.

[0084] Example 3: This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the safety production risk identification method based on multidimensional risk factor quantification described in Embodiment 1.

[0085] Example 4: This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. When the processor executes the program, it implements the steps of the safety production risk identification method based on multidimensional risk factor quantification described in Embodiment 1.

[0086] Example 5: This embodiment provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the steps of the safety production risk identification method based on multidimensional risk factor quantification described in Embodiment 1.

[0087] The above description is merely a preferred embodiment of this practice and is not intended to limit the scope of this practice. Various modifications and variations can be made to this practice by those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of this practice should be included within the protection scope of this practice.

Claims

1. A method for identifying safety production risks based on multidimensional risk factor quantification, characterized in that, include: Determine whether to perform quantitative identification based on the acquired identification type; If so, determine the risk type factor, single-choice factor, multiple-choice factor, equipment service life factor, and next inspection date factor under the risk point; otherwise, proceed to standardized identification. Among them, the risk type factor, single-choice factor, and multiple-choice factor are determined by quantifying the risk into two dimensions: severity and probability. The equipment service life factor is determined based on the relationship between the current year and the year of commencement of use. The next inspection date factor is determined based on the relationship between the current date and the next inspection date. The comprehensive risk score is determined by summing the risk type factor, single-choice factor, multiple-choice factor, equipment service life factor, and next inspection date factor. The risk level is determined by comparing the comprehensive risk score with the threshold.

2. The safety production risk identification method based on multidimensional risk factor quantification as described in claim 1, characterized in that, The risk type factor is determined by a combination of initial preset values ​​and user-selected values, with the contribution value of each part based on the product of severity and probability.

3. The safety production risk identification method based on multidimensional risk factor quantification as described in claim 2, characterized in that, The determination of the initial preset value includes: traversing all initialized preset data items; for each item, obtaining the corresponding severity value and probability value; if both exist, the product of the severity value and probability value is added to the total score; the determination of the user manually selected value includes: splitting the user-input string into multiple independent selection values ​​by commas, traversing the multiple independent selection values, for each selection value, obtaining the corresponding severity value and probability value; if both exist, the product of the severity value and probability value is added to the total score.

4. The safety production risk identification method based on multidimensional risk factor quantification as described in claim 1, characterized in that, The determination of the single-selection factor includes: retrieving the corresponding sub-item indicator data from the database according to the specified sub-item type and data value, and storing the results in a key-value pair set; extracting two numerical fields representing severity and probability from the obtained data set; converting the extracted severity and probability into numerical types and performing multiplication to obtain the current single-selection factor score.

5. The safety production risk identification method based on multidimensional risk factor quantification as described in claim 1, characterized in that, The determination of the multiple selection factors includes: splitting the input string by commas to obtain a string array; sequentially extracting each substring from the array and assigning it to a temporary variable for processing; for the current temporary variable, retrieving and returning a mapping object containing relevant information from the database; obtaining the values ​​with the keys of severity and probability from the mapping object; converting these two values ​​to decimal numbers and determining the individual score by multiplying severity by probability; and accumulating the individual scores.

6. The method for identifying safety production risks based on multidimensional risk factor quantification as described in claim 1, characterized in that, The equipment service life factor = ((current year - year of start of use) × severity parameter) / current year; the next inspection date factor = 1 / number of days; The number of days between the current date and the next inspection date is equal to the difference between the current date and the next inspection date.

7. The safety production risk identification method based on multidimensional risk factor quantification as described in claim 1, characterized in that, The determination of the risk level includes: if the comprehensive risk score is less than a first threshold, the risk level is low risk; if the comprehensive risk score is greater than or equal to the first threshold and less than the second threshold, the risk level is moderate risk; if the comprehensive risk score is greater than or equal to the second threshold and less than the third threshold, the risk level is relatively high risk; if the comprehensive risk score is greater than or equal to the third threshold, the risk level is significant risk; the first threshold is less than the second threshold, and the second threshold is less than the third threshold.

8. The method for identifying safety production risks based on multidimensional risk factor quantification as described in claim 1, characterized in that, The standardized identification includes: directly reading the pre-set standard risk level in the pre-assessed risk point template; and automatically mapping the risk level value corresponding to the standard risk level to the corresponding control level.

9. The method for identifying safety production risks based on multidimensional risk factor quantification as described in claim 1, characterized in that, The method further includes: after determining the risk level, checking all risk units bound to the current risk point one by one, and comparing the existing level of the risk unit with the level of the risk point; if the current level of the risk unit is lower than the level of the risk point, directly updating the level of the risk unit to the level of the risk point; if the risk unit has never been set with any level before, directly using the current level of the risk point to complete the initialization assignment.

10. A safety production risk identification system based on multi-dimensional risk factor quantification, characterized in that, include: The data acquisition module is configured to determine whether to perform quantitative identification based on the acquired identification type. The judgment module is configured to: if yes, determine the risk type factor, single-choice factor, multiple-choice factor, equipment service life factor, and next inspection date factor under the risk point; otherwise, proceed to standardized identification; among them, the risk type factor, single-choice factor, and multiple-choice factor are determined by quantifying the risk into two dimensions: severity and probability; the equipment service life factor is determined based on the relationship between the current year and the year of commencement of use; the next inspection date factor is determined based on the relationship between the current date and the next inspection date; The comprehensive risk score determination module is configured to determine the comprehensive risk score by accumulating the risk type factor, single-choice factor, multiple-choice factor, equipment service life factor, and next inspection date factor. The risk level determination module is configured to determine the risk level based on a comparison between the comprehensive risk score and a threshold.

Images