Five and four-parameter lifetime distributions for bathtub-shaped failure rate using Perks mortality equation. (August 2016)

• Record Type: Journal Article
• Authors: Zeng, Hongtao; Lan, Tian; Chen, Qiming
• Is Part Of: Reliability engineering & system safety. Volume 152(2016:Aug.)
• Page Start: 307
• Page End: 315
• Publication Date: 2016-08
• Subjects: CDF cumulative distribution function -- HRF hazard-rate function -- MLE maximum likelihood estimation -- PDF probability density function -- Perks4 probability distribution derived from 4-parameter Perks hazard rate function -- Perks5 probability distribution derived from 5-parameter Perks hazard rate function; Bathtub shape -- Hazard rate function -- Perks -- Parameter estimation; Reliability (Engineering) -- Periodicals; System safety -- Periodicals; Industrial safety -- Periodicals; Fiabilité -- Périodiques; Sécurité des systèmes -- Périodiques; Sécurité du travail -- Périodiques; 620.00452
• DOI: 10.1016/j.ress.2016.03.014
• Languages: English
• ISSNs: 0951-8320

Abstract

Abstract: Two lifetime distributions derived from Perks׳ mortality rate function, one with 4 parameters and the other with 5 parameters, for the modeling of bathtub-shaped failure rates are proposed in this paper. The Perks׳ mortality/failure rate functions have historically been used for human life modeling in life insurance industry. Although this distribution is no longer used in insurance industry, considering many nice and some unique features of this function, it is necessary to revisit it and introduce it to the reliability community. The parameters of the distributions can control the scale, shape, and location of the PDF. The 4-parameter distribution can be used to model the bathtub failure rate. This model is applied to three previously published groups of lifetime data. This study shows they fit very well. The 5-parameter version can potentially model constant hazard rates of the later life of some devices in addition to the good features of 4-parameter version. Both the 4 and 5-parameter versions have closed form PDF and CDF. The truncated distributions of both versions stay within the original distribution family with simple parameter transformation. This nice feature is normally considered to be only possessed by the simple exponential distribution Highlights: Two new distributions are proposed to model bathtub shaped hazard rate. Derive the close-form PDF, CDF and feature of scalability and truncatability. Perks4 is verified to be good to model common bathtub shapes through comparison. Perks5 has the potential to model the stabilization of hazard rate at later life.