There have been many recent developments in the application of data-basedudmethods to machine condition monitoring. A powerful methodology based on machine learningudhas emerged, where diagnostics are based on a two-step procedure: extraction of damage sensitiveudfeatures, followed by unsupervised learning (novelty detection) or supervised learningud(classification). The objective of the current pair of papers is simply to illustrate one state-of the-artudprocedure for each step, using synthetic data representative of reality in terms of sizeudand complexity. The first paper in the pair will deal with feature extraction.ududAlthough some papers have appeared in the recent past considering stochastic resonanceudas a means of amplifying damage information in signals, they have largely relied on ad hocudspecifications of the resonator used. In contrast, the current paper will adopt a principledudoptimisation-based approach to the resonator design. The paper will also show that a discreteuddynamical system can provide all the benefits of a continuous system, but also provide audconsiderable speed-up in terms of simulation time in order to facilitate the optimisationudapproach.
展开▼
