A railway information engineering safety risk prevention system and method

Abstract

This invention belongs to the field of industrial safety and risk management technology, specifically disclosing a safety risk prevention system and method for railway information engineering. The system includes a risk analysis module and a risk assessment and quantification module. The risk analysis module integrates the Work Breakdown Structure (WBS-RBS) method and the STPA model to identify and analyze risks in the entire workflow and system functions of railway information engineering, and constructs a risk classification system based on the analysis results. The risk assessment and quantification module defines each item in the risk classification system as a risk event, and uses the risk matrix method, Apriori association rule mining, and complex network theory to objectively calculate its probability, consequence value, and network weight, ultimately achieving a quantitative assessment of the comprehensive risk level. This invention solves the problems of subjective assessment and lack of systematic and forward-looking preventive measures in the field of railway information engineering due to the scarcity of accident cases, realizing precise prevention from systemic risk deduction to data-driven quantitative assessment.

Description

Technical Field

[0001] This invention belongs to the field of industrial safety and risk management technology, specifically relating to a safety risk prevention system and method for railway information engineering. Background Technology

[0002] Railway information technology projects are a crucial component of the nation's critical information infrastructure, and their security directly impacts railway transportation. Currently, risk management in this field faces a core dilemma: on the one hand, the systems are highly complex and tightly coupled, with numerous potential risk points; on the other hand, the scarcity of available historical accident cases limits the effectiveness of traditional risk prevention methods. Traditional security methods primarily suffer from the following limitations: 1. Limitations of the case study: It relies heavily on historical accident statistics, which makes it difficult to fully identify risks under sparse data conditions, and the analysis conclusions are easily affected by small sample bias.

[0003] 2. Risk assessment is highly subjective: In the risk assessment process, especially in determining risk weights, it often relies on subjective judgments such as expert scoring and the analytic hierarchy process, which lacks quantitative objectivity and makes it difficult to handle the complex relationships between a large number of risk events.

[0004] 3. Separation of analysis and assessment: Risk identification and risk quantification assessment are often treated as independent steps, lacking a unified data framework and model linkage, resulting in a lack of systematic risk management strategies.

[0005] Therefore, there is an urgent need for a new type of risk prevention technology that does not rely on massive amounts of accident data, can integrate systemic risk simulation with objective data assessment, and can self-iterate and improve. Summary of the Invention

[0006] To address the aforementioned technical challenges, this invention proposes a model-driven and data fusion-based railway information engineering safety risk prevention system. This system employs a model that combines system analysis of risks with multi-source data fusion for quantitative assessment. It effectively overcomes the scarcity of data in the early stages and is a novel risk prevention technology capable of self-iterative improvement. It effectively solves problems in the railway information field, such as incomplete risk identification, subjective assessment, and a lack of systematic and forward-looking preventive measures due to the scarcity of accident cases. It achieves precise prevention from systemic risk deduction to data-driven quantitative assessment.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: A railway information engineering safety risk prevention system, which includes a logically linked and data-interconnected risk analysis module and a risk assessment and quantification module; The risk analysis module integrates the WBS-RBS work decomposition method and the STPA model to identify and analyze risks in the entire workflow and system functions of railway information engineering, and constructs a risk classification system based on the analysis results; specifically, it includes: a data acquisition and preprocessing unit, a data analysis unit, and a risk classification system construction unit; The data acquisition and preprocessing unit is used to collect and filter risk data from multiple channels. Data sources include historical accident cases, near misses and failure reports, risk items found in security audits and assessments, system operation and maintenance logs and reports, and typical risk scenarios constructed by domain experts. The data analysis unit integrates the Work Breakdown Structure (WBS-RBS) and the STPA model to identify and analyze risks in target projects or systems. The WBS-RBS is used to deconstruct the entire workflow of railway information engineering projects and establish a corresponding risk decomposition structure to identify potential risk points and risk factors in each process and stage. The STPA model, based on the system components and control functions identified by WBS-RBS, analyzes possible unsafe control behaviors and traces the underlying causes of these behaviors, including software defects, incorrect requirements, human-computer interaction problems, environmental interference, and organizational management failures. The risk classification system construction unit, based on the output results of the data analysis unit, automatically merges, organizes, and structures risk factors to form and continuously maintain a standardized risk classification system for railway information engineering that covers multiple dimensions including technology, management, personnel, environment, and interaction. The risk assessment and quantification module is used to define each item in the risk classification system as a risk event, and to objectively calculate its probability, consequence value and network weight using the risk matrix method, Apriori association rule mining and complex network theory, so as to achieve a quantitative assessment of the comprehensive risk level. Specifically, it includes a risk event definition unit, a probability and consequence preliminary assessment unit, an objective weight analysis unit and a comprehensive risk calculation unit. The risk event definition unit is used to define each risk factor in the established standardized risk classification system as an assessable risk event; The probability and consequence preliminary assessment unit, based on the multi-source data provided by the risk data collection and preprocessing unit, uses the risk matrix method to calculate the occurrence frequency or probability estimate of each risk event, as well as the severity level of the possible consequences. The objective weight analysis unit uses the Apriori algorithm to mine association rules for the co-occurrence relationship of risk events in multi-source risk data. Based on the mined strong rules, a directed weighted complex network of risk events is constructed, and the objective weight of each risk event is determined by normalizing the node degree centrality index of the nodes in the network. The comprehensive risk calculation unit is used to integrate the probability value, severity value and objective weight of each risk event to construct a weighted risk assessment model, and calculate the risk value of a single risk event and the comprehensive risk level of the target project or system.

[0008] Preferably, the fusion analysis of WBS-RBS work decomposition method and STPA model in the data analysis unit is carried out in an iterative manner. First, WBS-RBS is used to identify high-risk stages and processes, and then in-depth STPA analysis is carried out on these key objects. New risk points discovered by STPA analysis can be fed back to the WBS-RBS matrix for supplementation.

[0009] Preferably, during the system initialization phase, the risk classification system construction unit pre-sets a basic framework based on a general risk database or expert knowledge; and during operation, it continuously integrates new results from the WBS-RBS work decomposition method and STPA model analysis, as well as newly collected risk data, to automatically classify, deduplicate, and adjust the structure of the classification system, thereby achieving gradual improvement of the classification system.

[0010] Preferably, when constructing a complex network, the objective weight analysis unit uses the preceding risk event of the association rule as the source node, the subsequent risk event as the target node, and the rule confidence as the edge weight; the total node degree of a node is the sum of the weights of all incoming and outgoing edges of that node.

[0011] In addition, the system also includes a risk management strategy generation module, which automatically generates a risk management strategy report containing key prevention and control points, priority governance items and specific action recommendations based on the high-risk stages and processes output by the risk analysis function module and the risk event ranking and comprehensive risk level output by the risk assessment and quantification function module.

[0012] Accordingly, this invention also proposes a method for preventing safety risks in railway information engineering, specifically including the following steps: S1: Multi-source risk data collection and construction: Systematically collect heterogeneous data related to railway information engineering, standardize and classify it for storage, and form an initial risk database; the heterogeneous data includes: historical accident investigation reports and fault records; near miss reports and system anomaly alarm information; risk items and vulnerability lists discovered by security level protection assessments, penetration tests and code audits; full operation logs and performance index data of information systems and network equipment; and a library of typical risk scenarios and cases built by domain experts based on their experience. S2: Systematic risk analysis based on WBS-RBS work breakdown method and STPA model: For the target railway information engineering or system, perform WBS-RBS work breakdown method analysis to identify risk points in the entire work process; and perform STPA analysis on key stages and processes to trace the deep-seated causes of unsafe control behaviors; S3: Construct and update the risk classification system: Integrate the analysis results and common risk factors from step S2 to form a structured, multi-dimensional standardized risk classification system for railway information engineering, and link this system with the risk database; S4: Risk Event Definition and Data Association: The system defines each risk in the standardized risk classification system for railway information engineering as a measurable risk event; and automatically associates and matches these risk events with all multi-source records in the risk database; S5: Quantitative assessment of the likelihood and severity of consequences; S6: Objective weighting analysis based on data mining: A data-driven objective weighting method is adopted. The system calls the Apriori association rule algorithm to mine the co-occurrence relationship of all risk events in the risk database. Based on all the mined rules, a directed weighted complex network is constructed. S7: Comprehensive Risk Value Calculation and Ranking: A weighted composite model is adopted, i.e., comprehensive risk value = probability quantification value × consequence quantification value × objective weight; the final quantified risk value of each risk event is calculated; then, all risk events are sorted in descending order according to their risk values, and the overall comprehensive risk index of the target project or system can be calculated by summarizing; a risk priority list is obtained, indicating which risks should be prioritized for prevention and control, and management decisions are proposed in order of priority. S8: Generate and implement risk management strategies: Based on the key risks output by the risk analysis module and the quantitative risk ranking list output by the risk assessment and quantification module, automatically generate a risk management strategy report; this report not only identifies high-risk items that need to be prioritized, but also proposes targeted and tiered prevention and control measures based on causal analysis. S9: Data Feedback and System Iteration Optimization: The effectiveness of control measures and the daily operation of the system continuously generate new status data, event records and feedback information; this information is automatically collected and incorporated into the risk database; the system automatically re-executes all analysis and evaluation processes from S2 to S7 according to a preset cycle or triggered by a major change event.

[0013] Furthermore, in step S1, for scarce major accident cases, the effective data sample is expanded by incorporating a large number of near misses, high-frequency failure modes, and risk scenarios constructed by experts, to ensure the breadth of the analysis basis.

[0014] Furthermore, in step S2, the WBS-RBS working decomposition analysis and the STPA model analysis form a closed loop. The WBS-RBS working decomposition method provides the analytical scope and focus for the STPA model, while the in-depth findings of the STPA model, in turn, supplement and refine the risk matrix of the WBS-RBS, specifically including: S2-1: Risk analysis using the WBS-RBS work breakdown structure: The system guides users or automatically imports the project's work breakdown structure (WBS), decomposing it to the work package level, which includes "requirement specification confirmation" and "trackside equipment modification"; at the same time, it links it to the standardized risk breakdown structure (RBS), covering technical, management, organizational, and external environment dimensions; by constructing a WBS-RBS matrix, the system systematically identifies potential risk points in each work process. S2-2: In-depth risk analysis using the STPA model: For high-risk components or functions identified by the WBS-RBS work decomposition method, the system initiates the STPA model; first, system-level safety constraints are defined; second, control loop diagrams are drawn; then, the four types of unsafe behaviors that may occur in control commands are systematically examined; finally, the deep-seated causes of the unsafe behavior in terms of technology, management, and organization are traced.

[0015] Furthermore, step S5 specifically includes: S5-1: Probability Calculation: For risk events supported by sufficient historical data, statistical calculations are performed directly based on the frequency of occurrence of related events in the database; for data-sparse or novel risk events, a comprehensive estimation is performed by fusing the expert Delphi method, the STPA scenario occurrence conditional probability, and reference values ​​from similar systems. S5-2: Quantification of the severity of consequences: Based on national railway industry standards and enterprise internal control guidelines, the possible consequences of risk events are mapped to quantifiable levels and assigned corresponding quantitative value ranges. S5-3: Preliminary Risk Rating: Combining the calculated probability level and the severity level of the consequences, a preliminary risk rating is generated for each risk event using a pre-set risk matrix model.

[0016] Furthermore, the system iterative optimization in step S9 includes: automatically discovering new risks based on new data, adjusting complex network structures, recalculating risk weights, and optimizing the structure of the risk classification system accordingly.

[0017] The beneficial effects of this invention are as follows: 1. This invention uses system theory analysis methods that do not rely on historical data, such as WBS-RBS work breakdown structure and STPA model, as the main body for risk identification. Combined with multi-source data such as near misses and audit reports, it constructs a stable and reliable foundation for risk identification and classification, solving the problem of insufficient or difficult-to-obtain historical incident data.

[0018] 2. In the risk analysis phase of this invention, the WBS-RBS work breakdown structure ensures the comprehensiveness of risk identification, covering the entire workflow, while the STPA model enables in-depth risk tracing, directly reaching the root causes of functional safety and interactive failures. The dynamic risk classification system constructed by integrating these two approaches allows for the structured accumulation and sharing of risk knowledge.

[0019] 3. In the risk assessment and quantification stage, this invention utilizes the Apriori algorithm and complex network theory to automatically mine risk correlations from multi-source data and calculate objective weights, which greatly reduces the bias of subjective judgment and makes risk ranking more realistically reflect the intrinsic connections and influence between risks.

[0020] 4. Starting from theoretical deduction and limited data, this invention continuously optimizes itself through continuous operation and data accumulation, ultimately forming a model with rich and accurate data, thus realizing dynamic prevention of risk management capabilities. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the core architecture of the system of the present invention; Figure 2 This is a flowchart of the risk analysis function module of the present invention; Figure 3 This is a flowchart of the risk assessment and quantification functional module of the present invention; Figure 4 This invention generates a risk management strategy report pattern diagram and an intelligent risk prevention and control closed-loop system diagram. Detailed Implementation

[0022] The principles and features of the present invention are described below with reference to the accompanying drawings. This description is for illustrative purposes only and is not intended to limit the scope of the invention.

[0023] This invention proposes a railway information engineering accident risk prevention system based on multi-model fusion. The core architecture of the system consists of a risk analysis module, a risk assessment and quantification module, and a risk control strategy generation module suitable for railway information engineering. It aims to overcome the problems of poor preventiveness, strong subjectivity, and inaccurate measures of traditional methods in the field of railway information engineering safety due to the scarcity of accident cases and complex system coupling.

[0024] Specifically, such as Figure 1 As shown, the risk analysis module integrates the WBS-RBS work breakdown structure and the STPA model to identify and analyze risks across the entire workflow and system functions of railway information engineering, and constructs a risk classification system based on the analysis results. Specifically, it includes: a data acquisition and preprocessing unit, a data analysis unit, and a risk classification system construction unit.

[0025] The data acquisition and preprocessing unit is used to collect and filter risk data from multiple channels. Data sources include historical accident cases, near misses and failure reports, risk items found in security audits and assessments, system operation and maintenance logs and reports, and typical risk scenarios constructed by domain experts. The data analysis unit integrates the Work Breakdown Structure (WBS-RBS) and the STPA model for risk identification and analysis of target projects or systems. The WBS-RBS method is used to deconstruct the entire workflow of railway information engineering projects and establish a corresponding risk decomposition structure to identify potential risk points and risk factors in each process and stage. The STPA model, based on the system components and control functions identified by WBS-RBS, analyzes possible unsafe control behaviors and traces the underlying causes of these behaviors, including software defects, incorrect requirements, human-computer interaction problems, environmental interference, and organizational management failures.

[0026] The integration analysis of WBS-RBS work decomposition method and STPA model is carried out in an iterative manner. First, WBS-RBS is used to identify high-risk stages and processes. Then, in-depth STPA analysis is conducted on these key objects. New risk points discovered by STPA analysis can be fed back into the WBS-RBS matrix for supplementation.

[0027] The risk classification system construction unit, based on the output results of the data analysis unit, automatically merges, organizes, and structures risk factors to form and continuously maintain a standardized risk classification system for railway information engineering that covers multiple dimensions, including technology, management, personnel, environment, and interaction.

[0028] In the system initialization phase, the risk classification system construction unit pre-sets a basic framework based on a general risk database or expert knowledge. During operation, it continuously integrates new results from the WBS-RBS work decomposition method and STPA model analysis, as well as newly collected risk data, to automatically classify, deduplicate, and adjust the structure of the classification system, thereby achieving gradual improvement of the classification system.

[0029] The risk assessment and quantification module defines each item in the risk classification system as a risk event. Using the risk matrix method, Apriori association rule mining, and complex network theory, it objectively calculates the probability, consequence value, and network weight of each event, ultimately achieving a quantitative assessment of the overall risk level. Specifically, it includes: a risk event definition unit, a preliminary probability and consequence assessment unit, an objective weight analysis unit, and a comprehensive risk calculation unit.

[0030] The risk event definition unit is used to define each risk factor in the established standardized risk classification system as an assessable risk event.

[0031] The probability and consequences preliminary assessment unit, based on multi-source data provided by the risk data collection and preprocessing unit, uses the risk matrix method to calculate the occurrence frequency or probability estimate of each risk event, as well as the severity level of the possible consequences.

[0032] The objective weight analysis unit uses the Apriori algorithm to mine association rules for the co-occurrence relationships of risk events in multi-source risk data. Based on the mined strong rules, a directed weighted complex network of risk events is constructed, and the objective weight of each risk event is determined by normalizing the node degree centrality index of the nodes in the network.

[0033] In constructing complex networks, the preceding risk event of an association rule is taken as the source node, the subsequent risk event as the target node, and the rule confidence is the weight of the edge. The total degree of a node is the sum of the weights of all incoming and outgoing edges of that node.

[0034] The comprehensive risk calculation unit is used to integrate the probability value, severity value, and objective weight of each risk event to construct a weighted risk assessment model, and calculate the risk value of a single risk event and the comprehensive risk level of the target project or system.

[0035] The risk management strategy generation module is used to automatically generate a risk management strategy report containing key prevention and control points, priority governance items, and specific action recommendations based on the high-risk stages and processes output by the risk analysis function module and the risk event ranking and comprehensive risk level output by the risk assessment and quantification function module.

[0036] like Figure 2 As shown, based on the above system, this invention also proposes a method for preventing accident risks in railway information engineering based on multi-model fusion, specifically including the following steps: S1: Multi-source Risk Data Collection and Preprocessing. Five types of heterogeneous data are collected from railway operating units' reporting systems, production and maintenance monitoring platforms, third-party security audit tools, and expert knowledge entry points: ① Historical accident investigation reports and fault records; ② Near miss reports and system anomaly alarm information; ③ Risk items and vulnerability lists discovered through security level protection assessments, penetration testing, and code audits; ④ Full operational logs and performance index data of information systems and network equipment; ⑤ A library of typical risk scenarios and cases built by domain experts based on experience. After structured extraction of key information such as time, related systems, fault phenomena, and preliminary attribution, this data is stored in a unified risk database. For scarce major accident cases, the effective data sample is expanded by incorporating a large number of near misses, high-frequency fault patterns, and expert-constructed risk scenarios, ensuring the breadth of the analytical foundation. This step solves the problem of scarce major accident samples in the railway safety field, providing multi-source data for subsequent model-driven analysis.

[0037] S2: Systematic risk analysis based on WBS-RBS work decomposition method and STPA model. The WBS-RBS work decomposition method and STPA model analysis form a closed loop. The WBS-RBS work decomposition method provides the analytical scope and focus for the STPA model, while the in-depth findings of the STPA model, in turn, supplement and refine the risk matrix of the WBS-RBS. Specifically, this includes: S2-1: Risk analysis using the WBS-RBS work breakdown structure. The system guides users or automatically imports the project's WBS (Work Breakdown Structure), breaking it down to work package levels such as "Requirements Specification Confirmation" and "Railwayside Equipment Modification." Simultaneously, a standardized RBS (Risk Breakdown Structure) is linked, covering dimensions such as technology, management, organization, and external environment. By constructing a WBS-RBS matrix, potential risk points in each work stage are systematically identified. This step comprehensively identifies risks throughout the entire project workflow.

[0038] S2-2: In-depth risk analysis using the STPA model. For high-risk components or functions identified by the WBS-RBS work decomposition method, the system initiates the STPA model. First, system-level safety constraints are defined; second, control loop diagrams are drawn; next, four types of unsafe behaviors that may occur in control commands are systematically examined; finally, the underlying causes of these unsafe behaviors in terms of technology, management, and organization are traced. The purpose of STPA model analysis is to elevate risk analysis from the static level of component failure to the dynamic level of system interaction and loss of control, revealing the deep-seated causes of complex system failures.

[0039] S3: Construction and Update of Dynamic Risk Classification System. The above analysis generated a large number of fragmented but semantically related risk descriptions. The system will automatically merge, refine, and structure risk factors to form a standardized risk classification system for railway information engineering, covering multiple dimensions such as technology, management, personnel, environment, and interaction. This system will then be linked to the risk database. This classification system is dynamic, continuously expanding and refining with the addition of new analysis results, providing a unified and shareable structured risk framework for the entire system.

[0040] like Figure 3 As shown, after risks are systematically identified and structured, the risk assessment and quantification module is activated, responsible for objectively measuring and prioritizing risks, specifically including the following steps: S4: Risk Event Definition and Data Association. The system defines each risk in the standardized risk classification system for railway information engineering as a measurable risk event. Subsequently, it automatically associates and matches these risk events with all multi-source records in the risk database.

[0041] S5: Quantitative assessment of the likelihood and severity of consequences.

[0042] S5-1: Probability Calculation. For risk events supported by sufficient historical data, the system directly performs statistical calculations based on the frequency of occurrence of its related events in the database. For sparse data or novel risk events, a comprehensive estimation is performed by fusing the expert Delphi method, the conditional probability of STPA scenario occurrence, and reference values ​​from similar systems.

[0043] S5-2: Quantification of the severity of consequences. Based on national railway industry standards and enterprise internal control guidelines, the potential consequences of risk events are mapped to quantifiable levels and assigned corresponding quantitative value ranges.

[0044] S5-3: Preliminary Risk Rating. Combining the calculated likelihood level and the severity level of the consequences, a preliminary risk rating is generated for each risk event using a pre-defined risk matrix model.

[0045] S6: Objective Weight Analysis Based on Data Mining. This invention employs a data-driven objective weighting method. The system calls the Apriori association rule algorithm to mine the co-occurrence relationships of all risk events in the risk database. Based on all the mined rules, a directed weighted complex network is constructed: nodes are risk events, directed edges represent association rules, the direction is from the antecedent to the consequent, and the weight of the edge is the confidence level of the rule. In this network, the node degree centrality of each node is calculated, which is the sum of the weights of all incoming and outgoing edges. This objectively measures the influence of the risk event in the overall risk network. The node degree of all nodes is normalized, and the result is the objective weight of each risk event. The core advantage of this method is that the weights are entirely derived from the real association patterns between risk events in historical and operational data, rather than from human scoring, thus improving the objectivity of the assessment.

[0046] S7: Comprehensive Risk Value Calculation and Ranking. The system employs a weighted composite model, where the comprehensive risk value = quantified probability value × quantified consequence value × objective weight, calculating the final quantified risk value for each risk event. Subsequently, all risk events are sorted in descending order of their risk values, and the overall comprehensive risk index of the target project or system can be calculated by aggregating the results. This step generates a risk priority list, indicating which risks should be prioritized for control, and proposes management decisions in order of priority.

[0047] like Figure 4 As shown, this invention ultimately forms a self-evolving and continuously optimized intelligent risk prevention and control closed-loop system, the specific process of which is as follows: S8: Generate and Implement Risk Management Strategies. Based on the key risks output by the risk analysis module and the quantitative risk ranking list output by the risk assessment and quantification module, the system automatically generates a risk management strategy report. This report not only identifies high-risk items that require priority handling but also proposes targeted, tiered prevention and control measures based on causal analysis. This report can be directly sent to relevant safety management personnel to guide on-site implementation.

[0048] S9: Data Feedback and System Iteration Optimization. The effectiveness of control measures and the daily operation of the system continuously generate new status data, event records, and feedback information. This information is automatically collected and imported into the risk database. The system automatically re-executes the entire analysis and evaluation process from S2 to S7 according to a preset cycle or triggered by a major change event. In the new iteration: Richer data enables more accurate association rule mining in step S6, more complete complex network structures, and more reliable weight calculations. New project or system change information triggers updates to the WBS-RBS work breakdown structure and STPA model analysis in step S2. Simultaneously, the dynamic risk classification system in step S3 integrates new risk descriptions and adaptively adjusts them. Finally, the risk assessment results in step S7 and the control strategies in step S8 are dynamically updated.

[0049] This invention employs a continuous closed loop of data summarization → model analysis → quantitative evaluation → precise decision-making → effect feedback → model updating. This allows the system to gradually become an increasingly accurate and intelligent early warning and prevention platform, even under conditions of limited data in the initial stages of construction. It fundamentally solves the deep-seated problems of passive response, subjective judgment, and lack of continuous improvement mechanisms in the safety risk management of railway information engineering. Starting from theoretical deduction and limited data, through continuous operation and data accumulation, it constantly self-optimizes, ultimately forming a model with rich and accurate data, achieving dynamic prevention capabilities for risk management.

Claims

1. A railway information engineering safety risk prevention system, characterized in that: The system includes a logically linked and data-interoperable risk analysis module and a risk assessment and quantification module; The risk analysis module integrates the WBS-RBS work decomposition method and the STPA model to identify and analyze risks in the entire workflow and system functions of railway information engineering, and constructs a risk classification system based on the analysis results; specifically, it includes: a data acquisition and preprocessing unit, a data analysis unit, and a risk classification system construction unit; The data acquisition and preprocessing unit is used to collect and filter risk data from multiple channels. Data sources include historical accident cases, near misses and failure reports, risk items found in security audits and assessments, system operation and maintenance logs and reports, and typical risk scenarios constructed by domain experts. The data analysis unit integrates the Work Breakdown Structure (WBS-RBS) and the STPA model to identify and analyze risks in target projects or systems. The WBS-RBS is used to deconstruct the entire workflow of railway information engineering projects and establish a corresponding risk decomposition structure to identify potential risk points and risk factors in each process and stage. The STPA model, based on the system components and control functions identified by WBS-RBS, analyzes possible unsafe control behaviors and traces the underlying causes of these behaviors, including software defects, incorrect requirements, human-computer interaction problems, environmental interference, and organizational management failures. The risk classification system construction unit, based on the output results of the data analysis unit, automatically merges, organizes, and structures risk factors to form and continuously maintain a standardized risk classification system for railway information engineering that covers multiple dimensions including technology, management, personnel, environment, and interaction. The risk assessment and quantification module is used to define each item in the risk classification system as a risk event, and to objectively calculate its probability, consequence value and network weight using the risk matrix method, Apriori association rule mining and complex network theory, so as to achieve a quantitative assessment of the comprehensive risk level. Specifically, it includes a risk event definition unit, a probability and consequence preliminary assessment unit, an objective weight analysis unit and a comprehensive risk calculation unit. The risk event definition unit is used to define each risk factor in the established standardized risk classification system as an assessable risk event; The probability and consequence preliminary assessment unit, based on the multi-source data provided by the risk data collection and preprocessing unit, uses the risk matrix method to calculate the occurrence frequency or probability estimate of each risk event, as well as the severity level of the possible consequences. The objective weight analysis unit uses the Apriori algorithm to mine association rules for the co-occurrence relationship of risk events in multi-source risk data. Based on the mined strong rules, a directed weighted complex network of risk events is constructed, and the objective weight of each risk event is determined by normalizing the node degree centrality index of the nodes in the network. The comprehensive risk calculation unit is used to integrate the probability value, severity value and objective weight of each risk event to construct a weighted risk assessment model, and calculate the risk value of a single risk event and the comprehensive risk level of the target project or system.

2. The railway information engineering safety risk prevention system according to claim 1, characterized in that: The data analysis unit integrates the WBS-RBS work breakdown method with the STPA model in an iterative manner. First, WBS-RBS is used to identify high-risk stages and processes. Then, in-depth STPA analysis is conducted on these key objects. New risk points discovered by STPA analysis can be fed back to the WBS-RBS matrix for supplementation.

3. The railway information engineering safety risk prevention system according to claim 1, characterized in that: During the system initialization phase, the risk classification system construction unit pre-sets a basic framework based on a general risk database or expert knowledge. During operation, it continuously integrates new results from the WBS-RBS work decomposition method and STPA model analysis, as well as newly collected risk data, to automatically classify, deduplicate, and adjust the structure of the classification system, thereby achieving gradual improvement of the classification system.

4. The railway information engineering safety risk prevention system according to claim 1, characterized in that: When constructing a complex network, the objective weight analysis unit uses the preceding risk event of the association rule as the source node, the subsequent risk event as the target node, and the rule confidence as the edge weight; the total node degree of a node is the sum of the weights of all incoming and outgoing edges of that node.

5. The railway information engineering safety risk prevention system according to any one of claims 1 to 4, characterized in that: It also includes a risk management strategy generation module, which automatically generates a risk management strategy report containing key prevention and control points, priority governance items and specific action recommendations based on the high-risk stages and processes output by the risk analysis function module and the risk event ranking and comprehensive risk level output by the risk assessment and quantification function module.

6. A method for preventing safety risks in railway information engineering, characterized in that: This method is based on the railway information engineering safety risk prevention system described in claim 5; the method specifically includes the following steps: S1: Multi-source risk data collection and construction: Systematically collect heterogeneous data related to railway information engineering, standardize and classify it for storage, and form an initial risk database; the heterogeneous data includes: historical accident investigation reports and fault records; near miss reports and system anomaly alarm information; risk items and vulnerability lists discovered by security level protection assessments, penetration tests and code audits; full operation logs and performance index data of information systems and network equipment; and a library of typical risk scenarios and cases built by domain experts based on their experience. S2: Systematic risk analysis based on WBS-RBS work breakdown method and STPA model: For the target railway information engineering or system, perform WBS-RBS work breakdown method analysis to identify risk points in the entire work process; and perform STPA analysis on key stages and processes to trace the deep-seated causes of unsafe control behaviors; S3: Construct and update the risk classification system: Integrate the analysis results and common risk factors from step S2 to form a structured, multi-dimensional standardized risk classification system for railway information engineering, and link this system with the risk database; S4: Risk Event Definition and Data Association: The system defines each risk in the standardized risk classification system for railway information engineering as a measurable risk event; and automatically associates and matches these risk events with all multi-source records in the risk database; S5: Quantitative assessment of the likelihood and severity of consequences; S6: Objective weighting analysis based on data mining: A data-driven objective weighting method is adopted. The system calls the Apriori association rule algorithm to mine the co-occurrence relationship of all risk events in the risk database. Based on all the mined rules, a directed weighted complex network is constructed. S7: Comprehensive Risk Value Calculation and Ranking: A weighted composite model is adopted, i.e., comprehensive risk value = probability quantification value × consequence quantification value × objective weight; the final quantified risk value of each risk event is calculated; then, all risk events are sorted in descending order according to their risk values, and the overall comprehensive risk index of the target project or system can be calculated by summarizing; a risk priority list is obtained, indicating which risks should be prioritized for prevention and control, and management decisions are proposed in order of priority. S8: Generate and implement risk management strategies: Based on the key risks output by the risk analysis module and the quantitative risk ranking list output by the risk assessment and quantification module, automatically generate a risk management strategy report; this report not only identifies high-risk items that need to be prioritized, but also proposes targeted and tiered prevention and control measures based on causal analysis. S9: Data Feedback and System Iteration Optimization: The effectiveness of control measures and the daily operation of the system continuously generate new status data, event records and feedback information; this information is automatically collected and incorporated into the risk database; the system automatically re-executes all analysis and evaluation processes from S2 to S7 according to a preset cycle or triggered by a major change event.

7. The method for preventing safety risks in railway information engineering according to claim 6, characterized in that: In step S1, for scarce major accident cases, the effective data sample is expanded by incorporating a large number of near misses, high-frequency failure modes, and risk scenarios constructed by experts to ensure the breadth of the analysis basis.

8. The method for preventing safety risks in railway information engineering according to claim 6, characterized in that: In step S2, the WBS-RBS working decomposition analysis and the STPA model analysis form a closed loop. The WBS-RBS working decomposition method provides the analytical scope and focus for the STPA model, while the in-depth findings of the STPA model, in turn, supplement and refine the risk matrix of the WBS-RBS, specifically including: S2-1: Risk analysis using the WBS-RBS work breakdown structure: The system guides users or automatically imports the project's work breakdown structure (WBS), decomposing it to the work package level, which includes "requirement specification confirmation" and "trackside equipment modification"; at the same time, it links it to the standardized risk breakdown structure (RBS), covering technical, management, organizational, and external environment dimensions; by constructing a WBS-RBS matrix, the system systematically identifies potential risk points in each work process. S2-2: In-depth risk analysis using the STPA model: For high-risk components or functions identified by the WBS-RBS work decomposition method, the system initiates the STPA model; first, system-level safety constraints are defined; second, control loop diagrams are drawn; then, the four types of unsafe behaviors that may occur in control commands are systematically examined; finally, the deep-seated causes of the unsafe behavior in terms of technology, management, and organization are traced.

9. The method for preventing safety risks in railway information engineering according to claim 6, characterized in that: Step S5 specifically includes: S5-1: Probability Calculation: For risk events supported by sufficient historical data, statistical calculations are performed directly based on the frequency of occurrence of related events in the database; for data-sparse or novel risk events, a comprehensive estimation is performed by fusing the expert Delphi method, the STPA scenario occurrence conditional probability, and reference values ​​from similar systems. S5-2: Quantification of the severity of consequences: Based on national railway industry standards and enterprise internal control guidelines, the possible consequences of risk events are mapped to quantifiable levels and assigned corresponding quantitative value ranges. S5-3: Preliminary Risk Rating: Combining the calculated probability level and the severity level of the consequences, a preliminary risk rating is generated for each risk event using a pre-set risk matrix model.

10. The method for preventing safety risks in railway information engineering according to claim 6, characterized in that: The system iterative optimization in step S9 includes: automatically discovering new risks based on new data, adjusting complex network structures, recalculating risk weights, and optimizing the structure of the risk classification system accordingly.

Images