Abstract
Hazardous sources are inherent to the characteristics or operational needs of complex systems and represent the fundamental cause of accidents. This paper introduces a structured method for identifying these sources and proposing corresponding safety requirements. By ensuring hazardous sources remain under control, system safety risk can be effectively reduced to acceptable levels. This process provides a more accurate and precise framework for safety analysis and assessment.
Introduction: Shifting the Focus to Hazardous Sources
In complex engineering—particularly within aerospace and weaponry—safety is traditionally viewed as a property of individual components. However, modern systems are highly interconnected, and accidents often emerge from the interactions between components, human factors, and management decisions, even when no single part fails.
Traditional safety methods often struggle to account for these emergent properties. This research argues that the root cause of any accident is an uncontrolled "hazardous source," such as high-pressure tanks, explosive propellants, or radiation. Since these sources are often essential for system operation, the challenge lies in identifying them comprehensively and designing "constraints" or safety requirements to maintain control.
Methodology: Control-Loop Modeling and Hazard Identification
The proposed framework utilizes a multi-stage process for safety analysis:
- Hazardous Source Identification: The method employs a taxonomy of 15 hazard categories to audit the system. It also integrates historical failure data and general safety standards.
- Safety Requirement Proposition: Requirements are developed by addressing component failures, modeling node behavior (using IF-THEN logic or functional models), and verifying requirements against a standardized flaw table.
- Scenario-Based Assessment: The paper leverages Control Loop Modeling to map the path from a controller to an actuator, through the controlled process, and back via sensors.
By viewing an accident as a failure within this control loop (e.g., "sensor failure → inadequate feedback → inaccurate order → system crash"), analysts can pinpoint exactly where a "feedback flaw" might transform a hazard into a mishap.
Key Results and Risk Assessment
The primary outcome of this research is a structured workflow for system safety, supported by standardized classification tables.
Control Flaw Classification
A critical component of the verification process is the Control Flaw Table, which categorizes inadequacies into three main types: Inadequate Enforcement of Constraints, Inadequate Execution, and Inadequate/Missing Feedback.
Quantitative Risk Calculation
To determine if a design is "safe enough," the authors define system risk ($R$) as the sum of the severity ($C_i$) of each scenario multiplied by its probability ($P_i$): $$R = \sum_{i=1}^{n} C_i \times P_i$$
Conclusion and Future Directions
While powerful, methods like Event Tree Analysis (ETA) remain subjective, relying heavily on the analyst's experience. To mitigate this, future work should focus on simulation-based analysis. By using computer simulations to model hazardous sources and control loops, researchers aim to move toward a more automated and exhaustive discovery of potential accident scenarios, reducing reliance on human intuition.
Summary updated