Feng, Qianmei2014-12-192014-12-19December 22012-12http://hdl.handle.net/10657/831Ship operators have collected condition monitoring data over 25 years, but the analysis of these vibration data for failure time prediction and maintenance management is sparse and critically needed within the marine industry. This thesis explores the degradation-based failure time estimation for electric motor pump units by using vibration analysis data provided from ships. The work is unique because the data is taken under non-homogeneous environmental conditions, varying vibration measuring hardware and onboard units’ characteristics, and small sample populations. The degradation/vibration data are stochastically modeled functions with a predetermined limit which is considered failure when exceeded. Two methods are applied. In the first approach, the times to failure for individual paths are modeled by a probability distribution. In the second two-stage approach, the path model is estimated by combining the individual model parameter estimates. Accordingly, ship operators can assess the remaining life of electric motor-pump combinations and make informed decisions concerning equipment shutdown and repair.application/pdfengThe author of this work is the copyright owner. UH Libraries and the Texas Digital Library have their permission to store and provide access to this work. Further transmission, reproduction, or presentation of this work is prohibited except with permission of the author(s).Condition monitoringDegradationFailure estimationReliabilityStochasticVibration monitoringIndustrial engineering2014-12-19Thesisborn digital