Utilizing unified progressive hybrid censored data in parametric inference under accelerated life tests

Area of Science:

• Reliability Engineering
• Statistical Modeling
• Accelerated Life Testing

Background:

• Progressive-stress accelerated life testing (PSALT) is crucial for product longevity assessment.
• Censored lifetime data is common in PSALT due to equipment and cost constraints.
• Existing models may not fully address the complexities of censored data in PSALT.

Purpose of the Study:

• To introduce a novel PSALT model incorporating Type-II unified progressive hybrid censoring.
• To model product lifespans using the truncated Cauchy power exponential (TCPE) distribution.
• To develop and evaluate methods for estimating TCPE parameters and acceleration factors under censoring.

Main Methods:

• Utilizing Type-II unified progressive hybrid censoring for data analysis.
• Applying the truncated Cauchy power exponential (TCPE) distribution for lifespan modeling.
• Employing maximum likelihood and Bayesian estimation techniques (Markov Chain Monte Carlo).

Main Results:

• Development of estimation methods for TCPE parameters and acceleration factors.
• Assessment of Bayesian estimates using symmetric and asymmetric loss functions.
• Evaluation of point and interval estimators through simulation and real-data analysis.

Conclusions:

• The proposed PSALT model effectively handles censored data with TCPE distributed lifespans.
• Both maximum likelihood and Bayesian methods provide viable estimation strategies.
• The study validates the model's efficacy using simulation and real-world application.