Reliability Growth Planning for Multi-Phase Missions Considering Epistemic Uncertainty

Abstract

During the reliability growth testing (RGT) phase of complex equipment, scientific planning is crucial for ensuring the rational allocation of resources and the achievement of acceptance objectives. In practical engineering, existing planning models are largely based on the assumption that systems experience continuous, uniform stress. These models often overlook the damage heterogeneity induced by phase-to-phase mission switching and typically assume that failure modes can be completely eliminated. Moreover, they fail to account for the epistemic uncertainty inherent in expert assessments of maintenance effectiveness. Although such idealized assumptions simplify computation, they systematically bias planning outcomes toward optimism. To address this issue, a robust planning framework tailored for multi-phase missions is introduced into reliability growth planning. The proposed methodology establishes a mission-driven effective time mapping mechanism and integrates the Crow-Extended concept with fuzzy set theory to formulate a heterogeneous growth evolution model that quantifies residual risks arising from imperfect repair. Furthermore, a robust optimization strategy based on risk tolerance is designed to circumvent pitfalls associated with phased stress oscillations. A case study demonstrates the significant applicability of this approach to practical problems. A core advantage of the method lies in its ability to elucidate the physical laws governing reliability evolution under the dual constraints of phased missions and imperfect maintenance, thereby providing a scientific decision-making basis for the robust development of complex mechanical equipment.