Fault tree automatic construction method based on multi-signal flow graph

Abstract

The invention discloses a fault tree automatic construction method based on a multi-signal flow graph. The fault tree automatic construction method aims at solving the problems that an existing method is low in construction efficiency, multiple in logic blind area, difficult in dynamic optimization and the like in a complex system. The fault tree automatic construction method comprises three steps: multi-signal flow graph entity extraction based on ontology constraint, entity structure normalization based on flow graph connectivity and fault tree data automatic layout based on depth first. According to the fault tree automatic construction method, a dynamic corresponding relation with fault tree nodes is established by utilizing nodes and attribute information in a multi-signal flow graph, and accurate mapping from an entity to the fault tree nodes is realized; the integrity of a logic link of the fault tree is ensured by skipping unrelated nodes, complementing an association relationship and automatically inserting'or 'nodes; and through depth-first branch statistics and breadth-first automatic layout, global information is obtained, top-down layered layout and inter-node connection are realized, and finally, automatic drawing of the fault tree is completed. According to the method, cross-component and cross-level complex system fault analysis requirements are supported, the efficiency and accuracy of fault tree construction are remarkably improved, and a new technical means is provided for modern reliability management.

Description

Technical Field

[0001] The invention relates to an automatic construction method of a fault tree, in particular to an automatic construction method of a fault tree based on a multi-signal flow graph. Background Art

[0002] Fault tree analysis is one of the core methods in the safety analysis of complex systems and is widely used in high-reliability fields such as aerospace, nuclear power, and transportation. By decomposing the logical relationship of potential faults into an intuitive tree structure, the fault tree can help engineers identify key failure modes and their propagation paths. However, with the increasing complexity of modern systems, traditional manual fault tree construction methods are gradually facing many challenges and are difficult to meet the needs of efficient and accurate analysis.

[0003] In complex systems, the interaction between functional modules is becoming increasingly close, and signal propagation paths often span multiple components and involve multi-level logical associations. Traditional fault tree analysis methods usually focus on the analysis of fault cases and capture fault modes through depth-first analysis of fault mechanisms. However, this method has certain blind spots, especially when faced with observable representations of faults, it is often difficult to fully capture all possible fault propagation chains in the system. In particular, when there are complex interactive relationships or multiple causal relationships between fault modes, traditional fault tree analysis methods often cannot effectively integrate these relationships, resulting in omissions or misjudgments of fault propagation paths, further limiting the depth and breadth of fault tree analysis. Therefore, existing methods show great limitations when dealing with multiple fault modes or cross-component and cross-level fault propagation, and cannot fully meet the fault analysis needs of complex systems.

[0004] In addition, existing fault trees are usually constructed based on fault cases. This analysis method based on individual fault conditions easily leads to the fault tree itself being lengthy and trivial. Since the fault tree is constructed with cases as the center, it lacks the ability to dynamically integrate. Especially in the cross-component fault propagation scenario, traditional methods often cannot flexibly integrate the failure modes of each component and their mutual influence into a comprehensive cross-component fault tree. This not only affects the efficiency of analysis, but also increases the difficulty of finding in use. In the case of a more complex fault tree, engineers often need to spend a lot of time looking for specific fault modes or related signals, which invisibly increases the difficulty of maintenance and troubleshooting. Therefore, how to simplify the fault tree structure and enable it to be dynamically adjusted according to actual maintenance and troubleshooting scenarios has become a key issue in improving the efficiency and operability of the fault tree.

[0005] Therefore, making full use of the knowledge of multi-signal flow graphs that evolve dynamically throughout the life cycle, enabling the construction of fault trees, and realizing the dynamic evolution and global construction of fault trees is an important task to achieve the upgrade of complex identification reliability assurance. The emergence of this method will provide a new technical path for the safety analysis of complex systems, significantly improve the analysis efficiency and the credibility of the results, and provide strong support for the modern reliability management system. Summary of the invention

[0006] In order to solve the problems of low construction efficiency, many logical blind spots, and difficulty in dynamic optimization in existing fault tree construction methods in complex systems, the present invention proposes a method for automatic construction of fault trees based on multi-signal flow graphs. This method makes full use of the characteristics of dynamic evolution of multi-signal flow graphs in the entire product life cycle, and realizes the automatic construction function from multi-signal flow graphs to fault trees through entity knowledge extraction, structure screening and tree layout generation, thereby meeting the needs of modern complex systems for rapid generation, automatic improvement and dynamic evolution of fault trees.

[0007] The method for automatically constructing a fault tree based on a multi-signal flow graph includes three main steps: entity extraction of a multi-signal flow graph based on ontology constraints, entity structure regularization based on flow graph connectivity, and automatic layout of fault tree data based on depth priority. The method for extracting a multi-signal flow graph entity based on ontology constraints establishes a dynamic correspondence with the fault tree node by parsing the type and attribute of the signal flow graph node, thereby realizing an accurate mapping of the signal flow graph entity to the fault tree node type; the method for regularizing the entity structure based on flow graph connectivity ensures the integrity and simplicity of the fault tree logical link by skipping irrelevant nodes, completing association relationships, and automatically inserting or nodes; the method for automatically layouting fault tree data based on depth priority utilizes the method of depth priority statistical branches and breadth iteration automatic layout to obtain global information and realize top-down hierarchical layout and connection between nodes, and finally completes the automatic drawing of the fault tree.

[0008] The characteristics of the present invention are:

[0009] (1) Through the ontology knowledge structure extraction method based on multi-signal flow graph, the association modeling between fault modes, signals and components is realized, which effectively makes up for the logical blind spots caused by manual inference in the traditional fault tree construction method;

[0010] (2) Through the automatic screening of fault propagation relationships in multi-signal flow graphs, redundant or invalid fault modes are eliminated to meet the efficiency and accuracy requirements of fault tree construction in industrial scenarios;

[0011] (3) By utilizing the dynamic characteristics of multi-signal flow graphs, the rapid generation and dynamic optimization of fault trees can be achieved through automated layout methods, supporting the evolution of fault trees throughout the product life cycle, thereby significantly improving the reliability analysis capabilities of complex systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 This is the overall flow chart of the automatic construction method of fault tree based on multi-signal flow graph.

[0013] Figure 2 Flowchart of the entity extraction method of multi-signal flow graph based on ontology constraints.

[0014] Figure 3 The flowchart of the entity structure regularization method based on flow graph connectivity.

[0015] Figure 4 Flowchart of the automatic layout method for fault tree data based on depth-first approach. DETAILED DESCRIPTION

[0016] A method for automatically constructing a fault tree based on a multi-signal flow graph provided by the present invention is described in detail below with reference to the accompanying drawings.

[0017] The present invention provides a method for automatically constructing a fault tree based on a multi-signal flow graph. Generally speaking, this method goes through three main stages: multi-signal flow graph entity extraction based on ontology constraints, entity structure regularization based on flow graph connectivity, and fault tree data automatic layout based on depth-first. According to the nodes and attribute information in the multi-signal flow graph, three entity types of "fault mode", "and", and "or" are extracted, a dynamic correspondence with the fault tree nodes is established, and accurate mapping of entities to fault tree nodes is achieved. Then, by skipping irrelevant nodes, completing associations, and automatically inserting "or" nodes, the integrity of the fault tree logical link is ensured. Finally, the acquired knowledge is used to obtain global information and realize top-down hierarchical layout and connections between nodes, so as to achieve the purpose of automatically drawing the fault tree. The overall flow chart of the method for automatically constructing a fault tree based on a multi-signal flow graph is as follows: Figure 1 shown.

[0018] 1. Multi-signal flow graph entity extraction based on ontology constraints

[0019] The multi-signal flow graph entity extraction step based on ontology constraints analyzes the types and attributes of signal flow graph nodes through two steps: the correspondence between multi-signal flow graph entities and fault tree structures and the graphical dynamic definition of fault tree entities, establishes a dynamic correspondence with fault tree nodes, and realizes accurate mapping of signal flow graph entities to fault tree node types. The flowchart of the multi-signal flow graph entity extraction method based on ontology constraints is as follows: Figure 2 shown.

[0020] 1.1. Correspondence between multi-signal flow graph entities and fault tree structures

[0021] There are mainly three types of entities in the multi-signal flow graph entity, namely "fault mode", "and" and "or", which correspond to the fault tree. Among them, "fault mode" corresponds to the fault propagation node in the fault tree, and "and" and "or" represent two conditional judgment nodes in the fault tree.

[0022] The corresponding steps between the multi-signal flow graph entities and the fault tree structure are based on the corresponding relationship between the entities to which they belong. By traversing the entity list of the multi-signal flow graph, the attributes of the corresponding types of entities are modified to the three fault tree nodes of "event", "and" and "or" and then loaded into the new list, thereby preliminarily extracting the entity list of the fault tree and providing structured data that is easy to call for subsequent steps.

[0023] 1.2. Graphical dynamic definition of fault tree entities

[0024] The step of dynamically defining the graphics of the fault tree entity provides support for dynamic graphics definition through the underlying node classification processing logic, and uses the systematic mapping between types determined in the previous step to further perform dynamic definition according to the node category identifier, and dynamically specifies specific icons or styles for different types of nodes during the mapping process. The dynamic definition of node graphics requires user interaction using editable image drawing software such as VISIO, and the user provides basic entity definition information as shown in Table 1 in the style. The corresponding steps of the multi-signal flow graph entity and the fault tree structure automatically add entity picture elements, input and output anchor points and other dynamic information during the loading process according to the comparison table, thereby realizing the adaptive function of the user-defined style in the subsequent fault tree drawing process.

[0025] Table 1 Entity basic definition information comparison table

[0026]

[0027] 2. Regularization of entity structure based on flow graph connectivity

[0028] The entity structure regularization step based on flow graph connectivity skips irrelevant nodes, completes the association relationship and automatically inserts the OR node through the two steps of association mining between entities and automatic completion of the "OR" node, ensuring the integrity and simplicity of the fault tree logic link. The flow chart of the entity structure regularization method based on flow graph connectivity is as follows: Figure 3 shown.

[0029] 2.1. Mining associations between entities

[0030] The step of mining the association between entities establishes a direct connection relationship from the source node to the target node by traversing the edge set of the signal flow graph. During the construction process, this step automatically ignores nodes that do not participate in the main logic (such as switches, input or output nodes) and focuses on the core connections between entities. The process is implemented by dynamically maintaining an entity jump relationship mapping table, which is used to record all upstream connection nodes of each target node. By reading the information in the table and traversing all the edges, the edges directly connected to the node are screened out, and these edges are retained to retain the relationship between the fault modes that conform to the logical rules and eliminate other connections that do not conform to the fault tree construction specifications. The method ensures that redundant information is automatically skipped while completing the actual logical relationship between entities, making the topological structure of the fault tree clearer and more refined.

[0031] 2.2. Automatic completion of "or" nodes

[0032] The automatic completion step of the or node dynamically generates the corresponding or node by traversing the upstream connection relationship of each target node, and inserts it into the fault tree. For each target node that needs to complete the or node, the code generates a unique identifier as the ID of the new or node, and connects all upstream nodes to the target node through the or node. During the generation, the method stores the source nodes of the target node that are not the edges of the node in a mapping table, and establishes a mapping relationship with all its input nodes for each target node to ensure that the or node can correctly combine multiple inputs to form an effective fault propagation path. This not only ensures logical integrity, but also greatly reduces the workload of manually completing the or node.

[0033] 3. Automatic layout of fault tree data based on depth-first approach

[0034] The automatic layout of fault tree data based on depth priority acquires global information and realizes top-down hierarchical layout and node connections through two steps: branch statistics based on depth priority and automatic layout based on breadth iteration, and finally completes the automatic drawing of the fault tree. The flowchart of the automatic layout method of fault tree data based on depth priority is as follows: Figure 4 3.1. Branch statistics based on depth-first

[0035] The depth-first-based branch counting step counts the total number of leaf nodes pointed to by each node according to the depth-first idea, thereby providing global information of each node for the distribution of the positions of nodes at the same level in the subsequent automatic layout process. Specifically, the depth-first-based branch counting step first constructs a visited node record table. Then traverse to the leaf nodes depth by depth. During each visit, if the node is already in the visited node record table, continue to update the downstream branch of the node, and let the number of branches of each node on the path except the node be added to the number of branches of the node; otherwise, add one to the number of branches of each node on the path. After traversing all leaf nodes, the statistics of the number of branches of each node are completed.

[0036] 3.2. Automatic layout based on breadth iteration

[0037] The automatic layout based on breadth iteration establishes a natural hierarchical relationship from top to bottom by recursively and dynamically inserting new nodes and generating edge connections for them. Specifically, the automatic layout method based on breadth iteration first constructs a layout node record table, and then layouts the nodes one by one at each level. The vertical position will calculate the vertical position of the current level node according to the depth level and the spacing of each level; at the same level, the automatic layout method based on breadth iteration accumulates and sums the number of branches of all nodes at the depth in order, and layouts each node in the corresponding horizontal position according to the proportion of the cumulative number of branches of the node to the total number of branches at the depth; after obtaining the height position, each node is placed. If the node is already in the layout node record table, the node position is not updated; otherwise, the node position is updated and the node is loaded into the layout node record table. After completing the layout of each depth node one by one, a node image element instance is created according to the position of each node, and the connection between nodes is completed according to the connection relationship, so that the automatic drawing of the fault tree is completed.

Claims

1. A method for automatically constructing a fault tree based on a multi-signal flow graph, characterized in that: Through entity extraction, structural regularization and automatic layout of multi-signal flow graphs, a logically complete and dynamically optimizable fault tree is quickly generated, thereby improving the efficiency and reliability of fault analysis of complex systems; the automatic fault tree construction method comprises a multi-signal flow graph entity extraction step based on ontology constraints, an entity structure regularization step based on flow graph connectivity, and a depth-first automatic layout step for fault tree data.

2. The method for automatically constructing a fault tree according to claim 1, characterized in that: The multi-signal flow graph entity extraction step in the fault tree automatic construction method analyzes the node types and attributes in the multi-signal flow graph, establishes a dynamic correspondence between the signal flow graph entity and the fault tree node using the ontology constraint, and realizes the accurate mapping of the node type. This step includes classifying the entities in the multi-signal flow graph into "event", "and" or "or" nodes, and dynamically defining the graphic style and connection relationship, so as to provide structured data support for the fault tree construction and realize the user visualization customization function.

3. The method for automatically constructing a fault tree according to claim 1, characterized in that: The entity structure regularization step in the automatic fault tree construction method analyzes the association relationship of nodes in the multi-signal flow graph, removes irrelevant nodes and completes the logical links to ensure the structural integrity of the fault tree. This step includes skipping irrelevant connection nodes, dynamically generating "or" nodes to supplement undefined upstream relationships, and automatically maintaining the connection mapping table to optimize the fault propagation path, thereby improving the logical clarity and construction efficiency of the fault tree.

4. The method for automatically constructing a fault tree according to claim 1, characterized in that: The automatic layout step of the fault tree data in the automatic fault tree construction method combines the depth-first branch statistics and breadth-first iterative layout method to complete the hierarchical layout and node connection of the fault tree. By counting the number of branches of each node to provide global information, and dynamically adjusting the vertical and horizontal positions according to the depth of the node in the tree and the branch ratio, a natural distribution and automatic layout from top to bottom are achieved, thereby quickly generating a logically rigorous fault tree graphical structure.

Images