Development and validation of machinery health monitoring, diagnostic and prognostic methodology and tools

To aid in developing and evaluating effective and reliable diagnostic and prognostic algorithms, a damage dynamic models based simulation methodology was developed. In addition, shaft life models based on Miner's law and Walker equation were developed and integrated with dynamic models to simulate the life progression of the shaft, and to develop and evaluate the prognostic algorithms.;In this thesis, a number of effective data mining based prognostic algorithms to predict the remaining useful life (RUL) for drive shafts were presented. The data used in developing and evaluating these data mining based prognostic algorithms were generated using the developed damage dynamic simulation tool.;Meanwhile, an effective regime recognition algorithm to help monitor the usage of aircraft was also provided instead of relying on the pilot's ability to estimate a particular flight profile. As an extension to regime recognition to improve safety and/or reduce maintenance by predicting if the aircraft will be flown in a damaging way, concept of regime prediction was first time proposed and an effective regime prediction approach was developed in this research. Potentially, this extension could be used to alert if the power required is greater than the power available for some maneuvers such as heavy lift.