A novel dynamic belief rule base method for aerospace equipment fault diagnosis

Abstract Fault diagnosis in aerospace equipment is critical for ensuring flight safety and operational reliability. Existing methods often struggle with dynamic fault propagation and uncertainty under sparse data conditions. This study proposes a dynamic fault tree analysis-enhanced belief rule base (DFTA-BRB) model for interpretable fault diagnosis in aerospace equipment. The framework integrates DFTA to formalize fault propagation mechanisms and employs a BRB with evidential reasoning (ER) for uncertainty modeling under sparse data. Key innovations include: (1) DFTA-based fault interaction modeling, (2) ER-based diagnostic inference, and (3) parameter optimization via the projected covariance matrix adaptation evolution strategy. Case studies demonstrate that DFTA-BRB significantly improves diagnostic interpretability and reliability under sparse fault data and dynamic failure propagation conditions. The approach effectively bridges dynamic fault mechanism modeling with uncertainty-aware reasoning, providing a systematic and reliable solution for spacecraft fault diagnosis.