Dynamics, Durability of Instruments and Equipment
DOI: https://dx.doi.org/10.24866/2227-6858/2019-2-3
Zhirabok A., Filatov A.
ALEXEY ZHIRABOK, Doctor of Engineering Sciences, Professor
of Department of Automation and Control, School of Engineering, e-mail: zhirabok@mail.ru
Far Eastern Federal University
8 Sukhanova St., Vladivostok, Russia, 690091
ARTUR FILATOV, Post Graduate Student in Institute of Marine Technology Problem, FEB RAS, e-mail: gryphon1@bk.ru
5a Sukhanova St., Vladivostok, Russia, 690091
Improvement of nonparametric method of diagnosis in mechatronic systems
Annotation: The problem of fault diagnosis in mechatronic systems and different devices, described by nonlinear dynamic models, by non-parametric method is considered. The peculiarity of such systems and devices is that their mathematical models contain such non-smooth nonlinearities as backlash, saturation, and Coulomb friction reflecting real physical processes in these systems. The feature of the non-parametric methods is that during diagnostic process the values of some parameters (in linear case – all parameters) may be unknown, but deviations from their values are detected by diagnostic tools. The logic-dynamic approach, which is used to implement this method in the present paper, is developed for diagnosis of nonlinear systems which may contain non-smooth nonlinearities. A solution of the problem based on this method is realized by two steps. On the first step, a linear solution is found which then is supplemented by nonlinear term. The feature of the logic-dynamic approach is that it uses the methods of linear algebra only that allows avoiding the use of methods of differential geometry and complex analytical mathematical packages. In this paper, we consider the methods to implement the non-parametric method allowing decreasing the computational complexity and obtaining calculating expressions for nonlinear systems. The theoretical results obtained during this studt are confirmed by illustrative examples.
Keywords: mechatronic systems, non-smooth nonlinearities, faults, diagnosis, non-parametric methods.
See the reference in English at the end of the article
Download article in PDF format