Safety Goal Violation Analysis Using Fault Tree-Bayesian Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Fault Tree Analysis (FTA) methods in functional safety are limited by boolean algebra and do not account for environmental conditions, while Bayesian Networks provide more detailed statistical relationships but are not integrated into safety evaluations.
Innovation Solution
Integrate Bayesian Networks with Fault Tree Analysis to model environmental influences and electrical/electronic faults, using nodes to describe system behavior and conditional dependencies, providing a combined quantitative evaluation of safety goal violations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If Fault Tree Analysis is used for safety evaluation, then the analysis structure is simple and clear, but it cannot model environmental conditions and only uses boolean algebra
Solution Approach 1:
The patent merges Fault Tree Analysis (FTA) with Bayesian Networks (BN) into a hybrid model. The FTA provides the top-down deductive structure for safety goal violations, while the BN integrates environmental condition modeling with conditional probabilities. This combination allows the system to maintain the structural clarity of FTA while gaining the environmental adaptability and statistical rigor of BN, resolving the contradiction between simple structure and versatile modeling capability.
Solution Approach 2:
The hybrid FTA-BN model serves multiple functions simultaneously: it performs traditional fault tree deductive analysis for electrical/electronic failures while also modeling environmental conditions and their probabilistic influences. The Bayesian network component acts as a universal interface that can incorporate various environmental factors (weather, terrain, lighting) and their complex interdependencies, making the safety analysis system adaptable to diverse operating conditions without requiring separate analysis frameworks.
2Measurement precision
If Bayesian Networks are used to model environmental conditions, then detailed statistical relationships are achieved, but integration with safety evaluation is lacking
Solution Approach 1:
The patent uses the hybrid FTA-BN model as an intermediary framework that connects Bayesian Network environmental modeling with traditional safety evaluation. The BN component captures detailed statistical relationships between environmental conditions and system performance, while the FTA component provides the formal safety evaluation structure. The interface between BN and FTA allows probabilistic outputs from the Bayesian network to feed into the fault tree analysis, enabling seamless integration of environmental statistics with safety goal violation assessment.
3Ease of operation
If only electrical and electronic faults are considered, then functional safety analysis is straightforward, but environmental influences are excluded
Solution Approach 1:
The patent segments the safety analysis into distinct components: traditional electrical/electronic fault analysis through FTA, and environmental condition analysis through BN. This segmentation allows each component to handle its specific domain effectively - FTA manages the straightforward fault propagation logic while BN handles complex environmental statistics. The segmented approach maintains analytical simplicity for fault analysis while systematically incorporating environmental factors that would otherwise be excluded.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Method for determining a probability of a violation of safety goals of a technical system (21), comprising: - providing a violated safety goal, - deriving a Fault Tree (30) top-down from the violated safety goal to an event corresponding to a component (22) of the technical system (21) causing the violated safety goal, - checking whether the event includes an external influence on the component's performance and/or an electrical and/or electronic fault of the component (22), - in case of one or more electrical and/or electronic faults of a component (22) outputting a propagated probability of respective components or events, - in case of an environmental influence on a component's performance, defining an interface to a Bayesian network (31, 38), - defining and causally connecting nodes of environmental conditions in the Bayesian network (31, 38), - providing quantification input on conditional dependencies into the Bayesian network (31, 38), - providing, by the Bayesian network (31, 38), a quantification on the environmental condition influence on the component's behavior to the interface or to the Fault Tree, - outputting the probability of a violation of a safety goal caused by environmental conditions and/or electrical and/or electronic faults in the technical system (21).