A Bayesian Nonparametric Test for Minimal Repair

作者：Li Li^*; Hanson Timothy; Damien Paul; Popova Elmira

来源：Technometrics, 2014, 56(3): 393-406.

DOI：10.1080/00401706.2013.842934

• Effective age
• Repairable system
• Tailfree process
• Truncated data

The article develops a Bayesian nonparametric reliability model for recurrent events where failure and truncated time-to-failure density shape is regressed on past maintenance decisions: perfect repair and minimal repair. By comparing the system interfailure lifetime distributions after minimal and perfect repair, we are able to test the minimal repair assumption of "good as old." Interfailure hazard functions after perfect and minimal repairs are estimated, shedding light on departures from minimal repair. The method is illustrated both on simulated data as well as failure time data from air-conditioning units at the South Texas Nuclear Operating Company near Bay City, Texas. This article has supplementary material online.

• 出版日期2014-8