Lista zgłoszonych artykułów

Multiresolution analysis of vibration signals acquired from locomotive Diesel engine for classification of engine states basing on signal statistical parameters Piotr Boguś1,2, Mateusz Cieszyński1, Jerzy Merkisz3 1Department of Physics and Biophysics, Medical University of Gdańsk, Poland, piotr.bogus@gumed.edu.pl, +48 604-126-032 2Rail Vehicles Institute „TABOR”, Poznań, Poland 3Poznań University of Technology, Faculty of Working Machines and Transportations, Institute of Internal Combustion Engines and Transport, Poznań, Poland Abstract. The paper presents a method of classification of locomotive Diesel engine states basing on vibration signals taken from an engine body and using chosen statistical parameters calculated for the original signal and it wavelet multiresolution components. The researches presented in the paper concern estimation of an engine states before and after a general repair. The target application of the presented researches is an on-line diagnostic system which can complement standard OBD systems. To this purpose the applied methods should not base on complex analysis of some spectral, time-frequency or scalogram plots but rather on choosing single diagnostic parameters which are suitable for the fast on-line diagnostic. The results have showed the significant difference in distinguishing of engine work before and after a general repair using some chosen statistical parameters applied to multiresolution components of vibration signals. Keywords: vibration signals, Diesel engine diagnostic, multiresolution analysis, classification.