Statistical analysis of linear degradation and failure time data with multiple failure modes

Area of Science:

• Engineering
• Reliability Engineering
• Statistical Modeling

Background:

• Complex systems often exhibit multiple failure modes.
• System degradation significantly influences failure times and intensities.
• Existing models may not fully capture the interplay between degradation and traumatic failures.

Purpose of the Study:

• To develop and analyze models for linear degradation and failure time with multiple failure modes.
• To incorporate the dependence of traumatic failure intensities on degradation levels.
• To propose estimators for cumulative intensities and reliability characteristics, including prediction of residual reliability.

Main Methods:

• Development of non-parametric, semiparametric, and parametric estimation methods.
• Formulation of theorems on the simultaneous asymptotic distribution of degradation and traumatic event intensity functions.
• Analysis of asymptotic properties for unconditional and residual reliability estimators.

Main Results:

• Proposed estimators for cumulative traumatic event intensities and reliability characteristics.
• Methods for predicting residual reliability based on current degradation levels.
• Theoretical results on the asymptotic behavior of degradation and failure intensity functions.

Conclusions:

• The proposed models and estimators provide a robust framework for analyzing systems with multiple failure modes and degradation.
• The study offers valuable tools for reliability assessment and prediction in complex engineering systems.
• Empirical analysis using real tire data demonstrates the practical applicability of the developed methods.