Diagnosing a Strong-Fault Model by Conflict and Consistency

Area of Science:

• * Fault diagnosis
• * Artificial intelligence
• * Systems engineering

Background:

• * Traditional fault diagnosis methods struggle with complex systems requiring strong-fault models.
• * Strong-fault models exhibit non-monotonic behavior, complicating diagnosis.
• * Existing methods often rely on conflict identification, which can be inefficient.

Purpose of the Study:

• * To develop an efficient method for diagnosing strong-fault models.
• * To propose a novel Logic-based Truth Maintenance System (LTMS) for this purpose.
• * To enhance diagnostic efficiency using conflict and consistency-based search strategies.

Main Methods:

• * Encoding strong-fault models into Boolean variables and Conjunctive Normal Form (CNF).
• * Employing a novel LTMS to reason over CNF, identifying minimal conflicts and maximal consistencies.
• * Implementing two search approaches based on conflict and consistency for efficient diagnosis.

Main Results:

• * The proposed LTMS effectively diagnoses strong-fault models.
• * The conflict and consistency-based search approaches improve diagnostic efficiency.
• * Theoretical analysis confirms the completeness, coverage, correctness, and complexity of the methods.
• * Demonstrated superior performance compared to existing methods (best first, conflict-directed A* search) on a spacecraft heat control unit.

Conclusions:

• * The novel LTMS provides an efficient and effective solution for strong-fault model diagnosis.
• * The proposed search strategies significantly enhance diagnostic performance.
• * The method shows practical applicability and superiority in real-world complex systems.