Diagnosis and Prognosis of Machinery Health based on Advanced Intelligent Computations

The data from machinery health monitoring contains high noise components and low information content. The research is concentrated on developing more advanced methods to analysis the data for more accurate diagnosis and prognosis of machinery health. Rolling bearings are the most common components used in different machines and the data from them are representative in terms of wide frequency bands, short impulses and random noise components. The method development is based on bearing systems at the beginning. Vibration signal is employed as the data sources for the analysis. Advanced intelligent computations which include non-linear time-series, various evolutionary algorithms, adaptive pattern algorithms, various neural networks and their ensembles, non-linear system based data conditioning will be applied to the data and their diagnosis performance will be investigated based on different degrees and types of faults from bearings. This research will produce a set of tools for accurate diagnosis of machines based on the advanced the intelligent methods.