Accelerated degradation models for failure based on geometric Brownian motion and gamma processes

Area of Science:

• Reliability Engineering
• Stochastic Processes
• Materials Science

Background:

• Traditional life testing often relies solely on failure data, potentially overlooking valuable degradation information.
• Accelerated life testing (ALT) is crucial for predicting product reliability under stress but requires robust modeling.

Purpose of the Study:

• To develop novel accelerated life test (ALT) models integrating both failure and degradation data.
• To provide a generalized cumulative damage approach for system lifetime inference.
• To introduce new ALT models based on geometric Brownian motion and gamma processes.

Main Methods:

• Utilized a generalized cumulative damage approach with stochastic degradation processes.
• Developed new accelerated degradation models incorporating an accelerated test variable.
• Employed geometric Brownian motion and gamma processes for failure modeling.
• Approximated proposed models using accelerated versions of Birnbaum-Saunders and inverse Gaussian distributions.

Main Results:

• Presented new ALT models that effectively incorporate both failure and degradation measures.
• Demonstrated that the proposed models closely approximate accelerated Birnbaum-Saunders and inverse Gaussian distributions.
• Showcased the application of these models through real-world data from carbon-film resistors and fatigue crack growth.

Conclusions:

• The developed ALT models offer a more comprehensive approach to system lifetime prediction by integrating diverse data types.
• The findings facilitate improved parametric inference and model selection in reliability analysis.
• The study provides practical tools for analyzing component reliability using accelerated testing methodologies.