Predictive modeling of reliability engineering data using a new version of the flexible Weibull model

Area of Science:

• Statistics
• Reliability Engineering
• Probability Distributions

Background:

• Hybrid censoring sampling approaches are crucial for efficient data analysis.
• Existing models may lack flexibility for diverse data patterns.
• Novel distributions are needed to enhance parameter estimation accuracy.

Purpose of the Study:

• To introduce a novel five-parameter distribution, the generalized Weibull-modified Weibull model.
• To apply a combined-unified hybrid censoring sampling approach for parameter estimation.
• To assess the flexibility and applicability of the new distribution using real-world data.

Main Methods:

• Development of a five-parameter generalized Weibull-modified Weibull distribution.
• Application of the combined-unified hybrid censoring sampling approach.
• Parameter estimation using the maximum likelihood method via Monte Carlo simulation.
• Utilizing Copula models for marginal univariate distributions and developing asymptotic confidence intervals.

Main Results:

• The proposed distribution exhibits flexibility, accommodating symmetric or right-skewed data.
• The risk function can model increasing or decreasing trends.
• Simulation results validate the theoretical findings.
• The model effectively analyzes failure times of electronic components.

Conclusions:

• The generalized Weibull-modified Weibull model offers a flexible and powerful tool for statistical analysis.
• The combined-unified hybrid censoring approach enhances parameter estimation.
• The model's applicability is demonstrated through a practical electronic component failure data analysis.