It is rare to find field data tracking for components over decades to compare and validate if their end-of-life reliability predictions at various time points during product life were accurate. This compressor reliability study was motivated by the need to understand the status quo reliability of a fielded compressor population made of distinct brands following a series of field failure events. Field reliability was first studied at the time when the compressor fleet was in service for 8 years. This age of fleet also happened to be its approximate mid-life. The expected design life as part of the reliability requirement provided to compressor suppliers was 15 years. After evaluating the 8 years of reliability data for current reliability status, 15th year reliability was predicted by extrapolation. The motivation behind this 15th year reliability prediction was to understand if the end-of-life compressor reliability goal is still achievable. Then, after passing another 7 years from the first analysis, the actual 15-year design life was reached. Finally, end-of-life reliability analysis of the same fleet of compressors was performed using the 15 years of service life data. This analysis was then compared with the 15th year predictions performed 7 years ago using the 8-year field failure data set. The reliability prediction at 15th year by extrapolation using 8-years of data fell close to the actual 15th year compressor reliability. Five distinct brands of compressors were selected to present in this paper. In addition to the added 7 years, the field population also had increased with more field failures since the field population and number of failures are functions of time. This paper discusses the data acquisition sources, time to first failures data filtering and field reliability analysis methodology. As the fielded population grew, shapes of failure data on Weibull plots visibly changed from the eighth-year analysis while the reliability perditions fell in the close ranges.
