Updating Time-to-Failure Distributions Based on Field Observations and Sensor Data

摘要

Enterprise level logistics and prognostics and health management (PHM) modeling efforts use reliability focused failure distributions to characterize the probability of failure over the lifetime of a component. This research characterized the Sandia National Laboratories developed combined lifecycle (CMBL) distribution and explored methods for updating this distribution as systems age and new failure data becomes available. The initial results obtained in applying a Bayesian sequential updating methodology to the CMBL distribution shows promise. This research also resulted in the development of a closed-form full life cycle (CFLC) distribution similar to the CMBL distribution but with slightly different, yet commonly recognized, input parameters. Further research is warranted to provide additional theoretical validation of the distributions, complete the updating methods for the CMBL distribution, evaluate a Bayesian updating methodology for the CFLC distribution, and determine which updating methods would be most appropriate for enterprise level logistics and PHM modeling.